https://bugs.winehq.org/show_bug.cgi?id=50813
Bug ID: 50813 Summary: X-Beyond The Frontier / X-Tension: crash on launch since 6.3 Product: Wine Version: 6.3 Hardware: x86-64 OS: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: -unknown Assignee: wine-bugs@winehq.org Reporter: 19non91@gmail.com Distribution: ---
Created attachment 69624 --> https://bugs.winehq.org/attachment.cgi?id=69624 wine log
Crash is present on 6.4.1 staging, 6.4.1, 6.3.1. Was downgrading and stopped at 6.2.1 - issue is not present there. I'm using up-to-date Arch Linux.
What other info I could provide to be more helpful ?
Thank you.
https://bugs.winehq.org/show_bug.cgi?id=50813
Non 19non91@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Distribution|--- |ArchLinux
https://bugs.winehq.org/show_bug.cgi?id=50813
Louis Lenders xerox.xerox2000x@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Keywords| |regression CC| |xerox.xerox2000x@gmail.com
--- Comment #1 from Louis Lenders xerox.xerox2000x@gmail.com --- (In reply to Non from comment #0)
Created attachment 69624 [details] wine log
What other info I could provide to be more helpful ?
So if I get it right it`s a regression in Staging? Or does it crash in vanlla wine as well?
I found a demo that seems to behave more or less the same: http://dl3.egosoft.com/download/x/files/xdemo03.zip
albeit it exits with an errorwindow (instead of crash) for me in Staging, but works fine in vanilla wine.
If the regression is only present in Staging, try see what patches are added since wineStaging-6.2 (it`s not that much I think) and try disable them one by one.
https://bugs.winehq.org/show_bug.cgi?id=50813
--- Comment #2 from Non 19non91@gmail.com --- (In reply to Louis Lenders from comment #1)
So if I get it right it`s a regression in Staging? Or does it crash in vanlla wine as well?
If the regression is only present in Staging, try see what patches are added since wineStaging-6.2 (it`s not that much I think) and try disable them one by one.
It is present in staging 6.4.1, and vanilla 6.4.1 and 6.3.1.
Yeah, it "exits" with an error window. I thought that was definition of a crash. :D
I guess the only option is to bisect 6.2.1 <> 6.3.1 then.
Is it meaningful to test staging 6.3.1 and 6.2.1 ?
https://bugs.winehq.org/show_bug.cgi?id=50813
--- Comment #3 from Louis Lenders xerox.xerox2000x@gmail.com --- (In reply to Non from comment #2)
(In reply to Louis Lenders from comment #1)
So if I get it right it`s a regression in Staging? Or does it crash in vanlla wine as well?
I guess the only option is to bisect 6.2.1 <> 6.3.1 then.
Is it meaningful to test staging 6.3.1 and 6.2.1 ?
If the bug is also present in regular wine there`s no need to test Staging.
Before starting bisecting could you check if it runs in current git (so after yesterdays commits) because that`s the wine version i run, and as I said the demo starts fine here then.
So the bug might already be fixed in current wine (git)
And also could you report what happens for you with the demo from the link when you run it in wine-6.3.1? Thanks
https://bugs.winehq.org/show_bug.cgi?id=50813
--- Comment #4 from Non 19non91@gmail.com --- Created attachment 69635 --> https://bugs.winehq.org/attachment.cgi?id=69635 xbtf errors git
https://bugs.winehq.org/show_bug.cgi?id=50813
--- Comment #5 from Non 19non91@gmail.com --- Tested Demo, X-Beyond The Frontier, X-Tension on today's git with fresh win32 wineprefix.
They all "exit" with an error message.
Demo works on 6.3.1-vanilla. X-Tension and X-Beyond The Frontier don't.
Compile time was around 2-3 hrs, should've run a timer on that one. :) Feels long.
Bisecting will be a long ride...
https://bugs.winehq.org/show_bug.cgi?id=50813
--- Comment #6 from Non 19non91@gmail.com --- Created attachment 69636 --> https://bugs.winehq.org/attachment.cgi?id=69636 xbtf errors 6.3.1 vanilla
https://bugs.winehq.org/show_bug.cgi?id=50813
--- Comment #7 from Non 19non91@gmail.com --- Created attachment 69637 --> https://bugs.winehq.org/attachment.cgi?id=69637 xt errors
https://bugs.winehq.org/show_bug.cgi?id=50813
--- Comment #8 from Non 19non91@gmail.com --- Created attachment 69638 --> https://bugs.winehq.org/attachment.cgi?id=69638 xt errors 6.3.1 vanilla
Should've cat'ed them. I'm sorry.
https://bugs.winehq.org/show_bug.cgi?id=50813
--- Comment #9 from Non 19non91@gmail.com --- Bisected from 6.2 to 6.3 and then from 6.3 to 6.4. 18 iterations. Everything was working. Strange...
