http://bugs.winehq.org/show_bug.cgi?id=13599
Summary: Can't install Autocad 2005 Product: Wine Version: unspecified Platform: PC-x86-64 OS/Version: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: -unknown AssignedTo: wine-bugs@winehq.org ReportedBy: max@veneto.com
Created an attachment (id=13568) --> (http://bugs.winehq.org/attachment.cgi?id=13568) warn+all last log lines
Install hangs on VBA setup, even if I disable vba setup in install options. I did a log with warn+all, attached the (imho) useful lines.
I can do relay logs if needed.
Regards
Max
http://bugs.winehq.org/show_bug.cgi?id=13599
max@veneto.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Version|unspecified |CVS/GIT
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #1 from Juan Lang juan_lang@yahoo.com 2008-06-02 10:21:33 --- Please don't post warn+all logs unless asked to, they're really distracting. The console output with no flags is usually best unless prompted for something else.
http://bugs.winehq.org/show_bug.cgi?id=13599
max@veneto.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Attachment #13568|0 |1 is obsolete| |
--- Comment #2 from max@veneto.com 2008-06-02 13:39:05 --- Created an attachment (id=13577) --> (http://bugs.winehq.org/attachment.cgi?id=13577) First install run
http://bugs.winehq.org/show_bug.cgi?id=13599
max@veneto.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Attachment #13577|application/octet-stream |text/plain mime type| |
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #3 from max@veneto.com 2008-06-02 13:41:28 --- Created an attachment (id=13578) --> (http://bugs.winehq.org/attachment.cgi?id=13578) repair installation log
http://bugs.winehq.org/show_bug.cgi?id=13599
max@veneto.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Attachment #13578|application/octet-stream |text/plain mime type| |
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #4 from max@veneto.com 2008-06-02 13:44:02 --- Previous 2 attachments are just console outputs; the first one is from a fresh install : I started with a fresh wine three, installed dotnet1.1 with winetricks, then setup autocad.
Second one is a second setup run, on which I could choose "repair" option.
Regards
Max
http://bugs.winehq.org/show_bug.cgi?id=13599
Louis Lenders xerox_xerox2000@yahoo.co.uk changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |xerox_xerox2000@yahoo.co.uk Status|UNCONFIRMED |NEW Ever Confirmed|0 |1
--- Comment #5 from Louis Lenders xerox_xerox2000@yahoo.co.uk 2008-06-08 06:48:17 --- Confirming, as this bug is also present in Autodesk 2006 and 2004 trial. In Autodesk 2006 , after you deselect "VBA" the installer finishes fine, and i got the app running. In Autodesk 2004, after deselecting vba it still tries to install it and the installer crashes similar as the srash from comment 3
http://bugs.winehq.org/show_bug.cgi?id=13599
Louis Lenders xerox_xerox2000@yahoo.co.uk changed:
What |Removed |Added ---------------------------------------------------------------------------- URL| |http://www.brothersoft.com/a | |utocad-78351.html Keywords| |download, Installer
--- Comment #6 from Louis Lenders xerox_xerox2000@yahoo.co.uk 2008-06-08 06:50:02 --- Added doenloadlink for version 2004, that has similar behaviour
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #7 from max@veneto.com 2008-06-08 07:55:29 --- (In reply to comment #5)
Confirming, as this bug is also present in Autodesk 2006 and 2004 trial. In Autodesk 2006 , after you deselect "VBA" the installer finishes fine, and i got the app running. In Autodesk 2004, after deselecting vba it still tries to install it and the installer crashes similar as the srash from comment 3
Well, you gave me really 2 good news ! 1) At leas one autocad after 2005 can be installed, so I can start testing it 2) The installer problem is on VBA installation side, so we could start solving it
Thanx
Max
http://bugs.winehq.org/show_bug.cgi?id=13599
Louis Lenders xerox_xerox2000@yahoo.co.uk changed:
What |Removed |Added ---------------------------------------------------------------------------- Component|-unknown |msi
--- Comment #8 from Louis Lenders xerox_xerox2000@yahoo.co.uk 2008-06-08 10:18:40 --- Then i'll change component to msi. Not 100% sure if the crash is msi-bug, but the deselecting of VBA in the installer doesn't work either, and i guess that's an msi-bug for sure.
I'll do some updating in the appdb of the autocad-trials. So far i got the trials of 2004 and 2006 running. Currently i'm downloading 2007 trial.
I cannot find a trial of AutoCAD2005 anywhere, you have an idea of it's still available?
http://bugs.winehq.org/show_bug.cgi?id=13599
James Hawkins truiken@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Component|msi |-unknown
--- Comment #9 from James Hawkins truiken@gmail.com 2008-06-08 10:48:23 --- Open a new bug for new issues.
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #10 from Louis Lenders xerox_xerox2000@yahoo.co.uk 2008-06-08 10:49:42 --- Created an attachment (id=13812) --> (http://bugs.winehq.org/attachment.cgi?id=13812) very dirty hack
As a last note, i attached the hack with which i was able to install AutoCAD2004. Maybe you could try if it works for your problem too, then at least you would be able to test the application yourself
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #11 from max@veneto.com 2008-06-08 12:28:49 --- (In reply to comment #9)
Open a new bug for new issues.
