align with C standard (optional C style); document differences to C style

for print builtin (align with ksh93 ipv GNU bash)

funcs.c

@@ -25,7 +25,7 @@
 
 #include "sh.h"
 
-__RCSID("$MirOS: src/bin/mksh/funcs.c,v 1.131 2009/09/19 15:16:02 tg Exp $");
+__RCSID("$MirOS: src/bin/mksh/funcs.c,v 1.132 2009/09/19 19:08:46 tg Exp $");
 
 #if HAVE_KILLPG
 /*
@@ -604,7 +604,7 @@ c_print(const char **wp)
 		while ((c = *s++) != '\0') {
 			Xcheck(xs, xp);
 			if ((flags & PO_EXPAND) && c == '\\') {
-				if ((c = unbksl(&s)) == -1) {
+				if ((c = unbksl(&s, false)) == -1) {
 					/* rejected by generic function */
 					switch ((c = *s++)) {
 					case 'c':

misc.c

@@ -29,7 +29,7 @@
 #include <grp.h>
 #endif
 
-__RCSID("$MirOS: src/bin/mksh/misc.c,v 1.119 2009/09/19 15:16:03 tg Exp $");
+__RCSID("$MirOS: src/bin/mksh/misc.c,v 1.120 2009/09/19 19:08:47 tg Exp $");
 
 #undef USE_CHVT
 /* XXX conditions correct? */
@@ -1456,7 +1456,7 @@ getrusage(int what, struct rusage *ru)
  * unprocessed character (unchanged if rv=-1)
  */
 int
-unbksl(const char **sp)
+unbksl(const char **sp, bool cstyle)
 {
 	int wc, i;
 	const char *cp = (*sp);
@@ -1490,7 +1490,21 @@ unbksl(const char **sp)
 		/* assume ASCII here as well */
 		wc = 11;
 		break;
+	case '1':
+	case '2':
+	case '3':
+	case '4':
+	case '5':
+	case '6':
+	case '7':
+	case '8':
+	case '9':
+		if (!cstyle)
+			return (-1);
+		/* FALLTHROUGH */
 	case '0':
+		if (cstyle)
+			--cp;
 		/*
 		 * look for an octal number with up to three
 		 * digits, not counting the leading zero;
@@ -1510,11 +1524,12 @@ unbksl(const char **sp)
 		if (0)
 		/* FALLTHROUGH */
 	case 'x':
-		i = 2;
+		i = cstyle ? -1 : 2;
 		/*
 		 * x:	look for a hexadecimal number with up to
-		 *    two digits; convert to raw octet
-		 * u: look for a hexadecimal number with up to
+		 *	two (C style: arbitrary) digits; convert
+		 *	to raw octet (C style: Unicode)
+		 * u/U:	look for a hexadecimal number with up to
 		 *	four (U: eight) digits; convert to Unicode
 		 */
 		wc = 0;
@@ -1531,7 +1546,7 @@ unbksl(const char **sp)
 				break;
 			}
 		}
-		if (**sp != 'x')
+		if (cstyle || **sp != 'x')
 			/* Unicode marker */
 			wc += 0x100;
 		break;

mksh.1

@@ -1,4 +1,4 @@
-.\" $MirOS: src/bin/mksh/mksh.1,v 1.186 2009/09/19 18:36:58 tg Exp $
+.\" $MirOS: src/bin/mksh/mksh.1,v 1.187 2009/09/19 19:08:47 tg Exp $
 .\" $OpenBSD: ksh.1,v 1.129 2009/05/28 06:09:06 jmc Exp $
 .\"-
 .\" Copyright © 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009
@@ -884,6 +884,11 @@ In places where backslashes are expanded, certain C and
 or GNU
 .Nm bash
 style escapes are translated.
+Explicitly excluded are
+.Ql \e" ,
+.Ql \e' ,
+and
+.Ql \e? .
 These include
 .Ql \ea ,
 .Ql \eb ,
@@ -3308,6 +3313,9 @@ above, as well as
 which is equivalent to using the
 .Fl n
 option.
+Contrary to C escapes, the backslash is retained if followed by an
+unsupported escape character; octal sequences must be preceded by a
+.Sq 0 .
 .Ql \e
 expansion may be inhibited with the
 .Fl r

sh.h

@@ -134,7 +134,7 @@
 #endif
 
 #ifdef EXTERN
-__RCSID("$MirOS: src/bin/mksh/sh.h,v 1.341 2009/09/19 15:16:05 tg Exp $");
+__RCSID("$MirOS: src/bin/mksh/sh.h,v 1.342 2009/09/19 19:08:48 tg Exp $");
 #endif
 #define MKSH_VERSION "R39 2009/09/07"
 
@@ -1587,7 +1587,7 @@ void set_current_wd(char *);
 char *strdup_(const char *, Area *);
 char *strndup_(const char *, size_t, Area *);
 #endif
-int unbksl(const char **);
+int unbksl(const char **, bool);
 /* shf.c */
 struct shf *shf_open(const char *, int, int, int);
 struct shf *shf_fdopen(int, int, struct shf *);