http://bugs.winehq.org/show_bug.cgi?id=25549
Summary: Newer Wine Fails to draw Bitmap Graphics (Regression) Product: Wine Version: 1.3.9 Platform: All OS/Version: Linux Status: UNCONFIRMED Severity: critical Priority: P2 Component: gdi32 AssignedTo: wine-bugs@winehq.org ReportedBy: Libertas.Mania@hotmail.com
I've tried the following in Ubuntu 10.04, 10.10, 11.04 Alpha 1
Wine 1.2 - 1.1.4.2 - ubuntu4 - Works Wine 1.2 - 1.2.0-lucid - ubuntu6 - Failed since above to below have failed Wine 1.3.9-0ubuntu1~lucid - Failed
Working Screenshot: http://img23.imageshack.us/img23/8637/mgpubuntu.png
Failed Screenshot: http://img651.imageshack.us/img651/7564/mgpubuntu2.png
Purpose of the Regression fix: http://www.ubuntugamer.com/2010/09/211-free-wine-compatible-games-in-one-dow...
After spending months working on the 211 Wine compatible Freeware games for the above pack. I really wanted as many people as possible to be able to play them with Wine, With the Games Installer and Launcher broken it makes it unusable.
Download the Problematic EXE package here: http://www.mediafire.com/?uq5i6zrnvded2ic
As the linked item is 9MB in size I thought it best to use a free file host.
Comments: I have a feeling the bitblitz routines in GDI have a protection routine that aborts the memory relocation or perhaps it's in the paint event, I am setting the backdrop of a picturebox for the screenshot and I set the backdrop/wallpaper property of the form to a Bitmap (which also is failing to draw).
Good luck with fixing this regression before the 1.4.0 release, I can only hope someone has the time an skill required to do so.
Thanks again
Glenn Chugg
http://bugs.winehq.org/show_bug.cgi?id=25549
Glenn Chugg Libertas.Mania@hotmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |Libertas.Mania@hotmail.com
http://bugs.winehq.org/show_bug.cgi?id=25549
--- Comment #1 from Glenn Chugg Libertas.Mania@hotmail.com 2010-12-18 06:15:03 CST --- I ran from the terminal to check output, here's the results:
fixme:wincodecs:JpegDecoder_Frame_GetResolution (0x189954, 0x32ee70,0x32ee68): stub
http://bugs.winehq.org/show_bug.cgi?id=25549
--- Comment #2 from Glenn Chugg Libertas.Mania@hotmail.com 2010-12-18 06:25:03 CST --- After googling about for a little while I came up with:
It's most likely the color conversion problem. This will have to be debugged in libjpeg.
I believe libjpeg errors currently result in a call to abort() instead of a failure return code.
So my initial guess was correct in the method of handling of the errors, just it wasn't the GDI, it was the libjpeg.
http://bugs.winehq.org/show_bug.cgi?id=25549
Nikolay Sivov bunglehead@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC|Libertas.Mania@hotmail.com | Component|gdi32 |-unknown Severity|critical |minor
--- Comment #3 from Nikolay Sivov bunglehead@gmail.com 2010-12-18 06:55:27 CST --- Please perform a regression test and specify official download location if it's available. Also don't use external image hosting sites, instead attach to this bug.
http://bugs.winehq.org/show_bug.cgi?id=25549
--- Comment #4 from Glenn Chugg Libertas.Mania@hotmail.com 2010-12-18 08:04:08 CST --- (In reply to comment #3)
Please perform a regression test and specify official download location if it's available. Also don't use external image hosting sites, instead attach to this bug.
I was reporting a bug in my own way, don't like it re post it in your own way.
if people don't have access to imageshack and/or mediafire and request it, I'd be willing to do as you ask, but to do it just for the sake of it is not in my agenda at all. I am just trying to help in the way I can, stop making things complected for new members and you might get a few more devs and testers.
