On 4/7/06, David Anderson <david.anderson@calixo.net> wrote:
> On a side-note, I haven't followed the progress of ra_serf. Does it
> now pass all tests, ie. can I build a ra_serf enabled svn and use
> that? If so, we may want to discuss whether to bundle Serf with the
> packages, in the same way that we currently bundle Neon.
ra_serf passes all regression tests except for lock and svnsync.
That's mainly because I haven't gotten around to writing those two yet
- those two aren't very high on my personal priority list. =) I can
have lock and svnsync implemented by the time we start the RC cycles.
So, yes, I feel ra_serf should be a go for 1.4 on a non-default basis.
I don't think it's prudent yet to talk about switching the default
from neon/ra_dav yet. An export of the serf tree is currently only
about 380KB (660KB after running buildconf - darn autoconf; in
contrast, neon-0.25.5 is 3.9MB), so we could think about bundling serf
to ease people trying it out. (Garrett's probably going to laugh, but
it'll mean that serf should then support APR 0.9. Grr.) I would like
at least one release cycle with actual users opt'ing-in before
considering switching the default; but that step isn't warranted yet.
-- justin
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Fri Apr 7 17:26:16 2006