fix(library/type_context): inconsistent unification of match-based definitions #141
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
The issue was that `nat.add` didn't get unified with `λ a b, nat.add a b`, because `nat.add` is defined using matches. If such a definition is applied to a term, only when the term matches one of the match cases the whole is unfolded. In contrast, if there is no application, the constant `nat.add` is happily unfolded all the way. Thus the elaborator decides unification fails. There are basically two solutions: unfolding more (i.e. also unfold `nat.add a b`) or unfolding less (i.e. don't unfold `nat.add`). Since unfolding `nat.add` would expose implementation details, namely the `nat.add._main` function, choosing to unfold less is better. But this leads to another issue: now `λ <a, b>, a` does not unify with `λ <a, b>, a` if they are defined separately. This is not a big issue since `example : let c := 0 in (λ <a, b>, a + c) = (λ <a, b>, a + c) := rfl` didn't work anyway. Hence the changed tests.
|
@Vierkantor I don't remember: did you test this change on mathlib? Were there any issues? |
|
Ok, good. There's no point in testing right now, everything's broken. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
The issue was that
nat.adddidn't get unified withλ a b, nat.add a b, becausenat.addis defined using matches. If such a definition is applied to a term, only when the term matches one of the match cases the whole is unfolded. In contrast, if there is no application, the constantnat.addis happily unfolded all the way. Thus the elaborator decides unification fails.There are basically two solutions: unfolding more (i.e. also unfold
nat.add a b) or unfolding less (i.e. don't unfoldnat.add). Since unfoldingnat.addwould expose implementation details, namely thenat.add._mainfunction, choosing to unfold less is better.But this leads to another issue: now
λ <a, b>, adoes not unify withλ <a, b>, aif they are defined separately. This is not a big issue sinceexample : let c := 0 in (λ <a, b>, a + c) = (λ <a, b>, a + c) := rfldidn't work anyway. Hence the changed tests.