http://bugs.winehq.org/show_bug.cgi?id=15534
Summary: Dragon NaturallySpeaking 10 crashes occasionally with seh error Product: Wine Version: CVS/GIT Platform: All URL: http://nuance.com OS/Version: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: -unknown AssignedTo: wine-bugs@winehq.org ReportedBy: susancragin@earthlink.net
Created an attachment (id=16502) --> (http://bugs.winehq.org/attachment.cgi?id=16502) winedebug seh log where program crashed right away
Every once in a while, DNS10 crashes with a seh error, caused, I believe, by either a page fault or stack handler problem. It crashes when using Dictation Box. Sometimes it crashes when DB is being opened, sometimes it crashes after DB has been used for a while. I will try to attach several different logs.
The first log is a WINEDEBUG=+seh. Last error message before crash is: trace:seh:IsBadReadPtr 0x3056affa caused page fault during read
I will also try to get the following when the program crashes at first instance and when it crashes later: WINEDEBUG=+seh+relay
Ubuntu Intrepid Studio with generic kernel wine version wine-1.1.5-399-g82f77cc includes 3 patches being developed by Nikolay including 2 oleacc and 1 gdiplus; however, the error pre-dates the application of these patches.
http://bugs.winehq.org/show_bug.cgi?id=15534
Dmitry Timoshkov dmitry@codeweavers.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Platform|All |Other Version|CVS/GIT |1.1.5
--- Comment #1 from Dmitry Timoshkov dmitry@codeweavers.com 2008-10-06 09:33:25 --- CVS/GIT version is deprecated, use what 'wine --version' reports instead.
http://bugs.winehq.org/show_bug.cgi?id=15534
Maarten Lankhorst m.b.lankhorst@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |m.b.lankhorst@gmail.com Platform|Other |All Version|1.1.5 |CVS/GIT
--- Comment #2 from Maarten Lankhorst m.b.lankhorst@gmail.com 2008-10-06 09:35:57 --- seh just means an exception occurs, there are unfortunately lots of reasons for this.
Could you add +tid and warn+all ?
http://bugs.winehq.org/show_bug.cgi?id=15534
Maarten Lankhorst m.b.lankhorst@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Platform|All |Other Version|CVS/GIT |1.1.5
--- Comment #3 from Maarten Lankhorst m.b.lankhorst@gmail.com 2008-10-06 09:36:49 --- Collision..
http://bugs.winehq.org/show_bug.cgi?id=15534
--- Comment #4 from Susan Cragin susancragin@earthlink.net 2008-10-06 10:04:23 --- Created an attachment (id=16503) --> (http://bugs.winehq.org/attachment.cgi?id=16503) seh relay log (surprisingly short)
http://bugs.winehq.org/show_bug.cgi?id=15534
--- Comment #5 from Susan Cragin susancragin@earthlink.net 2008-10-06 14:57:06 --- Created an attachment (id=16513) --> (http://bugs.winehq.org/attachment.cgi?id=16513) winedebug tid
Here's the tid. I realized I did the +seh,+relay log wrong, so tried to re-run it, but it doesn't work. It just keeps running and running, generating thousands of lines, but the program never comes up, even after waiting a few minutes.
http://bugs.winehq.org/show_bug.cgi?id=15534
--- Comment #6 from Susan Cragin susancragin@earthlink.net 2008-10-06 15:10:25 --- Created an attachment (id=16514) --> (http://bugs.winehq.org/attachment.cgi?id=16514) warn+all log
this is about the last thousand lines of that, in an instance where Dictation Box crashed immediately upon activation. (Usually it works for a while first.)
http://bugs.winehq.org/show_bug.cgi?id=15534
--- Comment #7 from Susan Cragin susancragin@earthlink.net 2008-10-06 21:52:34 --- This is the last 5 lines of terminal output before a typical crash.
