[llvm-commits] [PATCH] Optimistically analyse phi cycles in BasicAA

[Duncan Sands]

Hi Arnold,

On 20/11/12 23:32, Arnold Schwaighofer wrote:
> Analyse Phis under the starting assumption that they are NoAlias (we
> used to require the first operand to be NoAlias). This is correct
> because we recursively look at their inputs. If the Phis
> MayAlias/MustAlias there must be an input that makes them so.
>
> This addresses bug 14351.
>
> Bootstrapped clang passes the test-suite.

what happens if the recursion bails out due to hitting a recursion limit?
Will the conservatively correct MayAlias be returned in that case?

Ciao, Duncan.