Actually, it's the real POM which is tested, thus it's not possible to add a 
minimum set of elements.
But I just finished to commit a new tag in the artifact plugin which allow to 
generate an aggregated pom which resolves inheritence,
jelly variables, versions override, and we could use it to test the POM.

WDYT ?

Arnaud



> -----Message d'origine-----
> De : [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] De la 
> part de Carlos Sanchez
> Envoyé : mercredi 5 octobre 2005 00:31
> À : Maven Developers List
> Objet : Re: Maven xsd
> 
> I've opened an issue about POM not requiring groupId.
> My question is: the pom validates against schema before 
> checking extension or is a temp pom aggregating the parent 
> pom data the one that is checked against the schema?
> 
> On 10/4/05, Lukas Theussl <[EMAIL PROTECTED]> wrote:
> >
> > Thanks Tony for your remarks!
> >
> > So what you are saying is that we can drop the namespace 
> requirement 
> > if we make sure to check the pom version instead? Would 
> something like 
> > the following do:
> >
> > <xs:element name="pomVersion">
> > <xs:simpleType>
> >    <xs:restriction base="xs:string">
> >      <xs:enumeration value="3"/>
> >    </xs:restriction>
> > </xs:simpleType>
> > </xs:element>
> >
> >
> >
> > Another general question: I noticed that maven-project.xsd 
> defines a 
> > minOccurs="0" attribute for _all_ elements except the root 
> (project).
> > That means that a completely empty pom is perfectly valid? I would 
> > have thought that there should be a minimum set of elements 
> that make 
> > up a valid pom. It would also facilitate the tracking of 
> some obscure 
> > error messages that we sometimes get when a plugin tries to 
> access an 
> > inexistent pom element.
> >
> > Thanks again,
> > -Lukas
> >
> >
> >
> > Anthony B. Coates wrote:
> > > Since POM instances identify their POM version, there is 
> a case that 
> > > the  namespace does not need to change with every new POM version.
> > > However, it  sounds to me like there is a problem with the way 
> > > "pom:validate" is  conceived.
> > >
> > > If you are validate a POM instance for an older version 
> of the POM, 
> > > you should be using the older version of the POM Schema.  I 
> > > expect/assume/hope  that Maven takes note of the POM 
> version when it 
> > > reads a POM instance, and  doesn't just assume that any 
> POM instance 
> > > can be assumed to be compatible  with the latest POM 
> version.  The 
> > > same approach should apply to  "pom:validate".
> > >
> > > If you do remove the namespace, I hope that each Schema 
> version will 
> > > strictly check the value of the POM version, so that you 
> still get a 
> > > sanity check that you are using the right Schema version for the 
> > > right POM  version.
> > >
> > > Cheers, Tony.
> > >
> > > On Tue, 04 Oct 2005 16:41:00 +0100, Lukas Theussl 
> > > <[EMAIL PROTECTED]>
> > > wrote:
> > >
> > >>
> > >>
> > >> Carlos Sanchez wrote:
> > >>
> > >>> Hi,
> > >>>  I've poms without namespace which are working under 1.1b2, why 
> > >>> are you saying it's required?
> > >>>
> > >>
> > >> Sure they work but they don't validate. The problem that I have 
> > >> currently is that running 'pom:validate' on a pom 
> without namespace 
> > >> declaration seems to hang the msv verifier. This is 
> probably a bug 
> > >> in msv, but I find it quite inconvenient anyway to enforce this 
> > >> namespace  declaration.
> > >>
> > >> Cheers,
> > >> Lukas
> > >>
> > >>
> > >> 
> -------------------------------------------------------------------
> > >> -- To unsubscribe, e-mail: [EMAIL PROTECTED] For 
> > >> additional commands, e-mail: [EMAIL PROTECTED]
> > >>
> > >
> > >
> > >
> >
> > 
> ---------------------------------------------------------------------
> > To unsubscribe, e-mail: [EMAIL PROTECTED] For 
> > additional commands, e-mail: [EMAIL PROTECTED]
> >
> >
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED] For 
> additional commands, e-mail: [EMAIL PROTECTED]
> 




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

Reply via email to