Well, I think that one IS the issue. Installer hangs on same place on 2004, 2005, 2006, and 2007 (last one, IIRC). Setting to msi is the correct behavious, imho, at least if none better is found.
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #12 from James Hawkins truiken@gmail.com 2008-06-08 12:31:57 --- Are you referring to the softwareEmulation custom action?
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #13 from Louis Lenders xerox_xerox2000@yahoo.co.uk 2008-06-08 12:34:38 ---
Are you referring to the softwareEmulation custom action?
Can't speak for Max, but is it a msi-bug James?
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #14 from James Hawkins truiken@gmail.com 2008-06-08 12:36:18 --- No, not the softwareEmulation custom action. That's why I said open new bug for new issues (not being able to deselect the VBA feature).
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #15 from max@veneto.com 2008-06-08 12:43:05 --- (In reply to comment #11)
(In reply to comment #9)
Open a new bug for new issues.
Well, I think that one IS the issue. Installer hangs on same place on 2004, 2005, 2006, and 2007 (last one, IIRC). Setting to msi is the correct behavious, imho, at least if none better is found.
well, good job with your 'dirty hack', now autocad 2005 installs and starts, even if it gives an error on startup, then crashes. But that's good, at least we can start it and try to find bugs on it. It crashes because an unimplemented function on gdiplus, but here we could open another bug.... But I'm willing to test it a bit more before doing it.
About the installer, I'm not sure if the "vba deselection" removes completely the vba stuff or just the vba ide... I should test it on windows. My feelings are that parts of vba are still needed somewhere, so vba runtime gets installed even deselecting it.
Max
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #16 from max@veneto.com 2008-06-08 12:48:31 --- Well, Louis... you did a really great job !!! Installing gdiplus with winetricks now autocad 2005 starts... missing icons and so, but I can open a drawing and work on it !
Now, I think we should focus on an autocad release we all have access and file another bug... and try to solve all issues about it. What do you think about ?
Max
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #17 from Louis Lenders xerox_xerox2000@yahoo.co.uk 2008-06-08 12:51:10 ---
Now, I think we should focus on an autocad release we all have access and file another bug... and try to solve all issues about it. What do you think about ?
that's great, if you do testing, then please with autocad2006, as there's a download for the trial for that one. I'm not really familiar with the program itsself , so of you could do some thorough testing that would be great
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #18 from max@veneto.com 2008-06-08 12:57:39 --- (In reply to comment #17)
Now, I think we should focus on an autocad release we all have access and file another bug... and try to solve all issues about it. What do you think about ?
that's great, if you do testing, then please with autocad2006, as there's a download for the trial for that one. I'm not really familiar with the program itsself , so of you could do some thorough testing that would be great
Perfect, so I'll try to install autocad 2006 with your hack and file a new bug for it. Autocad2005 isn't missing to many stuffs, and seems to me not difficult to solve, so I guess we're on the right way :-)
Ciao
Max
http://bugs.winehq.org/show_bug.cgi?id=13599
Louis Lenders xerox_xerox2000@yahoo.co.uk changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |dominikowski@gmail.com
--- Comment #19 from Louis Lenders xerox_xerox2000@yahoo.co.uk 2008-06-08 13:01:02 --- *** Bug 12593 has been marked as a duplicate of this bug. ***
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #20 from max@veneto.com 2008-06-08 13:13:49 --- Sorry but AutoCAD 2006 crashes early with this one :
fixme:ntdll:NtPowerInformation Unimplemented NtPowerInformation action: 16 fixme:dbghelp:dump_system_info fill in CPU vendorID and feature set fixme:ntdll:NtPowerInformation Unimplemented NtPowerInformation action: 16 fixme:ntdll:NtPowerInformation Unimplemented NtPowerInformation action: 16 fixme:advapi:RegisterEventSourceW ((null),L".NET Runtime Optimization Service"): stub fixme:advapi:ReportEventW (0xcafe4242,0x0001,0x0000,0x00000457,(nil),0x0001,0x00000000,0x7e5d3254,(nil)): stub err:eventlog:ReportEventW L".NET Runtime Optimization Service (clr_optimization_v2.0.50727_32) - Service reached limit of transient errors. Will shut down. Last error returned from Service Manager: 0x80070005.\n" fixme:advapi:DeregisterEventSource (0xcafe4242) stub fixme:advapi:RegisterEventSourceW ((null),L".NET Runtime Optimization Service"): stub fixme:advapi:ReportEventW (0xcafe4242,0x0004,0x0000,0x00000454,(nil),0x0001,0x00000000,0x7e5d32c8,(nil)): stub fixme:advapi:DeregisterEventSource (0xcafe4242) stub
(again, posting here instead of attach because it's short !)
I'd prefere to work on 2005, which is starting and operating to some extent... But if you want I'll fill bug for 2006 and we try to cope with that one. Didn't you find a download site for 2005 trial ?
