Darwinbots Forum

Code center => Bugs and fixes => Bug reports => Topic started by: Henk on April 16, 2008, 03:50:10 AM

Title: Overflow error in longrunning internet simulation OPEN
Post by: Henk on April 16, 2008, 03:50:10 AM

Hi,

I ran an internet sim overnight. When I came back this morning I saw it overflowed somewhere in the night. I'm posting the sim file here. I hope the bug causing this can be found in such a large file with a longrunning sim.
If you have questions about the sim, please ask

regards,
Henk

[attachment=866:henkserror.rar]

Title: Overflow error in longrunning internet simulation OPEN
Post by: EricL on April 16, 2008, 11:04:38 AM

Could I ask you to post a zip of the sim?  I don't have winrar on this machine.  Thanks.

Title: Overflow error in longrunning internet simulation OPEN
Post by: EricL on April 21, 2008, 11:45:41 AM

bump.  I need a zip to proceed...

Title: Overflow error in longrunning internet simulation OPEN
Post by: Henk on April 22, 2008, 11:40:01 AM

I apologize. Here it is!

Title: Overflow error in longrunning internet simulation OPEN
Post by: Trafalgar on May 10, 2008, 09:48:24 PM

This just happened to me too - in this case, however, I was only away for about 2 hours.

Screenshot of DB after I came back:
http://screencast.com/t/DC3kiWEK (http://screencast.com/t/DC3kiWEK)

error.sim in a zip file: http://shadowlord13.googlepages.com/overflow.error.zip (http://shadowlord13.googlepages.com/overflow.error.zip)

In case it's relevant: The internet log has 138 lines in it, not including the blank line at the top of it.
I'm using windows XP home sp2 on an athlon 64 x2 4600+.

Title: Overflow error in longrunning internet simulation OPEN
Post by: bacillus on May 12, 2008, 03:25:11 AM

These errors seem to happen much faster in evolution sims; I'd look for any unnecessary data recordings.

Title: Overflow error in longrunning internet simulation OPEN
Post by: EricL on May 12, 2008, 11:34:28 AM

Found and fixed one issue having to do with populating the refvars when one bot is viewing another with negative nrg, which can occur mid-cycle before the bot is killed at the end of the cycle.  I'm running the sim long term now under the IDE to see if there are other problems.