beenull/ladybird
1
0
Fork 0
mirror of https://github.com/LadybirdBrowser/ladybird.git synced 2024-11-22 23:50:19 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 6
ladybird/Userland/Libraries/LibRegex/RegexByteCode.h

871 lines
33 KiB
C
Raw Normal View History

LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
/*
* Copyright (c) 2020, Emanuel Sprung <emanuel.sprung@gmail.com>
*
Everything: Move to SPDX license identifiers in all files. SPDX License Identifiers are a more compact / standardized way of representing file license information. See: https://spdx.dev/resources/use/#identifiers This was done with the `ambr` search and replace tool. ambr --no-parent-ignore --key-from-file --rep-from-file key.txt rep.txt *
2021-04-22 08:24:48 +00:00
* SPDX-License-Identifier: BSD-2-Clause
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
*/
#pragma once
LibRegex: Add a basic optimization pass This currently tries to convert forking loops to atomic groups, and unify the left side of alternations.
2021-09-12 13:00:27 +00:00
#include "RegexBytecodeStreamOptimizer.h"
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
#include "RegexMatch.h"
#include "RegexOptions.h"
LibRegex: Avoid excessive Vector copy when compiling regexps Previously we would've copied the bytecode instead of moving the chunks around, use the fancy new DisjointChunks<T> abstraction to make that happen automagically. This decreases vector copies and uses of memmove() by nearly 10x :^)
2021-09-13 19:15:22 +00:00
#include <AK/DisjointChunks.h>
LibRegex: Implement an ECMA262-compatible parser This also adds support for lookarounds and individually-negated comparisons. The only unimplemented part of the parser spec is the unicode stuff.
2020-11-27 16:03:53 +00:00
#include <AK/Format.h>
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
#include <AK/Forward.h>
#include <AK/HashMap.h>
#include <AK/NonnullOwnPtr.h>
#include <AK/OwnPtr.h>
#include <AK/Traits.h>
Userland: Replace VERIFY(is<T>) with verify_cast<T> Instead of doing a VERIFY(is<T>(x)) and *then* casting it to T, we can just do the cast right away with verify_cast<T>. :^)
2021-06-24 19:13:09 +00:00
#include <AK/TypeCasts.h>
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
#include <AK/Types.h>
#include <AK/Vector.h>
LibRegex+LibUnicode: Begin implementing Unicode property escapes This supports some binary property matching. It does not support any properties not yet parsed by LibUnicode, nor does it support value matching (such as Script_Extensions=Latin).
2021-07-29 18:18:51 +00:00
#include <LibUnicode/Forward.h>
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
namespace regex {
LibRegex: Change bytecode value type to a 64-bit value To allow storing unicode ranges compactly; this is not utilised at the moment, but changing this later would've been significantly more difficult. Also fixes a few debug logs.
2020-11-19 15:15:36 +00:00
using ByteCodeValueType = u64;
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
#define ENUMERATE_OPCODES \
__ENUMERATE_OPCODE(Compare) \
__ENUMERATE_OPCODE(Jump) \
LibRegex: Make infinite repetitions short-circuit on empty matches This makes (addmittedly weird) patterns like `(a*)*` work correctly without going into an infinite fork loop.
2021-09-05 22:08:47 +00:00
__ENUMERATE_OPCODE(JumpNonEmpty) \
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
__ENUMERATE_OPCODE(ForkJump) \
__ENUMERATE_OPCODE(ForkStay) \
LibRegex: Add a basic optimization pass This currently tries to convert forking loops to atomic groups, and unify the left side of alternations.
2021-09-12 13:00:27 +00:00
__ENUMERATE_OPCODE(ForkReplaceJump) \
__ENUMERATE_OPCODE(ForkReplaceStay) \
LibRegex: Implement an ECMA262-compatible parser This also adds support for lookarounds and individually-negated comparisons. The only unimplemented part of the parser spec is the unicode stuff.
2020-11-27 16:03:53 +00:00
__ENUMERATE_OPCODE(FailForks) \
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
__ENUMERATE_OPCODE(SaveLeftCaptureGroup) \
__ENUMERATE_OPCODE(SaveRightCaptureGroup) \
__ENUMERATE_OPCODE(SaveRightNamedCaptureGroup) \
__ENUMERATE_OPCODE(CheckBegin) \
__ENUMERATE_OPCODE(CheckEnd) \
LibRegex: Implement an ECMA262-compatible parser This also adds support for lookarounds and individually-negated comparisons. The only unimplemented part of the parser spec is the unicode stuff.
2020-11-27 16:03:53 +00:00
__ENUMERATE_OPCODE(CheckBoundary) \
__ENUMERATE_OPCODE(Save) \
__ENUMERATE_OPCODE(Restore) \
__ENUMERATE_OPCODE(GoBack) \
LibRegex: Clear previous capture group contents in ECMA262 mode ECMA262 requires that the capture groups only contain the values from the last iteration, e.g. `((c)(a)?(b))` should _not_ contain 'a' in the second capture group when matching "cabcb".
2021-07-23 15:07:18 +00:00
__ENUMERATE_OPCODE(ClearCaptureGroup) \
LibRegex: Implement and use a REPEAT operation for bytecode repetition Currently, when we need to repeat an instruction N times, we simply add that instruction N times in a for-loop. This doesn't scale well with extremely large values of N, and ECMA-262 allows up to N = 2^53 - 1. Instead, add a new REPEAT bytecode operation to defer this loop from the parser to the runtime executor. This allows the parser to complete sans any loops (for this instruction), and allows the executor to bail early if the repeated bytecode fails. Note: The templated ByteCode methods are to allow the Posix parsers to continue using u32 because they are limited to N = 2^20.
2021-08-12 15:02:46 +00:00
__ENUMERATE_OPCODE(Repeat) \
LibRegex: Correctly handle failing in the middle of explicit repeats - Make sure that all the Repeat ops are reset (otherwise the operation would not be correct when going over the Repeat op a second time) - Make sure that all matches that are allowed to fail are backed by a fork, otherwise the last failing fork would not have anywhere to return to. Fixes #9707.
2021-08-31 17:57:08 +00:00
__ENUMERATE_OPCODE(ResetRepeat) \
LibRegex: Make infinite repetitions short-circuit on empty matches This makes (addmittedly weird) patterns like `(a*)*` work correctly without going into an infinite fork loop.
2021-09-05 22:08:47 +00:00
__ENUMERATE_OPCODE(Checkpoint) \
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
__ENUMERATE_OPCODE(Exit)
LibRegex: Implement an ECMA262-compatible parser This also adds support for lookarounds and individually-negated comparisons. The only unimplemented part of the parser spec is the unicode stuff.
2020-11-27 16:03:53 +00:00
// clang-format off
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
enum class OpCodeId : ByteCodeValueType {
#define __ENUMERATE_OPCODE(x) x,
ENUMERATE_OPCODES
#undef __ENUMERATE_OPCODE
LibRegex: Implement an ECMA262-compatible parser This also adds support for lookarounds and individually-negated comparisons. The only unimplemented part of the parser spec is the unicode stuff.
2020-11-27 16:03:53 +00:00
First = Compare,
Last = Exit,
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
};
LibRegex: Implement an ECMA262-compatible parser This also adds support for lookarounds and individually-negated comparisons. The only unimplemented part of the parser spec is the unicode stuff.
2020-11-27 16:03:53 +00:00
// clang-format on
LibRegex: Use a match table for character classes Generate a sorted, compressed series of ranges in a match table for character classes, and use a binary search to find the matches. This is about a 3-4x speedup for character class match performance. :^)
2021-10-03 15:31:25 +00:00
#define ENUMERATE_CHARACTER_COMPARE_TYPES \
__ENUMERATE_CHARACTER_COMPARE_TYPE(Undefined) \
__ENUMERATE_CHARACTER_COMPARE_TYPE(Inverse) \
__ENUMERATE_CHARACTER_COMPARE_TYPE(TemporaryInverse) \
__ENUMERATE_CHARACTER_COMPARE_TYPE(AnyChar) \
__ENUMERATE_CHARACTER_COMPARE_TYPE(Char) \
__ENUMERATE_CHARACTER_COMPARE_TYPE(String) \
__ENUMERATE_CHARACTER_COMPARE_TYPE(CharClass) \
__ENUMERATE_CHARACTER_COMPARE_TYPE(CharRange) \
__ENUMERATE_CHARACTER_COMPARE_TYPE(Reference) \
__ENUMERATE_CHARACTER_COMPARE_TYPE(Property) \
__ENUMERATE_CHARACTER_COMPARE_TYPE(GeneralCategory) \
__ENUMERATE_CHARACTER_COMPARE_TYPE(Script) \
__ENUMERATE_CHARACTER_COMPARE_TYPE(ScriptExtension) \
__ENUMERATE_CHARACTER_COMPARE_TYPE(RangeExpressionDummy) \
__ENUMERATE_CHARACTER_COMPARE_TYPE(LookupTable)
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
enum class CharacterCompareType : ByteCodeValueType {
#define __ENUMERATE_CHARACTER_COMPARE_TYPE(x) x,
ENUMERATE_CHARACTER_COMPARE_TYPES
#undef __ENUMERATE_CHARACTER_COMPARE_TYPE
};
#define ENUMERATE_CHARACTER_CLASSES \
__ENUMERATE_CHARACTER_CLASS(Alnum) \
__ENUMERATE_CHARACTER_CLASS(Cntrl) \
__ENUMERATE_CHARACTER_CLASS(Lower) \
__ENUMERATE_CHARACTER_CLASS(Space) \
__ENUMERATE_CHARACTER_CLASS(Alpha) \
__ENUMERATE_CHARACTER_CLASS(Digit) \
__ENUMERATE_CHARACTER_CLASS(Print) \
__ENUMERATE_CHARACTER_CLASS(Upper) \
__ENUMERATE_CHARACTER_CLASS(Blank) \
__ENUMERATE_CHARACTER_CLASS(Graph) \
__ENUMERATE_CHARACTER_CLASS(Punct) \
LibRegex: Implement an ECMA262-compatible parser This also adds support for lookarounds and individually-negated comparisons. The only unimplemented part of the parser spec is the unicode stuff.
2020-11-27 16:03:53 +00:00
__ENUMERATE_CHARACTER_CLASS(Word) \
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
__ENUMERATE_CHARACTER_CLASS(Xdigit)
enum class CharClass : ByteCodeValueType {
#define __ENUMERATE_CHARACTER_CLASS(x) x,
ENUMERATE_CHARACTER_CLASSES
#undef __ENUMERATE_CHARACTER_CLASS
};
LibRegex: Implement an ECMA262-compatible parser This also adds support for lookarounds and individually-negated comparisons. The only unimplemented part of the parser spec is the unicode stuff.
