http://bugs.winehq.org/show_bug.cgi?id=17324
Summary: Lotus Notes Client 6.5.5 crashes after some minutes Product: Wine Version: 1.1.14 Platform: Other OS/Version: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: -unknown AssignedTo: wine-bugs@winehq.org ReportedBy: toralf.foerster@gmx.de
Compared to 1.1.13 if I run LN client longer than - say 10 minutes - it crashes. LN gives a popup windows saying it - that's all.
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #1 from Toralf Förster toralf.foerster@gmx.de 2009-02-09 06:31:31 --- BTW, this I got :
... err:win:DefWindowProcW called for other process window 0x10020 err:win:DefWindowProcW called for other process window 0x10020 err:win:DefWindowProcW called for other process window 0x10020 err:win:DefWindowProcW called for other process window 0x10020 err:win:DefWindowProcW called for other process window 0x10020 fixme:netapi32:NetUserEnum ((null),20, 0x2,0x33f064,-1,0x33f26c,0x33f270,0x33f058) stub! err:win:DefWindowProcW called for other process window 0x10020 err:win:DefWindowProcW called for other process window 0x10020 fixme:ntdll:NtQuerySystemInformation info_class SYSTEM_HANDLE_INFORMATION fixme:ntdll:NtQuerySystemInformation info_class SYSTEM_HANDLE_INFORMATION err:win:DefWindowProcW called for other process window 0x10020 err:win:DefWindowProcW called for other process window 0x10020 err:win:DefWindowProcW called for other process window 0x10020 err:win:DefWindowProcW called for other process window 0x10020 fixme:ntdll:NtQueryObject Unsupported information class 1 fixme:ntdll:NtQueryObject Unsupported information class 1
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #2 from Austin English austinenglish@gmail.com 2009-02-09 09:44:18 --- Please run a regression test: http://wiki.winehq.org/RegressionTesting
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #3 from Toralf Förster toralf.foerster@gmx.de 2009-02-10 03:43:13 --- Probably not a regression (some notes application were changed within last 2 weeks while I was away from my work place) - therefore it seems to be a general issue :-(
Today I got (after an hour or so) with 1.1.13:
... err:win:DefWindowProcW called for other process window 0x10020 fixme:netapi32:NetUserEnum ((null),20, 0x2,0x33f064,-1,0x33f26c,0x33f270,0x33f058) stub! err:win:DefWindowProcW called for other process window 0x10020 err:win:DefWindowProcW called for other process window 0x10020 err:win:DefWindowProcW called for other process window 0x10020 err:win:DefWindowProcW called for other process window 0x10020 err:win:DefWindowProcW called for other process window 0x10020 fixme:ntdll:NtQuerySystemInformation info_class SYSTEM_HANDLE_INFORMATION fixme:ntdll:NtQuerySystemInformation info_class SYSTEM_HANDLE_INFORMATION err:win:DefWindowProcW called for other process window 0x10020 err:win:DefWindowProcW called for other process window 0x10020 fixme:ntdll:NtQueryObject Unsupported information class 1 fixme:ntdll:NtQueryObject Unsupported information class 1
http://bugs.winehq.org/show_bug.cgi?id=17324
Toralf Förster toralf.foerster@gmx.de changed:
What |Removed |Added ---------------------------------------------------------------------------- Platform|Other |PC Summary|Lotus Notes Client 6.5.5 |Lotus Notes Client 6.5.5 |crashes after some minutes |crashed Version|1.1.14 |1.1.13
--- Comment #4 from Toralf Förster toralf.foerster@gmx.de 2009-02-11 04:00:47 --- verified, that 1.1.12 works fine, 1.1.13 not.
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #5 from Toralf Förster toralf.foerster@gmx.de 2009-02-13 06:27:47 --- Created an attachment (id=19417) --> (http://bugs.winehq.org/attachment.cgi?id=19417) stderr output
While currently bisecting, LN crashes for wine-1.1.12-270-g120584d, stderr is attached.
http://bugs.winehq.org/show_bug.cgi?id=17324
Toralf Förster toralf.foerster@gmx.de changed:
What |Removed |Added ---------------------------------------------------------------------------- Attachment #19417|stderr output |stderr output for wine- description| |1.1.12-270-g120584d
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #6 from Toralf Förster toralf.foerster@gmx.de 2009-02-17 03:57:51 --- Created an attachment (id=19509) --> (http://bugs.winehq.org/attachment.cgi?id=19509) stderr output for wine-1.1.12-217-g3fd7a47
another stderr output showing an issue
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #7 from Toralf Förster toralf.foerster@gmx.de 2009-02-18 02:31:39 --- Created an attachment (id=19526) --> (http://bugs.winehq.org/attachment.cgi?id=19526) stderr output for 11c1d7a
(In reply to comment #2)
Please run a regression test: http://wiki.winehq.org/RegressionTesting
Done :
tfoerste@n22 ~/devel/wine-git $ git bisect bad 11c1d7a0e7ff8d4713c8177dce4ba54781e69ead is first bad commit commit 11c1d7a0e7ff8d4713c8177dce4ba54781e69ead Author: Nikolay Sivov bunglehead@gmail.com Date: Wed Jan 7 11:58:29 2009 +0300
ole32: Fix return value for DefaultHandler_GetMiscStatus.
