maven-m2-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Maczka Michal <michal.mac...@imtf.ch>
Subject RE: [trygvela@ifi.uio.no: Re: [jira] Work Started: (MNG-69) m2 fr eaks offline becauseibiblio.org not found, even when a valid file:// repo sitory exists in theremote list]
Date Wed, 18 Aug 2004 15:21:34 GMT
> >
> > What's the plan regarding artifacts which don't have a pom or have 
> > invalid pom?
> 
> I don't know, there are some options:
> 
> a) Drop all artifacts without poms and make a report about 
> what artifacts
>    thats missing poms and make the vendor make the pom.
> b) Just make a pom by guessing the groupId, artifactId, type 
> and version.
> c) Make the poms ourselfs by checking out the v3 poms and use 
> the converter
>    on them.
> 
> I think the best way to go is to go a+b where we'll try to build a few
> opensource projects and make the missing poms. This will 
> assure that most
> projects will build but still a lot of old versions of a 
> artifact is going
> to be missing.
> 
> > 
> > Can you estimate the percentage of them?
> 
> Around 90% I'd guess, a lot for sure. I'm not sure if anyone 
> but us makes
> sure they are deployed.
> 
> > 
> > If we are going to follow the plan as I understand there is 
> no place for 
> > them in m2 repository,
> > Am I right?
> 
> Yes you are.


Citing Jason:
"No artifact will be released into the new repository without metadata.
We can stipulate this and we can enforce this. The POM is fundamental to
Maven and it's existence for each and every artifact in the repository
will be manditory for m2. Any sort of real analysis, not to mention
transitive dependencies, becomes impossible and it is something we can
prevent.

We don't have to do 

> We can create in memory a minimal pom object if the pom file doens't exist
> with this information :
> - groupId
> - artifactId
> - version

No. All artifacts will go in with metadata, there will be no choice."


-------------

So option b)is not really a valid choice ... as it is in this light even
worst then this what Emmanuel proposed 
("We can create in memory a minimal pom object if the pom file doens't
exist")

Taking into consideration that we will have new way of specifying artifact's
groupId 
and that new groupId (like www.apache.org/maven/m2) cannot be guessed from
current groupId (maven) I think that even valid poms which we have 
will be mostly or completly useless.

Only real option which I see is to start fresh with brand new repository and
slowly manually push reliable POMs to it.
This means a lot of work - so I don't think that we should start that before
we are sure that we have stable format of POMs v4.


Michal





Mime
View raw message