* use only macros for ctype stuff any more

XXX one of these uses a gcc extension, ok for now tho
* don't include <ctype.h> any more at all
* don't try nl_langinfo in small mode, just check locale

saves 171 .text, 4 .data, 256 .bss, 1 import

Build.sh

@@ -1,5 +1,5 @@
 #!/bin/sh
-# $MirOS: src/bin/mksh/Build.sh,v 1.74 2006/11/09 23:09:05 tg Exp $
+# $MirOS: src/bin/mksh/Build.sh,v 1.75 2006/11/10 07:18:56 tg Exp $
 #-
 # Environment: CC, CFLAGS, CPPFLAGS, LDFLAGS, LIBS, NROFF
 
@@ -18,8 +18,6 @@ ac_testn()
 {
 	f=$1
 	fu=`upper $f`
-	test 0 = "$HAVE_$fu" && return
-	test 1 = "$HAVE_$fu" && return
 	if test x"$2" = x""; then
 		ft=1
 	else
@@ -28,9 +26,21 @@ ac_testn()
 	fi
 	fd=$3
 	test x"$fd" = x"" && fd=$f
+	eval fv=\$HAVE_$fu
+	if test 0 = "$fv"; then
+		$e "==> $fd... no (cached)"
+		return
+	fi
+	if test 1 = "$fv"; then
+		$e "==> $fd... yes (cached)"
+		return
+	fi
 	if test 0 = "$ft"; then
-		eval HAVE_$fu=$2
+		fv=$2
-		$e "==> $fd... not checked ($2)"
+		eval HAVE_$fu=$fv
+		test 0 = "$fv" && fv=no
+		test 1 = "$fv" && fv=yes
+		$e "==> $fd... $fv (implied)"
 		return
 	fi
 	$e ... $fd
@@ -190,6 +200,8 @@ if test 0 = $HAVE_MKSH_FULL; then
 		int main(void) { return (0); }
 	EOF
 	test 1 = $HAVE_CAN_FNOINLINE || CFLAGS=$save_CFLAGS
+
+	HAVE_LANGINFO_CODESET=0
 fi
 
 save_CFLAGS=$CFLAGS

edit.c

@@ -5,7 +5,7 @@
 
 #include "sh.h"
 
-__RCSID("$MirOS: src/bin/mksh/edit.c,v 1.70 2006/11/10 06:53:25 tg Exp $");
+__RCSID("$MirOS: src/bin/mksh/edit.c,v 1.71 2006/11/10 07:18:56 tg Exp $");
 
 /* tty driver characters we are interested in */
 typedef struct {
@@ -497,7 +497,7 @@ x_locate_word(const char *buf, int buflen, int pos, int *startp,
 		int iscmd;
 
 		/* Figure out if this is a command */
-		for (p = start - 1; p >= 0 && isspace((unsigned char)buf[p]);
+		for (p = start - 1; p >= 0 && ksh_isspace(buf[p]);
 		    p--)
 			;
 		iscmd = p < 0 || strchr(";|&()`", buf[p]);
@@ -1049,8 +1049,9 @@ struct x_defbindings {
 
 /* Separator for completion */
 #define	is_cfs(c)	((c) == ' ' || (c) == '\t' || (c) == '"' || (c) == '\'')
-#define	is_mfs(c)	(!(isalnum((unsigned char)(c)) || (c) == '_' || \
+			/* Separator for motion */
-			    (c) == '$'))	/* Separator for motion */
+#define	is_mfs(c)	(!(ksh_isdigit(c) || ksh_islower(c) || \
+			    ksh_isupper(c) || (c) == '_' || (c) == '$'))
 
 #define CHARMASK	0xFF			/* 8-bit character mask */
 #define X_NTABS		3			/* normal, meta1, meta2 */
@@ -3209,9 +3210,9 @@ x_fold_case(int c)
 		 */
 		if (cp != xep) {
 			if (c == 'L')		/* lowercase */
-				*cp = _tolower((unsigned char)*cp);
+				*cp = ksh_tolower(*cp);
 			else			/* uppercase, capitalise */
-				*cp = _toupper((unsigned char)*cp);
+				*cp = ksh_toupper(*cp);
 			cp++;
 		}
 		/*
@@ -3219,9 +3220,9 @@ x_fold_case(int c)
 		 */
 		while (cp != xep && !is_mfs(*cp)) {
 			if (c == 'U')		/* uppercase */
-				*cp = _toupper((unsigned char)*cp);
+				*cp = ksh_toupper(*cp);
 			else			/* lowercase, capitalise */
-				*cp = _tolower((unsigned char)*cp);
+				*cp = ksh_tolower(*cp);
 			cp++;
 		}
 	}
@@ -3744,9 +3745,9 @@ vi_hook(int ch)
 			int i;
 			int n = srchlen;
 