:040000 040000 c714b9a367c8894e61a1c528a1e014e66e3a35c9 5e1fd0df7307b205054fdc2316e17b3b56dda223 M dlls
The test of each version was marked as "bad" as soon as a crash occured. If however notes run for longer than 1 to up to 4 hours it was marked as "good". Unfortunately IMHO this might implie false "good" but at least the issue was introduced between 1.1.12 and this commit.
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #8 from Austin English austinenglish@gmail.com 2009-02-18 11:17:06 --- (In reply to comment #7)
Created an attachment (id=19526)
--> (http://bugs.winehq.org/attachment.cgi?id=19526) [details]
stderr output for 11c1d7a
(In reply to comment #2)
Please run a regression test: http://wiki.winehq.org/RegressionTesting
Done :
tfoerste@n22 ~/devel/wine-git $ git bisect bad 11c1d7a0e7ff8d4713c8177dce4ba54781e69ead is first bad commit commit 11c1d7a0e7ff8d4713c8177dce4ba54781e69ead Author: Nikolay Sivov bunglehead@gmail.com Date: Wed Jan 7 11:58:29 2009 +0300
ole32: Fix return value for DefaultHandler_GetMiscStatus.:040000 040000 c714b9a367c8894e61a1c528a1e014e66e3a35c9 5e1fd0df7307b205054fdc2316e17b3b56dda223 M dlls
The test of each version was marked as "bad" as soon as a crash occured. If however notes run for longer than 1 to up to 4 hours it was marked as "good". Unfortunately IMHO this might implie false "good" but at least the issue was introduced between 1.1.12 and this commit.
That should probably be okay. To be sure, try reverting that patch and see if it helps.
http://bugs.winehq.org/show_bug.cgi?id=17324
Austin English austinenglish@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Keywords| |regression
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #9 from Toralf Förster toralf.foerster@gmx.de 2009-02-18 11:40:36 --- (In reply to comment #8)
That should probably be okay. To be sure, try reverting that patch and see if it helps.
Already done by the bisecting process - the previous version I marked as good.
BTW thi issue is still in wine-1.1.15.
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #10 from Austin English austinenglish@gmail.com 2009-02-18 12:03:59 --- (In reply to comment #9)
(In reply to comment #8)
That should probably be okay. To be sure, try reverting that patch and see if it helps.
Already done by the bisecting process - the previous version I marked as good.
BTW thi issue is still in wine-1.1.15.
No, I mean revert that patch in wine 1.1.15 and see if that helps.
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #11 from Toralf Förster toralf.foerster@gmx.de 2009-02-19 02:21:29 --- (In reply to comment #10)
No, I mean revert that patch in wine 1.1.15 and see if that helps.
Unfortunately not, notes still crashed with ... fixme:ntdll:NtQueryObject Unsupported information class 1 fixme:ntdll:NtQueryObject Unsupported information class 1
- so the whole work for bisecting was needless :-( OTOH at least it points out to an intervall where the issue was introduced.
http://bugs.winehq.org/show_bug.cgi?id=17324
kees duineveld kees.duineveld@xs4all.nl changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |kees.duineveld@xs4all.nl
--- Comment #12 from kees duineveld kees.duineveld@xs4all.nl 2009-02-23 12:44:09 --- Additional application/ On startup winBUGS gives a trap # 101 in a popup window. After pressing OK winbugs closes.
After a regression test I came to 11c1d7a0e7ff8d4713c8177dce4ba54781e69ead is first bad commit Which lead me to this bug.
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #13 from Toralf Förster toralf.foerster@gmx.de 2009-02-24 02:24:18 --- Seems to be a different issue b/c I'm currently within bisecting process and narrowed the notes crash down to a commit between 7c18851..d661725 (8 revisions left).
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #14 from Toralf Förster toralf.foerster@gmx.de 2009-02-24 05:34:54 --- I bisected the issue to :
tfoerste@n22 ~/devel/wine-git $ git bisect bad 8a5eae2122478f5f057b1f696523157003d0dbcc is first bad commit commit 8a5eae2122478f5f057b1f696523157003d0dbcc Author: Hans Leidekker hans@codeweavers.com Date: Tue Jan 6 11:20:41 2009 +0100
usp10: Handle CR/LF in ScriptItemize.
:040000 040000 ec9d2861749359e5f145a4af35189c37feff9010 2961a4bb15ac25e84452a4cc3626b410ef8e4d6b M dlls
Verified it by reverting that commit and applying the patch agains wine-1.1.15 :
tfoerste@n22 ~/devel/wine-git $ git diff 020456f..8a5eae2 | patch -p1 -R patching file dlls/usp10/tests/usp10.c patching file dlls/usp10/usp10.c
I currently run notes w/ this version, runs fine until now.
