Philipp Rudiger

Results 1206 comments of Philipp Rudiger

Yeah, this is awkward to fix. The only way to get all instances would be to keep (weak) references to all instances of a class or iterating over all objects...

Thanks, the test should probably just use an `int` then.

Seems to be to do with over-eagerly invalidating the root.

I think the only way forward we have here is to move to keyword only. There's no reasonable route to standardizing on consistently making the `default` argument the first positional...

I don't think it's ready, I want to revisit this as part of a more general plan for composite types as well as dataclass and pydantic integrations.

Actually this proposal is probably not sufficient for my purpose, what I really would need is a `preempt` keyword which basically says "execute me, cancel any existing events and process...

I like not having to import it though :/

No, I think the plan is for us to release a last patch or minor release for 1.x pretty soon and then create a `branch-1.x` which will be used for...

I like the idea of allowing the keyword specification to tell it to pass the parameters in explicitly but I absolutely don't follow the point of the react thing. Why...

So currently I'm a strong -1 on `react=True` and `@param.react()` but may be missing something.