Bug 207365 – Multiple GCC 10 build failures in ANGLE

• Component: ANGLE
• Status: RESOLVED FIXED

[Michael Catanzaro]

WebKitGTK 2.27.4 FTBFS with GCC 10:

In file included from ../Source/ThirdParty/ANGLE/src/common/event_tracer.h:9,
                 from ../Source/ThirdParty/ANGLE/src/common/event_tracer.cpp:5:
../Source/ThirdParty/ANGLE/include/platform/Platform.h:239:35: error: 'size_t' has not been declared
  239 |                                   size_t programSize,
      |                                   ^~~~~~
../Source/ThirdParty/ANGLE/include/platform/Platform.h:243:33: error: 'size_t' has not been declared
  243 |                                 size_t programSize,
      |                                 ^~~~~~


Problem is ANGLE's Platform.h only #includes <stdint.h> and <array>. size_t is declared in a bunch of different stdlib headers [1], but stdint.h is not one of them.

stdlib.h seems like the nicest one to use here.

[1] https://en.cppreference.com/w/c/types/size_t

[Michael Catanzaro]

In addition to the missing include, there is one more build failure:

../Source/ThirdParty/ANGLE/src/libANGLE/FrameCapture_mock.cpp
In file included from ../Source/ThirdParty/ANGLE/src/libANGLE/FrameCapture.h:13,
                 from ../Source/ThirdParty/ANGLE/src/libANGLE/FrameCapture_mock.cpp:10:
../Source/ThirdParty/ANGLE/src/libANGLE/angletypes.h:427:2:   in 'constexpr' expansion of 'angle::PackedEnumMap<gl::ComponentType, unsigned int>(std::initializer_list<std::pair<gl::ComponentType, unsigned int> >{((const std::pair<gl::ComponentType, unsigned int>*)(& const std::pair<gl::ComponentType, unsigned int> [3]{std::pair<gl::ComponentType, unsigned int>{gl::ComponentType::Float, 65537}, std::pair<gl::ComponentType, unsigned int>{gl::ComponentType::Int, 1}, std::pair<gl::ComponentType, unsigned int>{gl::ComponentType::UnsignedInt, 65536}})), 3})'
../Source/ThirdParty/ANGLE/src/common/PackedEnums.h:89:72: error: modifying a const object '(unsigned int&)(&(&(const Storage&)(&((angle::PackedEnumMap<gl::ComponentType, unsigned int>*)this)->angle::PackedEnumMap<gl::ComponentType, unsigned int>::mPrivateData))->std::array<unsigned int, 4>::operator[](((std::array<unsigned int, 4>::size_type)((angle::PackedEnumMap<gl::ComponentType, unsigned int>::UnderlyingType)it->std::pair<gl::ComponentType, unsigned int>::first))))' is not allowed in a constant expression
   88 |             const_cast<T &>(const_cast<const Storage &>(
      |             ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
   89 |                 mPrivateData)[static_cast<UnderlyingType>(it->first)]) = it->second;
      |                 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~
In file included from ../Source/ThirdParty/ANGLE/src/libANGLE/Caps.h:12,
                 from ../Source/ThirdParty/ANGLE/src/libANGLE/Context.h:20,
                 from ../Source/ThirdParty/ANGLE/src/libANGLE/FrameCapture.h:14,
                 from ../Source/ThirdParty/ANGLE/src/libANGLE/FrameCapture_mock.cpp:10:
../Source/ThirdParty/ANGLE/src/libANGLE/angletypes.h:427:2: note: originally declared 'const' here
  427 | }};
      |  ^


Which is easy to fix by removing the casts. The source code has a comment saying the casts are required until C++17, but in WebKit this file is guaranteed to be built using C++17. Not sure about upstreamability of this one.

[Michael Catanzaro]

Created attachment 390259 [details]
Patch

[EWS Watchlist]

Note that there are important steps to take when updating ANGLE. See http://trac.webkit.org/wiki/UpdatingANGLE

[Michael Catanzaro]

Created attachment 390260 [details]
Patch

[Michael Catanzaro]

Reported upstream:

https://bugs.chromium.org/p/angleproject/issues/detail?id=4389
https://bugs.chromium.org/p/angleproject/issues/detail?id=4390

[Carlos Garcia Campos]

*** Bug 207511 has been marked as a duplicate of this bug. ***

[Carlos Alberto Lopez Perez]

its 2.26.x also affected by this? should we backport this to 2.26.x?

[Michael Catanzaro]

(In reply to Carlos Alberto Lopez Perez from comment #7)
> its 2.26.x also affected by this? should we backport this to 2.26.x?

It's probably affected, yes.

I doubt any Linux distros will ever use GCC 10 to build 2.26.x, so a backport is probably not very important, but it might be useful for developers to check if bugs exist in 2.26 without having to manually apply the patch.

[WebKit Commit Bot]

Comment on attachment 390260 [details]
Patch

Rejecting attachment 390260 [details] from commit-queue.

Failed to run "['/Volumes/Data/EWS/WebKit/Tools/Scripts/webkit-patch', '--status-host=webkit-queues.webkit.org', '--bot-id=webkit-cq-02', 'apply-attachment', '--no-update', '--non-interactive', 390260, '--port=mac']" exit_code: 2 cwd: /Volumes/Data/EWS/WebKit

Logging in as [e-mail]...
Fetching: https://bugs.webkit.org/attachment.cgi?id=390260&action=edit
Fetching: https://bugs.webkit.org/show_bug.cgi?id=207365&ctype=xml&excludefield=attachmentdata
Processing 1 patch from 1 bug.
Processing patch 390260 from bug 207365.
Fetching: https://bugs.webkit.org/attachment.cgi?id=390260
Failed to run "[u'/Volumes/Data/EWS/WebKit/Tools/Scripts/svn-apply', '--force', '--reviewer', u'Carlos Alberto Lopez Perez']" exit_code: 1 cwd: /Volumes/Data/EWS/WebKit

Parsed 4 diffs from patch file(s).
patching file Source/ThirdParty/ANGLE/ChangeLog
Hunk #1 succeeded at 1 with fuzz 3.
patching file Source/ThirdParty/ANGLE/changes.diff
Hunk #1 FAILED at 138.
1 out of 1 hunk FAILED -- saving rejects to file Source/ThirdParty/ANGLE/changes.diff.rej
patching file Source/ThirdParty/ANGLE/include/platform/Platform.h
patching file Source/ThirdParty/ANGLE/src/common/PackedEnums.h

Failed to run "[u'/Volumes/Data/EWS/WebKit/Tools/Scripts/svn-apply', '--force', '--reviewer', u'Carlos Alberto Lopez Perez']" exit_code: 1 cwd: /Volumes/Data/EWS/WebKit

Full output: https://webkit-queues.webkit.org/results/13321487

[Michael Catanzaro]

Conflict is r256443

[Michael Catanzaro]

Created attachment 390571 [details]
Patch

[WebKit Commit Bot]

The commit-queue encountered the following flaky tests while processing attachment 390571 [details]:

editing/spelling/spellcheck-async-remove-frame.html bug 158401 (authors: [e-mail], [e-mail], and [e-mail])
The commit-queue is continuing to process your patch.

[WebKit Commit Bot]

Comment on attachment 390571 [details]
Patch

Clearing flags on attachment: 390571

Committed r256491: <https://trac.webkit.org/changeset/256491>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.

[Radar WebKit Bug Importer]

<rdar://problem/59406353>