Commit graph

485 commits

Author SHA1 Message Date
Andreas Kling
9889d170b9 JsonValue: Add as_string_or(String)
Return the contained string if the value *is* a string, otherwise it
returns the alternative string passed in the parameter.
2019-08-07 22:03:25 +02:00
Andreas Kling
f6998b1817 JSON: Templatize the JSON serialization code
This makes it possible to use something other than a StringBuilder for
serialization (and to produce something other than a String.) :^)
2019-08-07 21:29:32 +02:00
Andreas Kling
43ec733b61 AK: Add a basic unit test for FileSystemPath
Just to make sure that things are on the up-and-up.
2019-08-07 21:07:02 +02:00
Andreas Kling
2e416b1b87 Vector: Add a test for growing a Vector beyond its inline capacity 2019-08-07 20:43:13 +02:00
Andreas Kling
5096eaa845 AK: Add a FixedArray<T> container
This is a simple array wrapper that knows its size. It has begin/end
so you can use range-for. It also has a resize() that reallocates.
2019-08-07 20:05:36 +02:00
Andreas Kling
5487f81b5d Vector: Use memcpy when dynamically growing Vectors of trivial types 2019-08-07 15:35:23 +02:00
Andreas Kling
6da6ca64d2 Vector: Use TypedTransfer in more parts of Vector
Make more Vector-of-trivial-type operations go fast :^)
2019-08-07 15:25:34 +02:00
Andreas Kling
e8e85f5457 Vector: Use memcmp for comparing two vectors with trivial elements 2019-08-07 15:05:10 +02:00
Andreas Kling
6d97caf124 JsonParser: Scan ahead to find the first special char in quoted strings
This allows us to take advantage of the now-optimized (to do memmove())
Vector::append(const T*, int count) for collecting these strings.

This is a ~15% speedup on the load_4chan_catalog benchmark.
2019-08-07 11:57:51 +02:00
Andreas Kling
b48b6c0caa Vector: Use memmove() for moving trivial types around more
This can definitely be improved with better trivial type detection and
by using the TypedTransfer template in more places.

It's a bit annoying that we can't get <type_traits> in Vector.h since
it's included in the toolchain compilation before we have libstdc++.
2019-08-07 11:55:20 +02:00
Andreas Kling
60c25228ee AK: Fix -Wconsumed warnings in Optional move-ctor and move-assign
Our protocol says we have to call has_value() before release_value().
The code was already safe, but the compiler had no way of knowing that.
2019-08-07 07:17:52 +02:00
Andreas Kling
2abfab2892 ELFLoader: Remove an uninteresting debug log message
This is showing up at the boundary between kernel and userspace stack
frames in backtraces, and looks silly.
2019-08-06 20:49:36 +02:00
Conrad Pankoff
73c998dbfc Kernel: Refactor TCP/IP stack
This has several significant changes to the networking stack.

