[ tree-ssa] PRE problem

[Daniel Berlin]

>> In fact, no backwards subst PRE is performing should be between things
>> with different memory tags, since it only changes phi results into the
>> associated phi argument for that edge.
>> This shouldn't ever change the memory tag.
> Wrong again.
>

Then this was changed at some point.
Whoever changed it so that some of these things could have null memory 
tags should have mentioned this, and i would have pointed out that they 
needed to update PRE.

> jeff
>
>