http://bugs.winehq.org/show_bug.cgi?id=21384
Summary: Sins of a Solar Empire crash on exit Product: Wine Version: 1.1.36 Platform: x86 OS/Version: Linux Status: UNCONFIRMED Severity: trivial Priority: P2 Component: directx-d3d AssignedTo: wine-bugs@winehq.org ReportedBy: drakkk@centrum.cz CC: hverbeet@gmail.com
Created an attachment (id=25747) --> (http://bugs.winehq.org/attachment.cgi?id=25747) console output
Sins of a Solar Empire used to exit correctly in 1.1.35 (there was only this error massage: "err:d3d:wined3d_unregister_window Window 0x80030 is not registered with wined3d."), but in 1.1.36 it ends with SIGSEGV and shows some error dialog about Saved MiniDump. Console output attached. This is a regression introduced by this commit:
ce2c3ea6c74f7c28f0ec55e097d582b68c080d19 is the first bad commit commit ce2c3ea6c74f7c28f0ec55e097d582b68c080d19 Author: Henri Verbeet hverbeet@codeweavers.com Date: Sun Jan 3 21:18:26 2010 +0100
wined3d: Handle source and destination rectangles in IWineD3DSwapChainImpl_Present().
:040000 040000 c9f702ad05cb5c903597bd934605ce3d6037bd20 ae2ee3eeff4d13357d2eb9914baa8a201ee57d30 M dlls
CCing author.
http://bugs.winehq.org/show_bug.cgi?id=21384
Pavel Ondracka drakkk@centrum.cz changed:
What |Removed |Added ---------------------------------------------------------------------------- Keywords| |regression
http://bugs.winehq.org/show_bug.cgi?id=21384
Dmitry Timoshkov dmitry@codeweavers.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Severity|trivial |normal
http://bugs.winehq.org/show_bug.cgi?id=21384
--- Comment #1 from Henri Verbeet hverbeet@gmail.com 2010-01-19 04:00:57 --- Does this have a demo?
http://bugs.winehq.org/show_bug.cgi?id=21384
--- Comment #2 from Pavel Ondracka drakkk@centrum.cz 2010-01-19 05:26:47 --- (In reply to comment #1)
Does this have a demo?
Yes, there is a demo version available but it isn't affected by this bug.
http://bugs.winehq.org/show_bug.cgi?id=21384
Michael knightmb@knightmb.dyndns.org changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |knightmb@knightmb.dyndns.or | |g
--- Comment #3 from Michael knightmb@knightmb.dyndns.org 2010-03-07 20:45:01 --- I can still confirm this for the wine 1.1.39 and all releases of the Sins of Solar Empire games, doesn't matter if you run Sins, Sins + Entrenchment, or Sins + Diplomacy, the results is always the same for full screen mode.
Window mode, you don't get this crash I noticed. Can anyone else confirm if window mode is safe on exit?
http://bugs.winehq.org/show_bug.cgi?id=21384
--- Comment #4 from Pavel Ondracka drakkk@centrum.cz 2010-03-08 06:56:24 --- (In reply to comment #3)
Window mode, you don't get this crash I noticed. Can anyone else confirm if window mode is safe on exit?
I can't confirm this, because with virtual desktop there is Bug 21859 (menu does not respond to mouse clicks).
http://bugs.winehq.org/show_bug.cgi?id=21384
--- Comment #5 from Michael knightmb@knightmb.dyndns.org 2010-03-08 08:56:43 --- Can you disable the virtual desktop and manually set the resolution to match your monitor so that it will start in full screen mode? Manually I mean, you modify the game settings so that full screen mode matches your native screen resolution.
http://bugs.winehq.org/show_bug.cgi?id=21384
--- Comment #6 from Pavel Ondracka drakkk@centrum.cz 2010-03-08 10:04:17 --- (In reply to comment #5)
Can you disable the virtual desktop and manually set the resolution to match your monitor so that it will start in full screen mode? Manually I mean, you modify the game settings so that full screen mode matches your native screen resolution.
Ok, I'm not exactly sure what do you mean? Game start automatically fullscrean and with native resolution... however I did some testing involving changing resolution and windoved mode and there are really some configurations in which the crash on exit didn't happen. I will do more testing and I will post results then. BTW changing resolutions or selection window mode in game doesn't work with nvidia drivers (old known problem) but surprisingly works fine with r300 driver on my another machine. Don't know if this has anything to do with this bug.
http://bugs.winehq.org/show_bug.cgi?id=21384
--- Comment #7 from Henri Verbeet hverbeet@gmail.com 2010-03-16 16:02:23 --- What does a log with WINEDEBUG="+seh,+tid,+d3d" look like with current git?