2020-11-27 16:03:53 +00:00
#define ENUMERATE_BOUNDARY_CHECK_TYPES \
__ENUMERATE_BOUNDARY_CHECK_TYPE(Word) \
__ENUMERATE_BOUNDARY_CHECK_TYPE(NonWord)
enum class BoundaryCheckType : ByteCodeValueType {
#define __ENUMERATE_BOUNDARY_CHECK_TYPE(x) x,
ENUMERATE_BOUNDARY_CHECK_TYPES
#undef __ENUMERATE_BOUNDARY_CHECK_TYPE
};
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
struct CharRange {
LibRegex: Switch to east-const style
2021-07-23 15:55:14 +00:00
u32 const from;
u32 const to;
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
LibRegex: Change bytecode value type to a 64-bit value To allow storing unicode ranges compactly; this is not utilised at the moment, but changing this later would've been significantly more difficult. Also fixes a few debug logs.
2020-11-19 15:15:36 +00:00
CharRange(u64 value)
: from(value >> 32)
, to(value & 0xffffffff)
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
{
}
LibRegex: Change bytecode value type to a 64-bit value To allow storing unicode ranges compactly; this is not utilised at the moment, but changing this later would've been significantly more difficult. Also fixes a few debug logs.
2020-11-19 15:15:36 +00:00
CharRange(u32 from, u32 to)
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
: from(from)
, to(to)
{
}
LibRegex: Change bytecode value type to a 64-bit value To allow storing unicode ranges compactly; this is not utilised at the moment, but changing this later would've been significantly more difficult. Also fixes a few debug logs.
2020-11-19 15:15:36 +00:00
operator ByteCodeValueType() const { return ((u64)from << 32) | to; }
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
};
struct CompareTypeAndValuePair {
CharacterCompareType type;
ByteCodeValueType value;
};
class OpCode;
LibRegex: Avoid excessive Vector copy when compiling regexps Previously we would've copied the bytecode instead of moving the chunks around, use the fancy new DisjointChunks<T> abstraction to make that happen automagically. This decreases vector copies and uses of memmove() by nearly 10x :^)
2021-09-13 19:15:22 +00:00
class ByteCode : public DisjointChunks<ByteCodeValueType> {
using Base = DisjointChunks<ByteCodeValueType>;
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
public:
LibRegex: Avoid initialization checks in get_opcode_by_id()
2021-06-13 20:55:00 +00:00
ByteCode()
{
ensure_opcodes_initialized();
}
LibRegex: Switch to east-const style
2021-07-23 15:55:14 +00:00
ByteCode(ByteCode const&) = default;
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
virtual ~ByteCode() = default;
LibRegex: Give ByteCode a copy ctor and and a move assignment operator Previously all move assignments were actually copies. oops.
2021-04-10 04:00:48 +00:00
ByteCode& operator=(ByteCode&&) = default;
LibRegex: Avoid excessive Vector copy when compiling regexps Previously we would've copied the bytecode instead of moving the chunks around, use the fancy new DisjointChunks<T> abstraction to make that happen automagically. This decreases vector copies and uses of memmove() by nearly 10x :^)
2021-09-13 19:15:22 +00:00
ByteCode& operator=(Base&& value)
{
static_cast<Base&>(*this) = move(value);
return *this;
}
template<typename... Args>
void empend(Args&&... args)
{
if (is_empty())
Base::append({});
Base::last_chunk().empend(forward<Args>(args)...);
}
template<typename T>
void append(T&& value)
{
if (is_empty())
Base::append({});
Base::last_chunk().append(forward<T>(value));
}
template<typename T>
void prepend(T&& value)
{
if (is_empty())
return append(forward<T>(value));
Base::first_chunk().prepend(forward<T>(value));
}
void last_chunk() const = delete;
void first_chunk() const = delete;
LibRegex: Give ByteCode a copy ctor and and a move assignment operator Previously all move assignments were actually copies. oops.
2021-04-10 04:00:48 +00:00
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
void insert_bytecode_compare_values(Vector<CompareTypeAndValuePair>&& pairs)
{
LibRegex: Use a match table for character classes Generate a sorted, compressed series of ranges in a match table for character classes, and use a binary search to find the matches. This is about a 3-4x speedup for character class match performance. :^)
2021-10-03 15:31:25 +00:00
Optimizer::append_character_class(*this, move(pairs));
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
}
LibRegex: Implement an ECMA262-compatible parser This also adds support for lookarounds and individually-negated comparisons. The only unimplemented part of the parser spec is the unicode stuff.
2020-11-27 16:03:53 +00:00
void insert_bytecode_check_boundary(BoundaryCheckType type)
{
ByteCode bytecode;
bytecode.empend((ByteCodeValueType)OpCodeId::CheckBoundary);
bytecode.empend((ByteCodeValueType)type);
AK: Rename Vector::append(Vector) => Vector::extend(Vector) Let's make it a bit more clear when we're appending the elements from one vector to the end of another vector.
2021-06-12 11:24:45 +00:00
extend(move(bytecode));
LibRegex: Implement an ECMA262-compatible parser This also adds support for lookarounds and individually-negated comparisons. The only unimplemented part of the parser spec is the unicode stuff.
2020-11-27 16:03:53 +00:00
}
LibRegex: Clear previous capture group contents in ECMA262 mode ECMA262 requires that the capture groups only contain the values from the last iteration, e.g. `((c)(a)?(b))` should _not_ contain 'a' in the second capture group when matching "cabcb".
2021-07-23 15:07:18 +00:00
void insert_bytecode_clear_capture_group(size_t index)
{
empend(static_cast<ByteCodeValueType>(OpCodeId::ClearCaptureGroup));
empend(index);
}
LibRegex: Store 'String' matches inside the bytecode Also removes an unnecessary 'length' argument (StringView has a length!)
2020-12-06 13:32:03 +00:00
void insert_bytecode_compare_string(StringView view)
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
{
ByteCode bytecode;
bytecode.empend(static_cast<ByteCodeValueType>(OpCodeId::Compare));
AK: Use direct-list-initialization for Vector::empend() (#4564) clang trunk with -std=c++20 doesn't seem to properly look for an aggregate initializer here when the type being constructed is a simple aggregate (e.g. `struct Thing { int a; int b; };`). This template fails to compile in a usage added 12/16/2020 in `AK/Trie.h`. Both forms of initialization are supposed to call the aggregate-initializers but direct-list-initialization delegating to aggregate initializers is a new addition in c++20 that might not be implemented yet.
2020-12-27 22:06:37 +00:00
bytecode.empend(static_cast<u64>(1)); // number of arguments
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
ByteCode arguments;
arguments.empend(static_cast<ByteCodeValueType>(CharacterCompareType::String));
LibRegex: Store 'String' matches inside the bytecode Also removes an unnecessary 'length' argument (StringView has a length!)
2020-12-06 13:32:03 +00:00
arguments.insert_string(view);
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
bytecode.empend(arguments.size()); // size of arguments
AK: Rename Vector::append(Vector) => Vector::extend(Vector) Let's make it a bit more clear when we're appending the elements from one vector to the end of another vector.
2021-06-12 11:24:45 +00:00
bytecode.extend(move(arguments));
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
AK: Rename Vector::append(Vector) => Vector::extend(Vector) Let's make it a bit more clear when we're appending the elements from one vector to the end of another vector.
2021-06-12 11:24:45 +00:00
extend(move(bytecode));
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
}
void insert_bytecode_group_capture_left(size_t capture_groups_count)
{
empend(static_cast<ByteCodeValueType>(OpCodeId::SaveLeftCaptureGroup));
empend(capture_groups_count);
}
void insert_bytecode_group_capture_right(size_t capture_groups_count)
{
empend(static_cast<ByteCodeValueType>(OpCodeId::SaveRightCaptureGroup));
empend(capture_groups_count);
}
Everywhere: Pass AK::StringView by value
2021-11-10 23:55:02 +00:00
void insert_bytecode_group_capture_right(size_t capture_groups_count, StringView name)
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
{
empend(static_cast<ByteCodeValueType>(OpCodeId::SaveRightNamedCaptureGroup));
empend(reinterpret_cast<ByteCodeValueType>(name.characters_without_null_termination()));
empend(name.length());
LibRegex+LibJS: Combine named and unnamed capture groups in MatchState Combining these into one list helps reduce the size of MatchState, and as a result, reduces the amount of memory consumed during execution of very large regex matches. Doing this also allows us to remove a few regex byte code instructions: ClearNamedCaptureGroup, SaveLeftNamedCaptureGroup, and NamedReference. Named groups now behave the same as unnamed groups for these operations. Note that SaveRightNamedCaptureGroup still exists to cache the matched group name. This also removes the recursion level from the MatchState, as it can exist as a local variable in Matcher::execute instead.
2021-08-14 20:28:54 +00:00
empend(capture_groups_count);
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
}
LibRegex: Implement an ECMA262-compatible parser This also adds support for lookarounds and individually-negated comparisons. The only unimplemented part of the parser spec is the unicode stuff.
2020-11-27 16:03:53 +00:00
enum class LookAroundType {
LookAhead,
LookBehind,
NegatedLookAhead,
NegatedLookBehind,
};
void insert_bytecode_lookaround(ByteCode&& lookaround_body, LookAroundType type, size_t match_length = 0)
{
// FIXME: The save stack will grow infinitely with repeated failures
// as we do not discard that on failure (we don't necessarily know how many to pop with the current architecture).
switch (type) {
case LookAroundType::LookAhead: {
// SAVE
// REGEXP BODY
// RESTORE
empend((ByteCodeValueType)OpCodeId::Save);
AK: Rename Vector::append(Vector) => Vector::extend(Vector) Let's make it a bit more clear when we're appending the elements from one vector to the end of another vector.
2021-06-12 11:24:45 +00:00
extend(move(lookaround_body));
LibRegex: Implement an ECMA262-compatible parser This also adds support for lookarounds and individually-negated comparisons. The only unimplemented part of the parser spec is the unicode stuff.
2020-11-27 16:03:53 +00:00
empend((ByteCodeValueType)OpCodeId::Restore);
return;
}
case LookAroundType::NegatedLookAhead: {
// JUMP _A
// LABEL _L
// REGEXP BODY
LibRegex: Make FailForks fail all forks up to the last save point This makes negative lookarounds with more than one fork behave correctly. Fixes #11350.
2021-12-25 02:05:09 +00:00
// FAIL
LibRegex: Implement an ECMA262-compatible parser This also adds support for lookarounds and individually-negated comparisons. The only unimplemented part of the parser spec is the unicode stuff.
2020-11-27 16:03:53 +00:00
// LABEL _A
// SAVE
// FORKJUMP _L
// RESTORE
auto body_length = lookaround_body.size();
empend((ByteCodeValueType)OpCodeId::Jump);
LibRegex: Make FailForks fail all forks up to the last save point This makes negative lookarounds with more than one fork behave correctly. Fixes #11350.
2021-12-25 02:05:09 +00:00
empend((ByteCodeValueType)body_length + 1); // JUMP to label _A
AK: Rename Vector::append(Vector) => Vector::extend(Vector) Let's make it a bit more clear when we're appending the elements from one vector to the end of another vector.
