Closed adolfintel closed 9 months ago
I have experienced the same issue.. When i try compiling against the valve experimental and experimental bleeding edge. The issue it seems if it am not mistaken is that tue patch is aimed at plain wine. From what i read on github from valve the branch valve is based on plain wine, where valve-experimental and valve-experimental-bleeding-edge are based off wine staging. So two patches are needed one as it currently works for wine-proton-valve and then a second one that works on the staging branch for wine-proton-valve-exp-bleeding-edge and wine-proton-valve-exp. It used to all compile before the patch was included and forced.
Hi @Tk-Glitch I am not sure this should be closed as the fix seems to cause another exception:
../ValveSoftware-winegit/dlls/xinput1_3/main.c:541:17: error: ‘steam_overlay_event’ undeclared (first use in this function)
541 | CloseHandle(steam_overlay_event);
| ^~~~~~~~~~~~~~~~~~~
../ValveSoftware-winegit/dlls/xinput1_3/main.c:541:17: note: each undeclared identifier is reported only once for each function it appears in
../ValveSoftware-winegit/dlls/xinput1_3/main.c:542:17: error: ‘steam_keyboard_event’ undeclared (first use in this function)
542 | CloseHandle(steam_keyboard_event);
| ^~~~~~~~~~~~~~~~~~~~
make: *** [Makefile:269267: dlls/xinput1_2/x86_64-windows/main.o] Error 1
This looks minor. I'll fix that asap.
A little context: Because of the original bug I reverted to using proton_8.0
instead of experimental_8.0
because it would still build. After your fix I got the above exception building proton_8.0
but I just retried building against experimental_8.0
and while it's not done yet, it hasn't crashed with that exception and i'm almost 15 minutes further into the build. So hopefully that's useful information.
edit: confirmed it now builds on experimental_8.0
but not proton_8.0
.
The fix didn't work.
Hi @Tk-Glitch I am not sure this should be closed as the fix seems to cause another exception:
../ValveSoftware-winegit/dlls/xinput1_3/main.c:541:17: error: ‘steam_overlay_event’ undeclared (first use in this function) 541 | CloseHandle(steam_overlay_event); | ^~~~~~~~~~~~~~~~~~~ ../ValveSoftware-winegit/dlls/xinput1_3/main.c:541:17: note: each undeclared identifier is reported only once for each function it appears in ../ValveSoftware-winegit/dlls/xinput1_3/main.c:542:17: error: ‘steam_keyboard_event’ undeclared (first use in this function) 542 | CloseHandle(steam_keyboard_event); | ^~~~~~~~~~~~~~~~~~~~ make: *** [Makefile:269267: dlls/xinput1_2/x86_64-windows/main.o] Error 1
The fix didn't work as mentioned above.
Getting an error when applying the same patch while building valve-exp-bleeding prepare.log
Looks like it's broken again
Applying your own patch de-steamify-80-be.mypatch
patching file dlls/advapi32/advapi.c
patching file dlls/dbghelp/dwarf.c
patching file dlls/hidclass.sys/device.c
patching file dlls/hidclass.sys/hid.h
patching file dlls/hidclass.sys/pnp.c
patching file dlls/kernelbase/process.c
Hunk #1 succeeded at 1226 (offset 3 lines).
patching file dlls/ntdll/loader.c
Hunk #1 succeeded at 87 with fuzz 2 (offset 1 line).
patching file dlls/ntdll/unix/loader.c
Hunk #1 succeeded at 774 (offset 1 line).
Hunk #2 succeeded at 2113 (offset 6 lines).
Hunk #3 succeeded at 2125 (offset 6 lines).
Hunk #4 succeeded at 2137 (offset 6 lines).
Hunk #5 succeeded at 2151 (offset 6 lines).
Hunk #6 succeeded at 2159 (offset 6 lines).
patching file dlls/secur32/secur32.c
patching file dlls/winebus.sys/bus_sdl.c
Hunk #1 FAILED at 965.
