http://bugs.winehq.org/show_bug.cgi?id=13125
Summary: Fahrenheit (Indigo Prophecy) broken in newer versions of Wine. Product: Wine Version: 1.0-rc1 Platform: PC-x86-64 OS/Version: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: -unknown AssignedTo: wine-bugs@winehq.org ReportedBy: m_105@gmx.de
Created an attachment (id=12919) --> (http://bugs.winehq.org/attachment.cgi?id=12919) console output
Fahrenheit is broken in newer versions of Wine. I think it started somewhere about .55 (.59 is broken, .44 works great). It crashes reproducibly after the "Loading..." screen, whenever you start the game, a chapter or the tutorial.
http://bugs.winehq.org/show_bug.cgi?id=13125
--- Comment #1 from Ian Schwarz m_105@gmx.de 2008-05-11 05:29:34 --- .56 works. I can't test .57 and .58 because there are no SuSE RPMs left. .59 is broken.
http://bugs.winehq.org/show_bug.cgi?id=13125
--- Comment #2 from Austin English austinenglish@gmail.com 2008-05-11 23:59:47 --- Please run a regression test: http://wiki.winehq.org/RegressionTesting
is there a demo/download available?
http://bugs.winehq.org/show_bug.cgi?id=13125
--- Comment #3 from Ian Schwarz m_105@gmx.de 2008-05-12 04:29:43 --- Demo is here: http://www.gamershell.com/download_10034.shtml
http://bugs.winehq.org/show_bug.cgi?id=13125
--- Comment #4 from Ian Schwarz m_105@gmx.de 2008-05-12 05:47:31 --- Sorry, I can't do a bisect. Fahrenheit refuses to run with ANY of my home-compiled wine versions, good or bad.
http://bugs.winehq.org/show_bug.cgi?id=13125
--- Comment #5 from Ian Schwarz m_105@gmx.de 2008-05-12 10:02:46 --- (In reply to comment #4)
Sorry, I can't do a bisect. Fahrenheit refuses to run with ANY of my home-compiled wine versions, good or bad.
Disregard that, I su- uhh, I mean, I think I can do it.
http://bugs.winehq.org/show_bug.cgi?id=13125
--- Comment #6 from Ian Schwarz m_105@gmx.de 2008-05-12 11:53:46 --- Okay, here it goes: 02f6c9ece8b24d0161c5382e2f4dc6a74876afd5 is first bad commit commit 02f6c9ece8b24d0161c5382e2f4dc6a74876afd5 Author: Roderick Colenbrander thunderbird2k@gmx.net Date: Sun Mar 16 16:02:02 2008 +0000
wined3d: Move depth/stencil verification to CheckDepthStencilCapability.
:040000 040000 263d87834b5b7e74cfb7e03d05f5e7930606670f b0b14d052af3ac0ed2ae3a62efa95c2cfcb5fbd0 M dlls
But really, I am NOT sure if this is right. Why? Simple, the same damn bug did not show up ONCE during all that tests! There were many others that fu**ed it up so it became unplayable (crash during start happened most often, also no graphics visible, only sounds while still registering input) but it didn't crash ONCE in the desired opportunity. I'll probably have to retest it. *sigh*
http://bugs.winehq.org/show_bug.cgi?id=13125
Alexander Dorofeyev alexd4@inbox.lv changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |alexd4@inbox.lv
--- Comment #7 from Alexander Dorofeyev alexd4@inbox.lv 2008-05-12 14:14:18 --- (In reply to comment #3)
Demo is here: http://www.gamershell.com/download_10034.shtml
Does this demo install successfully for you? For me it fails early with
err:msi:ITERATE_Actions Execution halted, action L"DX9_Launcher.A545835B_7768_4D96_A5BC_930949616C2D" returned 1603
http://bugs.winehq.org/show_bug.cgi?id=13125
--- Comment #8 from Ian Schwarz m_105@gmx.de 2008-05-12 17:41:31 --- (In reply to comment #7) Okay, I think it boils down to this one: 7fd87f10cdf0d9345d909a4b5cc6130855e7ee0a is first bad commit commit 7fd87f10cdf0d9345d909a4b5cc6130855e7ee0a Author: Roderick Colenbrander thunderbird2k@gmx.net Date: Tue Mar 18 09:38:55 2008 +0100
wined3d: There is no need for the !Usage checks at the start of the resource type checks.
The return-path of the code will handle this properly and this also ensures that the R8G8B8 dxVersion test is reached.
