[2118] in Kerberos_V5_Development
Thoughts about CVS tagging conventions
daemon@ATHENA.MIT.EDU (Tom Yu)
Wed Dec 18 03:12:45 1996
Date: Wed, 18 Dec 1996 03:11:43 -0500
To: krbdev@MIT.EDU
From: Tom Yu <tlyu@MIT.EDU>
I was just thinking that perhaps a more consistent method of naming
various tags in our tree might not be a bad idea. I've come up with
the following:
Branches will be named FOO_BRANCH, with a base tag of FOO_BASE. This
will prevent losing the base point of branches, which could be
problematic.
Release branches will be named R1_0_BRANCH, with code freezes denoted
as R1_0_FREEZE1, R1_0_FREEZE2, etc. The actual release would be
tagged as R1_0_PUBLIC. Post-release patches would be tagged
R1_0_PATCH1, etc. Currently we're using V1_0_BRANCH, etc., but IMHO
it makes more sense to use R1_0_BRANCH, since it's *release* 1.0 of
krb5, and it happens to match the convention in our patchlevel.h.
Then again, maybe tagging code freezes as R1_0_BETA1, etc. would not
be a bad idea, since we do give snapshots out to selected parties as a
sort of limited distribution beta release program.
Comments? Much of this is fairly arbitrary, but some amount of
consistency and logic to tag naming in the future would be nice.
---Tom
