On Tue, 2005-09-20 at 11:52 -0400, Dave Neuer wrote:

> To which I have to say: why the hell did someone develop surefire in the
> first place? 

Short answer: classloader issues. Longer answer is that I wanted
something like SuiteRunner which Surefire is based on:

http://www.artima.com/suiterunner/

Surefire runs JUnit tests but also has its own notion of a unit of
testing called a Battery which easily allows things like scripted
testing. There's a Jython battery for example.

> There's already a perfectly good Ant junit task? 

It does work perfectly from Ant. Surefire was written a long time ago
and Ant's classloading might be fixed now but classloading was one of
the primary reasons.

> And why
> their own microcontainer? What the heck was wrong w/ Spring (which lots
> of people already use).

Spring was not on the radar when Plexus was started. Plexus came about
out of my experience with the Avalon project. That said we are looking
at things like Spring and OSGi.

> It seems to me to be a codehaus thing: a propensity to eschew reuse of
> other people's code.

To a certain extent sure, but there really weren't any mature containers
at the time m2 was started. It's been in development longer then most
think.

> So, the upshot is, my plugin doesn't work. It wouldn't work outside of
> m2 anyway (since m2 plugins don't rely on normal Java mechanisms -- like
> setter injection, to set their properties)

They do now. We felt that was important as other folks have asked that
too so we fixed it.

>  so it's not really general as
> I've heard claimed by some here as an argument why maven plugins are
> good - loosely coupled to maven. And to make it work, I might have to
> hack surefire. And plexus. And whatever other 20 wheels have been
> reinvented rather than reused.

We are earnestly trying to roll things in Plexus back into projects like
Jakarta Commons. We've done this with the compiler components we've made
(which I originally lifted from Cocoon) and our exec code. Slowly we
will integrate much of our code back into like Jakarta Commons where
things are more easily shared. Much of the reimplementation is due to me
lifting stuff and hacking it just to get things to work. Brett has
spearheaded the effort to move much of the code used in Maven back to
Jakarta from Plexus and it's not an easy task.

Ideally what we would like is to have all the utility code in Ant and
Plexus back in Jakarta Commons where it can all be maintained for
everyone's benefit.

> I realise that some of the above may be perceived as somewhat
> inflammatory, but it's really just born out of the frustration of seeing
> what seems like it should be an easy task -- one which I *can't imagine*
> I'm the only one requiring -- be so difficult.

Fair enough, this is the kind of feedback we need in order to correct
the deficiencies. I think the last couple posts by yourself and Ashley
are great as you've taken some care in expressing what difficulties
you're running into and we can't fix this stuff without feedback. So
thanks.

> And since I don't really have more time to steal from my project to
> devote to the maven plugin development task, I'm left looking for
> alternatives, or reluctantly planning to rewrite the build process in
> Ant buildfiles in the not too distant future.

We're usually pretty helpful in IRC and here if you want to toss around
ideas. I don't think what you're trying to do would stump us for long if
it's not already possible.

> Respectfully but w/ frustration and confusion,
> Dave
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
> 
> 
-- 
jvz.

Jason van Zyl
jason at maven.org
http://maven.apache.org

We know what we are, but know not what we may be.

  -- Shakespeare


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

Reply via email to