August 2018 - webkit-dev - lists.webkit.org

PSA: Bit fields won't be packed on Windows if you mix types
by Ryosuke Niwa 04 Sep '20

04 Sep '20

Hi, Unlike gcc and clang, MSVC pads each consecutive member variables of the same type in bitfields. e.g. if you have: struct AB { unsigned m_1 : 31; bool m_2 : 1; } then *MSVC pads m_1 and allocates sizeof(unsigned) * 2 for AB* whereas gcc and clang only allocate sizeof(unsigned) * 1 for AB. This is *not a compiler bug*. It's a spec. compliant behavior, and may in fact have a better run-time performance in some situations. However, for core objects like RenderObject and InlineBox, allocating extra 4-8 bytes per each object is prohibitory expensive. In such cases, please *use the same POD type for all bitfield member variables*. (Storage class classifiers and variable qualifiers seem to have no effect on how variables are packed; e.g. mutable, const, etc...). For example, MSVC will allocate sizeof(unsigned) * 1 for AB if we rewrite the above code as: struct AB { unsigned m_1 : 31; unsigned m_2 : 1; } When you're making this change, *be sure to audit all code that assigns a non-boolean value to m_2* because implicit type coercion into boolean is no longer in effect. For example, AB ab; ab.m_2 = 2; puts(ab.m_2 ? "true" : "false"); will print "true" before the change and will print "false" after the change. An easy way to ensure you've audited all such code is to add getters and setters for all bitfield member variables or wrap them in a special structure or class as done in http://trac.webkit.org/changeset/103353 and http://trac.webkit.org/changeset/104254. Best regards, Ryosuke Niwa Software Engineer Google Inc.

4 8

Support "registerProtocolHandler" in WebKit2
by Gyuyoung Kim 20 Apr '20

20 Apr '20

Hello, I would like to listen what do you think to support 'registerProtocolHandler' in WebKit2. This feature is to execute web content through registered custom protocol. - For example, web content can register "mailto" custom protocol using this feature, <script> navigator.registerProtocolHandler("mailto", "https://mail.naver.com/", "Web Mail"); </script> <html> <head> <title>Web Protocol Handler Sample - Test</title> </head> <body> <p>Send an email : <a href="mailto://">this !</a></p> </body> </html> Besides this feature has been supported by Mozilla and Chromium (From Mozilla 3.0, Chromium 13). - https://developer.mozilla.org/en-US/docs/Web/API/Navigator/registerProtocol… - https://developer.mozilla.org/en/docs/Web-based_protocol_handlers The feature is included in the W3C recommendation 28 released on Oct. 2014. - W3C spec : http://www.w3.org/TR/html5/webappapis.html#custom-handlers IIRC, some WebKit1 ports supported this feature though, those ports were deprecated. Now WebKit port supports this feature no more. There is a very old bug to support this feature though, it wasn't maintained so far. Recently I updated it based on latest WebKit. - https://bugs.webkit.org/show_bug.cgi?id=92749 Feel free to give me any feedback about this feature. Gyuyoung.

10 26

WebKit project in Coverity
by Carlos Alberto Lopez Perez 09 Jan '20

09 Jan '20

Hi, Coverity is an static analysis tool that allows to find bugs and vulnerabilities on the source code via static analysis. For open source projects, they offer free usage of their platform. The WebKit project is already registered there since a while. [1] To read the reports in detail or run new scans you have to be member of the WebKit project in Coverity. I happen to be one of the admins there, and I will happily grant you access to this platform if you are a WebKit committer (listed in contributors.json). So if you are interested in this, just send me an email requesting access. Regards ------- [1] https://scan.coverity.com/projects/webkit

2 1

C++17 is here. Should we use it?
by JF Bastien 26 Jun '19

26 Jun '19

Hello WebKilters, Our Chrome-y friends are considering the use of C++14 <https://groups.google.com/a/chromium.org/forum/?utm_medium=email&utm_source…>. I have to say that C++14 in WebKit has been quite amazing, and we should consider using C++17: it has many wonderful new things <https://github.com/tvaneerd/cpp17_in_TTs/blob/master/ALL_IN_ONE.md>, some of which we already use through WTF’s re-implementation of library features. By now (table as witness <http://en.cppreference.com/w/cpp/compiler_support>) most C++17 languages features are in clang and GCC, and MSVC isn’t doing too bad either. Language things can just come through WTF if we really want them. So how about it? JF

