From: H.G. Muller <h.g.muller@hccnet.nl>
Date: Sun, 13 Mar 2011 13:09:14 +0000 (+0100)
Subject: Fix o-o castling in new parser
X-Git-Url: http://winboard.nl/cgi-bin?p=xboard.git;a=commitdiff_plain;h=f87e38c6a6e20a385497aa947987a731baf5aedb

Fix o-o castling in new parser

The non-compliant notation of castling with lower-case 'o' should be
preferred over the non-compliant interpretation of moving a Pawn along
the o-file. Not doing so breaks ICS play, as ICS use o-o and o-o-o for
castling! So an exception is now made for the o-file in the 'fxg'
case.
---

diff --git a/parser.c b/parser.c
index 973d109..7afd348 100644
--- a/parser.c
+++ b/parser.c
@@ -242,7 +242,8 @@ if(appData.debugMode)fprintf(debugFP, "trial %d,%d,%d,%d  type %d%d%d%d\n", coor
 		return ImpossibleMove; // for now treat as invalid
 	    }
 	    // fxg stuff, but also things like 0-0, 0-1 and 1-0
-	    if(!piece && type[1] == NOTHING && type[0] == ALPHABETIC && type[2] == ALPHABETIC) {
+	    if(!piece && type[1] == NOTHING && type[0] == ALPHABETIC && type[2] == ALPHABETIC
+		 && (coord[0] != 14 || coord[2] != 14) /* reserve oo for castling! */ ) {
 		piece = 'P'; n = 4; // kludge alert: fake full to-square
 	    }
 	}