Commit | Line | Data |
---|---|---|
446ee5b1 WD |
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 | ** | |
7a1f46b6 WD |
8 | ** Modified by Wayne Davison to special-case '/' matching, to make '**' |
9 | ** work differently than '*', and to fix the character-class code. | |
446ee5b1 WD |
10 | */ |
11 | ||
12 | #include "rsync.h" | |
13 | ||
14 | /* What character marks an inverted character class? */ | |
e11c4251 WD |
15 | #define NEGATE_CLASS '!' |
16 | #define NEGATE_CLASS2 '^' | |
446ee5b1 | 17 | |
710faea9 WD |
18 | #define FALSE 0 |
19 | #define TRUE 1 | |
20b2e9ce WD |
20 | #define ABORT_ALL -1 |
21 | #define ABORT_TO_STARSTAR -2 | |
22 | ||
5bb92000 WD |
23 | #define CC_EQ(class, len, litmatch) ((len) == sizeof (litmatch)-1 \ |
24 | && *(class) == *(litmatch) \ | |
136c5c5e | 25 | && strncmp((char*)class, litmatch, len) == 0) |
e11c4251 | 26 | |
f2ac84c3 WD |
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 | ||
d5c973cc WD |
56 | #ifdef WILD_TEST_ITERATIONS |
57 | int wildmatch_iteration_count; | |
20b2e9ce | 58 | #endif |
446ee5b1 | 59 | |
8e744636 WD |
60 | static int force_lower_case = 0; |
61 | ||
390621a7 WD |
62 | /* Match pattern "p" against the a virtually-joined string consisting |
63 | * of "text" and any strings in array "a". */ | |
64 | static int dowild(const uchar *p, const uchar *text, const uchar*const *a) | |
446ee5b1 | 65 | { |
8e744636 | 66 | uchar p_ch; |
446ee5b1 | 67 | |
d5c973cc WD |
68 | #ifdef WILD_TEST_ITERATIONS |
69 | wildmatch_iteration_count++; | |
20b2e9ce WD |
70 | #endif |
71 | ||
8e744636 WD |
72 | for ( ; (p_ch = *p) != '\0'; text++, p++) { |
73 | int matched, special; | |
74 | uchar t_ch, prev_ch; | |
390621a7 WD |
75 | while ((t_ch = *text) == '\0') { |
76 | if (*a == NULL) { | |
77 | if (p_ch != '*') | |
78 | return ABORT_ALL; | |
79 | break; | |
80 | } | |
81 | text = *a++; | |
82 | } | |
8e744636 WD |
83 | if (force_lower_case && ISUPPER(t_ch)) |
84 | t_ch = tolower(t_ch); | |
85 | switch (p_ch) { | |
446ee5b1 WD |
86 | case '\\': |
87 | /* Literal match with following character. Note that the test | |
88 | * in "default" handles the p[1] == '\0' failure case. */ | |
8e744636 | 89 | p_ch = *++p; |
446ee5b1 WD |
90 | /* FALLTHROUGH */ |
91 | default: | |
8e744636 | 92 | if (t_ch != p_ch) |
710faea9 | 93 | return FALSE; |
446ee5b1 WD |
94 | continue; |
95 | case '?': | |
96 | /* Match anything but '/'. */ | |
8e744636 | 97 | if (t_ch == '/') |
710faea9 | 98 | return FALSE; |
446ee5b1 WD |
99 | continue; |
100 | case '*': | |
101 | if (*++p == '*') { | |
102 | while (*++p == '*') {} | |
710faea9 | 103 | special = TRUE; |
e725abcf | 104 | } else |
710faea9 | 105 | special = FALSE; |
446ee5b1 | 106 | if (*p == '\0') { |
710faea9 WD |
107 | /* Trailing "**" matches everything. Trailing "*" matches |
108 | * only if there are no more slash characters. */ | |
390621a7 WD |
109 | if (!special) { |
110 | do { | |
111 | if (strchr((char*)text, '/') != NULL) | |
112 | return FALSE; | |
113 | } while ((text = *a++) != NULL); | |
114 | } | |
115 | return TRUE; | |
446ee5b1 | 116 | } |
