Make LChar a distinct type from uint8_t so it can imply character encoding as char8/16/32_t do #48579
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
EWS run on previous version of this PR (hash 3ac13b1) |
🛠 wpe
darinadler
changed the title
darinadler
force-pushed
the
eng/Make-LChar-a-distinct-type-from-uint8_t-so-it-can-imply-character-encoding-as-char8-16-32_t-do
branch
from
3ac13b1 to
2f75985
Compare
|
EWS run on current version of this PR (hash 2f75985) |
|
EWS run on previous version of this PR (hash 2f75985) |
webkit-ews-buildbot
added
the
merging-blocked
darinadler
removed
the
merging-blocked
darinadler
force-pushed
the
eng/Make-LChar-a-distinct-type-from-uint8_t-so-it-can-imply-character-encoding-as-char8-16-32_t-do
branch
from
2f75985 to
bb9dd3c
Compare
|
EWS run on previous version of this PR (hash bb9dd3c) |
webkit-ews-buildbot
added
the
merging-blocked
darinadler
removed
the
merging-blocked
darinadler
force-pushed
the
eng/Make-LChar-a-distinct-type-from-uint8_t-so-it-can-imply-character-encoding-as-char8-16-32_t-do
branch
from
bb9dd3c to
0aa73b0
Compare
|
EWS run on previous version of this PR (hash 0aa73b0) |
webkit-ews-buildbot
added
the
merging-blocked
darinadler
removed
the
merging-blocked
darinadler
force-pushed
the
eng/Make-LChar-a-distinct-type-from-uint8_t-so-it-can-imply-character-encoding-as-char8-16-32_t-do
branch
from
0aa73b0 to
f550a51
Compare
|
EWS run on previous version of this PR (hash f550a51)
|
webkit-ews-buildbot
added
the
merging-blocked
darinadler
removed
the
merging-blocked
darinadler
force-pushed
the
eng/Make-LChar-a-distinct-type-from-uint8_t-so-it-can-imply-character-encoding-as-char8-16-32_t-do
branch
from
f550a51 to
f02d9e5
Compare
|
EWS run on previous version of this PR (hash f02d9e5)
|
webkit-ews-buildbot
added
the
merging-blocked
darinadler
removed
the
merging-blocked
darinadler
force-pushed
the
eng/Make-LChar-a-distinct-type-from-uint8_t-so-it-can-imply-character-encoding-as-char8-16-32_t-do
branch
from
f02d9e5 to
4903612
Compare
|
EWS run on previous version of this PR (hash 4903612) |
webkit-ews-buildbot
added
the
merging-blocked
darinadler
removed
the
merging-blocked
darinadler
force-pushed
the
eng/Make-LChar-a-distinct-type-from-uint8_t-so-it-can-imply-character-encoding-as-char8-16-32_t-do
branch
from
4903612 to
bfbf1fa
Compare
|
EWS run on previous version of this PR (hash bfbf1fa) |
darinadler
commented
Sep 14, 2025
cdumez
approved these changes
Sep 14, 2025
| } | ||
| if (frame.payload.size() >= 3) | ||
| m_closeEventReason = String::fromUTF8({ &frame.payload[2], frame.payload.size() - 2 }); | ||
| m_closeEventReason = String::fromUTF8(byteCast<char8_t>(frame.payload.subspan(2))); |
There was a problem hiding this comment.
We should no longer need ::fromUTF8() here right?
| static void createTestFile(const String& path) | ||
| { | ||
| auto written = FileSystem::overwriteEntireFile(path, FileSystemTestData.span8()); | ||
| auto written = FileSystem::overwriteEntireFile(path, byteCast<uint8_t>(FileSystemTestData.span8())); |
There was a problem hiding this comment.
I guess in the future we'll want to use std::byte for this sort of thing?
There was a problem hiding this comment.
Yes, I think that's right.
We also will probably want to rethink names like span8 since that could mean any 8-bit type. In the context of a String it can mean LChar but elsewhere maybe not.
There was a problem hiding this comment.
Maybe we want to use span<T>() with a template parameter for clarity in the future instead of span8() / span16().
There was a problem hiding this comment.
Yes, I think it’s really cool to have the actual type of the span be there at the call site. The implementation doesn’t really have to be generic, can just be the two functions but use template syntax.
| { | ||
| constexpr LChar buffer[] = "hello"; | ||
| constexpr size_t bufferStringLength = sizeof(buffer) - 1; | ||
| static constinit char buffer[] = "hello"; |
| constexpr LChar buffer[] = "hello"; | ||
| constexpr size_t bufferStringLength = sizeof(buffer) - 1; | ||
| static constinit char buffer[] = "hello"; | ||
| static constinit size_t bufferStringLength = sizeof(buffer) - 1; |
weinig
reviewed
Sep 15, 2025
Source/WTF/wtf/HexNumber.cpp
Outdated
| static const std::array<LChar, 16>& hexDigitsForMode(HexConversionMode mode) | ||
| { | ||
| static constinit std::array<LChar, 16> lowercaseHexDigits { '0', '1', '2', '3', '4', '5', '6', '7', '8', '9', 'a', 'b', 'c', 'd', 'e', 'f' }; | ||
| static constinit std::array<LChar, 16> uppercaseHexDigits { '0', '1', '2', '3', '4', '5', '6', '7', '8', '9', 'A', 'B', 'C', 'D', 'E', 'F' }; |
There was a problem hiding this comment.
