http://bugs.winehq.org/show_bug.cgi?id=9659
Summary: Guild Wars: Rendering of 3D models broken Product: Wine Version: 0.9.45. Platform: PC URL: http://www.guildwars.com/downloads/gwsetup.zip OS/Version: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: wine-directx AssignedTo: wine-bugs@winehq.org ReportedBy: nemonbox@hotmail.com
When updating from 0.9.44 to 0.9.45 all 3D models (characters and items/boxes/vegitation ect.) in Guild Wars become invisible, nothing there at all, everything else renders as it should altho text renders at first, after a few seconds/minutes depening on activity (more = faster) test disappears aswell.
Client can be downloaded altho you can't get far in without an account you can see the text problem just opening options menu at start and klicking back and forward between the general, grafic and sound tab without making any changes.
http://bugs.winehq.org/show_bug.cgi?id=9659
Austin English austinenglish@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Component|wine-directx |wine-misc Keywords| |download, regression
--- Comment #1 from Austin English austinenglish@gmail.com 2007-09-15 08:17:27 --- Please run a regression test:
http://bugs.winehq.org/show_bug.cgi?id=9659
--- Comment #2 from Austin English austinenglish@gmail.com 2007-09-15 14:16:55 --- Whoops, link might help:
http://wiki.winehq.org/RegressionTesting
http://bugs.winehq.org/show_bug.cgi?id=9659
--- Comment #3 from Ferrat nemonbox@hotmail.com 2007-09-16 21:43:00 --- Created an attachment (id=8112) --> (http://bugs.winehq.org/attachment.cgi?id=8112) Guild Wars Regression test
Regression test from 0.9.44 to 0.9.45
regression test of wine3d as result of 3D models not appering after update
http://bugs.winehq.org/show_bug.cgi?id=9659
Ferrat nemonbox@hotmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Component|wine-misc |wine-directx-d3d
http://bugs.winehq.org/show_bug.cgi?id=9659
Ferrat nemonbox@hotmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |stefandoesinger@gmx.at
http://bugs.winehq.org/show_bug.cgi?id=9659
--- Comment #4 from Ferrat nemonbox@hotmail.com 2007-09-16 22:08:06 --- don't know if this helps but noticed that some textures rendered ontop of others show even when the rest don't, like hats/bandanas ect.
http://bugs.winehq.org/show_bug.cgi?id=9659
Stefan Dösinger stefandoesinger@gmx.at changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |wine-bugs@winehq.org AssignedTo|wine-bugs@winehq.org |stefandoesinger@gmx.at
--- Comment #5 from Stefan Dösinger stefandoesinger@gmx.at 2007-09-17 05:05:04 --- Can you attach a d3d log? I am afraid that the game runs into some Z precision problems now. Maybe it uses D16 depth buffers, but Wine only has 24 bit ones due to some winex11.drv limitations. Before my patch, we lost half the Z precision because of the incorrect scaling, so it might have worked by chance.
http://bugs.winehq.org/show_bug.cgi?id=9659
--- Comment #6 from Ferrat nemonbox@hotmail.com 2007-09-17 05:57:51 --- pretty new at this, how do I get a d3dlog output? =)
http://bugs.winehq.org/show_bug.cgi?id=9659
--- Comment #7 from Stefan Dösinger stefandoesinger@gmx.at 2007-09-17 06:29:28 --- WINEDEBUG=d3d wine foo.exe 2> /some/output/file
Then compress the created file(bz2 -9, or with lrzip) and attach it here.
http://bugs.winehq.org/show_bug.cgi?id=9659
--- Comment #8 from Ferrat nemonbox@hotmail.com 2007-09-17 08:09:35 --- http://files-upload.com/files/508266/d3dlog.tar.bz2
log output for the d3d, hade to upload it to another place since it's 1,6MB and to big to upload here..
if you need anything else just say :)
http://bugs.winehq.org/show_bug.cgi?id=9659
volker.groeschel@sap.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |volker.groeschel@sap.com
--- Comment #9 from volker.groeschel@sap.com 2007-09-17 09:06:14 --- I guess it is the same problem I'm faced within Metin2 http://www.metin2.us/download.htm (I actually play the German version from http://www.metin2.de/). Metin2 is a free massive multiplayer online role-playing game (MMORPG) comparable with WoW. But did I mention it is free :-)
The situation got better and better until 0.9.44. But with 0.9.45 there are - a lot missing textures - very poor performance when I use my charms (blue light effects)
I can run WINEDEBUG=d3d this evening when you are interested.