2021-06-12 11:24:45 +00:00
extend(move(lookaround_body));
LibRegex: Implement an ECMA262-compatible parser This also adds support for lookarounds and individually-negated comparisons. The only unimplemented part of the parser spec is the unicode stuff.
2020-11-27 16:03:53 +00:00
empend((ByteCodeValueType)OpCodeId::FailForks);
empend((ByteCodeValueType)OpCodeId::Save);
empend((ByteCodeValueType)OpCodeId::ForkJump);
LibRegex: Make FailForks fail all forks up to the last save point This makes negative lookarounds with more than one fork behave correctly. Fixes #11350.
2021-12-25 02:05:09 +00:00
empend((ByteCodeValueType) - (body_length + 4)); // JUMP to label _L
LibRegex: Implement an ECMA262-compatible parser This also adds support for lookarounds and individually-negated comparisons. The only unimplemented part of the parser spec is the unicode stuff.
2020-11-27 16:03:53 +00:00
empend((ByteCodeValueType)OpCodeId::Restore);
return;
}
case LookAroundType::LookBehind:
// SAVE
// GOBACK match_length(BODY)
// REGEXP BODY
// RESTORE
empend((ByteCodeValueType)OpCodeId::Save);
empend((ByteCodeValueType)OpCodeId::GoBack);
empend((ByteCodeValueType)match_length);
AK: Rename Vector::append(Vector) => Vector::extend(Vector) Let's make it a bit more clear when we're appending the elements from one vector to the end of another vector.
2021-06-12 11:24:45 +00:00
extend(move(lookaround_body));
LibRegex: Implement an ECMA262-compatible parser This also adds support for lookarounds and individually-negated comparisons. The only unimplemented part of the parser spec is the unicode stuff.
2020-11-27 16:03:53 +00:00
empend((ByteCodeValueType)OpCodeId::Restore);
return;
case LookAroundType::NegatedLookBehind: {
// JUMP _A
// LABEL _L
// GOBACK match_length(BODY)
// REGEXP BODY
LibRegex: Make FailForks fail all forks up to the last save point This makes negative lookarounds with more than one fork behave correctly. Fixes #11350.
2021-12-25 02:05:09 +00:00
// FAIL
LibRegex: Implement an ECMA262-compatible parser This also adds support for lookarounds and individually-negated comparisons. The only unimplemented part of the parser spec is the unicode stuff.
2020-11-27 16:03:53 +00:00
// LABEL _A
// SAVE
// FORKJUMP _L
// RESTORE
auto body_length = lookaround_body.size();
empend((ByteCodeValueType)OpCodeId::Jump);
LibRegex: Make FailForks fail all forks up to the last save point This makes negative lookarounds with more than one fork behave correctly. Fixes #11350.
2021-12-25 02:05:09 +00:00
empend((ByteCodeValueType)body_length + 3); // JUMP to label _A
LibRegex: Implement an ECMA262-compatible parser This also adds support for lookarounds and individually-negated comparisons. The only unimplemented part of the parser spec is the unicode stuff.
2020-11-27 16:03:53 +00:00
empend((ByteCodeValueType)OpCodeId::GoBack);
empend((ByteCodeValueType)match_length);
AK: Rename Vector::append(Vector) => Vector::extend(Vector) Let's make it a bit more clear when we're appending the elements from one vector to the end of another vector.
2021-06-12 11:24:45 +00:00
extend(move(lookaround_body));
LibRegex: Implement an ECMA262-compatible parser This also adds support for lookarounds and individually-negated comparisons. The only unimplemented part of the parser spec is the unicode stuff.
2020-11-27 16:03:53 +00:00
empend((ByteCodeValueType)OpCodeId::FailForks);
empend((ByteCodeValueType)OpCodeId::Save);
empend((ByteCodeValueType)OpCodeId::ForkJump);
LibRegex: Make FailForks fail all forks up to the last save point This makes negative lookarounds with more than one fork behave correctly. Fixes #11350.
2021-12-25 02:05:09 +00:00
empend((ByteCodeValueType) - (body_length + 6)); // JUMP to label _L
LibRegex: Implement an ECMA262-compatible parser This also adds support for lookarounds and individually-negated comparisons. The only unimplemented part of the parser spec is the unicode stuff.
2020-11-27 16:03:53 +00:00
empend((ByteCodeValueType)OpCodeId::Restore);
return;
}
}
Everywhere: Rename ASSERT => VERIFY (...and ASSERT_NOT_REACHED => VERIFY_NOT_REACHED) Since all of these checks are done in release builds as well, let's rename them to VERIFY to prevent confusion, as everyone is used to assertions being compiled out in release. We can introduce a new ASSERT macro that is specifically for debug checks, but I'm doing this wholesale conversion first since we've accumulated thousands of these already, and it's not immediately obvious which ones are suitable for ASSERT.
2021-02-23 19:42:32 +00:00
VERIFY_NOT_REACHED();
LibRegex: Implement an ECMA262-compatible parser This also adds support for lookarounds and individually-negated comparisons. The only unimplemented part of the parser spec is the unicode stuff.
2020-11-27 16:03:53 +00:00
}
LibRegex: Avoid excessive Vector copy when compiling regexps Previously we would've copied the bytecode instead of moving the chunks around, use the fancy new DisjointChunks<T> abstraction to make that happen automagically. This decreases vector copies and uses of memmove() by nearly 10x :^)
2021-09-13 19:15:22 +00:00
void insert_bytecode_alternation(ByteCode&& left, ByteCode&& right)
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
{
LibRegex: Fix greedy/reluctant modifiers in PosixExtendedParser Also fixes the issue with assertions causing early termination when they fail.
2020-11-19 15:27:39 +00:00
// FORKJUMP _ALT
LibRegex: Test alternatives in the expected order That is, first try to match the left side of the alternation, and then the right side. Fixes part of #6042.
2021-04-01 14:07:51 +00:00
// REGEXP ALT2
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
// JUMP _END
// LABEL _ALT
LibRegex: Test alternatives in the expected order That is, first try to match the left side of the alternation, and then the right side. Fixes part of #6042.
2021-04-01 14:07:51 +00:00
// REGEXP ALT1
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
// LABEL _END
LibRegex: Add a basic optimization pass This currently tries to convert forking loops to atomic groups, and unify the left side of alternations.
2021-09-12 13:00:27 +00:00
// Optimisation: Eliminate extra work by unifying common pre-and-postfix exprs.
LibRegex: Avoid excessive Vector copy when compiling regexps Previously we would've copied the bytecode instead of moving the chunks around, use the fancy new DisjointChunks<T> abstraction to make that happen automagically. This decreases vector copies and uses of memmove() by nearly 10x :^)
2021-09-13 19:15:22 +00:00
Optimizer::append_alternation(*this, move(left), move(right));
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
}
LibRegex: Implement and use a REPEAT operation for bytecode repetition Currently, when we need to repeat an instruction N times, we simply add that instruction N times in a for-loop. This doesn't scale well with extremely large values of N, and ECMA-262 allows up to N = 2^53 - 1. Instead, add a new REPEAT bytecode operation to defer this loop from the parser to the runtime executor. This allows the parser to complete sans any loops (for this instruction), and allows the executor to bail early if the repeated bytecode fails. Note: The templated ByteCode methods are to allow the Posix parsers to continue using u32 because they are limited to N = 2^20.
2021-08-12 15:02:46 +00:00
template<typename T>
LibRegex: Correctly handle failing in the middle of explicit repeats - Make sure that all the Repeat ops are reset (otherwise the operation would not be correct when going over the Repeat op a second time) - Make sure that all matches that are allowed to fail are backed by a fork, otherwise the last failing fork would not have anywhere to return to. Fixes #9707.
2021-08-31 17:57:08 +00:00
static void transform_bytecode_repetition_min_max(ByteCode& bytecode_to_repeat, T minimum, Optional<T> maximum, size_t min_repetition_mark_id, size_t max_repetition_mark_id, bool greedy = true) requires(IsIntegral<T>)
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
{
LibRegex: Transform 0,1 min/unbounded max repetitions to * or + The implementation of transform_bytecode_repetition_min_max expects at least min=1, so let's also optimise on our way out of a bug where 'x{0,}' would cause a crash.
2021-10-09 13:48:21 +00:00
if (!maximum.has_value()) {
if (minimum == 0)
return transform_bytecode_repetition_any(bytecode_to_repeat, greedy);
if (minimum == 1)
return transform_bytecode_repetition_min_one(bytecode_to_repeat, greedy);
}
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
ByteCode new_bytecode;
LibRegex: Correctly handle failing in the middle of explicit repeats - Make sure that all the Repeat ops are reset (otherwise the operation would not be correct when going over the Repeat op a second time) - Make sure that all matches that are allowed to fail are backed by a fork, otherwise the last failing fork would not have anywhere to return to. Fixes #9707.
2021-08-31 17:57:08 +00:00
new_bytecode.insert_bytecode_repetition_n(bytecode_to_repeat, minimum, min_repetition_mark_id);
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
if (maximum.has_value()) {
LibRegex: Implement min/max repetition using the Repeat bytecode This makes repetitions with large max bounds work correctly. Also fixes an OOM issue found by OSS-Fuzz: https://oss-fuzz.com/testcase?key=4725721980338176
2021-08-30 20:25:35 +00:00
// (REPEAT REGEXP MIN)
LibRegex: Correctly handle failing in the middle of explicit repeats - Make sure that all the Repeat ops are reset (otherwise the operation would not be correct when going over the Repeat op a second time) - Make sure that all matches that are allowed to fail are backed by a fork, otherwise the last failing fork would not have anywhere to return to. Fixes #9707.
2021-08-31 17:57:08 +00:00
// LABEL _MAX_LOOP |
// FORK END |
// REGEXP |
// REPEAT _MAX_LOOP MAX-MIN | if max > min
// FORK END |
// REGEXP |
// LABEL END |
// RESET _MAX_LOOP |
LibRegex: Allow a '?' suffix for brace quantifiers This fixes another compat point in #6042.
2021-04-10 04:01:53 +00:00
auto jump_kind = static_cast<ByteCodeValueType>(greedy ? OpCodeId::ForkStay : OpCodeId::ForkJump);
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
if (maximum.value() > minimum) {
LibRegex: Allow a '?' suffix for brace quantifiers This fixes another compat point in #6042.
2021-04-10 04:01:53 +00:00
new_bytecode.empend(jump_kind);
LibRegex: Implement min/max repetition using the Repeat bytecode This makes repetitions with large max bounds work correctly. Also fixes an OOM issue found by OSS-Fuzz: https://oss-fuzz.com/testcase?key=4725721980338176
2021-08-30 20:25:35 +00:00
new_bytecode.empend((ByteCodeValueType)0); // Placeholder for the jump target.
auto pre_loop_fork_jump_index = new_bytecode.size();
LibRegex: Correctly handle failing in the middle of explicit repeats - Make sure that all the Repeat ops are reset (otherwise the operation would not be correct when going over the Repeat op a second time) - Make sure that all matches that are allowed to fail are backed by a fork, otherwise the last failing fork would not have anywhere to return to. Fixes #9707.
