Hi Brett --- No idea what the process is at Apache .... so maybe the Tapestry project was at fault maybe not ... but the impact on my project was several hours of work. (The developers impacted did not know what to do to solve the problem they are not maven experts.)
As you said ASF can set up any rules they want. For most of the places where we discovered the snapshots had disappeared - it was no big deal there was a more recent release -- but not everything. maven-repository-plugin and tapestry 4.1.6 for example. Still seems like if there is no later release to upgrade to then the snapshots no matter how old should not be deleted. But you are right ASF can make everyone scramble every time a new snapshot comes out to get their own copy. Who knows today it is snapshots older than 30 days. Tomorrow it could be older than 20,15,10,5 days. But hey rules are rules --- and if rules are important -- then community is less important .... All I am suggesting is that snapshots without a matching release not be deleted no matter how old they are. wrt going back to old builds .. I have worked at companies where we have gone back 6-7 months trying to figure out when a bug showed up -- but that is a little extreme! I am sure running maven is a little like herding cats and you do need *guidelines* but keep in mind the community using ASF code is >> people on ASF developer lists. -Pat P.S. Now about http://jira.codehaus.org/browse/SUREFIRE-446 ??? On Tue, Aug 5, 2008 at 8:00 PM, Brett Porter <[EMAIL PROTECTED]> wrote: > 2008/8/6 Patrick Moore <[EMAIL PROTECTED]>: > > Of course this means that the maven team is imposing their idea of good > > process on the other apache projects like tapestry which was also > impacted > > by this ... > > No, Apache is imposing rules on Apache projects. That makes sense, right? > > > > > I could understand this if the release occurs. (i.e. if 2.2 is released > then > > the 2.2-SNAPSHOT could be reasonably expected to be discarded) > > In the case of Surefire, that's exactly what happened... > > > > > part of the process of tracking down a bug involves going back to known > > states of the code. Using the snapshots to diagnose when a bug appeared > is > > useful as people move toward a release.....something to think about. > > Agreed, but as time increases that becomes much less meaningful over > time due to the number of changes. > > Cheers, > Bret > > -- > Brett Porter > Blog: http://blogs.exist.com/bporter/ > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [EMAIL PROTECTED] > For additional commands, e-mail: [EMAIL PROTECTED] > > -- Patrick Moore blog : http://www.sworddance.com/blog
