[gecode-users] Critical bug.

[Christian Schulte]

Dear Rafael,

thanks for insisting and take our apologies for being stubborn: but as you
did not hint at the fact that you actually observe the bug we were
understandibly defensive (not to mention that it was not reported to the bug
tracking system).

The bug is fixed and Gecode 1.3.1 should show up today on the web.

So that this was the first real bug in Gecode!

All the best
Christian

--
Christian Schulte, http://www.imit.kth.se/~schulte/ 

-----Original Message-----
From: Rafael Meneses [mailto:rafaparafernalia at gmail.com] 
Sent: Tuesday, October 24, 2006 9:50 PM
To: cschulte at kth.se
Cc: users
Subject: Re: [gecode-users] Critical bug.


On 10/24/06, Christian Schulte <cschulte at kth.se> wrote:
> This is not a bug! The invariant is that whatever branching 
> decsription is created the position always refers to the same 
> variable: In case some variables are removed from the branching, the 
> offset is updated accordingly.
>
> Do you have an example where things go wrong?

I attach an example.

With recomputation there are a "segmentation fault", without recomputation,
all is just fine.

Run the example five minutes, and there are a 100% of probability of crash,
take away recomputation, and the code never crash.

Saludos,


-- 
Rafael Meneses
http://www.zeke.cl
http://www.labsd.inf.utfsm.cl/~rmeneses