These should stay constexpr. constinit really should only be used when constant global initialization of a then mutable value is needed.
weinig
reviewed
Sep 15, 2025
| } | ||
|
|
||
| template<typename CharacterType1, typename CharacterType2, std::enable_if_t<std::is_integral_v<CharacterType1> && std::is_integral_v<CharacterType2>>* = nullptr> | ||
| template<IsFindableCharacter CharacterType1, IsFindableCharacter CharacterType2> |
There was a problem hiding this comment.
Loving the adopting of concepts here. Much more clear!
weinig
reviewed
Sep 15, 2025
| } | ||
|
|
||
| // Construct a string with Latin-1 data. | ||
| String::String(std::span<const char> characters) |
There was a problem hiding this comment.
Do we need this one? Ideally we would get rid of it and make callers be explicit here.
weinig
reviewed
Sep 15, 2025
| String json { buffer.span() }; | ||
|
|
||
| auto value = JSON::Value::parseJSON(json); | ||
| auto value = JSON::Value::parseJSON(byteCast<LChar>(buffer.span())); |
There was a problem hiding this comment.
Not new here, but worth a FIXME. The spec link above indicates this should be interpreted as UTF-8.
weinig
reviewed
Sep 15, 2025
| // Check if the third message is a multi-lines string, concatenating such message would look ugly in log events. | ||
| if (values.size() >= 3 && values[2].value.find("\r\n"_s) != notFound) | ||
| event = generateJSONLogEvent(MessageLogEvent { values[1].value, { values[2].value.span8() } }, false); | ||
| event = generateJSONLogEvent(MessageLogEvent { values[1].value, { byteCast<uint8_t>(values[2].value.span8()) } }, false); |
There was a problem hiding this comment.
Not new, and necessary to change now, but the name span8() leaves a lot to be desired (probably should call it something like spanLatin1()). I think using the helper span<LChar>() might help here, though I wonder why it is safe to assume this string is latin-1.
weinig
reviewed
Sep 15, 2025
| return false; | ||
| } | ||
|
|
||
| const CString headerValueData = headerValue.utf8(); |
There was a problem hiding this comment.
At some point we need to replace CString with explicit UTF8String and Latin1String classes and remove the need for this kind of casting.
weinig
reviewed
Sep 15, 2025
| return input; | ||
| return makeString(input.first(maxInputSampleSize), horizontalEllipsis); | ||
| return byteCast<LChar>(input); | ||
| return makeString(byteCast<LChar>(input.first(maxInputSampleSize)), horizontalEllipsis); |
There was a problem hiding this comment.
Why is LChar right here, as opposed to UTF-8? (not new, just if you know, might be worth a comment). Same below.
weinig
reviewed
Sep 15, 2025
|
|
||
| StringView statusCodeString(header.subspan(*firstSpaceIndex + 1, *secondSpaceIndex - *firstSpaceIndex - 1)); | ||
| StringView statusCodeString(byteCast<LChar>(header.subspan(*firstSpaceIndex + 1, *secondSpaceIndex - *firstSpaceIndex - 1))); | ||
| if (statusCodeString.length() != 3) // Status code must consist of three digits. |
There was a problem hiding this comment.
Is the StringView doing anything useful here other than taking a few branches? (Explicit use of String::length() or StringView::length() like this always makes me suspicious).
weinig
reviewed
Sep 15, 2025
| String jwkString(bytes.span()); | ||
| JSLockHolder locker(vm); | ||
| auto jwkObject = JSONParse(&state, jwkString); | ||
| auto jwkObject = JSONParse(&state, byteCast<LChar>(bytes.span())); |
There was a problem hiding this comment.
Another case where this should probably be UTF-8.
weinig
reviewed
Sep 15, 2025
| Vector<uint8_t> bodyData("body="_span); | ||
| FormDataBuilder::encodeStringAsFormData(bodyData, body.utf8()); | ||
| body = makeStringByReplacingAll(bodyData.span(), '+', "%20"_s); | ||
| body = makeStringByReplacingAll(byteCast<LChar>(bodyData.span()), '+', "%20"_s); |
There was a problem hiding this comment.
This is confusing, but I guess FormDataBuilder::encodeStringAsFormData() converts the utf-8 body into ascii or latin-1?
Should we make FormDataBuilder::encodeStringAsFormData just take a Vector<LChar> then?
weinig
approved these changes
Sep 15, 2025
There was a problem hiding this comment.
This has been fascinating to review. My major takeaways for future work here:
-
Most uses of
byteCast<LChar>probably need to looked at after this lands and checked to make sure they are right. While many are probably fine (passing along some base64 encoded string), I am not super confident in quite a few. The explicitness here is great! -
Would love for us to adopt
std::byteinstead ofuint8_t, just to make it more clear (uint8_tlooks a bit to close tochar8_tfor me). -
Would love to use long names for these character types, even the standard ones.
