http://bugs.winehq.org/show_bug.cgi?id=20922
Summary: Wine generates invalid ARB vertex programs Product: Wine Version: 1.1.34 Platform: PC OS/Version: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: directx-d3d AssignedTo: wine-bugs@winehq.org ReportedBy: m.cencora@gmail.com
Created an attachment (id=25082) --> (http://bugs.winehq.org/attachment.cgi?id=25082) Assassin's Creed game output, the bad program starts at line 979
When emulating sin/cos instructions in vertex programs, lines like these are created: MUL TA.x, R0.x, C4.w.w; MAD TA.x, TA.y, C4.w.x, TA.x;
The problem is the double swizzle (i.e. C4.w.w C4.w.x).
http://bugs.winehq.org/show_bug.cgi?id=20922
Stefan Dösinger stefandoesinger@gmx.at changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |stefandoesinger@gmx.at
--- Comment #1 from Stefan Dösinger stefandoesinger@gmx.at 2009-12-06 05:46:02 --- Hi, Can you run the program with WINEDEBUG=d3d_shader ? I suspect that the input shader is already incorrect. src1 and src2 - where the double swizzle occurs - are supposed to be 4 component constant registers. The double swizzle occurs because the app sets a swizzle on src1 and src2(2nd and third param) and the shader_addline code simply concatenates ".w" without looking at the input swizzle. The question is if it should. C4.w is a single scalar, it can't possibly work with what the helper constants are supposed to do.
I'll have to write a test case to figure out how Wine is supposed to behave here. I suspect we should reject the shader and return an error when the app tries to create it. Its also possible that we should simply ignore the swizzle and treat "C4.w" as "C4" in this specific instruction. Or, maybe we should honor the swizzle and return incorrect results because the taylor coefficients aren't correct.
http://bugs.winehq.org/show_bug.cgi?id=20922
--- Comment #2 from Maciej Cencora m.cencora@gmail.com 2009-12-06 07:41:39 --- Created an attachment (id=25090) --> (http://bugs.winehq.org/attachment.cgi?id=25090) Compressed log
Grep for "Failing program" when looking for relevant info.
http://bugs.winehq.org/show_bug.cgi?id=20922
--- Comment #3 from Austin English austinenglish@gmail.com 2011-10-02 11:45:34 CDT --- This is your friendly reminder that there has been no bug activity for 1 year. Is this still an issue in current (1.3.29 or newer) wine?
http://bugs.winehq.org/show_bug.cgi?id=20922
Austin English austinenglish@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|UNCONFIRMED |RESOLVED Resolution| |ABANDONED
--- Comment #4 from Austin English austinenglish@gmail.com 2013-11-13 15:38:34 CST --- No reply in over a year, and no download. Resolving abandoned, if anyone is still able to reproduce this in the current development version (currently 1.7.6) feel free to reopen.
http://bugs.winehq.org/show_bug.cgi?id=20922
Austin English austinenglish@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED
--- Comment #5 from Austin English austinenglish@gmail.com 2013-11-13 15:40:25 CST --- Closing.
-
wine-bugs＠winehq.org