Max
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #21 from Louis Lenders xerox_xerox2000@yahoo.co.uk 2008-06-08 13:25:34 ---
Sorry but AutoCAD 2006 crashes early with this one :
hmm weird, didn't crash here at all. Anyway, if you do some testing for 2005 please also add your findings to appdb: http://appdb.winehq.org/objectManager.php?sClass=version&iId=4105 Thamks
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #22 from max@veneto.com 2008-06-08 13:32:01 --- (In reply to comment #21)
Sorry but AutoCAD 2006 crashes early with this one :
hmm weird, didn't crash here at all. Anyway, if you do some testing for 2005 please also add your findings to appdb: http://appdb.winehq.org/objectManager.php?sClass=version&iId=4105 Thamks
Ok, I'll focus on 2005 for now, and I'll put an howto on appdb. I'm using autocad 2005 on my daily job so, once it can be used to some extents I can also make deep tests on it. BTW, I'd be happy to have the setup error solved sometimes... but I think I'll leave it to msi people :-) Once 2005 runs, we can go on on 2006+.
Ciao and thanx for all
Max
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #23 from max@veneto.com 2008-06-08 18:05:16 --- Luis, your "very dirty hack" uncovered the real Autocad stopper, I guess. It solved also my Autocad 2008 setup problem, but not the 2007 one, wich I think comes from a different bug.
So, up to now, this bug is related to 2004, 2005, 2006 and 2008 setup problems. Solving this one will allow (imho) to install any of the products above. I'll post more on appdb tomorrow.
So, now I can run 2005, which behaves quite well, and 2008, which is missing toolbars and other stuffs, but I guess it's easy solvable. 2006 installs but crashes, 2007 still don't install. Performances are quite good besides of text rendering, which is really bad... I think I'll file a bug about it on next days.
Ciao
Max
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #24 from Louis Lenders xerox_xerox2000@yahoo.co.uk 2008-06-14 04:28:53 --- It looks to me as if things go wrong here:
0064:Call opengl32.glGetString(00001f01) ret=1000cd58 0064:trace:seh:raise_exception code=c0000005 flags=0 addr=0x60ec12f6 0064:trace:seh:raise_exception info[0]=00000000 . . . . . 0064:trace:seh:__regs_RtlUnwind handler at 0x1004008b returned 1 0064:Ret ntdll.RtlUnwind() retval=00000000 ret=6034dc13 0064: eax=00000000 ebx=603b3ce4 ecx=6034dbb0 edx=7bc38020 esi=7ed097c8 edi=7ed09104 ebp=7ed08cd8 esp=7ed08cb8 ds=007b es=007b fs=0033 gs=003b flags=00000246 0064:err:msi:ACTION_CallDllFunction Custom action (L"C:\windows\temp\msifed2.tmp":L"softwareEmulation") caused a page fault: c0000005
Maybe AutoCad is not happy with what glGetString returns??
I'll do some more digging with opengl debug channels enabled, ans attach logs
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #25 from Dmitry Timoshkov dmitry@codeweavers.com 2008-06-14 04:44:30 --- It's glGetString who crashes.
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #26 from max@veneto.com 2008-06-14 04:55:22 --- (In reply to comment #25)
It's glGetString who crashes.
from what do you see that ?
Max
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #27 from Dmitry Timoshkov dmitry@codeweavers.com 2008-06-14 05:00:57 --- (In reply to comment #26)
(In reply to comment #25)
It's glGetString who crashes.
from what do you see that ?
From comment #24
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #28 from max@veneto.com 2008-06-14 05:35:08 --- (In reply to comment #27)
(In reply to comment #26)
(In reply to comment #25)
It's glGetString who crashes.
from what do you see that ?
From comment #24
const GLubyte * glGetString( GLenum name )
I don't understand from where in #24 can you see glGetString crashing... All what I see is
0064:Call opengl32.glGetString(00001f01) ret=1000cd58
glGetString gets 0x1f01 as parameter and return a pointer to a GLubyte, which is not null.... so no crash. Am I wrong ?
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #29 from Dmitry Timoshkov dmitry@codeweavers.com 2008-06-14 05:41:31 --- (In reply to comment #28)
0064:Call opengl32.glGetString(00001f01) ret=1000cd58
ret=xxx above is the caller's address. When an API calls returns you see something like
0064:Ret ntdll.RtlUnwind() retval=00000000 ret=6034dc13
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #30 from max@veneto.com 2008-06-14 05:45:15 --- (In reply to comment #29)
(In reply to comment #28)
0064:Call opengl32.glGetString(00001f01) ret=1000cd58
ret=xxx above is the caller's address. When an API calls returns you see something like
0064:Ret ntdll.RtlUnwind() retval=00000000 ret=6034dc13
so the missing 'retval=' indicates a crash inside function ? I didn't know that one... Thank you for your hint ! ;-)
Max
http://bugs.winehq.org/show_bug.cgi?id=13599
Louis Lenders xerox_xerox2000@yahoo.co.uk changed:
What |Removed |Added ---------------------------------------------------------------------------- Component|-unknown |opengl
--- Comment #31 from Louis Lenders xerox_xerox2000@yahoo.co.uk 2008-06-14 07:47:49 --- i'll change component to opengl. Apparently the problem is in glGetString. Simple hack there (return just 0 for GL_RENDERER ) made the installer finish fine. Maybe Roderick could have a look/help out?
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #32 from Roderick Colenbrander thunderbird2k@gmx.net 2008-06-14 08:34:26 --- Likely there is no opengl context around causing the call to crash. Try to attach a +wgl log (WINEDEBUG=+wgl wine appname.exe &> log). Of course also make sure that you have proper gl drivers installed on your system.
