incubator-ivy-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Xavier Hanin (JIRA)" <j...@apache.org>
Subject [jira] Commented: (IVY-383) LatestRevisionStrategy has split personality
Date Thu, 11 Jan 2007 07:15:27 GMT

    [ https://issues.apache.org/jira/browse/IVY-383?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12463824
] 

Xavier Hanin commented on IVY-383:
----------------------------------

Yes, the case is complex, and I agree this is an important bug.

The reason why Ivy does conflict management before having a version number for everything
is to optimize resolution, because if it can know that latest.integration will be evicted
by a revision forced to 1.0, for instance, it will avoid listing all revisions. But this problem
was quite easy to solve with first versions of Ivy which has no pluggable version matcher,
and only supported a few number of dynamic versions, but now it's more complex. LatestConflictManager
and LatestRevisionStrategy should work together with VersionMatcher, and be informed if versions
to sort can still be only version constraints and not real version numbers. Not an easy thing,
but if we have a junit test to reproduce the problem, it will be easier to find a solution.

> LatestRevisionStrategy has split personality
> --------------------------------------------
>
>                 Key: IVY-383
>                 URL: https://issues.apache.org/jira/browse/IVY-383
>             Project: Ivy
>          Issue Type: Bug
>          Components: Core
>    Affects Versions: 1.4.1
>         Environment: all
>            Reporter: John Williams
>            Priority: Critical
>         Attachments: IVY-383-LatestConflictManager.patch
>
>
> LatestRevisionStrategy seems to compare both revision numbers and revision number patterns,
but the implementation of patterns broken.  For instance, patterns like "1.0+" are treated
specially, but patterns like "[1.0,)" are treated like literal revision numbers!  I have a
case in which this causes conflict resolution to produce incorrect results.  I'll hopefully
be posting a patch soon to fix the particular resolution problem I've found, but I'm suspicious
that there are other cases in which patterns are erroneously passed to LatestRevisionStrategy.

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: https://issues.apache.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

Mime
View raw message