[Bug tree-optimization/102711] [9/10/11 Regression] CDDCE removes condition that might lead to an infinite loop causing an unconditional infinite loop

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102711

--- Comment #6 from Martin Liška <marxin at gcc dot gnu.org> ---
> I see - probably makes sense since that subtly changes post-dominators
> for backwards unreachable regions which reflects back on control
> dependence.  But there's nothing "wrong" here, I think it avoided
> placing the fake exit edges at the infinite loop headers as opposed
> to the latch but I don't remember exactly.

Do we still need more bisection?