Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We鈥檒l occasionally send you account related emails.

Already on GitHub? Sign in to your account

Give a canonical ordering for destructing terms in Wingman #1586

Merged
merged 4 commits into from
Mar 17, 2021

Conversation

isovector
Copy link
Collaborator

Today, Wingman tries to destruct terms in any order. For example, when trying to synthesize zip, it will try the following solutions:

\ l_a l_b
  -> case l_b of
       []
         -> case l_a of
              [] -> _
              (a : l_a3) -> _
       (b : l_b3)
         -> case l_a of
              [] -> _
              (a : l_a5) -> _

and

\ l_a l_b
  -> case l_a of
       []
         -> case l_b of
              [] -> _
              (b : l_b3) -> _
       (a : l_a3)
         -> case l_b of
              [] -> _
              (b : l_b5) -> _

These two solutions are equal modulo the order in which we did a pattern match. But it's stupid to explore both paths, since they cannot produce different answers.

Instead, this PR introduces a canonical ordering on the destructing on terms, using the natural ordering of introduction in the hypothesis. In particular, Wingman will not try to destruct anything that was introduced before something that has already been destructed.

This change has a remarkable O(n!) asymptotic increase at each hole 馃ぉ The search is still remains stupid, but at least it's no longer factorially stupid.

@isovector isovector added the merge me Label to trigger pull request merge label Mar 17, 2021
@mergify mergify bot merged commit a34d927 into haskell:master Mar 17, 2021
@isovector isovector deleted the orderly-destruct branch April 5, 2021 11:48
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
merge me Label to trigger pull request merge
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

2 participants