Phil Steitz wrote:
I agree with all of your points above. The issue is oversight. I have nothing against public RCs, betas, milestones or nightlies -- on the contrary I like and use all of these things. I just think that other than nightlies, these need to be endorsed in some way by the community if we are going to distribute them. Right now we have some jars out there that, to my knowledge, were never voted on and / or have no community support. I think that these should be removed.
I think that we have always agreed that they need to be removed. Its just been an issue of having the time to do it properly.
I would take issue with pushing nightlies or alphas to ibiblio.org. We do need to put some pressure on people to release.
So, you are OK with milestone builds, but not nightlies? Do we agree that the milestones should have some "status" before they get published and they should also be published with release notes, etc. to the main distribution site?
Note, theres already been allot of discussion concerning these issues across the [EMAIL PROTECTED] and [EMAIL PROTECTED] lists.
The MOST important issue with nightlies etc has also to do with the mirroring of www.apache.org/dist. Its important to be conservative concerning bandwidth, mirrors and rsync. So theres external pressure there not to be pumping auto build nightlies over rsync. Mainly because theres no need or want to have these things on the mirrors.
So to clarify, yes we need to get all interim builds out of java-repository. I don't think we need to work hard at this,
1.) just delete any interim builds (they are already archived on archive.apache.org)
2.) delete and any snapshot symlinks referring to them.
3.) delete and sanpshot-version files referring to an interim build.
-- Mark Diggory Software Developer Harvard MIT Data Center http://www.hmdc.harvard.edu