ant-notifications mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Nicolas Lalevée (JIRA) <>
Subject [jira] Created: (IVY-1227) UseCacheOnly doesn't respect the cache configuration in the ivysettings
Date Thu, 16 Sep 2010 12:18:32 GMT
UseCacheOnly doesn't respect the cache configuration in the ivysettings

                 Key: IVY-1227
             Project: Ivy
          Issue Type: Bug
          Components: Core
    Affects Versions: 2.2.0-RC1
            Reporter: Nicolas Lalevée

In my ivysettings.xml I have configured some cache which avoid having the resolver use the
common .ivy2 cache. If on the resolve task I set useCacheOnly="true", then my artifacts in
the cache won't be found.
I'll attach a complete test case for the different scenario:
# defaultResolver="ibiblio", useCacheOnly="false", dependency on common-logging-1.1: nominal
behaviour, everything works fine, artifacts get downloaded into .ivy2
# defaultResolver="ibiblio", useCacheOnly="true", dependency on common-logging-1.1: there
is a bug in the CacheResolver, artifacts are not "downloaded", they not found. It was a missing
ensuredConfigured(). I'll commit that as I am quite confident in that fix. With the fix, everything
goes well.
# defaultResolver="chain", useCacheOnly="false", dependency on common-logging-1.2: it should
be not be found in the ibiblio resolver but in the local one. Everything works fine, it get
cached correctly from the local repo.
# defaultResolver="chain", useCacheOnly="true", dependency on common-logging-1.2: it is not
found. As I was debugging it, in fact the CacheResolver isn't configured with the proper cache
manager, so it uses the default one. So it looks in the wrong cache.

My proposal as the last fix is to not use the cache resolver as a dictator resolver when we
want useCacheOnly. I don't see any obvious way to make cache resolver aware of the cache it
should look into, further more when there is chaining. I suggest to move this logic in the
configured resolvers. I tried a very simple patch and it seems to work nicely. See the attached
patch. I am not confident enough to commit it without review as it might be quite incomplete
or have some side effect. At least it works with some chaining and even if there is some useOrigin=true

This message is automatically generated by JIRA.
You can reply to this email to add a comment to the issue online.

View raw message