BTW the issue always happened (however unfortunately not at every try) by opening one of 3 specific documents, common was that all docsuments have attachements (*.xls or *.doc)
http://bugs.winehq.org/show_bug.cgi?id=17324
Hans Leidekker hans@meelstraat.net changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |hans@meelstraat.net
--- Comment #15 from Hans Leidekker hans@meelstraat.net 2009-02-24 06:18:51 --- Please attach a WINEDEBUG=+uniscribe trace with the patch applied.
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #16 from Toralf Förster toralf.foerster@gmx.de 2009-02-24 07:33:50 --- (In reply to comment #15)
Please attach a WINEDEBUG=+uniscribe trace with the patch applied.
Unfortunately not possible, the content of thoses docs are confidential .
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #17 from Austin English austinenglish@gmail.com 2009-02-24 10:03:34 --- (In reply to comment #16)
(In reply to comment #15)
Please attach a WINEDEBUG=+uniscribe trace with the patch applied.
Unfortunately not possible, the content of thoses docs are confidential .
Could you try with a test/non-confidential document?
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #18 from Toralf Förster toralf.foerster@gmx.de 2009-02-25 02:55:53 --- Created an attachment (id=19640) --> (http://bugs.winehq.org/attachment.cgi?id=19640) lotus notes stderr
(In reply to comment #17)
Could you try with a test/non-confidential document?
I created a dummy doc with 2 attachments, trace is attached.
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #19 from Hans Leidekker hans@meelstraat.net 2009-02-25 04:30:41 --- Thanks. Unfortunately there's not enough information in that trace. Can you please create another one with these debug channels turned on: +uniscribe,+relay,+seh ?
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #20 from Toralf Förster toralf.foerster@gmx.de 2009-02-25 04:45:29 --- (In reply to comment #19)
Thanks. Unfortunately there's not enough information in that trace. Can you please create another one with these debug channels turned on: +uniscribe,+relay,+seh ?
Yes, but b/c that the trace file is ~ 300 MB in size and compressed ~ 7 MB do you have any web spcae to upload it ?
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #21 from Hans Leidekker hans@meelstraat.net 2009-02-25 04:57:14 ---
Yes, but b/c that the trace file is ~ 300 MB in size and compressed ~ 7 MB do you have any web spcae to upload it ?
Just attach the tail end of the log that fits bugzilla's limit.
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #22 from Toralf Förster toralf.foerster@gmx.de 2009-02-25 05:48:33 --- Created an attachment (id=19646) --> (http://bugs.winehq.org/attachment.cgi?id=19646) split 1#7
splitted with:split --bytes=1000000 LotusNotes.stderr.gz
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #23 from Toralf Förster toralf.foerster@gmx.de 2009-02-25 05:49:09 --- Created an attachment (id=19647) --> (http://bugs.winehq.org/attachment.cgi?id=19647) split 2#7
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #24 from Toralf Förster toralf.foerster@gmx.de 2009-02-25 05:49:44 --- Created an attachment (id=19648) --> (http://bugs.winehq.org/attachment.cgi?id=19648) split 3#7
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #25 from Toralf Förster toralf.foerster@gmx.de 2009-02-25 05:50:21 --- Created an attachment (id=19649) --> (http://bugs.winehq.org/attachment.cgi?id=19649) split 4#7
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #26 from Toralf Förster toralf.foerster@gmx.de 2009-02-25 05:51:11 --- Created an attachment (id=19650) --> (http://bugs.winehq.org/attachment.cgi?id=19650) split 5#7
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #27 from Toralf Förster toralf.foerster@gmx.de 2009-02-25 05:51:46 --- Created an attachment (id=19651) --> (http://bugs.winehq.org/attachment.cgi?id=19651) split 6#7
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #28 from Toralf Förster toralf.foerster@gmx.de 2009-02-25 05:52:19 --- Created an attachment (id=19652) --> (http://bugs.winehq.org/attachment.cgi?id=19652) split 7#7
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #29 from Toralf Förster toralf.foerster@gmx.de 2009-02-25 05:54:12 --- Created an attachment (id=19653) --> (http://bugs.winehq.org/attachment.cgi?id=19653) split 8#7
oops - there are 8 parts due to the split into 1 MiB than 1 MB
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #30 from Hans Leidekker hans@meelstraat.net 2009-02-26 05:23:16 --- There's no crash in that log. Does it really fail in the same way?
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #31 from Toralf Förster toralf.foerster@gmx.de 2009-02-26 06:23:26 --- (In reply to comment #30)
There's no crash in that log. Does it really fail in the same way?
No, as already said, the crash occurs currently only for docs which are all confidential - therefore we cannot publish such a trace. The attached trace is w/o an crash got from a dummy document which follows the same structure as the confidential docs.
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #32 from Austin English austinenglish@gmail.com 2009-02-26 09:25:00 --- (In reply to comment #31)
(In reply to comment #30)
There's no crash in that log. Does it really fail in the same way?