Hunk #2 FAILED at 1021.
2 out of 2 hunks FAILED -- saving rejects to file dlls/winebus.sys/bus_sdl.c.rej
patching file dlls/winex11.drv/window.c
Hunk #1 succeeded at 1052 (offset -4 lines).
patching file dlls/xinput1_3/main.c
Hunk #1 succeeded at 122 with fuzz 2 (offset -2 lines).
Hunk #2 succeeded at 337 (offset -2 lines).
Hunk #3 succeeded at 357 (offset -2 lines).
Hunk #4 succeeded at 453 (offset -2 lines).
Hunk #5 FAILED at 483.
Hunk #6 succeeded at 562 (offset -2 lines).
Hunk #7 succeeded at 778 with fuzz 2 (offset -16 lines).
Hunk #8 succeeded at 864 (offset -26 lines).
Hunk #9 succeeded at 882 (offset -26 lines).
1 out of 9 hunks FAILED -- saving rejects to file dlls/xinput1_3/main.c.rej
patching file loader/wine.inf.in
Hunk #1 succeeded at 340 with fuzz 2 (offset 5 lines).
patching file programs/winedbg/debugger.h
patching file programs/winedbg/tgt_active.c
patching file programs/winedbg/winedbg.c
Was fixed with https://github.com/Frogging-Family/wine-tkg-git/commit/fd23f0bced821b784a55523342f6129636ef47e3
Please git pull
for latest changes.
This is a clean build, I cloned this repo to an empty folder. Does tkg cache something elsewhere?
Did you clone Frogging-Family/wine-tkg-git
or Tk-Glitch/PKGBUILDS
? The latter will need a pull in respective projects dirs as it's not getting sync'ed on every change (more like every couple weeks/months).
If it's indeed a clone of Frogging-Family/wine-tkg-git
, the patch applies fine on my end and your failing hunks definitely are what I fixed on that mentioned commit. Unless you're targeting an older proton-exp-be commit?
I cloned the repo manually to an empty folder and used the valve-exp-bleeding preset.
From what I can see it pulls valve's version of wine at commit bb66944d9d7a0fc7692f05d4a2db418b468d3021 (HEAD -> proton_8.0, tag: proton-wine-8.0-5c)
That does not seem to be the right commit, it didn't switch to experimental_8.0 nor bleeding-edge.
There's no tag set in the preset: https://github.com/Frogging-Family/wine-tkg-git/blob/master/wine-tkg-git/wine-tkg-profiles/wine-tkg-valve-exp-bleeding.cfg
:thinking:
Spun a build to check:
==> Starting prepare()...
HEAD is now at 9397ebe000d fixup! ntdll: Use kernel soft dirty flags for write watches support.
Looks in line with current bleeding edge: https://github.com/ValveSoftware/wine/tree/9397ebe000d4b7ee49f93a54f506262de63d6cde
Also tag is fine: Wine (plain) version: experimental.bleeding.edge.8.0.76225.20240127
Also tested on the non-makepkg path:
branch 'experimental_8.0' set up to track 'origin/experimental_8.0'.
HEAD is now at 9397ebe000d fixup! ntdll: Use kernel soft dirty flags for write watches support.
Is there some special behavior in your code for wine-tkg-valve-exp-bleeding.cfg? If I just make a copy of that file with a different name it fails to build
That's expected. The filename is the profile "id" and indeed there are specific behavior changes for Valve trees (and even more special handling for bleeding edge).
Ah, that explains it, I was making a copy and adding stuff to it.
So what would be the proper way to make a couple of changes to this profile? Do I just modify it?
That's fine indeed. That being said, having the profile id behind a variable set in the file might be worth investigating.
Maybe you could add a variable like TKG_QUIRKS="valve,experimental"
. That would make it more obvious that there's special behavior for some presets.
Started happening a couple days ago.
Here's the log: