| 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 "rsync.h" |
| 13 | |
| 14 | /* What character marks an inverted character class? */ |
| 15 | #define NEGATE_CLASS '!' |
| 16 | #define NEGATE_CLASS2 '^' |
| 17 | |
| 18 | #define FALSE 0 |
| 19 | #define TRUE 1 |
| 20 | #define ABORT_ALL -1 |
| 21 | #define ABORT_TO_STARSTAR -2 |
| 22 | |
| 23 | #define CC_EQ(class, len, litmatch) ((len) == sizeof (litmatch)-1 \ |
| 24 | && *(class) == *(litmatch) \ |
| 25 | && strncmp((char*)class, litmatch, len) == 0) |
| 26 | |
| 27 | #if defined STDC_HEADERS || !defined isascii |
| 28 | # define ISASCII(c) 1 |
| 29 | #else |
| 30 | # define ISASCII(c) isascii(c) |
| 31 | #endif |
| 32 | |
| 33 | #ifdef isblank |
| 34 | # define ISBLANK(c) (ISASCII(c) && isblank(c)) |
| 35 | #else |
| 36 | # define ISBLANK(c) ((c) == ' ' || (c) == '\t') |
| 37 | #endif |
| 38 | |
| 39 | #ifdef isgraph |
| 40 | # define ISGRAPH(c) (ISASCII(c) && isgraph(c)) |
| 41 | #else |
| 42 | # define ISGRAPH(c) (ISASCII(c) && isprint(c) && !isspace(c)) |
| 43 | #endif |
| 44 | |
| 45 | #define ISPRINT(c) (ISASCII(c) && isprint(c)) |
| 46 | #define ISDIGIT(c) (ISASCII(c) && isdigit(c)) |
| 47 | #define ISALNUM(c) (ISASCII(c) && isalnum(c)) |
| 48 | #define ISALPHA(c) (ISASCII(c) && isalpha(c)) |
| 49 | #define ISCNTRL(c) (ISASCII(c) && iscntrl(c)) |
| 50 | #define ISLOWER(c) (ISASCII(c) && islower(c)) |
| 51 | #define ISPUNCT(c) (ISASCII(c) && ispunct(c)) |
| 52 | #define ISSPACE(c) (ISASCII(c) && isspace(c)) |
| 53 | #define ISUPPER(c) (ISASCII(c) && isupper(c)) |
| 54 | #define ISXDIGIT(c) (ISASCII(c) && isxdigit(c)) |
| 55 | |
| 56 | #ifdef WILD_TEST_ITERATIONS |
| 57 | int wildmatch_iteration_count; |
| 58 | #endif |
| 59 | |
| 60 | static int force_lower_case = 0; |
| 61 | |
| 62 | /* Match pattern "p" against string "text". */ |
| 63 | static int dowild(const uchar *p, const uchar *text) |
| 64 | { |
| 65 | uchar p_ch; |
| 66 | |
| 67 | #ifdef WILD_TEST_ITERATIONS |
| 68 | wildmatch_iteration_count++; |
| 69 | #endif |
| 70 | |
| 71 | for ( ; (p_ch = *p) != '\0'; text++, p++) { |
| 72 | int matched, special; |
| 73 | uchar t_ch, prev_ch; |
| 74 | if ((t_ch = *text) == '\0' && p_ch != '*') |
| 75 | return ABORT_ALL; |
| 76 | if (force_lower_case && ISUPPER(t_ch)) |
| 77 | t_ch = tolower(t_ch); |
| 78 | switch (p_ch) { |
| 79 | case '\\': |
| 80 | /* Literal match with following character. Note that the test |
| 81 | * in "default" handles the p[1] == '\0' failure case. */ |
| 82 | p_ch = *++p; |
| 83 | /* FALLTHROUGH */ |
| 84 | default: |
| 85 | if (t_ch != p_ch) |
| 86 | return FALSE; |
| 87 | continue; |
| 88 | case '?': |
| 89 | /* Match anything but '/'. */ |
| 90 | if (t_ch == '/') |
| 91 | return FALSE; |
| 92 | continue; |
| 93 | case '*': |
| 94 | if (*++p == '*') { |
| 95 | while (*++p == '*') {} |
| 96 | special = TRUE; |
| 97 | } else |
| 98 | special = FALSE; |
| 99 | if (*p == '\0') { |
| 100 | /* Trailing "**" matches everything. Trailing "*" matches |
| 101 | * only if there are no more slash characters. */ |
| 102 | return special? TRUE : strchr((char*)text, '/') == NULL; |
| 103 | } |
| 104 | for ( ; t_ch; (t_ch = *++text)) { |
| 105 | if ((matched = dowild(p, text)) != FALSE) { |
| 106 | if (!special || matched != ABORT_TO_STARSTAR) |
| 107 | return matched; |
| 108 | } else if (!special && t_ch == '/') |
| 109 | return ABORT_TO_STARSTAR; |
| 110 | } |
| 111 | return ABORT_ALL; |
| 112 | case '[': |
| 113 | p_ch = *++p; |
| 114 | #ifdef NEGATE_CLASS2 |
| 115 | if (p_ch == NEGATE_CLASS2) |
| 116 | p_ch = NEGATE_CLASS; |
| 117 | #endif |
| 118 | /* Assign literal TRUE/FALSE because of "matched" comparison. */ |
