Mailing-List: contact ant-dev-help@jakarta.apache.org; run by ezmlm
Message-ID: <005401bff0d0$a563ace0$e900a8c0@oficina.com.ar>
From: "Carlos Pita" <carlos@aper.net>
To: <ant-dev@jakarta.apache.org>
References: 
 <02e301bfeda6$2db8c0e0$e900a8c0@oficina.com.ar><m3puogd8js.fsf@sbodewig.bost.de><004801bfedb1$acc8d360$e900a8c0@oficina.com.ar>
 <m3n1jfk742.fsf@sbodewig.bost.de>
Subject: RE: RV: Nesting Tasks
Date: Tue, 18 Jul 2000 12:55:43 -0300
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit

    Thank you for the info.

    Regards a NestingTask subclassing or not a Task I believe that despite
some differences there is a lot (well, actually a Task hasn't a lot of
features) of commonalty. But I never was completly sure about this, so I
accept that they could be thought as not being Tasks. In fact, the idea was
rounding my mind from the first time.

  About another kind of functionality for NestingTasks: they could implement
some kind of conditional structures (switch, if, etc.) nesting "dummy" else,
elseif, case tasks; they could implement some kind of procedures, the
special set being the parameters, the nested tasks the body. I think the
proposal is very general so a lot of control tasks with very different
semantics are allowed by it (it is a powerful hook, don't you think?).

Carlos.

----- Original Message -----
From: Stefan Bodewig <bodewig@bost.de>
To: <ant-dev@jakarta.apache.org>
Sent: Tuesday, July 18, 2000 8:31 AM
Subject: Re: RV: Nesting Tasks


> >>>>> "CP" == Carlos Pita <carlos@aper.net> writes:
>
>  CP> I had a feeling about the script Task being some kind of solution
>  CP> to the problems that originated my proposal (for example jar'ing
>  CP> classes in an arbitrary list of dirs given, for example, as a -D
>  CP> parameter to ant into a number of jars, one for each dir).
>
> XSLT can be another solution to this problem when the dirs are known
> up front.
>
>  CP> But I couldn't find out anything about BSF - the ibm site's open
>  CP> source section devoted to BSF has only a web access to the CVS
>  CP> and a broken become_a_contributor link-.
>
> I now see how you want to implement your NestingTask approach, it
> doesn't actually need to be a subclass of Task as ProjectHelper would
> have to handle it differently anyway.
>
> What other kind of functionality of a NestingTask would you envision,
> apart from the foreach thingy you've described?
>
> Stefan
>