Re: [Monotone-devel] Re: Bug in monotone lca

[Nathaniel Smith]

On Fri, May 20, 2005 at 09:29:22PM +0200, Wim Oudshoorn wrote:
> Nathaniel Smith <address@hidden> writes:
> 
> > On Fri, May 20, 2005 at 03:37:38PM +0200, Wim Oudshoorn wrote:
> >> However in that case lcad has a bug :-)
> >
> > Err, yep, sounds like it.  Sigh.
> >
> >> I have database with this example (artifically constructed).
> >
> > I guess I should ask if you could turn it into an automated test case?
> > Though part of me would rather just get on with throwing out lcad...
> 
> I can try.  But for this I need to find out how to write tests for
> monotone, so don't expect it overnight.  

Right, no problem.  tests/README is the place to start.

> >> At the moment I am trying to wrap my head around the merging problem as 
> >> well.  It is rather tricky and just creating a mathematical vocabulary
> >> is already quite challenging.  I am writing everything down and 
> >> when it comes to a stage that I can draw some conclusions and can
> >> be understood by more people than just me I will publish it.
> >
> > Yep.  You _are_ familiar with the precise cdv-merge work that Bram
> > (and I, and others) have done?
> 
> No, except a short discussion that started with:
> 
> Bram Cohen: The new Codeville merge algorithm
> 
> Can be found on gmane:
> 
>    http://article.gmane.org/gmane.comp.version-control.revctrl/2
> 
> I haven't found anything detailed on codeville merge.  
> If there is anything else to read I am interested.  

That's probably the best thing current available.

> Also, reading that e-mail I was left with some
> basic questions of what constitute a conflict and
> when it will cleanly merge.  My impression was
> that it will to easily merge changes, which I 
> would prefer to be flagged as conflicts. 
> 
> Maybe I have to read this again and reply in that thread
> with questions about the codeville merge algorithm.

That would be good.  Alternatively, you can go on IRC to #revctrl on
irc.freenode.net, and ask there.

> But I think I am approaching it along a slightly different
> line.  At the moment I restrict myself to the problem 
> of a single file containing a single character.  
> This seems already tricky enough.

Heh, indeed.  Well, I know how to solve it "the codeville way", I'll
be curious to hear what you come up with :-).

-- Nathaniel

-- 
"The problem...is that sets have a very limited range of
activities -- they can't carry pianos, for example, nor drink
beer."