No, as already said, the crash occurs currently only for docs which are all confidential - therefore we cannot publish such a trace. The attached trace is w/o an crash got from a dummy document which follows the same structure as the confidential docs.
You need to figure out what's triggering the bug in those documents then and get a trace with that information.
http://bugs.winehq.org/show_bug.cgi?id=17324
Toralf Förster toralf.foerster@gmx.de changed:
What |Removed |Added ---------------------------------------------------------------------------- Attachment #19646|0 |1 is obsolete| | Attachment #19647|0 |1 is obsolete| | Attachment #19648|0 |1 is obsolete| | Attachment #19649|0 |1 is obsolete| | Attachment #19650|0 |1 is obsolete| | Attachment #19651|0 |1 is obsolete| | Attachment #19652|0 |1 is obsolete| | Attachment #19653|0 |1 is obsolete| |
--- Comment #33 from Toralf Förster toralf.foerster@gmx.de 2009-03-02 04:15:36 --- Created an attachment (id=19741) --> (http://bugs.winehq.org/attachment.cgi?id=19741) stderr output for wine-1.1.16
(In reply to comment #32)
You need to figure out what's triggering the bug in those documents then and get a trace with that information.
attached is the end (last 100000 lines) of such a trace, if you need more (about 4.6 million lines), pls request it.
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #34 from Toralf Förster toralf.foerster@gmx.de 2009-03-02 04:51:43 --- BTW after taking a look at the soutrce code of the suspicious commit and - b/c wine-1.1.16 crashed 2 times in a row after opening a specific document - I reverted one specific line of that commit, until now it seems to work :
diff --git a/dlls/usp10/usp10.c b/dlls/usp10/usp10.c index fc2f8e9..6d8912c 100644 --- a/dlls/usp10/usp10.c +++ b/dlls/usp10/usp10.c @@ -556,7 +556,7 @@ HRESULT WINAPI ScriptItemize(const WCHAR *pwcInChars, int cInChars, int cMaxItem New_Script, pItems[index].a.eScript, index, cnt, pItems[index].iCharPos);
- for (cnt=1; cnt < cInChars; cnt++) + for (cnt=0; cnt < cInChars; cnt++) { if (pwcInChars[cnt] == '\r') New_Script = Script_CR;
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #35 from Toralf Förster toralf.foerster@gmx.de 2009-03-02 05:58:54 --- (In reply to comment #34)
reverted one specific line of that commit, until now it seems to work :
erm, no, it doesn#t help.
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #36 from Hans Leidekker hans@meelstraat.net 2009-03-02 07:00:32 --- That log file doesn't show any uniscribe calls. Please create a +relay,+seh,+uniscribe log that shows, say, the last 20 uniscribe calls and then as many lines as possible.
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #37 from Toralf Förster toralf.foerster@gmx.de 2009-03-02 07:43:18 --- Created an attachment (id=19745) --> (http://bugs.winehq.org/attachment.cgi?id=19745) uniscribe greped
attached:
tfoerste@n22 /var/tmp $ grep -B 5 -A 5 uniscribe LotusNotes.stderr > uniscribe.greped tfoerste@n22 /var/tmp $ gzip uniscribe.greped
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #38 from Toralf Förster toralf.foerster@gmx.de 2009-03-02 07:49:52 --- Created an attachment (id=19746) --> (http://bugs.winehq.org/attachment.cgi?id=19746) last 500 K Lines
tfoerste@n22 /var/tmp $ tail -n 500000 LotusNotes.stderr > last500Klines.log tfoerste@n22 /var/tmp $ gzip last500Klines.log
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #39 from Toralf Förster toralf.foerster@gmx.de 2009-03-16 05:03:15 --- Tested wine-1.1.17 - issue still here, applying $> git diff 020456f..8a5eae2 dlls/usp10/usp10.c | patch -R -p1 to the sources resolved it.
BTW I found an additional document within the same notes database where this issue occurs.
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #40 from Toralf Förster toralf.foerster@gmx.de 2009-04-14 02:31:37 --- FWIW the issue still happens with wine-1.1.18 however not relyable. This means w/o reverting the mentioned commit the issue only sometimes happens (always and only w/ documents from a specific Lotus Notes database) whereas reverting that commit never results into a crash.
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #41 from Austin English austinenglish@gmail.com 2009-10-12 14:55:50 --- Is this still an issue in current (1.1.31 or newer) wine?
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #42 from Toralf Förster toralf.foerster@gmx.de 2009-10-13 02:46:53 --- Created an attachment (id=24068) --> (http://bugs.winehq.org/attachment.cgi?id=24068) stderr
notes crashed few minutes ago while opening a document from that ln database (wine-1.1.31-78-gd95d7db)
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #43 from Dmitry Timoshkov dmitry@codeweavers.com 2009-10-13 03:51:31 --- (In reply to comment #42)
Created an attachment (id=24068)
--> (http://bugs.winehq.org/attachment.cgi?id=24068) [details]
stderr notes crashed few minutes ago while opening a document from that ln database (wine-1.1.31-78-gd95d7db)
There is no a crash indication in that output.
