Ian Romanick <i...@freedesktop.org> writes: > On 12/20/2012 05:05 PM, Eric Anholt wrote: >> Ian Romanick <i...@freedesktop.org> writes: >> >>> From: Ian Romanick <ian.d.roman...@intel.com> >> >> It looks to me like we really do have to support getting a >> binary with this extension/gles3: >> >> "Any program binary retrieved using GetProgramBinary and submitted us- >> ing ProgramBinary under the same configuration must be successful. Any >> programs loaded successfully by ProgramBinary must be run properly with >> any legal GL state vector." >> >> where I read configuration as "hardware and driver version" I don't see >> any of the text I would expect if they're letting length == 0 mean >> "sorry, can't save this one". > > I thought I had replied to this, but I don't see my e-mail. Weird. > Anyway, the spec allows us to expose zero binary formats. If there are > no binary formats, any value specified as binaryFormat for > glProgramBinary must be invalid. The older ARB_get_shader_binary spec > uses basically the same method to allow the same cowardly retreat. :)
Oh, of course. I was missing that.
pgpWL1nTUDga0.pgp
Description: PGP signature
_______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