fixme:gdiplus:GdipCreateHBITMAPFromBitmap stub err:ole:OleLoadPicture IPersistStream_Load failed fixme:gdiplus:GdipGetImageGraphicsContext not implemented for image type 1114360 fixme:gdiplus:GdipCreateHBITMAPFromBitmap stub err:seh:raise_exception Unhandled exception code e06d7363 flags 1 addr 0x7b845890
http://bugs.winehq.org/show_bug.cgi?id=15534
--- Comment #8 from Maarten Lankhorst m.b.lankhorst@gmail.com 2008-10-07 05:44:03 --- Does native gdiplus work?
http://bugs.winehq.org/show_bug.cgi?id=15534
--- Comment #9 from Susan Cragin susancragin@earthlink.net 2008-10-07 06:23:48 --- Created an attachment (id=16522) --> (http://bugs.winehq.org/attachment.cgi?id=16522) crash using winetricks gdiplus
I have been using wine's gdiplus. So I just removed .wine and reinstalled DNS10 using winetricks gdiplus. On first run, the program crashed with the attached log.
http://bugs.winehq.org/show_bug.cgi?id=15534
Susan Cragin susancragin@earthlink.net changed:
What |Removed |Added ---------------------------------------------------------------------------- Attachment #16503|0 |1 is obsolete| |
--- Comment #10 from Susan Cragin susancragin@earthlink.net 2008-10-15 09:21:12 --- Created an attachment (id=16654) --> (http://bugs.winehq.org/attachment.cgi?id=16654) winedebug relay output
It was suggested that I run WINEDEBUG=+seh,+relay wine natspeak. I tried, but the output went into an endless loop, and the program never opened. So I ran WINEDEBUG=+relay only, to see what was up. The program still looped endlessly. The last thousand or so lines (before I killed wineserver) are attached. wine-1.1.6-142-g2910b90
http://bugs.winehq.org/show_bug.cgi?id=15534
--- Comment #11 from Susan Cragin susancragin@earthlink.net 2008-10-15 13:15:45 --- Created an attachment (id=16659) --> (http://bugs.winehq.org/attachment.cgi?id=16659) winedebug seh log where program ran for a while before it crashed.
Just got the last thousand lines or so, of which this is just about the last: trace:seh:IsBadReadPtr 0x1 caused page fault during read wine-1.1.6-170-gf151180
http://bugs.winehq.org/show_bug.cgi?id=15534
--- Comment #12 from Susan Cragin susancragin@earthlink.net 2008-10-16 10:37:54 --- Was running DNS10 with today's git, which has changed the behavior of DNS10 somewhat. After running it for a while, the program still crashes, but now with this brief message. Crash still happens, but now with this brief message.
err:region:CombineRgn Invalid rgn=(nil) err:region:CombineRgn Invalid rgn=(nil) err:region:CombineRgn Invalid rgn=(nil) fixme:dbghelp:dump_system_info fill in CPU vendorID and feature set
http://bugs.winehq.org/show_bug.cgi?id=15534
--- Comment #13 from Susan Cragin susancragin@earthlink.net 2008-10-16 10:43:34 --- Created an attachment (id=16679) --> (http://bugs.winehq.org/attachment.cgi?id=16679) Oct 16 crash report, last few lines
Whoops. Ignore my earlier post right above. The program hadn't finished. This is the correct terminal output.
http://bugs.winehq.org/show_bug.cgi?id=15534
--- Comment #14 from Susan Cragin susancragin@earthlink.net 2008-11-01 13:09:17 --- The last few lines before a typical crash has changed in the last few days, so I thought I'd post it.
