[Top][All Lists]

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

Re: [gnugo-devel] view.pike enhancement

From: Gunnar Farneback
Subject: Re: [gnugo-devel] view.pike enhancement
Date: Mon, 19 Apr 2004 03:50:37 +0200
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)

Arend wrote:
> Yes the logic there had confused me and know I understand why. The
> if (has_value("1234...", testline[0..0]")) happens after
> sccanf has already stripped testline from the starting id, maybe I had
> moved it there by accident. How about this patch instead? Tries to strip
> down complete_testcase and full_testcase a bit more aggressively. (try
> viewing unconditional:25 to see the effect) It includes your patch.

Without analyzing exactly what your patch does, this is how I'd like
them to work:

complete_testcase: The commands needed to be guaranteed to run the
test case with the engine in the right state, excluding persistent
caches (but only because it tends to be too expensive).

full_testcase: The relevant loadsgf, possible other commands leading
up to the position, the test command, the correct answer, and
(important) relevant comments about the testcase. It is not critical
if this occasionally wouldn't be quite accurate. In situations where
we e.g. have an owl_attack followed by a corresponding owl_defense it
would be nice if both test cases were shown together.

The names could probably be chosen better but the semantics are that
complete_testcase contains the commands being sent to the engine while
full_testcase is the text shown in the control window.


reply via email to

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