Mailing-List: contact dev-help@felix.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@felix.apache.org
Received-SPF: pass (athena.apache.org: domain of gnodet@gmail.com designates
 66.249.92.173 as permitted sender)
DomainKey-Signature: a=rsa-sha1; c=nofws;
        d=gmail.com; s=gamma;
        h=message-id:date:from:to:subject:in-reply-to:mime-version:content-type:references;
        b=AyUFDtRorNJ7dNrqqk3BNIZemVZ3JFS5/XyoCCs2dfpET+jBEMG5pk1POPuAlnvM5kt5r1vNHLf2654HriV8yq2QXVYgLNb3+ORlm2qkLKxEFUrq1Gy5DKUGB0PkfTmrsX7ML0kYeLCXL5kunrm3cSzNr176iRJReXhKzoLVoCE=
Message-ID: <b23ecedc0712062307w5120c51pf2259bc53bbd3881@mail.gmail.com>
Date: Fri, 7 Dec 2007 08:07:10 +0100
From: "Guillaume Nodet" <gnodet@gmail.com>
To: dev@felix.apache.org
Subject: Re: Problem with the bundle plugin (and patch)
In-Reply-To: <b23ecedc0712062305n1063e668xa3223dbeb56075af@mail.gmail.com>
MIME-Version: 1.0
Content-Type: multipart/alternative;
	boundary="----=_Part_5555_18713442.1197011230781"
References: <b23ecedc0712060751y14fcefbav3a764c3c454cd201@mail.gmail.com>
	 <81f0d9c0712062259i4fd2399boef30abaf27b5195@mail.gmail.com>
	 <b23ecedc0712062305n1063e668xa3223dbeb56075af@mail.gmail.com>

------=_Part_5555_18713442.1197011230781
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

Note that the benefit of adding that to the maven-bundle-plugin is that you
can configure it on the top pom and you'd not have to add a new plugin /
config in each sub module (as it would be the case when using the dependency
plugin), so while I agree this is a workaround for a known maven bug, i
think the patch is still handy.

On Dec 7, 2007 8:05 AM, Guillaume Nodet <gnodet@gmail.com> wrote:

> Yeah, I understand the problem.
> FYI, the patch Hiram attached adds an optional flag to unpack the artifact
> (which is set to false by default to keep the existing behavior), so the
> patch is quite harmless and it seems this is want you are just talking about
> ;-)
>
>
> On Dec 7, 2007 7:59 AM, Stuart McCulloch <stuart.mcculloch@jayway.net>
> wrote:
>
> > On 06/12/2007, Guillaume Nodet <gnodet@gmail.com> wrote:
> > >
> > > While working on the ServiceMix project, Hiram found a bug and raise a
> > > JIRA
> > > with an attached patch [1].
> > > The problem is that in a reactor build (multiproject build), maven
> > uses
> > > the
> > > ./target/classes folder of dependant modules and add those in the
> > > classpath.
> >
> >
> > yes, this is actually a known bug with the core Maven code, specifically
> > with the
> > MavenProject.java class - it always uses 'target/classes' for projects
> > in
> > the same
> > reactor rather than the final artifact.
> >
> >   ( see the Felix mail archives for previous discussions on this topic )
> >
> > unfortunately because MavenProject.java is a core class, it's not easy
> > to
> > patch
> > or workaround this behaviour - so what people usually do is use the
> > dependency
> > plugin to unpack the bundle after it's been built.
> >
> > the maven-pax-plugin (as used in Pax-Construct) provides another
> > solution by
> > extending the Maven compiler mojo so it uses the original bundle, which
> > means
> > no unpacking is required (it also handles compiling against embedded
> > entries).
> >
> >  This causes a problem, because the bundle plugin does not
> > > extract the classes in that folder, which cause the classpath to not
> > > contain
> > > the neeed classes.
> > > Can someone review the patch and apply it please ?
> >
> >
> > so far I've been resisting the urge to add an unpack feature to the
> > bundle-plugin
> > because the same result is possible by adding the
> > maven-dependency-plugin to
> > the pom, and it's often better to combine plugins in the pom than have a
> > plugin
> > grow into a huge 'does-everything-you-ever-need' jar with duplicated
> > source.
> >
> > anyway, I'll take a look and think about making this an optional feature
> > -
> > because
> > some folks may not want the extra cost of unpacking the bundle, or might
> > want
> > the target/classes directory kept as-is (ie. not replaced with the
> > bundle
> > contents)
> >
> > Thanks!
> > >
> > > [1] https://issues.apache.org/jira/browse/FELIX-433
> > >
> > > --
> > > Cheers,
> > > Guillaume Nodet
> > > ------------------------
> > > Blog: http://gnodet.blogspot.com/
> > >
> >
> >
> >
> > --
> > Cheers, Stuart
> >
>
>
>
> --
>
> Cheers,
> Guillaume Nodet
> ------------------------
> Blog: http://gnodet.blogspot.com/
>


-- 
Cheers,
Guillaume Nodet
------------------------
Blog: http://gnodet.blogspot.com/

------=_Part_5555_18713442.1197011230781--