Refine implicit search fallbacks for better ClassTag handling #23532
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
It's not necessary to instantiate all type variables, deeply, since we are only interested in the outermost shape. Also, that way we do not instantiate to Nothing, which is something difficult to recover from.
We used to look at deep expected type when the result of an implicit was ambiguous. This could add more constraints which could resolve the ambiguity. We now do the same also if the search type has uninstantiated type variables. In that case, consulting more context might further constrain type variables, which might in turn enlarge the implicit scope so that a solution can be found. Fixes scala#23526
odersky
changed the title
lrytz
approved these changes
Jul 17, 2025
This was referenced Jul 27, 2025
odersky
added a commit
to dotty-staging/dotty
that referenced
this pull request
Aug 5, 2025
scala#23609 triggers an assertion error in TyperState. The relevant explanation seems to be in ProtoTypes.scala: ```scala // To respect the pre-condition of `mergeConstraintWith` and keep // `protoTyperState` committable we must ensure that it does not // contain any type variable which don't already exist in the passed // TyperState. This is achieved by instantiating any such type // variable. NOTE: this does not suffice to discard type variables // in ancestors of `protoTyperState`, if this situation ever // comes up, an assertion in TyperState will trigger and this code // will need to be generalized. ``` We should go to the bottom of it and fix the assertion. But before that's done this PR offers a temporary hack to catch the exception when it is triggered from a new code path created by PR scala#23532. This should fix the regression reported in scala#23609. We should leave the issue open as a reminder that we still need a better fix. Also: handle crash to missing span in a migration helper.
WojciechMazur
pushed a commit
that referenced
this pull request
Aug 5, 2025
#23609 triggers an assertion error in TyperState. The relevant explanation seems to be in ProtoTypes.scala: ```scala // To respect the pre-condition of `mergeConstraintWith` and keep // `protoTyperState` committable we must ensure that it does not // contain any type variable which don't already exist in the passed // TyperState. This is achieved by instantiating any such type // variable. NOTE: this does not suffice to discard type variables // in ancestors of `protoTyperState`, if this situation ever // comes up, an assertion in TyperState will trigger and this code // will need to be generalized. ``` We should go to the bottom of it and fix the assertion. But before that's done this PR offers a temporary hack to catch the exception when it is triggered from a new code path created by PR #23532. This should fix the regression reported in #23609. We should leave the issue open as a reminder that we still need a better fix. Also: handle crash due to missing span in a migration helper.
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.
It's not necessary to instantiate all type variables, deeply, since we are only interested in the outermost shape. Also, that way we do not instantiate to Nothing, which is something difficult to recover from. This step is necessary to enable (2)
We used to look at the deep expected type when the result of an implicit was ambiguous.
This could add more constraints which could resolve the ambiguity.
We now do the same also if the search type has uninstantiated type variables. In that
case, consulting more context might further constrain type variables, which might in
turn enlarge the implicit scope so that a solution can be found.
Fixes #23526