[Bug 24387] New: MS Money 2000: Crashes updating financial informations from the Internet
http://bugs.winehq.org/show_bug.cgi?id=24387 Summary: MS Money 2000: Crashes updating financial informations from the Internet Product: Wine Version: 1.3.2 Platform: x86 OS/Version: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: -unknown AssignedTo: wine-bugs(a)winehq.org ReportedBy: andrea.vai(a)unipv.it Created an attachment (id=30732) --> (http://bugs.winehq.org/attachment.cgi?id=30732) Terminal output. If I click on "update informations from the internet" in the main page of the application, the applications crashes (Application error window "MSMONEY encountered an error and has must be closed" appears. The only way to close the application is to kill it). Problem was not present in wine 1.1.15, so I performed a regression test and the result is: ---------------------------------------------------------------- c5a6b7b36674e8b7f7dc687b01773b88fa7d7312 is the first bad commit commit c5a6b7b36674e8b7f7dc687b01773b88fa7d7312 Author: Juan Lang <juan.lang(a)gmail.com> Date: Tue Mar 30 10:26:48 2010 -0700 rsaenh: Support setting salt via KP_SALT (resend). :040000 040000 d9a20e02d3637c41baadb3131a23907fb5603fe8 51eea86ad80613f612f64bd44a0517c0b89a4894 M dlls ---------------------------------------------------------------- As far as I can notice, the bad commit is the same involved in bug 23208. Fixing this regression would really bring this app towards complete functionality. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=24387 Andrea Vai <andrea.vai(a)unipv.it> changed: What |Removed |Added ---------------------------------------------------------------------------- Keywords| |regression -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=24387 Andrea Vai <andrea.vai(a)unipv.it> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |juan_lang(a)yahoo.com -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=24387 Andrea Vai <andrea.vai(a)unipv.it> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |andrea.vai(a)unipv.it -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=24387 Juan Lang <juan_lang(a)yahoo.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Component|-unknown |rsaenh --- Comment #1 from Juan Lang <juan_lang(a)yahoo.com> 2010-09-13 16:30:17 CDT --- Does reverting that commit fix the crash? If so, please attach a +crypt log. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #2 from Andrea Vai <andrea.vai(a)unipv.it> 2010-09-13 17:39:33 CDT --- (In reply to comment #1)
Does reverting that commit fix the crash? If so, please attach a +crypt log.
yes, I reverted the commit and the crash doesn't happen. Please tell me how to produce a +crypt log. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #3 from Juan Lang <juan_lang(a)yahoo.com> 2010-09-13 17:41:31 CDT --- http://wiki.winehq.org/FAQ#head-16da35b6327024d6ea576e3678488b16862d0f5e Replace +relay,+seh,+tid in the example with +crypt. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #4 from Andrea Vai <andrea.vai(a)unipv.it> 2010-09-13 18:04:41 CDT --- Created an attachment (id=30737) --> (http://bugs.winehq.org/attachment.cgi?id=30737) +Cc -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=24387 Andrea Vai <andrea.vai(a)unipv.it> changed: What |Removed |Added ---------------------------------------------------------------------------- Attachment #30737|+Cc |+crypt Log with the commit description| |*not* reverted -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #5 from Andrea Vai <andrea.vai(a)unipv.it> 2010-09-13 18:06:59 CDT --- Created an attachment (id=30738) --> (http://bugs.winehq.org/attachment.cgi?id=30738) +crypt log after having the commit reverted -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #6 from Juan Lang <juan_lang(a)yahoo.com> 2010-09-13 18:17:51 CDT --- (In reply to comment #5)
Created an attachment (id=30738) --> (http://bugs.winehq.org/attachment.cgi?id=30738) [details] +crypt log after having the commit reverted
Wait, this one also shows a crash: trace:crypt:CryptDecrypt (0x1a0218, 0x0, 1, 00000000, 0x32db30, 0x32d868) wine: Unhandled page fault on read access to 0x0000001c at address 0x29e05801 (thread 0009), starting debugger... Are you sure you attached the correct log? -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #7 from Andrea Vai <andrea.vai(a)unipv.it> 2010-09-14 07:49:25 CDT --- (In reply to comment #6)
(In reply to comment #5)
Created an attachment (id=30738) --> (http://bugs.winehq.org/attachment.cgi?id=30738) [details] [details] +crypt log after having the commit reverted
Wait, this one also shows a crash: trace:crypt:CryptDecrypt (0x1a0218, 0x0, 1, 00000000, 0x32db30, 0x32d868) wine: Unhandled page fault on read access to 0x0000001c at address 0x29e05801 (thread 0009), starting debugger...