My system: Ubuntu 7.04 with Nvidia Gforce 6600 binary driver.
http://bugs.winehq.org/show_bug.cgi?id=9659
--- Comment #10 from Ferrat nemonbox@hotmail.com 2007-10-01 10:54:06 --- Problem persists in 0.9.46
http://bugs.winehq.org/show_bug.cgi?id=9659
Ferrat nemonbox@hotmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|UNCONFIRMED |NEW Ever Confirmed|0 |1
--- Comment #11 from Ferrat nemonbox@hotmail.com 2007-10-01 10:55:15 --- *** This bug has been confirmed by popular vote. ***
http://bugs.winehq.org/show_bug.cgi?id=9659
--- Comment #12 from Stefan Dösinger stefandoesinger@gmx.at 2007-10-01 13:11:52 --- Does WineD3D write some FIXMEs about EQUAL or NOTEQUAL Z tests?
http://bugs.winehq.org/show_bug.cgi?id=9659
Geobert geobert@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |geobert@gmail.com
--- Comment #13 from Geobert geobert@gmail.com 2007-10-20 04:37:26 --- about the bug for metin2, it is still present in 0.9.47
http://bugs.winehq.org/show_bug.cgi?id=9659
--- Comment #14 from Ferrat nemonbox@hotmail.com 2007-10-20 05:05:09 --- I get no FIXME's from what I can see but it has come to my attention that not everyone experiences this according to people on the ubuntu forums but some do, might be an ATi/nVIDIA thing (Myself running on ATi) but havn't gotten reply's to confirm that it only applys to ATi
http://bugs.winehq.org/show_bug.cgi?id=9659
--- Comment #15 from Ferrat nemonbox@hotmail.com 2007-10-20 05:05:45 --- I get no FIXME's from what I can see but it has come to my attention that not everyone experiences this according to people on the ubuntu forums but some do, might be an ATi/nVIDIA thing (Myself running on ATi) but havn't gotten reply's to confirm that it only applys to ATi
and yes still the same problem in 0.9.47
http://bugs.winehq.org/show_bug.cgi?id=9659
--- Comment #16 from Stefan Dösinger stefandoesinger@gmx.at 2007-10-27 17:32:24 --- I have downloaded the guild wars client, but I cannot see any problems in the main menu, even in the options dialog. I do not have an account yet, so I cannot test in-game.
Can you retest this bug?
http://bugs.winehq.org/show_bug.cgi?id=9659
--- Comment #17 from Ferrat nemonbox@hotmail.com 2007-10-27 17:59:27 --- Are you using a nVIDIA card or ATi? sofar it only seems to manifest under ATi
http://bugs.winehq.org/show_bug.cgi?id=9659
--- Comment #18 from Ferrat nemonbox@hotmail.com 2007-11-06 02:08:28 --- Found that more games than Guild Wars has this problem, Neverwinter Nights 2 has the same problem with rendering the 3D models after 0.9.45 but it seems to be just this value, games that doesn't use it (I presume) like Baldur's Gate 2 works much more stable and faster, have yet to find out if it's only some ATi cards or all but from what I gathered no nVIDIA card has this problems.
http://bugs.winehq.org/show_bug.cgi?id=9659
--- Comment #19 from Stefan Dösinger stefandoesinger@gmx.at 2007-11-06 05:21:39 --- fwiw, I'm running guild wars without issues on my nvidia boxes and my mac with an X1600 under MacOS. So it looks like an fglrx-centric issue.