| 119 | special = p_ch == NEGATE_CLASS? TRUE : FALSE; |
| 120 | if (special) { |
| 121 | /* Inverted character class. */ |
| 122 | p_ch = *++p; |
| 123 | } |
| 124 | prev_ch = 0; |
| 125 | matched = FALSE; |
| 126 | do { |
| 127 | if (!p_ch) |
| 128 | return ABORT_ALL; |
| 129 | if (p_ch == '\\') { |
| 130 | p_ch = *++p; |
| 131 | if (!p_ch) |
| 132 | return ABORT_ALL; |
| 133 | if (t_ch == p_ch) |
| 134 | matched = TRUE; |
| 135 | } else if (p_ch == '-' && prev_ch && p[1] && p[1] != ']') { |
| 136 | p_ch = *++p; |
| 137 | if (p_ch == '\\') { |
| 138 | p_ch = *++p; |
| 139 | if (!p_ch) |
| 140 | return ABORT_ALL; |
| 141 | } |
| 142 | if (t_ch <= p_ch && t_ch >= prev_ch) |
| 143 | matched = TRUE; |
| 144 | p_ch = 0; /* This makes "prev_ch" get set to 0. */ |
| 145 | } else if (p_ch == '[' && p[1] == ':') { |
| 146 | const uchar *s; |
| 147 | int i; |
| 148 | for (s = p += 2; (p_ch = *p) && p_ch != ']'; p++) {} |
| 149 | if (!p_ch) |
| 150 | return ABORT_ALL; |
| 151 | i = p - s - 1; |
| 152 | if (i < 0 || p[-1] != ':') { |
| 153 | /* Didn't find ":]", so treat like a normal set. */ |
| 154 | p = s - 2; |
| 155 | p_ch = '['; |
| 156 | if (t_ch == p_ch) |
| 157 | matched = TRUE; |
| 158 | continue; |
| 159 | } |
| 160 | if (CC_EQ(s,i, "alnum")) { |
| 161 | if (ISALNUM(t_ch)) |
| 162 | matched = TRUE; |
| 163 | } else if (CC_EQ(s,i, "alpha")) { |
| 164 | if (ISALPHA(t_ch)) |
| 165 | matched = TRUE; |
| 166 | } else if (CC_EQ(s,i, "blank")) { |
| 167 | if (ISBLANK(t_ch)) |
| 168 | matched = TRUE; |
| 169 | } else if (CC_EQ(s,i, "cntrl")) { |
| 170 | if (ISCNTRL(t_ch)) |
| 171 | matched = TRUE; |
| 172 | } else if (CC_EQ(s,i, "digit")) { |
| 173 | if (ISDIGIT(t_ch)) |
| 174 | matched = TRUE; |
| 175 | } else if (CC_EQ(s,i, "graph")) { |
| 176 | if (ISGRAPH(t_ch)) |
| 177 | matched = TRUE; |
| 178 | } else if (CC_EQ(s,i, "lower")) { |
| 179 | if (ISLOWER(t_ch)) |
| 180 | matched = TRUE; |
| 181 | } else if (CC_EQ(s,i, "print")) { |
| 182 | if (ISPRINT(t_ch)) |
| 183 | matched = TRUE; |
| 184 | } else if (CC_EQ(s,i, "punct")) { |
| 185 | if (ISPUNCT(t_ch)) |
| 186 | matched = TRUE; |
| 187 | } else if (CC_EQ(s,i, "space")) { |
| 188 | if (ISSPACE(t_ch)) |
| 189 | matched = TRUE; |
| 190 | } else if (CC_EQ(s,i, "upper")) { |
| 191 | if (ISUPPER(t_ch)) |
| 192 | matched = TRUE; |
| 193 | } else if (CC_EQ(s,i, "xdigit")) { |
| 194 | if (ISXDIGIT(t_ch)) |
| 195 | matched = TRUE; |
| 196 | } else /* malformed [:class:] string */ |
| 197 | return ABORT_ALL; |
| 198 | p_ch = 0; /* This makes "prev_ch" get set to 0. */ |
| 199 | } else if (t_ch == p_ch) |
| 200 | matched = TRUE; |
| 201 | } while (prev_ch = p_ch, (p_ch = *++p) != ']'); |
| 202 | if (matched == special || t_ch == '/') |
| 203 | return FALSE; |
| 204 | continue; |
| 205 | } |
| 206 | } |
| 207 | |
| 208 | return *text == '\0'; |
| 209 | } |
| 210 | |
| 211 | /* Match the "pattern" against the "text" string. */ |
| 212 | int wildmatch(const char *pattern, const char *text) |
| 213 | { |
| 214 | #ifdef WILD_TEST_ITERATIONS |
| 215 | wildmatch_iteration_count = 0; |
| 216 | #endif |
| 217 | return dowild((const uchar*)pattern, (const uchar*)text) == TRUE; |
| 218 | } |
| 219 | |
| 220 | /* Match the "pattern" against the forced-to-lower-case "text" string. */ |
| 221 | int iwildmatch(const char *pattern, const char *text) |
| 222 | { |
| 223 | int ret; |
| 224 | #ifdef WILD_TEST_ITERATIONS |
| 225 | wildmatch_iteration_count = 0; |
| 226 | #endif |
| 227 | force_lower_case = 1; |
| 228 | ret = dowild((const uchar*)pattern, (const uchar*)text) == TRUE; |
| 229 | force_lower_case = 0; |
| 230 | return ret; |
| 231 | } |