monotone-devel
[Top][All Lists]
Advanced

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

Re: [Monotone-devel] Interesting bug in monotone-viz


From: Olivier Andrieu
Subject: Re: [Monotone-devel] Interesting bug in monotone-viz
Date: Fri, 12 Aug 2005 10:46:01 +0200

 Richard Levitte - VMS Whacker [Thursday 11 August 2005] :
 >
 > oliv__a> sure, run monotone-viz with the following environment variable set :
 > oliv__a> 
 > oliv__a>   MONOTONE_VIZ_DEBUG=dot 
 > 
 > Cool
 > 
 > oliv__a> it will produce two files agraph.in.dot and agraph.out.dot
 > oliv__a> (the input and output of the dot process).
 > 
 > Sending them to you privately, as attachments.

OK. So the ancestry graph for the main n.v.monotone branch has 3
components. Monotone-viz does a bit of exploration in the global
ancestry graph to reconnect those components. More precisely, it tries
to find a node at the "bottom" of one component that is an ancestor of
one at the "top" of another component (the path between the two going
through other branches) and adds a grey edge between these nodes.

On my repository, it adds these edges:
  058b0f -> f3005e
  118212 -> 925b3d
which look fine. But in your case, it picks a different second edge, a
much poorer choice because the target of the edge in not at the top of
a component but in the middle of it.

I guess I'll have to improve my algorithm :). Could you re-run
monotone-viz with MONOTONE_VIZ_DEBUG=comp and send me the output ?

-- 
   Olivier




reply via email to

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