:040000 040000 99a85b058cea69205001a75ab22fa5dd072c02c6 d5461793813bf5a85e26c8bcc324d9490aa1da03 M dlls This time I always marked it a "good" one if Fahrenheit was either playable OR crashed NOT when starting a new game, but right after launching it for example. I will test the demo tomorrow, it still hasn't not loaded fully.
http://bugs.winehq.org/show_bug.cgi?id=13125
--- Comment #9 from Alexander Dorofeyev alexd4@inbox.lv 2008-05-12 18:04:56 --- (In reply to comment #8)
(In reply to comment #7) This time I always marked it a "good" one if Fahrenheit was either playable OR crashed NOT when starting a new game, but right after launching it for example. I will test the demo tomorrow, it still hasn't not loaded fully.
I'm afraid this is not how this is done (if I'm reading you right). You mark the revision good or bad during bisect if and only if you can successfully get to the point in program where it usually shows up and then get consistent results there. Otherwise you end up with unreliable, probably wrong results.
Unfortunately, sometimes when bisecting you can't get to the required spot in the program, because, for example, the game crashes in that revision earlier than you can test (because of other bugs). That's where it gets much more difficult than usual routine. If it happens, the only (right) way is to find fixes for other bugs (which are not relevant to your bisect and are an interference) and then port those fixes to revision you test for bisect. I realize this is probably more than can be expected from users who are not developers themselves.
http://bugs.winehq.org/show_bug.cgi?id=13125
--- Comment #10 from Ian Schwarz m_105@gmx.de 2008-05-13 03:55:20 --- (In reply to comment #9) I followed the instructions as good as I could. The bisect was started with git bisect good wine-0.9.56 git bisect bad wine-0.9.59 If it crashed exactly where it does in the latest releases (1.0-rc1), it was marked "bad". If it did not and it ran either flawlessly or crashed differently, it was marked "good". I couldn't test it when I didn't even get to that point, sorry. Can you also get a version from git after or before a certain commit?
http://bugs.winehq.org/show_bug.cgi?id=13125
--- Comment #11 from Austin English austinenglish@gmail.com 2008-05-13 12:30:34 --- (In reply to comment #10)
(In reply to comment #9) I followed the instructions as good as I could. The bisect was started with git bisect good wine-0.9.56 git bisect bad wine-0.9.59 If it crashed exactly where it does in the latest releases (1.0-rc1), it was marked "bad". If it did not and it ran either flawlessly or crashed differently, it was marked "good". I couldn't test it when I didn't even get to that point, sorry. Can you also get a version from git after or before a certain commit?
git reset COMMIT-ID
To build before, use its parent.
http://bugs.winehq.org/show_bug.cgi?id=13125
--- Comment #12 from Ian Schwarz m_105@gmx.de 2008-05-13 13:56:22 --- Okay, thanks. Anyways, I think I'm not helpful here anymore. So, anything else that I could do about this bug?
http://bugs.winehq.org/show_bug.cgi?id=13125
--- Comment #13 from Alexander Dorofeyev alexd4@inbox.lv 2008-05-13 14:01:49 --- (In reply to comment #12)
Okay, thanks. Anyways, I think I'm not helpful here anymore. So, anything else that I could do about this bug?
Yes, you can try to find out if demo runs and if this is reproducible in demo. Unfortunately I can't install the demo. Maybe I should just file another report, but appdb tests indicate it's supposed to run, I wonder if I'm doing something wrong or what.
http://bugs.winehq.org/show_bug.cgi?id=13125
--- Comment #14 from Ian Schwarz m_105@gmx.de 2008-05-13 14:59:45 --- My demo (http://www.demonews.de/demos.php?demo=857, gamershell.com didn't work out) installed and ran flawlessly in my home-compiled v. 0.9.56 (vanilla) and ran in 1.0-rc1, but crashed in the same place as the full version.
http://bugs.winehq.org/show_bug.cgi?id=13125
Ian Schwarz m_105@gmx.de changed:
What |Removed |Added ---------------------------------------------------------------------------- Keywords| |regression
http://bugs.winehq.org/show_bug.cgi?id=13125
Alexander Dorofeyev alexd4@inbox.lv changed:
What |Removed |Added ---------------------------------------------------------------------------- URL| |http://www.gamershell.com/do | |wnload_10034.shtml Status|UNCONFIRMED |NEW Component|-unknown |directx-d3d Ever Confirmed|0 |1 Keywords| |download
--- Comment #15 from Alexander Dorofeyev alexd4@inbox.lv 2008-05-13 15:48:45 --- You are right. The first file didn't work in windows either. Re-downloading from same location (gamershell) helped. Yes, I confirm, crashes exactly like you describe.
http://bugs.winehq.org/show_bug.cgi?id=13125
--- Comment #16 from Ian Schwarz m_105@gmx.de 2008-05-13 15:57:38 --- Who set the component to "directx-d3d", did I accidentally? Or do you already know what the culprit is?
http://bugs.winehq.org/show_bug.cgi?id=13125
--- Comment #17 from Alexander Dorofeyev alexd4@inbox.lv 2008-05-13 16:52:00 --- (In reply to comment #16)
Who set the component to "directx-d3d", did I accidentally? Or do you already know what the culprit is?