Are you sure you attached the correct log?
Yes, I double checked and I am sure that it is the correct log. I was also aware of the crash in this log (call it crash #2), but it is not "the" crash (call it crash #1) I had reported in the bug description, so I didn't know if it should be better to take care of it, file another bug for crash#2, or whatever else. Crash#2 happens slightly after I launch the application: it pops up an "application crash" window, anyway the app itself keeps running, and if I close that window and ignore the crash, and follow steps reported to reproduce crash#1, crash#1 does not happen. That's why I said "I reverted the commit and the crash doesn't happen.". I don't know if crash#2 deserves being investigated or if we can say that crash#1 doesn't depend on it, so we don't care about it. Furthermore, the output of the commit command contains an error (says "1 out of 2 hunks FAILED"), and I attach that output. Last note: if I revert the commit, recompile, rm .wine, launch the setup program of the application, the setup program reports a crash too, but if I ignore it and go on with the setup, it finishes correctly. Should I post some information about this "crash", too??? Hope I didn't make too many mistakes with all of this. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #8 from Andrea Vai <andrea.vai(a)unipv.it> 2010-09-14 07:55:01 CDT --- Created an attachment (id=30749) --> (http://bugs.winehq.org/attachment.cgi?id=30749) Revert commit command output Terminal output of the command git show c5a6b7b36674e8b7f7dc687b01773b88fa7d7312 | patch -p1 -R -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #9 from Andrea Vai <andrea.vai(a)unipv.it> 2010-09-14 07:59:34 CDT --- Created an attachment (id=30750) --> (http://bugs.winehq.org/attachment.cgi?id=30750) rsaenh.c.rej content of dlls/rsaenh/tests/rsaenh.c.rej, mentioned in attachment (id=30749) -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #10 from Juan Lang <juan_lang(a)yahoo.com> 2010-09-14 11:35:11 CDT --- (In reply to comment #7)
I was also aware of the crash in this log (call it crash #2), but it is not "the" crash (call it crash #1) I had reported in the bug description, so I didn't know if it should be better to take care of it, file another bug for crash#2, or whatever else.
Ok. This is fine.
Furthermore, the output of the commit command contains an error (says "1 out of 2 hunks FAILED"), and I attach that output.
This failure is unimportant, it only affects the tests. Thanks for attaching the rsaenh.c.rej so I could verify it.
Hope I didn't make too many mistakes with all of this.
Nope, all seems fine. I have no idea why the crash is occurring, unfortunately. The ironic thing is, this patch was added expressly for MS Money. Could you attach a +relay,+crypt log of the crash? Perhaps that will shed more light. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #11 from Andrea Vai <andrea.vai(a)unipv.it> 2010-09-14 15:30:31 CDT --- Created an attachment (id=30756) --> (http://bugs.winehq.org/attachment.cgi?id=30756) +relay,+crypt log of the crash gzip splitted file, part 1 of 2 -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #12 from Andrea Vai <andrea.vai(a)unipv.it> 2010-09-14 15:31:28 CDT --- Created an attachment (id=30757) --> (http://bugs.winehq.org/attachment.cgi?id=30757) +relay,+crypt log of the crash gzip splitted file, part 2 of 2 -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #13 from Andrea Vai <andrea.vai(a)unipv.it> 2010-09-14 15:51:00 CDT --- (In reply to comment #10)
I have no idea why the crash is occurring, unfortunately. The ironic thing is, this patch was added expressly for MS Money.
I could add some other info if it can help: MS Money itself indeed works better as an "out of the box" app in wine, because without that patch winetricks/ies4linux are needed to install the app (as also pointed out in Bug 23208). At least, I never managed to install it without winetricks/ies4linux and without the patch. So, I would say that from that point of view the app works better (winetricks/ies4linux are also a mess), but is it worth it if the patch introduces so many bugs? I would say no.
Could you attach a +relay,+crypt log of the crash? Perhaps that will shed more light. File attached (37 MB uncompressed, 1.6M zipped, 1M+600k chunks) Thank you for your patience.
-- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #14 from Juan Lang <juan_lang(a)yahoo.com> 2010-09-15 12:58:32 CDT --- Created an attachment (id=30772) --> (http://bugs.winehq.org/attachment.cgi?id=30772) Patch: Revert c5a6b7b36674e8b7f7dc687b01773b88fa7d7312 This patch reverts the offending commit on latest git. Does it remove the crash? Unfortunately it's probably not correct, since it probably reverts bug 6698. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #15 from Andrea Vai <andrea.vai(a)unipv.it> 2010-09-15 15:58:42 CDT --- (In reply to comment #14)
This patch reverts the offending commit on latest git. Does it remove the crash? Yes, it does. I tested it applying it to a fresh git-wine downloaded today. Crash is present in HEAD, but after adding the patch, the crash doesn't occur anymore. Juan, does it definitely means that this is a regression?
Unfortunately it's probably not correct, since it probably reverts bug 6698.
BTW: I had to use winetricks/ies4linux to install the application in the patched git tree. If I don't use winetricks/ies4linux BEFORE running the setup program, then the bug 6698 occurs when I open a mny file. This is exactly the same behavior I mentioned in comment #13 (indeed, my expression "never managed to install" in comment #13 should be read "never managed to open an mny file without getting into bug 6698", because the installation itself works even without winetricks/ies4linux, but when the program is runned and an mny file is opened, the file does not open and the "money is not installed correctly" message appears). I don't know if I have to post more detailed information, maybe that's already too much and in the case sorry for that. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #16 from Juan Lang <juan_lang(a)yahoo.com> 2010-09-15 16:13:11 CDT --- (In reply to comment #15)
Juan, does it definitely means that this is a regression?
No. See your input below:
BTW: I had to use winetricks/ies4linux to install the application in the patched git tree. If I don't use winetricks/ies4linux BEFORE running the setup program, then the bug 6698 occurs when I open a mny file. This is exactly the same behavior I mentioned in comment #13 (indeed, my expression "never managed to install" in comment #13 should be read "never managed to open an mny file without getting into bug 6698", because the installation itself works even without winetricks/ies4linux, but when the program is runned and an mny file is opened, the file does not open and the "money is not installed correctly" message appears).
Basically what you're saying is that this crash doesn't exist without this patch because you can't get far enough in the application without the patch applied. In that case, this really isn't a regression: the patch appears to allow the app to get further than it did before, and this is really a new crash. I haven't been able to discover any relationship between the code and the crash, at any rate. On a hunch, I did the following: I used unpatched Wine (latest git), then I installed dcom98 with winetricks. I removed the DLL overrides for ole32 and rpcrt4, leaving only oleaut32. With this workaround in place, it no longer crashes for me. Can you confirm this works for you? -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=24387 Andrea Vai <andrea.vai(a)unipv.it> changed: What |Removed |Added ---------------------------------------------------------------------------- Keywords|regression | -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #17 from Andrea Vai <andrea.vai(a)unipv.it> 2010-09-15 17:18:08 CDT --- (In reply to comment #16)
Juan, does it definitely means that this is a regression?
No. See your input below:
...
Basically what you're saying is that this crash doesn't exist without this patch because you can't get far enough in the application without the patch applied. In that case, this really isn't a regression: the patch appears to allow the app to get further than it did before, and this is really a new crash. ok, got it, supposing that here you refer to "this patch" as c5a6b7b36674e8b7f7dc687b01773b88fa7d7312, and not as the patch that reverts it, am I right? Just to be sure I understand
On a hunch, I did the following: I used unpatched Wine (latest git), then I installed dcom98 with winetricks. I removed the DLL overrides for ole32 and rpcrt4, leaving only oleaut32. With this workaround in place, it no longer crashes for me. Can you confirm this works for you?
Yes, it works! Thank you! Does it means something useful for getting the bug fixed? -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=24387 Juan Lang <juan_lang(a)yahoo.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|UNCONFIRMED |NEW Component|rsaenh |oleaut32 Ever Confirmed|0 |1 --- Comment #18 from Juan Lang <juan_lang(a)yahoo.com> 2010-09-15 17:28:49 CDT --- (In reply to comment #17)
ok, got it, supposing that here you refer to "this patch" as c5a6b7b36674e8b7f7dc687b01773b88fa7d7312, and not as the patch that reverts it, am I right? Just to be sure I understand
Oops, yes, right you are. Sorry for the confusion.