2021-08-31 17:57:08 +00:00
new_bytecode.extend(bytecode_to_repeat);
LibRegex: Implement min/max repetition using the Repeat bytecode This makes repetitions with large max bounds work correctly. Also fixes an OOM issue found by OSS-Fuzz: https://oss-fuzz.com/testcase?key=4725721980338176
2021-08-30 20:25:35 +00:00
auto repetitions = maximum.value() - minimum;
LibRegex: Correctly handle failing in the middle of explicit repeats - Make sure that all the Repeat ops are reset (otherwise the operation would not be correct when going over the Repeat op a second time) - Make sure that all matches that are allowed to fail are backed by a fork, otherwise the last failing fork would not have anywhere to return to. Fixes #9707.
2021-08-31 17:57:08 +00:00
auto fork_jump_address = new_bytecode.size();
LibRegex: Implement min/max repetition using the Repeat bytecode This makes repetitions with large max bounds work correctly. Also fixes an OOM issue found by OSS-Fuzz: https://oss-fuzz.com/testcase?key=4725721980338176
2021-08-30 20:25:35 +00:00
if (repetitions > 1) {
new_bytecode.empend((ByteCodeValueType)OpCodeId::Repeat);
new_bytecode.empend(bytecode_to_repeat.size() + 2);
new_bytecode.empend(static_cast<ByteCodeValueType>(repetitions - 1));
LibRegex: Correctly handle failing in the middle of explicit repeats - Make sure that all the Repeat ops are reset (otherwise the operation would not be correct when going over the Repeat op a second time) - Make sure that all matches that are allowed to fail are backed by a fork, otherwise the last failing fork would not have anywhere to return to. Fixes #9707.
2021-08-31 17:57:08 +00:00
new_bytecode.empend(max_repetition_mark_id);
new_bytecode.empend(jump_kind);
new_bytecode.empend((ByteCodeValueType)0); // Placeholder for the jump target.
auto post_loop_fork_jump_index = new_bytecode.size();
new_bytecode.extend(bytecode_to_repeat);
fork_jump_address = new_bytecode.size();
new_bytecode[post_loop_fork_jump_index - 1] = (ByteCodeValueType)(fork_jump_address - post_loop_fork_jump_index);
new_bytecode.empend((ByteCodeValueType)OpCodeId::ResetRepeat);
new_bytecode.empend((ByteCodeValueType)max_repetition_mark_id);
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
}
LibRegex: Correctly handle failing in the middle of explicit repeats - Make sure that all the Repeat ops are reset (otherwise the operation would not be correct when going over the Repeat op a second time) - Make sure that all matches that are allowed to fail are backed by a fork, otherwise the last failing fork would not have anywhere to return to. Fixes #9707.
2021-08-31 17:57:08 +00:00
new_bytecode[pre_loop_fork_jump_index - 1] = (ByteCodeValueType)(fork_jump_address - pre_loop_fork_jump_index);
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
}
} else {
LibRegex: Make infinite repetitions short-circuit on empty matches This makes (addmittedly weird) patterns like `(a*)*` work correctly without going into an infinite fork loop.
2021-09-05 22:08:47 +00:00
// no maximum value set, repeat finding if possible:
// (REPEAT REGEXP MIN)
// LABEL _START
// CHECKPOINT _C
// REGEXP
// JUMP_NONEMPTY _C _START FORK
// Note: This is only safe because REPEAT will leave one iteration outside (see repetition_n)
new_bytecode.insert(new_bytecode.size() - bytecode_to_repeat.size(), (ByteCodeValueType)OpCodeId::Checkpoint);
LibRegex: Allow a '?' suffix for brace quantifiers This fixes another compat point in #6042.
2021-04-10 04:01:53 +00:00
auto jump_kind = static_cast<ByteCodeValueType>(greedy ? OpCodeId::ForkJump : OpCodeId::ForkStay);
LibRegex: Make infinite repetitions short-circuit on empty matches This makes (addmittedly weird) patterns like `(a*)*` work correctly without going into an infinite fork loop.
2021-09-05 22:08:47 +00:00
new_bytecode.empend((ByteCodeValueType)OpCodeId::JumpNonEmpty);
new_bytecode.empend(-bytecode_to_repeat.size() - 4 - 1); // Jump to the last iteration
new_bytecode.empend(-bytecode_to_repeat.size() - 4 - 1); // if _C is not empty.
LibRegex: Allow a '?' suffix for brace quantifiers This fixes another compat point in #6042.
2021-04-10 04:01:53 +00:00
new_bytecode.empend(jump_kind);
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
}
bytecode_to_repeat = move(new_bytecode);
}
LibRegex: Implement and use a REPEAT operation for bytecode repetition Currently, when we need to repeat an instruction N times, we simply add that instruction N times in a for-loop. This doesn't scale well with extremely large values of N, and ECMA-262 allows up to N = 2^53 - 1. Instead, add a new REPEAT bytecode operation to defer this loop from the parser to the runtime executor. This allows the parser to complete sans any loops (for this instruction), and allows the executor to bail early if the repeated bytecode fails. Note: The templated ByteCode methods are to allow the Posix parsers to continue using u32 because they are limited to N = 2^20.
2021-08-12 15:02:46 +00:00
template<typename T>
void insert_bytecode_repetition_n(ByteCode& bytecode_to_repeat, T n, size_t repetition_mark_id) requires(IsIntegral<T>)
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
{
LibRegex: Implement and use a REPEAT operation for bytecode repetition Currently, when we need to repeat an instruction N times, we simply add that instruction N times in a for-loop. This doesn't scale well with extremely large values of N, and ECMA-262 allows up to N = 2^53 - 1. Instead, add a new REPEAT bytecode operation to defer this loop from the parser to the runtime executor. This allows the parser to complete sans any loops (for this instruction), and allows the executor to bail early if the repeated bytecode fails. Note: The templated ByteCode methods are to allow the Posix parsers to continue using u32 because they are limited to N = 2^20.
2021-08-12 15:02:46 +00:00
// LABEL _LOOP
// REGEXP
// REPEAT _LOOP N-1
// REGEXP
if (n == 0)
return;
// Note: this bytecode layout allows callers to repeat the last REGEXP instruction without the
// REPEAT instruction forcing another loop.
extend(bytecode_to_repeat);
if (n > 1) {
empend(static_cast<ByteCodeValueType>(OpCodeId::Repeat));
empend(bytecode_to_repeat.size());
empend(static_cast<ByteCodeValueType>(n - 1));
empend(repetition_mark_id);
AK: Rename Vector::append(Vector) => Vector::extend(Vector) Let's make it a bit more clear when we're appending the elements from one vector to the end of another vector.
2021-06-12 11:24:45 +00:00
extend(bytecode_to_repeat);
LibRegex: Implement and use a REPEAT operation for bytecode repetition Currently, when we need to repeat an instruction N times, we simply add that instruction N times in a for-loop. This doesn't scale well with extremely large values of N, and ECMA-262 allows up to N = 2^53 - 1. Instead, add a new REPEAT bytecode operation to defer this loop from the parser to the runtime executor. This allows the parser to complete sans any loops (for this instruction), and allows the executor to bail early if the repeated bytecode fails. Note: The templated ByteCode methods are to allow the Posix parsers to continue using u32 because they are limited to N = 2^20.
2021-08-12 15:02:46 +00:00
}
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
}
LibRegex: Make the bytecode transformation functions static They were pretty confusing when compared with other non-transforming functions.
2021-07-10 09:48:04 +00:00
static void transform_bytecode_repetition_min_one(ByteCode& bytecode_to_repeat, bool greedy)
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
{
// LABEL _START = -bytecode_to_repeat.size()
LibRegex: Make infinite repetitions short-circuit on empty matches This makes (addmittedly weird) patterns like `(a*)*` work correctly without going into an infinite fork loop.
2021-09-05 22:08:47 +00:00
// CHECKPOINT _C
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
// REGEXP
LibRegex: Make infinite repetitions short-circuit on empty matches This makes (addmittedly weird) patterns like `(a*)*` work correctly without going into an infinite fork loop.
2021-09-05 22:08:47 +00:00
// JUMP_NONEMPTY _C _START FORKSTAY (FORKJUMP -> Greedy)
bytecode_to_repeat.prepend((ByteCodeValueType)OpCodeId::Checkpoint);
bytecode_to_repeat.empend((ByteCodeValueType)OpCodeId::JumpNonEmpty);
bytecode_to_repeat.empend(-bytecode_to_repeat.size() - 3); // Jump to the _START label...
bytecode_to_repeat.empend(-bytecode_to_repeat.size() - 2); // ...if _C is not empty
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
if (greedy)
bytecode_to_repeat.empend(static_cast<ByteCodeValueType>(OpCodeId::ForkJump));
LibRegex: Fix greedy/reluctant modifiers in PosixExtendedParser Also fixes the issue with assertions causing early termination when they fail.
2020-11-19 15:27:39 +00:00
else
bytecode_to_repeat.empend(static_cast<ByteCodeValueType>(OpCodeId::ForkStay));
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
}
LibRegex: Make the bytecode transformation functions static They were pretty confusing when compared with other non-transforming functions.
2021-07-10 09:48:04 +00:00
static void transform_bytecode_repetition_any(ByteCode& bytecode_to_repeat, bool greedy)
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
{
// LABEL _START
LibRegex: Fix greedy/reluctant modifiers in PosixExtendedParser Also fixes the issue with assertions causing early termination when they fail.
2020-11-19 15:27:39 +00:00
// FORKJUMP _END (FORKSTAY -> Greedy)
LibRegex: Make infinite repetitions short-circuit on empty matches This makes (addmittedly weird) patterns like `(a*)*` work correctly without going into an infinite fork loop.
2021-09-05 22:08:47 +00:00
// CHECKPOINT _C
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
// REGEXP
LibRegex: Make infinite repetitions short-circuit on empty matches This makes (addmittedly weird) patterns like `(a*)*` work correctly without going into an infinite fork loop.
2021-09-05 22:08:47 +00:00
// JUMP_NONEMPTY _C _START JUMP
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
// LABEL _END
// LABEL _START = m_bytes.size();
ByteCode bytecode;
if (greedy)
bytecode.empend(static_cast<ByteCodeValueType>(OpCodeId::ForkStay));
LibRegex: Fix greedy/reluctant modifiers in PosixExtendedParser Also fixes the issue with assertions causing early termination when they fail.
2020-11-19 15:27:39 +00:00
else
bytecode.empend(static_cast<ByteCodeValueType>(OpCodeId::ForkJump));
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
LibRegex: Make infinite repetitions short-circuit on empty matches This makes (addmittedly weird) patterns like `(a*)*` work correctly without going into an infinite fork loop.