-			while (n > 0 && isspace((unsigned char)locpat[n - 1]))
+			while (n > 0 && ksh_isspace(locpat[n - 1]))
 				n--;
-			while (n > 0 && !isspace((unsigned char)locpat[n - 1]))
+			while (n > 0 && !ksh_isspace(locpat[n - 1]))
 				n--;
 			for (i = srchlen; --i >= n; )
 				es->linelen -= char_len((unsigned char)locpat[i]);
@@ -4116,8 +4117,8 @@ vi_cmd(int argcnt, const char *cmd)
 					return -1;
 				if (*cmd == 'c' &&
 				    (cmd[1] == 'w' || cmd[1] == 'W') &&
-				    !isspace((unsigned char)(es->cbuf[es->cursor]))) {
+				    !ksh_isspace(es->cbuf[es->cursor])) {
-					while (isspace((unsigned char)(es->cbuf[--ncursor])))
+					while (ksh_isspace(es->cbuf[--ncursor]))
 						;
 					ncursor++;
 				}
@@ -4368,7 +4369,7 @@ vi_cmd(int argcnt, const char *cmd)
 				if (histnum(-1) < 0)
 					return -1;
 				p = *histpos();
-#define issp(c)		(isspace((unsigned char)(c)) || (c) == '\n')
+#define issp(c)		(ksh_isspace(c) || (c) == '\n')
 				if (argcnt) {
 					while (*p && issp(*p))
 						p++;
@@ -4425,14 +4426,14 @@ vi_cmd(int argcnt, const char *cmd)
 					return -1;
 				for (i = 0; i < argcnt; i++) {
 					p = &es->cbuf[es->cursor];
-					if (ksh_islower((unsigned char)*p)) {
+					if (ksh_islower(*p)) {
 						modified = 1;
 						hnum = hlast;
-						*p = _toupper((unsigned char)*p);
+						*p = ksh_toupper(*p);
-					} else if (ksh_isupper((unsigned char)*p)) {
+					} else if (ksh_isupper(*p)) {
 						modified = 1;
 						hnum = hlast;
-						*p = _tolower((unsigned char)*p);
+						*p = ksh_tolower(*p);
 					}
 					if (es->cursor < es->linelen - 1)
 						es->cursor++;
@@ -4577,7 +4578,7 @@ domove(int argcnt, const char *cmd, int sub)
 	case '^':
 		ncursor = 0;
 		while (ncursor < es->linelen - 1 &&
-		    isspace((unsigned char)(es->cbuf[ncursor])))
+		    ksh_isspace(es->cbuf[ncursor]))
 			ncursor++;
 		break;
 
@@ -4844,12 +4845,12 @@ forwword(int argcnt)
 			while (is_wordch(es->cbuf[ncursor]) &&
 			    ncursor < es->linelen)
 				ncursor++;
-		else if (!isspace((unsigned char)(es->cbuf[ncursor])))
+		else if (!ksh_isspace(es->cbuf[ncursor]))
 			while (!is_wordch(es->cbuf[ncursor]) &&
-			    !isspace((unsigned char)(es->cbuf[ncursor])) &&
+			    !ksh_isspace(es->cbuf[ncursor]) &&
 			    ncursor < es->linelen)
 				ncursor++;
-		while (isspace((unsigned char)(es->cbuf[ncursor])) &&
+		while (ksh_isspace(es->cbuf[ncursor]) &&
 		    ncursor < es->linelen)
 			ncursor++;
 	}
@@ -4863,7 +4864,7 @@ backword(int argcnt)
 
