🌐 AI搜索 & 代理 主页
blob: 9e9e2a2f955d242cf994c5451c537a7f88539379 [file] [log] [blame]
Nguyễn Thái Ngọc Duy5230f602012-10-15 06:25:521/*
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
Nguyễn Thái Ngọc Duyfeabcc12012-10-15 06:25:5512#include "cache.h"
13#include "wildmatch.h"
14
15typedef unsigned char uchar;
Nguyễn Thái Ngọc Duy5230f602012-10-15 06:25:5216
17/* What character marks an inverted character class? */
18#define NEGATE_CLASS '!'
19#define NEGATE_CLASS2 '^'
20
Nguyễn Thái Ngọc Duy5230f602012-10-15 06:25:5221#define CC_EQ(class, len, litmatch) ((len) == sizeof (litmatch)-1 \
22 && *(class) == *(litmatch) \
23 && strncmp((char*)class, litmatch, len) == 0)
24
25#if defined STDC_HEADERS || !defined isascii
26# define ISASCII(c) 1
27#else
28# define ISASCII(c) isascii(c)
29#endif
30
31#ifdef isblank
32# define ISBLANK(c) (ISASCII(c) && isblank(c))
33#else
34# define ISBLANK(c) ((c) == ' ' || (c) == '\t')
35#endif
36
37#ifdef isgraph
38# define ISGRAPH(c) (ISASCII(c) && isgraph(c))
39#else
40# define ISGRAPH(c) (ISASCII(c) && isprint(c) && !isspace(c))
41#endif
42
43#define ISPRINT(c) (ISASCII(c) && isprint(c))
44#define ISDIGIT(c) (ISASCII(c) && isdigit(c))
45#define ISALNUM(c) (ISASCII(c) && isalnum(c))
46#define ISALPHA(c) (ISASCII(c) && isalpha(c))
47#define ISCNTRL(c) (ISASCII(c) && iscntrl(c))
48#define ISLOWER(c) (ISASCII(c) && islower(c))
49#define ISPUNCT(c) (ISASCII(c) && ispunct(c))
50#define ISSPACE(c) (ISASCII(c) && isspace(c))
51#define ISUPPER(c) (ISASCII(c) && isupper(c))
52#define ISXDIGIT(c) (ISASCII(c) && isxdigit(c))
53
Nguyễn Thái Ngọc Duyb0e02872012-10-15 06:25:5354/* Match pattern "p" against "text" */
Nguyễn Thái Ngọc Duy0c528162013-01-01 02:44:0655static int dowild(const uchar *p, const uchar *text, unsigned int flags)
Nguyễn Thái Ngọc Duy5230f602012-10-15 06:25:5256{
Nguyễn Thái Ngọc Duy327f2f32012-10-15 06:25:5457 uchar p_ch;
Nguyễn Thái Ngọc Duy3a078de2013-01-01 02:44:0258 const uchar *pattern = p;
Nguyễn Thái Ngọc Duy5230f602012-10-15 06:25:5259
Nguyễn Thái Ngọc Duy327f2f32012-10-15 06:25:5460 for ( ; (p_ch = *p) != '\0'; text++, p++) {
Nguyễn Thái Ngọc Duyb6a3d332013-01-01 02:44:0461 int matched, match_slash, negated;
Nguyễn Thái Ngọc Duy327f2f32012-10-15 06:25:5462 uchar t_ch, prev_ch;
63 if ((t_ch = *text) == '\0' && p_ch != '*')
Nguyễn Thái Ngọc Duy9b3497c2013-01-01 02:44:0564 return WM_ABORT_ALL;
Nguyễn Thái Ngọc Duy0c528162013-01-01 02:44:0665 if ((flags & WM_CASEFOLD) && ISUPPER(t_ch))
Nguyễn Thái Ngọc Duy327f2f32012-10-15 06:25:5466 t_ch = tolower(t_ch);
Nguyễn Thái Ngọc Duy0c528162013-01-01 02:44:0667 if ((flags & WM_CASEFOLD) && ISUPPER(p_ch))
Nguyễn Thái Ngọc Duy164bf832012-10-15 06:25:5968 p_ch = tolower(p_ch);
Nguyễn Thái Ngọc Duy327f2f32012-10-15 06:25:5469 switch (p_ch) {
70 case '\\':
71 /* Literal match with following character. Note that the test
72 * in "default" handles the p[1] == '\0' failure case. */
Nguyễn Thái Ngọc Duy5230f602012-10-15 06:25:5273 p_ch = *++p;
Nguyễn Thái Ngọc Duy327f2f32012-10-15 06:25:5474 /* FALLTHROUGH */
75 default:
76 if (t_ch != p_ch)
Nguyễn Thái Ngọc Duy9b3497c2013-01-01 02:44:0577 return WM_NOMATCH;
Nguyễn Thái Ngọc Duy5230f602012-10-15 06:25:5278 continue;
Nguyễn Thái Ngọc Duy327f2f32012-10-15 06:25:5479 case '?':
80 /* Match anything but '/'. */
Nguyễn Thái Ngọc Duyc41244e2013-01-01 02:44:0781 if ((flags & WM_PATHNAME) && t_ch == '/')
Nguyễn Thái Ngọc Duy9b3497c2013-01-01 02:44:0582 return WM_NOMATCH;
Nguyễn Thái Ngọc Duy327f2f32012-10-15 06:25:5483 continue;
84 case '*':
85 if (*++p == '*') {
Nguyễn Thái Ngọc Duy40bbee02012-10-15 06:26:0086 const uchar *prev_p = p - 2;
Nguyễn Thái Ngọc Duy327f2f32012-10-15 06:25:5487 while (*++p == '*') {}
Nguyễn Thái Ngọc Duyc41244e2013-01-01 02:44:0788 if (!(flags & WM_PATHNAME))
89 /* without WM_PATHNAME, '*' == '**' */
90 match_slash = 1;
91 else if ((prev_p < pattern || *prev_p == '/') &&
Nguyễn Thái Ngọc Duy40bbee02012-10-15 06:26:0092 (*p == '\0' || *p == '/' ||
93 (p[0] == '\\' && p[1] == '/'))) {
Nguyễn Thái Ngọc Duy4c251e52012-10-15 06:26:0194 /*
95 * Assuming we already match 'foo/' and are at
96 * <star star slash>, just assume it matches
97 * nothing and go ahead match the rest of the
98 * pattern with the remaining string. This
99 * helps make foo/<*><*>/bar (<> because
100 * otherwise it breaks C comment syntax) match
101 * both foo/bar and foo/a/bar.
