You are certainly right about the separation of development and deployment time never the less I think people tent to go the comfort way therefore they probably end up using the repo file in production even though we told them not to :)
regards, Achim > Yes, and the point you raised are really important. > We definitely need to find a good mechanism to ensure we don't break > anything on existing karaf (well, maybe in production, we should just > tell people to disable this feature anyway, as it should just be a > call to features:removeurl somehow or modifying the correct > configuration file). > Also, pgp / signing and licensing are definitely good idea too, but > I'm slightly less worried about that, as I think the main goal is ease > of use at developement time and not really a provisioning mechanism to > be used in production (where you want to test before installing stuff > anyway, so you'd not use the global repo directly I think). > > On Sun, May 8, 2011 at 20:40, Jamie G. <[email protected]> wrote: >> To be clear the general concept I am ok with, I'm just reviewing >> potential issues that should be resolved. >> >> As an enhancement to the concept, I also think that each entry in a >> repo file should include a license notice of some sort so that we can >> adhere to categories A, B, and X under Apache licensing rules: >> http://www.apache.org/legal/3party.html#criteriaandcategories >> >> Cheers, >> Jamie >> >> On Sat, May 7, 2011 at 3:35 AM, Andreas Pieber <[email protected]> wrote: >>> On Sat, May 7, 2011 at 7:51 AM, Ioannis Canellos <[email protected]> wrote: >>>> Thanks Guillaume, >>>> >>>> >>>>> If we agree on that, I think we should also trim down a bit the >>>>> standard descriptor to remove any non core-karaf related features >>>>> (such as spring, spring-dm, spring-web, and even war). >>>>> >>>> I like the idea of the repository file, however I don't see war and cellar >>>> fall back to this category (imho this is a solution fit for external >>>> projects and not sub-projects). This could be a great idea for providing >>>> functionality to the minimal distribution, but not on standard. >>> TBH I'm with Guaillaume here (although I also share Jamies concerns >>> about stability). This is quite similar to what I've described on >>> another thread (extracting deployer (spring at least), management, >>> web, ...). This will allow us to keep the "karaf-kernel" as small as >>> possible. In addition, using this model we can release components >>> independent of Karaf. >>> >>> Kind regards, >>> Andreas >>> >>>> -- >>>> *Ioannis Canellos* >>>> * >>>> http://iocanel.blogspot.com >>>> >>>> Apache Karaf <http://karaf.apache.org/> Committer & PMC >>>> Apache ServiceMix <http://servicemix.apache.org/> Committer >>>> * >>>> > >
