Re: 'git notes merge' implementation questions

From: Johan Herland <johan@herland.net>
Date: Apr 22, 2010, 08:34 UTC
Message-ID: <201004221034.00639.johan@herland.net>
In-Reply-To: <7v8w8g8hvt.fsf@alter.siamese.dyndns.org>

On Thursday 22 April 2010, Junio C Hamano wrote:

Show 9 quoted lines

> Johan Herland <johan@herland.net> writes:
> > Ok, I'm just worried that it'll force us to re-implement much of the
> > three- way merge logic that's already implemented in the merge
> > machinery.
> 
> There is no way around it, as long as you have that variable fan-out in
> the notes structure.  Changing and unstabilizing "merge" for dubious
> benefit of code reuse is unacceptable, as the part that deals with
> variable fan-out has no benefit to the regular "merge".

Understood. At some point I merely contemplated whether a generic "tree_filter" callback could be added somewhere in the merge machinery, and then whether flattening the fanout could be implemented as such a "tree_filter" (with an inverse "tree_filter" for reconstructing the fanout in the merge result). But it seems the costs outweigh the benefits, even more so as I don't yet see any other use case for such a "tree_filter".

In any case, you have already given me lots of other/better options on how to implement 'git notes merge'.

Thanks! :)

...Johan

-- 
Johan Herland, <johan@herland.net>
www.herland.net