Thanks Richard, I think in this case I probably export th se.digia.sts.persistence from more than one bundle which I guess could cause this problem. I had a general setting (in my maven parent) for the maven-bundle-plugin that defaulted to exporting se.digia.sts*. This is probably not a good idea...
Thanks for your help, /Bengt 2010/9/20 Richard S. Hall <[email protected]> > On 9/20/10 16:23, Bengt Rodehav wrote: > >> * >> >> Since moving to Felix 3 I've noticed a lot of BLAMED ON logging I haven't >> seen before. I now use Felix 3.0.2. Can someone help me interpret the >> following exception? >> > > Yes, those are coming from the new resolver implementation. > > > >> Caused by: org.osgi.framework.BundleException: Constraint violation for >> >>> package 'se.digia.sts.persistence' when resolving module 32.0 between >>> existing import 28.0.se.digia.sts.persistence BLAMED ON [[32.0] package; >>> (&(package=se.digia.sts.persistence)(version>=1.0.0))] and uses >>> constraint >>> 29.0.se.digia.sts.persistence BLAMED ON [[32.0] package; >>> (package=se.digia.sts.refdata.domain)] >>> >>> > This is saying that while trying to resolve 32, we selected 28 as the > provider of se.degia.sts.persistence since 32 imported this package (this is > the first BLAMED ON, which is basically saying this choice was blamed on the > fact that bundle 32 imports the package). However, this choice was > problematic since we can see the same package from bundle 29 due to the fact > that 32 imports se.degia.sts.refdata.domain from it (this is the second > BLAMED ON). > > So, I would guess that this means that bundle 29 exports both the > persistence and domain packages and that its domain package uses its > persistence package. This would then expose 32 to two providers of the > persistence package which cannot be allowed. > > In theory, it seems like the original import for the persistence package > should be satisfiable by 29. If this is not happening, then maybe 29 isn't > exporting the package with version 1.0.0 or higher. Just a guess. > > > > In this case the bundle could not be started. Sometimes I get similar >> logging messages but the bundle is still started. I haven't fully >> understood >> what this means. >> >> > Some of these messages are just logged when the error happens, but an > alternative solution is found so the message is only informational at that > point. > > -> richard > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [email protected] > For additional commands, e-mail: [email protected] > >