http://bugs.winehq.org/show_bug.cgi?id=21384
--- Comment #8 from Andy Wilkinson andy@wtribe.com 2010-03-16 16:05:17 --- Created an attachment (id=26846) --> (http://bugs.winehq.org/attachment.cgi?id=26846) wine log with crash
This is a log of the console output from git master with the patch from bug 21859 applied. If I need to test it outside a desktop without the patch, let me know and I can disable a monitor to get it done.
http://bugs.winehq.org/show_bug.cgi?id=21384
Andy Wilkinson andy@wtribe.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |andy@wtribe.com
--- Comment #9 from Andy Wilkinson andy@wtribe.com 2010-03-16 16:17:02 --- http://hawkbats.com/logs/wine_soase_dbg+seh+tid+d3d.log.gz
Oops, looks like we cross-posted, Henri. I've uploaded a log with WINEDEBUG="+seh,+tid,+d3d" as requested. This is with the same setup as before, latest git with the patch from bug 21859. The log is 138 MB of text... wow.
FWIW, I am opening the game, clicking through the videos, then clicking the "Quit" button. I have to wineserver -k after clicking OK on the resulting error dialog.
http://bugs.winehq.org/show_bug.cgi?id=21384
--- Comment #10 from Henri Verbeet hverbeet@gmail.com 2010-03-16 16:38:25 --- Created an attachment (id=26847) --> (http://bugs.winehq.org/attachment.cgi?id=26847) debug patch
Does this patch make any difference? The game seems to die because we're making GL calls while the window is already destroyed. This patch avoids some of those, but it'll probably just die on the next one instead.
http://bugs.winehq.org/show_bug.cgi?id=21384
--- Comment #11 from Andy Wilkinson andy@wtribe.com 2010-03-16 16:45:55 --- (In reply to comment #10)
Created an attachment (id=26847)
--> (http://bugs.winehq.org/attachment.cgi?id=26847) [details]
debug patch
Does this patch make any difference? The game seems to die because we're making GL calls while the window is already destroyed. This patch avoids some of those, but it'll probably just die on the next one instead.
While it compiles, one question: do you want the same WINEDEBUG options on the log?
http://bugs.winehq.org/show_bug.cgi?id=21384
--- Comment #12 from Henri Verbeet hverbeet@gmail.com 2010-03-16 16:48:16 --- Yeah, unless it happens to fix the game.
http://bugs.winehq.org/show_bug.cgi?id=21384
--- Comment #13 from Andy Wilkinson andy@wtribe.com 2010-03-16 17:00:49 --- It didn't fix it, but we did get even more log output. 11 MB compressed this time:
http://hawkbats.com/logs/wine_soase_dbg+seh+tid+d3d.log.gz
http://bugs.winehq.org/show_bug.cgi?id=21384
--- Comment #14 from Henri Verbeet hverbeet@gmail.com 2010-03-16 17:17:51 --- Is that the right log? It's only 4.4MB.
http://bugs.winehq.org/show_bug.cgi?id=21384
--- Comment #15 from Andy Wilkinson andy@wtribe.com 2010-03-16 17:32:49 --- (In reply to comment #14)
Is that the right log? It's only 4.4MB.
Oh, whoops. http://hawkbats.com/logs/wine_soase_dbg+seh+tid+d3d_2.log.gz
Forgot to change the name after the copy/paste.
http://bugs.winehq.org/show_bug.cgi?id=21384
Henri Verbeet hverbeet@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Attachment #26847|0 |1 is obsolete| |
--- Comment #16 from Henri Verbeet hverbeet@gmail.com 2010-03-17 09:41:51 --- Created an attachment (id=26855) --> (http://bugs.winehq.org/attachment.cgi?id=26855) debug patch
How about this one?
http://bugs.winehq.org/show_bug.cgi?id=21384
--- Comment #17 from Andy Wilkinson andy@wtribe.com 2010-03-17 10:26:50 --- (In reply to comment #16)
Created an attachment (id=26855)
--> (http://bugs.winehq.org/attachment.cgi?id=26855) [details]
debug patch
How about this one?
Should I revert the previous patch first?
http://bugs.winehq.org/show_bug.cgi?id=21384
--- Comment #18 from Henri Verbeet hverbeet@gmail.com 2010-03-17 10:28:29 --- Yeah.
http://bugs.winehq.org/show_bug.cgi?id=21384
--- Comment #19 from Andy Wilkinson andy@wtribe.com 2010-03-17 10:54:04 --- (In reply to comment #16)
Created an attachment (id=26855)
--> (http://bugs.winehq.org/attachment.cgi?id=26855) [details]
debug patch
How about this one?
