half here, half in felix-dev I have made some improvements to Felix bundle plugin for manifest generation that you can see here https://issues.apache.org/jira/browse/FELIX-199
once that it's applied and hopefully released we'll give a try with our own maven artifacts On 3/28/07, Nathan Beyer <[EMAIL PROTECTED]> wrote:
On 3/28/07, Carlos Sanchez <[EMAIL PROTECTED]> wrote: > I'm working with the Felix guys to go through all these issues. As > usual Maven would provide a default manifest that you could override > with configuration in your pom. Where is this conversation happening? I'd like to hear more about the details. -Nathan > > On 3/28/07, Nathan Beyer <[EMAIL PROTECTED]> wrote: > > On 3/28/07, Jason van Zyl <[EMAIL PROTECTED]> wrote: > > > > > > We're just trying to help out the OSGi folks so that they aren't > > > repackaging everything which is what they are doing. It's basically > > > doing the minimum work so that JARs produced will function with OSGi > > > containers. If it has no impact on normal users there's no real harm > > > in trying to help. I just don't want Maven used as a vehicle to push > > > OSGi down people's throats. > > > > > > Jason. > > > > I must concur with Jason's note of caution. Be very careful in this > > realm. You can't just arbitrarily add OSGi entries to a manifest of > > every build. OSGi entries must be defined very carefully, as they are > > deployment descriptors and can have a significant impact on how a > > bundle is consumed. There are many factors to take into consideration. > > > > For example, there are multiple ways to declare dependencies, you can > > declare a dependency on a bundle, by name and you can declare a > > dependency on a Java package, which abstracts you from a particular > > packaging. Additionally, each dependency can be optional, meaning the > > bundle can be started, even if that dependency isn't available. The > > dependencies can be version-specific too, so a bundle can require a > > specific version of a bundle or specific version of a Java package. > > Note, though OSGi versions are similar to Maven2 versions, there are > > some major conflicts, such as the interpretation of qualifiers. > > > > Once OSGi information is added to a JAR to make it a proper OSGi > > bundle, those values make an API contract that must be strictly > > adhered to. > > > > It's my opinion that OSGi enablement must be taken on by > > component/project owners. I think that the only part the Maven > > community should play here is adding OSGi support and making it easy. > > The only other consideration might be to consider how Maven2 might be > > OSGi hosted, but I'd be very surprised if this turned out as simple as > > all Maven JARs adding "standard" manifest attributes. > > > > -Nathan Beyer > > > > --------------------------------------------------------------------- > > To unsubscribe, e-mail: [EMAIL PROTECTED] > > For additional commands, e-mail: [EMAIL PROTECTED] > > > > > > > -- > I could give you my word as a Spaniard. > No good. I've known too many Spaniards. > -- The Princess Bride > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [EMAIL PROTECTED] > For additional commands, e-mail: [EMAIL PROTECTED] > > --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
-- I could give you my word as a Spaniard. No good. I've known too many Spaniards. -- The Princess Bride --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]