more low-hanging fruits for EBCDIC; notes:

• ord() new, From: Daniel Richard G. <skunk@iSKUNK.ORG> ‣ used in some places • (c - '0') → ksh_numdig(c) # may take *x++ argument • (c - 'A') → ksh_numuc(c) # may NOT take *x+= argument ‣ idem for ksh_numlc(c) and 'a' ‣ these need changing for EBCDIC ‣ add testsuite for this • use macros more, they exist already often • use digits_lc[foo] instead of ('0' + foo), especially for letters • caught another ksh_eq case… • also caught a maybe-UB overflow check, but we don’t have TIME_T_MAX ☹
2015-04-29 20:07:35 +00:00
parent 3eb806b72b
commit 609b311919
10 changed files with 90 additions and 74 deletions
--- a/check.t
+++ b/check.t
@@ -1,4 +1,4 @@
-# $MirOS: src/bin/mksh/check.t,v 1.690 2015/04/19 19:18:03 tg Exp $
+# $MirOS: src/bin/mksh/check.t,v 1.691 2015/04/29 20:07:30 tg Exp $
 # -*- mode: sh -*-
 #-
 # Copyright © 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010,
@@ -4545,6 +4545,20 @@ expected-stdout:
 	64
 	64
 ---
+name: integer-base-8
+description:
+	Check that base-36 works (full span)
+stdin:
+	echo 1:$((36#109AZ)).
+	typeset -i36 x=1691675
+	echo 2:$x.
+	typeset -Uui36 x
+	echo 3:$x.
+expected-stdout:
+	1:1691675.
+	2:36#109az.
+	3:36#109AZ.
+---
 name: integer-base-check-flat
 description:
 	Check behaviour does not match POSuX (except if set -o posix),
--- a/edit.c
+++ b/edit.c
@@ -28,7 +28,7 @@

 #ifndef MKSH_NO_CMDLINE_EDITING

-__RCSID("$MirOS: src/bin/mksh/edit.c,v 1.284 2015/04/11 22:09:48 tg Exp $");
+__RCSID("$MirOS: src/bin/mksh/edit.c,v 1.285 2015/04/29 20:07:31 tg Exp $");

 /*
 * in later versions we might use libtermcap for this, but since external
@@ -2966,7 +2966,7 @@ x_set_arg(int c)
 	/* strip command prefix */
 	c &= 255;
 	while (c >= 0 && ksh_isdigit(c)) {
-		n = n * 10 + (c - '0');
+		n = n * 10 + ksh_numdig(c);
 		if (n > LINE)
 			/* upper bound for repeat */
 			goto x_set_arg_too_big;
@@ -3627,8 +3627,8 @@ vi_hook(int ch)
 				return (1);
 			cmdlen = 0;
 			argc1 = 0;
-			if (ch >= '1' && ch <= '9') {
-				argc1 = ch - '0';
+			if (ch >= ord('1') && ch <= ord('9')) {
+				argc1 = ksh_numdig(ch);
 				state = VARG1;
 			} else {
 				curcmd[cmdlen++] = ch;
@@ -3672,7 +3672,7 @@ vi_hook(int ch)

 	case VARG1:
 		if (ksh_isdigit(ch))
-			argc1 = argc1 * 10 + ch - '0';
+			argc1 = argc1 * 10 + ksh_numdig(ch);
 		else {
 			curcmd[cmdlen++] = ch;
 			state = nextstate(ch);
@@ -3681,8 +3681,8 @@ vi_hook(int ch)

 	case VEXTCMD:
 		argc2 = 0;
-		if (ch >= '1' && ch <= '9') {
-			argc2 = ch - '0';
+		if (ch >= ord('1') && ch <= ord('9')) {
+			argc2 = ksh_numdig(ch);
 			state = VARG2;
 			return (0);
 		} else {
@@ -3698,7 +3698,7 @@ vi_hook(int ch)

 	case VARG2:
 		if (ksh_isdigit(ch))
-			argc2 = argc2 * 10 + ch - '0';
+			argc2 = argc2 * 10 + ksh_numdig(ch);
 		else {
 			if (argc1 == 0)
 				argc1 = argc2;
--- a/funcs.c
+++ b/funcs.c
@@ -38,7 +38,7 @@
 #endif
 #endif

