Support completionItem/resolve to compute documentation of a code completion item #1938
Conversation
ahoppen
changed the title
ahoppen
force-pushed
the
completion-item-resolve
branch
from
bf7c123 to
dcf2b92
Compare
|
@swift-ci Please test |
| case is CompletionItemResolveRequest: | ||
| self = .freestanding |
There was a problem hiding this comment.
This is a request, not a notification, right? Also shouldn't it be considered a document request?
There was a problem hiding this comment.
Oh, good catch. Also just noticed that we do have it in the initializers for requests already.
Also shouldn't it be considered a document request?
I don’t think so. It’s not necessary to process any pending edits to resolve a code completion item. Either the session is still valid, in which can we can resolve independently of document edits or not, in which case we will error in either case.
| @@ -134,7 +158,6 @@ class CodeCompletionSession { | |||
| session.snapshot.uri == snapshot.uri | |||
| && session.position == completionPosition | |||
| && session.compileCommand == compileCommand | |||
| && session.clientSupportsSnippets == clientSupportsSnippets | |||
There was a problem hiding this comment.
Do we need to check whether the capabilities match, or does that not matter?
There was a problem hiding this comment.
I don’t think that we need to. The only case where this would change is if the client initiated a code completion session and then dynamically registered the snippet support capability, which would be very unlikely since any of that dynamic registration would happen all up-front. And if should get it wrong, it’s not going to be a fatal error, the results might just look a little off (eg. no snippets even though client supports them).
ahoppen
force-pushed
the
completion-item-resolve
branch
from
dcf2b92 to
f6b83db
Compare
|
@swift-ci Please test |
|
@swift-ci Please test Windows |
Fixes #1935