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

[Teun Burgers]

Gunnar Farneback wrote:

> Arend wrote:
> > On Sun, 28 Apr 2002, Daniel Bump wrote:
> > > It seems dangerous to define PASS_MOVE to be other than
> > > zero. It could be hard to track down all the places where
> > > this is assumed.

> Yes, PASS_MOVE == POS(-1, -1) is definitely assumed in various places

So that means PASS_MOVE should be BOARD_OFFSET

> and there are also many places where NO_MOVE or PASS_MOVE is assumed
> to be 0. While it's generally better style to use an explicit NO_MOVE
> or PASS_MOVE in such contexts

agreed

> I don't think this is a real problem.

Maybe not.
However every new approach to testing has revealed bugs. I
wouldn't be surprised if this approach would do so too
(after using PASS_MOVE consistently).

> 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.

Teun