Re: Is XDL_MERGE_ZEALOUS too zealous (or maybe not zealous enough)?

From: Johannes Schindelin <johannes.schindelin@gmx.de>
Date: Oct 20, 2008, 16:17 UTC
Message-ID: <alpine.DEB.1.00.0810201809100.22125@pacific.mpi-cbg.de.mpi-cbg.de>
In-Reply-To: <7v3ais5hb3.fsf_-_@gitster.siamese.dyndns.org>

Hi,

On Sun, 19 Oct 2008, Junio C Hamano wrote:

Show 19 quoted lines

> <<<<<<< ours
> 	/* --track without -b should DWIM */
> 	if (0 < opts.track && !opts.new_branch) {
> 		const char *argv0 = argv[0];
> 	...
> 		opts.new_branch = argv0 + 1;
> 	}
> 
> 	if (opts.track == BRANCH_TRACK_UNSPECIFIED)
> 		opts.track = git_branch_track;
> =======
> 	if (conflict_style) {
> 		opts.merge = 1; /* implied */
> 		git_xmerge_config("merge.conflictstyle", conflict_style, NULL);
> 	}
> 
> 	if (!opts.new_branch && (opts.track != git_branch_track))
> 		die("git checkout: --track and --no-track require -b");
> >>>>>>> theirs

This is that case that adjacent blocks A and B are changed to A' B and A B' in the to-be-merged branches.

I could imagine that you would have this automerged to A' B', but I actually advise against that.

My gut feeling tells me that a "for" statement as block "A" and some change ("B") in the _body_ of the "for" loop that takes advantage of the original version of the "for" statement would be _real_ hard to detect.

Better have trivial conflicts reported, but catch more non-trivial ones (instead of mismerging them).

(Another possibility would be that block "B" is actually an "else". Also in this case it is easy to think of cases that would wreak havoc; OTOH we would mismerge them _already_ if at least 1 common line is unchanged between the blocks. Tough.)

But I might be completely wrong.

Ciao, Dscho "who is a little jet-lagged, so you should trust him even less than normally"