monotone-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Monotone-devel] Re: [cdv-devel] more merging stuff (bit long...)


From: Joel Crisp
Subject: Re: [Monotone-devel] Re: [cdv-devel] more merging stuff (bit long...)
Date: Tue, 09 Aug 2005 12:16:29 +0100
User-agent: Mozilla Thunderbird 1.0 (Windows/20041206)

Not to forget that by 'change' we are talking about syntactical differences which may have completely borked semantic meaning and which may 'conflict' in a way a simple syntactic comparason can't determine. Roll on AI assisted merges...

There will always be an element of review and regression test needed. How far is it worth going to produce the 'perfect' merge and how dangerous is it to do a really good merge which reduces people's desire to review by increasing confidence to an unsafe level?

Just my cracksmoking thought for the day..

Joel

Nathaniel Smith wrote:
On Sun, Aug 07, 2005 at 04:40:17PM -0400, Kevin Smith wrote:

Maybe my preference, which might translate badly to an algorithm, is that any time my changes might win over someone elses, take mine, but whenever theirs might win, be conservative and generate a conflict.


That's a wonderful pull quote ;-).

From the point of view of a merge algorithm, it's not quite right, I
think.  My approach it is, when you say "merge that thing into my
tree", you're telling the merge algorithm that you do, in fact, want
the thing over there to end up in your tree.  The point of the
merge algorithm is not to make judgement over the quality of the
changes, but just to figure out what "put that thing in my tree" even
_means_.

The other thing is that merges aren't really in the position of
choosing which changes win over another... the tricky part is
determining what a change _is_ in the first place, and then if once
you've done th there are two of them in conflict, that's always a
conflict...

-- Nathaniel





reply via email to

[Prev in Thread] Current Thread [Next in Thread]