I did. Based on presense of this - IDirect3DDevice9Impl_AddRef() - in the backtrace, bug in d3d is the most likely assumption at this moment. I will attempt a regression test myself too.
http://bugs.winehq.org/show_bug.cgi?id=13125
--- Comment #18 from Alexander Dorofeyev alexd4@inbox.lv 2008-05-13 17:49:09 --- By trying a few versions I've stashed, it looks like
wine-0.9.57: game starts, i can 'start demo' without crash wine-0.9.58: game starts, 'start demo' gives a crash
85b024d5f71195b815d51ab990fdde02f6b66bed (somewhere in between those two): game doesn't even start.
So i would say the problem is somewhere between .57 and .58, now if i can find a fix for immediate crash at startup, maybe i'll be able to finish bisect.
http://bugs.winehq.org/show_bug.cgi?id=13125
Alexander Dorofeyev alexd4@inbox.lv changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |thunderbird2k@gmx.net
--- Comment #19 from Alexander Dorofeyev alexd4@inbox.lv 2008-05-14 11:22:43 --- After some voodoo rites, (I think) I found out more or less reliably the commit that introduces the "crash when start demo in main menu" bug:
commit d44c2956512117f8f03f77b38d35fb0503418962 Author: Roderick Colenbrander thunderbird2k@gmx.net Date: Sun Mar 16 16:37:38 2008 +0000
wined3d: Check the render target capabilities of a resource type.
--
near that time, two other problems were introduced that are a huge interference when regression testing, because they don't allow to get into main menu. I was able to get to the main menu in the above commit and the one prior to it, by backporting fixes from commits 7fd87f10cdf0d9345d909a4b5cc6130855e7ee0a and 0e9d02ded40810272f8cf8dfc429f8d07e9e6fb9. After doing so, I reproduced the "start demo" crash in the commit, but not in the one prior to it. That's why I think this particular problem was introduced by "check render target capabilities" commit.
Will try to check out, but it looks very complicated so I don't really expect to succeed :). CCing author.
http://bugs.winehq.org/show_bug.cgi?id=13125
--- Comment #20 from Alexander Dorofeyev alexd4@inbox.lv 2008-05-14 12:25:30 --- Some further findings: the critical part seems to be whether IWineD3DImpl_CheckDeviceFormat succeeds for 16 bit format like WINED3DFMT_A1R5G5B5 with WINED3DRTYPE_TEXTURE rtype and WINED3DUSAGE_RENDERTARGET usage. Before that commit by Roderick it apparently was succeding. After it, this call returns WINED3DERR_NOTAVAILABLE. Which, on first sight looks like it's correct. So maybe the real bug is not introduced by Roderick's comment but exposed by it. Interestingly, the game also checks other texture formats with rendertarget cap, some of which succeed. I would guess that maybe the app prefers 16bit, so if it sees WINED3DFMT_A1R5G5B5 then it chooses it and something different happens from what happens if it chooses some formats for which IWineD3DImpl_CheckDeviceFormat succeeds (like A8R8G8B8).
http://bugs.winehq.org/show_bug.cgi?id=13125
--- Comment #21 from Alexander Dorofeyev alexd4@inbox.lv 2008-05-14 12:49:24 --- Since bug seems to be related to offscreen render targets, I tried playing with OffscreenRenderingMode settings. And it appears setting OffscreenRenderingMode to fbo allows to avoid the crash. That's in versions around d44c2956512117f8f03f77b38d35fb0503418962.
http://bugs.winehq.org/show_bug.cgi?id=13125
Jan Niklas Hasse jhasse@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |jhasse@gmail.com
http://bugs.winehq.org/show_bug.cgi?id=13125
Peter Belm peterbelm@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |peterbelm@gmail.com
--- Comment #22 from Peter Belm peterbelm@gmail.com 2008-09-26 11:01:34 --- I'm having this exact same problem.
