http://bugs.winehq.org/show_bug.cgi?id=58513
Bug ID: 58513 Summary: Wine 10.9 completely broke (LGA775 Core2Quad) Product: Wine Version: 10.9 Hardware: x86-64 OS: Linux Status: UNCONFIRMED Severity: critical Priority: P2 Component: -unknown Assignee: wine-bugs@winehq.org Reporter: xeroxslayer@proton.me Distribution: ---
I have no other way to describe it except this. Everything was working fine up until Wine 10.8, then I upgrade to Wine 10.9 on Void Linux, and it doesn't work at all any more.
There are more details in this issue on GitHub. There is just way too much info to paste here.
https://github.com/void-linux/void-packages/issues/55697
Basically, it just doesn't work at all. You run `winecfg` or any other software and this is what you get.
``` 002c:err:seh:NtRaiseException Unhandled exception code c0000005 flags 0 addr 0x6fffffc4c133 0034:err:seh:NtRaiseException Unhandled exception code c0000005 flags 0 addr 0x6fffffc4c133 003c:err:seh:NtRaiseException Unhandled exception code c0000005 flags 0 addr 0x6fffffc4c133
```
Nothing works. This happens, and then Wine quietly exits.
From what little I understoo, basically, it can't locate binaries. The binaries are there, at that location, it just doesn't see them... or can't load them, IDK. You can find more relevant info in this comment, with some debug flags enabled.
https://github.com/void-linux/void-packages/issues/55697#issuecomment-299020...
I do have to note that Wine past 10.8 (>=10.9) does work on other rigs I have, newer ones, but this particular rig is an LGA775 Core2Quad. Now, I know that Wine is mostly focused towards gaming and gamers, but there are people out there that use it for simple things as well, like Winamp and similar things, and we really don't need brand new hardware to do that. In fact, this is my workshop rig and I basically just need it to browse some things on the web (still performs OKish regarding that) and maybe open a few documents, an IC programmer from time to time, but that's about it. So, this rig still suits my needs just fine.
I tried scrubbing the volume, check for errors, nothing, everything is peachy, no corruption problems whatsoever.
On Void's side, there are literally no changes in the template, except bumping up versions, hashes and removing `libOSMesa` as a dependency (which is depreciated, found that out later on), but I even tried adding `libOSMesa` as a dependency and building it like that, no changes whatsoever, it's still completely broken.
There are way too many commits between 10.8 and 10.9, and in all honesty, I wouldn't even know where to start looking (I presume something to do with CPU, but, I'm just guessing here), so if someone could shed some light, I would gladly test.
http://bugs.winehq.org/show_bug.cgi?id=58513
--- Comment #1 from xeroxslayer@proton.me --- Created attachment 78980 --> http://bugs.winehq.org/attachment.cgi?id=78980 Wine logs with some debug flags enabled, trying to run Winamp
http://bugs.winehq.org/show_bug.cgi?id=58513
--- Comment #2 from xeroxslayer@proton.me --- Created attachment 78981 --> http://bugs.winehq.org/attachment.cgi?id=78981 Wine logs with some debug flags enabled, trying to run Winamp
http://bugs.winehq.org/show_bug.cgi?id=58513
xeroxslayer@proton.me changed:
What |Removed |Added ---------------------------------------------------------------------------- Distribution|--- |Other Keywords| |hardware URL| |https://github.com/void-lin | |ux/void-packages/issues/556 | |97
http://bugs.winehq.org/show_bug.cgi?id=58513
--- Comment #3 from Dmitry Timoshkov dmitry@baikal.ru --- Probably duplicate of the bug 58335. Is it possible to perform a regression test?
http://bugs.winehq.org/show_bug.cgi?id=58513
--- Comment #4 from Dmitry Timoshkov dmitry@baikal.ru --- Also it's worth testing with wine-10.12 first.
http://bugs.winehq.org/show_bug.cgi?id=58513
--- Comment #5 from xeroxslayer@proton.me --- (In reply to Dmitry Timoshkov from comment #3)
Probably duplicate of the bug 58335. Is it possible to perform a regression test?
No problem, just give me link to a tree or a hash, I can clone and build from that.
http://bugs.winehq.org/show_bug.cgi?id=58513
--- Comment #6 from xeroxslayer@proton.me --- (In reply to Dmitry Timoshkov from comment #4)
Also it's worth testing with wine-10.12 first.
Tried latest verison, no dice, doesn't work. Haven't tried 10.10 and 10.11 though.
http://bugs.winehq.org/show_bug.cgi?id=58513
--- Comment #7 from xeroxslayer@proton.me --- (In reply to Dmitry Timoshkov from comment #3)
Probably duplicate of the bug 58335. Is it possible to perform a regression test?
Preferably from a tree within 10.9... I think it makes most sense, since that is the version in which everything broke.
http://bugs.winehq.org/show_bug.cgi?id=58513
--- Comment #8 from xeroxslayer@proton.me --- (In reply to Dmitry Timoshkov from comment #3)
Probably duplicate of the bug 58335. Is it possible to perform a regression test?
OK, I tried running winecfg with `WINEDEBUG=+heap`, started updating the index, that's good.
http://bugs.winehq.org/show_bug.cgi?id=58513
--- Comment #9 from xeroxslayer@proton.me --- OK, running with `WINEDEBUG=+heap` kinda worked for winecfg (not showing windows properly, no tabs in the window, selection menus are empty), but at least it loaded something. `WINEDEBUG=+heap` running Winamp with classic skin works just fine... but I presume this is because Winamp does it's own window theming.
http://bugs.winehq.org/show_bug.cgi?id=58513
--- Comment #10 from Dmitry Timoshkov dmitry@baikal.ru --- (In reply to xeroxslayer from comment #5)
(In reply to Dmitry Timoshkov from comment #3)
Probably duplicate of the bug 58335. Is it possible to perform a regression test?
