I’ve read this piece:
http://blog.regehr.org/archives/1307
And now I’m wondering if C code generated from ATS sources may contain UBs
due to strict aliasing?
I’ve read this piece:
http://blog.regehr.org/archives/1307
And now I’m wondering if C code generated from ATS sources may contain UBs
due to strict aliasing?
Actually, I am not so sure about this. Maybe in ATS1?
In ATS2. you essentially can only have void* pointers; casting
happens at the point where dereference is done. So something
like
void foo (int *x, long *y)
should not be generated. If it gets generated, I would definitely
look into the issue.On Tuesday, March 15, 2016 at 11:34:23 PM UTC-4, Artyom Shalkhakov wrote:
I’ve read this piece:
The Strict Aliasing Situation is Pretty Bad – Embedded in Academia
And now I’m wondering if C code generated from ATS sources may contain UBs
due to strict aliasing?