August 2023 - webkit-dev - lists.webkit.org

InjectedBundle deprecation
by Alex Christensen 01 Aug '25

01 Aug '25

Hello WebKit developers! We are ramping up to do some serious work on Site Isolation which includes putting cross-origin iframes in a different process than the parent frame. Similar efforts have been done in other browser engines and some related work has already been done in WebKit, but not enough. This will do strange things to the InjectedBundle APIs, which have fundamental assumptions that the whole DOM is in one process and that communication only happens with one process at a time. We have never exposed InjectedBundle APIs as public API, but some other distributors of WebKit have. As we make progress on this project, we anticipate that anything in Source/WebKit/WebProcess/InjectedBundle/API is subject to deprecation and removal. This will also allow us to tighten the sandbox on the web process, resolve security and performance issues, and have cleaner code. Would the maintainers of the GTK and WPE APIs be willing to assist in migrating from those APIs and designing replacements in the UI process? Alex Christensen

5 17

Site isolation
by Alex Christensen 01 Aug '25

01 Aug '25

The SiteIsolation feature is beginning to work in the simplest of cases, but I had to do some work in the Source/WebKit/Shared/RemoteLayerTree directory, which is platform-specific. Basically all the use of LayerHostingContextIdentifier in that directory needs an equivalent implementation for other graphics implementations, like in the CoordinatedGraphics directories. There is one layout test in LayoutTests/http/tests/site-isolation/basic-iframe.html that is currently skipped on all non-Cocoa platforms. Is someone willing to collaborate to design and implement the feature in non-Cocoa ports? We are in the early stages of the project and have a long way to go, so I imagine things will continue to come up, like deciding how to implement event handling and other platform dependent features. Alex

3 6

Re: [webkit-dev] Smart Pointers Usage Guidelines
by Ryosuke Niwa 22 Aug '23

22 Aug '23

> On Aug 21, 2023, at 4:50 PM, Tim Horton <thorton(a)apple.com> wrote: > > > >> On Aug 21, 2023, at 4:42 PM, Ryosuke Niwa <rniwa(a)apple.com> wrote: >> >> >> >>> On Aug 21, 2023, at 4:41 PM, Darin Adler <darin(a)apple.com> wrote: >>> >>>> On Aug 21, 2023, at 4:39 PM, Ryosuke Niwa <rniwa(a)apple.com> wrote: >>>> >>>> Alternatively, we could add a new member function which returns CheckedPtr like `pageChecked()`. >>> >>> Yes, I think that would be a good approach that would complement the static checker. >> >> Okay, let’s go with this solution since others have expressed concerns for ref churns in the past. > > So, to be clear, this example: > >> page()->document()->foo() > > > would become: > >> page()->documentChecked()->foo()? > > (only `checked` for the deepest getter before the complex call)? Yes.

1 1

Smart Pointers Usage Guidelines
by Ryosuke Niwa 21 Aug '23

21 Aug '23

Hi all, It has been a while since I last announced the plan to adopt smart pointers using clang static analyzer: https://lists.webkit.org/pipermail/webkit-dev/2020-September/031386.html Here are some updates. 1. We’ve made a progress in implementing all the rules including rules for local variables in clang static analyzer: https://github.com/WebKit/WebKit/wiki/Smart-Pointer-Usage-Guidelines#rules-… 2. We also have a new kind of smart pointers: CheckedRef <https://github.com/WebKit/WebKit/blob/main/Source/WTF/wtf/CheckedRef.h> / CheckedPtr <https://github.com/WebKit/WebKit/blob/main/Source/WTF/wtf/CheckedPtr.h>. These behave like Ref and RefPtr in that they increment & decrement a counter in an object but unlike them don’t extend the lifetime of the object. Instead, the destructor of the base object release asserts that there are no live CheckedRef / CheckedPtr left. I added a new section in the guide describing when to use each smart pointer type: https://github.com/WebKit/WebKit/wiki/Smart-Pointer-Usage-Guidelines#when-t… 3. I wanted to describe what applying these smart pointer rules mean. A lot of code changes needed for this work involves creating Ref / RefPtr / CheckedRef / CheckedPtr in stack: https://commits.webkit.org/267082@main One subtle thing is that even when a member variable is already Ref / RefPtr / CheckedRef / CheckedPtr, we must create another one in stack as seen here: https://commits.webkit.org/267108@main This is because these member variables can be cleared during the course of invoking a non-trivial function; or put it another way, it’s not immediately obvious from the code inspection that the object pointed to stays alive over the course of a non-trivial function call. - R. Niwa

3 7

ENABLE macro placement between includes
by Alex Christensen 16 Aug '23

16 Aug '23

Our coding style says to include config.h, then the primary header, then other headers. It does not say where to put any ENABLE macros in that list. We have about 1700 places where we follow this pattern: #include "config.h" #include "RemoteSampler.h" #if ENABLE(GPU_PROCESS) … #endif We have about 350 places where we follow this pattern: #include "config.h" #if ENABLE(WEBGL) #include "WebGLUniformLocation.h" … #endif The less common pattern has the advantage of not needing to search the directory where the header is if the feature is not enabled, but I like the more common pattern more and I think it matches our style guidelines better. Should we specify one or the other, or should we not specify this in our coding style?