LChar->Latin1Characterchar8_t-> UTF8Character`char16_t-> UTF16Character`
We have this historical exception for this type using abbreviations, but I don't think it warrants it.
geoffreygaren
approved these changes
Sep 16, 2025
|
I like the proposal to rename Using longer custom names for |
I don't know, |
darinadler
force-pushed
the
eng/Make-LChar-a-distinct-type-from-uint8_t-so-it-can-imply-character-encoding-as-char8-16-32_t-do
branch
from
3912a64 to
3a139dd
Compare
darinadler
requested review from
b-weinstein,
gsnedders and
xeenon
as code owners
|
EWS run on current version of this PR (hash 3a139dd) |
darinadler
commented
Sep 20, 2025
| String jsonData = String::fromUTF8(data); | ||
|
|
||
| auto messageValue = JSON::Value::parseJSON(jsonData); | ||
| auto messageValue = JSON::Value::parseJSON(String { byteCast<char8_t>(data.span()) }); |
There was a problem hiding this comment.
Just realized I could have simplified this further:
Suggested change
| auto messageValue = JSON::Value::parseJSON(String { byteCast<char8_t>(data.span()) }); | |
| auto messageValue = JSON::Value::parseJSON(byteCast<char8_t>(data.span())); |
…oding as char8/16/32_t do https://bugs.webkit.org/show_bug.cgi?id=296539 rdar://156856072 Reviewed by Chris Dumez, Sam Weinig, and Geoffrey Garen. LChar is now a struct so it can be a distinct type, so LChar means Latin-1, char8_t means UTF-8, and uint8_t and char remain ambiguous about encoding. Tried to mostly stay with the minimum to get things compiling, without a lot of "cleanup". As part of this made it possible to construct String directly from std::span<char8_t> without having to utter String::fromUTF8 since the type is unambiguous. We should follow up by removing more overloads and functions that interpret uint8_t, char, or even std::byte as particular encodings, and use byteCast to make encoding explicit. * Source/JavaScriptCore/API/JSScript.mm: (+[JSScript scriptOfType:memoryMappedFromASCIIFile:withSourceURL:andBytecodeCache:inVirtualMachine:error:]): Cast to LChar. * Source/JavaScriptCore/API/JSStringRefCF.cpp: (JSStringCreateWithCFString): Cast to UInt8. (JSStringCopyCFString): Ditto. * Source/JavaScriptCore/Scripts/xxd.pl: Use constexpr std::array<LChar> for C++. * Source/JavaScriptCore/heap/HeapSnapshotBuilder.cpp: (JSC::HeapSnapshotBuilder::json): Rely on serialization of enum as a numeral, rather than explicitly calling edgeTypeToNumber. * Source/JavaScriptCore/inspector/remote/socket/RemoteInspectorConnectionClient.cpp: (Inspector::RemoteInspectorConnectionClient::extractEvent): Cast to char8_t. * Source/JavaScriptCore/inspector/remote/socket/RemoteInspectorSocket.cpp: (Inspector::RemoteInspector::backendCommands const): Eliminate use of String::adopt. It doesn't really work for vectors any more, and likely we should remove it to avoid making a promise we can't keep. It doesn't work with byteCast, which is why we need to do this here now. * Source/JavaScriptCore/parser/Lexer.cpp: (JSC::Lexer<T>::peek const): Tweak so the conditional operator compiles. (JSC::Lexer<T>::parseString): Use SIMD::SizedUnsigned. (JSC::Lexer<T>::parseStringSlowCase): Removed a static assertion that's not super important and a bit tricky to write given that LChar is no longer a scalar. (JSC::Lexer<T>::lexWithoutClearingLineTerminator): Use SIMD::SizedUnsigned. * Source/JavaScriptCore/runtime/ISO8601.h: Cast to char. * Source/JavaScriptCore/runtime/IntlObject.cpp: (JSC::parseVariantCode): Removed a static assertion that's not super important and a bit tricky to write given that LChar is no longer a scalar. * Source/JavaScriptCore/runtime/JSGenericTypedArrayViewPrototype.cpp: (JSC::uint8ArrayPrototypeToHex): Cast to uint8_t. * Source/JavaScriptCore/runtime/JSONObject.cpp: (JSC::stringCopySameType): Use SIMD::SizedUnsigned. (JSC::stringCopyUpconvert): Use uint8_t directly. * Source/JavaScriptCore/runtime/LiteralParser.cpp: (JSC::reviverMode>::Lexer::lexString): Reduce mixing char with LChar a bit. Also use SIMD::SizedUnsigned. * Source/WTF/wtf/HexNumber.cpp: (WTF::Internal::hexDigitsForMode): Moved this here from the header since it's only used here. * Source/WTF/wtf/HexNumber.h: (WTF::Internal::appendHex): Added an overload that takes LChar and forwards all the other arguments. * Source/WTF/wtf/JSONValues.cpp: Use char16_t for local since that works with a switch statement, LChar would not since it's a struct rather than a scalar. * Source/WTF/wtf/PrintStream.h: (WTF::printInternal): Added char8_t overload. * Source/WTF/wtf/SIMDHelpers.h: Added SizedUnsigned, which works for LChar. (WTF::SIMD::find): Use SizedUnsigned. (WTF::SIMD::count): Ditto. * Source/WTF/wtf/SortedArrayMap.h: (WTF::foldForComparison): Use SIMD::SizedUnsigned. * Source/WTF/wtf/StdLibExtras.h: Update ByteType to work with LChar. Also tweaked it a little and renamed it IsByte and IsMutableByte. (WTF::ByteCastTraits<T>::cast): Ditto. (WTF::ByteCastTraits<T::cast): Ditto. (WTF::byteCast): Ditto. * Source/WTF/wtf/URLParser.cpp: (WTF::URLParser::appendNumberToASCIIBuffer): Cast to char. (WTF::URLParser::formURLDecode): Cast to char8_t. * Source/WTF/wtf/cf/CFURLExtras.cpp: (WTF::bytesAsString): Cast to UInt8. (WTF::isSameOrigin): Cast to LChar. * Source/WTF/wtf/cf/URLCF.cpp: (WTF::URL::createCFURL): Cast to UInt8. * Source/WTF/wtf/cocoa/NSURLExtras.mm: (WTF::userVisibleString): Cast to LChar. * Source/WTF/wtf/cocoa/SpanCocoa.h: (WTF::toNSData): Added overloads to make this work with all IsByte types. (WTF::toNSDataNoCopy): Ditto. * Source/WTF/wtf/persistence/PersistentCoders.cpp: (WTF::Persistence::Coder<String>::encodeForPersistence): Use asBytes to convert result of span8 into bytes, not LChar. * Source/WTF/wtf/persistence/PersistentDecoder.h: (WTF::Persistence::Decoder::bufferIsLargeEnoughToContain const): Update assertion to allow LChar. * Source/WTF/wtf/text/ASCIIFastPath.h: Redid the NonASCIIMask and NonLatin1Mask to not rely on specific types for CharacterType and depend on the size of the type instead. * Source/WTF/wtf/text/Base64.cpp: (WTF::base64DecodeInternal): Take the vector type as the template parameter instead of just the Malloc. Simplifies things a bit. (WTF::base64DecodeToString): Updated for the above, also simplified by removing the lambda since it reads well without it. * Source/WTF/wtf/text/IntegerToStringConversion.h: (WTF::numberToStringImpl): Cast to char. (WTF::writeIntegerToBufferImpl): Ditto. * Source/WTF/wtf/text/LChar.h: Make LChar a struct in the WTF namespace. Added constexpr functions to smooth the use of this in code that treats it as an integer. Added IsStringStorageCharacter concept so we can write templates that work with LChar and char16_t without accidentally allowing other types as well. * Source/WTF/wtf/text/ParsingUtilities.h: Added an overload of skipWhile for char8_t. * Source/WTF/wtf/text/StringBuilder.h: (WTF::StringBuilder::append): Added an overload for uint8_t, needed since there is code that depends on single argument append treating it as a character and other code that depends on variadic append treating it as a numeral! (WTF::StringBuilder::operator[] const): Tweak so the conditional operator compiles. * Source/WTF/wtf/text/StringCommon.h: Added an IsFindableCharacter concept so function templates compile for the correct types including LChar. Use SIMD::SizedUnsigned. (WTF::equalLettersIgnoringASCIICaseWithLength): (WTF::compareEach): Added a cast to int to silence a compiler warning about our intentional use of bitwise with booleans. * Source/WTF/wtf/text/StringHasher.h: (WTF::StringHasher::DefaultConverter::convert): Fixed this so it works with LChar. * Source/WTF/wtf/text/StringImpl.cpp: (WTF::StringImpl::create): Added overload for char8_t spans. * Source/WTF/wtf/text/StringImpl.h: Use IsStringStorageCharacter to make templates type check a bit better. (WTF::StringImpl::create): Added a new create function for UTF-8 strings to move the logic from WTF::String to here. (WTF::StringImpl::copyCharacters): Tweak to work with LChar. (WTF::StringImpl::tryCreateUninitialized): (WTF::StringImpl::at const): Tweak so the conditional