eliott/ladybird
1
0
Fork 0
mirror of https://github.com/LadybirdBrowser/ladybird synced 2026-05-03 13:02:09 +02:00
Code Issues Packages Projects Releases Wiki Activity

AK+Everywhere: Replace custom number parsers with fast_float

Browse Source 
Our floating point number parser was based on the fast_float library:
https://github.com/fastfloat/fast_float

However, our implementation only supports 8-bit characters. To support
UTF-16, we will need to be able to convert char16_t-based strings to
numbers as well. This works out-of-the-box with fast_float.

We can also use fast_float for integer parsing.
This commit is contained in:
Timothy Flynn
2025-06-26 19:06:46 -04:00
committed by Tim Flynn
parent 9fc3e72db2
commit 62d9a84b8d
Notes: github-actions[bot] 2025-07-03 13:53:10 +00:00 
Author: https://github.com/trflynn89
Commit: https://github.com/LadybirdBrowser/ladybird/commit/62d9a84b8d7
Pull-request: https://github.com/LadybirdBrowser/ladybird/pull/5228
Reviewed-by: https://github.com/ADKaster ✅
Reviewed-by: https://github.com/shannonbooth
30 changed files with 413 additions and 3034 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
Libraries/LibWeb/CSS/Parser/Tokenizer.cpp
View File

@@ -6,8 +6,8 @@
*/
#include <AK/Debug.h>
#include <AK/FloatingPointStringConversions.h>
#include <AK/SourceLocation.h>
#include <AK/StringConversions.h>
#include <AK/Vector.h>
#include <LibTextCodec/Decoder.h>
#include <LibWeb/CSS/CharacterTypes.h>
@@ -377,7 +377,7 @@ u32 Tokenizer::consume_escaped_code_point()
}
// Interpret the hex digits as a hexadecimal number.
auto unhexed = AK::StringUtils::convert_to_uint_from_hex<u32>(builder.string_view()).value_or(0);
auto unhexed = AK::parse_hexadecimal_number<u32>(builder.string_view()).value_or(0);
// If this number is zero, or is for a surrogate, or is greater than the maximum allowed
// code point, return U+FFFD REPLACEMENT CHARACTER (<28>).
if (unhexed == 0 || is_unicode_surrogate(unhexed) || is_greater_than_maximum_allowed_code_point(unhexed)) {

9
Libraries/LibWeb/CSS/Parser/ValueParsing.cpp
View File

@@ -16,6 +16,7 @@
#include <AK/Debug.h>
#include <AK/GenericLexer.h>
#include <AK/QuickSort.h>
#include <AK/StringConversions.h>
#include <AK/TemporaryChange.h>
#include <LibWeb/CSS/FontFace.h>
#include <LibWeb/CSS/Parser/Parser.h>
@@ -564,7 +565,7 @@ Optional<Gfx::UnicodeRange> Parser::parse_unicode_range(StringView text)
// with the U+003F QUESTION MARK (?) code points replaced by U+0030 DIGIT ZERO (0) code points.
// This is the start value.
auto start_value_string = start_value_code_points.replace("?"sv, "0"sv, ReplaceMode::All);
auto maybe_start_value = AK::StringUtils::convert_to_uint_from_hex<u32>(start_value_string);
auto maybe_start_value = AK::parse_hexadecimal_number<u32>(start_value_string);
if (!maybe_start_value.has_value()) {
dbgln_if(CSS_PARSER_DEBUG, "CSSParser: <urange> ?-converted start value did not parse as hex number.");
return {};
@@ -575,7 +576,7 @@ Optional<Gfx::UnicodeRange> Parser::parse_unicode_range(StringView text)
// with the U+003F QUESTION MARK (?) code points replaced by U+0046 LATIN CAPITAL LETTER F (F) code points.
// This is the end value.
auto end_value_string = start_value_code_points.replace("?"sv, "F"sv, ReplaceMode::All);
auto maybe_end_value = AK::StringUtils::convert_to_uint_from_hex<u32>(end_value_string);
auto maybe_end_value = AK::parse_hexadecimal_number<u32>(end_value_string);
if (!maybe_end_value.has_value()) {
dbgln_if(CSS_PARSER_DEBUG, "CSSParser: <urange> ?-converted end value did not parse as hex number.");
return {};
@@ -586,7 +587,7 @@ Optional<Gfx::UnicodeRange> Parser::parse_unicode_range(StringView text)
return make_valid_unicode_range(start_value, end_value);
}
// Otherwise, interpret the consumed code points as a hexadecimal number. This is the start value.
auto maybe_start_value = AK::StringUtils::convert_to_uint_from_hex<u32>(start_value_code_points);
auto maybe_start_value = AK::parse_hexadecimal_number<u32>(start_value_code_points);
if (!maybe_start_value.has_value()) {
dbgln_if(CSS_PARSER_DEBUG, "CSSParser: <urange> start value did not parse as hex number.");
return {};
@@ -625,7 +626,7 @@ Optional<Gfx::UnicodeRange> Parser::parse_unicode_range(StringView text)
}
// 7. Interpret the consumed code points as a hexadecimal number. This is the end value.
auto maybe_end_value = AK::StringUtils::convert_to_uint_from_hex<u32>(end_hex_digits);
auto maybe_end_value = AK::parse_hexadecimal_number<u32>(end_hex_digits);
if (!maybe_end_value.has_value()) {
dbgln_if(CSS_PARSER_DEBUG, "CSSParser: <urange> end value did not parse as hex number.");
return {};

