http://bugs.winehq.org/show_bug.cgi?id=11019
Summary: maltab r14 hangs on splash screen Product: Wine Version: 0.9.52. Platform: Other OS/Version: other Status: UNCONFIRMED Severity: normal Priority: P2 Component: wine-x11driver AssignedTo: wine-bugs@winehq.org ReportedBy: a.hurst@shef.ac.uk
Created an attachment (id=10013) --> (http://bugs.winehq.org/attachment.cgi?id=10013) +x11drv debug output
Matlab r14 hangs when starting up.
Caused by this commit:
http://source.winehq.org/git/wine.git/?a=commit;h=5461d505d727411f8de3f7846c...
I attach a +x11drv trace: any other flags I should try?
http://bugs.winehq.org/show_bug.cgi?id=11019
Andrew Charles Hurst a.hurst@shef.ac.uk changed:
What |Removed |Added ---------------------------------------------------------------------------- Keywords| |regression
http://bugs.winehq.org/show_bug.cgi?id=11019
Andrew Charles Hurst a.hurst@shef.ac.uk changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |stefan@codeweavers.com
http://bugs.winehq.org/show_bug.cgi?id=11019
Andrew Charles Hurst a.hurst@shef.ac.uk changed:
What |Removed |Added ---------------------------------------------------------------------------- Summary|maltab r14 hangs on splash |matlab r14 hangs on splash |screen |screen
http://bugs.winehq.org/show_bug.cgi?id=11019
--- Comment #1 from Andrew Charles Hurst a.hurst@shef.ac.uk 2008-01-03 10:47:28 --- Created an attachment (id=10014) --> (http://bugs.winehq.org/attachment.cgi?id=10014) +x11drv trace from previous working commit
I notice we're seeing CreateBitmap 32x32 instead of 1x32. relevant?
http://bugs.winehq.org/show_bug.cgi?id=11019
--- Comment #2 from Andrew Charles Hurst a.hurst@shef.ac.uk 2008-01-03 10:52:03 --- Sorry, I haven't the first idea how x11drv works. Anyway, for info, I run a 24bpp X desktop - something the first trace doesn't seem to mention.
http://bugs.winehq.org/show_bug.cgi?id=11019
--- Comment #3 from Andrew Charles Hurst a.hurst@shef.ac.uk 2008-01-05 06:33:30 --- I've found the problem. The switch (screen_depth) case 24: clause lies.
this fixes the bug:
------------------------------------------------------------------------------- diff --git a/dlls/winex11.drv/x11drv_main.c b/dlls/winex11.drv/x11drv_main.c index 11cd33a..2059045 100644 --- a/dlls/winex11.drv/x11drv_main.c +++ b/dlls/winex11.drv/x11drv_main.c @@ -516,7 +516,7 @@ static BOOL process_attach(void) * should check the framebuffer and/or opengl formats available to find out what the * framebuffer actually does */ - screen_bpp = 32; + screen_bpp = 24; break;
default: -------------------------------------------------------------------------------
http://bugs.winehq.org/show_bug.cgi?id=11019
--- Comment #4 from Andrew Charles Hurst a.hurst@shef.ac.uk 2008-01-05 07:33:03 --- Created an attachment (id=10058) --> (http://bugs.winehq.org/attachment.cgi?id=10058) output from xdpyinfo
the xdpyinfo output from my system.
http://bugs.winehq.org/show_bug.cgi?id=11019
--- Comment #5 from Andrew Charles Hurst a.hurst@shef.ac.uk 2008-01-15 19:10:13 --- oprofile says matlab's stuck in awt.dll in the java gui stuff.
# opreport -f CPU: CPU with timer interrupt, speed 0 MHz (estimated) Profiling through timer interrupt TIMER:0| samples| %| ------------------ 16968 67.3922 /home/ahurst/oldharddisk/d/matlabr14/sys/java/jre/win32/jre1.5.0/bin/awt.dll 6932 27.5320 /no-vmlinux 290 1.1518 /lib/libc-2.6.1.so 124 0.4925 /usr/lib/xorg/modules/drivers/nvidia_drv.so 108 0.4289 /usr/bin/Xorg
http://bugs.winehq.org/show_bug.cgi?id=11019
--- Comment #6 from Andrew Charles Hurst a.hurst@shef.ac.uk 2008-01-15 19:24:47 --- the +snoop output shows:
0009:CALL awt._Java_sun_awt_Win32GraphicsEnvironment_initDisplay@8(<unknown, check return>) ret=01ef826f
with no subsequent RET.
http://bugs.winehq.org/show_bug.cgi?id=11019
--- Comment #7 from Austin English austinenglish@gmail.com 2008-01-16 11:57:45 --- (In reply to comment #5)
/home/ahurst/oldharddisk/d/matlabr14/sys/java/jre/win32/jre1.5.0/bin/awt.dll
Are you running this from a windows partition? If so, this bug is invalid.
http://bugs.winehq.org/show_bug.cgi?id=11019
--- Comment #8 from Andrew Charles Hurst a.hurst@shef.ac.uk 2008-01-16 12:19:03 --- No this is not running from a mounted windows partition; it's a plain subdir on /home
I installed matlab here using wine, having helped nail down bug 5040 to get the installer working.
