This just updates the conversion rank for minimum precision types which are now represented in the Basic.types section.
I've iterated on this a few times and reduced it down to just the min16{int|uint|float} types because those are the only types DXC actually supports. Using any of the other types produces a warning and they get treated as the corresponding min16 type.
This just updates the conversion rank for minimum precision types which are now represented in the Basic.types section.
I've iterated on this a few times and reduced it down to just the min16{int|uint|float} types because those are the only types DXC actually supports. Using any of the other types produces a warning and they get treated as the corresponding min16 type.
The promotion orders are as follows:
min16int < int16_t < int32_t < int64_t min16uint < uint16_t < uint32_t < uint64_t min16float < half < float < double