You have to realize how many forums I've been banned from by posting links to other forums pages, life is not as black and white as you expect it to be.
http://www.lastos.org/index.php?forums/sswpi.7/
there's the homepage to the ssWPI installer/launcher, it has it's own screen shots of what it's supposed to look like, making the need to relocate the example images not required.
http://bugs.winehq.org/show_bug.cgi?id=25549
Wylda wylda@volny.cz changed:
What |Removed |Added ---------------------------------------------------------------------------- Keywords| |download, regression Status|UNCONFIRMED |NEW URL| |http://www.mediafire.com/?u | |q5i6zrnvded2ic CC| |julliard@winehq.org, | |wylda@volny.cz Component|-unknown |winex11.drv Version|1.3.9 |1.2-rc1 Ever Confirmed|0 |1
--- Comment #5 from Wylda wylda@volny.cz 2010-12-18 09:47:52 CST ---
I am just trying to help in the way I can
BTW: i believe, that filling bugs following way is what helps most. ...Not meant as a starting shot for flaming ;) Thanks for bug report and splitting the 9MB installer from the the whole 1.82GB pack!
1. I did a regression test between 1.1.44 and 1.2-rc1:
commit ea596f8a5f7d732bb56b5921ceb964137df0b930 Author: Alexandre Julliard julliard@winehq.org Date: Tue May 18 14:11:51 2010 +0200
winex11: Force alpha channel to 0 when blitting to depth 32 from a lower depth.
:040000 040000 369e8db88a7dd0c206f6b37d41bfcac27dfbc708 28700a3330faf630cb189926f9b335936f158ab9 M dlls
2. No other bug report suffers from this commit.
3. Revert of this patch after git checkout makes that problem go away.
4. Adding author of this patch to CC.
--private keyword: bisected
http://bugs.winehq.org/show_bug.cgi?id=25549
Wylda wylda@volny.cz changed:
What |Removed |Added ---------------------------------------------------------------------------- Summary|Newer Wine Fails to draw |Newer Wine Fails to draw |Bitmap Graphics |Bitmap Graphics |(Regression) |
http://bugs.winehq.org/show_bug.cgi?id=25549
--- Comment #6 from Glenn Chugg Libertas.Mania@hotmail.com 2010-12-18 11:38:48 CST --- http://www.winehq.org/pipermail/wine-cvs/2010-May/067073.html
I don't understand how to generate the fix, it would be great if I could see it in action before the next public release that has this fix applied, so that I may continue working on updating my tools to support the future wine releases, or at least offer a file solution for my users to update if they wish to use older versions of Wine as a temp solution until then.
Any help in to this matter would be appreciated, I have a git setup now but I'll wait until tomorrow to attempt to use it, looks like it'll be complicated to figure git out and I want to make sure I remember what I learn.
http://bugs.winehq.org/show_bug.cgi?id=25549
--- Comment #7 from Wylda wylda@volny.cz 2010-12-18 12:32:26 CST ---
I don't understand how to generate the fix
Fix does not exist yet and thus this bug is still in OPEN state. Probably only Alexandre knows how to fix this ;) so you have to wait for a fix or try to fix it yourself.
If you want to see your collection correctly running under latest wine-1.3.9 you can revert this trouble-making-patch by:
git show ea596f8a5f7d732bb56b5921ceb964137df0b930 | patch -p1 -R
but this is only work around, not proper fix. But works, tested now. What i do, i follow this link:
http://wiki.winehq.org/RegressionTesting
which points out what patch broke the application. For further question you should try forum and keep bugzilla only for bug fixing. Thank you.
http://bugs.winehq.org/show_bug.cgi?id=25549
--- Comment #8 from Glenn Chugg Libertas.Mania@hotmail.com 2010-12-18 14:38:04 CST ---
git show ea596f8a5f7d732bb56b5921ceb964137df0b930 | patch -p1 -R
but this is only work around, not proper fix.
Thanks for this, it is great to see that less than 40 lines of code in the xrender.c file is causing the conflict. I managed to use git to grab the latest and compile it.
First try it ran perfect, was cool to finally see Wine 1.3.9 able to use the ssWPI again.
The fixme:wincodecs:JpegDecoder... line is still shown in the terminal, meaning it's totally unrelated and the bug is instead caused by the bit depth up scaling patch given above.
I see why you said I might not have any luck fixing the bug myself, although the WinAPI's called in the patch make sense to me the C language does not, nor does any of the references and pointer uses.
