I realize discussion will be a bit of a problem with mail delays of more than an hour (at least for me ATM) ...
On Thu, 12 Jul 2001, Diane Holt <[EMAIL PROTECTED]> wrote: > A couple of weeks ago, the question of being able to set a property > based on derefencing another property came up again. -0 I'd go with Nico's solution (on the report) or use external property files to solve this (my standard answer, I know 8-) - I still couldn't get convinced that we need recursive property resolution. > Also, awhile back someone wanted to be able to require a particular > version of Ant, so I wrote a little <require> task. -0 You'll probably not just set a property but also make the build fail, right? This means we'd have two different types of <fail> tasks, in a sense this is not that different from adding a the fail functionality to available (which has been vetoed). A cleaner solution would be to (1) create the condition task - not that big a problem IMHO, (2) implement your require task as a condition in this task, and optionally (3) add if/unless attributes to fail. > And finally -- what should we do about the <foreach> task? Has been vetoed by many people - even though I'd not veto it any longer, others will probably stick to their vote. Count that as a +0. Stefan