Alexander: I've read through your comments, do you think this is a problem with the game not behaving correctly (but managed to get away with it under Windows) or another part of Wine causing problems? (You seemed to rule out IDirect3DDevice9Impl_AddRef being the problem)
http://bugs.winehq.org/show_bug.cgi?id=13125
Dennis Schridde lindevel@gmx.net changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |lindevel@gmx.net
--- Comment #23 from Dennis Schridde lindevel@gmx.net 2009-02-13 16:36:12 --- I am getting the same problem with the installer as Alexander in comment #7: err:msi:ITERATE_Actions Execution halted, action L"DX9_Launcher.A545835B_7768_4D96_A5BC_930949616C2D" returned 1603
Are any fixes/workarounds for that known?
Using wine-1.1.14.
http://bugs.winehq.org/show_bug.cgi?id=13125
Ken Sharp kennybobs@o2.co.uk changed:
What |Removed |Added ---------------------------------------------------------------------------- Depends on| |17536
http://bugs.winehq.org/show_bug.cgi?id=13125
Ken Sharp kennybobs@o2.co.uk changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |kennybobs@o2.co.uk
--- Comment #24 from Ken Sharp kennybobs@o2.co.uk 2009-03-18 22:43:13 --- Just tried the demo (ftp://ftp.download-center.com/international/fahrenheit/fahrenheit_demo_europe.zip) under 1.1.17 and got what appears to be a different bug. The same error exists in all the versions I tested: 1.1.17, 0.9.61, 0.9.57, 0.9.51 and 0.9.33.
Could someone else look at this to confirm one way or another?
(In reply to comment #23)
I am getting the same problem with the installer as Alexander in comment #7: err:msi:ITERATE_Actions Execution halted, action L"DX9_Launcher.A545835B_7768_4D96_A5BC_930949616C2D" returned 1603
Are any fixes/workarounds for that known?
You should be able to install this with Wine 1.0 to work around Bug 17536.
http://bugs.winehq.org/show_bug.cgi?id=13125
--- Comment #25 from Ian Schwarz m_105@gmx.de 2009-03-20 10:27:03 --- (In reply to comment #24)
Just tried the demo (ftp://ftp.download-center.com/international/fahrenheit/fahrenheit_demo_europe.zip) under 1.1.17 and got what appears to be a different bug. The same error exists in all the versions I tested: 1.1.17, 0.9.61, 0.9.57, 0.9.51 and 0.9.33.
Could someone else look at this to confirm one way or another?
As I said, 0.9.44 did and 0.9.56 still does work great for me. In fact, I keep a compiled vanilla 0.9.56 around in a subdirectory of the game just to be able to play it. (Sound does only work with OSS selected in the Audio tab, though.)
http://bugs.winehq.org/show_bug.cgi?id=13125
Bug 13125 depends on bug 17536, which changed state.
Bug 17536 Summary: DirectX 9 launcher causes Atari Scrabble Online installation to fail http://bugs.winehq.org/show_bug.cgi?id=17536
What |Old Value |New Value ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID
http://bugs.winehq.org/show_bug.cgi?id=13125
Bug 13125 depends on bug 17536, which changed state.
Bug 17536 Summary: DirectX 9 launcher causes Atari Scrabble Online installation to fail http://bugs.winehq.org/show_bug.cgi?id=17536
What |Old Value |New Value ---------------------------------------------------------------------------- Status|RESOLVED |REOPENED Resolution|INVALID |
http://bugs.winehq.org/show_bug.cgi?id=13125
Bug 13125 depends on bug 17536, which changed state.
Bug 17536 Summary: DirectX 9 launcher causes Atari Scrabble Online installation to fail http://bugs.winehq.org/show_bug.cgi?id=17536
What |Old Value |New Value ---------------------------------------------------------------------------- Status|REOPENED |RESOLVED Resolution| |FIXED
http://bugs.winehq.org/show_bug.cgi?id=13125
Pavel Ondracka drakkk@centrum.cz changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |drakkk@centrum.cz
--- Comment #26 from Pavel Ondracka drakkk@centrum.cz 2009-08-29 10:12:15 --- Demo works fine with Wine 1.1.28. Can someone test the full version please?
http://bugs.winehq.org/show_bug.cgi?id=13125
Ian Schwarz m_105@gmx.de changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED
--- Comment #27 from Ian Schwarz m_105@gmx.de 2009-08-30 06:44:42 --- Full version works as well.
http://bugs.winehq.org/show_bug.cgi?id=13125
Alexandre Julliard julliard@winehq.org changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED
--- Comment #28 from Alexandre Julliard julliard@winehq.org 2009-09-02 14:13:18 --- Closing bugs fixed in 1.1.29.
-
wine-bugs＠winehq.org