ant-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Peter Donald <dona...@apache.org>
Subject Re: Ant 1.5... loading tasks from jars?
Date Mon, 08 Oct 2001 08:54:42 GMT
On Mon, 8 Oct 2001 14:39, Steve Loughran wrote:
> ----- Original Message -----
> From: "Tim Dawson" <tdawson@wamnet.com>
> To: <ant-dev@jakarta.apache.org>
> Sent: Friday, October 05, 2001 9:09 AM
> Subject: Ant 1.5... loading tasks from jars?
>
> > I made the suggestion back in May to allow loading tasks from jars (a
> > tasklib, if you will), and I know it is currently listed in the 2.0
> > requested features, but it appears to be possible today...
> >
> > <taskdef resource="meta-inf/ant.tdf">
> >   <classpath>
> >     <pathelement location="mylib/mytasks.jar"/>
> >   </classpath>
> > </taskdef>
> >
> > Of course, this is a completely non-obvious solution, which is why its
>
> still
>
> > a requested feature. :-)
>
> sounds like a documentation bug to me; nobody has updated the developing
> for ant page for a while.

Probably good that it is not documented ;) It is an ugly hack IMHO 

> Maybe we could code up some very simple xml DTD for deploying ant1.x tasks,
> which includes a version identifier so future ant versions can recognise
> the version of the descriptor and react accordingly.
>
> <antdefs version="1.5">
> <taskdefs>
>  <taskdef name="..." class="..."/>
> </taskdefs>
> <typedefs/>

I would prefer something like

<antlib version="...">
  <task name="..." class="..."/>

  <datatype name="..." class="..."/>
</antlib>

"taskdef" is a task and has oodles of other cruft that should not be 
specified in descriptor. And I don't see any need to group the task type 
because in the future we will have oodles of types, most of which are 
dynamically registerable. ATM we can only have tasks, datatypes (and 
conditions?) but in Ant2 this will definetly not be the case.

> Does this seem like a good idea?

If you can get agreement. Considering the loons that came out last time such 
issues were raised I guess I am sceptical ;)

-- 
Cheers,

Pete

--------------------------------------------------
 Logic: The art of being wrong with confidence...
--------------------------------------------------


Mime
View raw message