http://bugs.winehq.org/show_bug.cgi?id=58827
Bug ID: 58827 Summary: Running Mahjong with Wine ends without displaying the game screen Product: Wine Version: 10.17 Hardware: x86-64 OS: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: win32u Assignee: wine-bugs@winehq.org Reporter: craigaschulstad@gmail.com Distribution: ---
Created attachment 79498 --> http://bugs.winehq.org/attachment.cgi?id=79498 Wine debug of no Mahjong game window display
After updating to Wine 10.17, my Windows game programs (such as Mahjong, Purble Place, and so forth) end without displaying the game in a window. This behavior seems very similar to the issue I submitted with bug 58321 a few months ago which had an issue identified and resolved in win32u. As such, I am attaching a debug log using the WINEDEBUG options suggested at that time. Please let me know if other information is needed. For now, I will reset my Linux virtual machine back to its state prior to installing Wine 10.17.
http://bugs.winehq.org/show_bug.cgi?id=58827
Austin English austinenglish@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Keywords| |regression
--- Comment #1 from Austin English austinenglish@gmail.com --- Please run a regression test: https://gitlab.winehq.org/wine/wine/-/wikis/Regression-Testing
http://bugs.winehq.org/show_bug.cgi?id=58827
--- Comment #2 from Craig Schulstad craigaschulstad@gmail.com --- I reviewed the request for regression testing and felt that I do not have the expertise or resources on my system to do that. Therefore, I am going to await the next release of Wine to see if other testing and feedback results in a resolution of this problem. As a further comment, the information in this bug report might be forwarded to Remi Bernon, as he corrected an issue I had previously submitted which was a part of revising EGL functionality.
--- Comment #3 from Craig Schulstad craigaschulstad@gmail.com --- I reviewed the request for regression testing and felt that I do not have the expertise or resources on my system to do that. Therefore, I am going to await the next release of Wine to see if other testing and feedback results in a resolution of this problem. As a further comment, the information in this bug report might be forwarded to Remi Bernon, as he corrected an issue I had previously submitted which was a part of revising EGL functionality.
http://bugs.winehq.org/show_bug.cgi?id=58827
--- Comment #2 from Craig Schulstad craigaschulstad@gmail.com --- I reviewed the request for regression testing and felt that I do not have the expertise or resources on my system to do that. Therefore, I am going to await the next release of Wine to see if other testing and feedback results in a resolution of this problem. As a further comment, the information in this bug report might be forwarded to Remi Bernon, as he corrected an issue I had previously submitted which was a part of revising EGL functionality.
--- Comment #3 from Craig Schulstad craigaschulstad@gmail.com --- I reviewed the request for regression testing and felt that I do not have the expertise or resources on my system to do that. Therefore, I am going to await the next release of Wine to see if other testing and feedback results in a resolution of this problem. As a further comment, the information in this bug report might be forwarded to Remi Bernon, as he corrected an issue I had previously submitted which was a part of revising EGL functionality.
http://bugs.winehq.org/show_bug.cgi?id=58827
--- Comment #4 from Craig Schulstad craigaschulstad@gmail.com --- After applying version 10.18 of Wine, the issue with no window display still exists. As I noted, this behavior mimics an issue I had encountered with a previous Wine release and was identified and resolved by Remi Bernon in program opengl.c in the win32u "DLL" folder. You might want to refer this issue to Remi. Thank you.
http://bugs.winehq.org/show_bug.cgi?id=58827
random-nick@mail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |random-nick@mail.com
--- Comment #5 from random-nick@mail.com --- Seems to be some kind of issue with using the swrast driver? I can't reproduce it normally but can reproduce the bug by using LIBGL_ALWAYS_SOFTWARE=1 or MESA_LOADER_DRIVER_OVERRIDE=swrast but it might just be that these options are somehow bugged in mesa because they give me errors (the first one gives me "libEGL warning: Not allowed to force software rendering when API explicitly selects a hardware device." with wine while the second one makes eglgears crash and gives me "glx: failed to create dri3 screen" "failed to load driver: swrast" errors with glxgears) so the behaviour might differ on setups which really don't have graphics cards. Also for me the EGL is loaded successfully (the attached log shows EGL failing to load, maybe an issue with the mesa installation?) but the game still doesn't work. I haven't tested if other programs are affected.
http://bugs.winehq.org/show_bug.cgi?id=58827
--- Comment #6 from Craig Schulstad craigaschulstad@gmail.com --- Sounds like you might have found a clue. Again, I will await the next release of Wine to see if possibly this gets fixed.
http://bugs.winehq.org/show_bug.cgi?id=58827
--- Comment #7 from Craig Schulstad craigaschulstad@gmail.com --- After applying version 10.19 of Wine, the issue still persists. With that, I decided to do a few more experiments using my previous Linux/Wine version that works with my Mahjong game.
