Do we prefer plugin developers to use List and get ClassCastExceptions at
runtime or sue type-safe collections to help them create stronger code ?Anyway,
few plugins allready use Java5. Java 1.4 based one will not be broken as the
generics signature is only a compile-time check.
2009/5/1 Brian Fo
I will take a look shortly
--Brian (mobile)
On May 1, 2009, at 4:59 PM, Arnaud HERITIER wrote:
I'm not administrator to verify but perhaps, the job to fix metadata
isn't
scheduled ?
On Fri, May 1, 2009 at 6:09 PM, Nord, James wrote:
Hi all,
The metadata served by nexus for http://repo
Hi everyone,
It's that time again. For the 2.2.0 release of Maven, we're trying to
keep the number of issues fairly limited to regressions, issues with
quite a few votes that are low-hanging fruit, and some sort of
"strategic" issues. We're moving Maven to JDK 1.5 with this release,
though th
Ok, I was kind of wondering about that. When I first tried trunk it was not
building and had some outdated stuff in there, but I committed changes in both
places to be safe.
I'll stop committing to trunk, and then we can start fresh after 2.0.10 like you
suggested.
Hervé BOUTEMY wrote:
I d
as explained in [1] or in [3], these namespaces are not only namespace
conventions, but they are 2 different ways of deploying the tasks:
- either add a typedef in each Ant build file with a reference to
maven-ant-tasks.jar: "urn:maven-artifact-ant" namespace
- or copy maven-ant-tasks.jar to Ant'
Ok, that makes a little more sense, thanks for the explanation. I think those
are the only places in the docs where those are used.
I guess I still don't understand why you don't just put
"antlib:org.apache.maven.artifact.ant" for both examples. That URI works fine
whether you use the ant li
I'm not administrator to verify but perhaps, the job to fix metadata isn't
scheduled ?
On Fri, May 1, 2009 at 6:09 PM, Nord, James wrote:
> Hi all,
>
> The metadata served by nexus for http://repository.apache.org/snapshots/
> is incorrect for the archetype plugin.
>
> (https://repository.apache
I don't think merging changes to ant-tasks/trunk is worth the effort now.
Trunk was created in r524632, in april 2007, when Maven 2.0.x and Maven 2.1.x
started 2 different branches. Changes were done by Jason in ant-tasks/trunk
to match Maven 2.1.x structural changes, but AFAIK they stopped whe
There are two different xml namespaces used in the docs and examples of the
maven ant tasks [1]. They are "urn:maven-artifact-ant" and
"antlib:org.apache.maven.artifact.ant". The second one seems to better match
the example on the ant site [2].
Anyone have preferences one way or the other?
The maven ant tasks currently creates a shaded jar that combines several of the
dependencies into one jar. While working on MANTTASKS-106 (using ant tasks
inside ant-run) I noticed that the plexus utils classes are relocated to
hidden.org.plexus.utils.
This causes classloader problems when ca
What exactly distinct having a List that only contains Artifact and
internally cast to this type, and use List as method signature ?
2009/5/1 Brian Fox
> I'm not sure if this is in scope of what John is trying to do wrt to 2.2.
>
>
> Jason van Zyl wrote:
>
>> I don't believe anyone actually agre
I'm not sure if this is in scope of what John is trying to do wrt to 2.2.
Jason van Zyl wrote:
I don't believe anyone actually agreed to this yet. Are you sure this
is not going to cause problems for users?
On 1-May-09, at 1:04 AM, nico...@apache.org wrote:
Author: nicolas
Date: Fri May 1 0
> I actually think jamming all this into the JDK is a very
> bad idea ...
Yes, I also strongly second this. What I ment was: OWB will need nothing more
than a simple JRE and doesn't need additional J2EE libraries as it is defined
in the JSR-299 Spec.
All the headers are in javax.inject which h
On 1-May-09, at 12:48 AM, Mark Struberg wrote:
A little bit of self advertisement:
why not consider OpenWebBeans [1]?
You can work on whatever you like, you have access to a sandbox.
Personally nothing interests me except XBR and Guice. Ultimately in
two years it's not going to matter,
I don't believe anyone actually agreed to this yet. Are you sure this
is not going to cause problems for users?
On 1-May-09, at 1:04 AM, nico...@apache.org wrote:
Author: nicolas
Date: Fri May 1 08:04:48 2009
New Revision: 770570
URL: http://svn.apache.org/viewvc?rev=770570&view=rev
Log:
us
nicolas de loof wrote:
is there some CI server set for maven 2.2 branch ?
Just added
https://grid.sonatype.org/ci/view/Maven%202.2.x/
Benjamin
-
To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org
For additional comm
Hi
is there some CI server set for maven 2.2 branch ?
Sonatype grid hasn't (yet) and the POM don't set the element, so I
suppose not
As a side note, did we allredy tested "unbreakable builds" using Teamcity
(or equivalent) on apache infrastructure ? It may avoid me to commit the
ugly regressions
A little bit of self advertisement:
why not consider OpenWebBeans [1]?
It is a best of all worlds approach, will get the official Java standard for
dependency injection and is an Apache project (so no worry about licensing
issues).
OpenWebBeans extends the spec in the way that it also can be
18 matches
Mail list logo
