AK+Everywhere: Replace custom number parsers with fast_float
Some checks failed
CI / macOS, arm64, Sanitizer_CI, Clang (push) Waiting to run
CI / Linux, x86_64, Fuzzers_CI, Clang (push) Waiting to run
CI / Linux, x86_64, Sanitizer_CI, GNU (push) Waiting to run
CI / Linux, x86_64, Sanitizer_CI, Clang (push) Waiting to run
Package the js repl as a binary artifact / Linux, arm64 (push) Waiting to run
Package the js repl as a binary artifact / macOS, arm64 (push) Waiting to run
Package the js repl as a binary artifact / Linux, x86_64 (push) Waiting to run
Run test262 and test-wasm / run_and_update_results (push) Waiting to run
Lint Code / lint (push) Waiting to run
Label PRs with merge conflicts / auto-labeler (push) Waiting to run
Push notes / build (push) Waiting to run
Build Dev Container Image / build (push) Has been cancelled

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
commit 62d9a84b8d
Notes: github-actions[bot] 2025-07-03 13:53:10 +00:00
30 changed files with 413 additions and 3034 deletions

View file

@ -6,7 +6,6 @@
*/
#include <AK/CharacterTypes.h>
#include <AK/FloatingPointStringConversions.h>
#include <AK/Function.h>
#include <AK/Optional.h>
#include <AK/Utf16View.h>

View file

@ -7,8 +7,8 @@
#include <AK/BuiltinWrappers.h>
#include <AK/CharacterTypes.h>
#include <AK/FloatingPointStringConversions.h>
#include <AK/Hex.h>
#include <AK/StringConversions.h>
#include <AK/UnicodeUtils.h>
#include <AK/Utf16View.h>
#include <AK/Utf8View.h>
@ -269,12 +269,10 @@ JS_DEFINE_NATIVE_FUNCTION(GlobalObject::parse_float)
// 6. Assert: parsedNumber is a Parse Node.
// 7. Return StringNumericValue of parsedNumber.
auto trimmed_string_view = trimmed_string.bytes_as_string_view();
auto const* begin = trimmed_string_view.characters_without_null_termination();
auto const* end = begin + trimmed_string_view.length();
auto parsed_number = parse_first_floating_point<double>(begin, end);
if (parsed_number.parsed_value())
return parsed_number.value;
auto parsed_number = AK::parse_first_number<double>(trimmed_string_view, TrimWhitespace::No);
if (parsed_number.has_value())
return parsed_number->value;
auto first_code_point = *trimmed_string.code_points().begin();
if (first_code_point == '-' || first_code_point == '+')

View file

@ -5,6 +5,7 @@
*/
#include <AK/StringBuilder.h>
#include <AK/StringConversions.h>
#include <AK/StringUtils.h>
#include <LibJS/Runtime/AbstractOperations.h>
#include <LibJS/Runtime/TypedArray.h>
@ -502,7 +503,7 @@ DecodeResult from_hex(VM& vm, StringView string, Optional<size_t> max_length)
// d. Let byte be the integer value represented by hexits in base-16 notation, using the letters A-F and a-f
// for digits with values 10 through 15.
// NOTE: We do this early so that we don't have to effectively parse hexits twice.
auto byte = AK::StringUtils::convert_to_uint_from_hex<u8>(hexits, AK::TrimWhitespace::No);
auto byte = AK::parse_hexadecimal_number<u8>(hexits, TrimWhitespace::No);
// b. If hexits contains any code units which are not in "0123456789abcdefABCDEF", then
if (!byte.has_value()) {

View file

@ -10,7 +10,6 @@
#include <AK/Assertions.h>
#include <AK/ByteString.h>
#include <AK/CharacterTypes.h>
#include <AK/FloatingPointStringConversions.h>
#include <AK/StringBuilder.h>
#include <AK/StringFloatingPointConversions.h>
#include <AK/Utf8View.h>

View file

@ -8,7 +8,6 @@
#include "Token.h"
#include <AK/Assertions.h>
#include <AK/CharacterTypes.h>
#include <AK/FloatingPointStringConversions.h>
#include <AK/GenericLexer.h>
#include <AK/StringBuilder.h>