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #33 from Louis Lenders xerox_xerox2000@yahoo.co.uk 2008-06-14 09:50:03 --- I don't even see a call to glGetString in a +wgl log:
trace:wgl:wglGetProcAddress func: 'wglFinish' trace:wgl:X11DRV_wglGetProcAddress ('wglFinish'): (0x60a62f90) - WineGL trace:wgl:wglGetProcAddress func: 'wglFlush' trace:wgl:X11DRV_wglGetProcAddress ('wglFlush'): (0x60a64d60) - WineGL err:msi:ACTION_CallDllFunction Custom action (L"C:\windows\temp\msi58df.tmp":L"softwareEmulation") caused a page fault: c0000005 err:ntdll:RtlpWaitForCriticalSection section 0x60a9fb20 "x11drv_main.c: X11DRV_CritSection" wait timed out in thread 0028, blocked by 0064, retrying (60 sec)
However in a +relay log it's right there before the crash. So what now?
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #34 from Louis Lenders xerox_xerox2000@yahoo.co.uk 2008-06-14 09:51:35 --- Created an attachment (id=14003) --> (http://bugs.winehq.org/attachment.cgi?id=14003) exception caught
Just to make sure the bug is really in glGetString: adding an exception handler by attached patch makes AutoCad finish just fine
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #35 from Louis Lenders xerox_xerox2000@yahoo.co.uk 2008-06-14 10:06:15 --- Created an attachment (id=14005) --> (http://bugs.winehq.org/attachment.cgi?id=14005) +wgl trace
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #36 from max@veneto.com 2008-06-14 10:53:18 --- (In reply to comment #34)
Created an attachment (id=14003)
--> (http://bugs.winehq.org/attachment.cgi?id=14003) [details]
exception caught
Just to make sure the bug is really in glGetString: adding an exception handler by attached patch makes AutoCad finish just fine
I'm wondering why autocad installer needs opengl at all... at least, with autocad2005 the installer doesn't do graphics at all. Maybe it just asks for opengl info to set up display configuration... and maybe on windows it doesn't crash even with no gl context active, as does on linux.
But imho the proposed solution should be good... in glGetString() call documentation there's no hint on valid opengl context.
Maybe a good thing would be to code a small test app and run on windows, to see if it crashes too or the error is silently ignored.
OTOH, you patch does no harm, just avoiding a crash that would otherwise occurr, so i guess it could be added for 1.0 !
Ciao
Max
http://bugs.winehq.org/show_bug.cgi?id=13599
Austin English austinenglish@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Keywords| |patch
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #37 from Dmitry Timoshkov dmitry@codeweavers.com 2008-06-14 23:22:50 --- (In reply to comment #36)
I'm wondering why autocad installer needs opengl at all... at least, with autocad2005 the installer doesn't do graphics at all.
Most likely the installer does some checks whether the system is suitable for running Autocad itself. I suspect that Autocad will crash for you as well once it's installed. Do you have a working opengl setup? Does glxgears work?
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #38 from Louis Lenders xerox_xerox2000@yahoo.co.uk 2008-06-15 02:03:14 ---
(In reply to comment #37) I suspect that Autocad will crash for you as well once it's installed.
No it doesn't, and moreover it seems to run just fine, though i'm not really familiar with how the app works, but max did some indeep testing and it seems to run quite stable.
That's why i think this bug needs some more attention from the hardcore wine-developers;
Quote from the wiki-page:
"Autocad was the number two most requested application in Novell's LinuxApplicatonRequestSurvey"
Now if even i can make the application work with a simple hack, it shouldn't be too difficult for one of the wine-developers to solve this bug in git i'd say ;) (I tried to write a simple test myself, but that wouldn't crash, so i'm currently stuck)
This bug has prevented people to run all versions from autocad-2004 up to 2008. If this would be fixed more people would be happy again...
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #39 from max@veneto.com 2008-06-15 03:12:22 --- (In reply to comment #37)
(In reply to comment #36)
I'm wondering why autocad installer needs opengl at all... at least, with autocad2005 the installer doesn't do graphics at all.
Most likely the installer does some checks whether the system is suitable for running Autocad itself. I suspect that Autocad will crash for you as well once it's installed.----
Nope. I'm using, thanx to Louis's hack, Autocad for my daily job. I'd say that I tested 100% of it. Before the hack I had to use a virtual machine just to run it.
What's the worst is that, from answers in forums and here, it seemed to me that running a decent version of autocad was almost impossible in wine, and I was giving up.... then Louis solved it at once with a really simple hack.
I can't really believe that an application that has the second/third place on user's requests deserves so few attention from wine devels. There are tons of civil ingeneers/architects that would to go to linux and can't because of lack of an Autocad alternative.... and we're focusing on making M$ Access/Outlook or IE ??? That's what I really can't understand.
Well, back to Autocad. Once installed, it runs perfectly, see at AppDB, I posted a little howto. You need some dll overrides to make it ALL work, and there are a couple of (IMHO) dumb bugs to solve before it can go to gold/platinum status.
The OpenGL problem is ininfluent after setup. And, IMHO, Louis solution of catching the exception on glGetString() code is right, as glGetString() shouldn't crash anyways, just return an error code if called in wrong environment.