First off, I manually launched the Mahjong game, launched the "winecfg" command, and launched another Windows 95 pinball game in the previous Linux/Wine version, redirecting the error output to "before" log files. All of the programs worked and displayed their respective window.
I then, ran the Linux update process to install Wine 10.19. After that, I again manually ran the Wine command for the Mahjong game, the Pinball game, and "winecfg", redirecting the error output to "after" log files. The Pinball game and the "winecfg" command displayed their respective windows; whereas, the Mahjong game just fell over.
For reference, I am attaching the three, "before" logs and the three "after" logs for your reference. Let me know if I need to include any different parameters in the debugging portion of the testing. One other thing to note. As I indicated, the Pinball game is a Windows 95 game, but the Mahjong game is a Windows XP game.
http://bugs.winehq.org/show_bug.cgi?id=58827
--- Comment #8 from Craig Schulstad craigaschulstad@gmail.com --- Created attachment 79723 --> http://bugs.winehq.org/attachment.cgi?id=79723 Before log of Pinball game
This is the log of the Windows 95 Pinball game using the previous version of Wine.
http://bugs.winehq.org/show_bug.cgi?id=58827
--- Comment #9 from Craig Schulstad craigaschulstad@gmail.com --- Created attachment 79724 --> http://bugs.winehq.org/attachment.cgi?id=79724 Pinball game log after applying Wine 10.19
Log of Windows 95 Pinball game after applying Wine 10.19 update
http://bugs.winehq.org/show_bug.cgi?id=58827
--- Comment #10 from Craig Schulstad craigaschulstad@gmail.com --- Created attachment 79725 --> http://bugs.winehq.org/attachment.cgi?id=79725 winecfg log run from previous Wine version
winecfg log output indicating successful display of window
http://bugs.winehq.org/show_bug.cgi?id=58827
--- Comment #11 from Craig Schulstad craigaschulstad@gmail.com --- Created attachment 79726 --> http://bugs.winehq.org/attachment.cgi?id=79726 winecfg log after applying Wine 10.19
winecfg log after applying Wine 10.19 indicating that the window still successfully displays
http://bugs.winehq.org/show_bug.cgi?id=58827
--- Comment #12 from Craig Schulstad craigaschulstad@gmail.com --- Created attachment 79727 --> http://bugs.winehq.org/attachment.cgi?id=79727 Mahjong game log when run from previous version of Wine
This log indicates the traces when the Mahjong game successfully displays the game window. You might note the amount of trace output.
http://bugs.winehq.org/show_bug.cgi?id=58827
Craig Schulstad craigaschulstad@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Attachment #79498|0 |1 is obsolete| |
--- Comment #13 from Craig Schulstad craigaschulstad@gmail.com --- Created attachment 79729 --> http://bugs.winehq.org/attachment.cgi?id=79729 Mahjong game log after applying latest version of Wine (10.19)
Updated log of the Mahjong game after applying the latest version of Wine. Note that the size of the log is substantially smaller as the failure seems to occur right away.
http://bugs.winehq.org/show_bug.cgi?id=58827
--- Comment #14 from Craig Schulstad craigaschulstad@gmail.com --- Having some time to do a bit of research in the Mahjong logs, the few information lines that stood out to me focused on an incapability of creating a display context noted by the following log lines almost at the end of the "after" log and right before the window gets destroyed.
0024:trace:opengl:wglCreateContext hDc 40010062 0024:trace:wgl:win32u_wgl_context_reset context 0x7c4cd4e0, hdc 0x40010062, share (nil), attribs (nil) 0024:trace:wgl:get_dc_pixel_format 0x40010062/0x40068 -> 0 0024:trace:x11drv:destroy_whole_window win 0x40068 xwin 5600003/0 0024:err:d3d:wined3d_adapter_gl_init Failed to get a GL context for adapter 0089F788.
Working my way back through the various program files in the "win32u" dll folder, the program of interest seems to be "opengl.c". In comparing the previous version of Wine that works for me (10.15) to subsequent versions, there is a new function named "init_device_info" that updates a function pointer "compat_context = funcs->p_eglCreateContext( egl->display, config, EGL_NO_CONTEXT, context_attribs );". That might be a spot to investigate.