Justin Erenkrantz <justin@erenkrantz.com> writes:
> I honestly don't think we should concern ourselves with admins who
> downgrade across minor versions. I believe this turns out to be a
> really minor issue. Plus, when we set out our versioning rules, we
> explicitly agreed that the file system would always be
> forwards-compatible, but not backwards-compatible across minor
> releases (see HACKING). I see no reason to alter that policy. I
> think we're being overly conservative when there is little pragmatic
> need to do so.
>
> Now, I'll agree with Greg Hudson on IRC that we could do ourselves a
> favor and delay the automatic format upgrade until a user performs a
> locking function. But, I'm really against requiring admins to take
> explicit action to 'upgrade' a repository to enable a feature. I feel
> this is being way too burdensome on real-world admins. -- justin
These two paragraphs seem contradictory to me.
Concern for admins who downgrade across minor versions is the only
reason to delay the automatic format upgrade until a user performs a
locking function. If we're not going to concern ourselves with such
admins, then there is no reason to delay the format upgrade.
(Delaying the upgrade doesn't "do ourselves a favor"; it's not about
us, it's about the admins. From our point of view, it's simpler to
just do the upgrade unconditionally.)
-Karl
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sat Mar 12 19:37:48 2005