gnugo-devel
[Top][All Lists]
Advanced

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

Re: [gnugo-devel] gnugo weirdness


From: Gunnar Farnebäck
Subject: Re: [gnugo-devel] gnugo weirdness
Date: Tue, 16 May 2006 14:10:57 +0200
User-agent: EMH/1.14.1 SEMI/1.14.3 (Ushinoya) FLIM/1.14.2 (Yagi-Nishiguchi) APEL/10.3 Emacs/21.3 (sparc-sun-solaris2.9) MULE/5.0 (SAKAKI)

Hena wrote:
> Im new to the list and could not find any proper policy on how to apply 
> bug reports to the list (maybe i'm blind).

Just sending them here is fine. Even better is to post them as tickets
at http://trac.gnugo.org. In neither case there's any guarantee that
they will be fixed or even responded to quickly but in the latter case
they will at least not be easily forgotten.

> But i have few bugs that 
> i bumped into. these occure in the cvs version.
> 
> I have links to the sgf's
> 
> these are prolly the same bug
> http://hack.fi/~hena/tmp/gnugo-lifeanddeath-error.sgf

Looks like a semeai mistake, trying too hard to kill white. But I fail
to reproduce it by just loading the position.

> http://hack.fi/~hena/tmp/gnugo-seki-error.sgf

At move 45 there's a problem with semeai reading related to the
question whether running away with a tail of a dragon should count as
a success. (http://trac.gnugo.org/gnugo/ticket/114)

The mistake at move 49 is caused by a bad interaction between tactical
reading and owl (life and death) reading. This might require some
redesign of the owl reading to solve.

> also, this seems to apply to some life and death logic also but is 
> a bit funnier :)
> http://hack.fi/~hena/tmp/gnugo-superko-error.sgf

This is fixed by the patch gunnar_7_10.12 attached to
http://trac.gnugo.org/gnugo/ticket/132.

/Gunnar




reply via email to

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