Re: roles, restricted types

[Peter Reilly]

Stefan Bodewig wrote:
I think we need that in order to decompose ant.jar.
     

If this means making multiple jars from ant.jar, I would be against.
   

Why?
Even "core tasks" contains a bunch of tasks that really should go into
a separate antlib - the cvs tasks for example.  I didn't say we'd stop
distributing them with Ant.
 

I meant that the core - conditions, filters, mappers can be in separate 
namespaces, but
that these could still be in ant.jar, one can have multiple antlib.xml 
's in ant.jar.

Of course the cvs tasks could/should be in a separate antlib.
However as a user, I do not like the current gallizion number of jars in 
$ANT_HOME/lib,
I would rather have one ant.jar containing all that tasks and types 
including optional ones.

Currently I favor a model where we have major releases of Ant's core
together with a number of antlibs but the antlibs themselves can have
releases independent of Ant.  I'd like to be able to release a new
version of the <javah> task without requiring a full Ant release.
 

Having a ant-javah.jar, ant-javac.jar ... would seem to be a bit 
excessive. - However
I assume you mean that some of the core tasks/types could be in logical 
separate
jars.

A three to six month cycle for dot releases, with a 9 month to 18 month 
cycle for
major releases does not seem to be too bad.

 

And I'd like to have shorter names, somehow. role="condition".
     

This becomes a naming problem... One would have to define the
mapping between the role and the class the type would need be an
instanceof.  This mapping may need to be unique through a number of
third party antlibs - a world of pain unless use is made of xml
namespaces.
   

role="antcore:condition" xmlns:antcore="antlib:org.apache.tools.ant"
works for me as well.  If this is what you meant.
 

Yes, however this is not easy to do at the moment with the current ant 
core code.
It (mapping of antcore to antlib:org.apache.tools.ant) can be done at 
parse time
but some infrastructure needs to be added to allow it to be done at run 
time.
[The current code does this mapping for ant-type at parse time].

My current feeling is that restricted types/roles are not necessary, one 
can use
namespaces to distingish condition <or> from selector <or>.

Peter
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]