Bug 5071 - TOT WebKit doesn't compile with GCC 4.0.1 (5226)
Summary: TOT WebKit doesn't compile with GCC 4.0.1 (5226)
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: Tools / Tests (show other bugs)
Version: 420+
Hardware: Mac OS X 10.4
: P1 Normal
Assignee: Justin Garcia
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2005-09-20 15:15 PDT by Rosyna
Modified: 2005-09-24 04:02 PDT (History)
1 user (show)

See Also:


Attachments
Patch (1.30 KB, text/plain)
2005-09-21 11:42 PDT, Justin Garcia
eric: review+
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Rosyna 2005-09-20 15:15:56 PDT
/Volumes/Apostasy/rosyna/WebCore/kwq/KWQPainter.mm: In member function 'void 
QPainter::drawFloatPixmap(float, float, float, float, const QPixmap&, float, float, float, float, int, 
CGContext*)':
/Volumes/Apostasy/rosyna/WebCore/kwq/KWQPainter.mm:518: warning: argument 'w' might be 
clobbered by 'longjmp' or 'vfork'
/Volumes/Apostasy/rosyna/WebCore/kwq/KWQPainter.mm:518: warning: argument 'h' might be 
clobbered by 'longjmp' or 'vfork'
/Volumes/Apostasy/rosyna/WebCore/kwq/KWQPainter.mm:518: warning: argument 'sw' might be 
clobbered by 'longjmp' or 'vfork'
/Volumes/Apostasy/rosyna/WebCore/kwq/KWQPainter.mm:518: warning: argument 'sh' might be 
clobbered by 'longjmp' or 'vfork'

Kind of odd considering they are all arguments to the function
Comment 1 Rosyna 2005-09-20 15:18:14 PDT
GCC 4.0.1 build 5226 is used.
Comment 2 Darin Adler 2005-09-20 15:22:16 PDT
Most people have an older version of gcc, so this doesn't affect everyone.
Comment 3 Justin Garcia 2005-09-20 15:26:11 PDT
xenon and I made a workaround for this, I'll post it
Comment 4 Justin Garcia 2005-09-21 11:42:49 PDT
Created attachment 3989 [details]
Patch

This is compiler bug <rdar://problem/4266497>.	Only happens with -Os on 5226+.
 I made temporary variables for sw, sh, w, and h.  I had to move
KWQ_BLOCK_EXCEPTIONS as well, I'm not sure if that's OK.
Comment 5 Eric Seidel (no email) 2005-09-21 13:25:33 PDT
Comment on attachment 3989 [details]
Patch

Looks fine.  I still need to get around to investigating the @try/@catch change
more, as hopefully that would allow us to avoid these silly warnings.  r=me.
Comment 6 Darin Adler 2005-09-23 18:08:18 PDT
Eric, note that Justin and Tim *did* investigate the @try/@catch approach. When they tried, it caused even 
more of these types of problems.