Chris0mos opened a new issue, #2196:
URL: https://github.com/apache/maven-resolver/issues/2196

   ### Affected version
   
   Maven 3.10.0 with Maven Resolver 2.0.24
   
   ### Bug description
   
   Since #2082, `DefaultUpdateCheckManager` logs a warning whenever an artifact 
is downloaded from one repository while another repository holds a not-found 
marker for it.
   
   This also fires for the normal fall-through between repositories within a 
single build. With two repositories where the first one does not host some 
artifacts, the resolver tries the first repository, writes a not-found marker, 
finds the artifact in the second repository in the same session, and then warns 
about the marker it has just written:
   
       Downloading from repo-a: org.example:lib:pom:1.0
       Downloading from repo-b: org.example:lib:pom:1.0
       Downloaded from repo-b: org.example:lib:pom:1.0
       [WARNING] org.example:lib:pom:1.0 was downloaded from <repo-b> while a 
cached not-found from a previous attempt suppresses re-checking <repo-a>; 
expiring the cached not-found so that repository is checked again on the next 
resolution
   
   `repo-a` is requested right before the warning, so the marker cannot be from 
a previous attempt. In a build that downloads many such artifacts, this 
produces hundreds of warnings and drowns out relevant ones.
   
   ### Expected behaviour
   
   Only a not-found marker from a previous session should warn, since that is 
the case the expiry exists for. A marker written during the current session is 
the regular fall-through and could be logged at debug level.
   
   ### Notes
   
   #2112 proposed exactly this distinction and was approved, but it was closed 
without being merged, as was the 1.9.x port that included it (#2119). On 
`master` and in 2.0.24, `getStaleSiblingNotFounds` still logs every expiry at 
WARN. A comment on #2112 from 3 October asking why it was not merged has not 
been answered yet.
   
   Is the session-aware logging from #2112 still planned for 2.0.x, or is there 
a different approach?


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to