operator compiles. (WTF::StringImpl::createByReplacingInCharacters): Ditto. * Source/WTF/wtf/text/WTFString.cpp: (WTF::String::String): Added a constructor that takes a std::span<const char8_t> for use instead of String::fromUTF8. Use StringImpl::create. (WTF::String::ascii const): Tweak so the conditional operator compiles. (WTF::String::fromUTF8ReplacingInvalidSequences): Use StringImpl::create. (WTF::fromUTF8Impl): Deleted. Uses StringImpl instead. (WTF::String::fromUTF8): Deleted the overload that takes const char8_t; instead the template allows any byte type. * Source/WTF/wtf/text/WTFString.h: Added the new constructor for span<char8_t> and rearranged from functions so they work with more types. Since the explicitly state the encoding, we can allow them to take any byte type. * Source/WTF/wtf/text/cf/StringCF.cpp: (WTF::String::String): Cast to UInt8. * Source/WTF/wtf/text/cf/StringImplCF.cpp: (WTF::StringImpl::createCFString): Ditto. * Source/WTF/wtf/text/cf/StringViewCF.cpp: (WTF::StringView::createCFString const): Ditto. (WTF::StringView::createCFStringWithoutCopying const): Ditto. * Source/WebCore/DerivedSources-input.xcfilelist: Updated for scripts from JavaScriptCore that were always used but are now dependencies. * Source/WebCore/DerivedSources.make: Added missing dependencies in the rules that produce XMLViewerCSS.h and XMLViewerJS.h so they are regenerated if the scripts that produce them change, as with this patch that changes xxd.pl. * Source/WebCore/Modules/encryptedmedia/InitDataRegistry.cpp: (WebCore::extractKeyIDsKeyids): Cast to LChar to pass to parseJSON and remove the unnecessary copy into a temporary String. * Source/WebCore/Modules/mediastream/PeerConnectionBackend.cpp: (WebCore::PeerConnectionBackend::handleLogMessage): Cast to uint8_t. * Source/WebCore/Modules/mediastream/RTCRtpSFrameTransformerCocoa.cpp: (WebCore::RTCRtpSFrameTransformer::computeSaltKey): Ditto. (WebCore::createBaseSFrameKey): Ditto. (WebCore::RTCRtpSFrameTransformer::computeAuthenticationKey): Ditto. (WebCore::RTCRtpSFrameTransformer::computeEncryptionKey): Ditto. * Source/WebCore/Modules/mediastream/gstreamer/GStreamerDtlsTransportBackend.cpp: (WebCore::GStreamerDtlsTransportBackendObserver::stateChanged): Ditto. * Source/WebCore/Modules/push-api/PushMessageCrypto.cpp: (WebCore::PushCrypto::decryptAES128GCMPayload): Ditto. (WebCore::PushCrypto::decryptAESGCMPayload): Ditto. * Source/WebCore/Modules/url-pattern/URLPatternParser.cpp: (WebCore::URLPatternUtilities::escapeRegexStringForCharacters): Specify the type and size of the array to make it compile. (WebCore::URLPatternUtilities::escapePatternStringForCharacters): Ditto. * Source/WebCore/Modules/websockets/WebSocketExtensionDispatcher.cpp: (WebCore::WebSocketExtensionDispatcher::processHeaderValue): Use char8_t since the code currently parses a UTF-8 representation of the header value. * Source/WebCore/Modules/websockets/WebSocketExtensionParser.cpp: (WebCore::isSpaceOrTab): Use char8_t, for now at least. * Source/WebCore/Modules/websockets/WebSocketExtensionParser.h: Ditto. * Source/WebCore/Modules/websockets/WebSocketHandshake.cpp: (WebCore::trimInputSample): Cast to LChar. (WebCore::WebSocketHandshake::readStatusLine): Ditto. * Source/WebCore/PAL/PAL.xcodeproj/project.pbxproj: Removed Gunzip.cpp/h. * Source/WebCore/PAL/pal/CMakeLists.txt: Removed Gunzip.h * Source/WebCore/PAL/pal/Gunzip.h: Removed. * Source/WebCore/PAL/pal/PlatformMac.cmake: Removed Gunzip.cpp. * Source/WebCore/PAL/pal/cocoa/Gunzip.cpp: Removed. * Source/WebCore/bindings/js/ScriptBufferSourceProvider.h: Cast to LChar. * Source/WebCore/bindings/js/SerializedScriptValue.cpp: (WebCore::CloneDeserializer::readString): Ditto. * Source/WebCore/contentextensions/DFABytecodeInterpreter.cpp: (WebCore::ContentExtensions::DFABytecodeInterpreter::interpretJumpTable): Cast to char so the conditional operator compiles. (WebCore::ContentExtensions::DFABytecodeInterpreter::interpret): DIto. * Source/WebCore/crypto/SubtleCrypto.cpp: (WebCore::SubtleCrypto::unwrapKey): Cast to