Yes, it works! Thank you! Does it means something useful for getting the bug fixed?
Good! At least it means we have a better idea where to look. Maybe a +ole,+typelib,+seh,+tid log might be of use, but it's rather outside my area of expertise, so hopefully someone else will take a look. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #19 from Andrea Vai <andrea.vai(a)unipv.it> 2010-09-15 17:39:30 CDT --- (In reply to comment #18)
Maybe a +ole,+typelib,+seh,+tid log might be of use I will post the log asap. Shall I post the log of the "crashing" version, or the one with dcom98 installed? Or maybe both? Thank you
-- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #20 from Juan Lang <juan_lang(a)yahoo.com> 2010-09-15 17:43:09 CDT --- (In reply to comment #19)
I will post the log asap. Shall I post the log of the "crashing" version, or the one with dcom98 installed? Or maybe both? Thank you
The crashing one only, please. (With dcom98 installed, there won't be any output from these channels, as only Wine's ole32/oleaut32 libraries provide these debug channels.) -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #21 from Andrea Vai <andrea.vai(a)unipv.it> 2010-09-17 06:41:58 CDT --- Created an attachment (id=30815) --> (http://bugs.winehq.org/attachment.cgi?id=30815) +ole,+typelib,+seh,+tid log -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #22 from butraxz(a)gmail.com 2013-06-30 08:52:24 CDT --- This ticket has not been updated for over 900 days. Is this still an issue in wine version 1.6-rc4 or higher or is this to be closed as abandoned ? -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #23 from Andrea Vai <andrea.vai(a)unipv.it> 2013-06-30 15:34:40 CDT --- (In reply to comment #22)
This ticket has not been updated for over 900 days.
Is this still an issue in wine version 1.6-rc4 or higher or is this to be closed as abandoned ?
I can try and let know here in the next days. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #24 from Andrea Vai <andrea.vai(a)unipv.it> 2013-07-01 14:42:47 CDT --- The bug still exists in wine 1.6-rc4. I attach a backtrace. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #25 from Andrea Vai <andrea.vai(a)unipv.it> 2013-07-01 14:43:54 CDT --- Created attachment 45068 --> http://bugs.winehq.org/attachment.cgi?id=45068 Backtrace in wine 1.6-rc4 -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #26 from Andrea Vai <andrea.vai(a)unipv.it> --- Bug is still present in wine-1.7.18. Backtrace attached. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #27 from Andrea Vai <andrea.vai(a)unipv.it> --- Created attachment 49142 --> http://bugs.winehq.org/attachment.cgi?id=49142 backtrace in wine-1.7.18 -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #28 from Austin English <austinenglish(a)gmail.com> --- (In reply to Andrea Vai from comment #27)
Created attachment 49142 [details] backtrace in wine-1.7.18
Could you install debugging symbols and attach a new backtrace please? -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #29 from Andrea Vai <andrea.vai(a)unipv.it> --- (In reply to Austin English from comment #28)
Could you install debugging symbols and attach a new backtrace please?
Ok, how can I do that? -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #30 from Austin English <austinenglish(a)gmail.com> --- (In reply to Andrea Vai from comment #29)
(In reply to Austin English from comment #28)
Could you install debugging symbols and attach a new backtrace please?
Ok, how can I do that?
See 'Backtrace Quality' on http://wiki.winehq.org/Backtraces -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #31 from Andrea Vai <andrea.vai(a)unipv.it> --- (In reply to Austin English from comment #30)
Could you install debugging symbols and attach a new backtrace please?
Ok, how can I do that?