I initially tested the latest git using https://aur.archlinux.org/packages/wine-git
Bisecting was done using the following script:
#!/bin/sh export WINE=$HOME/downloads/wine-git/wine export WINEARCH=win32 export WINEPREFIX=/tmp/wine export CC="ccache gcc -m32" export CFLAGS="-g -O0" export CROSSCFLAGS="-g -O0" ./configure --verbose --disable-tests || exit 125 make || exit 125 rm -rf $WINEPREFIX mkdir -p $WINEPREFIX $WINE regedit $HOME/games/xfiles/wine-xbtf-xt-config.reg
Every iteration I did this (compile-wine is a script above):
git bisect good git clean -f && git reset --hard && time compile-wine
and the game was launched like this:
WINEARCH=win32 WINEPREFIX=/tmp/wine WINE=$HOME/downloads/wine-git/wine $HOME/downloads/wine-git/wine x-tension.exe
-O0 reduced build times significantly, ranging from 20 to 40 mins, and when cache kicked in to something funny like 5-10 mins.
So based on what I've seen I could blame Arch packaging, -O2 or ccache, which hid the bugs somehow.
I think I should try latest commit with -O0, then Arch PKGBUILD with -O0 and then git with -O2.
Thoughts?
https://bugs.winehq.org/show_bug.cgi?id=50813
--- Comment #10 from Non 19non91@gmail.com --- Latest commit with -O0: no errors. Arch PKGBUILD with -O0: no errors. Latest commit without specifying -O0 or -g: no errors. Latest commit without specifying anything through Arch's PKGBUILD: no errors.
Looks like I can't reproduce.
Also it is funny that I completely forgot about make clean and used git clean -f + git reset --hard instead, which had no effect due to autogenerated .gitignore. It seems that bisection was a time waste due to that.
https://bugs.winehq.org/show_bug.cgi?id=50813
--- Comment #11 from Louis Lenders xerox.xerox2000x@gmail.com --- (In reply to Non from comment #10)
Latest commit with -O0: no errors. Arch PKGBUILD with -O0: no errors. Latest commit without specifying -O0 or -g: no errors. Latest commit without specifying anything through Arch's PKGBUILD: no errors.
Looks like I can't reproduce.
Also it is funny that I completely forgot about make clean and used git clean -f + git reset --hard instead, which had no effect due to autogenerated .gitignore. It seems that bisection was a time waste due to that.
Huh, that`s indeed wast of time and frustrating i guess. So is the bug than only present if you _don`'t_ compile wine yourself? (so with your distro`s wine package?)
For the demo i found that is is present in self compiled _and_ my distro`s wine-staging, but not in my self compiled wine vanilla. So gets more and more confusing.... Maybe someone else could test/try
https://bugs.winehq.org/show_bug.cgi?id=50813
--- Comment #12 from Non 19non91@gmail.com --- The only time self-compiled wine didn't work for me is when I tested latest commit through PKGBUILD using it's regular bare repo fetch.
Then I fetched the repo myself and adjusted PKGBUILD so it could use same repo too (1).
Also I skipped prepare step to prevent repo redownloading. Prepare step contains this:
sed 's|OpenCL/opencl.h|CL/opencl.h|g' -i wine/configure*
Maybe "exits" with errors have something to do with changed header paths.
1. Didn't know about bare repos and the possibility to enable a worktree for them. Thought PKGBUILD just downloads stripped repo with one latest commit to compile, like --depth 1 or something. So I just downloaded it again manually to be able to use bisect.
https://bugs.winehq.org/show_bug.cgi?id=50813
--- Comment #13 from Non 19non91@gmail.com --- Out of curiosity just pulled and compiled today's commit with -O0 and that OpenCL -> CL header path change from PKGBUILD - still works.
https://github.com/archlinux/svntogit-community/blob/59757e5a881745acf5e036b...
There's nothing esotheric in the regular wine PKGBUILD compared to wine-git PKGBUILD from AUR...
I don't know what else can be done.
https://bugs.winehq.org/show_bug.cgi?id=50813
Zebediah Figura z.figura12@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |z.figura12@gmail.com
--- Comment #14 from Zebediah Figura z.figura12@gmail.com --- There's a good chance that this was just a race condition to begin with. Does it work consistently with current Wine?
https://bugs.winehq.org/show_bug.cgi?id=50813
Gijs Vermeulen gijsvrm@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Summary|X-Beyond The Frontier / |X-Beyond The Frontier & |X-Tension: crash on launch |X-Tension crash on start |since 6.3 |
https://bugs.winehq.org/show_bug.cgi?id=50813
--- Comment #15 from Non 19non91@gmail.com --- (In reply to Zebediah Figura from comment #14)
There's a good chance that this was just a race condition to begin with. Does it work consistently with current Wine?
Could be, because disabling compiler opts fixed it. Tested on Wine 7.0 rc2, works just fine.
https://bugs.winehq.org/show_bug.cgi?id=50813
Zebediah Figura z.figura12@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Resolution|--- |FIXED Status|UNCONFIRMED |RESOLVED
--- Comment #16 from Zebediah Figura z.figura12@gmail.com --- (In reply to Non from comment #15)
(In reply to Zebediah Figura from comment #14)
There's a good chance that this was just a race condition to begin with. Does it work consistently with current Wine?
Could be, because disabling compiler opts fixed it. Tested on Wine 7.0 rc2, works just fine.
Okay, I'm going to go ahead and mark as fixed then. Please reopen if you encounter the bug again.
https://bugs.winehq.org/show_bug.cgi?id=50813
Alexandre Julliard julliard@winehq.org changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED
--- Comment #17 from Alexandre Julliard julliard@winehq.org --- Closing bugs fixed in 7.0-rc6.