2021-09-05 22:08:47 +00:00
bytecode.empend(bytecode_to_repeat.size() + 1 + 4); // Jump to the _END label
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
LibRegex: Make infinite repetitions short-circuit on empty matches This makes (addmittedly weird) patterns like `(a*)*` work correctly without going into an infinite fork loop.
2021-09-05 22:08:47 +00:00
auto c_label = bytecode.size();
bytecode.empend(static_cast<ByteCodeValueType>(OpCodeId::Checkpoint));
bytecode.extend(bytecode_to_repeat);
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
LibRegex: Make infinite repetitions short-circuit on empty matches This makes (addmittedly weird) patterns like `(a*)*` work correctly without going into an infinite fork loop.
2021-09-05 22:08:47 +00:00
bytecode.empend(static_cast<ByteCodeValueType>(OpCodeId::JumpNonEmpty));
bytecode.empend(-bytecode.size() - 3); // Jump(...) to the _START label...
bytecode.empend(c_label - bytecode.size() - 2); // ...only if _C passes.
bytecode.empend((ByteCodeValueType)OpCodeId::Jump);
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
// LABEL _END = bytecode.size()
bytecode_to_repeat = move(bytecode);
}
LibRegex: Make the bytecode transformation functions static They were pretty confusing when compared with other non-transforming functions.
2021-07-10 09:48:04 +00:00
static void transform_bytecode_repetition_zero_or_one(ByteCode& bytecode_to_repeat, bool greedy)
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
{
LibRegex: Fix greedy/reluctant modifiers in PosixExtendedParser Also fixes the issue with assertions causing early termination when they fail.
2020-11-19 15:27:39 +00:00
// FORKJUMP _END (FORKSTAY -> Greedy)
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
// REGEXP
// LABEL _END
ByteCode bytecode;
if (greedy)
bytecode.empend(static_cast<ByteCodeValueType>(OpCodeId::ForkStay));
LibRegex: Fix greedy/reluctant modifiers in PosixExtendedParser Also fixes the issue with assertions causing early termination when they fail.
2020-11-19 15:27:39 +00:00
else
bytecode.empend(static_cast<ByteCodeValueType>(OpCodeId::ForkJump));
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
bytecode.empend(bytecode_to_repeat.size()); // Jump to the _END label
LibRegex: Avoid excessive Vector copy when compiling regexps Previously we would've copied the bytecode instead of moving the chunks around, use the fancy new DisjointChunks<T> abstraction to make that happen automagically. This decreases vector copies and uses of memmove() by nearly 10x :^)
2021-09-13 19:15:22 +00:00
bytecode.extend(move(bytecode_to_repeat));
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
// LABEL _END = bytecode.size()
bytecode_to_repeat = move(bytecode);
}
LibRegex: Make get_opcode() return a reference Previously this would return a pointer which could be null if the requested opcode was invalid. This should never be the case though so let's VERIFY() that instead.
2021-06-13 20:18:54 +00:00
OpCode& get_opcode(MatchState& state) const;
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
private:
Everywhere: Pass AK::StringView by value
2021-11-10 23:55:02 +00:00
void insert_string(StringView view)
LibRegex: Store 'String' matches inside the bytecode Also removes an unnecessary 'length' argument (StringView has a length!)
2020-12-06 13:32:03 +00:00
{
empend((ByteCodeValueType)view.length());
for (size_t i = 0; i < view.length(); ++i)
empend((ByteCodeValueType)view[i]);
}
LibRegex: Avoid initialization checks in get_opcode_by_id()
2021-06-13 20:55:00 +00:00
void ensure_opcodes_initialized();
LibRegex: Make get_opcode() return a reference Previously this would return a pointer which could be null if the requested opcode was invalid. This should never be the case though so let's VERIFY() that instead.
2021-06-13 20:18:54 +00:00
ALWAYS_INLINE OpCode& get_opcode_by_id(OpCodeId id) const;
LibRegex: Use a plain array to store opcodes Using a hash map is unnecessary because the number of opcodes and their IDs never change.
2021-06-13 20:10:25 +00:00
static OwnPtr<OpCode> s_opcodes[(size_t)OpCodeId::Last + 1];
static bool s_opcodes_initialized;
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
};
#define ENUMERATE_EXECUTION_RESULTS \
__ENUMERATE_EXECUTION_RESULT(Continue) \
__ENUMERATE_EXECUTION_RESULT(Fork_PrioHigh) \
__ENUMERATE_EXECUTION_RESULT(Fork_PrioLow) \
__ENUMERATE_EXECUTION_RESULT(Failed) \
__ENUMERATE_EXECUTION_RESULT(Failed_ExecuteLowPrioForks) \
__ENUMERATE_EXECUTION_RESULT(Succeeded)
enum class ExecutionResult : u8 {
#define __ENUMERATE_EXECUTION_RESULT(x) x,
ENUMERATE_EXECUTION_RESULTS
#undef __ENUMERATE_EXECUTION_RESULT
};
LibRegex: Switch to east-const style
2021-07-23 15:55:14 +00:00
char const* execution_result_name(ExecutionResult result);
char const* opcode_id_name(OpCodeId opcode_id);
char const* boundary_check_type_name(BoundaryCheckType);
char const* character_compare_type_name(CharacterCompareType result);
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
class OpCode {
public:
LibRegex: Avoid initialization checks in get_opcode_by_id()
2021-06-13 20:55:00 +00:00
OpCode() = default;
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
virtual ~OpCode() = default;
virtual OpCodeId opcode_id() const = 0;
virtual size_t size() const = 0;
LibRegex: Remove (mostly) unused regex::MatchOutput This struct holds a counter for the number of executed operations, and vectors for matches, captures groups, and named capture groups. Each of the vectors is unused. Remove the struct and just keep a separate counter for the executed operations.
2021-08-14 21:00:58 +00:00
virtual ExecutionResult execute(MatchInput const& input, MatchState& state) const = 0;
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
ALWAYS_INLINE ByteCodeValueType argument(size_t offset) const
{
Everywhere: Rename ASSERT => VERIFY (...and ASSERT_NOT_REACHED => VERIFY_NOT_REACHED) Since all of these checks are done in release builds as well, let's rename them to VERIFY to prevent confusion, as everyone is used to assertions being compiled out in release. We can introduce a new ASSERT macro that is specifically for debug checks, but I'm doing this wholesale conversion first since we've accumulated thousands of these already, and it's not immediately obvious which ones are suitable for ASSERT.
2021-02-23 19:42:32 +00:00
VERIFY(state().instruction_position + offset <= m_bytecode->size());
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
return m_bytecode->at(state().instruction_position + 1 + offset);
}
LibRegex: Switch to east-const style
2021-07-23 15:55:14 +00:00
ALWAYS_INLINE char const* name() const;
LibRegex: Remove some meaningless/useless const-qualifiers Also replace String creation from `""` with `String::empty()`
2021-12-21 17:06:24 +00:00
static char const* name(OpCodeId);
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
LibRegex: Remove return value for setters
2021-06-13 20:13:49 +00:00
ALWAYS_INLINE void set_state(MatchState& state) { m_state = &state; }
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
LibRegex: Remove return value for setters
2021-06-13 20:13:49 +00:00
ALWAYS_INLINE void set_bytecode(ByteCode& bytecode) { m_bytecode = &bytecode; }
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
LibRegex: Switch to east-const style
2021-07-23 15:55:14 +00:00
ALWAYS_INLINE MatchState const& state() const
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
{
LibRegex: Use a plain pointer for OpCode::m_state
2021-06-13 20:59:36 +00:00
VERIFY(m_state);
return *m_state;
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
}
LibRegex: Remove some meaningless/useless const-qualifiers Also replace String creation from `""` with `String::empty()`
2021-12-21 17:06:24 +00:00
String to_string() const
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
{
Everywhere: Prefer using {:#x} over 0x{:x} We have a dedicated format specifier which adds the "0x" prefix, so let's use that instead of adding it manually.
2021-07-21 23:30:24 +00:00
return String::formatted("[{:#02X}] {}", (int)opcode_id(), name(opcode_id()));
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
}
LibRegex: Remove some meaningless/useless const-qualifiers Also replace String creation from `""` with `String::empty()`
2021-12-21 17:06:24 +00:00
virtual String arguments_string() const = 0;
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
LibRegex: Switch to east-const style
2021-07-23 15:55:14 +00:00
ALWAYS_INLINE ByteCode const& bytecode() const { return *m_bytecode; }
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
protected:
LibRegex: Avoid initialization checks in get_opcode_by_id()
2021-06-13 20:55:00 +00:00
ByteCode* m_bytecode { nullptr };
LibRegex: Use a plain pointer for OpCode::m_state
2021-06-13 20:59:36 +00:00
MatchState* m_state { nullptr };
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
};
class OpCode_Exit final : public OpCode {
public:
LibRegex: Remove (mostly) unused regex::MatchOutput This struct holds a counter for the number of executed operations, and vectors for matches, captures groups, and named capture groups. Each of the vectors is unused. Remove the struct and just keep a separate counter for the executed operations.
2021-08-14 21:00:58 +00:00
ExecutionResult execute(MatchInput const& input, MatchState& state) const override;
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
ALWAYS_INLINE OpCodeId opcode_id() const override { return OpCodeId::Exit; }
ALWAYS_INLINE size_t size() const override { return 1; }
LibRegex: Remove some meaningless/useless const-qualifiers Also replace String creation from `""` with `String::empty()`
2021-12-21 17:06:24 +00:00
String arguments_string() const override { return String::empty(); }
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
};
LibRegex: Implement an ECMA262-compatible parser This also adds support for lookarounds and individually-negated comparisons. The only unimplemented part of the parser spec is the unicode stuff.
2020-11-27 16:03:53 +00:00
class OpCode_FailForks final : public OpCode {
public:
LibRegex: Remove (mostly) unused regex::MatchOutput This struct holds a counter for the number of executed operations, and vectors for matches, captures groups, and named capture groups. Each of the vectors is unused. Remove the struct and just keep a separate counter for the executed operations.
2021-08-14 21:00:58 +00:00
ExecutionResult execute(MatchInput const& input, MatchState& state) const override;
LibRegex: Implement an ECMA262-compatible parser This also adds support for lookarounds and individually-negated comparisons. The only unimplemented part of the parser spec is the unicode stuff.
2020-11-27 16:03:53 +00:00
ALWAYS_INLINE OpCodeId opcode_id() const override { return OpCodeId::FailForks; }
LibRegex: Make FailForks fail all forks up to the last save point This makes negative lookarounds with more than one fork behave correctly. Fixes #11350.
2021-12-25 02:05:09 +00:00
ALWAYS_INLINE size_t size() const override { return 1; }
String arguments_string() const override { return String::empty(); }
LibRegex: Implement an ECMA262-compatible parser This also adds support for lookarounds and individually-negated comparisons. The only unimplemented part of the parser spec is the unicode stuff.
2020-11-27 16:03:53 +00:00
};
class OpCode_Save final : public OpCode {
public:
LibRegex: Remove (mostly) unused regex::MatchOutput This struct holds a counter for the number of executed operations, and vectors for matches, captures groups, and named capture groups. Each of the vectors is unused. Remove the struct and just keep a separate counter for the executed operations.
