[llvm-dev] funnel shift, select, and poison

[Krzysztof Parzyszek via llvm-dev]

On 2/25/2019 11:15 AM, John Regehr via llvm-dev wrote:
> I'd just like to add that the general question here is "where does 
> poison stop propagating" and this question needs to be definitively 
> answered by this community.

Does a call stop poison?

Whatever the decision is may be contradicted after inlining, so what 
should such a call return?  A superposition of poison and non-poison, 
until the function is inlined? It's a serious question.

-Krzysztof