[
https://issues.apache.org/jira/browse/MRESOLVER-633?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17905459#comment-17905459
]
Tamas Cservenak commented on MRESOLVER-633:
-------------------------------------------
And to summarize:
* unresolvable artifacts as "remote repo said is not here (404)" are cached
UNLESS user uses -canf=false
* the other (transport) errors are not cached (by def) nor option for it
exposed via mvn cli.
* The CLI -U (misleading, as we do not talk about snapshots) solves these
cases: will refetch (otherwise will try after midnight again)
* proper local repo hygiene is recommended
> Clarify caching of unresolvable artifacts in local repository
> -------------------------------------------------------------
>
> Key: MRESOLVER-633
> URL: https://issues.apache.org/jira/browse/MRESOLVER-633
> Project: Maven Resolver
> Issue Type: Improvement
> Components: Resolver
> Reporter: Konrad Windszus
> Priority: Major
> Fix For: 2.0.5
>
>
> A dedicated section in
> https://maven.apache.org/resolver/local-repository.html about local caching
> of unresolvable artifacts would be beneficial.
> IIUC it respects the update policy and all files not found in any remote repo
> are stored as 0 byte artifacts in the local repo.
> It is unclear to me how to deal with edge cases like adding a new repo which
> provides that previously unresolvable release artifact? What do I need to do
> from a consumer perspective to resolve that, if the 0 byte artifact is
> already cached in the local repo...
> Other edge cases include a repository which is only available via a VPN
> connection, which is not always established, so the remote repo may not be
> available at all times and the availability may change without any changes on
> remote repositories on Maven side ({{settings.xml}} or some {{pom.xml}})
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
