On Tue, 28 Nov 2006, Matt Benson <[EMAIL PROTECTED]> wrote:
> manipulations, for 1.7.0. Are there still objections
> before I do this with lazy consensus?
like Peter I recall we decided against oing this for 1.7.0.
Stefan
-
To
On 11/28/06, Matt Benson <[EMAIL PROTECTED]> wrote:
--- Peter Reilly <[EMAIL PROTECTED]> wrote:
> On 10/14/06, Matt Benson <[EMAIL PROTECTED]>
> wrote:
> > --- Peter Reilly <[EMAIL PROTECTED]>
> wrote:
> > the uber-antlib could be called ant-contrib!,
> > > I think that this task could belong t
--- Peter Reilly <[EMAIL PROTECTED]> wrote:
> On 10/14/06, Matt Benson <[EMAIL PROTECTED]>
> wrote:
> > --- Peter Reilly <[EMAIL PROTECTED]>
> wrote:
> > the uber-antlib could be called ant-contrib!,
> > > I think that this task could belong to
> ant-contrib,
> > > along with some of Rainer's oth
On 10/14/06, Matt Benson <[EMAIL PROTECTED]> wrote:
--- Peter Reilly <[EMAIL PROTECTED]> wrote:
> the uber-antlib could be called ant-contrib!,
> I think that this task could belong to ant-contrib,
> along with some of Rainer's other classloader code.
Are you volunteering to put it in ac?
I
--- Peter Reilly <[EMAIL PROTECTED]> wrote:
> On 10/13/06, Matt Benson <[EMAIL PROTECTED]>
> wrote:
> > As promised, I reworked Peter's task to use
> > ResourceCollection stuff better, made a stupid
> error
> > when writing my tests, and forgot to come back to
> it
> > until now. The thing works
On 10/13/06, Matt Benson <[EMAIL PROTECTED]> wrote:
As promised, I reworked Peter's task to use
ResourceCollection stuff better, made a stupid error
when writing my tests, and forgot to come back to it
until now. The thing works great. Pasting inline...
are we still thinking it's possible to pu
As promised, I reworked Peter's task to use
ResourceCollection stuff better, made a stupid error
when writing my tests, and forgot to come back to it
until now. The thing works great. Pasting inline...
are we still thinking it's possible to put this in 1.7
? Or this thing is self-contained enoug
2006 18:57:04 -0500
Von: "Dominique Devienne" <[EMAIL PROTECTED]>
An: "Ant Developers List"
Betreff: Re: classloader for 1.7
> > I had a problem with a custom task which I had written which was using
> JNDI. This task had a runtime dependency on a JNDI driver f
I had a problem with a custom task which I had written which was using JNDI.
This task had a runtime dependency on a JNDI driver for MQ Series. Adding the
JNDI driver to the classpath of the taskdef never worked, the only thing that
worked was to start ant with a -lib fullpathtoMQJNDI.jar. I as
--- Peter Reilly <[EMAIL PROTECTED]> wrote:
> There is the task:
Thanks!
> I am not too sure if I am using the new
> ResourceCollection stuff
> correctly!
>
I don't think the URLElement is necessary, as Ant 1.7
has URLResource mapped to . There may also be a
problem with the reference handl
hricht
Datum: Tue, 12 Sep 2006 21:24:30 +0100
Von: "Peter Reilly" <[EMAIL PROTECTED]>
An: "Ant Developers List"
Betreff: Re: classloader for 1.7
> On 9/12/06, Antoine Levy-Lambert <[EMAIL PROTECTED]> wrote:
> >
> > Hi Peter,
> >
> >
t;Ant Developers List"
Betreff: Re: classloader for 1.7
> On 9/12/06, Antoine Levy-Lambert <[EMAIL PROTECTED]> wrote:
> >
> > Hi Peter,
> >
> > I would add another question : does this appendcomponentpath task allow
> to
> > add stuff to the classpath where
ntoine
Original-Nachricht
Datum: Tue, 12 Sep 2006 07:29:37 -0700 (PDT)
Von: Matt Benson <[EMAIL PROTECTED]>
An: Ant Developers List
Betreff: Re: classloader for 1.7
> The subject line is manually entered b/c I no longer
> have the mails from the original thread; I&#
There is the task:
I am not too sure if I am using the new ResourceCollection stuff
correctly!
