On Monday 21 March 2005 12:05, Ben Collins-Sussman wrote:
> cmpilato and I are having a bit of debate. We'd like opinions on this.
>
> For a very long time, autoversioning (that is, the ability for generic
> DAV clients to write to an svn repository) has been considered an
> experimental feature. Mainly because it only half-worked... because we
> had no locking support.
>
> Now that locking support is real, so is the autoversioning feature.
> The question is:
>
> When installing mod_dav_svn, should autoversioning be 'on' or
> 'off' by default?
I vote for off by default.
> * cmpilato feels that one the original motivations for using
> apache/HTTP/DAV was the long-term promise of being able to use generic
> DAV clients against a repository. Therefore, as of svn 1.2, he thinks
> autoversioning should be considered a "core feature" of mod_dav_svn,
> and that it should be turned on by default.
That may be the original motivations, but now mod_dav_svn offers a bunch
of other motivations. For instance, path-based access control, the
ability to authenticate users against different backends (ldap, kerberos,
digest, basic, etc), and I personally find it nice that I can browse the
head revision. :-)
> * I'm still a bit recicent about that. I feel like opening a
> repository to a bunch of DAV clients might lead to a lot of extra
> commits, lots of extra noise in the commit logs. I feel like an
> adminstrator should understand the implications of autoversioning, and
> thus be forced to turn it on.
Agreed.
> Do others on this list have strong opinions either way? We'd like to
> hear them.
-John
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Wed Mar 23 02:56:46 2005