Re: [Monotone-devel] Re: lcs mystery

[William Uther]

On 20/08/2007, at 2:51 PM, Graydon Hoare wrote:

> Alvaro Herrera wrote:
> > Markus Schiltknecht wrote:
> > Hi Markus,
> > > To make things even worse, all the links to papers cited in  
> > > differ.scm are not valid anymore ([2] and [3]). And google didn't  
> > > turn up anything useful either.
> > This paper is here:
> > http://www.cs.arizona.edu/~gene/PAPERS/np_diff.ps
>
> That's the Myers algorithm. Curiously, Google can't find the Jaffer  
> paper I read. But Yahoo found it. Read these:
>
> http://www-swiss.ai.mit.edu/~jaffer/ONPLSASES/LSASES.pdf
> http://www-swiss.ai.mit.edu/~jaffer/ONPLSASES/LSASES-slides.pdf

How do these compare with: http://revctrl.org/Resolution and its  
pointer to: http://bramcohen.livejournal.com/37690.html ?

In these posts Bram Chohen claims that he has a way of doing 2-way  
diff that leads to more sensible diffs than simple LCS.

As another aside, we currently have a bug in our 3-way diff that one  
particular ambiguous merge does not generate a conflict.  This is  
because we allow "replace" as an option in a diff, rather than just  
add and delete.  See tests/test_a_merge_9.

I was planning to get to this some time...

Will        :-}