Peter
/*
* Copyright 2006 The Apache Software Foundation
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* Y
2006 07:29:37 -0700 (PDT)
Von: Matt Benson <[EMAIL PROTECTED]>
An: Ant Developers List
Betreff: Re: classloader for 1.7
> The subject line is manually entered b/c I no longer
> have the mails from the original thread; I'm hoping
> the mail archive will do the right thing anyway.
The subject line is manually entered b/c I no longer
have the mails from the original thread; I'm hoping
the mail archive will do the right thing anyway. So
(and I make no pretense that this isn't DIRECTLY
related to my other thread)...
Peter, what's the status of your simplified extension
task p
Opps, I hit "Send" instead of "Save Now",
to finish:
We do not have control of the classes loaded in
the Project classloader when ant is called from
the java api, so we cannot make a "the transitive closure of Project".
I think however that ant-launcher can initialize the componentclassloader
wi
I do not want this to be too complicated.
The component classloader should delegate everything to its parent
classloader (the Project.class.getClassLoader()) except for ant's optional
tasks and types.
Ant code has no control over the contents of the classlaoder for
project, it is set up differentl
Peter Reilly wrote:
I have done a prototype and the results look good:
Does indeed look safer to me.
b) it intercepts classes with "optional" or "ScriptRunner" in the
name and loads these from itself rather than from the parent
classloader.
Note that you could exclude all of .taskdefs.** i
On 8/30/06, Jesse Glick <[EMAIL PROTECTED]
> wrote:
Peter Reilly wrote:> So what would be the hier for the secondary class loader?> if it is bootstrap->ext->system->project->secondary,Yes.> it would meant that the classes in project would not be able to see the
> secondary classes.Correct. Why is t
Peter Reilly wrote:
So what would be the hier for the secondary class loader?
if it is bootstrap->ext->system->project->secondary,
Yes.
it would meant that the classes in project would not be able to see the
secondary classes.
Correct. Why is this a problem? Ant core classes should not be r
_
From: Peter Reilly [mailto:[EMAIL PROTECTED]
Sent: Wednesday, 30 August 2006 2:20 AM
To: Ant Developers List
Subject: Re: classloader for 1.7
On 8/25/06, Stephen McConnell <[EMAIL PROTECTED]> wrote:
Given that a solution that limits mutation to the system class
On 8/25/06, Stephen McConnell <[EMAIL PROTECTED]> wrote:
Given that a solution that limits mutation to the system classloaderprovides the potential solution for all requirements - I think it is thesafer next step (i.e. update the class in SCN to handle a URLClassLoader,
with formal documented e
I can see some problems with this approach.
The classloader hierarchy currently is: (using
bootstrap (rt.jar etc)
extension (jre/lib/ext/*.jar)
system loader (ant-launcher.jar)
project (ant.jar + contents of $ANT_HOME/lib + ~/.ant/lib)
taskdef ...
For netbeans the hierarchy is similar:
bootstr
Peter Reilly wrote:
The antlibs solution would be ok, but it does not solve ant optional tasks
Perhaps we could arrange for tasks in the standard distribution not to
be loaded until actually used. (This might be a performance win anyway.)
Rather, the Ant core would create a new AntClassLoader
I'm currently digging into the classloader expansion issue with the
objective of establish how system classloader state modification can be
achieved without breaking in applications that establish Ant as an embedded
application.
Background:
JDK 1.4 introduced the java.system.class.loader proper
> -Original Message-
> From: Conor MacNeill [mailto:[EMAIL PROTECTED]
> Sent: Friday, 25 August 2006 9:42 AM
> To: Ant Developers List
> Subject: Re: classloader for 1.7
>
> I'm +1 on having the classloader task and being able to add
> paths to the mai
The 1.6 classloader uses the "core" classloader of project. The
core classloader was an attempt to set the classloader for loading
tasks in an early alpha 1.6 ant
development cycle - it did not work and was not used and was never removed
from
the code base.
A simple classloader extender may be a
I'm +1 on having the classloader task and being able to add paths to the
main Ant classloader. Internally, we have been using a very simple task
to extend the classpath and it is a very elegant solution for users who
don't want to modify their ant install, etc. It is not without its
hazards but th
On Fri, 25 Aug 2006, Stephen McConnell <[EMAIL PROTECTED]> wrote:
>> From: Stefan Bodewig [mailto:[EMAIL PROTECTED]
>
>> It has been included since 1.6.0 we just never told anybody.