I just happen to have installed it into the directory structure which I copied across from my dead windows hard disk.
http://bugs.winehq.org/show_bug.cgi?id=11019
--- Comment #9 from Andrew Charles Hurst a.hurst@shef.ac.uk 2008-01-20 05:02:58 --- Created an attachment (id=10376) --> (http://bugs.winehq.org/attachment.cgi?id=10376) tar gz file of output from WINEDEBUG=+snoop,+x11drv
http://bugs.winehq.org/show_bug.cgi?id=11019
Stefan Dösinger stefandoesinger@gmx.at changed:
What |Removed |Added ---------------------------------------------------------------------------- CC|stefan@codeweavers.com |stefandoesinger@gmx.at
--- Comment #10 from Stefan Dösinger stefandoesinger@gmx.at 2008-03-01 04:30:42 --- Is this bug fixed with recent wine? Alexandre fixed a few places that didn't handle 32 bpp properly
http://bugs.winehq.org/show_bug.cgi?id=11019
--- Comment #11 from Andrew Charles Hurst a.hurst@shef.ac.uk 2008-03-02 05:52:58 --- If 0.9.56 counts as recent, then no, I'm afraid not.
http://bugs.winehq.org/show_bug.cgi?id=11019
Andrew Charles Hurst a.hurst@shef.ac.uk changed:
What |Removed |Added ---------------------------------------------------------------------------- Summary|matlab r14 hangs on splash |matlab r14 hangs on splash |screen |screen if X in 24bpp mode
--- Comment #12 from Andrew Charles Hurst a.hurst@shef.ac.uk 2008-03-03 09:02:39 --- this bug also affects matlab r16 (2006b/7.3.0). Should I add this fact to the summary?
http://bugs.winehq.org/show_bug.cgi?id=11019
--- Comment #13 from Saulius K. saulius2@gmail.com 2008-03-04 02:30:50 --- Andrew, broadening app version in a summary is probably right thing to do.
Although following Java bugreport [2] is too old (2002) and the versions differ (1.5.0 vs 1.4.1), it may be usefull for investigators, as it speaks about infinite loop inside Win32GraphicsEnv.initDisplay() or AwtWin32GraphicsDevice::Initialize() and is related to color depths.
It would be nice to get JRE 1.5.0 sources, and IMHO it is quite doable. OpenJDK writes [3]:
| Sources | | AWT has a very old code base, and has a complicated structure. ... | ... | The native parts of the code base are written in C and C++. Most of the code is | platform-specific and resides in the corresponding directories: src/solaris/ or | src/windows/
As this code base is old, it probably isn't changing too much. Nevertheless, I found a download page [4] for JDK 5.0, which seems to offer a source of J2SDK. See link "Download(SCSL Source)", probably.
I am not planning to investigate this ATM, but maybe someone is.
[2] http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=4684966 [3] http://openjdk.java.net/groups/awt/ [4] http://www.sun.com/software/communitysource/j2se/java2/download.xml#t3c1
http://bugs.winehq.org/show_bug.cgi?id=11019
Andrew Charles Hurst a.hurst@shef.ac.uk changed:
What |Removed |Added ---------------------------------------------------------------------------- Summary|matlab r14 hangs on splash |matlab r14 and r16 (7.0.4 |screen if X in 24bpp mode |and 7.3.0) hang on splash | |screen if X in 24bpp mode
http://bugs.winehq.org/show_bug.cgi?id=11019
Dan Kegel dank@kegel.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |dank@kegel.com URL| |http://www.ravensheadservice | |s.com Status|UNCONFIRMED |NEW Ever Confirmed|0 |1 Summary|matlab r14 and r16 (7.0.4 |matlab r14 and r16 (7.0.4 |and 7.3.0) hang on splash |and 7.3.0) and |screen if X in 24bpp mode |WriteItNow3.1.0s hang if X | |in 24bpp mode
--- Comment #14 from Dan Kegel dank@kegel.com 2008-03-17 17:45:03 --- Also seems to affect WriteItNow from http://www.ravensheadservices.com which is a free download, see http://winehq.org/pipermail/wine-users/2008-March/030296.html Verified that wine-0.9.51 doesn't hang there.