See 'Backtrace Quality' on http://wiki.winehq.org/Backtraces
Ok thanks, I hope this is what you need... new backtrace attached -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=24387 Andrea Vai <andrea.vai(a)unipv.it> changed: What |Removed |Added ---------------------------------------------------------------------------- Attachment #45068|0 |1 is obsolete| | Attachment #49142|0 |1 is obsolete| | --- Comment #32 from Andrea Vai <andrea.vai(a)unipv.it> --- Created attachment 49170 --> https://bugs.winehq.org/attachment.cgi?id=49170 Backtrace in 1.7.23 -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #33 from Austin English <austinenglish(a)gmail.com> --- This is your friendly reminder that there has been no bug activity for over a year. Is this still an issue in current (1.7.51 or newer) wine? -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #34 from Andrea Vai <andrea.vai(a)unipv.it> --- I can try and let know here in the next days. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #35 from Andrea Vai <andrea.vai(a)unipv.it> --- Bug is still present in wine-1.7.51 I am sorry, but I am having some problems in attaching a new backtrace: following instructions at http://wiki.winehq.org/Backtraces I find this output from winedbg (*): Wine-dbg>info process pid threads executable (all id:s are in hex) 0000000e 7 'services.exe' 00000022 3 \_ 'winedevice.exe' 0000001b 3 \_ 'plugplay.exe' 00000014 4 \_ 'winedevice.exe' and I don't know which pid to use for attaching the process, as none of them seems related to the program that, indeed, is currently running (msmoney.exe). (*) I now realize that I am (perhaps (**)) using winedbg 1.7.18 (and I can't find a winedbg executable in the compiled 1.7.51 version, except in programs/winedbg/winedbg.exe.so, which gives me segmentation fault if I launch it). Don't know if this matters. (**) I can't find a command line option of winedbg to get its version, so I can only guess that the version is 1.7.18 because this is the version number reported at the bottom of the page generated by typing "man winedbg". Don't know if this is a correct guess. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #36 from Bruno Jesus <00cpxxx(a)gmail.com> --- Please attach a new +rsaenh,+crypt from wine 1.7.50 or later. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #37 from Andrea Vai <andrea.vai(a)unipv.it> --- Created attachment 52736 --> https://bugs.winehq.org/attachment.cgi?id=52736 +rsaenh,+crypt log, from wine-1.7.51 Log attached. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #38 from Bruno Jesus <00cpxxx(a)gmail.com> --- (In reply to Andrea Vai from comment #37)
Created attachment 52736 [details] +rsaenh,+crypt log, from wine-1.7.51
Log attached.
Thanks for the log, as expected there is nothing wrong with the cryptography part. The problem seems to be inside OLE as pointed in comment 18. I cannot help in that area. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #39 from Andrea Vai <andrea.vai(a)unipv.it> --- Still present in 1.8.2 -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=24387 Alistair Leslie-Hughes <leslie_alistair(a)hotmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |leslie_alistair(a)hotmail.com --- Comment #40 from Alistair Leslie-Hughes <leslie_alistair(a)hotmail.com> --- (In reply to Andrea Vai from comment #39)
Still present in 1.8.2
Does wine-staging have the same issue? -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #41 from Andrea Vai <andrea.vai(a)unipv.it> --- (In reply to Alistair Leslie-Hughes from comment #40)
(In reply to Andrea Vai from comment #39)
Still present in 1.8.2
Does wine-staging have the same issue?
Yes, I have verified it right now, using wine-2.13 (Staging) -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #42 from Wolfgang Hartl <wolfo.dev@wolke7.net> --- Created attachment 80396 --> http://bugs.winehq.org/attachment.cgi?id=80396 backtrace after cash Still present in wine 11.0. