http://bugs.winehq.org/show_bug.cgi?id=19315
Summary: Call Of Duty 4: regression, graphical glitches in 1.1.25 (post-processing works intermittently) Product: Wine Version: 1.1.25 Platform: PC OS/Version: Linux Status: UNCONFIRMED Severity: major Priority: P2 Component: directx-d3d AssignedTo: wine-bugs@winehq.org ReportedBy: amoswenger@gmail.com
In Wine 1.1.24, Call Of Duty 4 has absolutely no graphical glitches, once Depth of Field and Glow are disabled.
But since Wine 1.1.25, often there are glitches where post-processing seems to be temporarily disabled for a few milliseconds, regularly. This results in quick flashes and it makes it nearly impossible to play certain missions (especially President Fulani-something cinematic, near the beginning). Sometimes, the graphic post-processing is blur and again it flickers between a blurred and a non-blurred image. Sometimes also the post-processing disabling stick, and it only disappears when getting back to the menu (Esc) and back again to the game (Esc again).
http://bugs.winehq.org/show_bug.cgi?id=19315
Amos Wenger amoswenger@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- URL| |http://www.gamershell.com/d | |ownload_21400.shtml
http://bugs.winehq.org/show_bug.cgi?id=19315
--- Comment #1 from Austin English austinenglish@gmail.com 2009-07-13 15:45:43 --- Regression test? http://wiki.winehq.org/RegressionTesting
http://bugs.winehq.org/show_bug.cgi?id=19315
--- Comment #2 from Amos Wenger amoswenger@gmail.com 2009-07-13 15:48:48 --- (In reply to comment #1)
Regression test? http://wiki.winehq.org/RegressionTesting
Seems tedious to me, but yeah, way to go.
I'm on it.
http://bugs.winehq.org/show_bug.cgi?id=19315
--- Comment #3 from Amos Wenger amoswenger@gmail.com 2009-07-13 15:54:14 --- By the way, I've reported it as being in wined3d, but is there any chance it could be an opengl bug? (since wine implements direct3d on top of opengl)
I'm asking cause git bisect accepts a path as an argument so I don't want to be looking in the wrong place.
http://bugs.winehq.org/show_bug.cgi?id=19315
Vitaliy Margolen vitaliy@kievinfo.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Keywords| |regression Summary|Call Of Duty 4: regression, |Call Of Duty 4 - graphical |graphical glitches in |glitches (post-processing |1.1.25 (post-processing |works intermittently) |works intermittently) | Severity|major |normal
--- Comment #4 from Vitaliy Margolen vitaliy@kievinfo.com 2009-07-13 19:41:01 --- http://bugs.winehq.org/page.cgi?id=fields.html#bug_severity
Just test the entire tree if you not sure of the component.
http://bugs.winehq.org/show_bug.cgi?id=19315
--- Comment #5 from Amos Wenger amoswenger@gmail.com 2009-07-13 19:43:22 --- Found the offending commit:
===================================================================== ca69ec0e852a6f2e877f1c93d7a37d6f857bf9c7 is first bad commit commit ca69ec0e852a6f2e877f1c93d7a37d6f857bf9c7 Author: Henri Verbeet hverbeet@codeweavers.com Date: Wed Jul 1 09:46:16 2009 +0200
wined3d: Recognize geometry shaders in shader_trace_init().
:040000 040000 ac174b4b20d52bce9c38dcb13f88081f29704667 86e8a0f50e0989a916d89c6b606702bd3f145c48 M dlls =====================================================================
(Easy) steps to reproduce: - Install the full version - Apply a No-CD crack - Launch the game, choose 'Select mission', choose 'Prologue -> The Coup' - There is flickering because the post-processing isn't applied on every frame
'The Coup' is the mission where it's the most visible. You can probably see the bug in a mission from the demo (see the URL)
Is there anything else I can do?
http://bugs.winehq.org/show_bug.cgi?id=19315
--- Comment #6 from Henri Verbeet hverbeet@gmail.com 2009-07-14 04:28:20 --- (In reply to comment #5)
Found the offending commit:
===================================================================== ca69ec0e852a6f2e877f1c93d7a37d6f857bf9c7 is first bad commit commit ca69ec0e852a6f2e877f1c93d7a37d6f857bf9c7 Author: Henri Verbeet hverbeet@codeweavers.com Date: Wed Jul 1 09:46:16 2009 +0200
wined3d: Recognize geometry shaders in shader_trace_init().
