Re: [Monotone-devel] multiple (3+) ancestors

[Timothy Brownawell]

On Fri, 2007-02-09 at 10:53 -0800, Zack Weinberg wrote:
> On 2/8/07, Markus Schiltknecht <address@hidden> wrote:
> > Hi,
> >
> > http://www.venge.net/monotone/wiki/MultiParentWorkspaceFallout on the
> > wiki currently states: "A revision can't have more than two parents."
> > While Graydon recently answered me, that monotone itself does not pose
> > such a limit.
> 
> Graydon is mistaken.  revision_t::check_sane() reads in part
> 
>  if (edges.size() == 1)
>     ...
>   else if (edges.size() == 2)
>     ...
>   else
>     // revisions must always have either 1 or 2 edges
>     I(false);

That's not a limitation of monotone, just our current code. There's no
real reason not to remove that limitation... we have know how to do
n-parent merging for scalars (deterministic *-merge), and once we figure
out a good way to handle/present n-parent text merging then fixing the
code is a simple matter of programming.

It's even simpler if we only care about accepting such revisions rather
than being able to generate them; then all we have to do is replace a
few I()'s with something sensible.


-- 
Timothy

Free (experimental) public monotone hosting: http://mtn-host.prjek.net