That was better, at the very least. No minidump thus time, or any other error dialog. Wine still didn't let go of my terminal, and I had to wineserver -k, but that's not entirely unheard of. Regardless, here's the new log:
http://hawkbats.com/logs/wine_soase_dbg+seh+tid+d3d_3.log.gz
http://bugs.winehq.org/show_bug.cgi?id=21384
--- Comment #20 from Henri Verbeet hverbeet@gmail.com 2010-03-17 11:08:36 --- (In reply to comment #19)
That was better, at the very least. No minidump thus time, or any other error dialog. Wine still didn't let go of my terminal, and I had to wineserver -k, but that's not entirely unheard of. Regardless, here's the new log:
http://hawkbats.com/logs/wine_soase_dbg+seh+tid+d3d_3.log.gz
Nothing obvious in the log this time. Wined3d appears to shutdown correctly now, so it may be an unrelated problem.
http://bugs.winehq.org/show_bug.cgi?id=21384
--- Comment #21 from Andy Wilkinson andy@wtribe.com 2010-03-17 11:34:02 --- (In reply to comment #20)
(In reply to comment #19)
That was better, at the very least. No minidump thus time, or any other error dialog. Wine still didn't let go of my terminal, and I had to wineserver -k, but that's not entirely unheard of. Regardless, here's the new log:
http://hawkbats.com/logs/wine_soase_dbg+seh+tid+d3d_3.log.gz
Nothing obvious in the log this time. Wined3d appears to shutdown correctly now, so it may be an unrelated problem.
I would agree with that. Wine sometimes shuts everything down but still doesn't let go of my terminal in situations that I otherwise associate with successful runs.
Let me know what else I can do to help.
http://bugs.winehq.org/show_bug.cgi?id=21384
--- Comment #22 from Henri Verbeet hverbeet@gmail.com 2010-03-17 11:49:43 --- (In reply to comment #21)
Let me know what else I can do to help.
The application may be deadlocked or waiting for something to happen that never happens. You should probably attach with winedbg and get some backtraces, I think the wiki has some pages on how to do that exactly, somewhere.
http://bugs.winehq.org/show_bug.cgi?id=21384
--- Comment #23 from Henri Verbeet hverbeet@gmail.com 2010-03-17 11:50:48 --- Also, if that hang *does* occur with the demo it would make it easier to debug for a Wine developer.
http://bugs.winehq.org/show_bug.cgi?id=21384
--- Comment #24 from Austin English austinenglish@gmail.com 2010-03-17 12:04:16 --- (In reply to comment #22)
(In reply to comment #21)
Let me know what else I can do to help.
The application may be deadlocked or waiting for something to happen that never happens. You should probably attach with winedbg and get some backtraces, I think the wiki has some pages on how to do that exactly, somewhere.
http://wiki.winehq.org/Backtraces
http://bugs.winehq.org/show_bug.cgi?id=21384
--- Comment #25 from Andy Wilkinson andy@wtribe.com 2010-03-17 12:16:45 --- (In reply to comment #22)
(In reply to comment #21)
Let me know what else I can do to help.
The application may be deadlocked or waiting for something to happen that never happens. You should probably attach with winedbg and get some backtraces, I think the wiki has some pages on how to do that exactly, somewhere.
I will need to fool with it more when I'm sitting at my computer and not running through VNC. However, when I turned off the extra WINEDEBUG flags and stopped piping to the log file, wine quit and let go of my terminal as expected, so this is probably unrelated or a false alarm.
If it's still something that we want to continue tackling, I can get some debug logs and backtraces, but it's probably something for a different bug. Let me know if I should create a new bug, or CC me on a relevant bug and I can provide testing.
For now it seems to me like the latest patch fixed this bug.
http://bugs.winehq.org/show_bug.cgi?id=21384
--- Comment #26 from Pavel Ondracka drakkk@centrum.cz 2010-03-17 15:20:09 --- Confirming, with patch from Comment #16 SoaSe now exit correctly.
http://bugs.winehq.org/show_bug.cgi?id=21384
--- Comment #27 from Henri Verbeet hverbeet@gmail.com 2010-03-18 12:00:25 --- Should be fixed by 7c930589bc40097eb0ebb84310a3ea0ae6999edb.
http://bugs.winehq.org/show_bug.cgi?id=21384
--- Comment #28 from Henri Verbeet hverbeet@gmail.com 2010-03-18 12:03:34 --- (In reply to comment #25)
If it's still something that we want to continue tackling, I can get some debug logs and backtraces, but it's probably something for a different bug. Let me know if I should create a new bug, or CC me on a relevant bug and I can provide testing.
Yeah, please create a new bug for that. (Provided it still happens, isn't filed already, etc.)
http://bugs.winehq.org/show_bug.cgi?id=21384
Pavel Ondracka drakkk@centrum.cz changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|UNCONFIRMED |RESOLVED Resolution| |FIXED
--- Comment #29 from Pavel Ondracka drakkk@centrum.cz 2010-03-20 14:08:15 --- Fixed in Wine 1.1.41.
http://bugs.winehq.org/show_bug.cgi?id=21384
Alexandre Julliard julliard@winehq.org changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED
--- Comment #30 from Alexandre Julliard julliard@winehq.org 2010-04-02 12:52:19 --- Closing bugs fixed in 1.1.42.