permit inference variables to be unified with unnormalized projections #305
Conversation
This file contains hidden or 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
This test dies with an overflow. With CHALK_DEBUG=1 this death is quick, without it, it is slow and painful. The problem is that some of the types in the environment are "truncated" and replaced with inference variables, but those types are unnormalized projections. When we then go to unify the environment found in the answer, we end up needing to unify inference variables with unnormalized projections -- but the current code creates subgoals instead, which leads to an infinite recursion.
I have a dim memory that this violates some invariant I wanted to maintain, but I can't right now be sure that that is. I suppose it means that answers can contain unnormalized projections, in theory, which we might want to prevent. The only time this is needed is when equating answers after truncation. It's a rather special case and we might be able to do a "smaller" fix.
|
So, what's interesting about this approach is how it might overlap with the solution to the associated types double answer (#234). Like, the change to the other test output is interesting. |
|
Decided to close for now. Opened #306 to track my concerns with the current workaround. |
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Outdated suggestions cannot be applied.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.
This fixes the rust-analyzer problems found in #280 and serves as an alternative to #294. I'm not really sure which approach I prefer -- there might even be a "third way" that is better still.
If nothing else, though, I think that the regression tests found on this branch is the correct form. I had to make a few small fixes to the test macro so that we could select the solver and specify the "max size" correctly.