2021-08-14 21:00:58 +00:00
ExecutionResult execute(MatchInput const& input, MatchState& state) const override;
LibRegex: Implement an ECMA262-compatible parser This also adds support for lookarounds and individually-negated comparisons. The only unimplemented part of the parser spec is the unicode stuff.
2020-11-27 16:03:53 +00:00
ALWAYS_INLINE OpCodeId opcode_id() const override { return OpCodeId::Save; }
ALWAYS_INLINE size_t size() const override { return 1; }
LibRegex: Remove some meaningless/useless const-qualifiers Also replace String creation from `""` with `String::empty()`
2021-12-21 17:06:24 +00:00
String arguments_string() const override { return String::empty(); }
LibRegex: Implement an ECMA262-compatible parser This also adds support for lookarounds and individually-negated comparisons. The only unimplemented part of the parser spec is the unicode stuff.
2020-11-27 16:03:53 +00:00
};
class OpCode_Restore final : public OpCode {
public:
LibRegex: Remove (mostly) unused regex::MatchOutput This struct holds a counter for the number of executed operations, and vectors for matches, captures groups, and named capture groups. Each of the vectors is unused. Remove the struct and just keep a separate counter for the executed operations.
2021-08-14 21:00:58 +00:00
ExecutionResult execute(MatchInput const& input, MatchState& state) const override;
LibRegex: Implement an ECMA262-compatible parser This also adds support for lookarounds and individually-negated comparisons. The only unimplemented part of the parser spec is the unicode stuff.
2020-11-27 16:03:53 +00:00
ALWAYS_INLINE OpCodeId opcode_id() const override { return OpCodeId::Restore; }
ALWAYS_INLINE size_t size() const override { return 1; }
LibRegex: Remove some meaningless/useless const-qualifiers Also replace String creation from `""` with `String::empty()`
2021-12-21 17:06:24 +00:00
String arguments_string() const override { return String::empty(); }
LibRegex: Implement an ECMA262-compatible parser This also adds support for lookarounds and individually-negated comparisons. The only unimplemented part of the parser spec is the unicode stuff.
2020-11-27 16:03:53 +00:00
};
class OpCode_GoBack final : public OpCode {
public:
LibRegex: Remove (mostly) unused regex::MatchOutput This struct holds a counter for the number of executed operations, and vectors for matches, captures groups, and named capture groups. Each of the vectors is unused. Remove the struct and just keep a separate counter for the executed operations.
2021-08-14 21:00:58 +00:00
ExecutionResult execute(MatchInput const& input, MatchState& state) const override;
LibRegex: Implement an ECMA262-compatible parser This also adds support for lookarounds and individually-negated comparisons. The only unimplemented part of the parser spec is the unicode stuff.
2020-11-27 16:03:53 +00:00
ALWAYS_INLINE OpCodeId opcode_id() const override { return OpCodeId::GoBack; }
ALWAYS_INLINE size_t size() const override { return 2; }
ALWAYS_INLINE size_t count() const { return argument(0); }
LibRegex: Remove some meaningless/useless const-qualifiers Also replace String creation from `""` with `String::empty()`
2021-12-21 17:06:24 +00:00
String arguments_string() const override { return String::formatted("count={}", count()); }
LibRegex: Implement an ECMA262-compatible parser This also adds support for lookarounds and individually-negated comparisons. The only unimplemented part of the parser spec is the unicode stuff.
2020-11-27 16:03:53 +00:00
};
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
class OpCode_Jump final : public OpCode {
public:
LibRegex: Remove (mostly) unused regex::MatchOutput This struct holds a counter for the number of executed operations, and vectors for matches, captures groups, and named capture groups. Each of the vectors is unused. Remove the struct and just keep a separate counter for the executed operations.
2021-08-14 21:00:58 +00:00
ExecutionResult execute(MatchInput const& input, MatchState& state) const override;
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
ALWAYS_INLINE OpCodeId opcode_id() const override { return OpCodeId::Jump; }
ALWAYS_INLINE size_t size() const override { return 2; }
ALWAYS_INLINE ssize_t offset() const { return argument(0); }
LibRegex: Remove some meaningless/useless const-qualifiers Also replace String creation from `""` with `String::empty()`
2021-12-21 17:06:24 +00:00
String arguments_string() const override
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
{
LibRegex: Convert String::format() => String::formatted()
2021-04-21 21:02:27 +00:00
return String::formatted("offset={} [&{}]", offset(), state().instruction_position + size() + offset());
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
}
};
LibRegex: Add a basic optimization pass This currently tries to convert forking loops to atomic groups, and unify the left side of alternations.
2021-09-12 13:00:27 +00:00
class OpCode_ForkJump : public OpCode {
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
public:
LibRegex: Remove (mostly) unused regex::MatchOutput This struct holds a counter for the number of executed operations, and vectors for matches, captures groups, and named capture groups. Each of the vectors is unused. Remove the struct and just keep a separate counter for the executed operations.
2021-08-14 21:00:58 +00:00
ExecutionResult execute(MatchInput const& input, MatchState& state) const override;
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
ALWAYS_INLINE OpCodeId opcode_id() const override { return OpCodeId::ForkJump; }
ALWAYS_INLINE size_t size() const override { return 2; }
ALWAYS_INLINE ssize_t offset() const { return argument(0); }
LibRegex: Remove some meaningless/useless const-qualifiers Also replace String creation from `""` with `String::empty()`
2021-12-21 17:06:24 +00:00
String arguments_string() const override
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
{
LibRegex: Convert String::format() => String::formatted()
2021-04-21 21:02:27 +00:00
return String::formatted("offset={} [&{}], sp: {}", offset(), state().instruction_position + size() + offset(), state().string_position);
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
}
};
LibRegex: Add a basic optimization pass This currently tries to convert forking loops to atomic groups, and unify the left side of alternations.
2021-09-12 13:00:27 +00:00
class OpCode_ForkReplaceJump final : public OpCode_ForkJump {
public:
ExecutionResult execute(MatchInput const& input, MatchState& state) const override;
ALWAYS_INLINE OpCodeId opcode_id() const override { return OpCodeId::ForkReplaceJump; }
};
class OpCode_ForkStay : public OpCode {
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
public:
LibRegex: Remove (mostly) unused regex::MatchOutput This struct holds a counter for the number of executed operations, and vectors for matches, captures groups, and named capture groups. Each of the vectors is unused. Remove the struct and just keep a separate counter for the executed operations.
2021-08-14 21:00:58 +00:00
ExecutionResult execute(MatchInput const& input, MatchState& state) const override;
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
ALWAYS_INLINE OpCodeId opcode_id() const override { return OpCodeId::ForkStay; }
ALWAYS_INLINE size_t size() const override { return 2; }
ALWAYS_INLINE ssize_t offset() const { return argument(0); }
LibRegex: Remove some meaningless/useless const-qualifiers Also replace String creation from `""` with `String::empty()`
2021-12-21 17:06:24 +00:00
String arguments_string() const override
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
{
LibRegex: Convert String::format() => String::formatted()
2021-04-21 21:02:27 +00:00
return String::formatted("offset={} [&{}], sp: {}", offset(), state().instruction_position + size() + offset(), state().string_position);
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
}
};
LibRegex: Add a basic optimization pass This currently tries to convert forking loops to atomic groups, and unify the left side of alternations.
2021-09-12 13:00:27 +00:00
class OpCode_ForkReplaceStay final : public OpCode_ForkStay {
public:
ExecutionResult execute(MatchInput const& input, MatchState& state) const override;
ALWAYS_INLINE OpCodeId opcode_id() const override { return OpCodeId::ForkReplaceStay; }
};
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
class OpCode_CheckBegin final : public OpCode {
public:
LibRegex: Remove (mostly) unused regex::MatchOutput This struct holds a counter for the number of executed operations, and vectors for matches, captures groups, and named capture groups. Each of the vectors is unused. Remove the struct and just keep a separate counter for the executed operations.
2021-08-14 21:00:58 +00:00
ExecutionResult execute(MatchInput const& input, MatchState& state) const override;
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
ALWAYS_INLINE OpCodeId opcode_id() const override { return OpCodeId::CheckBegin; }
ALWAYS_INLINE size_t size() const override { return 1; }
LibRegex: Remove some meaningless/useless const-qualifiers Also replace String creation from `""` with `String::empty()`
2021-12-21 17:06:24 +00:00
String arguments_string() const override { return String::empty(); }
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
};
class OpCode_CheckEnd final : public OpCode {
public:
LibRegex: Remove (mostly) unused regex::MatchOutput This struct holds a counter for the number of executed operations, and vectors for matches, captures groups, and named capture groups. Each of the vectors is unused. Remove the struct and just keep a separate counter for the executed operations.
2021-08-14 21:00:58 +00:00
ExecutionResult execute(MatchInput const& input, MatchState& state) const override;
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
ALWAYS_INLINE OpCodeId opcode_id() const override { return OpCodeId::CheckEnd; }
ALWAYS_INLINE size_t size() const override { return 1; }
LibRegex: Remove some meaningless/useless const-qualifiers Also replace String creation from `""` with `String::empty()`
2021-12-21 17:06:24 +00:00
String arguments_string() const override { return String::empty(); }
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
};
LibRegex: Implement an ECMA262-compatible parser This also adds support for lookarounds and individually-negated comparisons. The only unimplemented part of the parser spec is the unicode stuff.
2020-11-27 16:03:53 +00:00
class OpCode_CheckBoundary final : public OpCode {
public:
LibRegex: Remove (mostly) unused regex::MatchOutput This struct holds a counter for the number of executed operations, and vectors for matches, captures groups, and named capture groups. Each of the vectors is unused. Remove the struct and just keep a separate counter for the executed operations.
2021-08-14 21:00:58 +00:00
ExecutionResult execute(MatchInput const& input, MatchState& state) const override;
LibRegex: Implement an ECMA262-compatible parser This also adds support for lookarounds and individually-negated comparisons. The only unimplemented part of the parser spec is the unicode stuff.
2020-11-27 16:03:53 +00:00
ALWAYS_INLINE OpCodeId opcode_id() const override { return OpCodeId::CheckBoundary; }
ALWAYS_INLINE size_t size() const override { return 2; }
ALWAYS_INLINE size_t arguments_count() const { return 1; }
ALWAYS_INLINE BoundaryCheckType type() const { return static_cast<BoundaryCheckType>(argument(0)); }
LibRegex: Remove some meaningless/useless const-qualifiers Also replace String creation from `""` with `String::empty()`
2021-12-21 17:06:24 +00:00
String arguments_string() const override { return String::formatted("kind={} ({})", (long unsigned int)argument(0), boundary_check_type_name(type())); }
LibRegex: Implement an ECMA262-compatible parser This also adds support for lookarounds and individually-negated comparisons. The only unimplemented part of the parser spec is the unicode stuff.