:040000 040000 ac174b4b20d52bce9c38dcb13f88081f29704667 86e8a0f50e0989a916d89c6b606702bd3f145c48 M dlls =====================================================================
It's extremely unlikely that that's the commit that caused this. The code in question isn't executed unless you have the d3d_shader debug channel enabled.
http://bugs.winehq.org/show_bug.cgi?id=19315
--- Comment #7 from Amos Wenger amoswenger@gmail.com 2009-07-14 05:32:48 --- Hm then it's the commit just before or just after. I don't know, I just ran git bisect as told on the Regression wiki page, and this is what came out.
If this commit's the not the right, how do I find the right one then ?
http://bugs.winehq.org/show_bug.cgi?id=19315
--- Comment #8 from Amos Wenger amoswenger@gmail.com 2009-07-14 06:25:00 --- I'll try to test regressions on dlls/opengl, then.
http://bugs.winehq.org/show_bug.cgi?id=19315
--- Comment #9 from Amos Wenger amoswenger@gmail.com 2009-07-14 07:19:02 --- Tested regression on the whole tree once again, and this time this commit comes out:
==================================================================== 50419daa11b921237a0203e713450b7776ea3463 is first bad commit commit 50419daa11b921237a0203e713450b7776ea3463 Author: Henri Verbeet hverbeet@codeweavers.com Date: Wed Jul 1 09:46:17 2009 +0200
wined3d: Try to activate the original context for queries.
:040000 040000 86e8a0f50e0989a916d89c6b606702bd3f145c48 b1bc85b9206e2c305a4c7eb39e58300c150c1f8b M dlls ====================================================================
(makes more sense to me, actually)
http://bugs.winehq.org/show_bug.cgi?id=19315
Amos Wenger amoswenger@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |hverbeet@gmail.com
http://bugs.winehq.org/show_bug.cgi?id=19315
--- Comment #10 from Henri Verbeet hverbeet@gmail.com 2009-07-14 07:38:02 --- That one's more plausible, although it doesn't sound directly related. If possible, revert the commit on top of current to make sure. Also, if possible, could you verify it happens in the demo as well. (I could do that as well of course, but you know what to look for, and it would save me some time.)
http://bugs.winehq.org/show_bug.cgi?id=19315
--- Comment #11 from Amos Wenger amoswenger@gmail.com 2009-07-14 07:51:05 --- Created an attachment (id=22374) --> (http://bugs.winehq.org/attachment.cgi?id=22374) My version of query.c, manually merged after I reverted the suspected commit.
http://bugs.winehq.org/show_bug.cgi?id=19315
--- Comment #12 from Amos Wenger amoswenger@gmail.com 2009-07-14 07:52:48 --- I did: $ git bisect reset $ git revert 50419daa11b921237a0203e713450b7776ea3463
Then it said it couldn't revert automatically, I did $ git status
The conflict was in query.c lines 333 to 354 approx, so I commented one of the two versions (see the attachment), compiled everything again, and the bug is gone =) So it's seems it's in this commit after all (except if I did anything wrong)
I'm gonna test the demo right away.
http://bugs.winehq.org/show_bug.cgi?id=19315
--- Comment #13 from Amos Wenger amoswenger@gmail.com 2009-07-14 15:09:09 --- Ok, bad news, the bug isn't visible in the Demo (the only mission it contains seems to be unaffected from the bug), but the full version is.
