[llvm-commits] [PATCH][SOPT] Fix assertion in Reassociate

Shuxin Yang shuxin.llvm at gmail.com
Thu Nov 15 10:54:32 PST 2012


Chandler:
    It is not simply walk forward. It need to pop from the front.

Shuxin

On 11/15/12 10:51 AM, Chandler Carruth wrote:
> On Thu, Nov 15, 2012 at 10:41 AM, Shuxin Yang <shuxin.llvm at gmail.com> wrote:
>>     2) not only fix (you may think it is merely sidestep) the problem.
>>        It include the enhancement to the worklist, making it possible to go
>> through the worklist in forward order.
> I've tried to explain this before -- it is possible to walk worklists
> based on SetVector in forward order. We don't need a new abstraction
> for this.

>>        It also remove the cost the erasing element in the middle of the
>> container.
>>        It also include the optimization for "t = -a; t2 = t * 7 (where t has
>> multiple uses)" => t2 = a * -7"
>>
>>      How about all this changes?
>>
>>     Down the load, please let us know if you want to own the Reassociate.cpp
>> and have  *EXCLUSIVELY" write-permission or not.
> Clearly Duncan doesn't have exclusive write permission -- no one does.
> He was just fixing the root cause of the issue, and felt that the root
> cause should be fixed (at least initially) rather than avoiding the
> crash without fixing the root cause. I think that's a reasonable
> stance to take as a reviewer in general.




More information about the llvm-commits mailing list