Rewrite progress handling to allow for debouncing messages #571
Conversation
|
@soulomoon what do you think? I think the |
|
I also need to actually test this with a live HLS and see if I can actually notice the difference! |
michaelpj
force-pushed
the
mpj/progress-threads
branch
from
c228abb to
afedfb1
Compare
This had to be redone in order to allow us to "wake up" and notice that there are pending messages. I also wrote it so there can be a stateful interface (the `ProgressTracker`) which I think might make it easier to use in that weird case in `ghcide`. I haven't exposed that yet, though.
michaelpj
force-pushed
the
mpj/progress-threads
branch
from
afedfb1 to
b81f25c
Compare
| @@ -138,6 +140,10 @@ data LanguageContextEnv config = LanguageContextEnv | |||
| resState :: !(LanguageContextState config) | |||
| , resClientCapabilities :: !L.ClientCapabilities | |||
| , resRootPath :: !(Maybe FilePath) | |||
| , resProgressStartDelay :: Int | |||
There was a problem hiding this comment.
Maybe this can be set for each progress instead of globally? Since we might want different delay for different progress
There was a problem hiding this comment.
I can't actually think of a use case for this. I think it should be pretty generic?
There was a problem hiding this comment.
I'm inclined to leave it as global for now unless we have a need for it to be different.
This piece of code in HLS is quite delicate, the previous version did not work well. |
michaelpj
force-pushed
the
mpj/progress-threads
branch
2 times, most recently
from
0a6728c to
e65065a
Compare
michaelpj
force-pushed
the
mpj/progress-threads
branch
from
e65065a to
1b9cc1d
Compare
|
Turns out the stateful interface wasn't really that helpful and complicated things, I think haskell/haskell-language-server#4218 is actually perfectly fine with the normal interface, arguably even clearer. |
michaelpj
force-pushed
the
mpj/progress-threads
branch
from
b05ee42 to
e4ab8c5
Compare
michaelpj
force-pushed
the
mpj/progress-threads
branch
from
4f63f44 to
f45f29c
Compare
This had to be redone in order to allow us to "wake up" and notice that there are pending messages. I also wrote it so there can be a stateful interface (the
ProgressTracker) which I think might make it easier to use in that weird case inghcide. I haven't exposed that yet, though.