Max
http://bugs.winehq.org/show_bug.cgi?id=13599
guillermo guillemarchant@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |guillemarchant@gmail.com
--- Comment #40 from guillermo guillemarchant@gmail.com 2008-06-18 18:07:06 --- Max:
I`ve been looking at APPDb for the HOWTO that you mentioned at your last post and I can`t find it. Can you post the link to the HOWTO here?
thanks
Guillermo
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #41 from TomaszD dominikowski@gmail.com 2008-06-19 12:01:52 --- (In reply to comment #40)
Max:
I`ve been looking at APPDb for the HOWTO that you mentioned at your last post and I can`t find it. Can you post the link to the HOWTO here?
thanks
Guillermo
It's here:
http://appdb.winehq.org/objectManager.php?sClass=version&iId=4105&iT...
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #41 from TomaszD dominikowski@gmail.com 2008-06-19 12:01:52 --- (In reply to comment #40)
Max:
I`ve been looking at APPDb for the HOWTO that you mentioned at your last post and I can`t find it. Can you post the link to the HOWTO here?
thanks
Guillermo
It's here:
http://appdb.winehq.org/objectManager.php?sClass=version&iId=4105&iT...
--- Comment #42 from TomaszD dominikowski@gmail.com 2008-06-19 12:19:09 --- So is Luis's patch headed towards upstream or do we have to wait for something or someone?
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #41 from TomaszD dominikowski@gmail.com 2008-06-19 12:01:52 --- (In reply to comment #40)
Max:
I`ve been looking at APPDb for the HOWTO that you mentioned at your last post and I can`t find it. Can you post the link to the HOWTO here?
thanks
Guillermo
It's here:
http://appdb.winehq.org/objectManager.php?sClass=version&iId=4105&iT...
--- Comment #42 from TomaszD dominikowski@gmail.com 2008-06-19 12:19:09 --- So is Luis's patch headed towards upstream or do we have to wait for something or someone?
--- Comment #43 from max@veneto.com 2008-06-22 10:30:21 --- (In reply to comment #42)
So is Luis's patch headed towards upstream or do we have to wait for something or someone?
I'd like to know it too... It seems to me that Louis patch is correct and can't cause regressions, so I guess it could be applied.
There's also another problem preventing correct install : http://bugs.winehq.org/show_bug.cgi?id=13838 I made a patch for it, but it should be tested on windows for correctness, which I can't.... any volunteer ? :-) Last but not least.... we still need native shlwapi.dll to show "register now" dialog; after that one autocad 2005 could gain status of platinum, imho.
Max
http://bugs.winehq.org/show_bug.cgi?id=13599
max@veneto.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |max@veneto.com
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #44 from max@veneto.com 2008-06-26 03:34:39 --- Any news on that one ? I've got many requests from users interested to run autocad on a unpatched wine. What should I do to have the 2 needed patches inside wine ?
Best Regards
Max
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #45 from TomaszD dominikowski@gmail.com 2008-06-26 03:40:26 --- (In reply to comment #44)
Any news on that one ? I've got many requests from users interested to run autocad on a unpatched wine. What should I do to have the 2 needed patches inside wine ?
Best Regards
Max
I don't know, send them to wine-patches mailing list maybe? :]
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #46 from Louis Lenders xerox_xerox2000@yahoo.co.uk 2008-06-26 03:54:55 --- i already sent the 1st one, let's just see what happens
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #47 from Louis Lenders xerox_xerox2000@yahoo.co.uk 2008-06-27 14:26:19 --- Apparently the patch is not going to be accepted, so that's where it ends for now i guess. I fear we can just wait a whole long time, before anyone else will take this up.
I will attach at least the crash log, and a log where the setup ended succesfully after removing the GLENTER/GLLEAVE calls around glGetString (but according to roderick that wasn't an acceptable patch either)
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #48 from Louis Lenders xerox_xerox2000@yahoo.co.uk 2008-06-27 14:29:11 --- Created an attachment (id=14398) --> (http://bugs.winehq.org/attachment.cgi?id=14398) crash log (Install unsuccessfull)
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #49 from Louis Lenders xerox_xerox2000@yahoo.co.uk 2008-06-27 14:30:29 --- Created an attachment (id=14399) --> (http://bugs.winehq.org/attachment.cgi?id=14399) log after removing GLENTER/GLLEAVE (installation successfull)
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #50 from max@veneto.com 2008-06-27 14:36:31 --- (In reply to comment #47)
Apparently the patch is not going to be accepted, so that's where it ends for now i guess. I fear we can just wait a whole long time, before anyone else will take this up.
as usual... That's becoming ridiculous. Autocad is one of most requested apps and still no way to progress from 2000 version without hand-made patches.
I will attach at least the crash log, and a log where the setup ended succesfully after removing the GLENTER/GLLEAVE calls around glGetString (but according to roderick that wasn't an acceptable patch either)
I'll try to make a crashing testcase for that one. But I still don't understand which kind or harm can your patch bring to wine.... It just avoid a crash on an undefined situation, it can't change working apps behaviour.
Max
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #51 from max@veneto.com 2008-06-28 03:25:50 --- (In reply to comment #49)
Created an attachment (id=14399)
--> (http://bugs.winehq.org/attachment.cgi?id=14399) [details]
log after removing GLENTER/GLLEAVE (installation successfull)
The problem is indeed that one... You can see removing the ENTER:GL() call. GL_ENTER() sets up a lock on X11 which is removed by LEAVE_GL(). In case of an exception after GL_ENTER() but before GL_LEAVE(), the lock is not freed and any following X11 call will fail.
