I don't have commit rights, could somebody please land this for me?

Thanks for the feedback.

Addressed @nikic's comments, thanks for the feedback.

Abandoning revision since the transformation is not safe.

In D124153#3468447, @zero9178 wrote:

• A debug build on Windows using MSVC can take around 70 to roughly 100 GB of storage space, which are always surprising and may also lead to out of disk space errors. Release builds are comparatively just a few Gigabytes. I believe they are also multiply tens of Gigabytes on other OSs

Addressed comments, I kept the lambda.

Rebased

Updated the patch with update_test_checks.py run with the changes applied on top of https://reviews.llvm.org/rGfd9026131e6c5de6f4e582d5c85c5b5588672dfb
Looks like what I'd expect. It's optimizing it through the opaque use call, but that doesn't look like a problem to me.

Updated the diff with the 2 tests to ensure the opt is not applied on non-exact ashr/lshr ashr_add_negative_shift_no_exact and lshr_add_negative_shift_no_exact.

I collapsed all three into InstCombinerImpl::commonShiftTransforms, right below the pre-existing pre-shift combine.

Updating the diff with the condition change and more tests. Will see about avoiding the duplication and moving into commonShiftTransforms

In D124369#3471465, @spatel wrote:

1. Please write a more general proof for these transforms. See examples in the comments here: https://github.com/llvm/llvm-project/issues/54890
2. Does the lshr transform require "exact" or not?