wildmatch.c

   1 /*
   2 **  Do shell-style pattern matching for ?, \, [], and * characters.
   3 **  It is 8bit clean.
   4 **
   5 **  Written by Rich $alz, mirror!rs, Wed Nov 26 19:03:17 EST 1986.
   6 **  Rich $alz is now <rsalz@bbn.com>.
   7 **
   8 **  Modified by Wayne Davison to special-case '/' matching, to make '**'
   9 **  work differently than '*', and to fix the character-class code.
  10 */
  11
  12 #include "cache.h"
  13 #include "wildmatch.h"
  14
  15 typedef unsigned char uchar;
  16
  17 /* What character marks an inverted character class? */
  18 #define NEGATE_CLASS    '!'
  19 #define NEGATE_CLASS2   '^'
  20
  21 #define FALSE 0
  22 #define TRUE 1
  23 #define ABORT_ALL -1
  24 #define ABORT_TO_STARSTAR -2
  25
  26 #define CC_EQ(class, len, litmatch) ((len) == sizeof (litmatch)-1 \
  27                                     && *(class) == *(litmatch) \
  28                                     && strncmp((char*)class, litmatch, len) == 0)
  29
  30 #if defined STDC_HEADERS || !defined isascii
  31 # define ISASCII(c) 1
  32 #else
  33 # define ISASCII(c) isascii(c)
  34 #endif
  35
  36 #ifdef isblank
  37 # define ISBLANK(c) (ISASCII(c) && isblank(c))
  38 #else
  39 # define ISBLANK(c) ((c) == ' ' || (c) == '\t')
  40 #endif
  41
  42 #ifdef isgraph
  43 # define ISGRAPH(c) (ISASCII(c) && isgraph(c))
  44 #else
  45 # define ISGRAPH(c) (ISASCII(c) && isprint(c) && !isspace(c))
  46 #endif
  47
  48 #define ISPRINT(c) (ISASCII(c) && isprint(c))
  49 #define ISDIGIT(c) (ISASCII(c) && isdigit(c))
  50 #define ISALNUM(c) (ISASCII(c) && isalnum(c))
  51 #define ISALPHA(c) (ISASCII(c) && isalpha(c))
  52 #define ISCNTRL(c) (ISASCII(c) && iscntrl(c))
  53 #define ISLOWER(c) (ISASCII(c) && islower(c))
  54 #define ISPUNCT(c) (ISASCII(c) && ispunct(c))
  55 #define ISSPACE(c) (ISASCII(c) && isspace(c))
  56 #define ISUPPER(c) (ISASCII(c) && isupper(c))
  57 #define ISXDIGIT(c) (ISASCII(c) && isxdigit(c))
  58
  59 static int force_lower_case = 0;
  60
  61 /* Match pattern "p" against "text" */
  62 static int dowild(const uchar *p, const uchar *text)
  63 {
  64         uchar p_ch;
  65
  66         for ( ; (p_ch = *p) != '\0'; text++, p++) {
  67                 int matched, special;
  68                 uchar t_ch, prev_ch;
  69                 if ((t_ch = *text) == '\0' && p_ch != '*')
  70                         return ABORT_ALL;
  71                 if (force_lower_case && ISUPPER(t_ch))
  72                         t_ch = tolower(t_ch);
  73                 switch (p_ch) {
  74                 case '\\':
  75                         /* Literal match with following character.  Note that the test
  76                          * in "default" handles the p[1] == '\0' failure case. */
  77                         p_ch = *++p;
  78                         /* FALLTHROUGH */
  79                 default:
  80                         if (t_ch != p_ch)
  81                                 return FALSE;
  82                         continue;
  83                 case '?':
  84                         /* Match anything but '/'. */
  85                         if (t_ch == '/')
  86                                 return FALSE;
  87                         continue;
  88                 case '*':
  89                         if (*++p == '*') {
  90                                 while (*++p == '*') {}
  91                                 special = TRUE;
  92                         } else
  93                                 special = FALSE;
  94                         if (*p == '\0') {
  95                                 /* Trailing "**" matches everything.  Trailing "*" matches
  96                                  * only if there are no more slash characters. */
  97                                 if (!special) {
  98                                         if (strchr((char*)text, '/') != NULL)
  99                                                 return FALSE;
 100                                 }
 101                                 return TRUE;
 102                         }
 103                         while (1) {
 104                                 if (t_ch == '\0')
 105                                         break;
 106                                 if ((matched = dowild(p, text)) != FALSE) {
 107                                         if (!special || matched != ABORT_TO_STARSTAR)
 108                                                 return matched;
 109                                 } else if (!special && t_ch == '/')
 110                                         return ABORT_TO_STARSTAR;
 111                                 t_ch = *++text;
 112                         }
 113                         return ABORT_ALL;
 114                 case '[':
 115                         p_ch = *++p;
 116 #ifdef NEGATE_CLASS2
 117                         if (p_ch == NEGATE_CLASS2)
 118                                 p_ch = NEGATE_CLASS;
 119 #endif
 120                         /* Assign literal TRUE/FALSE because of "matched" comparison. */
 121                         special = p_ch == NEGATE_CLASS? TRUE : FALSE;
 122                         if (special) {
 123                                 /* Inverted character class. */
 124                                 p_ch = *++p;
 125                         }
 126                         prev_ch = 0;
 127                         matched = FALSE;
 128                         do {
 129                                 if (!p_ch)
 130                                         return ABORT_ALL;
 131                                 if (p_ch == '\\') {
 132                                         p_ch = *++p;
 133                                         if (!p_ch)
 134                                                 return ABORT_ALL;
 135                                         if (t_ch == p_ch)
