libbb: robustify isXXXX(). +39 bytes

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

diff --git a/coreutils/sort.c b/coreutils/sort.c
index fad6d124465b27b8b6e8770a80794f5261a0a8e7..e2e7983a1ebd173f1a6b62343cac88c3ce03091a 100644 (file)
--- a/coreutils/sort.c
+++ b/coreutils/sort.c
@@ -128,7 +128,7 @@ static char *get_key(char *str, struct sort_key *key, int flags)
        /* Handle -i */
        if (flags & FLAG_i) {
                for (start = end = 0; str[end]; end++)
-                       if (isprint(str[end]))
+                       if (isprint_asciionly(str[end]))
                                str[start++] = str[end];
                str[start] = '\0';
        }

diff --git a/coreutils/tr.c b/coreutils/tr.c
index d89b80bec104cf069a86dbb10a94f4822a39fb9f..6d4cb4a14f6f9eb578d3e6a837ab78b9365700f4 100644 (file)
--- a/coreutils/tr.c
+++ b/coreutils/tr.c
@@ -153,7 +153,7 @@ static unsigned expand(const char *arg, char **buffer_p)
                                }
                                if (j == CLASS_punct || j == CLASS_cntrl) {
                                        for (i = '\0'; i < ASCII; i++) {
-                                               if ((j == CLASS_punct && isprint(i) && !isalnum(i) && !isspace(i))
+                                               if ((j == CLASS_punct && isprint_asciionly(i) && !isalnum(i) && !isspace(i))
                                                 || (j == CLASS_cntrl && iscntrl(i))
                                                ) {
                                                        buffer[pos++] = i;

diff --git a/editors/diff.c b/editors/diff.c
index e977f4d14cb077cddd6159de9c55dfd3affda00d..745ef0a334be90b25e12a68b0f56b1d783f50ce7 100644 (file)
--- a/editors/diff.c
+++ b/editors/diff.c
@@ -749,7 +749,7 @@ static int asciifile(FILE *f)
        rewind(f);
        cnt = fread(g_read_buf, 1, COMMON_BUFSIZE, f);
        for (i = 0; i < cnt; i++) {
-               if (!isprint(g_read_buf[i])
+               if (!isprint_asciionly(g_read_buf[i])
                 && !isspace(g_read_buf[i])
                ) {
                        return 0;

diff --git a/editors/vi.c b/editors/vi.c
index 6a6722875e48c590bd5b776800c5087ea702b134..7d83db642eeb6e05f7f38478098d76c4dc8c8f6b 100644 (file)
--- a/editors/vi.c
+++ b/editors/vi.c
@@ -30,9 +30,9 @@
 #if ENABLE_LOCALE_SUPPORT
 
 #if ENABLE_FEATURE_VI_8BIT
-#define Isprint(c) isprint(c)
+# define Isprint(c) isprint(c)
 #else
-#define Isprint(c) (isprint(c) && (unsigned char)(c) < 0x7f)
+# define Isprint(c) (isprint(c) && (unsigned char)(c) < 0x7f)
 #endif
 
 #else

diff --git a/include/libbb.h b/include/libbb.h
index d95be5c51eab8452c7b5e8a4ddc15249c8c5073d..77c9e288858a08b72768c42ce53ce13eb97984af 100644 (file)
--- a/include/libbb.h
+++ b/include/libbb.h
@@ -1606,10 +1606,11 @@ extern const char bb_default_login_shell[];
 
 /* We save ~500 bytes on isdigit alone.
  * BTW, x86 likes (unsigned char) cast more than (unsigned). */
-#define isdigit(a) ((unsigned char)((a) - '0') <= 9)
+
+/* These work the same for ASCII and Unicode,
+ * assuming no one asks "is this a *Unicode* letter?" using isalpha(letter) */
 #define isascii(a) ((unsigned char)(a) <= 0x7f)
-#define isgraph(a) ((unsigned char)(a) > ' ')
-#define isprint(a) ((unsigned char)(a) >= ' ')
+#define isdigit(a) ((unsigned char)((a) - '0') <= 9)
 #define isupper(a) ((unsigned char)((a) - 'A') <= ('Z' - 'A'))
 #define islower(a) ((unsigned char)((a) - 'a') <= ('z' - 'a'))
 #define isalpha(a) ((unsigned char)(((a)|0x20) - 'a') <= ('z' - 'a'))
@@ -1619,9 +1620,9 @@ extern const char bb_default_login_shell[];
  * "\t\n\v\f\r" happen to have ASCII codes 9,10,11,12,13.
  */
 #define isspace(a) ({ unsigned char bb__isspace = (a) - 9; bb__isspace == (' ' - 9) || bb__isspace <= (13 - 9); })
-
-// Bigger code:
-//#define isalnum(a) ({ unsigned char bb__isalnum = (a) - '0'; bb__isalnum <= 9 || ((bb__isalnum - ('A' - '0')) & 0xdf) <= 25; })
+// Unsafe wrt NUL: #define ispunct(a) (strchr("!\"#$%&'()*+,-./:;<=>?@[\\]^_`{|}~", (a)) != NULL)
+#define ispunct(a) (strchrnul("!\"#$%&'()*+,-./:;<=>?@[\\]^_`{|}~", (a))[0])
+// Bigger code: #define isalnum(a) ({ unsigned char bb__isalnum = (a) - '0'; bb__isalnum <= 9 || ((bb__isalnum - ('A' - '0')) & 0xdf) <= 25; })
 #define isalnum(a) bb_ascii_isalnum(a)
 static ALWAYS_INLINE int bb_ascii_isalnum(unsigned char a)
 {
@@ -1640,11 +1641,6 @@ static ALWAYS_INLINE int bb_ascii_isxdigit(unsigned char a)
        b = (a|0x20) - 'a';
        return b <= 'f' - 'a';
 }
-
-// Unsafe wrt NUL!
-//#define ispunct(a) (strchr("!\"#$%&'()*+,-./:;<=>?@[\\]^_`{|}~", (a)) != NULL)
-#define ispunct(a) (strchrnul("!\"#$%&'()*+,-./:;<=>?@[\\]^_`{|}~", (a))[0])
-
 #define toupper(a) bb_ascii_toupper(a)
 static ALWAYS_INLINE unsigned char bb_ascii_toupper(unsigned char a)
 {
@@ -1662,6 +1658,14 @@ static ALWAYS_INLINE unsigned char bb_ascii_tolower(unsigned char a)
        return a;
 }
 