2 1

constinit: C++20 tip about constexpr initialization for global variables
by Yusuke Suzuki 06 Aug '23

06 Aug '23

Hello WebKittens, Now, our baseline is GCC 10, so I would like to introduce useful C++20 feature here, constinit. https://en.cppreference.com/w/cpp/language/constinit You can annotate the global variables etc. with `constinit`, like, ``` static constinit StaticStringImpl aboutBlankString { "about:blank" }; ``` Tthe benefit is that, compiler knows the strong requirement that this variable must be initialized at compile-time (constexpr constructors). Our global constructor warning is just a warning, so compilers typically allow compiling and at the very end, it emits a warning as an error. So, at that time, we don't know what is breaking constexpr condition. For example, ``` /Volumes/WebKit/macOS/OpenSource/Source/WTF/wtf/URL.cpp:932:25: error: declaration requires a global constructor [-Werror,-Wglobal-constructors] static StaticStringImpl aboutBlankString { "about:blank" }; ``` You can see that *something* in StaticStringImpl breaks `constexpr`, but we cannot see what is breaking it, because, when emitting this warning, compiler already lost that information (and compiler just already compiled it via non constexpr, and at the very end, it emits warnings like this). But if you annotate this variable with `constinit`, then compiler knows this requirement a-priori, thus, ``` /Users/yusukesuzuki/dev/OpenSource/Source/WTF/wtf/URL.cpp:932:35: error: variable does not have a constant initializer static constinit StaticStringImpl aboutBlankString { "about:blank" }; /Users/yusukesuzuki/dev/OpenSource/Source/WTF/wtf/URL.cpp:932:8: note: required by 'constinit' specifier here static constinit StaticStringImpl aboutBlankString { "about:blank" }; In file included from /Users/yusukesuzuki/dev/OpenSource/Source/WTF/wtf/URL.cpp:28: In file included from /Users/yusukesuzuki/dev/OpenSource/Source/WTF/wtf/URL.h:28: In file included from /Users/yusukesuzuki/dev/OpenSource/WebKitBuild/Release/usr/local/include/wtf/text/WTFString.h:28: In file included from /Users/yusukesuzuki/dev/OpenSource/WebKitBuild/Release/usr/local/include/wtf/text/StringImpl.h:43: /Users/yusukesuzuki/dev/OpenSource/WebKitBuild/Release/usr/local/include/wtf/text/WYHasher.h:248:28: note: cast that performs the conversions of a reinterpret_cast is not allowed in a constant expression const uint8_t* p = (const uint8_t*)characters; /Users/yusukesuzuki/dev/OpenSource/WebKitBuild/Release/usr/local/include/wtf/text/WYHasher.h:64:40: note: in call to 'computeHashImpl(&"about:blank"[0], 11, 0)' return finalizeAndMaskTop8Bits(computeHashImpl<T, Converter>(data, characterCount, 0)); /Users/yusukesuzuki/dev/OpenSource/WebKitBuild/Release/usr/local/include/wtf/text/WYHasher.h:76:16: note: in call to 'computeHashAndMaskTop8Bits(&"about:blank"[0], 11)' return computeHashAndMaskTop8Bits<T, DefaultConverter>(characters, charactersCount - 1); In file included from /Users/yusukesuzuki/dev/OpenSource/Source/WTF/wtf/URL.cpp:28: In file included from /Users/yusukesuzuki/dev/OpenSource/Source/WTF/wtf/URL.h:28: In file included from /Users/yusukesuzuki/dev/OpenSource/WebKitBuild/Release/usr/local/include/wtf/text/WTFString.h:28: /Users/yusukesuzuki/dev/OpenSource/WebKitBuild/Release/usr/local/include/wtf/text/StringImpl.h:1251:89: note: in call to 'computeLiteralHashAndMaskTop8Bits("about:blank")' s_hashFlag8BitBuffer | s_hashFlagDidReportCost | stringKind | BufferInternal | (WYHasher::computeLiteralHashAndMaskTop8Bits(characters) << s_flagCount), ConstructWithConstExpr) ``` The compiler can tell directly what is breaking `constexpr` condition. `constinit` is available from GCC 10, so it is now supported in all WebKit-supporting compilers. Best regards, -Yusuke

1 0