2
Libraries/LibWeb/HTML/HTMLFontElement.cpp
View File

@@ -54,7 +54,7 @@ Optional<CSS::Keyword> HTMLFontElement::parse_legacy_font_size(StringView string
lexer.consume_while(is_ascii_digit);
size_t end_index = lexer.tell();
auto digits = lexer.input().substring_view(start_index, end_index - start_index);
auto value_or_empty = AK::StringUtils::convert_to_int<i32>(digits);
auto value_or_empty = digits.to_number<i32>();
// 7. If digits is the empty string, there is no presentational hint. Return.
if (!value_or_empty.has_value())

2
Libraries/LibWeb/HTML/HTMLTableColElement.cpp
View File

@@ -34,7 +34,7 @@ WebIDL::UnsignedLong HTMLTableColElement::span() const
// The span IDL attribute must reflect the content attribute of the same name. It is clamped to the range [1, 1000], and its default value is 1.
if (auto span_string = get_attribute(HTML::AttributeNames::span); span_string.has_value()) {
if (auto span_digits = parse_non_negative_integer_digits(*span_string); span_digits.has_value()) {
auto span = AK::StringUtils::convert_to_int<i64>(*span_digits);
auto span = span_digits->to_number<i64>();
// NOTE: If span has no value at this point, the value must be larger than NumericLimits<i64>::max(), so return the maximum value of 1000.
if (!span.has_value())
return 1000;

5
Libraries/LibWeb/HTML/Numbers.cpp
View File

@@ -5,6 +5,7 @@
*/
#include <AK/GenericLexer.h>
#include <AK/StringConversions.h>
#include <LibWeb/HTML/Numbers.h>
#include <LibWeb/Infra/CharacterTypes.h>
#include <math.h>
@@ -173,7 +174,7 @@ Optional<double> parse_floating_point_number(StringView string)
lexer.consume_while(is_ascii_digit);
size_t end_index = lexer.tell();
auto digits = lexer.input().substring_view(start_index, end_index - start_index);
auto optional_value = AK::StringUtils::convert_to_floating_point<double>(digits, TrimWhitespace::No);
auto optional_value = digits.to_number<double>(TrimWhitespace::No);
value *= optional_value.value();
}
@@ -274,7 +275,7 @@ fraction_exit:
lexer.consume_while(is_ascii_digit);
size_t end_index = lexer.tell();
auto digits = lexer.input().substring_view(start_index, end_index - start_index);
auto optional_value = AK::StringUtils::convert_to_int<i32>(digits);
auto optional_value = digits.to_number<i32>();
exponent *= optional_value.value();
}

15
Libraries/LibWeb/SVG/AttributeParser.cpp
View File

@@ -7,10 +7,10 @@
* SPDX-License-Identifier: BSD-2-Clause
*/
#include "AttributeParser.h"
#include <AK/FloatingPointStringConversions.h>
#include <AK/GenericShorthands.h>
#include <AK/StringBuilder.h>
#include <AK/StringConversions.h>
#include <LibWeb/SVG/AttributeParser.h>
#include <ctype.h>
namespace Web::SVG {
@@ -439,14 +439,11 @@ ErrorOr<float> AttributeParser::parse_nonnegative_number()
if (match('+') || match('-') || !match_number())
return Error::from_string_literal("Expected number");
auto remaining_source_text = m_lexer.remaining();
char const* start = remaining_source_text.characters_without_null_termination();
auto parse_result = AK::parse_first_number<float>(m_lexer.remaining(), TrimWhitespace::No);
VERIFY(parse_result.has_value());
auto maybe_float = parse_first_floating_point<float>(start, start + remaining_source_text.length());
VERIFY(maybe_float.parsed_value());
m_lexer.ignore(maybe_float.end_ptr - start);
return maybe_float.value;
m_lexer.ignore(parse_result->characters_parsed);
return parse_result->value;
}
ErrorOr<float> AttributeParser::parse_flag()