mirror of
https://github.com/LadybirdBrowser/ladybird.git
synced 2024-11-25 17:10:23 +00:00
5d170810db
Prior to this, it'd try to stuff them into an i64, which could fail and give us nothing. Even though this is an extension we've made to JSON, the parser should be able to correctly round-trip from whatever our serialiser has generated.
356 lines
8.3 KiB
C++
356 lines
8.3 KiB
C++
/*
|
|
* Copyright (c) 2018-2020, Andreas Kling <kling@serenityos.org>
|
|
*
|
|
* SPDX-License-Identifier: BSD-2-Clause
|
|
*/
|
|
|
|
#include <AK/CharacterTypes.h>
|
|
#include <AK/JsonArray.h>
|
|
#include <AK/JsonObject.h>
|
|
#include <AK/JsonParser.h>
|
|
|
|
namespace AK {
|
|
|
|
constexpr bool is_space(int ch)
|
|
{
|
|
return ch == '\t' || ch == '\n' || ch == '\r' || ch == ' ';
|
|
}
|
|
|
|
String JsonParser::consume_and_unescape_string()
|
|
{
|
|
if (!consume_specific('"'))
|
|
return {};
|
|
StringBuilder final_sb;
|
|
|
|
for (;;) {
|
|
size_t peek_index = m_index;
|
|
char ch = 0;
|
|
for (;;) {
|
|
if (peek_index == m_input.length())
|
|
break;
|
|
ch = m_input[peek_index];
|
|
if (ch == '"' || ch == '\\')
|
|
break;
|
|
if (is_ascii_c0_control(ch))
|
|
return {};
|
|
++peek_index;
|
|
}
|
|
|
|
while (peek_index != m_index) {
|
|
final_sb.append(m_input[m_index]);
|
|
m_index++;
|
|
}
|
|
|
|
if (m_index == m_input.length())
|
|
break;
|
|
if (ch == '"')
|
|
break;
|
|
if (ch != '\\') {
|
|
final_sb.append(consume());
|
|
continue;
|
|
}
|
|
ignore();
|
|
if (next_is('"')) {
|
|
ignore();
|
|
final_sb.append('"');
|
|
continue;
|
|
}
|
|
|
|
if (next_is('\\')) {
|
|
ignore();
|
|
final_sb.append('\\');
|
|
continue;
|
|
}
|
|
|
|
if (next_is('/')) {
|
|
ignore();
|
|
final_sb.append('/');
|
|
continue;
|
|
}
|
|
|
|
if (next_is('n')) {
|
|
ignore();
|
|
final_sb.append('\n');
|
|
continue;
|
|
}
|
|
|
|
if (next_is('r')) {
|
|
ignore();
|
|
final_sb.append('\r');
|
|
continue;
|
|
}
|
|
|
|
if (next_is('t')) {
|
|
ignore();
|
|
final_sb.append('\t');
|
|
continue;
|
|
}
|
|
|
|
if (next_is('b')) {
|
|
ignore();
|
|
final_sb.append('\b');
|
|
continue;
|
|
}
|
|
|
|
if (next_is('f')) {
|
|
ignore();
|
|
final_sb.append('\f');
|
|
continue;
|
|
}
|
|
|
|
if (next_is('u')) {
|
|
ignore();
|
|
if (tell_remaining() < 4)
|
|
return {};
|
|
|
|
auto code_point = AK::StringUtils::convert_to_uint_from_hex(consume(4));
|
|
if (code_point.has_value()) {
|
|
final_sb.append_code_point(code_point.value());
|
|
continue;
|
|
} else {
|
|
return {};
|
|
}
|
|
}
|
|
|
|
return {};
|
|
}
|
|
if (!consume_specific('"'))
|
|
return {};
|
|
|
|
return final_sb.to_string();
|
|
}
|
|
|
|
Optional<JsonValue> JsonParser::parse_object()
|
|
{
|
|
JsonObject object;
|
|
if (!consume_specific('{'))
|
|
return {};
|
|
for (;;) {
|
|
ignore_while(is_space);
|
|
if (peek() == '}')
|
|
break;
|
|
ignore_while(is_space);
|
|
auto name = consume_and_unescape_string();
|
|
if (name.is_null())
|
|
return {};
|
|
ignore_while(is_space);
|
|
if (!consume_specific(':'))
|
|
return {};
|
|
ignore_while(is_space);
|
|
auto value = parse_helper();
|
|
if (!value.has_value())
|
|
return {};
|
|
object.set(name, value.release_value());
|
|
ignore_while(is_space);
|
|
if (peek() == '}')
|
|
break;
|
|
if (!consume_specific(','))
|
|
return {};
|
|
ignore_while(is_space);
|
|
if (peek() == '}')
|
|
return {};
|
|
}
|
|
if (!consume_specific('}'))
|
|
return {};
|
|
return JsonValue { move(object) };
|
|
}
|
|
|
|
Optional<JsonValue> JsonParser::parse_array()
|
|
{
|
|
JsonArray array;
|
|
if (!consume_specific('['))
|
|
return {};
|
|
for (;;) {
|
|
ignore_while(is_space);
|
|
if (peek() == ']')
|
|
break;
|
|
auto element = parse_helper();
|
|
if (!element.has_value())
|
|
return {};
|
|
array.append(element.release_value());
|
|
ignore_while(is_space);
|
|
if (peek() == ']')
|
|
break;
|
|
if (!consume_specific(','))
|
|
return {};
|
|
ignore_while(is_space);
|
|
if (peek() == ']')
|
|
return {};
