first of all, thanks stefan for triggering the discussion.  I very much agree with you Dominique.  Where did you post that code you were talking about.  I wasn't sure how to limit dependencies on an external jar but found out very quickly I wanted that too after experiencing it in one source tree(I thought it was great and wanted to extend it further).  Perfect example is a common.jar which my project uses and I only want certain components depending on certain packages of the common.jar.

I kind of gave up on people responding so I have some performance fixes for the task and some other fix.  I didn't right the test cases yet.  Let me do that, and then I will submit them.  thanks for posting this task.  I really appreciate it.

ps.  I am at a conference next week, and then on vacation two, so you may not hear from me for a while.
thanks again, I hope to post fixes for this task,
Dean

Dominique Devienne wrote:
+1. I don't like the name too much though... Any other suggestions?

Once it's in, it will be easier to make it evolve. The XSL-based system I
currently have, and would very much like to replace with something like this
task, reads the definition of what the project sub-parts are, and which
dependencies they each have, between each other, but also with external
JARs. I enable part of this functionality thru a <buildpath> element that
extends Path, and allows to plug-in a custom BuildPathResolver to determine
the build order. (I've posted that code).

The ability to compile a project's subpart independently is great, but the
exact dependencies between these parts, and for each part of external JARs
is something that I (and I suspect other) really need.

My $0.02. --DD

  
-----Original Message-----
From: Stefan Bodewig [mailto:bodewig@apache.org]
Sent: Thursday, September 04, 2003 10:11 AM
To: ant-contrib-developers@lists.sourceforge.net
Subject: Re: [Ant-contrib-developers] new ant task

I was waiting for anybody else to comment on Dean's compilewithwalls
task.  I really like the idea and would like to see it somewhere.

Ant is more or less in a no-new-tasks-accepted state and I don't
expect it to change until after the release of Ant 1.6.  Maybe we'll
see multiple antlibs as subprojects of Ant after that, who knows.

Dean's task is really useful, well documented and tested IMHO and I
see it as a good fit for ant-contrib.  What do others think?

Stefan
    


-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
Ant-contrib-developers mailing list
Ant-contrib-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ant-contrib-developers