I think this might be relevant now in 1.2.2 as indexes are downloadable.
Typical use-case is using m2eclipse, with maven configured on a repository group in Archiva. In this case indexes are incomplete, and search is very limited.
Jeremie BOUSQUET
added a comment - I think this might be relevant now in 1.2.2 as indexes are downloadable.
Typical use-case is using m2eclipse, with maven configured on a repository group in Archiva. In this case indexes are incomplete, and search is very limited.
Something to take care is karma on repos. Generating a full index is possible but in some cases users will have access to only a limited number of repositories mapped in the group.
So in this case an index will be generated on the fly and not incremental.
Olivier Lamy
added a comment - - edited Something to take care is karma on repos. Generating a full index is possible but in some cases users will have access to only a limited number of repositories mapped in the group.
So in this case an index will be generated on the fly and not incremental.
The idea is to have a cache value to not regenerate the merge index for all requests.
And to cleanup temporary directories having a session listener which will cleanup existing tmp index when user session end.
Olivier Lamy
added a comment - The idea is to have a cache value to not regenerate the merge index for all requests.
And to cleanup temporary directories having a session listener which will cleanup existing tmp index when user session end.
Olivier Lamy
added a comment - better cleanup solution implemented.
I leave the issue open until configurable cache value per repo group implemented and unit test added!
New Feature
Closed
Major
this will only be relevant if the index is downloadable - at present this is not done