In autocad installer the exception itself isn't harmful, but the lock prevents the installer to continue; because of that there's a long delay (waiting for lock to go away) before installer abort.
I'll try to make a patch that frees the lock in case of an exception.
Max
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #52 from max@veneto.com 2008-06-28 04:46:20 --- Created an attachment (id=14419) --> (http://bugs.winehq.org/attachment.cgi?id=14419) Free X11 lock after an OpenGL exception
Louis, I made a "more generic" patch on wgl that frees the lock made by GL_ENTER() on any exception caused by opengl code. That one will avoid the endless waiting for freed lock (and finally the crash) found in autocad installer.
Patch is already posted on wine-patches.... let's see if that one is accepted ! :-)
Max
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #53 from max@veneto.com 2008-06-28 09:58:47 --- (In reply to comment #47)
Apparently the patch is not going to be accepted, so that's where it ends for now i guess. I fear we can just wait a whole long time, before anyone else will take this up.
I will attach at least the crash log, and a log where the setup ended succesfully after removing the GLENTER/GLLEAVE calls around glGetString (but according to roderick that wasn't an acceptable patch either)
Well, I start agreeing with you. Your patch didn't go, and also mine didn't.
I give up and run my autocad on my patched wine, and leave the 100-lines howto on appdb for other people that wanna try the same.
Max
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #54 from Roderick Colenbrander thunderbird2k@gmx.net 2008-06-29 04:37:28 --- Created an attachment (id=14441) --> (http://bugs.winehq.org/attachment.cgi?id=14441) +wgl,+opengl,+seh,+x11drv
Very strange stuff is happening. For instance the app is receiving a gl_drawable but then a next call it can be gone again: 0013:trace:x11drv:X11DRV_ExtEscape SET_DRAWABLE hdc 0x51d8 drawable 3600045 gl_drawable 0 pf 0 dc_rect (72,95)-(423,134) drawable_rect (268,209 )-(756,560) 0013:trace:x11drv:X11DRV_ExtEscape SET_DRAWABLE hdc 0x51d8 drawable 3600045 gl_drawable 0 pf 0 dc_rect (72,95)-(423,134) drawable_rect (268,209 )-(756,560) 0013:trace:x11drv:X11DRV_ExtEscape SET_DRAWABLE hdc 0x51d8 drawable 3600045 gl_drawable 3600045 pf 0 dc_rect (0,0)-(488,351) drawable_rect (268 ,209)-(756,560) 0013:trace:x11drv:X11DRV_ExtEscape SET_DRAWABLE hdc 0x51d8 drawable 3600045 gl_drawable 0 pf 0 dc_rect (72,95)-(423,134) drawable_rect (268,209 )-(756,560) 0013:trace:x11drv:X11DRV_ExtEscape SET_DRAWABLE hdc 0x51d8 drawable 3600045 gl_drawable 0 pf 0 dc_rect (72,95)-(423,134) drawable_rect (268,209 )-(756,560)
Further I never see a call to X11DRV_set_win_format which is the call which must create the drawable. I don't see the glGetString call here but I see softwareRenderer getting called and it is 0xc0000005 error. I explicitly set WINEDEBUG as an environment variable but still it doesn't help. I get the impression that somehow not all debug info appears and that some strange stuff is happening. I doubt it is just GL.
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #55 from max@veneto.com 2008-06-29 05:18:05 --- (In reply to comment #54)
Created an attachment (id=14441)
--> (http://bugs.winehq.org/attachment.cgi?id=14441) [details]
+wgl,+opengl,+seh,+x11drv
Very strange stuff is happening. For instance the app is receiving a gl_drawable but then a next call it can be gone again: 0013:trace:x11drv:X11DRV_ExtEscape SET_DRAWABLE hdc 0x51d8 drawable 3600045 gl_drawable 0 pf 0 dc_rect (72,95)-(423,134) drawable_rect (268,209 )-(756,560) 0013:trace:x11drv:X11DRV_ExtEscape SET_DRAWABLE hdc 0x51d8 drawable 3600045 gl_drawable 0 pf 0 dc_rect (72,95)-(423,134) drawable_rect (268,209 )-(756,560) 0013:trace:x11drv:X11DRV_ExtEscape SET_DRAWABLE hdc 0x51d8 drawable 3600045 gl_drawable 3600045 pf 0 dc_rect (0,0)-(488,351) drawable_rect (268 ,209)-(756,560) 0013:trace:x11drv:X11DRV_ExtEscape SET_DRAWABLE hdc 0x51d8 drawable 3600045 gl_drawable 0 pf 0 dc_rect (72,95)-(423,134) drawable_rect (268,209 )-(756,560) 0013:trace:x11drv:X11DRV_ExtEscape SET_DRAWABLE hdc 0x51d8 drawable 3600045 gl_drawable 0 pf 0 dc_rect (72,95)-(423,134) drawable_rect (268,209 )-(756,560)
Further I never see a call to X11DRV_set_win_format which is the call which must create the drawable. I don't see the glGetString call here but I see softwareRenderer getting called and it is 0xc0000005 error. I explicitly set WINEDEBUG as an environment variable but still it doesn't help. I get the impression that somehow not all debug info appears and that some strange stuff is happening. I doubt it is just GL.
