[llvm] de18fa9 - Revert "[InstSimplify] Bypass no-op `and`-mask, using known bits (PR49543)"

[Nikita Popov via llvm-commits]

On Wed, Apr 21, 2021 at 6:21 PM Roman Lebedev <lebedev.ri at gmail.com> wrote:

> 1. I don't recall getting a message "hey, this is causing xyz, any
> thoughts, should revert?"
> 2. I don't recall seeing a follow-up mail to the original commit that
> it was reverted
>

You replied to the revert email before I had a chance to message the
original commit. From your reply to the revert I assumed you were already
aware of the revert. Of course, now that you mention it, mailing the
original commit may still be useful for others.

As you are a very heavy user of post-commit review, please expect that
reverts of your commits will be more common than if you go through
pre-commit review. If you had gone through review with this change, I would
have left my concerns on the review. As you did not go through a review, I
reverted to status quo, indicated my primary point of contention, and
suggested that the patch go through review.

In most cases I would simply leave comments on specific parts of the patch,
with the expectation that they be addressed in due time. In this instance,
I disagree with the patch in concept, so that wouldn't work.

Possibly I misunderstood our review/revert policy, but I was under the
impression that it's always possible to request that a post-commit review
be converted into a pre-commit review, with the change being reverted in
the meantime.

Regards,
Nikita
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20210421/ed7d8f4d/attachment.html>