Bug 89463 – Buildfix for newer Qt5 with C++11 support

• Status: RESOLVED
• Resolution: DUPLICATE
• Component: Web Template Framework

[Csaba Osztrogonác]

After bf84d1a1fe1550fb7cb86d686bf0138c1c41e87b in qtbase
"Qt is compiled with C++11 support enabled by default, provided the compiler supports C++11."

But unfortunately building WebKit with -std=c++0x doesn't work:
In file included from /home/webkitbuildbot/slaves/release64bitWebKit2_EC2/buildslave/qt-linux-64-release-webkit2/build/Source/WebCore/platform/graphics/IntPoint.h:30:0,
                 from /home/webkitbuildbot/slaves/release64bitWebKit2_EC2/buildslave/qt-linux-64-release-webkit2/build/Source/WebKit/qt/Api/qwebpage_p.h:35,
                 from /home/webkitbuildbot/slaves/release64bitWebKit2_EC2/buildslave/qt-linux-64-release-webkit2/build/Source/WebKit/qt/Api/qwebsettings.cpp:24:
/home/webkitbuildbot/slaves/release64bitWebKit2_EC2/buildslave/qt-linux-64-release-webkit2/build/Source/WTF/wtf/MathExtras.h: In function ‘bool std::wtf_isinf(float)’:
/home/webkitbuildbot/slaves/release64bitWebKit2_EC2/buildslave/qt-linux-64-release-webkit2/build/Source/WTF/wtf/MathExtras.h:285:63: error: ‘bool std::isinf(float)’ is not ‘constexpr’
/home/webkitbuildbot/slaves/release64bitWebKit2_EC2/buildslave/qt-linux-64-release-webkit2/build/Source/WTF/wtf/MathExtras.h: In function ‘bool std::wtf_isinf(double)’:
/home/webkitbuildbot/slaves/release64bitWebKit2_EC2/buildslave/qt-linux-64-release-webkit2/build/Source/WTF/wtf/MathExtras.h:286:64: error: ‘bool std::isinf(double)’ is not ‘constexpr’
/home/webkitbuildbot/slaves/release64bitWebKit2_EC2/buildslave/qt-linux-64-release-webkit2/build/Source/WTF/wtf/MathExtras.h: In function ‘bool std::wtf_isnan(float)’:
/home/webkitbuildbot/slaves/release64bitWebKit2_EC2/buildslave/qt-linux-64-release-webkit2/build/Source/WTF/wtf/MathExtras.h:287:63: error: ‘bool std::isnan(float)’ is not ‘constexpr’
/home/webkitbuildbot/slaves/release64bitWebKit2_EC2/buildslave/qt-linux-64-release-webkit2/build/Source/WTF/wtf/MathExtras.h: In function ‘bool std::wtf_isnan(double)’:
/home/webkitbuildbot/slaves/release64bitWebKit2_EC2/buildslave/qt-linux-64-release-webkit2/build/Source/WTF/wtf/MathExtras.h:288:64: error: ‘bool std::isnan(double)’ is not ‘constexpr’
...

What should we do? Fix the -std=c++0x build of WebKit (I have no idea how) or build Qt with -no-c++11?

[Balazs Kelemen]

WebKit should (and partially do) support c++ 11, so I think we should fix this.

[Csaba Osztrogonác]

(In reply to comment #1)
> WebKit should (and partially do) support c++ 11, so I think we should fix this.

Could you pick up this bug? As far as I remember you are c++11 expert. ;-)

[Balazs Kelemen]

(In reply to comment #2)
> (In reply to comment #1)
> > WebKit should (and partially do) support c++ 11, so I think we should fix this.
> 
> Could you pick up this bug? As far as I remember you are c++11 expert. ;-)

You should be kidding that I'm an expert :) But yes, I can pick this up.

