Re: [gnugo-devel] Re: testing the 1-D boardcode: BOARD_OFFSET

[Heikki Levanto]

On Mon, Apr 29, 2002 at 07:47:19PM +0200, Teun Burgers wrote:
> However every new approach to testing has revealed bugs. 

Ok, here is a simple one:
#define POS so that it asserts that i,j is either on board, or is PASS_MOVE
or any other acceptable values. Slows the program down a lot, but may find
the occasional case where i or j get out of bounds.


> > I suggest not using the patch.
> It doesn't affect anything at BOARD_OFFSET = 0.
> Why not leave it in, if only as an idea for the future.

Although I like the idea of future references, they better be commented
clearly as such, and explained why they are left in, and especially why
they are not being used.

-H

-- 
Heikki Levanto  LSD - Levanto Software Development   <address@hidden>