Thanks again, at least I can continue to develop and test my projects using the work around.
http://bugs.winehq.org/show_bug.cgi?id=25549
Dmitry Timoshkov dmitry@codeweavers.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Platform|All |Other Summary|Newer Wine Fails to draw |Mega Games Pack fails to |Bitmap Graphics |draw bitmap graphics
http://bugs.winehq.org/show_bug.cgi?id=25549
--- Comment #9 from Glenn Chugg Libertas.Mania@hotmail.com 2011-05-08 01:40:24 CDT --- Created an attachment (id=34542) --> (http://bugs.winehq.org/attachment.cgi?id=34542) This file contains a manual fix that save you compiling from source code
This file contains a manual fix that save you compiling from source code
I can't believe how long this bug has stayed active, after using the patch I found which file the code compiled in to and made a simple bash script to place that modded file over the top of the default file. This saves approx 40 minutes compile time to get the fix in place on any Linux OS. The codebase was from git 8th may 2011, Wine v1.3.19
I am not sure why this problem has been left, but this fix gets a LOT more games working for me, so I'll use this work around until someone with the power to remove the problematic code from the actual public releases, if it's still in v1.4 then I will cry, cause it's such a simple change with no side effects.
http://bugs.winehq.org/show_bug.cgi?id=25549
Glenn Chugg Libertas.Mania@hotmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Attachment #34542|0 |1 is obsolete| |
--- Comment #10 from Glenn Chugg Libertas.Mania@hotmail.com 2011-05-18 12:57:58 CDT --- Created an attachment (id=34767) --> (http://bugs.winehq.org/attachment.cgi?id=34767) Precompiled Fix Updated to 1.3.20 Wine
Precompiled Fix Updated to 1.3.20 Wine
http://bugs.winehq.org/show_bug.cgi?id=25549
--- Comment #11 from Glenn Chugg Libertas.Mania@hotmail.com 2011-05-18 13:09:13 CDT --- Lastest 1.3.20 has added a large delay to starting ssWPI under WINE, but it does become usable after a short wait.
http://www.lastos.org/index.php?threads/sswpi-v8-11-05-13-rc4.284/#post-4278
LastOS.org is the home of the ssWPI tool I use to make packs of apps/games that work with Windows and WINE, the bug makes it white as shown here:
http://forum.winehq.org/viewtopic.php?p=61237#61237
I've noticed many of the Freeware games I included also suffer from lack of the graphics being drawn, mostly the GUI Game menu systems that are in 2d.
http://bugs.winehq.org/show_bug.cgi?id=25549
Glenn Chugg Libertas.Mania@hotmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- URL|http://www.mediafire.com/?u |http://www.lastos.org/index |q5i6zrnvded2ic |.php?threads/sswpi-v8-11-05 | |-13-rc4.284/#post-4324
http://bugs.winehq.org/show_bug.cgi?id=25549
--- Comment #12 from Glenn Chugg Libertas.Mania@hotmail.com 2011-05-29 06:52:36 CDT --- The bug still exists in 1.3.21, haven't got around to compiling a new patched one and don't think anyone bothered getting it before anyway :D
Ahh well,I tried my best to get this bug fixed, but it must be a cockroach, because it wont die.
http://bugs.winehq.org/show_bug.cgi?id=25549
André H. nerv@dawncrow.de changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |nerv@dawncrow.de Regression SHA1| |ea596f8a5f7d732bb56b5921ceb | |964137df0b930
http://bugs.winehq.org/show_bug.cgi?id=25549
Glenn Chugg glennlchugg@hotmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |glennlchugg@hotmail.com
--- Comment #13 from Glenn Chugg glennlchugg@hotmail.com 2011-09-16 02:01:56 CDT --- This bug is still in 1.3.28, I am still hoping it'll be fixed/reverted before the next stable release.
http://bugs.winehq.org/show_bug.cgi?id=25549
--- Comment #14 from Glenn Chugg glennlchugg@hotmail.com 2011-09-16 02:03:09 CDT --- This bug is still in 1.3.28, I am still hoping it'll be fixed/reverted before the next stable release.