390621a7 WD |
117 | while (1) { |
118 | if (t_ch == '\0') { | |
119 | if ((text = *a++) == NULL) | |
120 | break; | |
121 | t_ch = *text; | |
122 | continue; | |
123 | } | |
124 | if ((matched = dowild(p, text, a)) != FALSE) { | |
20b2e9ce WD |
125 | if (!special || matched != ABORT_TO_STARSTAR) |
126 | return matched; | |
8e744636 | 127 | } else if (!special && t_ch == '/') |
20b2e9ce | 128 | return ABORT_TO_STARSTAR; |
390621a7 | 129 | t_ch = *++text; |
446ee5b1 | 130 | } |
20b2e9ce | 131 | return ABORT_ALL; |
446ee5b1 | 132 | case '[': |
8e744636 | 133 | p_ch = *++p; |
e11c4251 | 134 | #ifdef NEGATE_CLASS2 |
8e744636 WD |
135 | if (p_ch == NEGATE_CLASS2) |
136 | p_ch = NEGATE_CLASS; | |
e11c4251 | 137 | #endif |
c9a59880 | 138 | /* Assign literal TRUE/FALSE because of "matched" comparison. */ |
8e744636 | 139 | special = p_ch == NEGATE_CLASS? TRUE : FALSE; |
446ee5b1 WD |
140 | if (special) { |
141 | /* Inverted character class. */ | |
8e744636 | 142 | p_ch = *++p; |
446ee5b1 | 143 | } |
8e744636 | 144 | prev_ch = 0; |
710faea9 | 145 | matched = FALSE; |
c9a59880 | 146 | do { |
8e744636 | 147 | if (!p_ch) |
5bb92000 | 148 | return ABORT_ALL; |
8e744636 WD |
149 | if (p_ch == '\\') { |
150 | p_ch = *++p; | |
151 | if (!p_ch) | |
5bb92000 | 152 | return ABORT_ALL; |
8e744636 | 153 | if (t_ch == p_ch) |
e11c4251 | 154 | matched = TRUE; |
8e744636 WD |
155 | } else if (p_ch == '-' && prev_ch && p[1] && p[1] != ']') { |
156 | p_ch = *++p; | |
157 | if (p_ch == '\\') { | |
158 | p_ch = *++p; | |
159 | if (!p_ch) | |
5bb92000 | 160 | return ABORT_ALL; |
e11c4251 | 161 | } |
8e744636 | 162 | if (t_ch <= p_ch && t_ch >= prev_ch) |
e11c4251 | 163 | matched = TRUE; |
8e744636 WD |
164 | p_ch = 0; /* This makes "prev_ch" get set to 0. */ |
165 | } else if (p_ch == '[' && p[1] == ':') { | |
5f238db2 | 166 | const uchar *s; |
e11c4251 | 167 | int i; |
8e744636 WD |
168 | for (s = p += 2; (p_ch = *p) && p_ch != ']'; p++) {} |
169 | if (!p_ch) | |
5bb92000 | 170 | return ABORT_ALL; |
fc96552d WD |
171 | i = p - s - 1; |
172 | if (i < 0 || p[-1] != ':') { | |
173 | /* Didn't find ":]", so treat like a normal set. */ | |
174 | p = s - 2; | |
8e744636 WD |
175 | p_ch = '['; |
176 | if (t_ch == p_ch) | |
fc96552d WD |
177 | matched = TRUE; |
178 | continue; | |
179 | } | |
5bb92000 | 180 | if (CC_EQ(s,i, "alnum")) { |
8e744636 | 181 | if (ISALNUM(t_ch)) |
5bb92000 | 182 | matched = TRUE; |
e725abcf | 183 | } else if (CC_EQ(s,i, "alpha")) { |
8e744636 | 184 | if (ISALPHA(t_ch)) |
5bb92000 | 185 | matched = TRUE; |
e725abcf | 186 | } else if (CC_EQ(s,i, "blank")) { |
8e744636 | 187 | if (ISBLANK(t_ch)) |
5bb92000 | 188 | matched = TRUE; |
e725abcf | 189 | } else if (CC_EQ(s,i, "cntrl")) { |
8e744636 | 190 | if (ISCNTRL(t_ch)) |
5bb92000 | 191 | matched = TRUE; |
e725abcf | 192 | } else if (CC_EQ(s,i, "digit")) { |
8e744636 | 193 | if (ISDIGIT(t_ch)) |
5bb92000 | 194 | matched = TRUE; |
e725abcf | 195 | } else if (CC_EQ(s,i, "graph")) { |
8e744636 | 196 | if (ISGRAPH(t_ch)) |
5bb92000 | 197 | matched = TRUE; |
