On Sun, 21 Dec 2025 22:04:58 GMT, Michael Strauß <[email protected]> wrote:

> > I do however have a bit of problem with `isAttached`; what does this have 
> > to do with properties in general? The description doesn't clarify what it 
> > means really, or what it even means outside the very specific case where 
> > HBox can "donate" properties to its children.
> > So to me, the attached part feels out of place for a low-level property 
> > API, unless there are compelling reasons and other use cases we'd envision 
> > outside of having CSS stylable child properties. If not, I don't see why we 
> > shouldn't make this a method on `Stylable` or something more Node/CSS 
> > specific like in this PR: #1714
> 
> I think attached-ness is a fundamental aspect of properties, and making it 
> visible in the type system is a logical choice. Compare this to WPF or 
> Avalonia, where attached properties are also explicitly modelled. It's a 
> piece of information that you can't reliably get from a property's metadata 
> otherwise.

I never used either framework, so "attached" is completely meaningless to me, 
and the description didn't enlighten me.  What would this mean when I have just 
a bunch of properties (like a model)?  It seems to me it is giving some extra 
information related to some kind of parent/child relationship that's outside 
the definition of the properties themselves.

I'm only looking for a use case where this would be useful if I only have 
imported `javafx.base` to better understand it, and how it may benefit 
properties in general.

I did some checking, and this was suggested by AI:
- A `Task` is a simple model with a permanent property (name).
- A `Workflow` defines an attached property `activeInWorkflow`, which exists 
only while the task is in the workflow.
- `isAttached()` signals that this property is context-dependent, not intrinsic 
to `Task`.
- Listeners or code can use `isAttached()` to safely determine whether the 
property is meaningful.

-------------

PR Comment: https://git.openjdk.org/jfx/pull/2015#issuecomment-3679626131

Reply via email to