http://bugs.winehq.org/show_bug.cgi?id=17971
Summary: GCC optimisations cause workaround for Slingplayer bug 13371 to fail Product: Wine Version: 1.1.17 Platform: PC URL: http://download.slingmedia.com/player/pc/SlingPlayer- Setup-EU-1.5.1.343.exe OS/Version: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: ole32 AssignedTo: wine-bugs@winehq.org ReportedBy: kennybobs@o2.co.uk
The workaround for Bug 13371 is to use native qcap.dll and quartz.dll. Up to Wine 1.1.16, the workarounds work fine. With 1.1.17 (and 1.1.18) the workarounds fail (resulting in the same DLL Function Crash error).
While attempting a git bisect I found that the workaround still worked when using "gcc -O0" to compile, but "gcc -O1" causes the workaround to fail.
Result of bisect with gcc -O1:
7b6dd2c9f8339a0bc14aa7f466f5c5a0bb03da06 is first bad commit commit 7b6dd2c9f8339a0bc14aa7f466f5c5a0bb03da06 Author: Alexandre Julliard julliard@winehq.org Date: Wed Mar 11 17:28:21 2009 +0100
ole32: Get rid of WINE_StringFromCLSID and A->W conversions.
:040000 040000 ccd95ffc3e797d4978e121994715610434851ca5 dac4595337bddbc841d737265254e1bc52ccea03 M dlls
Reverting to previous commit allows the workaround to work with optimisations on.
This is kind of a blocker to Bug 13371 as the pre-compiled .debs from the WineHQ repository uses -O2 optimisation.
To complete the bisect, I had to move from GCC 4.3.2-1ubuntu12 to GCC 4.3.3 I compiled myself, as the results were random (GCC bug).
Tested in latest gitwine wine-1.1.18-199-ga9c0c24 and an unhandled exception occurs with -O1, workaround works with -O0.
http://bugs.winehq.org/show_bug.cgi?id=17971
Ken Sharp kennybobs@o2.co.uk changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |julliard@winehq.org Blocks| |13371 Keywords| |download, Installer, | |regression
http://bugs.winehq.org/show_bug.cgi?id=17971
--- Comment #1 from Ken Sharp kennybobs@o2.co.uk 2009-04-07 18:29:47 --- CC Alexandre
http://bugs.winehq.org/show_bug.cgi?id=17971
Ken Sharp kennybobs@o2.co.uk changed:
What |Removed |Added ---------------------------------------------------------------------------- Blocks|13371 |
http://bugs.winehq.org/show_bug.cgi?id=17971
--- Comment #2 from Ken Sharp kennybobs@o2.co.uk 2009-05-19 20:29:28 --- Still apparent in 1.1.21.
http://bugs.winehq.org/show_bug.cgi?id=17971
Ken Sharp kennybobs@o2.co.uk changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|UNCONFIRMED |NEW Ever Confirmed|0 |1
--- Comment #3 from Ken Sharp kennybobs@o2.co.uk 2009-06-17 18:32:01 --- Confirmed in test results.
Will test further with individual options.
Possibly a GCC bug? Will close invalid if so.
http://bugs.winehq.org/show_bug.cgi?id=17971
Ken Sharp kennybobs@o2.co.uk changed:
What |Removed |Added ---------------------------------------------------------------------------- Depends on| |18536
http://bugs.winehq.org/show_bug.cgi?id=17971
Ken Sharp kennybobs@o2.co.uk changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |ASSIGNED Component|ole32 |-unknown AssignedTo|wine-bugs@winehq.org |kennybobs@o2.co.uk
--- Comment #4 from Ken Sharp kennybobs@o2.co.uk 2009-07-10 19:34:13 --- OK, tried again...
Tested gcc -O1 again just to make sure, and the workaround fails.
So, decided to try the individual option that -O1 is supposed to enable according to the manual: http://gcc.gnu.org/onlinedocs/gcc-4.3.3/gcc/Optimize-Options.html#Optimize-O...
Annoyingly, this bug doesn't pop up when enabling all the options listed under -O1. So either -O1 is doing something that isn't documented, or it is doing something it shouldn't.
