Philipp Rudiger
Philipp Rudiger
I fully agree, and strongly regret adding `watch` to `bind`. The main problem we face though, is that deprecating it is quite difficult and adding a diverging version of `bind`...
This new implementation allows providing hooks that convert arguments to dependencies so you could do something like this: ```python def transform_pane_obj_dep(obj): return obj.param.object if isinstance(obj, PaneBase) else obj param.parameterized.register_depends_transform(transform_pane_obj_dep) ```...
That's the whole point of this PR, to have one version of depends and bind with the same behavior. I'm just saying that while param defines the ability to define...
Got you, I was referring to this part of your post: > why cant you use a Pane (or even Layout) as input to pn.bind/ pn.depends in the same way...
@MarcSkovMadsen, @maximlt and @jbednar I've now added a description of the full public API surface of reactive. Please consider the naming and the potential for those names to clash with...
That part is doable but the last time @jbednar discussed this the problem was that the semantics of depends and bind are still quite different.
>Could we maybe put them all into a namespace like .rv so there is only one point of conflict, so that people can detect when we're using one of our...
In a meeting just now we agreed that the "terminating methods" that the original `hvplot.interactive` implementation used do not really make sense in a Param context and we really want...
Based on our discussion the other day I've moved all custom methods that were originally on `reactive` itself on a new namespace called `rx` which is short for "**R**eactive e**X**pression"....
@jbednar @maximlt and anyone else who is interested, please review.