[jira] [Commented] (MRESOLVER-381) Undo MRESOLVER-373 as it was fixed by other means

["ASF GitHub Bot (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/MRESOLVER-381?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17739197#comment-17739197 ] 

ASF GitHub Bot commented on MRESOLVER-381:
------------------------------------------

cstamas opened a new pull request, #314:
URL: https://github.com/apache/maven-resolver/pull/314

   Partially undo MRESOLVER-346: 1acec9baf6fcb8eb010d64e8009778a13a88d38e
   Also, make existing "static" name mapper obey the rule (distinguish between artifacts and metadata) and make not about this on NameMapper interface as we..
   
   ---
   
   https://issues.apache.org/jira/browse/MRESOLVER-381




> Undo MRESOLVER-373 as it was fixed by other means
> -------------------------------------------------
>
>                 Key: MRESOLVER-381
>                 URL: https://issues.apache.org/jira/browse/MRESOLVER-381
>             Project: Maven Resolver
>          Issue Type: Task
>            Reporter: Tamas Cservenak
>            Priority: Major
>             Fix For: 1.9.14
>
>
> The MRESOLVER-373 made two important operations serialized (again, since existence of locks), but the proper fix was actually this one (also part of MRESOLVER-373):
> https://github.com/apache/maven-resolver/commit/4fb1cf93b478ab1eeb563c3e69ad888017265e54
> And given that artifacts and metadata does not use same lock name, the code undone would work. Still, errors would occur again, with "static" name mapper (maps always to same name), hence some smarter change is needed: when artifact resolver invokes metadata resolver, maybe it needs to unlock and then invoke?
> The ultimate goal: to not serialize on "happy path", but make it use "shared" locking so multiple thread can work together.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)