[Mlir-commits] [mlir] [MLIR] Fix computeTiedSuccessorInputs for single-iteration scf.for (PR #188986)

[Mehdi Amini]

joker-eph wrote:

We already canonicalize these by performing the inline you mention, however canonicalization is optional and we can't afford to miscompile when it isn't performed.

I didn't add a test because this is fixing MLIR_ENABLE_EXPENSIVE_PATTERN_API_CHECKS failures, so the test actually already exists.

https://github.com/llvm/llvm-project/pull/188986