Had problems compiling GCC 4.4.0, but I'll give it another go later and try and figure out what is going on. I'm still leaning toward a GCC bug, but I still need to prove it.
http://bugs.winehq.org/show_bug.cgi?id=17971
Jerome Leclanche adys.wh@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |adys.wh@gmail.com
--- Comment #5 from Jerome Leclanche adys.wh@gmail.com 2009-12-01 12:21:24 --- (In reply to comment #0)
The workaround for Bug 13371 is to use native qcap.dll and quartz.dll. Up to Wine 1.1.16, the workarounds work fine. With 1.1.17 (and 1.1.18) the workarounds fail (resulting in the same DLL Function Crash error).
While attempting a git bisect I found that the workaround still worked when using "gcc -O0" to compile, but "gcc -O1" causes the workaround to fail.
Result of bisect with gcc -O1:
7b6dd2c9f8339a0bc14aa7f466f5c5a0bb03da06 is first bad commit commit 7b6dd2c9f8339a0bc14aa7f466f5c5a0bb03da06 Author: Alexandre Julliard julliard@winehq.org Date: Wed Mar 11 17:28:21 2009 +0100
ole32: Get rid of WINE_StringFromCLSID and A->W conversions.
:040000 040000 ccd95ffc3e797d4978e121994715610434851ca5 dac4595337bddbc841d737265254e1bc52ccea03 M dlls
Reverting to previous commit allows the workaround to work with optimisations on.
This is kind of a blocker to Bug 13371 as the pre-compiled .debs from the WineHQ repository uses -O2 optimisation.
To complete the bisect, I had to move from GCC 4.3.2-1ubuntu12 to GCC 4.3.3 I compiled myself, as the results were random (GCC bug).
Tested in latest gitwine wine-1.1.18-199-ga9c0c24 and an unhandled exception occurs with -O1, workaround works with -O0.
I can reproduce bug 13371's "1628: Failed to complete installation." with CC="ccache gcc-4.4" CFLAGS="-O0".
http://bugs.winehq.org/show_bug.cgi?id=17971
--- Comment #6 from Ken Sharp kennybobs@o2.co.uk 2009-12-08 08:31:23 --- That's no help whatsoever.
http://bugs.winehq.org/show_bug.cgi?id=17971
Bug 17971 depends on bug 18536, which changed state.
Bug 18536 Summary: Slingplayer EU 1.5.1 installer exits silently part way through http://bugs.winehq.org/show_bug.cgi?id=18536
What |Old Value |New Value ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED
http://bugs.winehq.org/show_bug.cgi?id=17971
Ken Sharp kennybobs@o2.co.uk changed:
What |Removed |Added ---------------------------------------------------------------------------- Depends on| |19068
http://bugs.winehq.org/show_bug.cgi?id=17971
Bug 17971 depends on bug 19068, which changed state.
Bug 19068 Summary: Slingplayer 1.5 installshield error 1628 http://bugs.winehq.org/show_bug.cgi?id=19068
What |Old Value |New Value ---------------------------------------------------------------------------- Status|REOPENED |RESOLVED Resolution| |FIXED
http://bugs.winehq.org/show_bug.cgi?id=17971
Ken Sharp kennybobs@o2.co.uk changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED
--- Comment #7 from Ken Sharp kennybobs@o2.co.uk 2011-03-26 23:16:16 CDT --- AFAICT this has been fixed / fixed itself / disappeared. It's hard to tell as the source of Bug 13371 has been fleeting between quartz, qcap and devenum.
With "winetricks quartz devenum" installation now completes with optimisations on.
Annoyingly, it now doesn't work with them off, and the failure is completely different, but as this is the OPPOSITE of this bug I cannot be bothered finding out why.
Fixed.
http://bugs.winehq.org/show_bug.cgi?id=17971
Alexandre Julliard julliard@winehq.org changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED
--- Comment #8 from Alexandre Julliard julliard@winehq.org 2011-04-01 12:40:35 CDT --- Closing bugs fixed in 1.3.17.