Mailing-List: contact ant-user-help@jakarta.apache.org; run by ezmlm
Precedence: bulk
Reply-To: ant-user@jakarta.apache.org
Content-Type: text/plain;
  charset="iso-8859-1"
From: Peter Donald <donaldp@apache.org>
To: ant-user@jakarta.apache.org
Subject: Re: for-each (another proposal)
Date: Fri, 22 Jun 2001 04:14:08 +1000
References: <001001c0fa7a$c46b62e0$0201a8c0@knightsAPC.knightsofthenet.com>
 <3B32361D.78027DA5@healthlanguage.com>
In-Reply-To: <3B32361D.78027DA5@healthlanguage.com>
MIME-Version: 1.0
Message-Id: <01062204140800.02060@helm.realityforge.org>
Content-Transfer-Encoding: 8bit

On Fri, 22 Jun 2001 03:59, Larry V. Streepy, Jr. wrote:
> Brett Knights wrote:
> > Wouldn't it be nice to use a fileset to produce a comma separated list of
> > filenames needed by the current javadoc thread?
>
> Actually, I was just thinking the same thing (having been in the middle
> of that thread)! :-)
>
> A good start would be to allow filesets to have referencable ids.  Then
> the property task could take the resulting file list and stuff it into a
> property, hopefully with a way to specify a delimeter.  This has
> application well beyond javadoc.
>
> Just an thought I had this morning.

Thats something that will probably be in Ant2. We used the term PathConvert 
on antdev to describe it. This would allow you to convert it to all different 
conventions: comma deliminated, colon deliminated for unix or unix layers (ie 
cygwin), semi-colon deliminated for win32, space separated for some tasks etc.

No ones actually implemented it yet though - I am sure if someone was to 
offer a similar task for ant1 there would not be any complaints ;)

Cheers,

Pete

*-----------------------------------------------------*
| "Faced with the choice between changing one's mind, |
| and proving that there is no need to do so - almost |
| everyone gets busy on the proof."                   |
|              - John Kenneth Galbraith               |
*-----------------------------------------------------*