----- Michael Haggerty <mhagger@alum.mit.edu> wrote:
> Daniel Berlin wrote:
> > On 1/31/07, Michael Haggerty <mhagger@alum.mit.edu> wrote:
> >> Hi,
> >>
> >> [Insert usual disclaimer about only being the cvs2svn guy, pretty
> >> ignorant of svn internals]
> >>
> >> This discussion reminded me again that the text-base part of a
> working
> >> copy resembles an incomplete, read-only svn repository. Perhaps
> both
> >> could be accessed via a common API?
> >
> >
> > This is actually what i suggested during the summit, that the WC
> > become just another repository .
> > People seemed generally okay with the idea.
> > That said, if David doesn't want to tackle this, i don't blame him.
>
> I wouldn't blame him either, because that would be a lot of work.
>
> But much less work would be required now. One would only need to
> abstract out a ra-like interface for talking to the text-base
> information that is already in the repository (including allowing the
> "data incomplete" errors). David's new WC code could still decide to
> implement only a small fraction of the repository functionality.
>
> The point is, it sounds like the WC API has to be changed anyway, so,
> perhaps this would be a way to make it more general at the same time
> and
> open a lot of doors for future expansion but without requiring new
> functionality to be implemented now.
FWIW, I've long thought it would be nice to have WC access work like the ra layers. I think it would ease the burden in the client code, and remove a lot of edges. However, I never sat down to look at the RA interface enough to know whether it was efficient enough for fast local use.
-John
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Thu Feb 1 12:39:55 2007