I've been trying to do screenshots of the bug, and videos, but I can't capture the exact frames where it's wrong, and when capturing video using xvidcap with CoD4 in "windowed mode" (from wine configuration) the bug isn't visible either !
http://bugs.winehq.org/show_bug.cgi?id=19315
rocko rockorequin@hotmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |rockorequin@hotmail.com
--- Comment #14 from rocko rockorequin@hotmail.com 2009-07-18 01:10:29 --- I confirm this in both 1.1.25 and 1.1.26.
I can reproduce it in normal play and also by pausing the game and then moving the mouse around. When paused, it looks like the 'flashes' might be caused by a single frame from earlier being inserted. When I stop moving the mouse, the flashes stop.
http://bugs.winehq.org/show_bug.cgi?id=19315
rocko rockorequin@hotmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|UNCONFIRMED |NEW Ever Confirmed|0 |1
--- Comment #15 from rocko rockorequin@hotmail.com 2009-07-18 01:11:07 --- *** This bug has been confirmed by popular vote. ***
http://bugs.winehq.org/show_bug.cgi?id=19315
--- Comment #16 from rocko rockorequin@hotmail.com 2009-07-18 01:46:41 --- I can confirm that the query.c file Amos posted fixes the glitches for me. I built it in the 1.1.26 git tree.
http://bugs.winehq.org/show_bug.cgi?id=19315
--- Comment #17 from Amos Wenger amoswenger@gmail.com 2009-07-18 05:16:26 --- I'm pleased to see the bug confirmed.
I'd like to emphasize that the query.c I attached is just the 1.1.24 version adapted to the current codebase, it's not a really a "fix", just a revert. If there's another, proper way to fix it, it'd be a lot better of course.
http://bugs.winehq.org/show_bug.cgi?id=19315
--- Comment #18 from Henri Verbeet hverbeet@gmail.com 2009-07-24 11:16:58 --- Could you retry with current git?
http://bugs.winehq.org/show_bug.cgi?id=19315
--- Comment #19 from Amos Wenger amoswenger@gmail.com 2009-07-24 11:35:06 --- Seems to works for me, I'll try again with a clean checkout of the HEAD.
http://bugs.winehq.org/show_bug.cgi?id=19315
Amos Wenger amoswenger@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED
--- Comment #20 from Amos Wenger amoswenger@gmail.com 2009-07-24 13:20:38 --- I confirm that the bug is fixed in the latest git =)
Thanks a million, Henri!
http://bugs.winehq.org/show_bug.cgi?id=19315
--- Comment #21 from rocko rockorequin@hotmail.com 2009-07-24 18:04:54 --- It's fixed for me too, thanks.
http://bugs.winehq.org/show_bug.cgi?id=19315
Alexandre Julliard julliard@winehq.org changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED
--- Comment #22 from Alexandre Julliard julliard@winehq.org 2009-08-07 12:56:48 --- Closing bugs fixed in 1.1.27.
http://bugs.winehq.org/show_bug.cgi?id=19315
Oli Warner oli@thepcspy.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |oli@thepcspy.com
--- Comment #23 from Oli Warner oli@thepcspy.com 2010-03-02 08:14:34 --- This bug is back with an ugly vengeance.
Both git and recent dev releases (tested 1.1.29 through 1.1.39) are exhibiting serious flickering with any graphical settings. I tried almost all the possible permutations. This is on the latest stable Nvidia drivers on Ubuntu.
Rolling back to a 1.1.24 release fixes it.
Any ideas?
http://bugs.winehq.org/show_bug.cgi?id=19315
--- Comment #24 from rocko rockorequin@hotmail.com 2010-03-27 19:54:10 --- I'm noticing this bug again in wine 1.1.41 with nvidia 195.36.15.
http://bugs.winehq.org/show_bug.cgi?id=19315
--- Comment #25 from Jeff Zaroyko jeffz@jeffz.name 2010-03-27 20:05:42 --- Please file a new bug, this one was reported to be fixed. If a new bug with the same symptoms reappears we do not go back and open old bugs.