On Mon, 2008-05-12 at 19:58 -0300, Edgar J. De Cleene wrote: > > > El 5/12/08 5:28 PM, "Ken Causey" <[EMAIL PROTECTED]> escribió: > > > It seems to me that when a release team has a policy and has followed it > > for the bulk of the process, then changing it for the minor releases is > > not the best idea. And especially changing it unilaterally and without > > any notice. > You got a point. I could do the .mcz for the change sets changing things. > As I said, we just start to work together. > My model for a team is of "The CRC Card Book" of Bellin and Suchman Simone. > See page 32 Two is only compay, ten is a Crowd. > > If we agree on work together, each must have some rol and all should agree > and be comfortable with. > > Edgar
OK, here's how I see it. The old-style changeset-only method has some points in its favor. I'm not totally oppposed to changing it although that really complicates things for anyone in the future who tries to track down a problem if the history of the problem includes the transition point (from ReleaseBuilder/MCZ to changesets). But really we must decide now before issuing any more updates what we are going to do. 1. Change back to ReleaseBuilder. This would involve updating the relevant packages in the 3.10 source.sqf.repository with the changes that you recently made and then issuing at least one new update that loads the relevant packages. This update wouldn't really do anything other than update the MCZ history as they would overwrite methods with the exact same versions. 2. Continue using only changesets. I don't have a strong opinion which way we go, but I must insist that we decide now and stick to it for the future of all 3.10 development. Does anyone else on this list have an opinion on the subject? Ken
signature.asc
Description: This is a digitally signed message part
_______________________________________________ V3dot10 mailing list [email protected] http://lists.squeakfoundation.org/mailman/listinfo/v3dot10