What happens in Guild Wars if you use the -noshaders launch option?
http://bugs.winehq.org/show_bug.cgi?id=9659
--- Comment #20 from Ferrat nemonbox@hotmail.com 2007-11-06 06:17:32 --- Same problem without shaders, with -dx8 most of the models render but text is still unstable disappering after a few seconds
Screens <a href="http://img204.imageshack.us/my.php?image=dx8ug3.jpg" target="_blank"><img src="http://img204.imageshack.us/img204/8195/dx8ug3.th.jpg" border="0" alt="Free Image Hosting at www.ImageShack.us" />[TAGS: -dsound -dx8 -noshaders]</a> as you can see the models show
<a href="http://img84.imageshack.us/my.php?image=dx8shadersku6.jpg" target="_blank"><img src="http://img84.imageshack.us/img84/4364/dx8shadersku6.th.jpg" border="0" alt="Free Image Hosting at www.ImageShack.us" />[TAGS: -dsound -dx8]</a> as you can see the models doesn't show
<a href="http://img204.imageshack.us/my.php?image=noshaderslr8.jpg" target="_blank"><img src="http://img204.imageshack.us/img204/3227/noshaderslr8.th.jpg" border="0" alt="Free Image Hosting at www.ImageShack.us" />[TAGS: -dsound -noshaders]</a> as you can see the models doesn't show
<a href="http://img411.imageshack.us/my.php?image=shadersto3.jpg" target="_blank"><img src="http://img411.imageshack.us/img411/2117/shadersto3.th.jpg" border="0" alt="Free Image Hosting at www.ImageShack.us" />[TAGS: -dsound]</a> as you can see the models doesn't show BUT! removing the shaders reveals the Bandana on the model?
http://bugs.winehq.org/show_bug.cgi?id=9659
--- Comment #21 from Stefan Dösinger stefandoesinger@gmx.at 2007-11-06 08:40:21 --- Please attach screenshots to the bug report instead of linking them from an external site. This will make sure that they remain accessible in the future.
"-dx8 -noshader" should be the same as "-noshader", and it is indeed as far as I can see. It looks like there is an fglrx bug here. The d3d9 testsuit has a test for the depth buffer behavior which shows the correctness of the patch that was found to cause the regression. If this test is broken on fglrx it might help debugging this bug.
You can run the test from dlls/d3d9/tests, by running wine "d3d9_test.exe.so visual" . Perhaps enable the virtual desktop before doing so, and *make sure that no window overlaps the test window* .
http://bugs.winehq.org/show_bug.cgi?id=9659
--- Comment #22 from Ferrat nemonbox@hotmail.com 2007-11-06 08:59:48 --- Sorry will fix that, just trying to save some band :)
Well ran wine d3d9_test.exe.so visual
fixme:win:EnumDisplayDevicesW ((null),0,0x34f060,0x00000000), stub! fixme:d3d_draw:drawPrimitive Using software emulation because manual fog coordinates are provided fixme:d3d_draw:drawPrimitive Using software emulation because manual fog coordinates are provided fixme:d3d_draw:drawPrimitive Using software emulation because manual fog coordinates are provided fixme:d3d:IWineD3DVolumeImpl_LockBox (0x6c4d98) : pBox=(nil) stub visual.c:3720: Tests skipped: No vs_2_0 support visual.c:2809: Test failed: D3DFMT_X8L8V8U8 = 0x112131ca returns color 008363ca, expected 0x008262ca visual.c:3740: Tests skipped: No ps_3_0 support visual: 1785 tests executed (0 marked as todo, 1 failure), 2 skipped.
That's all that comes out when running it, so just one failure, not sure if' it's the one you're looking for :)
http://bugs.winehq.org/show_bug.cgi?id=9659
--- Comment #23 from Ferrat nemonbox@hotmail.com 2007-11-06 09:04:04 --- Created an attachment (id=8990) --> (http://bugs.winehq.org/attachment.cgi?id=8990) -dx8 -noshaders
This is when running the tags -dx8 -noshaders
http://bugs.winehq.org/show_bug.cgi?id=9659
--- Comment #24 from Ferrat nemonbox@hotmail.com 2007-11-06 09:06:03 --- Created an attachment (id=8991) --> (http://bugs.winehq.org/attachment.cgi?id=8991) -dx8
This is when running -dx8 the ground is gone model is gone the frames around the pictures is gone
http://bugs.winehq.org/show_bug.cgi?id=9659
--- Comment #25 from Ferrat nemonbox@hotmail.com 2007-11-06 09:08:06 --- Created an attachment (id=8992) --> (http://bugs.winehq.org/attachment.cgi?id=8992) dx9 -noshaders
this is when running with -noshader tag
The frame for the picture is there The gound is there The Models ain't there
http://bugs.winehq.org/show_bug.cgi?id=9659
--- Comment #26 from Ferrat nemonbox@hotmail.com 2007-11-06 09:11:45 --- Created an attachment (id=8993) --> (http://bugs.winehq.org/attachment.cgi?id=8993) dx9 no tags (except -dsound)
this is just more or less running the Gw.exe without tags
The ground is there The picture frames are there The models ain't there BUT if you look close, you see the bandana both in the picture frame and in the fullview area
http://bugs.winehq.org/show_bug.cgi?id=9659
--- Comment #27 from Stefan Dösinger stefandoesinger@gmx.at 2007-11-06 12:27:58 --- Indeed, looks like the Z test runs fine. The one failure is a bug in the test, and I'll send a patch. 0x008363ca is very close to the expected 0x008262ca, so it is just a precision problem and the test is too picky. 3D rendering isn't promised to be exact. So tracking that down will be a bit more tricky than just running a test.
