On Sat, Apr 26, 2014 at 10:00:53PM +0200, Christian Couder wrote:
> This patch series comes from what Peff sent in the following thread:
Thanks. As I recall, these were in pretty good shape, and I just read
over them again and didn't see anything wrong.
> I added the following fixes:
> - add "strbuf_release(&result);" in import_object(); this was suggested
> by Eric Sunshine
> - use MODE_LIST instead of MODE_DELETE if no arguments are passed; this
> makes the test suite pass
> - add "--no-replace-objects" when calling "git cat-file" in export_object();
> so that we edit the original object if an object is already replaced
All sensible, I think.
> I am not happy with the fact that if the user doesn't modify the object when
> editing it, then a replace ref can still be created that points to the
> original object. I think something should be done to avoid that.
Yeah, it should be easy to just hashcmp the sha1s after calling
import_object. In fact, I think we can just erase any existing replace
ref in that case (the user might have started with a replace ref and
converted it _back_ to the original object, for example).
> Once that is fixed, I plan to add some tests and documentation, but I wanted
> first to let you know that I am looking at this.
Great. Thanks for working on this.
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html