8 23

For your consideration: Naming scheme for fooIfExists/ensureFoo
by Darin Adler 22 Feb '19

22 Feb '19

Lets bike shed! For some time, functions with names like fooIfExists and ensureFoo have been bothering me. I find both names kind of opaque and unpleasant. Here’s an example: StyleResolver* styleResolverIfExists(); StyleResolver* ensureStyleResolver() What do you think of these names instead? StyleResolver* optionalStyleResolver(); StyleResolver& requiredStyleResolver(); I like them better. Note also that I think the requiredStyleResolver function should return a reference so nobody is tempted to do null checks. It seems like the C++ community likes the name optional for this concept; isn’t there some kind of std::optional template? -- Darin

15 25

CSS Logical Properties and Values
by Oriol Brufau 02 Nov '18

02 Nov '18

https://drafts.csswg.org/css-logical-1/defines properties and values to control layout through logical, rather than physical, direction and dimension mappings. I have been working on implementing this in Blink, now I'm starting with WebKit. Various properties are already supported but using non-standard names with a -webkit- prefix. The standard properties have already been shipped in Firefox and Blink, so I will start with implementing them and converting the prefixed ones into aliases. This will happen in https://bugs.webkit.org/show_bug.cgi?id=188386, the specific list of properties is: * margin-{block,inline}-{start,end} * padding-{block,inline}-{start,end} * border-{block,inline}-{start,end} * border-{block,inline}-{start,end}-{width,style,color} * {block,inline}-size * {min,max}-{block,inline}-size See Blink's Intent to Ship: https://groups.google.com/a/chromium.org/forum/#!topic/blink-dev/tkN-r119eNc <https://groups.google.com/a/chromium.org/forum/#%21topic/blink-dev/tkN-r119…> After this I willcontinue withthe remaining flow-relative box model properties and the flow-relative values for existing properties: * margin-{block,inline} * padding-{block,inline} * border-{block,inline} * border-{block,inline}-{width,style,color} * float: inline-{start,end} * clear: inline-{start,end} * resize: {block,inline} This will be behind an experimentalflag, like it's being done in Blink. Firefox shipped some. See Blink's Intent to Implement: https://groups.google.com/a/chromium.org/d/msg/blink-dev/48OwfwZrbvI/A1XZFG… I would like to get your feedback and comments about this topic, if you are fine with the plan we will start to land patches next week.

3 6

Planned Maintenance: svn.webkit.org downtime this weekend
by Lucas Forschler 15 Oct '18

15 Oct '18

Hello WebKit Developers! We are planning maintenance to our webkit.org continuous integration infrastructure this weekend. The last time we had a lengthy outage, many regressions crept into the tree and it took a while to get them all resolved. Therefore, starting at 5pm on Friday August 24th (PST), svn.webkit.org will be made read only for the duration of the maintenance. I apologize for the downtime and inconvenience. We are doing our best to keep it as short as possible and over a weekend/non-core hours. We anticipate having everything back online under normal operation before Monday at 8am, PST. I will reply to this email when svn is back open for business. Please email admin(a)webkit.org if you have any questions or concerns. Thanks, Lucas

3 11

CSSOM View behavior for scrollLeft, scrollTop, scrollWidth etc
by Frédéric Wang 24 Sep '18

24 Sep '18

Hi everybody, I'd like to announce that I started to implement the behavior for scrollLeft, scrollTop, scrollWidth, scrollHeight, scrollTo, scrollBy and scrollingElement [1] as specified by the CSSOM View specification [2]. To summarize, the main difference is which element to use in order to set or get the scroll properties of the viewport in standard (non-Quirks) mode: * The CSSOM View spec says it should be document.documentElement and most browsers implement it that way. * Chromium used to use document.body but this has recently been changed to follow the CSSOM View spec [3]. * WebKit still uses document.body and I'm proposing to change this behavior. Note that this is likely to break existing content. The change was attempted in WebKit in the past but broke some non-regression tests [4]. Chromium people also had some issues when trying to ship the change but things went well at the end [5]. For now I'm developing it under a CSSOMViewScrollingAPI developer flag. I was able to run the tests with that flag enabled, modulo a few test adjustments. [1] https://bugs.webkit.org/show_bug.cgi?id=5991 [2] https://drafts.csswg.org/cssom-view [3] https://groups.google.com/a/chromium.org/forum/#!topic/blink-dev/X64Sg16RhT4 [4] https://bugs.webkit.org/show_bug.cgi?id=106133, https://bugs.webkit.org/show_bug.cgi?id=121876 [5] https://groups.google.com/a/chromium.org/forum/#!msg/blink-dev/2iOz5-fgD8Y/… ; https://bugs.webkit.org/show_bug.cgi?id=5991#c20 -- Frédéric Wang - frederic-wang.fr