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #44 from Toralf Förster toralf.foerster@gmx.de 2009-10-13 03:57:45 --- (In reply to comment #43)
There is no a crash indication in that output.
which WINEDEBUG settings would you advice to improve such a output ?
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #45 from Dmitry Timoshkov dmitry@codeweavers.com 2009-10-13 04:04:58 --- (In reply to comment #44)
There is no a crash indication in that output.
which WINEDEBUG settings would you advice to improve such a output ?
I just wonder why did you claim that there is a crash.
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #46 from Toralf Förster toralf.foerster@gmx.de 2009-10-13 04:23:43 --- (In reply to comment #45)
I just wonder why did you claim that there is a crash.
Oh - Lotus Notes crashes with the crash dialog.
I know that reverting the mentioned commit in the bug report makes Lotus Notes working fine w/o any problems.
I'm wondering how I can help you to identify the root cause. Unfortunately the crash can't be easily reproduced and I've do revert that commit for current git sources to get a wrking wine+ln solution - but for the next few days I could run the unpatched wine software and send you the appropriate output if wanted.
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #47 from Jeff Zaroyko jeffz@jeffz.name 2009-10-13 04:51:09 --- (In reply to comment #46)
(In reply to comment #45)
I just wonder why did you claim that there is a crash.
Oh - Lotus Notes crashes with the crash dialog.
http://wiki.winehq.org/Backtraces
http://bugs.winehq.org/show_bug.cgi?id=17324
Toralf Förster toralf.foerster@gmx.de changed:
What |Removed |Added ---------------------------------------------------------------------------- Attachment #19417|0 |1 is obsolete| | Attachment #19509|0 |1 is obsolete| | Attachment #19526|0 |1 is obsolete| | Attachment #19640|0 |1 is obsolete| | Attachment #19741|0 |1 is obsolete| | Attachment #19745|0 |1 is obsolete| | Attachment #19746|0 |1 is obsolete| | Attachment #24068|0 |1 is obsolete| |
--- Comment #48 from Toralf Förster toralf.foerster@gmx.de 2009-10-13 07:43:48 --- Created an attachment (id=24072) --> (http://bugs.winehq.org/attachment.cgi?id=24072) winedbg bt
backtrace of ln crash
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #49 from Toralf Förster toralf.foerster@gmx.de 2009-10-26 08:35:32 --- Created an attachment (id=24362) --> (http://bugs.winehq.org/attachment.cgi?id=24362) backtrace of versino 1.1.32
http://bugs.winehq.org/show_bug.cgi?id=17324
Toralf Förster toralf.foerster@gmx.de changed:
What |Removed |Added ---------------------------------------------------------------------------- Attachment #24362|backtrace of versino 1.1.32 |backtrace of version 1.1.32 description| |
http://bugs.winehq.org/show_bug.cgi?id=17324
Dan Kegel dank@kegel.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |dank@kegel.com
--- Comment #50 from Dan Kegel dank@kegel.com 2009-10-27 09:09:11 --- a +synchronous,+relay,+xrender log might be of some interest, too.
compress big logs with rzip (rzip.samba.org) rather than splitting, if possible.
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #51 from Toralf Förster toralf.foerster@gmx.de 2009-10-27 09:25:12 --- Created an attachment (id=24388) --> (http://bugs.winehq.org/attachment.cgi?id=24388) backtrace w/ export WINEDEBUG=+synchronous,+relay,+xrender
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #52 from Dan Kegel dank@kegel.com 2009-10-27 12:48:30 --- That looks like the same crash as bug 12307. Hmm.
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #53 from Toralf Förster toralf.foerster@gmx.de 2010-01-05 09:49:02 --- Created an attachment (id=25555) --> (http://bugs.winehq.org/attachment.cgi?id=25555) revert part of commit which crashes lotus notes.
There's now another scenario where Lotus Notes crashed under wine-1.1.35.
If I open an invitation within the calendar and press "Request information ...." then notes crashes. With the (attached) patch however an input window is opened w/o problems.
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #54 from Dan Kegel dank@kegel.com 2010-01-05 10:11:10 --- Now the big challenge: can you come up with a miminal testcase that needs your change to pass?
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #55 from Dmitry Timoshkov dmitry@codeweavers.com 2010-01-06 00:12:53 --- 'pItems[index].iCharPos = cnt' assignments in TRACE()s look wrong.
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #56 from Toralf Förster toralf.foerster@gmx.de 2010-01-06 03:10:11 --- (In reply to comment #55)
'pItems[index].iCharPos = cnt' assignments in TRACE()s look wrong.
Probably you#re right, after I bisected the wrong commit (comment #14) I tried to play with that line but w/o luck (maybe b/c the bug wasn't 100% repeatable with the used test case).
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #57 from Toralf Förster toralf.foerster@gmx.de 2010-01-08 03:55:13 --- (In reply to comment #54)
Now the big challenge: can you come up with a miminal testcase that needs your change to pass?
