Fixes: nir: add missing vec opcodes in lower_bool_to_float
This is required to avoid assert failures on debug builds.
I copied this from lower_bool_to_int32, but perhaps we should return false on these opcodes if the bit_size is already 32?
This is required to avoid assert failures on debug builds.
I copied this from lower_bool_to_int32, but perhaps we should return false on these opcodes if the bit_size is already 32?