http://bugs.winehq.org/show_bug.cgi?id=11210
Summary: test_arb_vs_offset_limit() vertex program error Product: Wine Version: unspecified Platform: Other OS/Version: other Status: UNCONFIRMED Severity: enhancement Priority: P2 Component: directx-d3d AssignedTo: wine-bugs@winehq.org ReportedBy: eric@anholt.net
While test_arb_vs_offset_limit() appears to intend to test the limits of constant offsets from an address register, it also (I think) unintentionally tests whether ARL accepts a floating point constant as the source value, which is outside of the spec. Excerpts from ARB_vertex_program spec:
<ARL_instruction> ::= "ARL" <maskedAddrReg> "," <scalarSrcReg> <scalarSrcReg> ::= <optionalSign> <srcReg> <scalarSuffix> <srcReg> ::= <vertexAttribReg> | <temporaryReg> | <progParamReg> <scalarSuffix> ::= "." <component> <component> ::= "x" | "y" | "z" | "w"
While it may be that Mesa should also allow constants as ARL arguments, the program should probably be fixed to test only the intended out-of-spec behavior.
http://bugs.winehq.org/show_bug.cgi?id=11210
Stefan Dösinger stefandoesinger@gmx.at changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |stefandoesinger@gmx.at Status|UNCONFIRMED |RESOLVED Resolution| |FIXED
--- Comment #1 from Stefan Dösinger stefandoesinger@gmx.at 2008-03-13 16:07:28 --- I fixed that bug some time ago. The program still tests the invalid offset, but the ARL offset is fixed.
Note that real vertex programs generated from D3D shaders might still use "0.0" in ARL. That isn't yet fixed, I'd be happy if Mesa would accept that, as catching such situations would be a bit of pain, and we're pretty short on free PARAMs, and we can't really throw away a float to load 0.0 into the shader.
http://bugs.winehq.org/show_bug.cgi?id=11210
Alexandre Julliard julliard@winehq.org changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED
--- Comment #2 from Alexandre Julliard julliard@winehq.org 2008-03-21 10:54:24 --- Closing bugs fixed in 0.9.58.