[PATCH] D11149: Fix memcheck interval ends for pointers with negative strides

[silviu.baranga at arm.com]

sbaranga added a comment.

In http://reviews.llvm.org/D11149#204727, @anemet wrote:

> Is this a preexisting bug to grouping pointers?  IOW, did we handle swapping start and end like this for pointers before your change?


Yes, this was a pre-existing bug. Now the pointer grouping would just make it more difficult for someone to debug if they would hit this issue (besides also relying on this).

> If this was a bug even before, we probably want to benchmark this to avoid surprises.  (In theory we could have had a case where the memchecks should have failed but did not but do after your change.)


Correct. I'm running benchmarks now. Hopefully we won't get any regressions.

> Please also add a test for the non-constant case.

> 

> Thanks!


Thanks,
Silviu


http://reviews.llvm.org/D11149