>> Or is the one you want any different from the classloader
>> task that is already part of the code base.
>
>
> -Original Message-
> From: Antoine Levy-Lambert [mailto:[EMAIL PROTECTED]
> Sent: Friday, 25 August 2006 3:49 AM
> To: Ant Developers List
> Subject: Re: RE: classloader for 1.7
>
> Hello Steve,
>
> the classloader which is in the codebase now is not the
hricht
Datum: Fri, 25 Aug 2006 01:50:54 +0930
Von: "Stephen McConnell \\(DPML\\)" <[EMAIL PROTECTED]>
An: "\'Ant Developers List\'"
Betreff: RE: classloader for 1.7
>
>
> > -Original Message-
> > From: Stefan Bodewig [mailto:[EMAIL
> -Original Message-
> From: Stefan Bodewig [mailto:[EMAIL PROTECTED]
> It has been included since 1.6.0 we just never told anybody.
> Or is the one you want any different from the classloader
> task that is already part of the code base.
Stefan:
Can you provide a summary of what
Peter Reilly wrote:
> Subject: classloader for 1.7
>
> If it is not a little too late, I would like to get a vote on
> including classloader into ant 1.7.
> (http://issues.apache.org/bugzilla/show_bug.cgi?id=28228)
I've been reading through the documentation on the proposed classloader
data
On Wed, 2006-08-23 at 10:33 +0100, Steve Loughran wrote:
> Peter Reilly wrote:
> > If it is not a little too late, I would like to get a vote
> > on including classloader into ant 1.7.
> > (http://issues.apache.org/bugzilla/show_bug.cgi?id=28228)
> > I have been playing with it the last week or so.
On Wed, 23 Aug 2006, Peter Reilly <[EMAIL PROTECTED]> wrote:
> The one in 1.6 does not work.
Didn't know that.
> The proposed one is totally different.
But its intention is the same and so my opinion is the same, add a
warning to the docs and it's OK.
The only thing I'd be concerned about is t
Peter Reilly wrote:
If it is not a little too late, I would like to get a vote
on including classloader into ant 1.7.
(http://issues.apache.org/bugzilla/show_bug.cgi?id=28228)
I have been playing with it the last week or so. It really makes
working with antlibs, and scripting different languages
Hello Peter,
I have just had a look at the link provided. It looks powerful.
Cool that it already has documentation.
Regards,
Antoine
On Aug 23, 2006, at 2:26 PM, Peter Reilly wrote:
The one in 1.6 does not work.
The proposed one is totally different.
see http://enitsys.sourceforge.net/ant-c
Yes, classloader issues do give me headaches.
I see a number of use cases for the classloader task;
1) stop the need to do add third-party jars to ~/.ant/lib for
the ant optional tasks.
This could be avoided if all the optional task are converted a la junit,
but that would be a big job.
The one in 1.6 does not work.
The proposed one is totally different.
see http://enitsys.sourceforge.net/ant-classloadertask/
Peter
On 8/23/06, Stefan Bodewig <[EMAIL PROTECTED]> wrote:
On Tue, 22 Aug 2006, Peter Reilly <[EMAIL PROTECTED]> wrote:
> If it is not a little too late, I would like
On Tue, 22 Aug 2006, Peter Reilly <[EMAIL PROTECTED]> wrote:
> If it is not a little too late, I would like to get a vote on
> including classloader into ant 1.7.
It has been included since 1.6.0 we just never told anybody. Or is
the one you want any different from the classloader task that is
a
Peter Reilly wrote:
If it is not a little too late, I would like to get a vote
on including classloader into ant 1.7.
+0.5 I guess, if someone is testing it carefully (URLClassLoaderAdapter
in particular is a little unnerving). Really the advantage #1,
"to avoid the need to either change Ant
Hello Peter,
I am +1,
Regards,
Antoine
Original-Nachricht
Datum: Tue, 22 Aug 2006 14:24:13 +0100
Von: "Peter Reilly" <[EMAIL PROTECTED]>
An: "Ant Developers List"
Betreff: classloader for 1.7
> If it is not a little too late, I would like to get a vote
> on including classloa
43 matches
Mail list logo
