[resubmitted having pulled from sources and
fixed careless mistake caused by manually entering
a change that i'd made on another system, without testing it.]

for the record, these are the changes:
	printf("%c", 0);
		could produce unexpected results (it read uninitialised memory)

	the utf function was in the wrong place, so never worked.

Notes:
Fri Mar 10 10:41:55 EST 2006 rsc
    I don't see why you need
    
    	else{
    		*p++ = 0;
    		*p = 0;
    	}
    
    why can't you just use *p = 0 alone?
    Do you really want an embedded NUL in your string?
    Is that really going to work?
    

Fri Mar 10 11:44:41 EST 2006 rsc
    never mind, i see that it's fine.


Reference: /n/sources/patch/applied/awkbugs.again
Date: Thu Mar  9 17:39:25 CET 2006
Reviewed-by: rsc

--- /sys/src/cmd/awk/run.c	Thu Mar  9 16:45:54 2006
+++ /sys/src/cmd/awk/run.c	Thu Mar  9 16:45:36 2006
@@ -903,8 +903,10 @@
 			if (isnum(x)) {
 				if (getfval(x))
 					sprintf(p, fmt, (int) getfval(x));
-				else
+				else{
 					*p++ = '\0';
+					*p = '\0';
+				}
 			} else
 				sprintf(p, fmt, getsval(x)[0]);
 			break;
--- /sys/src/cmd/awk/lex.c	Thu Mar  9 16:47:13 2006
+++ /sys/src/cmd/awk/lex.c	Thu Mar  9 16:46:46 2006
@@ -86,8 +86,8 @@
 	{ "system",	FSYSTEM,	BLTIN },
 	{ "tolower",	FTOLOWER,	BLTIN },
 	{ "toupper",	FTOUPPER,	BLTIN },
-	{ "while",	WHILE,		WHILE },
 	{ "utf",	FUTF,		BLTIN },
+	{ "while",	WHILE,		WHILE },
 };
 
 #define DEBUG