fixme:gdiplus:GdipCreateHBITMAPFromBitmap stub err:region:CombineRgn Invalid rgn=(nil) fixme:bitmap:SetDIBits shouldn't require a DC for DIB_RGB_COLORS err:ole:OleLoadPicture IPersistStream_Load failed fixme:gdiplus:GdipGetImageGraphicsContext not implemented for image type 1114288 fixme:gdiplus:GdipCreateHBITMAPFromBitmap stub err:seh:raise_exception Unhandled exception code e06d7363 flags 1 addr 0x7b8456f0
http://bugs.winehq.org/show_bug.cgi?id=15534
Susan Cragin susancragin@earthlink.net changed:
What |Removed |Added ---------------------------------------------------------------------------- Attachment #16502|0 |1 is obsolete| | Attachment #16522|0 |1 is obsolete| | Attachment #16659|0 |1 is obsolete| | Attachment #16679|0 |1 is obsolete| |
--- Comment #15 from Susan Cragin susancragin@earthlink.net 2008-11-09 15:54:45 --- Created an attachment (id=17183) --> (http://bugs.winehq.org/attachment.cgi?id=17183) winedebug +seh+relay wine-1.1.8
This is a new seh relay log, just the last 10000 lines. wine-1.1.8
http://bugs.winehq.org/show_bug.cgi?id=15534
Susan Cragin susancragin@earthlink.net changed:
What |Removed |Added ---------------------------------------------------------------------------- Attachment #16654|0 |1 is obsolete| |
--- Comment #16 from Susan Cragin susancragin@earthlink.net 2008-11-09 15:57:02 --- Created an attachment (id=17184) --> (http://bugs.winehq.org/attachment.cgi?id=17184) winedebug +user32 wine-1.1.8
Since the last line of seh relay referred to an incomplete call to user32, I ran that next.
http://bugs.winehq.org/show_bug.cgi?id=15534
Susan Cragin susancragin@earthlink.net changed:
What |Removed |Added ---------------------------------------------------------------------------- Attachment #16514|0 |1 is obsolete| |
--- Comment #17 from Susan Cragin susancragin@earthlink.net 2008-11-09 16:38:07 --- Created an attachment (id=17185) --> (http://bugs.winehq.org/attachment.cgi?id=17185) winedebug warn+all wine-1.1.8
Replaces obsolete attachment.
http://bugs.winehq.org/show_bug.cgi?id=15534
--- Comment #18 from Susan Cragin susancragin@earthlink.net 2008-11-09 16:58:04 --- Created an attachment (id=17186) --> (http://bugs.winehq.org/attachment.cgi?id=17186) winedebug +rpcrt4 wine-1.1.8
There doesn't seem to be much here. I ran this report because the last messages in warn+all referenced rpcrt4.
http://bugs.winehq.org/show_bug.cgi?id=15534
--- Comment #19 from Susan Cragin susancragin@earthlink.net 2009-01-28 21:21:27 --- As of 28-jan-09, DNS10S runs for about 12 minutes and then crashes. The error messages are as follows.
fixme:storage:StgCreateDocfile Storage share mode not implemented. fixme:storage:StgCreateDocfile Transacted mode not implemented. fixme:menu:GetSystemMenu not supported on other process window 0x140094 err:seh:raise_exception Unhandled exception code e06d7363 flags 1 addr 0x7b8457f0
A storage error makes logical sense because the program does store a lot of information temporarily, including corrections and voice data. The winedebug runs that I did for 1.1.8 (above) show several instances of the above errors before the program crashed. ?? But this may make sense because the program used to hang a little bit periodically, and it didn't do that this time.
wine-1.1.13-480-gd0e77f8
http://bugs.winehq.org/show_bug.cgi?id=15534
Saulius K. saulius2@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |saulius2@gmail.com
--- Comment #20 from Saulius K. saulius2@gmail.com 2009-06-28 23:38:17 --- Susan, I am not into this debug session much, but as the attachment 16654 refers to lots of heap operation, could you please test DNS with WINEDEBUG=warn+heap?