e725abcf | 198 | } else if (CC_EQ(s,i, "lower")) { |
8e744636 | 199 | if (ISLOWER(t_ch)) |
5bb92000 | 200 | matched = TRUE; |
e725abcf | 201 | } else if (CC_EQ(s,i, "print")) { |
8e744636 | 202 | if (ISPRINT(t_ch)) |
5bb92000 | 203 | matched = TRUE; |
e725abcf | 204 | } else if (CC_EQ(s,i, "punct")) { |
8e744636 | 205 | if (ISPUNCT(t_ch)) |
5bb92000 | 206 | matched = TRUE; |
e725abcf | 207 | } else if (CC_EQ(s,i, "space")) { |
8e744636 | 208 | if (ISSPACE(t_ch)) |
5bb92000 | 209 | matched = TRUE; |
e725abcf | 210 | } else if (CC_EQ(s,i, "upper")) { |
8e744636 | 211 | if (ISUPPER(t_ch)) |
5bb92000 | 212 | matched = TRUE; |
e725abcf | 213 | } else if (CC_EQ(s,i, "xdigit")) { |
8e744636 | 214 | if (ISXDIGIT(t_ch)) |
5bb92000 | 215 | matched = TRUE; |
e725abcf | 216 | } else /* malformed [:class:] string */ |
5bb92000 | 217 | return ABORT_ALL; |
8e744636 WD |
218 | p_ch = 0; /* This makes "prev_ch" get set to 0. */ |
219 | } else if (t_ch == p_ch) | |
710faea9 | 220 | matched = TRUE; |
8e744636 WD |
221 | } while (prev_ch = p_ch, (p_ch = *++p) != ']'); |
222 | if (matched == special || t_ch == '/') | |
710faea9 | 223 | return FALSE; |
446ee5b1 WD |
224 | continue; |
225 | } | |
226 | } | |
227 | ||
390621a7 WD |
228 | do { |
229 | if (*text) | |
230 | return FALSE; | |
231 | } while ((text = *a++) != NULL); | |
232 | ||
233 | return TRUE; | |
234 | } | |
235 | ||
236 | /* Match literal string "s" against the a virtually-joined string consisting | |
237 | * of "text" and any strings in array "a". */ | |
238 | static int doliteral(const uchar *s, const uchar *text, const uchar*const *a) | |
239 | { | |
240 | for ( ; *s != '\0'; text++, s++) { | |
241 | while (*text == '\0') { | |
242 | if ((text = *a++) == NULL) | |
243 | return FALSE; | |
244 | } | |
245 | if (*text != *s) | |
246 | return FALSE; | |
247 | } | |
248 | ||
249 | do { | |
250 | if (*text) | |
251 | return FALSE; | |
252 | } while ((text = *a++) != NULL); | |
253 | ||
254 | return TRUE; | |
255 | } | |
256 | ||
257 | /* Return the last "count" path elements from the concatenated string. | |
258 | * We return a string pointer to the start of the string, and update the | |
259 | * array pointer-pointer to point to any remaining string elements. */ | |
260 | static const uchar *trailing_N_elements(const uchar*const **a_ptr, int count) | |
261 | { | |
262 | const uchar*const *a = *a_ptr; | |
263 | const uchar*const *first_a = a; | |
264 | ||
265 | while (*a) | |
266 | a++; | |
267 | ||
268 | while (a != first_a) { | |
269 | const uchar *s = *--a; | |
270 | s += strlen((char*)s); | |
271 | while (--s >= *a) { | |
272 | if (*s == '/' && !--count) { | |
273 | *a_ptr = a+1; | |
274 | return s+1; | |
275 | } | |
276 | } | |
277 | } | |
278 | ||
279 | if (count == 1) { | |
280 | *a_ptr = a+1; | |
281 | return *a; | |
282 | } | |
283 | ||
284 | return NULL; | |
446ee5b1 | 285 | } |
710faea9 | 286 | |
e725abcf WD |
287 | /* Match the "pattern" against the "text" string. */ |
288 | int wildmatch(const char *pattern, const char *text) | |
710faea9 | 289 | { |
390621a7 | 290 | static const uchar *nomore[1]; /* A NULL pointer. */ |
