Peter Vogel wrote: > Basically an alteration of the patch I sent earlier for > expansion of properties in attributes of the target task.
And I rejected this patch because it lets you expand properties in *all* attributes of a target. I haven't looked at your patch specifically, so I'm not going to try to comment on it; I'm only willing to look like an idiot once a day. :-) > One last point in defense of my earlier patch: > > Ant is swarming (like an anthill :-)) in conceptual inconsistencies, > IMHO, one of these inconsistencies is the failure to evaluate > properties for attributes of some elements and not others. I would > hope, even if my patch isn't committed, that someone would take > the time to iron these out, BEFORE ant 2.0 I agree that there are a lot of inconsistencies in the way Ant works. Gotta love backwards compatibility. :-) You are absolutely correct, we need to iron out when its appropriate to evaluate properties and be consistent about it. The rub seems to be that there is a lot of disagreement as to exactly when properties need to be evaluated. The evaluation is often requested to implement a "missing feature" of ant. A lot of what people currently consider to be missing will be introduced in Ant 2 (if we can ever stop arguing about it), including mutable properties, a <condition> task, and cross project DAGs. Glenn McAllister SOMA Networks, Inc.
