Bug 96889 – [Qt]REGRESSION(r128653): It broke the build on Mountain Lion

• Status: RESOLVED
• Resolution: INVALID
• Component: New Bugs

[Csaba Osztrogonác]

http://trac.webkit.org/changeset/128653 broke the build
on the Qt Mountain Lion bot:

In file included from /Users/indt/Development/build-slaves/indt-mountainlion/qt-mountainlion-release/build/Source/WebKit2/UIProcess/qt/QtPageClient.cpp:39:
In file included from ../include/WebCore/Cursor.h:1:
In file included from /Users/indt/Development/build-slaves/indt-mountainlion/qt-mountainlion-release/build/Source/WebCore/platform/Cursor.h:29:
In file included from /Users/indt/Development/build-slaves/indt-mountainlion/qt-mountainlion-release/build/Source/WebCore/platform/graphics/Image.h:39:
In file included from /Users/indt/Development/build-slaves/indt-mountainlion/qt-mountainlion-release/build/Source/WTF/wtf/RetainPtr.h:30:
In file included from /System/Library/Frameworks/CoreFoundation.framework/Headers/CoreFoundation.h:38:
In file included from /System/Library/Frameworks/CoreFoundation.framework/Headers/CFBase.h:146:
/usr/include/MacTypes.h:185:9: error: reference to 'Fixed' is ambiguous
typedef Fixed *                         FixedPtr;
        ^
/usr/include/MacTypes.h:184:41: note: candidate found by name lookup is 'Fixed'
typedef SInt32                          Fixed;
                                        ^
/Users/indt/Development/build-slaves/indt-mountainlion/qt-mountainlion-release/build/Source/WebCore/platform/Length.h:37:30: note: candidate found by name lookup is 'WebCore::LengthType::Fixed'
    Auto, Relative, Percent, Fixed,
                             ^

[Tor Arne Vestbø]

r128653 was not the cause if this. Root cause fixed in r128735

[Csaba Osztrogonác]

(In reply to comment #1)
> r128653 was not the cause if this. Root cause fixed in r128735

Strange ... But how is it possible?
- r128652 - build passed - http://build.webkit.org/builders/Qt%20Mountain%20Lion%20Release/builds/1903
- r128653 - build failed - http://build.webkit.org/builders/Qt%20Mountain%20Lion%20Release/builds/1901

Is it a build system bug too? Or what?

[Tor Arne Vestbø]

(In reply to comment #2)
> (In reply to comment #1)
> > r128653 was not the cause if this. Root cause fixed in r128735
> 
> Strange ... But how is it possible?
> - r128652 - build passed - http://build.webkit.org/builders/Qt%20Mountain%20Lion%20Release/builds/1903
> - r128653 - build failed - http://build.webkit.org/builders/Qt%20Mountain%20Lion%20Release/builds/1901
> 
> Is it a build system bug too? Or what?

Not sure, the buildbot history seems a bit funky around those revisions:

Sep 14 16:02	128664	failure	#1885	Failed compile-webkit
Sep 14 15:09	128658	failure	#1884	Failed compile-webkit
Sep 14 14:19	128651	success	#1883	Build successful
Sep 14 13:41	128648	success	#1882	Build successful

Build 1884 was r128652-r128658, and might have exposed the root cause.

Sep 15 15:26	128693	failure	#1905	Failed compile-webkit
Sep 15 12:20	128692	failure	#1904	Failed compile-webkit
Sep 15 09:59	128652	success	#1903	Build successful
Sep 15 09:32	128691	exception	#1902	Exception interrupted
Sep 15 08:39	128653	failure	#1901	Failed compile-webkit
Sep 15 08:14	128690	failure	#1900	Failed compile-webkit

I have no idea what's going on here. Are you triggering manual builds? Were they clean builds?

[Csaba Osztrogonác]

(In reply to comment #3)
> I have no idea what's going on here. Are you triggering manual builds? Were they clean builds?

Yes, I triggered clean builds manually to bisect this build failure.
(r128652-r128658 ) Unfortunately they weren't clean build, because
triggereing clean build on build.webkit.org slave is impossible
without shell access for the build slave.

So it might be an incremental build issue too.