 	ncursor = es->cursor;
 	while (ncursor > 0 && argcnt--) {
-		while (--ncursor > 0 && isspace((unsigned char)(es->cbuf[ncursor])))
+		while (--ncursor > 0 && ksh_isspace(es->cbuf[ncursor]))
 			;
 		if (ncursor > 0) {
 			if (is_wordch(es->cbuf[ncursor]))
@@ -4873,7 +4874,7 @@ backword(int argcnt)
 			else
 				while (--ncursor >= 0 &&
 				    !is_wordch(es->cbuf[ncursor]) &&
-				    !isspace((unsigned char)(es->cbuf[ncursor])))
+				    !ksh_isspace(es->cbuf[ncursor]))
 					;
 			ncursor++;
 		}
@@ -4889,7 +4890,7 @@ endword(int argcnt)
 	ncursor = es->cursor;
 	while (ncursor < es->linelen && argcnt--) {
 		while (++ncursor < es->linelen - 1 &&
-		    isspace((unsigned char)(es->cbuf[ncursor])))
+		    ksh_isspace(es->cbuf[ncursor]))
 			;
 		if (ncursor < es->linelen - 1) {
 			if (is_wordch(es->cbuf[ncursor]))
@@ -4899,7 +4900,7 @@ endword(int argcnt)
 			else
 				while (++ncursor < es->linelen &&
 				    !is_wordch(es->cbuf[ncursor]) &&
-				    !isspace((unsigned char)(es->cbuf[ncursor])))
+				    !ksh_isspace(es->cbuf[ncursor]))
 					;
 			ncursor--;
 		}
@@ -4914,10 +4915,10 @@ Forwword(int argcnt)
 
 	ncursor = es->cursor;
 	while (ncursor < es->linelen && argcnt--) {
-		while (!isspace((unsigned char)(es->cbuf[ncursor])) &&
+		while (!ksh_isspace(es->cbuf[ncursor]) &&
 		    ncursor < es->linelen)
 			ncursor++;
-		while (isspace((unsigned char)(es->cbuf[ncursor])) &&
+		while (ksh_isspace(es->cbuf[ncursor]) &&
 		    ncursor < es->linelen)
 			ncursor++;
 	}
@@ -4932,10 +4933,10 @@ Backword(int argcnt)
 	ncursor = es->cursor;
 	while (ncursor > 0 && argcnt--) {
 		while (--ncursor >= 0 &&
-		    isspace((unsigned char)(es->cbuf[ncursor])))
+		    ksh_isspace(es->cbuf[ncursor]))
 			;
 		while (ncursor >= 0 &&
-		    !isspace((unsigned char)(es->cbuf[ncursor])))
+		    !ksh_isspace(es->cbuf[ncursor]))
 			ncursor--;
 		ncursor++;
 	}
@@ -4950,11 +4951,11 @@ Endword(int argcnt)
 	ncursor = es->cursor;
 	while (ncursor < es->linelen - 1 && argcnt--) {
 		while (++ncursor < es->linelen - 1 &&
-		    isspace((unsigned char)(es->cbuf[ncursor])))
+		    ksh_isspace(es->cbuf[ncursor]))
 			;
 		if (ncursor < es->linelen - 1) {
 			while (++ncursor < es->linelen &&
-			    !isspace((unsigned char)(es->cbuf[ncursor])))
+			    !ksh_isspace(es->cbuf[ncursor]))
 				;
 			ncursor--;
 		}

expr.c

@@ -2,7 +2,7 @@
 
 #include "sh.h"
 
-__RCSID("$MirOS: src/bin/mksh/expr.c,v 1.4 2006/05/10 18:54:10 tg Exp $");
+__RCSID("$MirOS: src/bin/mksh/expr.c,v 1.5 2006/11/10 07:18:57 tg Exp $");
 
 /* The order of these enums is constrained by the order of opinfo[] */
 enum token {
@@ -448,7 +448,7 @@ token(Expr_state *es)
 	char *tvar;
 
 	/* skip white space */
-	for (cp = es->tokp; (c = *cp), isspace(c); cp++)
+	for (cp = es->tokp; (c = *cp), ksh_isspace(c); cp++)
 		;
 	es->tokp = cp;
 

lex.c

@@ -2,7 +2,7 @@
 
 #include "sh.h"
 
-__RCSID("$MirOS: src/bin/mksh/lex.c,v 1.22 2006/11/05 15:31:36 tg Exp $");
+__RCSID("$MirOS: src/bin/mksh/lex.c,v 1.23 2006/11/10 07:18:57 tg Exp $");
 
