Micha Reiser
Micha Reiser
> I'm confused about the unused-type-ignore-comment rule. Is this intended to be a new rule, separate from our existing unused-ignore-comment? If so, then a) I don't think we need this...
Let's start with the option, and split the rule based on user feedback.
I started to work on this.
Thank you. With not working, I assume you mean ty doesn't provide any LSP features within a notebook? Yes, I think this is the expected result of their "mitigation" mentioned...
We can also see this in profiles where most threads finish after 180ms, but then there's one thread that just keeps going https://share.firefox.dev/3KAw91E
We see the same pattern on macos and linux. It's more notable on my Linux machine because it has more cores, but each core has worse single-core performance compared to...
One thing that stands out to me is that we call `is_redundant_with` a lot... Like a ton: - Total queries executed: 56,481 - is_redundant_with queries: 42,628 - Percentage: ~75.5% (42,628...
> Ideally, we would avoid this scenario. One possibility is that semantic indexing collects dependency modules and names depended on from those modules, and these also become part of a...
Another thing I played with is spawning a task per scope in `check_types`. This helps a lot for projects like discord.py where a single file takes very long to type...
I assigned this to the Stable milestone for now but I think we should prioritize this based on user feedback. Most clients support pull diagnostics and/or workspace diagnostics, in which...