On Wed, Jul 28, 2010 at 10:55:39PM +0300, Daniel Shahaf wrote: > Stefan Sperling wrote on Wed, Jul 28, 2010 at 16:05:49 +0200: > > But there's no harm in making svn patch interpret existing move information > > in git diffs. We can carry out a corresponding copy + delete. > > We won't be generating move git diff headers ourselves, but we should > > be interoperable with git diff and hg diff as much as we can. > > hg diff? I don't recall hearing of 'hg diff' on this list before... are we > trying to be able to parse 'hg diff' output as well? Or how does 'hg diff' > come into the picture?
hg can produce the same unidiff format extensions as git diff: $ hg rename alpha alpha-renamed $ hg diff -g diff --git a/alpha b/alpha-renamed rename from alpha rename to alpha-renamed $ So with us adopting the format as well, it becomes more and more of a de-facto standard for expressing tree changes in unidiff. Stefan
