|
From: | Gunnar Farneback |
Subject: | Re: [gnugo-devel] assertion failure in the board code |
Date: | Fri, 07 Mar 2003 18:18:23 +0100 |
User-agent: | EMH/1.14.1 SEMI/1.14.3 (Ushinoya) FLIM/1.14.2 (Yagi-Nishiguchi) APEL/10.3 Emacs/20.7 (sparc-sun-solaris2.7) (with unibyte mode) |
I wrote: > The position is somewhat extreme so it's possible that some assumption > has turned out to be too optimistic. Unfortunately I haven't been able > to reduce the test case further and just removing a single genmove > seems to suffice to no longer trigger the bug. I think I know what's going on. In any case the test case complexity can be radically reduced by replacing all "genmove white" with "play white pass". It's a safe bet that the move history collapsing around line 1000 is the triggering factor. I have a feeling that the bug may have been introduced with the patch arend_3_13.1. http://mail.gnu.org/archive/html/gnugo-devel/2002-11/msg00335.html /Gunnar
[Prev in Thread] | Current Thread | [Next in Thread] |