http://bugs.winehq.org/show_bug.cgi?id=25549
Alexandre Julliard julliard@winehq.org changed:
What |Removed |Added ---------------------------------------------------------------------------- Keywords|regression | Component|winex11.drv |gdiplus
--- Comment #15 from Alexandre Julliard julliard@winehq.org 2011-09-20 10:58:49 CDT --- The x11 change is correct, the bug is in gdiplus. It loads a jpg into a 24-bpp DIB, so when blitting it into a 32-bpp DIB the alpha channel is set to 0. Not sure where in gdiplus the alpha should be fixed up.
http://bugs.winehq.org/show_bug.cgi?id=25549
--- Comment #16 from Vincent Povirk madewokherd@gmail.com 2011-09-20 11:42:50 CDT --- Gdiplus only uses DIBs for images without alpha, so the content of the alpha channel shouldn't matter. Also, gdiplus only blits to HDCs or HWNDs it gets from the program, never to DIBs that it creates.
So that suggests to me that in this case gdiplus doesn't know anything more about the alpha channel than gdi32 does.
http://bugs.winehq.org/show_bug.cgi?id=25549
--- Comment #17 from Vincent Povirk madewokherd@gmail.com 2011-09-20 11:59:04 CDT --- If for some reason native gdiplus treats HDC's with 32-bit DIBs selected into them as having alpha data (and we'd need a test to show that), then we need to create a GpBitmap object for the DIB in GdipCreateFromHDC2.
http://bugs.winehq.org/show_bug.cgi?id=25549
Alexandre Julliard julliard@winehq.org changed:
What |Removed |Added ---------------------------------------------------------------------------- URL|http://www.lastos.org/index |http://www.lastos.org/index |.php?threads/sswpi-v8-11-05 |.php?forums/sswpi.7/ |-13-rc4.284/#post-4324 |
--- Comment #18 from Alexandre Julliard julliard@winehq.org 2011-09-20 12:47:11 CDT --- It looks like the app expects an alpha channel in the bitmap returned from GdipCreateHBITMAPFromBitmap (which does blit to a DIB created by gdiplus).
The app displays correctly with native gdiplus.
http://bugs.winehq.org/show_bug.cgi?id=25549
--- Comment #19 from Glenn Chugg glennlchugg@hotmail.com 2011-09-21 01:10:29 CDT --- it is very strange how common this problem is, out of the many hundreds of games I've tested under WINE, approx 1/4 of them don't have their menu's displayed correctly (same bug as ssWPI gets), I understand a lot of users don't use wine for low end games, but I enjoy them. Also Realstudio is used for windows development under WINE to save me having to run one for Linux and one for Windows, they compile and design the same. When I execute any Picture handling in RealStudio it creates the same problems, stopped me from pursuing any other WINE based projects as it's my language of choice (for speedy tool development). I just ask if the problem can not be fix that it revert to the patch above, as that at least allows it to function normally for 2d and 3d (all tests were 100% successful with the patch applied), has got to be better than the side effects it causes as in the last year or so of wine releases.
please and thanks
Glenn
http://bugs.winehq.org/show_bug.cgi?id=25549
--- Comment #20 from Glenn Chugg glennlchugg@hotmail.com 2011-09-22 04:24:38 CDT --- Patch is useless with the latest git build :(
git show ea596f8a5f7d732bb56b5921ceb964137df0b930 | patch -p1 -R patching file dlls/winex11.drv/xrender.c Unreversed patch detected! Ignore -R? [n] Apply anyway? [n] y Hunk #1 FAILED at 603. Hunk #2 FAILED at 2311. Hunk #3 FAILED at 2319. 3 out of 3 hunks FAILED -- saving rejects to file dlls/winex11.drv/xrender.c.rej
I'll just have to wait an see now.
http://bugs.winehq.org/show_bug.cgi?id=25549
--- Comment #21 from Vincent Povirk madewokherd@gmail.com 2011-09-29 12:05:27 CDT --- Well, that information makes things easy for me. Patch sent: http://www.winehq.org/pipermail/wine-patches/2011-September/107360.html
http://bugs.winehq.org/show_bug.cgi?id=25549
--- Comment #22 from Glenn Chugg glennlchugg@hotmail.com 2011-09-29 15:56:32 CDT --- Thanks Vincent, but how does one apply such a new patch that hasn't made it in to the current git yet?
