https://bugs.freedesktop.org/show_bug.cgi?id=53256
Chris Wolfe <cwo...@chromium.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
AssignedTo|i...@freedesktop.org |mesa-dev@lists.freedesktop.
| |org
--- Comment #1 from Chris Wolfe <cwo...@chromium.org> 2012-08-09 15:41:30 UTC
---
Re-aiming at the list, guess the CC isn't automatic. Please see the bug for
context.
Summary: GLSL optimizer mangles "vec4 tmp; tmp[3] = 1.0; tmp.xyz =
some_3d_vector;"
Walked through the optimization behavior in opt_dead_local_code (and friends)
again. Feels like the best fix is to consider assignment to a deference_array
of a vector as a write to an unknown component.
That way the assignment will get removed only if there is an overriding
whole_variable_written to the vector. If the array index is constant it will
get converted to a swizzle (vec_index_to_swizzle), and hopefully become a
deference_variable for the next pass (haven't found the code for that yet).
--
Configure bugmail: https://bugs.freedesktop.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.
_______________________________________________
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/mesa-dev