2020-11-27 16:03:53 +00:00
};
LibRegex: Clear previous capture group contents in ECMA262 mode ECMA262 requires that the capture groups only contain the values from the last iteration, e.g. `((c)(a)?(b))` should _not_ contain 'a' in the second capture group when matching "cabcb".
2021-07-23 15:07:18 +00:00
class OpCode_ClearCaptureGroup final : public OpCode {
public:
LibRegex: Remove (mostly) unused regex::MatchOutput This struct holds a counter for the number of executed operations, and vectors for matches, captures groups, and named capture groups. Each of the vectors is unused. Remove the struct and just keep a separate counter for the executed operations.
2021-08-14 21:00:58 +00:00
ExecutionResult execute(MatchInput const& input, MatchState& state) const override;
LibRegex: Clear previous capture group contents in ECMA262 mode ECMA262 requires that the capture groups only contain the values from the last iteration, e.g. `((c)(a)?(b))` should _not_ contain 'a' in the second capture group when matching "cabcb".
2021-07-23 15:07:18 +00:00
ALWAYS_INLINE OpCodeId opcode_id() const override { return OpCodeId::ClearCaptureGroup; }
ALWAYS_INLINE size_t size() const override { return 2; }
ALWAYS_INLINE size_t id() const { return argument(0); }
LibRegex: Remove some meaningless/useless const-qualifiers Also replace String creation from `""` with `String::empty()`
2021-12-21 17:06:24 +00:00
String arguments_string() const override { return String::formatted("id={}", id()); }
LibRegex: Clear previous capture group contents in ECMA262 mode ECMA262 requires that the capture groups only contain the values from the last iteration, e.g. `((c)(a)?(b))` should _not_ contain 'a' in the second capture group when matching "cabcb".
2021-07-23 15:07:18 +00:00
};
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
class OpCode_SaveLeftCaptureGroup final : public OpCode {
public:
LibRegex: Remove (mostly) unused regex::MatchOutput This struct holds a counter for the number of executed operations, and vectors for matches, captures groups, and named capture groups. Each of the vectors is unused. Remove the struct and just keep a separate counter for the executed operations.
2021-08-14 21:00:58 +00:00
ExecutionResult execute(MatchInput const& input, MatchState& state) const override;
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
ALWAYS_INLINE OpCodeId opcode_id() const override { return OpCodeId::SaveLeftCaptureGroup; }
ALWAYS_INLINE size_t size() const override { return 2; }
ALWAYS_INLINE size_t id() const { return argument(0); }
LibRegex: Remove some meaningless/useless const-qualifiers Also replace String creation from `""` with `String::empty()`
2021-12-21 17:06:24 +00:00
String arguments_string() const override { return String::formatted("id={}", id()); }
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
};
class OpCode_SaveRightCaptureGroup final : public OpCode {
public:
LibRegex: Remove (mostly) unused regex::MatchOutput This struct holds a counter for the number of executed operations, and vectors for matches, captures groups, and named capture groups. Each of the vectors is unused. Remove the struct and just keep a separate counter for the executed operations.
2021-08-14 21:00:58 +00:00
ExecutionResult execute(MatchInput const& input, MatchState& state) const override;
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
ALWAYS_INLINE OpCodeId opcode_id() const override { return OpCodeId::SaveRightCaptureGroup; }
ALWAYS_INLINE size_t size() const override { return 2; }
ALWAYS_INLINE size_t id() const { return argument(0); }
LibRegex: Remove some meaningless/useless const-qualifiers Also replace String creation from `""` with `String::empty()`
2021-12-21 17:06:24 +00:00
String arguments_string() const override { return String::formatted("id={}", id()); }
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
};
class OpCode_SaveRightNamedCaptureGroup final : public OpCode {
public:
LibRegex: Remove (mostly) unused regex::MatchOutput This struct holds a counter for the number of executed operations, and vectors for matches, captures groups, and named capture groups. Each of the vectors is unused. Remove the struct and just keep a separate counter for the executed operations.
2021-08-14 21:00:58 +00:00
ExecutionResult execute(MatchInput const& input, MatchState& state) const override;
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
ALWAYS_INLINE OpCodeId opcode_id() const override { return OpCodeId::SaveRightNamedCaptureGroup; }
LibRegex+LibJS: Combine named and unnamed capture groups in MatchState Combining these into one list helps reduce the size of MatchState, and as a result, reduces the amount of memory consumed during execution of very large regex matches. Doing this also allows us to remove a few regex byte code instructions: ClearNamedCaptureGroup, SaveLeftNamedCaptureGroup, and NamedReference. Named groups now behave the same as unnamed groups for these operations. Note that SaveRightNamedCaptureGroup still exists to cache the matched group name. This also removes the recursion level from the MatchState, as it can exist as a local variable in Matcher::execute instead.
2021-08-14 20:28:54 +00:00
ALWAYS_INLINE size_t size() const override { return 4; }
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
ALWAYS_INLINE StringView name() const { return { reinterpret_cast<char*>(argument(0)), length() }; }
ALWAYS_INLINE size_t length() const { return argument(1); }
LibRegex+LibJS: Combine named and unnamed capture groups in MatchState Combining these into one list helps reduce the size of MatchState, and as a result, reduces the amount of memory consumed during execution of very large regex matches. Doing this also allows us to remove a few regex byte code instructions: ClearNamedCaptureGroup, SaveLeftNamedCaptureGroup, and NamedReference. Named groups now behave the same as unnamed groups for these operations. Note that SaveRightNamedCaptureGroup still exists to cache the matched group name. This also removes the recursion level from the MatchState, as it can exist as a local variable in Matcher::execute instead.
2021-08-14 20:28:54 +00:00
ALWAYS_INLINE size_t id() const { return argument(2); }
LibRegex: Remove some meaningless/useless const-qualifiers Also replace String creation from `""` with `String::empty()`
2021-12-21 17:06:24 +00:00
String arguments_string() const override
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
{
LibRegex: Convert String::format() => String::formatted()
2021-04-21 21:02:27 +00:00
return String::formatted("name={}, length={}", name(), length());
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
}
};
class OpCode_Compare final : public OpCode {
public:
LibRegex: Remove (mostly) unused regex::MatchOutput This struct holds a counter for the number of executed operations, and vectors for matches, captures groups, and named capture groups. Each of the vectors is unused. Remove the struct and just keep a separate counter for the executed operations.
2021-08-14 21:00:58 +00:00
ExecutionResult execute(MatchInput const& input, MatchState& state) const override;
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
ALWAYS_INLINE OpCodeId opcode_id() const override { return OpCodeId::Compare; }
ALWAYS_INLINE size_t size() const override { return arguments_size() + 3; }
ALWAYS_INLINE size_t arguments_count() const { return argument(0); }
ALWAYS_INLINE size_t arguments_size() const { return argument(1); }
LibRegex: Remove some meaningless/useless const-qualifiers Also replace String creation from `""` with `String::empty()`
2021-12-21 17:06:24 +00:00
String arguments_string() const override;
Vector<String> variable_arguments_to_string(Optional<MatchInput> input = {}) const;
LibRegex: Add a basic optimization pass This currently tries to convert forking loops to atomic groups, and unify the left side of alternations.
2021-09-12 13:00:27 +00:00
Vector<CompareTypeAndValuePair> flat_compares() const;
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
private:
LibRegex: Switch to east-const style
2021-07-23 15:55:14 +00:00
ALWAYS_INLINE static void compare_char(MatchInput const& input, MatchState& state, u32 ch1, bool inverse, bool& inverse_matched);
Everywhere: Pass AK::StringView by value
2021-11-10 23:55:02 +00:00
ALWAYS_INLINE static bool compare_string(MatchInput const& input, MatchState& state, RegexStringView str, bool& had_zero_length_match);
LibRegex: Switch to east-const style
2021-07-23 15:55:14 +00:00
ALWAYS_INLINE static void compare_character_class(MatchInput const& input, MatchState& state, CharClass character_class, u32 ch, bool inverse, bool& inverse_matched);
ALWAYS_INLINE static void compare_character_range(MatchInput const& input, MatchState& state, u32 from, u32 to, u32 ch, bool inverse, bool& inverse_matched);
LibRegex+LibUnicode: Begin implementing Unicode property escapes This supports some binary property matching. It does not support any properties not yet parsed by LibUnicode, nor does it support value matching (such as Script_Extensions=Latin).
2021-07-29 18:18:51 +00:00
ALWAYS_INLINE static void compare_property(MatchInput const& input, MatchState& state, Unicode::Property property, bool inverse, bool& inverse_matched);
LibRegex: Support property escapes of Unicode General Categories This changes LibRegex to parse the property escape as a Variant of Unicode Property & General Category values. A byte code instruction is added to perform matching based on General Category values.
2021-07-31 21:46:05 +00:00
ALWAYS_INLINE static void compare_general_category(MatchInput const& input, MatchState& state, Unicode::GeneralCategory general_category, bool inverse, bool& inverse_matched);
LibRegex: Support property escapes of the Unicode script property Note that unlike binary properties and general categories, scripts must be specified in the non-binary (Script=Value) form.
2021-08-04 10:35:48 +00:00
ALWAYS_INLINE static void compare_script(MatchInput const& input, MatchState& state, Unicode::Script script, bool inverse, bool& inverse_matched);
LibRegex: Support property escapes of Unicode script extensions
2021-08-04 11:26:25 +00:00
ALWAYS_INLINE static void compare_script_extension(MatchInput const& input, MatchState& state, Unicode::Script script, bool inverse, bool& inverse_matched);
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
};
LibRegex: Implement and use a REPEAT operation for bytecode repetition Currently, when we need to repeat an instruction N times, we simply add that instruction N times in a for-loop. This doesn't scale well with extremely large values of N, and ECMA-262 allows up to N = 2^53 - 1. Instead, add a new REPEAT bytecode operation to defer this loop from the parser to the runtime executor. This allows the parser to complete sans any loops (for this instruction), and allows the executor to bail early if the repeated bytecode fails. Note: The templated ByteCode methods are to allow the Posix parsers to continue using u32 because they are limited to N = 2^20.
2021-08-12 15:02:46 +00:00
class OpCode_Repeat : public OpCode {
public:
ExecutionResult execute(MatchInput const& input, MatchState& state) const override;
ALWAYS_INLINE OpCodeId opcode_id() const override { return OpCodeId::Repeat; }
ALWAYS_INLINE size_t size() const override { return 4; }
ALWAYS_INLINE size_t offset() const { return argument(0); }
ALWAYS_INLINE u64 count() const { return argument(1); }
ALWAYS_INLINE size_t id() const { return argument(2); }
LibRegex: Remove some meaningless/useless const-qualifiers Also replace String creation from `""` with `String::empty()`
2021-12-21 17:06:24 +00:00
String arguments_string() const override
LibRegex: Implement and use a REPEAT operation for bytecode repetition Currently, when we need to repeat an instruction N times, we simply add that instruction N times in a for-loop. This doesn't scale well with extremely large values of N, and ECMA-262 allows up to N = 2^53 - 1. Instead, add a new REPEAT bytecode operation to defer this loop from the parser to the runtime executor. This allows the parser to complete sans any loops (for this instruction), and allows the executor to bail early if the repeated bytecode fails. Note: The templated ByteCode methods are to allow the Posix parsers to continue using u32 because they are limited to N = 2^20.
