[Mesa-dev] [PATCH 1/7] compiler: Add enums for blend state

[Alyssa Rosenzweig]

> Interesting, I wouldn't have split blend_factor from inverting the
> factor.  I'm fine with it, though.

The rationale was that the code logic needs it split anyway, so the
split has to happen in one place or another... at least this way there's
less redundant enum'ing. TBH, I'd prefer Gallium had it this way too,
and GL for that matter, but it's too late for those :)

Thanks for the ack.