The glGetString is indeed called by the custom action (softwareemulation in log). I put some traces inside wine_glGetString() and it's called, crashing on glGetString call inside it. I've noticed too that it doesn't appear on logs, but I guess it's because of missing traces in wine_glGetString(). You can see it with a +relay, or adding traces. The crash IS in glGetString(), but I think too that something is happening before it. Context IS null on glGetString call.
Max
http://bugs.winehq.org/show_bug.cgi?id=13599
Anastasius Focht focht@gmx.net changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |focht@gmx.net
--- Comment #56 from Anastasius Focht focht@gmx.net 2008-06-29 09:04:30 --- Hello,
ideally it should work without gl context (returning NULL) - though the call is questionable because it makes no sense at all. That custom action is either a leftover or the developer didn't understand at all what he wrote.
On the wine/linux side it's most likely a misconception in nvidia's opengl driver which leads to the crash.
The custom action runs in a secondary msi thread (by design) and that's why the bug is uncovered that late.
The following sample code illustrates the problem, calling gl without context from primary and secondary thread:
--- snip sample --- /* winegcc -o gltest gltest.c -lopengl32 */ #include <windows.h> #include <stdio.h> #include <GL/gl.h>
void print_info() { printf("(%x, %p)\n", GetCurrentThreadId(), glGetString(GL_VERSION)); }
static DWORD WINAPI thread_proc( PVOID arg) { print_info(); return 0; }
int main() { print_info(); WaitForSingleObject( CreateThread( NULL,0, thread_proc, NULL, 0, NULL), INFINITE); return 0; } --- snip sample ---
Ideally it shouldn't crash, printing out two times "(<tid>, (nil))".
--- snip winedbg session --- winedbg gltest.exe.so WineDbg starting on pid 001e 0x7b87de58 DbgBreakPoint+0x4 in kernel32: popl %ebp Wine-dbg>b wine_glGetString Breakpoint 1 at 0x7d1eeaad wine_glGetString in opengl32 Wine-dbg>c Stopped on breakpoint 1 at 0x7d1eeaad wine_glGetString in opengl32 .. 0x7d1eead3 wine_glGetString+0x26 in opengl32: call 0x7d1ecd74 enter_gl in opengl32 Wine-dbg> 0x7d1eead8 wine_glGetString+0x2b in opengl32: movl 0x8(%ebp),%eax Wine-dbg> 0x7d1eeadb wine_glGetString+0x2e in opengl32: movl %eax,0x0(%esp) Wine-dbg> 0x7d1eeade wine_glGetString+0x31 in opengl32: call 0x7d1a7d1c Wine-dbg>si 0x7d1a7d1c: jmp *0x4ec(%ebx) .. Wine-dbg>x/10x $ebx 0x7d20701c _GLOBAL_OFFSET_TABLE_: 0007fee8 00000000 00000000 60505ed0 0x7d20702c _GLOBAL_OFFSET_TABLE_+0x10: 60505510 60506010 605042d0 60504e10 0x7d20703c _GLOBAL_OFFSET_TABLE_+0x20: 604ffb10 60505670 .. Wine-dbg>x/x _GLOBAL_OFFSET_TABLE_ Many symbols with name '_GLOBAL_OFFSET_TABLE_', choose the one you want (<cr> to abort): [1]: 0x616cedd0 _GLOBAL_OFFSET_TABLE_ in winex11 [2]: 0x616f6130 _GLOBAL_OFFSET_TABLE_ in imm32 [3]: 0x7d20701c _GLOBAL_OFFSET_TABLE_ in opengl32 .. 0x7d2074fc _GLOBAL_OFFSET_TABLE_+0x4e0: 60505210 605027f0 605053f0 605052f0 .. Wine-dbg>disas 0x605052f0 0x605052f0: movl %gs:0xffffffb0,%eax 0x605052f6: jmp *0x390(%eax) 0x605052fc: int $3 .. Wine-dbg>si 0x605052f0: movl %gs:0xffffffb0,%eax Wine-dbg> 0x605052f6: jmp *0x390(%eax) Wine-dbg> 0x60a1c9c0: xorl %eax,%eax Wine-dbg> 0x60a1c9c2: ret .. --- snip winedbg session ---
The GOT entry for glGetString is at _GLOBAL_OFFSET_TABLE_+0x4ec (opengl32.dll.so, section .GOT)
The target address contains a R_386_TLS_LE (reloc) type instruction which resolves jump destination for the thunk. Let's see where this target address belongs to ...
--- snip --- cat /proc/17753/maps
.. 6045e000-604f2000 r-xp 00000000 fd:00 19081266 /usr/lib/nvidia/libGL.so.173.14.09 604f2000-6050d000 rwxp 00094000 fd:00 19081266 /usr/lib/nvidia/libGL.so.173.14.09 .. 6052f000-611d4000 r-xp 00000000 fd:00 19081262 /usr/lib/nvidia/libGLcore.so.173.14.09 611d4000-61361000 rwxp 00ca4000 fd:00 19081262 /usr/lib/nvidia/libGLcore.so.173.14.09 .. --- snip ---
Now lets verify it's really the correct destination:
--- snip --- objdump -T /usr/lib/nvidia/libGL.so.173.14.09 | grep glGetString 002ad2f0 g DF .writetext 00000000 Base glGetString --- snip ---
Seems so. The thunks will be written at runtime (hence the ".writetext" section name).
