Re: Is anybody actually using git-cherry.sh?

[Jeff King <peff@peff.net>]

On Fri, Jun 23, 2006 at 06:22:05PM +0200, Johannes Schindelin wrote:

> from git-cherry.sh:
> 
> -- snip --
> for c in $inup
> do
>         git-diff-tree -p $c
> done | git-patch-id |
> while read id name
> do
>         echo $name >>$patch/$id
> done
> -- snap --
>
> AFAICS this _must_ be broken.  git-diff-tree -p <ent> does not emit
> "diff-tree <sha1>", and neither "commit <sha1>" lines. So this code
> would yield just one file, treating all diffs as one huge diff. A
> quick fix would be this change (without the patch I sent out earlier):

Maybe I don't understand what you're saying, but it works fine here
(using latest head of master):
  $ for c in origin origin^; do git-diff-tree -p $c; done | git-patch-id
  1511e4e276ccc98ecf0ea31dad1bc9010869fdaf f60349aa786d519368938d7b6e5bb2006eccb0cf
  86ce7eeedd87a78cd8cac79adb6d4d968ece9e53 50f575fc9836704d45a5f732125b8f58103425a4

It looks like patch-id does a flush whenever a sha1 is found at the
beginning of a line; diff-tree lines simply have the 'diff-tree ' part
ignored.

-Peff