Re: [gnugo-devel] arend_3_3.1: mixed patch

[Gunnar Farneback]

Arend wrote:
> - refutation of hoshi non-joseki added
> [...]
> -(;B[pj];C[#kosumi tsuke for 6-9 stone games];W[qf]PL[B]
> +(;B[pj];C[#kosumi tsuke for 6-9 stone games];W[qf]PL[1]
>  (;B[qe]MA[mk];W[pf]C[S]MA[mi]
> -(;B[nd]MA[mk]C[U]PL[W];W[ph]C[j];B[re]MA[lk]C[j])
> +(;B[nd]MA[mk]C[U]PL[2];W[ph]C[j];B[re]MA[lk]C[j])

The PL[B]/PL[W] to PL[1]/PL[2] changes are meaningless and bad. I
guess this is a good time to repost a message from April last year:

> > The PL property takes color as value, either B or W, not 1 or 2
> 
> This is not quite as clear as one would want it to be. The problem is
> that the "Alphabetical list of properties" in the FF[3] specification
> (http://www.red-bean.com/sgf/ff1_3/ff3.html#SGF4List) includes the
> unfortunate line
> 
> PL     Player whose turn it is      PL[color], 1 = Black, 2 = White
> 
> which contradicts the previously given definition of "color". This is
> probably the reason why certain applications use PL[1] and PL[2]
> rather than PL[B] and PL[W]. In particular CGoban 1 does this, which
> explains why this is occuring in many of the game records in the GNU
> Go regression/games directory. Appended is a patch to CGoban 1.9.11
> which fixes this. GNU Go developers who use CGoban 1.9 are strongly
> encouraged to apply the patch.

/Gunnar

--- sgfOut.c.orig       Fri Apr  6 15:55:26 2001
+++ sgfOut.c    Fri Apr  6 15:56:32 2001
@@ -147,7 +147,7 @@
 
     case sgfType_whoseMove:
       assert(goStone_isStone(me->gVal));
-      str_print(nodeOut, "PL[%d]", (int)(me->gVal == goStone_white) + 1);
+      str_print(nodeOut, "PL[%s]", (me->gVal == goStone_white) ? "W" : "B");
       break;
     case sgfType_move:
       str_clip(nodeOut, 0);