I commented out line 607 and now the crash went away.
http://bugs.winehq.org/show_bug.cgi?id=17324
Toralf Förster toralf.foerster@gmx.de changed:
What |Removed |Added ---------------------------------------------------------------------------- Attachment #24072|0 |1 is obsolete| | Attachment #24362|0 |1 is obsolete| | Attachment #24388|0 |1 is obsolete| |
--- Comment #58 from Toralf Förster toralf.foerster@gmx.de 2010-01-08 04:07:44 --- Created an attachment (id=25608) --> (http://bugs.winehq.org/attachment.cgi?id=25608) fix a crash
I tested this patch with some documents from a confidential lotus notes database where notes crashed before - now it works there too.
http://bugs.winehq.org/show_bug.cgi?id=17324
Dan Kegel dank@kegel.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Keywords| |patch Status|UNCONFIRMED |NEW CC| |lats@yless4u.com.au Ever Confirmed|0 |1
--- Comment #59 from Dan Kegel dank@kegel.com 2010-01-08 08:06:39 --- git blame says that line was added by http://source.winehq.org/git/wine.git/?a=commitdiff;h=ff421fde8e88243b8d78e2... cc'ing the author.
That change was way before wine-1.1.13, so I don't know why it's biting you now. Maybe you just got lucky :-)
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #60 from Toralf Förster toralf.foerster@gmx.de 2010-01-11 03:45:04 --- (In reply to comment #59)
That change was way before wine-1.1.13, so I don't know why it's biting you now. Maybe you just got lucky :-)
With wine-1.1.36 I had to apply the full patch, commenting out only that single line brought the issue back. I'm arguing that this patch isn't the culprit itself but it does unhide an issue elsewhere.
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #61 from Toralf Förster toralf.foerster@gmx.de 2010-01-13 03:36:17 --- With "$>export WINEDEBUG=+uniscribe" I run Lotus Notes 6.5.5 with the current git version wine-1.1.36 and with the git version patched with http://bugs2.winehq.org/attachment.cgi?id=25555 . I'll attach the output files.
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #62 from Toralf Förster toralf.foerster@gmx.de 2010-01-13 03:37:43 --- Created an attachment (id=25703) --> (http://bugs.winehq.org/attachment.cgi?id=25703) WINEDEBUG=+uniscribe output of wine-1.1.36
output of
tfoerste@n22 ~ $ wine start /Unix "/home/tfoerste/.wine/drive_c/Program Files/lotus/notes/notes.exe" 1>out.txt 2>&1
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #63 from Toralf Förster toralf.foerster@gmx.de 2010-01-13 03:38:42 --- Created an attachment (id=25704) --> (http://bugs.winehq.org/attachment.cgi?id=25704) WINEDEBUG=+uniscribe output of wine-1.1.36 + patch
output of
tfoerste@n22 ~ $ wine start /Unix "/home/tfoerste/.wine/drive_c/Program Files/lotus/notes/notes.exe" 1>out_patched.txt 2>&1
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #64 from Jeff L lats@yless4u.com.au 2010-02-28 06:02:05 --- I have submitted a patch to clean up the TRACE statement to remove the assignment:
diff --git a/dlls/usp10/usp10.c b/dlls/usp10/usp10.c index 523bd4f..a8ae4ce 100644 --- a/dlls/usp10/usp10.c +++ b/dlls/usp10/usp10.c @@ -604,7 +604,7 @@ HRESULT WINAPI ScriptItemize(const WCHAR *pwcInChars, int cInChars, int cMaxItem * checks the n+1 structure */ memset(&pItems[index+1].a, 0, sizeof(SCRIPT_ANALYSIS));
- TRACE("index=%d cnt=%d iCharPos=%d\n", index+1, cnt, pItems[index+1].iCharPos = cnt); + TRACE("index=%d cnt=%d iCharPos=%d\n", index+1, cnt, pItems[index+1].iCharPos);
/* Set one SCRIPT_STATE item being returned */ if (pcItems) *pcItems = index + 1;
you might like to test it and see if there is any improvement.
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #65 from Toralf Förster toralf.foerster@gmx.de 2010-03-01 02:20:18 --- (In reply to comment #64)
I have submitted a patch to clean up the TRACE statement to remove the assignment:
...
you might like to test it and see if there is any improvement.
No, bug is still in 1.1.39.
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #66 from Toralf Förster toralf.foerster@gmx.de 2010-03-01 03:40:02 --- This makes the bug away - but I dunno why :
diff --git a/dlls/usp10/usp10.c b/dlls/usp10/usp10.c index 523bd4f..b4896c8 100644 --- a/dlls/usp10/usp10.c +++ b/dlls/usp10/usp10.c @@ -558,9 +558,9 @@ HRESULT WINAPI ScriptItemize(const WCHAR *pwcInChars, int cInChars, int cMaxItem
for (cnt=1; cnt < cInChars; cnt++) { - if (pwcInChars[cnt] == '\r') - New_Script = Script_CR; - else +// if (pwcInChars[cnt] == '\r') +// New_Script = Script_CR; +// else if (pwcInChars[cnt] == '\n') New_Script = Script_LF; else
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #67 from Jeff L lats@yless4u.com.au 2010-03-01 05:44:51 --- Created an attachment (id=26559) --> (http://bugs.winehq.org/attachment.cgi?id=26559) Extra trace for debugging
I want to check if the problem is a buffer overrun.
