Feat: basic record dot completions #3080
Conversation
coltenwebb
force-pushed
the
completions-work
branch
2 times, most recently
from
898e577 to
5f4be06
Compare
coltenwebb
force-pushed
the
completions-work
branch
from
5f4be06 to
795ad8e
Compare
coltenwebb
force-pushed
the
completions-work
branch
2 times, most recently
from
b0ad658 to
d61dac3
Compare
coltenwebb
force-pushed
the
completions-work
branch
from
d61dac3 to
d69114b
Compare
|
@pepeiborra Would you mind doing another review? It should be good to go. |
coltenwebb
force-pushed
the
completions-work
branch
from
d69114b to
0bc2cfe
Compare
| let uses_overloaded_record_dot _ = False | ||
| #endif | ||
| ms <- fmap fst <$> useWithStaleFast GetModSummaryWithoutTimestamps npath | ||
| astres <- case ms of |
There was a problem hiding this comment.
Possibly stupid question, but why guard here? What happens if we just guard getting the HIE AST on whether or not we have record dot?
There was a problem hiding this comment.
If I understand you correctly, that's exactly what I'm trying to do here. Maybe there's a simpler way to do this?
There was a problem hiding this comment.
No, it was a stupid question: I missed that the uses_overloaded_record_dot predicate needs the module summary!
| case (pfix, completionContext) of | ||
| (Just (VFS.PosPrefixInfo _ "" _ _), Just CompletionContext { _triggerCharacter = Just "."}) | ||
| ((PosPrefixInfo _ "" _ _), Just CompletionContext { _triggerCharacter = Just "."}) |
There was a problem hiding this comment.
OOI, what is happening in this case? I guess this is basically preventing us from giving completions when people write . which usually corresponds to function composition?
There was a problem hiding this comment.
That's how I read this too. getCompletionPrefix doesn't find a prefix word before the cursor so it returns emptiness.
| , typeText = Nothing | ||
| , label = label | ||
| , isInfix = Nothing | ||
| , docs = emptySpanDoc |
There was a problem hiding this comment.
Well, fields have haddock usually, you know
data Foo {
hello :: Int -- ^ This is the doc for foo
}
I would expect to get the field haddock in the completion doc!
This might get automatically sorted out for us with #3204, not sure.
|
Looks like the new test failed on 9.0.2? |
coltenwebb
force-pushed
the
completions-work
branch
2 times, most recently
from
4231579 to
e021ae1
Compare
coltenwebb
force-pushed
the
completions-work
branch
from
e021ae1 to
6077694
Compare
|
🎉 |
|
Last thing: could you make some issues for any remaining user-facing problems, like maybe the documentation one? |
When a user dots a record type with OverloadedRecordDot enabled, shows all possible record selectors for that type.
Todo:
Notes