* Significant refactoring of the TCP state machine. Right now it's
  probably more fragile than it used to be, but handles quite a lot
  more of the handshake process.
* `TCPSocket` holds a `NetworkAdapter*`, assigned during `connect()` or
  `bind()`, whichever comes first.
* `listen()` is now virtual in `Socket` and intended to be implemented
  in its child classes
* `listen()` no longer works without `bind()` - this is a bit of a
  regression, but listening sockets didn't work at all before, so it's
  not possible to observe the regression.
* A file is exposed at `/proc/net_tcp`, which is a JSON document listing
  the current TCP sockets with a bit of metadata.
* There's an `ETHERNET_VERY_DEBUG` flag for dumping packet's content out
  to `kprintf`. It is, indeed, _very debug_.
2019-08-06 16:21:17 +02:00
Andreas Kling
151e6a1818 AK: Fix leak in Optional(Optional&&)
We were *copying* the other Optional's value and then marking it as not
having a value. This prevented us from ever destroying the original.
2019-08-05 22:27:47 +02:00
Andreas Kling
9553ecfe01 AK: Optional::operator=(Optional&&) should clear movee's has_value bit
We were forgetting to clear m_has_value in the Optional being moved
from when using operator=(Optional&&).
2019-08-05 21:47:36 +02:00
Sergey Bugaev
55d7810fab AK: Do not keep an open FD in MmappedFile
We only need an open FD while establishing the mapping,
and can close it immediately after mmap() call.
2019-08-05 16:04:31 +02:00
Andreas Kling
0d0230ab10 JsonParser: Fold extract_while() into parse_number()
It wasn't unsed anywhere else anyway, and this is actually ~1% faster
on the load_4chan_catalog benchmark.
2019-08-04 20:23:46 +02:00
Andreas Kling
a846ee76ea AK: Add VectorIterator::index()
A simple helper function that extracts the index of an iterator.
Note that the index is not valid if the iterator is end().
2019-08-04 20:18:18 +02:00
Andreas Kling
2a89bb2ac4 Vector: Add find() and some iterator improvements
Vector now has find() just like HashTable. I also made the iterator
comparison functions const-correct.
2019-08-04 19:21:08 +02:00
Andreas Kling
539985f4fe HashTable: Use the Bucket type in some places over SinglyLinkedList<T>
This is just for consistency, and we might want to switch to another
bucket type some day.
2019-08-04 19:20:20 +02:00
Andreas Kling
72b69b82bb JsonParser: Oops, fix build. 2019-08-04 18:59:06 +02:00
Andreas Kling
4e004a664f JsonParser: Cache the last seen string starting with each possible char
Keep a 256-entry string cache during parse to avoid creating some new
strings when possible. This cache is far from perfect but very cheap.
Since none of the strings are transient, this only costs us a couple of
pointers and a bit of ref-count manipulation.

The cache hit rate on 4chan_catalog.json is ~33% and the speedup on
the load_4chan_catalog benchmark is ~7%.
2019-08-04 18:41:24 +02:00
Andreas Kling
3eb1a7f8f8 AK: Add a benchmark for parsing 4chan catalog JSON
I was able to get parsing time down to about 1/3 of the original time
by using callgrind+kcachegrind. There's definitely more improvements
that can be made here, but I'm gonna be happy with this for now. :^)
2019-08-04 11:57:32 +02:00
Andreas Kling
b62a12c687 JsonParser: Some minor optimizations
- Return more specific types from parse_array() and parse_object().
- Don't create a throwaway String in extract_while().
- Use a StringView in parse_number() to avoid a throwaway String.
2019-08-04 11:47:21 +02:00
Andreas Kling
93596dc00d JsonObject: Add move constructors for JsonObject(Json{Array,Object}&&)
This also helps avoid JsonValue copying during parse.
2019-08-04 11:46:31 +02:00
Andreas Kling
82826104e0 Json: Add efficient copy and move constructors for Json{Array,Object}
This helps avoid copying JsonValues during parsing.
2019-08-04 11:45:16 +02:00
Andreas Kling
cce2ea9bb0 AK: Add StringView::to_int()
This is a shameless copy-paste of String::to_int(). We should find some
way to share this code between String and StringView instead of having
two duplicate copies like this.
2019-08-04 11:44:20 +02:00
Andreas Kling
c55129e573 JsonParser: Use Vector<char, 1024> instead of StringBuilder in parsing
This is a 10-12% speedup on the 4chan thread catalog JSON.
2019-08-04 10:38:15 +02:00
Andreas Kling
210550d4b3 JsonArray: Expose the value vector to the outside for convenience
Sometimes it's easier to just work with a const Vector<JsonValue>&,
so give clients the option of doing that.
2019-08-04 10:05:02 +02:00
Andreas Kling
6797f71e73 JsonParser: When encountering \uXXXX, just emit a "?" for now. 2019-08-04 08:58:45 +02:00
Andreas Kling
18aeda2e0d BufferStream: It's not an error to read up to the very end of a buffer
Don't signal an error before we've reached the very end of the buffer.
2019-08-03 19:20:05 +02:00
Andreas Kling
6e956f2a0b BufferStream: Add operator<<(const char*) overload
Ext2FS was relying on the presence of a (StringView) overload here and
when I added a (bool) overload, it started choosing that for literals.
2019-08-03 19:02:48 +02:00
Andreas Kling
b04dae7faa BufferStream: Support "bool" as a streaming type 2019-08-03 17:02:42 +02:00
Andreas Kling
c5903ec4bb AK: Remove two redundant RefPtr constructors.
We already have constructors for "const T*" and "const T&" so we don't
need to have non-const variants.
2019-08-03 14:16:05 +02:00
Andreas Kling
fe25426ee4 AK: Simplify RefPtr and NonnullRefPtr's leak_ref() functions
Use AK::exchange() to switch out the internal storage. Also mark these
functions with [[nodiscard]] to provoke an compile-time error if they
are called without using the return value.
2019-08-02 12:05:09 +02:00
Andreas Kling
15866714da AK: Add anti-null assertions in RefPtr.
This gives us better error messages when dereferencing null RefPtrs.
2019-08-02 12:00:43 +02:00
Andreas Kling
cbc1272810 AK: Fix ref leaks in RefPtr assignment operators.
Many of the RefPtr assignment operators would cause ref leaks when we
call them to assign a pointer that's already the one kept.
2019-08-02 11:56:55 +02:00
Andreas Kling
06f82901b7 AK: Explicitly delete NonnullRefPtr::operator=(RefPtr).
This gives us much better error messages when you try to use them.
Without this change, it would complain about the absence of functions
named ref() and deref() on RefPtr itself. With it, we instead get a
"hey, this function is deleted" error.