-__RCSID("$MirOS: src/bin/mksh/funcs.c,v 1.269 2015/04/29 18:32:43 tg Exp $");
+__RCSID("$MirOS: src/bin/mksh/funcs.c,v 1.270 2015/04/29 20:07:32 tg Exp $");

 #if HAVE_KILLPG
 /*
@@ -1666,8 +1666,11 @@ c_umask(const char **wp)
 		mode_t new_umask;

 		if (ksh_isdigit(*cp)) {
-			for (new_umask = 0; *cp >= '0' && *cp <= '7'; cp++)
-				new_umask = new_umask * 8 + (*cp - '0');
+			new_umask = 0;
+			while (*cp >= ord('0') && *cp <= ord('7')) {
+				new_umask = new_umask * 8 + ksh_numdig(*cp);
+				++cp;
+			}
 			if (*cp) {
 				bi_errorf("bad number");
 				return (1);
--- a/lex.c
+++ b/lex.c
@@ -23,7 +23,7 @@

 #include "sh.h"

-__RCSID("$MirOS: src/bin/mksh/lex.c,v 1.200 2015/04/19 18:50:36 tg Exp $");
+__RCSID("$MirOS: src/bin/mksh/lex.c,v 1.201 2015/04/29 20:07:33 tg Exp $");

 /*
 * states while lexing word
@@ -920,7 +920,7 @@ yylex(int cf)
 				goto no_iop;
 			if (!ksh_isdigit(dp[c2 + 1]))
 				goto no_iop;
-			iop->unit = (iop->unit * 10) + dp[c2 + 1] - '0';
+			iop->unit = iop->unit * 10 + ksh_numdig(dp[c2 + 1]);
 			if (iop->unit >= FDBASE)
 				goto no_iop;
 		}
--- a/main.c
+++ b/main.c
@@ -34,7 +34,7 @@
 #include <locale.h>
 #endif

-__RCSID("$MirOS: src/bin/mksh/main.c,v 1.292 2015/04/19 18:50:37 tg Exp $");
+__RCSID("$MirOS: src/bin/mksh/main.c,v 1.293 2015/04/29 20:07:33 tg Exp $");

 extern char **environ;

@@ -1459,7 +1459,7 @@ check_fd(const char *name, int mode, const char **emsgp)
 	if (name[0] == 'p' && !name[1])
 		return (coproc_getfd(mode, emsgp));
 	while (ksh_isdigit(*name)) {
-		fd = (fd * 10) + *name - '0';
+		fd = fd * 10 + ksh_numdig(*name);
 		if (fd >= FDBASE) {
 			if (emsgp)
 				*emsgp = "file descriptor too large";
@@ -1613,28 +1613,20 @@ maketemp(Area *ap, Temp_type type, struct temp **tlist)
 	memcpy(cp, "/shXXXXXX.tmp", 14);
 	/* point to the first of six Xes */
 	cp += 3;
-	/* generate random part of filename */
-	len = -1;
-	do {
-		i = rndget() % 36;
-		cp[++len] = i < 26 ? 'a' + i : '0' + i - 26;
-	} while (len < 5);

 	/* cyclically attempt to open a temporary file */
-	while ((i = open(tp->tffn, O_CREAT | O_EXCL | O_RDWR | O_BINARY,
-	    0600)) < 0) {
-		if (errno != EEXIST)
+	do {
+		/* generate random part of filename */
+		len = 0;
+		do {
+			cp[len++] = digits_lc[rndget() % 36];
+		} while (len < 6);
+
+		/* check if this one works */
+		if ((i = open(tp->tffn, O_CREAT | O_EXCL | O_RDWR | O_BINARY,
+		    0600)) < 0 && errno != EEXIST)
 			goto maketemp_out;
-		/* count down from z to a then from 9 to 0 */
-		while (cp[len] == '0')
-			if (!len--)
-				goto maketemp_out;
-		if (cp[len] == 'a')
-			cp[len] = '9';
-		else
-			--cp[len];
-		/* do another cycle */
-	}
+	} while (i < 0);

 	if (type == TT_FUNSUB) {
 		/* map us high and mark as close-on-exec */
--- a/misc.c
+++ b/misc.c
@@ -30,7 +30,7 @@
 #include <grp.h>
 #endif