AFAIK it would make some heap places less random so the crash could probably take place earlier.
http://bugs.winehq.org/show_bug.cgi?id=15534
--- Comment #21 from Susan Cragin susancragin@earthlink.net 2009-06-29 07:59:42 --- Created an attachment (id=22085) --> (http://bugs.winehq.org/attachment.cgi?id=22085) winedebug warn heap as requested
Thanks for looking into this. I will also do new seh / relay logs because all the winedebugs are old and DNS has changed (and runs slightly longer before crashing). My current wine-1.1.24-292-ga2dd6fb
Actually, if I could get rid of the following line, which repeats hundreds of times in all my logs, my logs would be much shorter.
fixme:gdiplus:GdipCreateHBITMAPFromBitmap stub
http://bugs.winehq.org/show_bug.cgi?id=15534
--- Comment #22 from Austin English austinenglish@gmail.com 2009-06-29 08:51:56 --- (In reply to comment #21)
Actually, if I could get rid of the following line, which repeats hundreds of times in all my logs, my logs would be much shorter.
fixme:gdiplus:GdipCreateHBITMAPFromBitmap stub
A) WINEDEBUG=fixme-gdiplus B) 'winetricks gdiplus' C) Remove it in the source D) Implement it ;-)
http://bugs.winehq.org/show_bug.cgi?id=15534
--- Comment #23 from Susan Cragin susancragin@earthlink.net 2009-06-29 21:21:35 --- Created an attachment (id=22094) --> (http://bugs.winehq.org/attachment.cgi?id=22094) warn heap -- second try
I tried it again, just to make sure. And I scrubbed it of those hundreds of repeating gdiplus errors.
The crash happens every single time I run DNS, and at almost exactly the same time--15 minutes 30 seconds after I initialize the program. Variations seem to be linked to the amount of dictation I do. The more words I say, the sooner it crashes. (Now the crash is about 15 minutes with almost-steady dictation.)
Because of this, I suspect that the crash is connected to some instruction that moves blocks of recognized text from volatile memory (RAM or Linux swap) to the temporary hard drive swap space that DNS creates for itself.
Below is my Suspicious Error of the Day.
err:clipboard:X11DRV_CLIPBOARD_ReadSelectionData Timed out waiting for SelectionNotify event err:clipboard:X11DRV_CLIPBOARD_RenderFormat Failed to cache clipboard data owned by another process. Format=c097 err:ole:local_server_thread ConnectNamedPipe failed with error 23
http://bugs.winehq.org/show_bug.cgi?id=15534
--- Comment #24 from Vincent Povirk madewokherd@gmail.com 2009-06-29 23:45:20 --- Hey, neat, I added that X11DRV_CLIPBOARD_ReadSelectionData line. Didn't expect to actually see it anywhere. It suggests that some other program you're pasting data from is being slow. I don't think it's likely to be relevant.
Anyway, I think you should try to get a backtrace. You should be able to get one by doing this: * Make sure your Wine has debugging symbols (either build it from source or install the appropriate debug package). * Start DNS. * Start winedbg. * Type "info process" to get a list of running processes. * Attach winedbg to the running DNS process (something like "attach 0x12" if the pid of the process is 12). This should freeze the process. * Type "set $BreakOnFirstChance = 0" so that winedbg will only break for unhandled exceptions, not for all of them (most are harmless). * Type "cont". * Wait for the program to crash.
When it crashes, winedbg should automatically stop the program and print out a backtrace. This can take a while to finish, so be sure to wait for "WineDbg>" prompt to return before assuming it's done.
http://bugs.winehq.org/show_bug.cgi?id=15534
--- Comment #25 from Susan Cragin susancragin@earthlink.net 2009-06-30 06:37:46 --- Created an attachment (id=22101) --> (http://bugs.winehq.org/attachment.cgi?id=22101) backtrace, following Vincent's directions
Vincent, Well, the results are unexpected, at least to me. Let me know if you want any other tests. Susan
http://bugs.winehq.org/show_bug.cgi?id=15534
--- Comment #26 from Vincent Povirk madewokherd@gmail.com 2009-06-30 14:28:01 --- Strange, the only division in OLEPictureImpl_SetBitmap is by GetDeviceCaps(hdcRef, LOGPIXELSX) or GetDeviceCaps(hdcRef, LOGPIXELSY). That really shouldn't be 0.