--- snip --- readelf -x .writetext /usr/lib/nvidia/libGL.so.173.14.09 | grep 002ad2f0 0x002ad2f0 00000000 00000000 00000000 00000000 ................ --- snip ---
As already indicated by instruction type, all API entries are *TLS* aware.
No gl context:
If called from main thread, the resulting TLS value points to a small stub in libGLcore.so. If called from a different thread context, the resulting TLS value is zero, pointing to nirvana, resulting in crash. This seems to be a misconception in nvidia's opengl driver, not properly handling context-less calls from different threads.
With valid gl context:
The resulting thunk address will point to real implementation.
It's probably a very rare case that an application calls gl API without context so I suggest to fix only wine_glGetString() for now, returning NULL for empty context and not calling drivers glGetString (making a big NOTE in source code, referring to nvidia brain damage). The fix with SEH guard would only hide other bugs.
Regards
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #57 from max@veneto.com 2008-06-29 09:38:24 --- Thanx for the detailed analysis :-)
I'll drop a new patch for it.
Ciao
Max
http://bugs.winehq.org/show_bug.cgi?id=13599
max@veneto.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Attachment #14419|0 |1 is obsolete| |
--- Comment #58 from max@veneto.com 2008-06-29 09:56:00 --- Created an attachment (id=14448) --> (http://bugs.winehq.org/attachment.cgi?id=14448) ade wine_glGetString() return NULL on null gl context
Here the patch that solves this bug. We're at -1 from install on an unpatched wine :-)
Max
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #59 from Roderick Colenbrander thunderbird2k@gmx.net 2008-06-29 13:29:53 --- Once again thanks Anastasius Focht :)
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #60 from Louis Lenders xerox_xerox2000@yahoo.co.uk 2008-06-29 17:01:24 --- I repeat one more time: Focht for president! Thanks indeed for the analysis
@Max, the patch you sent does not apply clean here:
patch -Np1 -i uy patching file dlls/opengl32/wgl.c Hunk #1 FAILED at 632. 1 out of 1 hunk FAILED -- saving rejects to file dlls/opengl32/wgl.c.rej
Maybe you could rediff it and send again (or am i doing something wrong?)
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #61 from max@veneto.com 2008-06-29 17:04:53 --- (In reply to comment #60)
I repeat one more time: Focht for president! Thanks indeed for the analysis
@Max, the patch you sent does not apply clean here:
patch -Np1 -i uy patching file dlls/opengl32/wgl.c Hunk #1 FAILED at 632. 1 out of 1 hunk FAILED -- saving rejects to file dlls/opengl32/wgl.c.rej
Maybe you could rediff it and send again (or am i doing something wrong?)
Did you reset your git before applying ? On mine work. BTW, I'll send you my ICQ number by mail, if you want to contact me directly.
Max
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #62 from Louis Lenders xerox_xerox2000@yahoo.co.uk 2008-06-29 17:15:57 ---
Did you reset your git before applying ?
yes, i copied and pasted it from my browser, maybe things went wrong there, i'll have a look later
BTW, I'll send you my ICQ number by mail, if you want to contact me directly.
That's ok, i'll search for it when i need it, but don't try contact me by sending mail to that address, i only use it for sending mails, and never read the mails i receive there, i think i have about 65000 unread messages in that mailbox now ;)
http://bugs.winehq.org/show_bug.cgi?id=13599
--- Comment #63 from max@veneto.com 2008-06-29 17:33:11 --- Well, if Anastasius Focht could give a try to that one, that'd be perfect :-) http://bugs.winehq.org/show_bug.cgi?id=13801
The other remaining bug is quite simple and I guess will be solved in short time.
Max
http://bugs.winehq.org/show_bug.cgi?id=13599
max@veneto.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED
--- Comment #64 from max@veneto.com 2008-07-04 15:01:00 --- Well, finally fixed in today wine git !
Ciao
Max
http://bugs.winehq.org/show_bug.cgi?id=13599
Alexandre Julliard julliard@winehq.org changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED
--- Comment #65 from Alexandre Julliard julliard@winehq.org 2008-07-11 11:10:34 --- Closing bugs fixed in 1.1.1.
http://bugs.winehq.org/show_bug.cgi?id=13599
Alexandre Julliard julliard@winehq.org changed:
What |Removed |Added ---------------------------------------------------------------------------- Target Milestone|--- |1.0.1
--- Comment #66 from Alexandre Julliard julliard@winehq.org 2008-10-16 07:44:56 --- Fixed by d6a6b88f1904783157245cae0b3f5a8d3de1ec20, nominating for 1.0.1.
http://bugs.winehq.org/show_bug.cgi?id=13599
Austin English austinenglish@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Version|CVS/GIT |unspecified
http://bugs.winehq.org/show_bug.cgi?id=13599
Anastasius Focht focht@gmx.net changed:
What |Removed |Added ---------------------------------------------------------------------------- Fixed by SHA1| |d6a6b88f1904783157245cae0b3 | |f5a8d3de1ec20 Version|unspecified |1.0-rc3