Some things are UB, because they are UB in regular D, C and C++ too.

[p0nce]

• shifting by more bits than the data type size(it's defined in some intel-intrinsic operations, but not sure if legitimate)
• cast out-of-range float to an integer type that cannot hold it

=> document them

[p0nce]

The spec classify UB as illegal, say it in the DDoc.

[p0nce]

intel-intrinsics allows some instruction to shift by more than size in bits, because it has actually no runtime cost in inline optimized code. Not sure how evil that is.

[p0nce]

The question (with regards to shifting) is that D says it's UB, but the x86 instruction is well defined. And I've said at Dconf that "intel-intrinsics semantics is always modelled on the instruction they are supposed to emulate" but that may not be true, some instruction are different from their intrinsic (_mm_comieq_ss) MMMmmmmm...

[p0nce]

Also same question with _mm_alignr_pi8 / _mm_alignr_epi8 with larger counts. If we make this UB, might be better for arm vs x86.

[p0nce]

Need to see if the shift instruction agree between arm and x86, making it UB would be more correct in arm perhaps