add completions for specific MatchType cases #14639
Conversation
| @@ -362,6 +363,12 @@ object Completion { | |||
| implicitConversionTargets(qual)(using ctx.fresh.setExploreTyperState()).flatMap(accessibleMembers) | |||
| membersFromConversion.toSeq.groupByName | |||
|
|
|||
| /** Completions for derived members of `MatchType`'s type. */ | |||
| def matchTypeCompletions(qual: Tree)(using Context): CompletionMap = | |||
| qual.tpe.widenDealias match | |||
There was a problem hiding this comment.
I think we should only dealias here but not widen
| @@ -302,6 +302,7 @@ object Completion { | |||
| def selectionCompletions(qual: Tree)(using Context): CompletionMap = | |||
| implicitConversionMemberCompletions(qual) ++ | |||
| extensionCompletions(qual) ++ | |||
| matchTypeCompletions(qual) ++ | |||
There was a problem hiding this comment.
I think this shouldn't be a separate case but we should rather ensure that match types used as method receivers are properly reduced in all other cases. We should also add test cases for them. For some reason in case of extension methods surprisingly this seemed to work even without this PR but still we should add a test to avoid regressions. We should also make sure this works well with implicit conversions.
rochala
force-pushed
the
match-type-completion
branch
from
5df04b3 to
61822d2
Compare
| .withSource.completion(m1, expected) | ||
| } | ||
|
|
||
| @Test def higherKindedMatchTypeLazyCompletion: Unit = { |
There was a problem hiding this comment.
What's the essential difference between this test case and higherKindedMatchTypeDeclaredCompletion?
There was a problem hiding this comment.
higherKindedMatchTypeDeclaredCompletion is reduced and was working properly without this fix, added this one as a regression test.
Fixes scalameta/metals#2810