2021-08-12 15:02:46 +00:00
{
auto reps = id() < state().repetition_marks.size() ? state().repetition_marks.at(id()) : 0;
return String::formatted("offset={} count={} id={} rep={}, sp: {}", offset(), count() + 1, id(), reps + 1, state().string_position);
}
LibRegex: Correctly handle failing in the middle of explicit repeats - Make sure that all the Repeat ops are reset (otherwise the operation would not be correct when going over the Repeat op a second time) - Make sure that all matches that are allowed to fail are backed by a fork, otherwise the last failing fork would not have anywhere to return to. Fixes #9707.
2021-08-31 17:57:08 +00:00
};
class OpCode_ResetRepeat : public OpCode {
public:
ExecutionResult execute(MatchInput const& input, MatchState& state) const override;
ALWAYS_INLINE OpCodeId opcode_id() const override { return OpCodeId::ResetRepeat; }
ALWAYS_INLINE size_t size() const override { return 2; }
ALWAYS_INLINE size_t id() const { return argument(0); }
LibRegex: Remove some meaningless/useless const-qualifiers Also replace String creation from `""` with `String::empty()`
2021-12-21 17:06:24 +00:00
String arguments_string() const override
LibRegex: Correctly handle failing in the middle of explicit repeats - Make sure that all the Repeat ops are reset (otherwise the operation would not be correct when going over the Repeat op a second time) - Make sure that all matches that are allowed to fail are backed by a fork, otherwise the last failing fork would not have anywhere to return to. Fixes #9707.
2021-08-31 17:57:08 +00:00
{
auto reps = id() < state().repetition_marks.size() ? state().repetition_marks.at(id()) : 0;
return String::formatted("id={} rep={}", id(), reps + 1);
}
LibRegex: Make infinite repetitions short-circuit on empty matches This makes (addmittedly weird) patterns like `(a*)*` work correctly without going into an infinite fork loop.
2021-09-05 22:08:47 +00:00
};
class OpCode_Checkpoint final : public OpCode {
public:
ExecutionResult execute(MatchInput const& input, MatchState& state) const override;
ALWAYS_INLINE OpCodeId opcode_id() const override { return OpCodeId::Checkpoint; }
ALWAYS_INLINE size_t size() const override { return 1; }
LibRegex: Remove some meaningless/useless const-qualifiers Also replace String creation from `""` with `String::empty()`
2021-12-21 17:06:24 +00:00
String arguments_string() const override { return String::empty(); }
LibRegex: Make infinite repetitions short-circuit on empty matches This makes (addmittedly weird) patterns like `(a*)*` work correctly without going into an infinite fork loop.
2021-09-05 22:08:47 +00:00
};
class OpCode_JumpNonEmpty final : public OpCode {
public:
ExecutionResult execute(MatchInput const& input, MatchState& state) const override;
ALWAYS_INLINE OpCodeId opcode_id() const override { return OpCodeId::JumpNonEmpty; }
ALWAYS_INLINE size_t size() const override { return 4; }
ALWAYS_INLINE ssize_t offset() const { return argument(0); }
ALWAYS_INLINE ssize_t checkpoint() const { return argument(1); }
ALWAYS_INLINE OpCodeId form() const { return (OpCodeId)argument(2); }
LibRegex: Remove some meaningless/useless const-qualifiers Also replace String creation from `""` with `String::empty()`
2021-12-21 17:06:24 +00:00
String arguments_string() const override
LibRegex: Make infinite repetitions short-circuit on empty matches This makes (addmittedly weird) patterns like `(a*)*` work correctly without going into an infinite fork loop.
2021-09-05 22:08:47 +00:00
{
return String::formatted("{} offset={} [&{}], cp={} [&{}]",
opcode_id_name(form()),
offset(), state().instruction_position + size() + offset(),
checkpoint(), state().instruction_position + size() + checkpoint());
}
LibRegex: Implement and use a REPEAT operation for bytecode repetition Currently, when we need to repeat an instruction N times, we simply add that instruction N times in a for-loop. This doesn't scale well with extremely large values of N, and ECMA-262 allows up to N = 2^53 - 1. Instead, add a new REPEAT bytecode operation to defer this loop from the parser to the runtime executor. This allows the parser to complete sans any loops (for this instruction), and allows the executor to bail early if the repeated bytecode fails. Note: The templated ByteCode methods are to allow the Posix parsers to continue using u32 because they are limited to N = 2^20.
2021-08-12 15:02:46 +00:00
};
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
template<typename T>
LibRegex: Switch to east-const style
2021-07-23 15:55:14 +00:00
bool is(OpCode const&);
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
template<typename T>
LibRegex: Switch to east-const style
2021-07-23 15:55:14 +00:00
ALWAYS_INLINE bool is(OpCode const&)
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
{
return false;
}
template<typename T>
LibRegex: Switch to east-const style
2021-07-23 15:55:14 +00:00
ALWAYS_INLINE bool is(OpCode const* opcode)
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
{
return is<T>(*opcode);
}
template<>
LibRegex: Switch to east-const style
2021-07-23 15:55:14 +00:00
ALWAYS_INLINE bool is<OpCode_ForkStay>(OpCode const& opcode)
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
{
return opcode.opcode_id() == OpCodeId::ForkStay;
}
template<>
LibRegex: Switch to east-const style
2021-07-23 15:55:14 +00:00
ALWAYS_INLINE bool is<OpCode_Exit>(OpCode const& opcode)
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
{
return opcode.opcode_id() == OpCodeId::Exit;
}
template<>
LibRegex: Switch to east-const style
2021-07-23 15:55:14 +00:00
ALWAYS_INLINE bool is<OpCode_Compare>(OpCode const& opcode)
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
{
return opcode.opcode_id() == OpCodeId::Compare;
}
template<typename T>
Userland: Replace VERIFY(is<T>) with verify_cast<T> Instead of doing a VERIFY(is<T>(x)) and *then* casting it to T, we can just do the cast right away with verify_cast<T>. :^)
2021-06-24 19:13:09 +00:00
ALWAYS_INLINE T const& to(OpCode const& opcode)
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
{
Userland: Replace VERIFY(is<T>) with verify_cast<T> Instead of doing a VERIFY(is<T>(x)) and *then* casting it to T, we can just do the cast right away with verify_cast<T>. :^)
2021-06-24 19:13:09 +00:00
return verify_cast<T>(opcode);
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
}
template<typename T>
ALWAYS_INLINE T* to(OpCode* opcode)
{
Userland: Replace VERIFY(is<T>) with verify_cast<T> Instead of doing a VERIFY(is<T>(x)) and *then* casting it to T, we can just do the cast right away with verify_cast<T>. :^)
2021-06-24 19:13:09 +00:00
return verify_cast<T>(opcode);
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
}
template<typename T>
Userland: Replace VERIFY(is<T>) with verify_cast<T> Instead of doing a VERIFY(is<T>(x)) and *then* casting it to T, we can just do the cast right away with verify_cast<T>. :^)
2021-06-24 19:13:09 +00:00
ALWAYS_INLINE T const* to(OpCode const* opcode)
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
{
Userland: Replace VERIFY(is<T>) with verify_cast<T> Instead of doing a VERIFY(is<T>(x)) and *then* casting it to T, we can just do the cast right away with verify_cast<T>. :^)
2021-06-24 19:13:09 +00:00
return verify_cast<T>(opcode);
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
}
template<typename T>
ALWAYS_INLINE T& to(OpCode& opcode)
{
Userland: Replace VERIFY(is<T>) with verify_cast<T> Instead of doing a VERIFY(is<T>(x)) and *then* casting it to T, we can just do the cast right away with verify_cast<T>. :^)
2021-06-24 19:13:09 +00:00
return verify_cast<T>(opcode);
LibRegex: Add a regular expression library This commit is a mix of several commits, squashed into one because the commits before 'Move regex to own Library and fix all the broken stuff' were not fixable in any elegant way. The commits are listed below for "historical" purposes: - AK: Add options/flags and Errors for regular expressions Flags can be provided for any possible flavour by adding a new scoped enum. Handling of flags is done by templated Options class and the overloaded '|' and '&' operators. - AK: Add Lexer for regular expressions The lexer parses the input and extracts tokens needed to parse a regular expression. - AK: Add regex Parser and PosixExtendedParser This patchset adds a abstract parser class that can be derived to implement different parsers. A parser produces bytecode to be executed within the regex matcher. - AK: Add regex matcher This patchset adds an regex matcher based on the principles of the T-REX VM. The bytecode pruduced by the respective Parser is put into the matcher and the VM will recursively execute the bytecode according to the available OpCodes. Possible improvement: the recursion could be replaced by multi threading capabilities. To match a Regular expression, e.g. for the Posix standard regular expression matcher use the following API: ``` Pattern<PosixExtendedParser> pattern("^.*$"); auto result = pattern.match("Well, hello friends!\nHello World!"); // Match whole needle EXPECT(result.count == 1); EXPECT(result.matches.at(0).view.starts_with("Well")); EXPECT(result.matches.at(0).view.end() == "!"); result = pattern.match("Well, hello friends!\nHello World!", PosixFlags::Multiline); // Match line by line EXPECT(result.count == 2); EXPECT(result.matches.at(0).view == "Well, hello friends!"); EXPECT(result.matches.at(1).view == "Hello World!"); EXPECT(pattern.has_match("Well,....")); // Just check if match without a result, which saves some resources. ``` - AK: Rework regex to work with opcodes objects This patchsets reworks the matcher to work on a more structured base. For that an abstract OpCode class and derived classes for the specific OpCodes have been added. The respective opcode logic is contained in each respective execute() method. - AK: Add benchmark for regex - AK: Some optimization in regex for runtime and memory - LibRegex: Move regex to own Library and fix all the broken stuff Now regex works again and grep utility is also in place for testing. This commit also fixes the use of regex.h in C by making `regex_t` an opaque (-ish) type, which makes its behaviour consistent between C and C++ compilers. Previously, <regex.h> would've blown C compilers up, and even if it didn't, would've caused a leak in C code, and not in C++ code (due to the existence of `OwnPtr` inside the struct). To make this whole ordeal easier to deal with (for now), this pulls the definitions of `reg*()` into LibRegex. pros: - The circular dependency between LibC and LibRegex is broken - Eaiser to test (without accidentally pulling in the host's libc!) cons: - Using any of the regex.h functions will require the user to link -lregex - The symbols will be missing from libc, which will be a big surprise down the line (especially with shared libs). Co-Authored-By: Ali Mohammad Pur <ali.mpfard@gmail.com>
2020-04-26 12:45:10 +00:00
}
}
Reference in a new issue Copy permalink