On Tue, Mar 13, 2007 at 06:52:55PM +0100, Charles Acknin wrote:
> Right, why not having one single 'svn diff' that outputs one single
> rich-format that is backward compatible with patch(1)? What's the
> point of having one more "other format" if the rich-format fulfills
> patch(1) and 'svn patch' needs?
I answered this more fully elsethread, but briefly: the current output
from 'svn diff' isn't compatible with patch(1) in some cases, so we need
at least two formats. We've deliberately made those decisions so that
'diff' is more useful for reviewing some of the tree changes you can
make, before you commit.
It may also make sense to make the extra tree-delta information optional
in the output, which gives us a third format.
Received on Wed Mar 14 10:03:29 2007
- application/pgp-signature attachment: stored