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

Re: augmented diff design patch format

From: Malcolm Rowe <malcolm-svn-dev_at_farside.org.uk>
Date: 2007-06-26 11:59:40 CEST

On Sun, Jun 24, 2007 at 01:02:43PM -0700, Eric Gillespie wrote:
> "Charles Acknin" <charlesacknin@gmail.com> writes:
>
> > - we don't want 'svn diff' default behaviour (designed more for
> > patch review than for patch(1) use) to be changed -- 'default'
> > - because of 'default' design, we need another format that really
> > fits with a patch(1) use -- 'unidiff'
>
> The existing format already works with patch, as can any new
> format we come up with, so long as we still use unidiff for
> actual text diffs; patch ignores everything else. Since the

(Note that the existing format does not produce valid patch(1) input for
e.g. copy-with-history operations. I'm not sure whether you were
suggesting that it did, or whether you were just saying that we could
make the unidiff parts unidiff, and use the tree-deltas part in preference).

As Charles said, there are effectively three use cases for diff:

- the current output, suitable for code reviews (and semi-working for
  patch(1)).
- patch(1)-specific output, differing from the previous version in
  only a few minor ways (directory deletes, copy-with-history).
- an svn-specific-tree-aware patch output.

Regards,
Malcolm

  • application/pgp-signature attachment: stored
Received on Tue Jun 26 11:59:24 2007

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