On 23/05/07, Tim Moloney <[EMAIL PROTECTED]> wrote:
Richard S. Hall wrote: > Stuart McCulloch wrote: >> On 22/05/07, Richard S. Hall <[EMAIL PROTECTED]> wrote: >>> In summary, the dilemma we had was that everyone wanted the artifactId >>> to be the short name and the JAR file to be the long name. But if we >>> made the artifactId the short name, then we also got the short name for >>> the JAR. >> >> FYI, it is possible to alter the final name of the jarfile from the >> default. >> >> The maven super pom, which all top-level poms inherit from, defines: >> >> <build> >> <finalName>${artifactId}-${version}</finalName> >> ... etc ... >> >> but afaik there's no reason why the top-level felix pom couldn't say: >> >> <build> >> <finalName>${groupId}.${artifactId}-${version}</finalName> >> ... etc ... >> >> then any jars in target folders would have the long form of the name. >> >> However, the jarfile installed in the maven repository will still have >> the standard ${artifactId}-${version} name, because that's defined >> by the repository layout. > > I guess that could be okay, although I still think it could lead to > confusion for people who might manually download artifacts from a > maven repository. What does everyone else think? > > -> richard >If I understand this correctly, bundle jar files in the Maven repository would have short filenames and they would have long filenames everywhere else. This causes problems for implementing FELIX-219 (Update maven-bundle-plugin to install bundles to a local OBR) since the OBR would have to know whether it is part of a Maven repository to adjust the filename accordingly. This seems overly complicated to me. I think that we should just stick with the long filenames.
just to be clear, I'm ok with the long filenames and keeping things simple - was just pointing out the final target jar name isn't fixed ;)
Tim
-- Cheers, Stuart
