Re: [SVN-DEV] Re: [Issue 422] Changed - "svn diff" needs to be fleshed out.

"C. Scott Ananian" <cananian@lesser-magoo.lcs.mit.edu> writes:

> On 4 Sep 2001, Ben Collins-Sussman wrote:
>
> > 1. svn diff foo.c
> > 2. svn diff -r 5 -r 7 foo.c
> > 3. svn diff -r 5 foo.c
>
> So, are you saying that if the user has a (client-side) custom diff
> program installed, that case 1 will behave (possibly radically)
> differently from 2 and 3?

Yeah, possibly.

>
> This seems like a misfeature. Why not check out the revisions into
> temporary files and then run the client-side diff? Slightly more
> inefficient, perhaps, but cooperates much better with the user's
> preferences.

True. Maybe consistency is important here.

> (The 'efficient' way to do this is to generate *svn* diffs between
> revisions, apply them to the working copy to create a pair of
> temporary files, and *then* run the user's 'diff' program on the
> pair. This way you get the efficiency of the server-side approach
> (since the server "knows" the diffs) with the consistency of the
> client-side model.)

Yeah, that could work. I wonder how Greg/Karl feel about this.

Still, our RA vtable has no way (at the moment) to request either the
plaintext (or svndiff data) for a particular file. The latter is done
all the time during drives of "editors", but there's no way to ask for
that info standalone. Should be easy to do.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sat Oct 21 14:36:39 2006