On Thu, 28 Mar 2002, Conor MacNeill <[EMAIL PROTECTED]> wrote: > Stefan Bodewig wrote: > >> On Tue, 26 Mar 2002, Paul Kilroy <[EMAIL PROTECTED]> wrote:
>>>I think the is related to the fact that ant caches >>>targets (and their tasks), but the task is "cleaned >>>up" at the end of the execute method. This pretty much >>>invalidates the cache for the target. >>> >> Yes. > > Well, this task cleans itself up (kills itself really) but the core > does not do any general task cleanup. I think this is what Paul talks about (the task cleaning up itself), yes. >>>-Don't run the target again, it's already been run. >>> >> or alternatively, don't use cached tasks. > > These two statements amount to the same thing, in general, don't > they? No. My alternative is to run a target of that name but use fresh instances, while Paul talks about not running the target twice at all. >> We've always said that "ant foo moo" is supposed to be the same as >> "ant foo; ant moo" which it obviously is not. > > This has never really been true that these are equivalent due to > property immutability. So my alternative is not viable. Stefan -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
