Re: #50218 - Notification on write-only properties is questionable/problematic



Bill Haneman <Bill Haneman Sun COM> writes:

> >> I think you were in agreement that notification on write-only 
> properties
> >> was a bad idea. 
> 
> I'm a little worried about preventing this notification...
> 
> If write-only properties "result in some state change" to
> the GUI/UI, probably a change to the visuals, in some
> cases an assistive technology may need to know that
> "property X has changed" even if it's not ordinarily
> a readable property, in order to keep consistent state
> within the AT's view of the UI.
> 
> We don't know what the RO properties are offhand,
> what should we grep for? :-)  Then maybe we can
> identify when if ever this situation might arise.

The point here is: what would you do when you got
notification of a change to a writeonly property?  You can't
retrieve the new value...

The only writeonly property I know of in GTK+ at the
moment is GtkLabel::pattern, which is an alternate
method of setting the ::attributes property, which
is read-write.

(::attributes takes a PangoAttrList object; ::pattern is
much more limited and takes a string pattern with _
characters where underscores should be.)

When you set the ::pattern attribute, you get notification
that ::attributes has changed.

I expect that most write-only attributes will be like
this - an alternate way of setting a readwrite property.7

Regards,
                                        Owen




[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]