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #43 from Wolfgang Hartl <wolfo.dev@wolke7.net> --- Created attachment 80397 --> http://bugs.winehq.org/attachment.cgi?id=80397 Log of output with WINEDEBUG=warn+all, exec in wine 11.0 I have the program on Windows 10/11 in parallel and still use it there. Anything else I could provide additionally? -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=24387 Wolfgang Hartl <wolfo.dev@wolke7.net> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |wolfo.dev@wolke7.net --- Comment #44 from Wolfgang Hartl <wolfo.dev@wolke7.net> --- I reproduced this issue and have the following steps and comments: # setup money2000 in fresh wine env $ env WINEPREFIX="/home/wolfgang/.wine-mny" wine SETUP.EXE 2>&1 | tee -a mny2000.setup.log # start money 2000 (1) $ env WINEPREFIX="/home/wolfgang/.wine-mny" wine C:\\Program\ Files\ \(x86\)\\Microsoft\ Money\\System\\MSMONEY.EXE 2>&1 | tee -a mny2000.start1.log # program crashes with click on "update internet information" # log shows also this: 0190:err:typelib:sltg_get_typelib_ref Unable to find reference 0190:fixme:ole:ITypeLib2_Constructor_SLTG Not processing typekind 7 # installing oleaut32 with winetricks $ env WINEPREFIX="/home/wolfgang/.wine-mny" winetricks oleaut32 $ env WINEPREFIX="/home/wolfgang/.wine-mny" winetricks list-installed Using winetricks 20240105 - sha256sum: 17da748ce874adb2ee9fed79d2550c0c58e57d5969cc779a8779301350625c55 with wine-11.0 and WINEARCH=win64 oleaut32 # start money 2000 (2) env WINEPREFIX="/home/wolfgang/.wine-mny" wine C:\\Program\ Files\ \(x86\)\\Microsoft\ Money\\System\\MSMONEY.EXE 2>&1 | tee -a mny2000.start2.log # click on "update internet information" does not crash anymore # ms-money has an INV7.OCX which as a TYPELIB of type TKIND_UNION $ winedump dump INV7.OCX -j resource >mny2000.dump $ grep "typekind 7" mny2000.dump.txt | misc: unknown1 02, flags 0000, unknown2 02, typekind 7 (TKIND_UNION) # checked in the code of wine 11.2 in oleaut32/typelib.c, lines 7474+ # There is the implementation of TKIND_UNION missing This might be the issue for the crash? What is needed to be able to add this typekind and to test it? Anything I can provide out of this INV7.OCX? -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #45 from Wolfgang Hartl <wolfo.dev@wolke7.net> --- Created attachment 80471 --> http://bugs.winehq.org/attachment.cgi?id=80471 oleaut32 typelib - add SLTG_ProcessUnion for TKIND_UNION -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #46 from Wolfgang Hartl <wolfo.dev@wolke7.net> --- Created attachment 80473 --> http://bugs.winehq.org/attachment.cgi?id=80473 oleaut32 typelib - VT_RECORD, VT_UNION handing -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #47 from Wolfgang Hartl <wolfo.dev@wolke7.net> --- I’ve added an implementation for TKIND_UNION in typelib.c. A UNION has the same structure as a RECORD, so I introduced a call to SLTG_ProcessUnion() and implemented the function itself, mirroring the behavior of SLTG_ProcessRecord(). The SLTG_DoVars() function simply reads whatever data is present in the union. Since all typekinds in the switch() are now handled, I’m not sure whether the FIXME() in the default: case is still appropriate. Additionally, in function typedescvt_to_variantvt() / userdefined_to_variantvt(), I also replaced the FIXME for TKIND_RECORD and TKIND_UNION with the necessary assignments. I git-cloned the Wine repository, applied these two fixes, and created patches (see attachments). I then tried to test the changes. I finally was able to compile Wine locally, but I’m still struggling. I can run the locally compiled wine, though I’m unsure whether doing so interferes with my stable Wine 11 installation and everything installed there. Additionally, the compiled Wine (even without my fix) does not behave the same way as the version installed via apt. My compiled version shows other issues, such as: "libEGL warning: pci id...", it does not behave well. At least, when installing Microsoft Money 2000 with the patched, self-compiled version, the error: "fixme:ole:ITypeLib2_Constructor_SLTG Not processing typekind 7" no longer appears. I also haven’t yet figured out the best way to do unit or component testing for the specific changes I made. It would be great if one could apply the patches and test them (hopefully more easily). In the meantime, I’ll keep digging and learning, any hints or guidance are appreciated. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #48 from Wolfgang Hartl <wolfo.dev@wolke7.net> --- I had some progress. I managed now to have a proper own compiled version 11.0 of wine. Seems, the issue was that I took 11.3 as code base, which behave differently. Anyway. My assumptions about the missing TKIND_UNION implementation were wrong! I was able to fix a missing implementation, so that the FIXME log is gone; but it does not solve this crash. Neither was the crash solved by another fix that I was able to make for TKIND_RECORD, that solved multiple errors of "0170:err:typelib:sltg_get_typelib_ref Unable to find reference" when the record has a reference to another user-defined type. It seems that referencing the GUID {3e9cf0bc-53f5-11d2-9f2d-0000f87a581e} causes the crash.