http://bugs.winehq.org/show_bug.cgi?id=9659
--- Comment #28 from Ferrat nemonbox@hotmail.com 2007-11-06 17:17:03 --- Hmm ok, don't ask me how or exactly what I did, was just fiddeling around with settings in regedit and winecfg for another game and when I started Guild Wars I got 3D models? o.O working with shaders and no black/pink boxes around the sprits far away... really weird part is that the settings Im using now is AFAIK the same as I had before...
Text still disappears like it's not rendered over and over again but the game otherwise is smooth and better looking than ever, running on max settings..
http://bugs.winehq.org/show_bug.cgi?id=9659
volker.groeschel@sap.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC|volker.groeschel@sap.com |
http://bugs.winehq.org/show_bug.cgi?id=9659
--- Comment #29 from Ferrat nemonbox@hotmail.com 2007-11-15 12:50:57 --- Well HATE ATi right now >_<
DisabledExtensions > GL_ARB_draw_buffers
seems to work in Guild Wars to get the models to render but just tried GTA:SA and the models there are gone aswell and DisabledExtensions > GL_ARB_draw_buffers doesn't help :(
http://bugs.winehq.org/show_bug.cgi?id=9659
--- Comment #30 from Jean Michel jean.bruenn@ip-minds.de 2007-11-24 10:28:25 --- Created an attachment (id=9314) --> (http://bugs.winehq.org/attachment.cgi?id=9314) It contains an example of broken 3d moddeling in the game
Hello, i have this bug, too. But only when using -dx9 - with directx8 moddeling seems fine.
http://bugs.winehq.org/show_bug.cgi?id=9659
--- Comment #31 from Stefan Dösinger stefandoesinger@gmx.at 2007-11-24 10:54:59 --- Are you using GLSL or ARB shaders? I spotted a compiler bug on MacOS I could not reproduce in a test app in the code sequence that modifies the Z coordinate. This bug caused the skybox to appear in front of most objects.
http://bugs.winehq.org/show_bug.cgi?id=9659
Geobert geobert@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC|geobert@gmail.com |
http://bugs.winehq.org/show_bug.cgi?id=9659
Joseph Davis narcberry@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |narcberry@gmail.com
--- Comment #32 from Joseph Davis narcberry@gmail.com 2007-12-06 10:58:27 --- I am having this problem still using an ATI card. Using directx8, thanks for that information, I can get models but text still disappears.
I hope I'm not being the peanut gallery here, but I looked through the changelog in 0.9.45 where the problem started and found this:
wine-0.9.45 wine/ChangeLog:
614 2007-08-31 Stefan Dösinger stefandoesinger@gmx.at 615 616 * dlls/d3d9/tests/visual.c, dlls/wined3d/arb_program_shader.c, 617 dlls/wined3d/glsl_shader.c, dlls/wined3d/pixelshader.c, 618 dlls/wined3d/wined3d_private.h: 619 wined3d: Fix texdepth instruction. 620 621 * dlls/d3d9/tests/visual.c, dlls/wined3d/state.c, 622 dlls/wined3d/vertexshader.c: 623 wined3d: Fix the z range.