I guess CreateCompatibleDC(0) is failing.
http://bugs.winehq.org/show_bug.cgi?id=15534
--- Comment #27 from Vincent Povirk madewokherd@gmail.com 2009-06-30 14:40:38 --- I'd say either the process is running out of virtual memory or it's running out of GDI handles. Either way, I'm not sure what the next step is in tracking this down.
http://bugs.winehq.org/show_bug.cgi?id=15534
--- Comment #28 from Saulius K. saulius2@gmail.com 2009-06-30 14:56:19 --- Vincent, would a +dc trace be helpfull in pinpointing the exact issue (of those two) or the whole +relay one (or even +all) is needed?
http://bugs.winehq.org/show_bug.cgi?id=15534
--- Comment #29 from Alexandre Julliard julliard@winehq.org 2009-06-30 15:01:32 --- Judging by some of the other traces, it's very likely a GDI handle leak.
A +gdi,+relay trace would probably be the best way. It doesn't have to be run until it crashes, just long enough that we can see which handle values are consistently increasing.
http://bugs.winehq.org/show_bug.cgi?id=15534
--- Comment #30 from Alexandre Julliard julliard@winehq.org 2009-07-01 05:08:23 --- Created an attachment (id=22130) --> (http://bugs.winehq.org/attachment.cgi?id=22130) Don't delete cache DCs
Give this a try.
http://bugs.winehq.org/show_bug.cgi?id=15534
--- Comment #31 from Susan Cragin susancragin@earthlink.net 2009-07-01 12:26:34 --- Don't mean to ignore you. Will test as soon as my audio starts functioning again. Incoming sound had been getting crazier on Ubuntu Studio with pulse purged, and this morning it collapsed entirely. Sigh.
http://bugs.winehq.org/show_bug.cgi?id=15534
--- Comment #32 from Alexandre Julliard julliard@winehq.org 2009-07-01 12:43:01 --- The patch is in git now, if that helps.
http://bugs.winehq.org/show_bug.cgi?id=15534
--- Comment #33 from Susan Cragin susancragin@earthlink.net 2009-07-02 11:23:31 --- To the tune of "The Lone Ranger" It works it works it works it works, It works it works it works it works, It works it works it works it works, IT WORKS It-works-it-works.
At least, it's worked for over the last half-hour. Thanks!
http://bugs.winehq.org/show_bug.cgi?id=15534
Austin English austinenglish@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|UNCONFIRMED |RESOLVED Resolution| |FIXED
--- Comment #34 from Austin English austinenglish@gmail.com 2009-07-02 12:04:32 --- (In reply to comment #33)
To the tune of "The Lone Ranger" It works it works it works it works, It works it works it works it works, It works it works it works it works, IT WORKS It-works-it-works.
At least, it's worked for over the last half-hour. Thanks!
Excellent. Marking fixed.
http://bugs.winehq.org/show_bug.cgi?id=15534
Alexandre Julliard julliard@winehq.org changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED
--- Comment #35 from Alexandre Julliard julliard@winehq.org 2009-07-03 12:21:46 --- Closing bugs fixed in 1.1.25.
http://bugs.winehq.org/show_bug.cgi?id=15534
--- Comment #36 from Susan Cragin susancragin@earthlink.net 2009-07-06 07:41:58 --- *** Bug 17669 has been marked as a duplicate of this bug. ***
http://bugs.winehq.org/show_bug.cgi?id=15534
Roland Tobler bugzilla@rols.ch changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |bugzilla@rols.ch
http://bugs.winehq.org/show_bug.cgi?id=15534
Anastasius Focht focht@gmx.net changed:
What |Removed |Added ---------------------------------------------------------------------------- Fixed by SHA1| |70e4611e733029964fba040c768 | |158b58008069c CC| |focht@gmx.net Component|-unknown |oleaut32 Hardware|Other |x86
-
wine-bugs＠winehq.org