err:ole:CoGetClassObject no class object {3e9cf0bc-53f5-11d2-9f2d-0000f87a581e} could be created for context 0x1
It is not registered, not in wine and not in windows too. I checked. So, windows ignores the error gracefully, wine with native oleaut32 also. Digging deeper... -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=24387 Wolfgang Hartl <wolfo.dev@wolke7.net> changed: What |Removed |Added ---------------------------------------------------------------------------- Attachment #80397|0 |1 is obsolete| | --- Comment #49 from Wolfgang Hartl <wolfo.dev@wolke7.net> --- Created attachment 80585 --> http://bugs.winehq.org/attachment.cgi?id=80585 Crash log with relevant root cause Finally, I found the issue! with the following run: env "WINEPREFIX=/home/wolfgang/.wine110wow" "WINEDEBUG=+ole,+typelib" ./wine "C:\\Program Files (x86)\\Microsoft Money\\System\\MSMONEY.EXE" 2>&1 | tee -a ~/Local/log/run18.wow.ole.typelib.log The INV7.OCX calls ITypeInfo_fnInvoke "CreateProperties(pva, retval)", the arg 0 requests type VT_VARIANT|VT_ARRAY|VT_BYREF. The ITypeInfo_fnInvoke() gets it wrong and gives the DispCallFunc() the arg 0 type VT_VARIANT|VT_BYREF to VT_ERROR. The inv7.ocx crashes due to the wrong ptr. 0140:trace:ole:DispCallFunc (07CD0490, 44, 4, 10, 2, 064AC610, 064AC608, 0792FB48 (vt=10)) 0140:trace:ole:DispCallFunc arg 0: type VT_VARIANT|VT_ARRAY|VT_BYREF 064AC5C8 {VT_VARIANT|VT_BYREF: 064AC5E8 {VT_ERROR: 80020004}} 0140:trace:ole:DispCallFunc arg 1: type VT_VARIANT|VT_BYREF 064AC5D8 {VT_VARIANT|VT_BYREF: 0792FB38 {VT_EMPTY}} wine: Unhandled page fault on read access to 00000016 at address 600CDC79 (thread 0140), starting debugger... I have a fix that gives the correct argument and fixes the crash, but I am still not sure, if this is the correct way. Analyzing further. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #50 from Wolfgang Hartl <wolfo.dev@wolke7.net> --- Okay, one step forward - two steps back. The "found" issue again might be a side effect of a different issue. Another test failed, I was testing if the requested "Optional" Ptr to safearray of variants can be given back. No, it can not. In the call in ITypeInfo_fnInvoke "CreateProperties(pva, retval)", the arg 0 is built as OPTIONAL type VT_VARIANT|VT_ARRAY|VT_BYREF from the SLTG-typelib. The assumption was: The ITypeInfo_fnInvoke() gets it wrong and gives the DispCallFunc() the arg 0 type VT_VARIANT|VT_BYREF to VT_ERROR. Actually, that is correct - according to MS specs: "The [optional] attribute is valid only if the parameter is of type VARIANT or VARIANTA" I see are following options: 1) The implementation of INV7 did not follow the specs and is somehow mail-functioned, but the native OLEAUT32 still handles this correctly. 2) The marker "optional" as delivered or read by wine via the SLTG lib is not correct or there is special case not handled correctly. Tending now more to the second scenario. It would be great if anyone is available and interested to discuss this with me. To be continued... -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=24387 --- Comment #51 from Wolfgang Hartl <wolfo.dev@wolke7.net> --- The MS-Money lib mnyolinv (inv7.ocx) calls CreateProperties(). In the SLTG typelib of inv7.ocx it is coded that CreateProperties() wants to have vararg, but cParamsOpt in STLG is coded as 6 bit field, and inv7.ocx gives "111111" = 63 as cParamsOpt. Wine wrongly prepares the optional Parameter (VT_ERROR in VT_VARIANT instead of an empty SafeArray) with assuming - all parameters are "optional" inv7 crashes immediately when accessing the wrong delivered parameter and the CreateProperties() gets never executed. VARARG in MSFT is defined with cParamsOpt = -1, that means that the delivered cParamsOpt=63 from SLTG needs to be interpreted as sentinel value for -1. When that is done, the call for CreateProperties() works as requested, a VT_VARIANT|VT_ARRAY|VT_BYREF - a pointer to a SafeArray of VARIANTs is given to the function and the crash does not happen anymore. I'll prepare a patch. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
participants (2)
-
wine-bugs＠winehq.org -
WineHQ Bugzilla