Re: RFC: CHANGES

Greg Stein <gstein@lyra.org> writes:
> A good policy is that once an RM constructs the branch, *only* the RM can
> commit to the branch. This prevents a bunch of yahoos from dropping changes
> all over the branch while the RM is trying to test the thing. We've been
> using this strategy in httpd land for a while.
>
> [ in that setup, the RM lays down CVS tags across the codebase; the RM can
> then adjust tags to incorporate fixes; we haven't really done release
> branches so far in httpd, but that's just cuz cvs sucks ]

Oooh, excellent idea, yes.

I'm tweaking our release procedure docs right now.

> I'm with Branko. We want to keep that saying "0.18.0+ (dev build)". If we
> integrate svnversion, then we can include that info, too. But I wouldn't
> recommend dropping the public version number.

Yup. I didn't mean to drop the public release number, just to include
the (possibly mixed) revision number as well.

-K

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Thu Feb 20 18:07:52 2003