1 3

Re: [webkit-dev] Can we drop supporting mixed Wasm::MemoryMode in one process?
by Yusuke Suzuki 28 Aug '18

28 Aug '18

Posted this mail to webkit-dev mailing list too :) On Wed, Aug 29, 2018 at 3:19 AM Yusuke Suzuki <yusukesuzuki(a)slowstart.org> wrote: > Hi JSC folks, > > In Wasm supported environment, our MemoryMode is a bit dynamic. > When we fail to allocate WasmMemory for signaling mode, we fall back to > the bound checking memory instead. > > But Wasm code compiled for signaling / bound checking is incompatible. If > the code is compiled > as signaling mode, and if we attach the memory for bound checking, we need > to recompile the > code for bound checking mode. This introduces significant complexity to > our wasm compilation. > And our WebAssembly.compile is not basically compiling: it is just > validating. > Actual compiling needs to be deferred until the memory is attached by > instantiating. > It is not good when we would like to share WasmModule among multiple wasm > threads / workers in the future, since the "compiled" Wasm module is not > actually compiled. > > So, my proposal is, can we explore the way to exclusively support one of > MemoryMode in a certain architecture? > For example, in x64, enable signaling mode, and we report OOM errors if we > fail to allocate WasmMemory with signaling mode. > > Best regards, > Yusuke Suzuki >

3 10

Proposal: macros to explicitly ignore warnings
by Guillaume Emont 28 Aug '18

28 Aug '18

Hi all, Through my work to clean up compilation warnings on some platforms[1], I realized that there are many places in the code where we explicitly ignore compiler warnings. This means that we have a lot of repetition of boilerplate code that typically looks something like: #if COMPILER(CLANG) && defined(__has_warning) #pragma clang diagnostic push #if __has_warning("-Wfoo-bar") #pragma clang diagnostic ignored "-Wfoo-bar" #endif #endif // Code that triggers the warning. #if COMPILER(CLANG) && defined(__has_warning) #pragma clang diagnostic pop #endif There are also some inconsistencies on how we do this (e.g. sometimes we don't check for the compiler, or use __has_warning). So, I would like to replace that boilerplate with a set of macros that do all the necessary checks, so that the above would become: IGNORE_WARNING_CLANG("-Wfoo-bar"); // Code that triggers the warning. IGNORE_WARNING_CLANG_END("-Wfoo-bar"); Since we want to ignore warnings for gcc too, and sometimes for both gcc and clang, I also made IGNORE_WARNING_GCC[_END]() and IGNORE_WARNING_GCC_OR_CLANG[_END](). On top of that, since the places we ignore warnings happen both inside and outside of function definitions, I made _TOP_LEVEL variants of all the above macros, the difference being that the non-_TOP_LEVEL versions are wrapped inside a "do { ... } while(false)", so that they behave like an instruction. Though I am starting to agree with Michael[2] that this might be overcomplicating things, and we might be better off with only one variant. In that case, I would prefer to add "PRAGMA" in the name of the macro to make things a little more explicit as to why we wouldn't have semicolons, so that the above example would become: PRAGMA_IGNORE_WARNING_CLANG("-Wfoo-bar") // Code that triggers the warning. PRAGMA_IGNORE_WARNING_CLANG_END("-Wfoo-bar") Note the absence of semicolons. Also, in such a case, I'm not sure whether we would want to indent the macro like the surrounding code, or at the beginning of the line like a #pragma (I would naturally opt for the former option). So, this is my proposal, I'd like to hear opinions, and ideally reach consensus, on the best way to do this, or if it is a good idea at all. I filed a bug[3] and already made a patch, though as discussed above, I'm happy to change things a bit depending on the consensus. Cheers, Guij [1] https://webkit.org/b/188598 [2] https://bugs.webkit.org/show_bug.cgi?id=188996#c3 [3] https://bugs.webkit.org/show_bug.cgi?id=188996

2 1