[PATCH] D70091: [BranchFolding] Fix PR43964 about branch folder not being debug invariant

[Bjorn Pettersson via Phabricator via llvm-commits]

bjope marked an inline comment as done.
bjope added a comment.

In D70091#1753302 <https://reviews.llvm.org/D70091#1753302>, @jmorse wrote:

> This looks good to me and great refactor. To make sure I understand what's going on: by keeping a temporary iterator in ComputeCommonTailLength, and only "committing" the advance to I1/I2 when we're sure we've found some additional identical tail, this avoids all the concerns about I1/I2 ending up pointing at pseudo instructions?


Yes, simply scan for the next non-pseudo instructions to compare. If they are identical "commit" the advance and continue searching. Otherwise we are done.
ComputeCommonTailLength should give the same result as the old code using that algorithm. But the old code was really hard to read and used goto.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D70091/new/

https://reviews.llvm.org/D70091