Hi,
Il 13/05/22 01:10, Zebediah Figura ha scritto:
This uses more storage than necessary for e.g. sm1 scalars, though. I think we can use hlsl_type_component_count() instead.
Ok, makes sense.
(All of the RA logic desperately needs to be rewritten, but oh well...)
Do we consider this a realistic/sensible target for 1.4? Or should we postpone that to 1.5 (or whatever comes next)? Given a number of things we'd like to have in 1.4 we have already informally discussed, I guess we can postpone this.
Gio.