mathlib4
mathlib4 copied to clipboard
feat: generalize_proofs tactic
trafficstars
There are still a couple of things to do to please the linters, but I'd be glad to receive feedback on this PR at this stage
There are still some TODO items for this tactic remaining, but I think it's fine to merge now.
- It does not handle hypotheses that depend on binders correctly, it gives a
result is not type correcterror. (A commented-out test displays this issue.) It should forall-quantify the hypotheses to lift them out of whatever context they appear in. - It should reuse hypotheses in the context when available rather than introducing a new hypothesis.
bors merge