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #68 from Jeff L lats@yless4u.com.au 2010-03-01 05:46:24 --- Can you apply the patch posted last and rerun the trace so that I see if this a buffer overrun?
http://bugs.winehq.org/show_bug.cgi?id=17324
Toralf Förster toralf.foerster@gmx.de changed:
What |Removed |Added ---------------------------------------------------------------------------- Attachment #25703|0 |1 is obsolete| | Attachment #25704|0 |1 is obsolete| |
--- Comment #69 from Toralf Förster toralf.foerster@gmx.de 2010-03-01 07:57:39 --- Created an attachment (id=26562) --> (http://bugs.winehq.org/attachment.cgi?id=26562) last 300 lines w/ exra TRACE at wine-1.1.39
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #70 from Toralf Förster toralf.foerster@gmx.de 2010-03-01 09:57:18 --- Created an attachment (id=26563) --> (http://bugs.winehq.org/attachment.cgi?id=26563) complete trace file
This is the complete trace file as requested for a dummy invitation with these settings :
Subject: "test" When: "Tue 03/02/2010 9:00 AM - 10:00 AM" Location: "somewhere" Rooms: "over" Resources: "the rainbow" Description: "hier könnte ihre werbung stehen !"
http://bugs.winehq.org/show_bug.cgi?id=17324
Toralf Förster toralf.foerster@gmx.de changed:
What |Removed |Added ---------------------------------------------------------------------------- Attachment #25608|0 |1 is obsolete| |
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #71 from Jeff L lats@yless4u.com.au 2010-03-06 17:39:41 --- Created an attachment (id=26648) --> (http://bugs.winehq.org/attachment.cgi?id=26648) Candidate patch to prevent running out of script items on last item
ScriptItemize needs an extra script_item to store the end pointer. It looks like it is possible to cause a write past the end of the block which is probably triggers an exception randomly depending on where the block is located in memory.
Can you give this a test.
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #72 from Toralf Förster toralf.foerster@gmx.de 2010-03-08 02:43:18 --- (In reply to comment #71)
Can you give this a test.
Doesn't help.
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #73 from Jeff L lats@yless4u.com.au 2010-03-08 02:57:55 --- That's unfortunate as I have taken some strings from the trace you provided and run test cases that was the only area where there was a weakness. You say that the trace is of a run that does not fail? If that is the case, are you able to provide last part of the trace where a failure occurs with substitution of the text but not changing numerics, \n, \r or unicode text?
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #74 from Toralf Förster toralf.foerster@gmx.de 2010-03-08 07:42:28 --- Created an attachment (id=26692) --> (http://bugs.winehq.org/attachment.cgi?id=26692) another creash trace example
This is another example of a Lotus Notes calendar invitation I got where LN crashes as soon as I click on "Request information".
The trace is made of wine-1.1.40 with this patch :
tfoerste@n22 ~/devel/wine-git $ git diff diff --git a/dlls/usp10/usp10.c b/dlls/usp10/usp10.c index 28984e0..fcbedb1 100644 --- a/dlls/usp10/usp10.c +++ b/dlls/usp10/usp10.c @@ -558,6 +558,7 @@ HRESULT WINAPI ScriptItemize(const WCHAR *pwcInChars, int cInChars, int cMaxItem
for (cnt=1; cnt < cInChars; cnt++) { + TRACE("debug: Index=%d cnt=%d iCharPos=%d\n", index, cnt, pItems[index].iCharPos); if (pwcInChars[cnt] == '\r') New_Script = Script_CR; else @@ -605,6 +606,7 @@ HRESULT WINAPI ScriptItemize(const WCHAR *pwcInChars, int cInChars, int cMaxItem TRACE("index=%d cnt=%d iCharPos=%d\n", index+1, cnt, pItems[index+1].iCharPos);
/* Set one SCRIPT_STATE item being returned */ + if (index + 2 > cMaxItems) return E_OUTOFMEMORY; if (pcItems) *pcItems = index + 1;
/* Set SCRIPT_ITEM */
wine was configured with this line :
export CFLAGS="-O2 -march=native -pipe -g -ggdb"; export CXXFLAGS="$CFLAGS="; export CC="ccache gcc"; ./configure
Before that I made a "make distclean" and for the trace a "export WINEDEBUG=+uniscribe".
