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

Re: storing URLs in SVN/ (was: Re: absolute paths)

From: Ben Collins-Sussman <sussman_at_newton.collab.net>
Date: 2000-12-20 21:21:39 CET

Greg Stein <gstein@lyra.org> writes:

> Why separate that stuff? Why not just store the URL? The repository_URL
> passed to RA->open is simply the URL associated with the topmost, common
> directory.
>
> The separate "repository" and "path" is following CVS's pattern for no good
> reason... it is an artificial separation.

Hmmmm... perhaps these are hyserical raisins. The raisin that comes
immediately to mind is that we've been working with nothing but XML
files so far. Regular Unix pathnames have been most intuitive.

> This will be especially important if we end up with a file that comes from a
> different repository. I'm not sure whether you store the URL for a subdir
> in "./SVN/entries" or in "subdir/SVN/entries", but we definitely mix
> repositories on a per-directory basis. If the URL is stored in ".", then
> they definitely have to be mixable.

Yeah, I like this idea. We need the mixability. Karl? Do you agree
that we should change this?

In our `entries' file, we're currently storing ancestor_path and
ancestor_revnum as two separate arguments. Maybe we can blend them
together as something like
"http://www.lyra.org/reposname/67/glorb/foo.c".

(I can't remember, did we come up with a URL schema that includes tree
revisions?)

> [ hmm. it appears that you store ancestry for a subdir in "." ]

Right. Our policy has been that a directory's ancestry info is stored
only once: within the directory's own `entries' file in the "."
entry. Parent `entries' files list the subdir as a child, but give no
other info.
Received on Sat Oct 21 14:36:17 2006

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

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.