-__RCSID("$MirOS: src/bin/mksh/misc.c,v 1.228 2015/04/29 18:38:52 tg Exp $");
+__RCSID("$MirOS: src/bin/mksh/misc.c,v 1.229 2015/04/29 20:07:34 tg Exp $");

 #define KSH_CHVT_FLAG
 #ifdef MKSH_SMALL
@@ -95,10 +95,8 @@ initctypes(void)
 {
 	int c;

-	for (c = 'a'; c <= 'z'; c++)
-		chtypes[c] |= C_ALPHA;
-	for (c = 'A'; c <= 'Z'; c++)
-		chtypes[c] |= C_ALPHA;
+	setctypes(digits_uc, C_ALPHA);
+	setctypes(digits_lc, C_ALPHA);
 	chtypes['_'] |= C_ALPHA;
 	setctypes("0123456789", C_DIGIT);
 	/* \0 added automatically */
@@ -545,7 +543,7 @@ getn(const char *s, int *ai)
 		if (num.u > 214748364U)
 			/* overflow on multiplication */
 			return (0);
-		num.u = num.u * 10U + (unsigned int)(c - '0');
+		num.u = num.u * 10U + (unsigned int)ksh_numdig(c);
 		/* now: num.u <= 2147483649U */
 	} while ((c = *s++));

@@ -2194,8 +2192,8 @@ unbksl(bool cstyle, int (*fg)(void), void (*fp)(int))
 		wc = 0;
 		i = 3;
 		while (i--)
-			if ((c = (*fg)()) >= '0' && c <= '7')
-				wc = (wc << 3) + (c - '0');
+			if ((c = (*fg)()) >= ord('0') && c <= ord('7'))
+				wc = (wc << 3) + ksh_numdig(c);
 			else {
 				(*fp)(c);
 				break;
@@ -2221,12 +2219,12 @@ unbksl(bool cstyle, int (*fg)(void), void (*fp)(int))
 		wc = 0;
 		while (i--) {
 			wc <<= 4;
-			if ((c = (*fg)()) >= '0' && c <= '9')
-				wc += c - '0';
-			else if (c >= 'A' && c <= 'F')
-				wc += c - 'A' + 10;
-			else if (c >= 'a' && c <= 'f')
-				wc += c - 'a' + 10;
+			if ((c = (*fg)()) >= ord('0') && c <= ord('9'))
+				wc += ksh_numdig(c);
+			else if (c >= ord('A') && c <= ord('F'))
+				wc += ksh_numuc(c) + 10;
+			else if (c >= ord('a') && c <= ord('f'))
+				wc += ksh_numlc(c) + 10;
 			else {
 				wc >>= 4;
 				(*fp)(c);
--- a/sh.h
+++ b/sh.h
@@ -169,7 +169,7 @@
 #endif

 #ifdef EXTERN
-__RCSID("$MirOS: src/bin/mksh/sh.h,v 1.727 2015/04/29 19:11:57 tg Exp $");
+__RCSID("$MirOS: src/bin/mksh/sh.h,v 1.728 2015/04/29 20:07:34 tg Exp $");
 #endif
 #define MKSH_VERSION "R51 2015/04/19"

@@ -466,6 +466,8 @@ EXTERN const char initvsn[] E_INIT("KSH_VERSION=@(#)" KSH_VERSIONNAME \

 EXTERN const char digits_uc[] E_INIT("0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ");
 EXTERN const char digits_lc[] E_INIT("0123456789abcdefghijklmnopqrstuvwxyz");
+#define letters_uc (digits_uc + 10)
+#define letters_lc (digits_lc + 10)