I want to review your changes but don't know how to build it.
http://bugs.winehq.org/show_bug.cgi?id=25549
--- Comment #23 from Glenn Chugg glennlchugg@hotmail.com 2011-09-29 16:14:59 CDT --- patch -p1 < ~/wine-git/patch.diff
Figured it out, but the results are not so good:
http://img62.imageshack.us/img62/2615/bug2e.jpg
All the "Now displayed" images are upside down.
so the patch is 1/2 done :)
http://bugs.winehq.org/show_bug.cgi?id=25549
--- Comment #24 from Vincent Povirk madewokherd@gmail.com 2011-09-30 14:35:23 CDT --- Created attachment 36615 --> http://bugs.winehq.org/attachment.cgi?id=36615 patch, try 2
Thanks for catching my error. Any improvement with this new version?
http://bugs.winehq.org/show_bug.cgi?id=25549
--- Comment #25 from Glenn Chugg glennlchugg@hotmail.com 2011-09-30 18:40:52 CDT --- OMG!
http://img51.imageshack.us/img51/1479/happyje.jpg
PERFECT!
I am so grateful for your time you've put in to fixing this bug, I have been stuck with it not working for so long now that I'd given up on some wine projects, but now I can start them up again, as Unity now doesn't show wine installed games (very well anyway), I intend to make ssWPI optimized for Linux use again for installing and launching games from a desktop shortcut.
Anyway good work and please submit this patch into the public version so 1.3.30+ will work for everyone again.
Thank you. Your free to close this bug once submitted.
http://bugs.winehq.org/show_bug.cgi?id=25549
--- Comment #26 from Vincent Povirk madewokherd@gmail.com 2011-10-03 09:25:21 CDT --- Patch sent: http://www.winehq.org/pipermail/wine-patches/2011-October/107417.html
http://bugs.winehq.org/show_bug.cgi?id=25549
Austin English austinenglish@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED
--- Comment #27 from Austin English austinenglish@gmail.com 2011-10-03 17:28:38 CDT --- (In reply to comment #26)
Patch sent: http://www.winehq.org/pipermail/wine-patches/2011-October/107417.html
http://source.winehq.org/git/wine.git/commitdiff/5f946531593a0f934cccc22eb1f...
http://bugs.winehq.org/show_bug.cgi?id=25549
--- Comment #28 from Glenn Chugg glennlchugg@hotmail.com 2011-10-04 05:41:31 CDT --- I just fetched the latest git tree and the fix isn't implemented, just curious how long it takes for a committed patch to become integrated in the public tree?
http://bugs.winehq.org/show_bug.cgi?id=25549
Bruno Jesus 00cpxxx@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |00cpxxx@gmail.com
--- Comment #29 from Bruno Jesus 00cpxxx@gmail.com 2011-10-04 07:08:48 CDT --- You can follow the patch status here: http://source.winehq.org/patches/ If the patch is simple it usually takes one day for it to get accepted if Alexandre thinks it's right and no one else complained about it.
After doing a "git pull" inside the repository folder to update it to the last revision do a "git describe" to get the version you are in.
The URL Austin posted in the comment above proves the patch was commited as you can also see in the patch list url from this comment.
http://bugs.winehq.org/show_bug.cgi?id=25549
--- Comment #30 from Glenn Chugg glennlchugg@hotmail.com 2011-10-04 08:03:01 CDT --- Thanks, I didn't even know about "git pull", it's all new to me. As soon as I did that and waited for a big compile session (compared to the patch only compile), it worked as I expected.
I am slowly catching on to how all this works here at winehq. I like how organized things are, just it's very confusing for a new user to follow what you actually have to do (Real world usage). I think I can make do with what I've picked up now, the patches list is very informative one who is doing what and where the focus is. So I appreciate you taking the time to answer my questions.
http://bugs.winehq.org/show_bug.cgi?id=25549
Alexandre Julliard julliard@winehq.org changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED
--- Comment #31 from Alexandre Julliard julliard@winehq.org 2011-10-10 13:13:13 CDT --- Closing bugs fixed in 1.3.30.
https://bugs.winehq.org/show_bug.cgi?id=25549
André H. nerv@dawncrow.de changed:
What |Removed |Added ---------------------------------------------------------------------------- Fixed by SHA1| |5f946531593a0f934cccc22eb1f | |67f27d268501e