[PATCH] Revert "[PECOFF] Add atoms to the PassManager file"
Shankar Easwaram
shankarke at gmail.com
Wed Oct 30 18:35:58 PDT 2013
It became more of a discussion post this revert as I was unclear on how the atom got inserted into the file twice and all tests were still passing. I think there is a lack of test otherwise this could have been caught earlier.
We should add a test along with the revert to make sure the revert actually fixes something.
Sent from my iPhone
> On Oct 30, 2013, at 19:32, Chandler Carruth <chandlerc at gmail.com> wrote:
>
>
> Also, if you want to change the design of how atoms get added, I really don't see why that should happen in this change.
>
> It seems much better to revert a change which introduced a regression, and *then* consider new or different designs that would be simpler or more efficient for some reason.
>
> http://llvm-reviews.chandlerc.com/D2069
More information about the llvm-commits
mailing list