[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: CVS update: subversion/subversion/libsvn_ra_local ra_plugin.c

From: Greg Stein <gstein_at_lyra.org>
Date: 2001-03-10 17:00:24 CET

On Sat, Mar 10, 2001 at 05:51:24AM -0600, Ben Collins-Sussman wrote:
> Greg Stein <gstein@lyra.org> writes:
> > On Thu, Mar 08, 2001 at 06:00:17PM -0000, sussman@tigris.org wrote:
>...
> > > * svn_ra.h (get_commit_editor): take base_path argument, for the same
> > > reason that svn_fs_get_editor now does. Namely, the client already
> > > knows the "grandaddy" dir where it will call replace root. It's
> > > silly to force the editor-driver to always begin changes at the very
> > > top of a revision tree.
> >
> > I don't understand this part. Why isn't the root just placed at BASE_PATH?
> >
> > In other words, don't open the session with "/A" and then pass "B/C/D" as
> > BASE_PATH. Just open the session with "/A/B/C/D".
>
> Ohhhhh, I think see what you're saying!
>
> The fs commit editor *requires* that you specify a REVISION:BASE_PATH
> pair whet you fetch it. This is the place where you want to start
> editing the filesystem (what replace_root() will affect.)
>
> But: this doesn't mean the client needs to see this requirement. As a
> user of RA, it already *has* a way to say where it wants to start
> editing: the ra->open() call takes a URL. Let the client provide
> BASE_PATH within this URL, and provide REVISION to
> ra->get_commit_editor() as usual. Then the RA layer can put 1 and 1
> together when fetching the fs commit editor
>
> Am I understanding correctly? :)

Yessir! Spot on :-)

You have all the information that you need passed into RA (via two call
points: open, get_commit), so adding more parameters to the interface is
just adding complexity.

Cheers,
-g

-- 
Greg Stein, http://www.lyra.org/
Received on Sat Oct 21 14:36:25 2006

This is an archived mail posted to the Subversion Dev mailing list.