Maybe "updatedWhen" would work, although I still like "activeWhen" or
simply "when" better. The problem that has been raised about
"updateWhen" is that it isn't really the right verb tense. What we want
is a binding that is updated (or active) when the condition evaluates to
true. The value itself isn't necessarily updated unless the source of
the binding has been updated, and the name "updateWhen" might imply that
it is.
John: of the various choices, which one(s) do you like best?
-- Kevin
On 11/21/2022 3:18 PM, Nir Lisker wrote:
My proposal in the PR was 'updateWhen', which I prefer over
observedWhen and activeWhen. Just 'when' is also fine by me and
prefered over 'when'.
My only problem with 'when'/'whenever' is that they don't say what
happens "when"/"whenever". However, since these are bindings, and what
bindings do is update a bound value based on the binding, it's rather
hinted what happens.
On Tue, Nov 22, 2022 at 1:03 AM Kevin Rushforth
<kevin.rushfo...@oracle.com> wrote:
My initial reaction is that I like the name "activeWhen" at least as
well as any of the alternatives discussed so far. It's less wordy
than
"observedWhen" (which I suggested), and probably easier to
describe. I
don't really care for using the term "scope".
I also think "when" or "whenever" are acceptable, but I know some
don't
like them.
I think the leading candidates are:
activeWhen
whenever
when
observedWhen
Unless someone can come up with a better name that can be easily
described, I recommend picking one of these.
-- Kevin
On 11/21/2022 2:36 PM, Michael Strauß wrote:
> Thanks for your clarifications.
> Maybe the actual problem is that we don't have a good name for "gets
> the current value, but doesn't subscribe to updates".
> We could call bindings "active" when changes of the source value are
> processed, and "inactive" if the binding exists, but doesn't process
> changes.
> With a documented definition of "active", the method could simply be
> named `activeWhen`.
>
>
> On Mon, Nov 21, 2022 at 10:57 PM John Hendrikx
<john.hendr...@gmail.com> wrote:
>> Hi Michael,
>>
>> Thanks for your suggestion.
>>
>> The effect is not quite what you describe however, as the
initial value
>> when the operation is first invoked is retained. It's true
however that
>> when the condition is always `false` that the value will be a
constant,
>> and that when it is always `true` it effectively is just a
duplicate of
>> the left hand observable. Let me illustrate:
>>
>> public static void main(String[] args) {
>> StringProperty sp = new SimpleStringProperty("foo");
>> BooleanProperty active = new
SimpleBooleanProperty(false); //
>> inactive
>> ObservableValue<String> x = sp.when(active); // holds
"foo"
>> despite being inactive
>>
>> System.out.println(x.getValue()); // prints "foo"
>>
>> sp.set("bar");
>>
>> System.out.println(x.getValue()); // still prints "foo"
>>
>> active.set(true);
>>
>> System.out.println(x.getValue()); // prints "bar"
>> }
>>
>> This behavior doesn't violate the rule that the new binding
shouldn't
>> observe its source when the condition is false as no listener was
>> involved to get the initial value. The initial value is
important as
>> all bindings must have some kind of value. The docs do describe
this in
>> the first sentence:
>>
>> "Returns an {@code ObservableValue} that holds this value and
is updated
>> only when {@code condition} holds {@code true}"
>>
>> I think `withScope` could work (or `scopedTo`) but not sure if
"scope"
>> itself is a good description -- we'd need to update the
description to
>> use the word scope in a way that makes clear what it does
preferably
>> without having to resort to "updated only when" or "only
observes when":
>> ie: "Returns an ObservableValue that is scoped to the given
condition" ?
>>
>> --John