LChar to pass to JSONParse and remove the unnecessary copy into a temporary String. * Source/WebCore/editing/cocoa/WebContentReaderCocoa.mm: (WebCore::replaceRichContentWithAttachments): Cast to LChar. * Source/WebCore/fileapi/FileReaderLoader.cpp: (WebCore::FileReaderLoader::stringResult): Ditto. * Source/WebCore/html/FTPDirectoryDocument.cpp: (WebCore::FTPDirectoryDocumentParser::loadDocumentTemplate): Ditto. * Source/WebCore/html/parser/HTMLEntityParser.cpp: (WebCore::StringParsingBufferSource::currentCharacter const): Tweak so the conditional operator compiles. * Source/WebCore/html/track/VTTScanner.h: (WebCore::VTTScanner::currentChar const): Ditto. * Source/WebCore/html/track/WebVTTParser.cpp: (WebCore::WebVTTParser::fileFinished): Cast to uint8_t. * Source/WebCore/loader/FTPDirectoryParser.cpp: (WebCore::parseOneFTPLine): Cast to LChar. * Source/WebCore/loader/FormSubmission.cpp: (WebCore::appendMailtoPostFormDataToURL): Cast to LChar. (WebCore::FormSubmission::create): Ditto. * Source/WebCore/loader/TextResourceDecoder.cpp: (WebCore::findXMLEncoding): Cast to uint8_t. (WebCore::TextResourceDecoder::checkForCSSCharset): Cast to uint8_t and LChar. (WebCore::TextResourceDecoder::checkForHeadCharset): Ditto. * Source/WebCore/loader/cache/CachedScript.cpp: (WebCore::CachedScript::script): Cast to LChar. (WebCore::CachedScript::codeBlockHashConcurrently): Ditto. * Source/WebCore/platform/encryptedmedia/CDMUtilities.cpp: (WebCore::CDMUtilities::parseJSONObject): Cast to LChar to pass to parseJSON and remove the unnecessary copy into a temporary String. * Source/WebCore/platform/graphics/avfoundation/CDMFairPlayStreaming.cpp: (WebCore::extractSinfData): Ditto. (WebCore::CDMPrivateFairPlayStreaming::extractKeyIDsMpts): Ditto. * Source/WebCore/platform/graphics/avfoundation/objc/CDMInstanceFairPlayStreamingAVFObjC.mm: (WebCore::parseJSONValue): Cast to LChar. * Source/WebCore/platform/graphics/freetype/FontCacheFreeType.cpp: (WebCore::fontNameMapName): Cast to LChar. * Source/WebCore/platform/graphics/gstreamer/eme/CDMThunder.cpp: (WebCore::ParsedResponseMessage::ParsedResponseMessage): Cast to LChar. (WebCore::CDMInstanceSessionThunder::loadSession): Ditto. * Source/WebCore/platform/graphics/iso/ISOVTTCue.cpp: Use char8_t for the buffer we are treating as UTF-8. * Source/WebCore/platform/gstreamer/GStreamerElementHarness.cpp: (WebCore::MermaidBuilder::span const): Cast to uint8_t. * Source/WebCore/platform/image-decoders/png/PNGImageDecoder.cpp: (WebCore::decodingWarning): Cast to char. (WebCore::PNGImageDecoder::readChunks): Ditto. * Source/WebCore/platform/mediarecorder/MediaRecorderPrivateMock.cpp: (WebCore::MediaRecorderPrivateMock::fetchData): Cast to uint8_t. * Source/WebCore/platform/network/HTTPParsers.cpp: (WebCore::trimInputSample): Make into a non-template function. (WebCore::parseHTTPHeader): Cast to LChar. * Source/WebCore/platform/network/curl/OpenSSLHelper.cpp: (OpenSSL::BIO::getDataAsString const): Cast to LChar. (OpenSSL::toString): Ditto. * Source/WebCore/platform/network/soup/CertificateInfoSoup.cpp: (WebCore::CertificateInfo::summary const): Ditto. * Source/WebCore/platform/text/SegmentedString.h: (WebCore::SegmentedString::Substring::currentCharacter const): Tweak to compile conditional operator. * Source/WebCore/rendering/BreakLines.h: (WebCore::BreakLines::nextBreakablePosition): Switch to use a simpler scoped CharacterInfo struct instead of a struct template with conversions. * Source/WebCore/testing/MockCDMFactory.cpp: (WebCore::MockCDM::sanitizeResponse const): Cast to LChar. (WebCore::MockCDMInstance::setServerCertificate): Ditto. (WebCore::MockCDMInstanceSession::updateLicense): Ditto. * Source/WebGPU/WGSL/Lexer.cpp: (WGSL::Lexer<CharacterType>::makeToken): Moved from header. (WGSL::Lexer<CharacterType>::makeFloatToken): Ditto. (WGSL::Lexer<CharacterType>::makeIntegerToken): Ditto. (WGSL::Lexer<CharacterType>::makeIdentifierToken): Ditto. * Source/WebGPU/WGSL/Lexer.h: Simplified constructor to take a span instead of a String, and moved some private