 /* Structure to keep track of the lexing state and the various pieces of info
  * needed for each particular state. */
@@ -899,7 +899,7 @@ getsc__(void)
 				source->flags |= s->flags & SF_ALIAS;
 				s = source;
 			} else if (*s->u.tblp->val.s &&
-			    isspace((unsigned char)strchr(s->u.tblp->val.s, 0)[-1])) {
+			    ksh_isspace(strchr(s->u.tblp->val.s, 0)[-1])) {
 				source = s = s->next;	/* pop source stack */
 				/* Note that this alias ended with a space,
 				 * enabling alias expansion on the following

misc.c

@@ -3,7 +3,7 @@
 
 #include "sh.h"
 
-__RCSID("$MirOS: src/bin/mksh/misc.c,v 1.36 2006/11/10 06:53:27 tg Exp $\t"
+__RCSID("$MirOS: src/bin/mksh/misc.c,v 1.37 2006/11/10 07:18:57 tg Exp $\t"
 	MKSH_SH_H_ID);
 
 #undef USE_CHVT
@@ -429,7 +429,7 @@ getn(const char *s, int *ai)
 
 	do {
 		c = *s++;
-	} while (isspace(c));
+	} while (ksh_isspace(c));
 	if (c == '-') {
 		neg = true;
 		c = *s++;

sh.h

@@ -8,7 +8,7 @@
 /*	$OpenBSD: c_test.h,v 1.4 2004/12/20 11:34:26 otto Exp $	*/
 /*	$OpenBSD: tty.h,v 1.5 2004/12/20 11:34:26 otto Exp $	*/
 
-#define MKSH_SH_H_ID "$MirOS: src/bin/mksh/sh.h,v 1.78 2006/11/10 06:53:27 tg Exp $"
+#define MKSH_SH_H_ID "$MirOS: src/bin/mksh/sh.h,v 1.79 2006/11/10 07:18:57 tg Exp $"
 #define MKSH_VERSION "R29 2006/11/10"
 
 #if HAVE_SYS_PARAM_H
@@ -40,7 +40,6 @@
 #include <sys/resource.h>
 #include <sys/stat.h>
 #include <sys/wait.h>
-#include <ctype.h>
 #include <dirent.h>
 #include <errno.h>
 #include <fcntl.h>
@@ -105,6 +104,13 @@
 #define ksh_tolower(c)	(((c) >= 'A') && ((c) <= 'Z') ? (c) - 'A' + 'a' : (c))
 #define ksh_toupper(c)	(((c) >= 'a') && ((c) <= 'z') ? (c) - 'a' + 'A' : (c))
 
+/* this macro must not evaluate its arguments several times */
+#define ksh_isspace(c)	__extension__({					\
+		unsigned char ksh_isspace_c = (c);			\
+		(ksh_isspace_c >= 0x09 && ksh_isspace_c <= 0x0D) ||	\
+		    (ksh_isspace_c == 0x20);				\
+	})
+
 #ifndef S_ISTXT
 #define S_ISTXT 0001000
 #endif

var.c

@@ -2,7 +2,7 @@
 
 #include "sh.h"
 
-__RCSID("$MirOS: src/bin/mksh/var.c,v 1.29 2006/11/10 06:53:27 tg Exp $");
+__RCSID("$MirOS: src/bin/mksh/var.c,v 1.30 2006/11/10 07:18:58 tg Exp $");
 
 /*
  * Variables
@@ -505,7 +505,7 @@ formatstr(struct tbl *vp, const char *s)
 		if (vp->flag & RJUST) {
 			const char *qq = s + olen;
 			/* strip trailing spaces (at&t uses qq[-1] == ' ') */
-			while (qq > s && isspace((unsigned char)qq[-1]))
+			while (qq > s && ksh_isspace(qq[-1]))
 				--qq;
 			slen = qq - s;
 			if (slen > vp->u2.field) {
@@ -520,7 +520,7 @@ formatstr(struct tbl *vp, const char *s)
 			    "%.*s", slen, s);
 		} else {
 			/* strip leading spaces/zeros */
-			while (isspace((unsigned char)*s))
+			while (ksh_isspace(*s))
 				s++;
 			if (vp->flag & ZEROFIL)
 				while (*s == '0')
@@ -533,10 +533,10 @@ formatstr(struct tbl *vp, const char *s)
 
 	if (vp->flag & UCASEV_AL) {
 		for (q = p; *q; q++)
-			*q = _toupper((unsigned char)*q);
+			*q = ksh_toupper(*q);
 	} else if (vp->flag & LCASEV) {
 		for (q = p; *q; q++)
-			*q = _tolower((unsigned char)*q);
+			*q = ksh_tolower(*q);
 	}
 
 	return p;