+/* In ASCII and Unicode, these are likely to be very different.
+ * Let's prevent ambiguous usage from the start */
+#define isgraph(a) isgraph_is_ambiguous_dont_use(a)
+#define isprint(a) isprint_is_ambiguous_dont_use(a)
+/* NB: must not treat EOF as isgraph or isprint */
+#define isgraph_asciionly(a) ((unsigned)((a) - 0x21) <= 0x7e - 0x21)
+#define isprint_asciionly(a) ((unsigned)((a) - 0x20) <= 0x7e - 0x20)
+
 
 POP_SAVED_FUNCTION_VISIBILITY
 

diff --git a/libbb/dump.c b/libbb/dump.c
index bef485eff3491dda8c4599eb04c5eddaf130f6c3..49e5e26f81e52fc68a2cbcd01623a82e7cfa6da5 100644 (file)
--- a/libbb/dump.c
+++ b/libbb/dump.c
@@ -492,13 +492,13 @@ static void conv_c(PR *pr, unsigned char *p)
                str += 4;
        } while (*str);
 
-       if (isprint(*p)) {
+       if (isprint_asciionly(*p)) {
                *pr->cchar = 'c';
                printf(pr->fmt, *p);
        } else {
                sprintf(buf, "%03o", (int) *p);
                str = buf;
-         strpr:
+ strpr:
                *pr->cchar = 's';
                printf(pr->fmt, str);
        }
@@ -519,7 +519,7 @@ static void conv_u(PR *pr, unsigned char *p)
        } else if (*p == 0x7f) {
                *pr->cchar = 's';
                printf(pr->fmt, "del");
-       } else if (isprint(*p)) {
+       } else if (*p < 0x7f) { /* isprint() */
                *pr->cchar = 'c';
                printf(pr->fmt, *p);
        } else {
@@ -609,7 +609,7 @@ static void display(priv_dumper_t* dumper)
                                                        break;
                                                }
                                                case F_P:
-                                                       printf(pr->fmt, isprint(*bp) ? *bp : '.');
+                                                       printf(pr->fmt, isprint_asciionly(*bp) ? *bp : '.');
                                                        break;
                                                case F_STR:
                                                        printf(pr->fmt, (char *) bp);

diff --git a/loginutils/getty.c b/loginutils/getty.c
index f44d387b63ed6d6fe3db2ead0866d58bcd1040a2..20411b04c92a5f4b0f59569816ab552e54d33065 100644 (file)
--- a/loginutils/getty.c
+++ b/loginutils/getty.c
@@ -487,7 +487,7 @@ static char *get_logname(char *logname, unsigned size_logname,
                        case CTL('D'):
                                exit(EXIT_SUCCESS);
                        default:
-                               if (!isprint(ascval)) {
+                               if (ascval < ' ') {
                                        /* ignore garbage characters */
                                } else if ((int)(bp - logname) >= size_logname - 1) {
                                        bb_error_msg_and_die("%s: input overrun", op->tty);

diff --git a/miscutils/strings.c b/miscutils/strings.c
index fea9edbed57f6ddf5de9f62b1fbf3149f67fa27e..b4c5854cf2525156aeff93c2954afa710e3bcb31 100644 (file)
--- a/miscutils/strings.c
+++ b/miscutils/strings.c
@@ -49,7 +49,7 @@ int strings_main(int argc UNUSED_PARAM, char **argv)
                count = 0;
                do {
                        c = fgetc(file);
-                       if (isprint(c) || c == '\t') {
+                       if (isprint_asciionly(c) || c == '\t') {
                                if (count > n) {
                                        bb_putchar(c);
                                } else {

diff --git a/networking/inetd.c b/networking/inetd.c
index 391bb9ba65fcbfbd736b863b885fd10ac49fd742..a45573396af776ed1d17a6d9e9d07db36a38ed2d 100644 (file)
--- a/networking/inetd.c
+++ b/networking/inetd.c
@@ -1475,9 +1475,8 @@ static void init_ring(void)
        int i;
 
        end_ring = ring;
-       for (i = 0; i <= 128; ++i)
-               if (isprint(i))
-                       *end_ring++ = i;
+       for (i = ' '; i < 127; i++)
+               *end_ring++ = i;
 }
 /* Character generator. MMU arches only. */
 /* ARGSUSED */