IDE Minutes 2008 11-18
- Jon Ferraiolo, IBM
- Rich Thompson, IBM
- Nitin Dahyabhai, IBM
- Lori Hylan-Cho, Aptana
- Javier Pedemonte, IBM
- Bertrand Le Roy, Microsoft
Topic: Some advanced Dojo cases
- Jon: http://openajax.org/pipermail/ide/2008q4/000885.html
- Bertrand: http://openajax.org/pipermail/ide/2008q4/000886.html
- Jon: http://openajax.org/pipermail/ide/2008q4/000892.html.
Jon: I believe we should address the two use cases that Dojo has that we aren't supporting yet.
Bertrand: Singleton case is really common
Jon: How about the single property mixin case?
Bertrand: Not very common. Can accomplish by making a mixin out of a property.
Jon: I'm in favor of addressing the mixin case because if you have a foo class and a bar class, the foo developer cannot know what the bar class wants to mixin one of its properties.
Bertrand: Yes, sure.
Rich: If trying to accomplish finer-grain control, is there a more general mechanism?
Jon/Rich: One at a time is likely to be more straightforward.
Rich: I can buy this.
Jon: One proposal is 'fromProperty' and 'toProperty' attributes, where 'toProperty' defaults to 'fromProperty'. How does this sound?
RESOLUTION: Add 'fromProperty' and 'toProperty' attributes to <mix>
(discussion of singleton classes)
Jon: Three options come to mind:
- (a) <object> element as a sibling to <class>
- (b) add static="true|false" to <class>
- (c) <staticClass> element as a sibling to <class>
Lori: ScriptDoc has (b)
Rich: I prefer (b) since static class uses all of the same things as an instantiated class
RESOLUTION: Add 'static' attribute to <class>
ACTION: Jon to send email to Tom Trenka
Topic: Spec and schema changes for <config>
Jon: If anyone see anything wrong with the changes I made, please speak up.
Topic: Please review new spec sections on <handlerFunction>
Jon: Is the proposed write-up OK?
Rich: Does it need to be a pattern?
Bertrand: Yes, you want to say it once and apply it to random properties
Jon: It's an inherited property, like our getters and setters.
Bertrand: Also need a pattern for unregister
Jon: OK, I will add that
Jon: OK that Template:Callback is a function?
Bertrand: That's fine
Rich: Need to say crisply that function parameters are required and that define the signature of the callback.
RESOLUTION: Add 'unregisterCallbackPattern' attribute
Topic: Cleanups to the write-up on 'datatype'
- Jon: http://openajax.org/pipermail/ide/2008q4/000889.html
- Bertrand: http://openajax.org/pipermail/ide/2008q4/000890.html
- Jon: http://openajax.org/pipermail/ide/2008q4/000891.html.
- Bertrand said:
I can see datatype="[Object]|String" being useful: the method here could take a comma-separated list that it knows how to parse to an array.
RESOLUTION: We will support "[Object]|String"