Backport "Update unreducible match types error reporting" to LTS #21024
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
[Cherry-picked 7b69d33]
This reverts commit 9ae1598 Note that the changes in Typer: ``` val unsimplifiedType = result.tpe simplify(result, pt, locked) result.tpe.stripTypeVar match case e: ErrorType if !unsimplifiedType.isErroneous => errorTree(xtree, e.msg, xtree.srcPos) case _ => result ``` cannot be reverted yet since the MatchReducer now also reduces to an `ErrorType` for MatchTypeLegacyPatterns, introduced after 9ae1598. [Cherry-picked 7460ab3][modified]
i18488.scala was only passing because of the bug in the MatchReducer,
as we can see in the subtyping trace:
```
==> isSubType TableQuery[BaseCrudRepository.this.EntityTable] <:< Query[BaseCrudRepository.this.EntityTable, E[Option]]?
==> isSubType Query[BaseCrudRepository.this.EntityTable, Extract[BaseCrudRepository.this.EntityTable]] <:<
Query[BaseCrudRepository.this.EntityTable, E[Option]] (left is approximated)?
==> isSubType E[Option] <:< Extract[BaseCrudRepository.this.EntityTable]?
==> isSubType [T[_$1]] =>> Any <:< Extract?
==> isSubType Any <:< Extract[T]?
==> isSubType Any <:< T match { case AbstractTable[t] => t } <: t (right is approximated)?
==> isSubType Any <:< <error Match type reduction failed since selector T
matches none of the cases
case AbstractTable[t] => t> (right is approximated)?
<== isSubType Any <:< <error Match type reduction failed since selector T
matches none of the cases
case AbstractTable[t] => t> (right is approximated) = true
<== isSubType Any <:< T match { case AbstractTable[t] => t } <: t (right is approximated) = true
<== isSubType Any <:< Extract[T] = true
<== isSubType [T[_$1]] =>> Any <:< Extract = true
...
<== isSubType Extract[BaseCrudRepository.this.EntityTable] <:< E[Option] = true
<== isSubType Query[BaseCrudRepository.this.EntityTable, Extract[BaseCrudRepository.this.EntityTable]] <:<
Query[BaseCrudRepository.this.EntityTable, E[Option]] (left is approximated) = true
<== isSubType TableQuery[BaseCrudRepository.this.EntityTable] <:< Query[BaseCrudRepository.this.EntityTable, E[Option]] = true
```
[Cherry-picked 5becaac]
Modify the MatchReducer to return NoType in the case of no matches, rather than throwing a MatchTypeReductionError. This makes it consistent with the other match type reduction failures, where being stuck does not result in an error, but simply in an unreduced match type. We still get the explanations of the underlying error in the MatchTypeTrace, but in positions which need the reduction for conformance, rather than at application site of the match type. [Cherry-picked d7946bf][modified]
The diff in neg/10349.scala is quite interesting. With a few intermediate values: ```scala type First[X] = X match case Map[_, v] => First[Option[v]] def first[X](x: X): First[X] = x match case x: Map[k, v] => val hdOpt: Option[v] = x.values.headOption first(hdOpt): First[Option[v]] // error only before changes ``` This now type-checks but will fail at runtime because of the in-exhaustivity of the match expression. Perhaps we should add some additional condition in `isMatchTypeShaped` to account for this, or at least emit a warning ? [Cherry-picked 61c2832]
`recurseWith` can be called with the same scrutinee (even if match type reduction is cached) if it is an applied match alias For example, `Tuple.Head[Tuple.Tail[T]]` will attempt to reduce `Tuple.Tail[T]` twice: - once as an argument of the match alias `Head`, and - once as a scrutinee in body of `Head` (after the substitution). [Cherry-picked 2329236][modified]
If a match type pattern is an opaque type, we use its bounds when checking the validity of the pattern. Following the ElimOpaque phase however, the pattern is beta-reduced (as normal applied type aliases), which may result in an illegal pattern. [Cherry-picked 2beb67e][modified]
|
Can't be backported to the LTS - courses regression in multiple compilation tests, probably due to missing dependant PRs backports |
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.
Backports #19954 to the LTS branch.
PR submitted by the release tooling.
[skip ci]