Other point: we have an implementation in the sandbox
https://svn.apache.org/repos/asf/ant/sandbox/parallelexecutor

Maybe you could use some code ;-)


Jan 

>-----Ursprüngliche Nachricht-----
>Von: Dominique Devienne [mailto:ddevie...@gmail.com] 
>Gesendet: Donnerstag, 24. Juni 2010 23:12
>An: Ant Developers List
>Betreff: Re: [Proposal] Capture attributes in unknown namespaces
>
>On Thu, Jun 24, 2010 at 2:50 PM, Danny Yates 
><da...@codeaholics.org> wrote:
>> What would be kind of cool would be that if the parser 
>encounters attributes
>> in a namespace that it doesn't recognise, then instead of 
>ignoring them (as
>> it does now), it records them and makes them available 
>through an API on the
>> Project and Target objects. This would allow the executor to 
>inspect them.
>
>Something related to that was discussed in the past to allow arbitrary
>"cross-cutting" attributes to be added to tasks which wouldn't
>normally support them, typically in the context of adding global
>ant:if and ant:unless attributes (or maybe I just thought about it,
>I'm no longer sure ;)
>
>> I realise this is very specific to my parallel executor 
>project, but I think
>> adding it would be a non-breaking change that wouldn't have 
>any impact on
>> existing consumers of the API.
>
>True. But I'd be more in favor of an opt-in framework, where you
>explicitly tell Ant to record attributes from specific namespaces.
>From the list, I think other people also use "extra" markup in their
>builds for "external" tools, so in their case they don't need to
>record those attributes for examples.
>
>> What do you folks think?
>
>I think it's a good idea. There are design considerations like, should
>it be free form string=string recorded as-is, or should these
>attributes be processed like ordinary attributes Ant deals with to do
>property expansion and conversion to Java types (like boolean
>accepting "true"/"on"/"yes" as a true). For example, imagine you group
>your extra attributes into a DataType part of an AntLib bound to that
>extra namespace, the DataType could be instantiated "as usual" by Ant
>provided it looked at the UnknownElement specifically for a given
>namespace, and then the DataType is "bound" somehow to the task or
>type or target. That way everything is typed as usual, and could be
>made to reuse a lot of the existing code/facilities. See what I mean?
>--DD
>
>---------------------------------------------------------------------
>To unsubscribe, e-mail: dev-unsubscr...@ant.apache.org
>For additional commands, e-mail: dev-h...@ant.apache.org
>
>

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@ant.apache.org
For additional commands, e-mail: dev-h...@ant.apache.org

Reply via email to