Ben Collins-Sussman <sussman@collab.net> writes:
> Greg Stein <gstein@lyra.org> writes:
>
> > On Mon, May 20, 2002 at 11:36:20AM -0500, Ben Collins-Sussman wrote:
> > >...
> > > - on the server side, we get a free fs 'walk' from dir_delta by
> > > comparing the switched tree against revision 0. dir_delta is
> > > driving the xml-output editor in a secondary <resource-walk> mode,
> > > which produces a *flat* list of <resource> elements that contain
> > > just full paths and vsn-rsc-urls.
> >
> > The flat list seems fine. I dislike using dir_delta as a mechanism to do a
> > walk, though. It smacks too much of "everything is a nail for dir_delta to
> > smack."
>
> I guess I'm following svnlook's lead... it passes rev 0 to dir_delta
> as a way of printing full trees. Honestly, I've never understood why
> we don't use this method in our RA layers for performing checkouts.
> Why does each RA layer's do_checkout() have its own implementation of
> manually walking an fs tree? Is it *that* much faster than using
> dir_delta against rev 0?
Much faster? I doubt it. At least, from the server side, dir_deltas
is doing the the same things: svn_fs_dir_entries; loop; recurse.
I personally think that dir_delta is not such a wild hammer that it
requires special caution in its use -- it's not a Hole Hawg; just a
perfectly good DeWalt commercial-strength power drill.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Mon May 20 22:23:46 2002