d5c973cc WD |
291 | #ifdef WILD_TEST_ITERATIONS |
292 | wildmatch_iteration_count = 0; | |
20b2e9ce | 293 | #endif |
390621a7 | 294 | return dowild((const uchar*)pattern, (const uchar*)text, nomore) == TRUE; |
710faea9 | 295 | } |
8e744636 WD |
296 | |
297 | /* Match the "pattern" against the forced-to-lower-case "text" string. */ | |
298 | int iwildmatch(const char *pattern, const char *text) | |
299 | { | |
390621a7 | 300 | static const uchar *nomore[1]; /* A NULL pointer. */ |
8e744636 WD |
301 | int ret; |
302 | #ifdef WILD_TEST_ITERATIONS | |
303 | wildmatch_iteration_count = 0; | |
304 | #endif | |
305 | force_lower_case = 1; | |
390621a7 | 306 | ret = dowild((const uchar*)pattern, (const uchar*)text, nomore) == TRUE; |
8e744636 WD |
307 | force_lower_case = 0; |
308 | return ret; | |
309 | } | |
390621a7 WD |
310 | |
311 | /* Match pattern "p" against the a virtually-joined string consisting | |
312 | * of all the pointers in array "texts" (which has a NULL pointer at the | |
313 | * end). The int "where" can be 0 (normal matching), > 0 (match only | |
314 | * the trailing N slash-separated filename components of "texts"), or < 0 | |
315 | * (match the "pattern" at the start or after any slash in "texts"). */ | |
316 | int wildmatch_array(const char *pattern, const char*const *texts, int where) | |
317 | { | |
318 | const uchar *p = (const uchar*)pattern; | |
319 | const uchar*const *a = (const uchar*const*)texts; | |
320 | const uchar *text; | |
321 | int matched; | |
322 | ||
323 | #ifdef WILD_TEST_ITERATIONS | |
324 | wildmatch_iteration_count = 0; | |
325 | #endif | |
326 | ||
327 | if (where > 0) | |
328 | text = trailing_N_elements(&a, where); | |
329 | else | |
330 | text = *a++; | |
331 | if (!text) | |
332 | return FALSE; | |
333 | ||
334 | if ((matched = dowild(p, text, a)) != TRUE && where < 0 | |
335 | && matched != ABORT_ALL) { | |
336 | while (1) { | |
337 | if (*text == '\0') { | |
338 | if ((text = (uchar*)*a++) == NULL) | |
339 | return FALSE; | |
340 | continue; | |
341 | } | |
342 | if (*text++ == '/' && (matched = dowild(p, text, a)) != FALSE | |
343 | && matched != ABORT_TO_STARSTAR) | |
344 | break; | |
345 | } | |
346 | } | |
347 | return matched == TRUE; | |
348 | } | |
349 | ||
350 | /* Match literal string "s" against the a virtually-joined string consisting | |
351 | * of all the pointers in array "texts" (which has a NULL pointer at the | |
352 | * end). The int "where" can be 0 (normal matching), or > 0 (match | |
353 | * only the trailing N slash-separated filename components of "texts"). */ | |
354 | int litmatch_array(const char *string, const char*const *texts, int where) | |
355 | { | |
356 | const uchar *s = (const uchar*)string; | |
357 | const uchar*const *a = (const uchar* const*)texts; | |
358 | const uchar *text; | |
359 | ||
360 | if (where > 0) | |
361 | text = trailing_N_elements(&a, where); | |
362 | else | |
363 | text = *a++; | |
364 | if (!text) | |
365 | return FALSE; | |
366 | ||
367 | return doliteral(s, text, a) == TRUE; | |
368 | } |