[
https://issues.apache.org/jira/browse/MNG-7805?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17732885#comment-17732885
]
ASF GitHub Bot commented on MNG-7805:
-------------------------------------
gnodet commented on PR #1148:
URL: https://github.com/apache/maven/pull/1148#issuecomment-1592472126
> Maybe we can just drop consumer stuff, since as we proved, model can
"move" with existing tech as well? Am just worried about gigantic projects and
their heap requirements..
That's really not what the discussions on
[dev@](https://lists.apache.org/thread/t0jj2mnwzxyd6oln3r67s78qvw4hdj4n) seem
to lead to. The problem does not seem to be maven, but "other"(?) tools which
we would not want to break by upgrading the model. So we'll have to forever
produce a consumer pom, which means introducing a bom packaging (to
differentiate between parent poms and consumed boms). However, I think we
should flatten it, so that the used schema is minimal and it will also make
things much faster in the long term when all dependencies are flattened.
> The modelVersion should be inferred from the namespace used
> -----------------------------------------------------------
>
> Key: MNG-7805
> URL: https://issues.apache.org/jira/browse/MNG-7805
> Project: Maven
> Issue Type: Improvement
> Reporter: Guillaume Nodet
> Priority: Major
>
> The model reader currently completely ignores the namespace, but it should be
> quite easy to get it and, if the modelVersion is not set, derive it from the
> namespace.
> This would make modelVersion effectively optional.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