Looks like the place to start, I'm hoping to have some time by this weekend to run a diff. I hope I'm being helpful.
http://bugs.winehq.org/show_bug.cgi?id=9659
--- Comment #33 from Joseph Davis narcberry@gmail.com 2007-12-06 11:04:00 --- Also, you can get the text back on the screen by changing either your resolution or your ui size. There might be other methods. I'm currently running the game without text and spam changing my ui size when I need text, like when I'm selling items. F11 can be your friend for this.
http://bugs.winehq.org/show_bug.cgi?id=9659
--- Comment #34 from Dustin McCartney dustin.mccartney.public@gmail.com 2007-12-09 17:04:34 --- Created an attachment (id=9578) --> (http://bugs.winehq.org/attachment.cgi?id=9578) Geometry rendering error
Wine 0.9.50 - Guildwars often renders models incorrectly. It looks like some of the vertices have invalid (possibly un-initialized) positions, which causes polys to render with one or more of their position coordinates with unreasonable values.
I am using an Nvidia Geforce 7600 GT, so contrary to earlier reports, this is not an ATI specific issue.
http://bugs.winehq.org/show_bug.cgi?id=9659
Dustin McCartney dustin.mccartney.public@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |dustin.mccartney.public@gmai | |l.com
--- Comment #35 from Dustin McCartney dustin.mccartney.public@gmail.com 2007-12-09 17:27:00 --- In the past, I have seen similar behavior occur under 2 conditions:
1) One or more values in an index buffer references vertices which are out of range. Ie: A vertex buffer contains 100 vertices (0 thru 99) but one or more values in the OpenGL index buffer reference values outside this range (for example, an index value of 100 or 101) which attempts to use a vertex that is undefined. This would be a problem with invalid indice(s) in the index buffer.
...or...
2) A vertex buffer is being created from a set of data that fails to fully fill the vertex buffer. Ie: A vertex buffer expects 100 vertices... The source data that fills the vertex buffer only fills (or copies data into) the first 96 vertices instead of filling all 100. The index buffer then refers to vertices in the range between 0-99, however whenever an index references the last 4 vertices (which have not been initialized/stale), invalid vertex positions are read which cause rendering artifacts. This would be a problem filling the vertex buffer with the proper/expected values.
I do not _believe_ that this is a "z-depth" issue as discussed earlier.
http://bugs.winehq.org/show_bug.cgi?id=9659
--- Comment #36 from Stefan Dösinger stefandoesinger@gmx.at 2007-12-09 17:41:23 --- The last report is a different bug, fixed in git already. It was caused by incorrect rounding of floats to ints in GLSL shaders. We applied the shader model 2.0 / 3.0 round to nearest, while 1.x shaders that guild wars uses perform a floor(), even though the docs say otherwise.
http://bugs.winehq.org/show_bug.cgi?id=9659
--- Comment #37 from Dustin McCartney dustin.mccartney.public@gmail.com 2007-12-09 17:43:53 --- (In reply to comment #36)
The last report is a different bug, fixed in git already. It was caused by incorrect rounding of floats to ints in GLSL shaders. We applied the shader model 2.0 / 3.0 round to nearest, while 1.x shaders that guild wars uses perform a floor(), even though the docs say otherwise.
Well then... my bad. :-) Thanks for the update. I'll git and build to verify this is fixed then. :-D
Thanks! Dustin
http://bugs.winehq.org/show_bug.cgi?id=9659
--- Comment #38 from Timo-Heikki Mäkelä imaxfun@gmail.com 2008-03-12 13:23:15 --- FIXED?
http://bugs.winehq.org/show_bug.cgi?id=9659
Matze Sternler starmatz71@yahoo.de changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |starmatz71@yahoo.de
--- Comment #39 from Matze Sternler starmatz71@yahoo.de 2008-09-13 14:31:25 --- (In reply to comment #38)
FIXED?
Fixed for me.
http://bugs.winehq.org/show_bug.cgi?id=9659
Austin English austinenglish@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED
--- Comment #40 from Austin English austinenglish@gmail.com 2008-09-13 14:43:17 --- Reported fixed.
http://bugs.winehq.org/show_bug.cgi?id=9659
Alexandre Julliard julliard@winehq.org changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED
--- Comment #41 from Alexandre Julliard julliard@winehq.org 2008-09-19 11:16:37 --- Closing bugs fixed in 1.1.5.
-
wine-bugs＠winehq.org