 /*
 * Evil hack for const correctness due to API brokenness
@@ -924,16 +926,20 @@ extern unsigned char chtypes[];
 #define ctype(c, t)	tobool( ((t) == C_SUBOP2) ?			\
 			    (((c) == '#' || (c) == '%') ? 1 : 0) :	\
 			    (chtypes[(unsigned char)(c)] & (t)) )
+#define ord(c)		((int)(unsigned char)(c))
 #define ksh_isalphx(c)	ctype((c), C_ALPHA)
 #define ksh_isalnux(c)	ctype((c), C_ALPHA | C_DIGIT)
 #define ksh_isdigit(c)	(((c) >= '0') && ((c) <= '9'))
 #define ksh_islower(c)	(((c) >= 'a') && ((c) <= 'z'))
 #define ksh_isupper(c)	(((c) >= 'A') && ((c) <= 'Z'))
-#define ksh_tolower(c)	(((c) >= 'A') && ((c) <= 'Z') ? (c) - 'A' + 'a' : (c))
-#define ksh_toupper(c)	(((c) >= 'a') && ((c) <= 'z') ? (c) - 'a' + 'A' : (c))
+#define ksh_tolower(c)	(ksh_isupper(c) ? (c) - 'A' + 'a' : (c))
+#define ksh_toupper(c)	(ksh_islower(c) ? (c) - 'a' + 'A' : (c))
 #define ksh_isdash(s)	(((s)[0] == '-') && ((s)[1] == '\0'))
 #define ksh_isspace(c)	((((c) >= 0x09) && ((c) <= 0x0D)) || ((c) == 0x20))
 #define ksh_eq(c,u,l)	(((c) | 0x20) == (l))
+#define ksh_numdig(c)	((c) - ord('0'))
+#define ksh_numuc(c)	((c) - ord('A'))
+#define ksh_numlc(c)	((c) - ord('a'))

 EXTERN int ifs0 E_INIT(' ');	/* for "$*" */

--- a/shf.c
+++ b/shf.c
@@ -25,7 +25,7 @@

 #include "sh.h"

-__RCSID("$MirOS: src/bin/mksh/shf.c,v 1.64 2015/02/06 10:09:07 tg Exp $");
+__RCSID("$MirOS: src/bin/mksh/shf.c,v 1.65 2015/04/29 20:07:35 tg Exp $");

 /* flags to shf_emptybuf() */
 #define EB_READSW	0x01	/* about to switch to reading */
@@ -847,11 +847,11 @@ shf_vfprintf(struct shf *shf, const char *fmt, va_list args)
 			if (ksh_isdigit(c)) {
 				bool overflowed = false;

-				tmp = c - '0';
+				tmp = ksh_numdig(c);
 				while (c = *fmt++, ksh_isdigit(c)) {
 					if (notok2mul(2147483647, tmp, 10))
 						overflowed = true;
-					tmp = tmp * 10 + c - '0';
+					tmp = tmp * 10 + ksh_numdig(c);
 				}
 				--fmt;
 				if (overflowed)
@@ -872,7 +872,7 @@ shf_vfprintf(struct shf *shf, const char *fmt, va_list args)
 			/* nasty format */
 			break;

-		if (c >= 'A' && c <= 'Z') {
+		if (ksh_isupper(c)) {
 			flags |= FL_UPPER;
 			c = ksh_tolower(c);
 		}
@@ -917,7 +917,7 @@ shf_vfprintf(struct shf *shf, const char *fmt, va_list args)
 				/* FALLTHROUGH */
 			case 'u':
 				do {
-					*--cp = lnum % 10 + '0';
+					*--cp = digits_lc[lnum % 10];
 					lnum /= 10;
 				} while (lnum);

@@ -933,7 +933,7 @@ shf_vfprintf(struct shf *shf, const char *fmt, va_list args)

 			case 'o':
 				do {
-					*--cp = (lnum & 0x7) + '0';
+					*--cp = digits_lc[lnum & 0x7];
 					lnum >>= 3;
 				} while (lnum);

