[PATCH] D51861: [LSR] Combine unfolded offset into invariant register

[Quentin Colombet via Phabricator via llvm-commits]

qcolombet added a comment.

> Since testing for optimized or illegal code in these cases isn't an option in these cases, such a LIT could check that the incorrect code is not generated, that the incorrect formula is not generated, or that the correct formula is generated - which all seem fragile and overfitted to this specific bug, so I'm reluctant to add any of them (especially since the bug is covered by 4 runtime tests). What do you think?

As long as what we check if that produces the correct sequence for the problematic case, that doesn't seem more fragile than any of the existing test cases.
I would rather have this small test than looking in runtime issues in general.


https://reviews.llvm.org/D51861