Change operator=(T&) to operator=T(const T&) also, to keep assigning
a const T& to a NonnullRefPtr working.
2019-08-02 11:51:28 +02:00
Andreas Kling
6db879ee66 AK: Fix ref leak in NonnullRefPtr::operator=(T&).
We would leak a ref when assigning a T& to a NonnullRefPtr that already
contains that same T.
2019-08-02 11:35:05 +02:00
Andreas Kling
d9cc3e453c AK: Add assertions when dereferencing an OwnPtr.
This will make it immediately obvious what the problem is when you're
dereferencing a null OwnPtr.
2019-08-02 10:34:40 +02:00
Andreas Kling
a9a1a5dfa9 AK: Add a test for iterating a HashTable during clear (should assert)
Ideally we should also verify that the assertion actually happens,
but we need some support in the TestSuite framework for that.
2019-08-02 09:25:35 +02:00
Andreas Kling
6560116b67 TestSuite: Hijack the ASSERT macros during unit tests.
Instead of aborting the program when we hit an assertion, just print a
message and keep going.

This allows us to write tests that provoke assertions on purpose.
2019-08-02 09:23:03 +02:00
Andreas Kling
31793b8f3a AK: Fix typo in the WeakPtr test. Behavior was actually correct.
Also remove an unused variable.
2019-08-02 09:21:42 +02:00
Andreas Kling
4e59300650 AK: Fix typo in TestVector.cpp, oops. 2019-08-02 08:59:26 +02:00
Andreas Kling
9230b42f28 AK: Use Vector::empend() a bit in the unit tests, and fix a bug.
There was a bug in the "prepend_vector_object" test but it was masked
by us not printing failures. (The bug was that we were adding three
elements to the "objects" vector and then checking that another
vector called "more_objects" indeed had three elements. Oops!)
2019-08-01 16:24:31 +02:00
Andreas Kling
8434548f14 TestSuite: Actually print failed comparions.. :^) 2019-08-01 16:22:50 +02:00
Andreas Kling
eeff0cd570 AK: Don't allow constructing an OwnPtr from a const NonnullOwnPtr&
OwnPtr's must move around, they can't be copy constructed.
2019-08-01 15:46:18 +02:00
Andreas Kling
79ce75d862 AK: Add Vector::empend().
This is a complement to append() that works by constructing the new
element in-place via placement new and forwarded constructor arguments.

The STL calls this emplace_back() which looks ugly, so I'm inventing
a nice word for it instead. :^)
2019-08-01 15:35:45 +02:00
Andreas Kling
a3ee35510f AK: Make Bitmap movable but not copyable.
We were falling back to an incorrect compiler-generated copy ctor for
this class, and let's not do that.

Found by PVS-Studio.
2019-08-01 11:34:36 +02:00
Andreas Kling
2923d39106 JsonParser: Merge the parsing of '\n' and '\r' in quoted strings 2019-08-01 10:45:37 +02:00