Commit 2025-07-15 08:10 42d7516d
View on Github →feat(Aesop): Improve SetLike ruleset (#25961)
- Make some Aesop rules in the SetLike ruleset unsafe
- Register new rules with the Aesop SetLike ruleset, by analogy with existing ones
- Add new rules like
mem_closure_of_mem
by analogy with #26162 to improve performance in proving membership of closures. Before this PR, to prove something likex ∈ s ⊢ x ∈ closure s
, Aesop would have to try the rulemem_of_subset
, creating a metavariable that it would then try to fill withsubset_closure
. Obviously this was pretty inefficient, and the new lemmamem_closure_of_mem
is a much more usable Aesop rule.
- Add new rules like
- Update some Aesop rules' probabilities
- Add new library note [SetLike Aesop ruleset] explaining the principles behind the new design
The design note stipulates that
safe
-phase Aesop rules are to also besimp
rules; this change has been split into #26175