No problem, just give me link to a tree or a hash, I can clone and build from that.
https://bugs.winehq.org/show_bug.cgi?id=58335#c97 mentions 745a11a231816a7d82490156c72a18c39216a14c Could you please verify whether that's the problematic commit? If not, then do https://gitlab.winehq.org/wine/wine/-/wikis/Regression-Testing
http://bugs.winehq.org/show_bug.cgi?id=58513
--- Comment #11 from Dmitry Timoshkov dmitry@baikal.ru --- (In reply to xeroxslayer from comment #5)
(In reply to Dmitry Timoshkov from comment #3)
Probably duplicate of the bug 58335. Is it possible to perform a regression test?
No problem, just give me link to a tree or a hash, I can clone and build from that.
https://bugs.winehq.org/show_bug.cgi?id=58335#c97 mentions 745a11a231816a7d82490156c72a18c39216a14c Could you please verify whether that's the problematic commit? If not, then do https://gitlab.winehq.org/wine/wine/-/wikis/Regression-Testing
http://bugs.winehq.org/show_bug.cgi?id=58513
--- Comment #12 from Dmitry Timoshkov dmitry@baikal.ru --- (In reply to xeroxslayer from comment #5)
(In reply to Dmitry Timoshkov from comment #3)
Probably duplicate of the bug 58335. Is it possible to perform a regression test?
No problem, just give me link to a tree or a hash, I can clone and build from that.
https://bugs.winehq.org/show_bug.cgi?id=58335#c97 mentions 745a11a231816a7d82490156c72a18c39216a14c Could you please verify whether that's the problematic commit? If not, then do https://gitlab.winehq.org/wine/wine/-/wikis/Regression-Testing
http://bugs.winehq.org/show_bug.cgi?id=58513
--- Comment #13 from Dmitry Timoshkov dmitry@baikal.ru --- Sorry for the duplicate comments, bugzilla seems to behave a little bit flaky.
http://bugs.winehq.org/show_bug.cgi?id=58513
--- Comment #14 from xeroxslayer@proton.me --- (In reply to Dmitry Timoshkov from comment #13)
Sorry for the duplicate comments, bugzilla seems to behave a little bit flaky.
No problem, I accidentally attached the logs twice, lol. You (or the admin) can safely remove one, they're the same file.
Just to confirm once more, 10.12 doesn't work. On the other hand, running winecfg with `WINEDEBUG=+heap` at least showed a window in 10.9, now it doesn't even do that. What I can tell is that it does a lot of work in the background, the CPU is doing something, but eventually, it also exits. On the other hand, running Winamp with `WINEDEBUG=+heap` works just fine, loads and plays music (haven't tried other functionality). I also tried this old app written in Delphi, DoubleKiller (dated 2005 or 2006 I think), that also works fine with `WINEDEBUG=+heap`, the window shows up, buttons work, that seems to work fine as well. I tried running an older version of PassMark's KeyboardTest (dated 2003), written in MSVC++, classic Windows theme shows up, buttons work, doesn't seem to be any problems.
So, basically, the bug is still there, but you just bypass it with `WINEDEBUG=+heap`... except for winecfg, that at least showed some window in 10.9, in 10.12 it just quietly exits after a while. Whatever got merged in 10.12, didn't fix the problem.
Regarding commit 745a11a231816a7d82490156c72a18c39216a14c, I basically build with whatever is committed up until commit 745a11a231816a7d82490156c72a18c39216a14c, without that commit, correct? I presume GitLab makes a zip archive of the source after every commit? If so, where do I find this archive?
Sorry for the (probably) stupid questions, I'm more used to working with GitHub, lol.
http://bugs.winehq.org/show_bug.cgi?id=58513
--- Comment #15 from Dmitry Timoshkov dmitry@baikal.ru --- (In reply to xeroxslayer from comment #14)
Regarding commit 745a11a231816a7d82490156c72a18c39216a14c, I basically build with whatever is committed up until commit 745a11a231816a7d82490156c72a18c39216a14c, without that commit, correct? I presume GitLab makes a zip archive of the source after every commit? If so, where do I find this archive?
There's no archive, you would need to clone the git, and perform usual things like $ git reset --hard 745a11a231816a7d82490156c72a18c39216a14c build, make sure that the bug persists. $ git show 745a11a231816a7d82490156c72a18c39216a14c | patch -p1 -R build, see if the bug is fixed.
http://bugs.winehq.org/show_bug.cgi?id=58513
--- Comment #16 from xeroxslayer@proton.me --- (In reply to Dmitry Timoshkov from comment #15)
There's no archive, you would need to clone the git, and perform usual things like $ git reset --hard 745a11a231816a7d82490156c72a18c39216a14c build, make sure that the bug persists. $ git show 745a11a231816a7d82490156c72a18c39216a14c | patch -p1 -R build, see if the bug is fixed.
Mhm, got it. Was hoping to do this in xbps's chroot, but we'll see... I'll try and post back results.
http://bugs.winehq.org/show_bug.cgi?id=58513
--- Comment #17 from Dmitry Timoshkov dmitry@baikal.ru --- Created attachment 78984 --> http://bugs.winehq.org/attachment.cgi?id=78984 patch
Does this patch help?