[Webkit-unassigned] [Bug 37750] New: [Qt] Build failure with Visual Studio 2010

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=37750

           Summary: [Qt] Build failure with Visual Studio 2010
           Product: WebKit
           Version: 528+ (Nightly build)
          Platform: PC
        OS/Version: Windows 7
            Status: UNCONFIRMED
          Severity: Normal
          Priority: P3
         Component: New Bugs
        AssignedTo: webkit-unassigned at lists.webkit.org
        ReportedBy: webkit at dereferenced.net


Created an attachment (id=53594)
 --> (https://bugs.webkit.org/attachment.cgi?id=53594)
Partial patch against Qt 4.7

The following applies to webkit in Qt 4.7 git (HEAD
140a96d0b860b045c18d53c1ac96e77b3893d31c). I have no idea how that translates
to webkit versions; my concern is only with the build errors.

JavaScriptCore/runtime/Structure.cpp passes literal 0 as a template parameter
for a pointer, which due to a regression in MSVC 2010. This issue is discussed
in bugs #30718 and #34800 - both of which are superseded by this bug (as the
location of the error has since changed). I've attached a very simple patch
that fixes this error.

JavaScriptCore/wtf/MathExtras.h, when built with MSVC, defines the following
function in the global scope:

inline long long abs(long long num) { return _abs64(num); }

This is a redefinition of a built-in overload from stdlib.h (in the Platform
SDK 7.0; I can't easily prove or disprove that it exists in older SDKs),
defined as inline __int64 abs(__int64 i). The error can be fixed by simply
removing this definition, if older versions do provide an overload of abs() for
__int64.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.