From: Junio C Hamano <junkio@cox.net>
Date: Sun, 24 Sep 2006 17:47:29 GMT
Subject: Re: [PATCH] Fixes git-cherry algorithmic flaws
Message-ID: <7vodt59mxa.fsf@assigned-by-dhcp.cox.net>
In-Reply-To: <20060924111737.GL20017@pasky.or.cz>

Petr Baudis <pasky@suse.cz> writes:

> Hmm, well, what's curious is that the documentation says
>
> 	Every commit with a changeset that doesn't exist in the other branch
> 	has its id (sha1) reported, prefixed by a symbol.  Those existing only
> 	in the <upstream> branch are prefixed with a minus (-) sign, and those
> 	that only exist in the <head> branch are prefixed with a plus (+)
> 	symbol.
>
> which is in contradiction of Ilpo's description of the old algorithm
> (and also your description of it). It would seem he just wants to fix it
> according to the documented behaviour.
>
> I guess the documentation is what's broken then?

Ah I did not realize that, but yes the documentation is
incorrect.

I wonder if we can kill it by introducing a new rev notation and
using regular rev-list family of commands instead.

What we want here is a way to say "give me commits that are in B
but not in A, but before returning a commit see if there is an
equivalent change in the set of commits that are in A but not in
B, and filter it out".

Time for "rev-list A....B"? ;-)