http://bugs.winehq.org/show_bug.cgi?id=17324
Toralf Förster toralf.foerster@gmx.de changed:
What |Removed |Added ---------------------------------------------------------------------------- Attachment #26692|another creash trace |another crash trace example description|example |
http://bugs.winehq.org/show_bug.cgi?id=17324
Toralf Förster toralf.foerster@gmx.de changed:
What |Removed |Added ---------------------------------------------------------------------------- Attachment #26563|0 |1 is obsolete| | Attachment #26692|0 |1 is obsolete| |
--- Comment #75 from Toralf Förster toralf.foerster@gmx.de 2010-03-09 03:50:34 --- Created an attachment (id=26703) --> (http://bugs.winehq.org/attachment.cgi?id=26703) trace log containing only the relevant part before the crash occurs
therefore much more handy I think.
http://bugs.winehq.org/show_bug.cgi?id=17324
Jeff L lats@yless4u.com.au changed:
What |Removed |Added ---------------------------------------------------------------------------- Attachment #26648|0 |1 is obsolete| |
--- Comment #76 from Jeff L lats@yless4u.com.au 2010-03-10 06:13:15 --- Created an attachment (id=26726) --> (http://bugs.winehq.org/attachment.cgi?id=26726) Patch to change invalid glyphs to 0
The default glyph from an mapped char is 0xffff. ScriptShape does the look up of the glyphs and marks missing glyphs with 0xffff. This causes a crash in xrender in test cases.
I am not sure that this will be a valid patch even if it fixes the problem but if it works we will see as I have a valid test exhibiting the problem.
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #77 from Toralf Förster toralf.foerster@gmx.de 2010-03-10 07:23:30 --- (In reply to comment #76)
Created an attachment (id=26726)
--> (http://bugs.winehq.org/attachment.cgi?id=26726) [details]
Patch to change invalid glyphs to 0
The default glyph from an mapped char is 0xffff. ScriptShape does the look up of the glyphs and marks missing glyphs with 0xffff. This causes a crash in xrender in test cases.
I am not sure that this will be a valid patch even if it fixes the problem but if it works we will see as I have a valid test exhibiting the problem.
The crash went away applying this patch on top on 1.1.40
:-)
http://bugs.winehq.org/show_bug.cgi?id=17324
Jeff L lats@yless4u.com.au changed:
What |Removed |Added ---------------------------------------------------------------------------- Component|-unknown |usp10
--- Comment #78 from Jeff L lats@yless4u.com.au 2010-03-11 07:13:01 --- Patch and test case at http://www.winehq.org/pipermail/wine-patches/2010-March/085655.html
This patch fixes bug 17324 http://bugs.winehq.org/show_bug.cgi?id=17324 The problem is that ScriptShape convert characters to glyphs and there is no translation for '\r'. ScriptTextOut is called and X11DRV_XRender_ExtTextOut is called eventually. The glyph is used as an index into an array but as the glyph value is 0xffff it causes an exception.
This patch converts the NONEXISTING_GLYPHS to 0 (this is the same behaviour as Windows).
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #79 from Toralf Förster toralf.foerster@gmx.de 2010-03-11 07:54:46 --- (In reply to comment #78)
This patch fixes bug 17324 http://bugs.winehq.org/show_bug.cgi?id=17324
Thx Jeff for your work, this nasty bug did annoy me a long time.
http://bugs.winehq.org/show_bug.cgi?id=17324
Toralf Förster toralf.foerster@gmx.de changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED
--- Comment #80 from Toralf Förster toralf.foerster@gmx.de 2010-03-12 02:28:10 --- fixed in svn
http://bugs.winehq.org/show_bug.cgi?id=17324
Hans Leidekker hans@meelstraat.net changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |REOPENED Resolution|FIXED |
--- Comment #81 from Hans Leidekker hans@meelstraat.net 2010-03-12 02:44:05 --- There's no fix in git yet.
http://bugs.winehq.org/show_bug.cgi?id=17324
Jeff L lats@yless4u.com.au changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|REOPENED |RESOLVED Resolution| |FIXED
--- Comment #82 from Jeff L lats@yless4u.com.au 2010-03-17 04:10:48 --- Patch resolving this is at http://www.winehq.org/pipermail/wine-patches/2010-March/085794.html
http://bugs.winehq.org/show_bug.cgi?id=17324
--- Comment #83 from Jeff L lats@yless4u.com.au 2010-03-17 04:13:59 --- Patch resolving this is at http://www.winehq.org/pipermail/wine-cvs/2010-March/064993.html
http://bugs.winehq.org/show_bug.cgi?id=17324
Paul Vriens Paul.Vriens.Wine@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |Paul.Vriens.Wine@gmail.com
--- Comment #84 from Paul Vriens Paul.Vriens.Wine@gmail.com 2010-03-17 07:51:59 --- Fixed by ee891d3b59199daf294c55c7e980035d27d5907f
http://bugs.winehq.org/show_bug.cgi?id=17324
Alexandre Julliard julliard@winehq.org changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED
--- Comment #85 from Alexandre Julliard julliard@winehq.org 2010-03-19 14:10:26 --- Closing bugs fixed in 1.1.41.
-
wine-bugs＠winehq.org