Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Most obscure bug *ever*

Author: Jon Dart

Date: 13:08:19 10/23/02

Go up one level in this thread


On October 23, 2002 at 15:25:04, Colin Frayn wrote:

>
>How annoying is that?  :)
>Can anyone beat this story?

It's a good story.

I once found a single wrong instruction in a 4-megabyte code
image using an assembly-language debugger, and then patched the
compiler that generated. I also spent a few days "blindly" fixing
an assembly-language runtime that I couldn't run a debugger on,
because it switched itself into protected mode and did other
nasty system-level things (it was basically like an operating
system loader). Try something, watch it crash, try something,
watch it hang, bang head against wall .. repeat.

But those weren't chess problems. The worst chess problems I've
had were hashing errors where the actual symptom was thousands
of nodes away from the real cause (the bad hash data that was
inserted). Hashing is evil to debug because it makes the current
program state heavily dependent on the past search history.

--Jon



This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.