Fixes a possible infinite loop in glsl optimization due to interaction
between copy_propagation_elements and swizzle_swizzle. This approach is
based on handle_rvalue appearing to already handle a swizzle's
referenced value. As a result it does not seem correct to also call
handle_rvalue with the re
Hi all,
So this is my attempt at a fix for
https://bugs.freedesktop.org/show_bug.cgi?id=34006. Any feedback on if this
looks right to anyone is much appreciated. Sven Arvidsson did confirm on
bugzilla that the patch does not cause any piglit regressions and does also fix
the infinite loop issu