[PATCH] [Polly][Fix] Dead memory accesses should not confuse RTC generation

[Tobias Grosser]

On 06/10/2014 14:32, Johannes Doerfert wrote:
> Thanks, will be updated.
>
> ================
> Comment at: lib/Analysis/ScopInfo.cpp:1278
> @@ +1277,3 @@
> +      // Skip accesses with an empty domain as they will never be executed.
> +      isl_set *MADomain = MA->getStatement()->getDomain();
> +      bool MADomainEmpty = isl_set_is_empty(MADomain);
> ----------------
> simbuerg wrote:
>> You are iterating over all memory accesses of one statement, so I assume MA->getStatement() is always equal to Stmt, right?
>> With that in mind, you can do the empty check outside this loop and skip the whole loop, if true.
> 100% correct... my bad.
>
> http://reviews.llvm.org/D5623

Assuming you address Andreas' comment, this LGTM.

Tobias