http://bugs.winehq.org/show_bug.cgi?id=11019
Dan Kegel dank@kegel.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Target Milestone|--- |1.0.0
--- Comment #15 from Dan Kegel dank@kegel.com 2008-03-17 17:45:53 --- This affects multiple apps, and seems like a pretty basic problem, so nominating for 1.0.
http://bugs.winehq.org/show_bug.cgi?id=11019
Jyrki Muukkonen jvtm@kruu.org changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |jvtm@kruu.org
--- Comment #16 from Jyrki Muukkonen jvtm@kruu.org 2008-03-28 06:53:33 --- This affects other Java/AWT programs too, eg. FRWD Replayer (which has a few native JNI windows components).
http://www.frwd.com/downloads/replayer/demo/frwdsetup_2-5_build90.exe
The program either hangs or spews out AWT backtrace about invalid colour mask, which looks like a bpp/depth issue.
I'll attach the backtrace later, and also might try it out with 16bpp.
My system is running 64-bit Ubuntu Hardy with wine-0.9.58-0ubuntu2 package.
http://bugs.winehq.org/show_bug.cgi?id=11019
--- Comment #17 from Jyrki Muukkonen jvtm@kruu.org 2008-03-28 06:57:02 --- (In reply to comment #16)
This affects other Java/AWT programs too, eg. FRWD Replayer (which has a few native JNI windows components).
http://www.frwd.com/downloads/replayer/demo/frwdsetup_2-5_build90.exe
The program either hangs or spews out AWT backtrace about invalid colour mask, which looks like a bpp/depth issue.
I'll attach the backtrace later, and also might try it out with 16bpp.
My system is running 64-bit Ubuntu Hardy with wine-0.9.58-0ubuntu2 package.
BTW, this FRWD Replayer needs JRE6 when run under Wine. After installing it should be launched from "Program Files/FRWD Replayer" directory with something like "wine ../Java/jre-1.6.0/bin/javaw.exe -jar FRWD.jar".
http://bugs.winehq.org/show_bug.cgi?id=11019
--- Comment #18 from Jyrki Muukkonen jvtm@kruu.org 2008-03-28 07:41:22 --- Created an attachment (id=11710) --> (http://bugs.winehq.org/attachment.cgi?id=11710) Exception in thread "AWT-EventQueue-0" java.lang.IllegalArgumentException: Noncontiguous red mask
Got this with wine_0.9.58-0ubuntu2_amd64.deb and JRE 1.6.0_5
http://bugs.winehq.org/show_bug.cgi?id=11019
--- Comment #19 from Andrew Charles Hurst a.hurst@shef.ac.uk 2008-03-28 08:21:19 --- I've seen the noncontiguous red mask exception with matlab too.
http://bugs.winehq.org/show_bug.cgi?id=11019
--- Comment #20 from Alexandre Julliard julliard@winehq.org 2008-04-21 08:22:16 --- Please retest with latest git.
http://bugs.winehq.org/show_bug.cgi?id=11019
--- Comment #21 from Andrew Charles Hurst a.hurst@shef.ac.uk 2008-04-21 09:44:38 --- Retested with 0.9.60 - matlab still hangs.
Explicitly setting "screen_bpp = 24" still fixes/circumvents the problem.
http://bugs.winehq.org/show_bug.cgi?id=11019
--- Comment #22 from Andrew Charles Hurst a.hurst@shef.ac.uk 2008-04-21 09:53:18 --- Haven't tested since the '32 bit bitmask in GetDIBits' patch... will do when I get home.
http://bugs.winehq.org/show_bug.cgi?id=11019
Andrew Charles Hurst a.hurst@shef.ac.uk changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED
--- Comment #23 from Andrew Charles Hurst a.hurst@shef.ac.uk 2008-04-21 19:34:27 --- Fixed in git by:
http://source.winehq.org/git/wine.git/?a=commit;h=baceb8d07e0ae08a00aaaac39c...
Thanks Alexandre.
Marking fixed, assuming WriteItNow is ok too.
http://bugs.winehq.org/show_bug.cgi?id=11019
Alexandre Julliard julliard@winehq.org changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED
--- Comment #24 from Alexandre Julliard julliard@winehq.org 2008-05-02 12:37:40 --- Closing bugs fixed in 0.9.61.
-
wine-bugs＠winehq.org