[Balazs Kelemen]

Created attachment 148335 [details]
Patch

[WebKit Review Bot]

Attachment 148335 [details] did not pass style-queue:

Failed to run "['Tools/Scripts/check-webkit-style', '--diff-files', u'Source/WTF/ChangeLog', u'Source/WTF/wtf/Ma..." exit_code: 1
Source/WTF/wtf/MathExtras.h:285:  Code inside a namespace should not be indented.  [whitespace/indent] [4]
Source/WTF/wtf/MathExtras.h:285:  wtf_isinf is incorrectly named. Don't use underscores in your identifier names.  [readability/naming] [4]
Source/WTF/wtf/MathExtras.h:286:  wtf_isinf is incorrectly named. Don't use underscores in your identifier names.  [readability/naming] [4]
Source/WTF/wtf/MathExtras.h:287:  wtf_isnan is incorrectly named. Don't use underscores in your identifier names.  [readability/naming] [4]
Source/WTF/wtf/MathExtras.h:288:  wtf_isnan is incorrectly named. Don't use underscores in your identifier names.  [readability/naming] [4]
Total errors found: 5 in 2 files


If any of these errors are false positives, please file a bug against check-webkit-style.

[Balazs Kelemen]

(In reply to comment #5)
> Attachment 148335 [details] did not pass style-queue:
> 
> Failed to run "['Tools/Scripts/check-webkit-style', '--diff-files', u'Source/WTF/ChangeLog', u'Source/WTF/wtf/Ma..." exit_code: 1
> Source/WTF/wtf/MathExtras.h:285:  Code inside a namespace should not be indented.  [whitespace/indent] [4]
> Source/WTF/wtf/MathExtras.h:285:  wtf_isinf is incorrectly named. Don't use underscores in your identifier names.  [readability/naming] [4]
> Source/WTF/wtf/MathExtras.h:286:  wtf_isinf is incorrectly named. Don't use underscores in your identifier names.  [readability/naming] [4]
> Source/WTF/wtf/MathExtras.h:287:  wtf_isnan is incorrectly named. Don't use underscores in your identifier names.  [readability/naming] [4]
> Source/WTF/wtf/MathExtras.h:288:  wtf_isnan is incorrectly named. Don't use underscores in your identifier names.  [readability/naming] [4]
> Total errors found: 5 in 2 files
> 
> 
> If any of these errors are false positives, please file a bug against check-webkit-style.

This is the style used in this file for such functions.

[Alexey Proskuryakov]

Duplicate of 88721.

This patch is for cross platform code. Please do not ever use [Qt] prefix for such bugs.

*** This bug has been marked as a duplicate of bug 88721 ***

[Alexey Proskuryakov]

Comment on attachment 148335 [details]
Patch

Clearing r? because this will be tracked in another bug.

[Allan Sandfeld Jensen]

Comment on attachment 148335 [details]
Patch

The patch looks good, but resubmit to bug #88721, and don't mind the style warnings.

[Csaba Osztrogonác]

(In reply to comment #8)
> (From update of attachment 148335 [details])
> Clearing r? because this will be tracked in another bug.

Why this one did you mark as duplicated when there were proposed fix in this bug, and there wasn't in the another one? Is there a rule that the younger bug should be the duplicated one?

[Balazs Kelemen]

(In reply to comment #7)
> Duplicate of 88721.
> 
> This patch is for cross platform code. Please do not ever use [Qt] prefix for such bugs.
> 
> *** This bug has been marked as a duplicate of bug 88721 ***

Resubmitted to 88721.

[Alexey Proskuryakov]

> Why this one did you mark as duplicated when there were proposed fix in this bug, and there wasn't in the another one?

Three reasons:

1. This bug had a misleading [Qt] prefix, so reviewers who are C++ experts have likely skipped over the patch in review queue already.

2. EWS was unhappy (purple bubbles for cr-linux and efl), so it looked like patch needed to be updated and resubmitted anyway.

3. The older bug had more discussion of the issue.

> Is there a rule that the younger bug should be the duplicated one?

Generally, it is indeed best to dupe to older bug, mainly to show respect to the person who found the problem first, but also to keep better history of when the problem started, and besides, having such a rule is better than no rule. We sometimes do reverse duping to a newer bug to make sure that the active bug is the one with the most productive discussion, and is nearing resolution.

In this case, the choice of which direction to dupe in was not very clear.