function implementations out of the header. * Source/WebGPU/WGSL/Parser.cpp: (WGSL::parse): Changed to take CharacterType as the template argument rather than the Lexer type, then we can pass a span into the Lexer instead of a String. * Source/WebGPU/WGSL/ParserPrivate.h: Added include. * Source/WebGPU/WebGPU/Pipeline.mm: (WebKit::printToFileForPsoRepro): Cast to uint8_t. * Source/WebKit/NetworkProcess/cache/NetworkCache.cpp: (WebKit::NetworkCache::Cache::dumpContentsToFile): Ditto. * Source/WebKit/NetworkProcess/curl/WebSocketTaskCurl.cpp: (WebKit::WebSocketTask::didReceiveData): Cast to char8_t. Simplify some code that is extracting bytes. (WebKit::WebSocketTask::sendClosingHandshakeIfNeeded): Ditto. * Source/WebKit/NetworkProcess/storage/CacheStorageManager.cpp: (WebKit::readSizeFile): Cast to LChar. * Source/WebKit/Platform/IPC/DaemonCoders.h: (WebKit::Daemon::Coder<WTF::String>::encode): Cast to uint8_t. * Source/WebKit/Platform/IPC/DaemonDecoder.h: (WebKit::Daemon::Decoder::bufferIsLargeEnoughToContain const): Update assertion to allow LChar. * Source/WebKit/Shared/API/c/cf/WKStringCF.mm: (WKStringCopyCFString): Cast to UInt8. * Source/WebKit/Shared/Cocoa/SandboxExtensionCocoa.mm: (WebKit::SandboxExtensionImpl::SandboxExtensionImpl): Cast to LChar. * Source/WebKit/UIProcess/API/APIContentRuleListStore.cpp: (API::getContentRuleListSourceFromMappedFile): Cast to LChar. * Source/WebKit/UIProcess/API/C/WKPage.cpp: (dataFrom): Cast to uint8_t. * Source/WebKit/UIProcess/Cocoa/WebPasteboardProxyCocoa.mm: (WebKit::WebPasteboardProxy::testIPCSharedMemory): Cast to LChar. * Source/WebKit/UIProcess/Inspector/glib/RemoteInspectorClient.cpp: (WebKit::RemoteInspectorClient::setBackendCommands): Cast to std::byte. * Source/WebKit/UIProcess/Inspector/mac/RemoteWebInspectorUIProxyMac.mm: (WebKit::RemoteWebInspectorUIProxy::platformLoad): Cast to LChar. * Source/WebKit/UIProcess/Inspector/mac/WebInspectorUIProxyMac.mm: (WebKit::WebInspectorUIProxy::platformLoad): Cast to LChar. * Source/WebKit/UIProcess/wpe/WebPasteboardProxyWPE.cpp: (WebKit::WebPasteboardProxy::readURLFromPasteboard): Cast to LChar. * Source/WebKit/WebProcess/Network/webrtc/RTCDataChannelRemoteManager.cpp: (WebKit::RTCDataChannelRemoteManager::sendData): Cast to LChar. * Source/WebKit/WebProcess/cocoa/WebProcessCocoa.mm: (WebKit::registerLogClient): Cast to uint8_t. * Source/WebKitLegacy/WebCoreSupport/WebSocketChannel.cpp: (WebCore::WebSocketChannel::processFrame): Cast to char8_t. * Tools/Scripts/webkitpy/style/checkers/cpp.py: (check_spacing): Add missing operator| check. * Tools/TestWebKitAPI/Tests/WGSL/ConstLiteralTests.cpp: (parseLCharPrimaryExpression): Use ASCIILiteral. * Tools/TestWebKitAPI/Tests/WGSL/LexerTests.cpp: (TestWGSLAPI::TestLexer::TestLexer): Use ASCIILiteral. (TestWGSLAPI::checkSingleToken): Ditto. (TestWGSLAPI::checkSingleIntegerLiteral): Ditto. (TestWGSLAPI::checkSingleFloatLiteral): Ditto. * Tools/TestWebKitAPI/Tests/WTF/Base64.cpp: (TestWebKitAPI::TEST(Base64, Encode)): Cast to uint8_t. (TestWebKitAPI::TEST(Base64, EncodeOmitPadding)): Ditto. (TestWebKitAPI::TEST(Base64, EncodeURL)): Ditto. (TestWebKitAPI::TEST(Base64, EncodeURLOmitPadding)): Ditto. * Tools/TestWebKitAPI/Tests/WTF/FileSystem.cpp: (TestWebKitAPI::createTestFile): Cast to uint8_t. (TestWebKitAPI::TEST_F(FileSystemTest, openExistingFileTruncate)): Ditto. (TestWebKitAPI::TEST_F(FileSystemTest, openExistingFileReadWrite)): Ditto. (TestWebKitAPI::TEST_F(FileSystemTest, deleteEmptyDirectoryContainingDSStoreFile)): Ditto. (TestWebKitAPI::TEST_F(FileSystemTest, deleteEmptyDirectoryOnNonEmptyDirectory)): Ditto. (TestWebKitAPI::TEST_F(FileSystemTest, moveDirectory)): Ditto. (TestWebKitAPI::runGetFileModificationTimeTest): Ditto. (TestWebKitAPI::TEST_F(FileSystemTest, readEntireFile)): Ditto. * Tools/TestWebKitAPI/Tests/WTF/StringImpl.cpp: (TestWebKitAPI::TEST(WTF, ExternalStringImplCreate8bit)): Use char and cast to LChar. (TestWebKitAPI::TEST(WTF, ExternalStringAtom)): Ditto. * Tools/TestWebKitAPI/Tests/WTF/StringView.cpp: (TestWebKitAPI::TEST(WTF, StringViewEqualIgnoringASCIICaseWithLatin1Characters)): Use byteCast instead of reinterpret_cast. * Tools/TestWebKitAPI/Tests/WTF/cocoa/URLExtras.mm: (TestWebKitAPI::dataAsString): Pass a character instead of an int. * Tools/TestWebKitAPI/Tests/WebCore/FileMonitor.cpp: (TestWebKitAPI::readContentsOfFile): Cast to LChar. * Tools/TestWebKitAPI/Tests/WebCore/PushMessageCrypto.cpp: (TestWebKitAPI::TEST(PushMessageCrypto, AES128GCMPayloadWithMinimalPadding)): Eliminate use of String::adopt. (TestWebKitAPI::TEST(PushMessageCrypto, AES128GCMPayloadWithPadding)): Ditto. (TestWebKitAPI::TEST(PushMessageCrypto, AESGCMPayloadWithMinimalPadding)): Ditto. (TestWebKitAPI::TEST(PushMessageCrypto, AESGCMPayloadWithPadding)): Ditto. * Tools/TestWebKitAPI/Tests/WebCore/SharedBuffer.cpp: (TestWebKitAPI::TEST_F(FragmentedSharedBufferTest, createWithContentsOfExistingFile)): Cast to LChar. (TestWebKitAPI::TEST_F(FragmentedSharedBufferTest, read)): Ditto. (TestWebKitAPI::TEST_F(SharedBufferChunkReaderTest, includeSeparator)): Do not mix uint8_t with LChar. (TestWebKitAPI::TEST_F(SharedBufferChunkReaderTest, peekData)): Cast to LChar. * Tools/TestWebKitAPI/Tests/WebCore/SharedBufferTest.cpp: (TestWebKitAPI::FragmentedSharedBufferTest::SetUp): Cast to uint8_t. * Tools/TestWebKitAPI/Tests/WebCore/curl/CurlMultipartHandleTests.cpp: (TestWebKitAPI::Curl::TEST(CurlMultipartHandleTests, SimpleMessage)): Ditto. (TestWebKitAPI::Curl::TEST(CurlMultipartHandleTests, NoHeader)): Ditto. (TestWebKitAPI::Curl::TEST(CurlMultipartHandleTests, NoBody)): Ditto. (TestWebKitAPI::Curl::TEST(CurlMultipartHandleTests, TransportPadding)): Ditto. (TestWebKitAPI::Curl::TEST(CurlMultipartHandleTests, NoEndOfBoundary)): Ditto. (TestWebKitAPI::Curl::TEST(CurlMultipartHandleTests, NoEndOfBoundaryAfterCompleted)): Ditto. (TestWebKitAPI::Curl::TEST(CurlMultipartHandleTests, NoCloseDelimiter)): Ditto. (TestWebKitAPI::Curl::TEST(CurlMultipartHandleTests, NoCloseDelimiterAfterCompleted)): Ditto. (TestWebKitAPI::Curl::TEST(CurlMultipartHandleTests, CloseDelimiter)): Ditto. (TestWebKitAPI::Curl::TEST(CurlMultipartHandleTests, CloseDelimiterAfterCompleted)): Ditto. (TestWebKitAPI::Curl::TEST(CurlMultipartHandleTests, DivideFirstDelimiter)): Ditto. (TestWebKitAPI::Curl::TEST(CurlMultipartHandleTests, DivideSecondDelimiter)): Ditto. (TestWebKitAPI::Curl::TEST(CurlMultipartHandleTests, DivideLastDelimiter)): Ditto. (TestWebKitAPI::Curl::TEST(CurlMultipartHandleTests, DivideCloseDelimiter)): Ditto. (TestWebKitAPI::Curl::TEST(CurlMultipartHandleTests, DivideTransportPadding)): Ditto. (TestWebKitAPI::Curl::TEST(CurlMultipartHandleTests, DivideHeader)): Ditto. (TestWebKitAPI::Curl::TEST(CurlMultipartHandleTests, DivideBody)): Ditto. (TestWebKitAPI::Curl::TEST(CurlMultipartHandleTests, CompleteWhileHeaderProcessing)): Ditto. * Tools/TestWebKitAPI/Tests/WebKitCocoa/IndexedDBPersistence.mm: (-[IndexedDBOpenPanelUIDelegate webView:runOpenPanelWithParameters:initiatedByFrame:completionHandler:]): Ditto. * Tools/TestWebKitAPI/Tests/WebKitCocoa/WebPushDaemon.mm: Cast to LChar. * Tools/WebKitTestRunner/StringFunctions.h: (WTR::toWTFString): Use char8_t and char. Canonical link: https://commits.webkit.org/300282@main
webkit-commit-queue
force-pushed
the
eng/Make-LChar-a-distinct-type-from-uint8_t-so-it-can-imply-character-encoding-as-char8-16-32_t-do
branch
from
3a139dd to
774c0dc
Compare
|
Committed 300282@main (774c0dc): https://commits.webkit.org/300282@main Reviewed commits have been landed. Closing PR #48579 and removing active labels. |
webkit-commit-queue
removed
the
merge-queue
darinadler
deleted the
eng/Make-LChar-a-distinct-type-from-uint8_t-so-it-can-imply-character-encoding-as-char8-16-32_t-do
branch
After spending a bit of time in the char*_t weeds, I again more strongly suggest that we stick with standard C++ types. They're a little new, but we will eventually get used to them. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
9 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
774c0dc
3a139dd
🧪 wpe-wk2🧪 mac-AS-debug-wk2