steakhal wrote:
I wanted to come back with another example:
```c++
int ternary_in_assume(int a, int b) {
[[assume(a > 10 ? b == 4 : b == 10)]];
if (a > 20)
return b + 100; // expecting 104
if (a > 10)
return b + 200; // expecting 204
return b + 300; // expecting 310
}
```
Actually, neither gcc, clang, msvc, or icc optimizes the return paths to
constants - even though I think they would be allowed to.
https://compiler-explorer.com/z/1ajvcvG9h
I'd need to think about of the implications for us though. I think it should be
safe to eval these subexpressions and do state splits if and only if none of
those expressions have sideeffects. In that case, they wouldn't be emitted into
the CFG, so they shouldn't be a problem.
https://github.com/llvm/llvm-project/pull/116462
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