 136                                                 matched = TRUE;
 137                                 } else if (p_ch == '-' && prev_ch && p[1] && p[1] != ']') {
 138                                         p_ch = *++p;
 139                                         if (p_ch == '\\') {
 140                                                 p_ch = *++p;
 141                                                 if (!p_ch)
 142                                                         return ABORT_ALL;
 143                                         }
 144                                         if (t_ch <= p_ch && t_ch >= prev_ch)
 145                                                 matched = TRUE;
 146                                         p_ch = 0; /* This makes "prev_ch" get set to 0. */
 147                                 } else if (p_ch == '[' && p[1] == ':') {
 148                                         const uchar *s;
 149                                         int i;
 150                                         for (s = p += 2; (p_ch = *p) && p_ch != ']'; p++) {} /*SHARED ITERATOR*/
 151                                         if (!p_ch)
 152                                                 return ABORT_ALL;
 153                                         i = p - s - 1;
 154                                         if (i < 0 || p[-1] != ':') {
 155                                                 /* Didn't find ":]", so treat like a normal set. */
 156                                                 p = s - 2;
 157                                                 p_ch = '[';
 158                                                 if (t_ch == p_ch)
 159                                                         matched = TRUE;
 160                                                 continue;
 161                                         }
 162                                         if (CC_EQ(s,i, "alnum")) {
 163                                                 if (ISALNUM(t_ch))
 164                                                         matched = TRUE;
 165                                         } else if (CC_EQ(s,i, "alpha")) {
 166                                                 if (ISALPHA(t_ch))
 167                                                         matched = TRUE;
 168                                         } else if (CC_EQ(s,i, "blank")) {
 169                                                 if (ISBLANK(t_ch))
 170                                                         matched = TRUE;
 171                                         } else if (CC_EQ(s,i, "cntrl")) {
 172                                                 if (ISCNTRL(t_ch))
 173                                                         matched = TRUE;
 174                                         } else if (CC_EQ(s,i, "digit")) {
 175                                                 if (ISDIGIT(t_ch))
 176                                                         matched = TRUE;
 177                                         } else if (CC_EQ(s,i, "graph")) {
 178                                                 if (ISGRAPH(t_ch))
 179                                                         matched = TRUE;
 180                                         } else if (CC_EQ(s,i, "lower")) {
 181                                                 if (ISLOWER(t_ch))
 182                                                         matched = TRUE;
 183                                         } else if (CC_EQ(s,i, "print")) {
 184                                                 if (ISPRINT(t_ch))
 185                                                         matched = TRUE;
 186                                         } else if (CC_EQ(s,i, "punct")) {
 187                                                 if (ISPUNCT(t_ch))
 188                                                         matched = TRUE;
 189                                         } else if (CC_EQ(s,i, "space")) {
 190                                                 if (ISSPACE(t_ch))
 191                                                         matched = TRUE;
 192                                         } else if (CC_EQ(s,i, "upper")) {
 193                                                 if (ISUPPER(t_ch))
 194                                                         matched = TRUE;
 195                                         } else if (CC_EQ(s,i, "xdigit")) {
 196                                                 if (ISXDIGIT(t_ch))
 197                                                         matched = TRUE;
 198                                         } else /* malformed [:class:] string */
 199                                                 return ABORT_ALL;
 200                                         p_ch = 0; /* This makes "prev_ch" get set to 0. */
 201                                 } else if (t_ch == p_ch)
 202                                         matched = TRUE;
 203                         } while (prev_ch = p_ch, (p_ch = *++p) != ']');
 204                         if (matched == special || t_ch == '/')
 205                                 return FALSE;
 206                         continue;
 207                 }
 208         }
 209
 210         return *text ? FALSE : TRUE;
 211 }
 212
 213 /* Match the "pattern" against the "text" string. */
 214 int wildmatch(const char *pattern, const char *text)
 215 {
 216         return dowild((const uchar*)pattern, (const uchar*)text) == TRUE;
 217 }
 218
 219 /* Match the "pattern" against the forced-to-lower-case "text" string. */
 220 int iwildmatch(const char *pattern, const char *text)
 221 {
 222         int ret;
 223         force_lower_case = 1;
 224         ret = dowild((const uchar*)pattern, (const uchar*)text) == TRUE;
 225         force_lower_case = 0;
 226         return ret;
 227 }