@@ -945,7 +945,7 @@ shf_vfprintf(struct shf *shf, const char *fmt, va_list args)
 				const char *digits = (flags & FL_UPPER) ?
 				    digits_uc : digits_lc;
 				do {
-					*--cp = digits[lnum & 0xf];
+					*--cp = digits[lnum & 0xF];
 					lnum >>= 4;
 				} while (lnum);

@@ -1013,7 +1013,7 @@ shf_vfprintf(struct shf *shf, const char *fmt, va_list args)
 						s++;
 						nwritten++;
 						if (--precision > 0 &&
-						    (*s | 0x20) == 'x') {
+						    ksh_eq(*s, 'X', 'x')) {
 							shf_putc(*s, shf);
 							s++;
 							precision--;
@@ -1025,8 +1025,10 @@ shf_vfprintf(struct shf *shf, const char *fmt, va_list args)
 					c = flags & FL_ZERO ? '0' : ' ';
 				if (field < 0) {
 					nwritten += -field;
-					for ( ; field < 0 ; field++)
+					while (field < 0) {
 						shf_putc(c, shf);
+						++field;
+					}
 				}
 			} else
 				c = ' ';
--- a/syn.c
+++ b/syn.c
@@ -23,7 +23,7 @@

 #include "sh.h"

-__RCSID("$MirOS: src/bin/mksh/syn.c,v 1.100 2015/04/11 22:03:32 tg Exp $");
+__RCSID("$MirOS: src/bin/mksh/syn.c,v 1.101 2015/04/29 20:07:35 tg Exp $");

 struct nesting_state {
 	int start_token;	/* token than began nesting (eg, FOR) */
@@ -214,10 +214,10 @@ synio(int cf)

 		if (iop->unit > 9) {
 			*cp++ = CHAR;
-			*cp++ = '0' + (iop->unit / 10);
+			*cp++ = digits_lc[iop->unit / 10];
 		}
 		*cp++ = CHAR;
-		*cp++ = '0' + (iop->unit % 10);
+		*cp++ = digits_lc[iop->unit % 10];
 		*cp = EOS;

 		iop->ioflag &= ~IOBASH;
@@ -1074,7 +1074,8 @@ parse_usec(const char *s, struct timeval *tv)
 	tv->tv_sec = 0;
 	/* parse integral part */
 	while (ksh_isdigit(*s)) {
-		tt.tv_sec = tv->tv_sec * 10 + (*s++ - '0');
+		tt.tv_sec = tv->tv_sec * 10 + ksh_numdig(*s++);
+		/*XXX this overflow check maybe UB */
 		if (tt.tv_sec / 10 != tv->tv_sec) {
 			errno = EOVERFLOW;
 			return (true);
@@ -1095,7 +1096,7 @@ parse_usec(const char *s, struct timeval *tv)
 	/* parse decimal fraction */
 	i = 100000;
 	while (ksh_isdigit(*s)) {
-		tv->tv_usec += i * (*s++ - '0');
+		tv->tv_usec += i * ksh_numdig(*s++);
 		if (i == 1)
 			break;
 		i /= 10;
--- a/var.c
+++ b/var.c
@@ -28,7 +28,7 @@
 #include <sys/sysctl.h>
 #endif

-__RCSID("$MirOS: src/bin/mksh/var.c,v 1.191 2015/04/29 18:38:54 tg Exp $");
+__RCSID("$MirOS: src/bin/mksh/var.c,v 1.192 2015/04/29 20:07:35 tg Exp $");

 /*-
 * Variables
@@ -553,11 +553,11 @@ getint(struct tbl *vp, mksh_ari_u *nump, bool arith)
 			continue;
 		}
 		if (ksh_isdigit(c))
-			c -= '0';
+			c = ksh_numdig(c);
 		else if (ksh_isupper(c))
-			c -= 'A' - 10;
+			c = ksh_numuc(c) + 10;
 		else if (ksh_islower(c))
-			c -= 'a' - 10;
+			c = ksh_numlc(c) + 10;
 		else
 			return (-1);
 		if (c >= base)