|
|
}
|
|
ignore_while(is_space);
|
|
if (!consume_specific(']'))
|
|
return {};
|
|
return JsonValue { move(array) };
|
|
}
|
|
|
|
Optional<JsonValue> JsonParser::parse_string()
|
|
{
|
|
auto result = consume_and_unescape_string();
|
|
if (result.is_null())
|
|
return {};
|
|
return JsonValue(result);
|
|
}
|
|
|
|
Optional<JsonValue> JsonParser::parse_number()
|
|
{
|
|
JsonValue value;
|
|
Vector<char, 128> number_buffer;
|
|
Vector<char, 128> fraction_buffer;
|
|
|
|
bool is_double = false;
|
|
for (;;) {
|
|
char ch = peek();
|
|
if (ch == '.') {
|
|
if (is_double)
|
|
return {};
|
|
|
|
is_double = true;
|
|
++m_index;
|
|
continue;
|
|
}
|
|
if (ch == '-' || (ch >= '0' && ch <= '9')) {
|
|
if (is_double) {
|
|
if (ch == '-')
|
|
return {};
|
|
|
|
fraction_buffer.append(ch);
|
|
} else {
|
|
if (number_buffer.size() > 0) {
|
|
if (number_buffer.at(0) == '0')
|
|
return {};
|
|
}
|
|
|
|
if (number_buffer.size() > 1) {
|
|
if (number_buffer.at(0) == '-' && number_buffer.at(1) == '0')
|
|
return {};
|
|
}
|
|
|
|
number_buffer.append(ch);
|
|
}
|
|
++m_index;
|
|
continue;
|
|
}
|
|
break;
|
|
}
|
|
|
|
StringView number_string(number_buffer.data(), number_buffer.size());
|
|
StringView fraction_string(fraction_buffer.data(), fraction_buffer.size());
|
|
|
|
#ifndef KERNEL
|
|
if (is_double) {
|
|
// FIXME: This logic looks shaky.
|
|
int whole = 0;
|
|
auto to_signed_result = number_string.to_uint();
|
|
if (to_signed_result.has_value()) {
|
|
whole = to_signed_result.value();
|
|
} else {
|
|
auto number = number_string.to_int();
|
|
if (!number.has_value())
|
|
return {};
|
|
whole = number.value();
|
|
}
|
|
|
|
auto fraction_string_uint = fraction_string.to_uint();
|
|
if (!fraction_string_uint.has_value())
|
|
return {};
|
|
int fraction = fraction_string_uint.value();
|
|
fraction *= (whole < 0) ? -1 : 1;
|
|
|
|
auto divider = 1;
|
|
for (size_t i = 0; i < fraction_buffer.size(); ++i) {
|
|
divider *= 10;
|
|
}
|
|
value = JsonValue((double)whole + ((double)fraction / divider));
|
|
} else {
|
|
#endif
|
|
auto to_unsigned_result = number_string.to_uint<u64>();
|
|
if (to_unsigned_result.has_value()) {
|
|
auto number = *to_unsigned_result;
|
|
if (number <= NumericLimits<u32>::max())
|
|
value = JsonValue((u32)number);
|
|
else
|
|
value = JsonValue(number);
|
|
} else {
|
|
auto number = number_string.to_int<i64>();
|
|
if (!number.has_value())
|
|
return {};
|
|
if (number.value() <= NumericLimits<i32>::max()) {
|
|
value = JsonValue((i32)number.value());
|
|
} else {
|
|
value = JsonValue(number.value());
|
|
}
|
|
}
|
|
#ifndef KERNEL
|
|
}
|
|
#endif
|
|
|
|
return value;
|
|
}
|
|
|
|
Optional<JsonValue> JsonParser::parse_true()
|
|
{
|
|
if (!consume_specific("true"))
|
|
return {};
|
|
return JsonValue(true);
|
|
}
|
|
|
|
Optional<JsonValue> JsonParser::parse_false()
|
|
{
|
|
if (!consume_specific("false"))
|
|
return {};
|
|
return JsonValue(false);
|
|
}
|
|
|
|
Optional<JsonValue> JsonParser::parse_null()
|
|
{
|
|
if (!consume_specific("null"))
|
|
return {};
|
|
return JsonValue(JsonValue::Type::Null);
|
|
}
|
|
|
|
Optional<JsonValue> JsonParser::parse_helper()
|
|
{
|
|
ignore_while(is_space);
|
|
auto type_hint = peek();
|
|
switch (type_hint) {
|
|
case '{':
|
|
return parse_object();
|
|
case '[':
|
|
return parse_array();
|
|
case '"':
|
|
return parse_string();
|
|
case '-':
|
|
case '0':
|
|
case '1':
|
|
case '2':
|
|
case '3':
|
|
case '4':
|
|
case '5':
|
|
case '6':
|
|
case '7':
|
|
case '8':
|
|
case '9':
|
|
return parse_number();
|
|
case 'f':
|
|
return parse_false();
|
|
case 't':
|
|
return parse_true();
|
|
case 'n':
|
|
return parse_null();
|
|
}
|
|
|
|
return {};
|
|
}
|
|
|
|
Optional<JsonValue> JsonParser::parse()
|
|
{
|
|
auto result = parse_helper();
|
|
if (!result.has_value())
|
|
return {};
|
|
ignore_while(is_space);
|
|
if (!is_eof())
|
|
return {};
|
|
return result;
|
|
}
|
|
|
|
}
|