102 */
103 if (p[0] == '/' &&
Nguyễn Thái Ngọc Duy0c528162013-01-01 02:44:06104 dowild(p + 1, text, flags) == WM_MATCH)
Nguyễn Thái Ngọc Duy9b3497c2013-01-01 02:44:05105 return WM_MATCH;
106 match_slash = 1;
Nguyễn Thái Ngọc Duye5bbe092018-10-27 08:48:23107 } else /* WM_PATHNAME is set */
108 match_slash = 0;
Nguyễn Thái Ngọc Duy327f2f32012-10-15 06:25:54109 } else
Nguyễn Thái Ngọc Duyc41244e2013-01-01 02:44:07110 /* without WM_PATHNAME, '*' == '**' */
111 match_slash = flags & WM_PATHNAME ? 0 : 1;
Nguyễn Thái Ngọc Duy327f2f32012-10-15 06:25:54112 if (*p == '\0') {
113 /* Trailing "**" matches everything. Trailing "*" matches
114 * only if there are no more slash characters. */
Nguyễn Thái Ngọc Duyb6a3d332013-01-01 02:44:04115 if (!match_slash) {
Nguyễn Thái Ngọc Duy327f2f32012-10-15 06:25:54116 if (strchr((char*)text, '/') != NULL)
Nguyễn Thái Ngọc Duy9b3497c2013-01-01 02:44:05117 return WM_NOMATCH;
Nguyễn Thái Ngọc Duy327f2f32012-10-15 06:25:54118 }
Nguyễn Thái Ngọc Duy9b3497c2013-01-01 02:44:05119 return WM_MATCH;
Nguyễn Thái Ngọc Duy46983442013-01-01 02:44:09120 } else if (!match_slash && *p == '/') {
121 /*
122 * _one_ asterisk followed by a slash
123 * with WM_PATHNAME matches the next
124 * directory
125 */
126 const char *slash = strchr((char*)text, '/');
127 if (!slash)
128 return WM_NOMATCH;
129 text = (const uchar*)slash;
130 /* the slash is consumed by the top-level for loop */
131 break;
Nguyễn Thái Ngọc Duy327f2f32012-10-15 06:25:54132 }
133 while (1) {
134 if (t_ch == '\0')
135 break;
Nguyễn Thái Ngọc Duy6f1a31f2013-01-01 02:44:10136 /*
137 * Try to advance faster when an asterisk is
138 * followed by a literal. We know in this case
Li Peng832c0e52016-05-06 12:36:46139 * that the string before the literal
Nguyễn Thái Ngọc Duy6f1a31f2013-01-01 02:44:10140 * must belong to "*".
141 * If match_slash is false, do not look past
142 * the first slash as it cannot belong to '*'.
143 */
144 if (!is_glob_special(*p)) {
145 p_ch = *p;
146 if ((flags & WM_CASEFOLD) && ISUPPER(p_ch))
147 p_ch = tolower(p_ch);
148 while ((t_ch = *text) != '\0' &&
149 (match_slash || t_ch != '/')) {
150 if ((flags & WM_CASEFOLD) && ISUPPER(t_ch))
151 t_ch = tolower(t_ch);
152 if (t_ch == p_ch)
153 break;
154 text++;
155 }
156 if (t_ch != p_ch)
157 return WM_NOMATCH;
158 }
Nguyễn Thái Ngọc Duy0c528162013-01-01 02:44:06159 if ((matched = dowild(p, text, flags)) != WM_NOMATCH) {
Nguyễn Thái Ngọc Duy9b3497c2013-01-01 02:44:05160 if (!match_slash || matched != WM_ABORT_TO_STARSTAR)
Nguyễn Thái Ngọc Duy327f2f32012-10-15 06:25:54161 return matched;
Nguyễn Thái Ngọc Duyb6a3d332013-01-01 02:44:04162 } else if (!match_slash && t_ch == '/')
Nguyễn Thái Ngọc Duy9b3497c2013-01-01 02:44:05163 return WM_ABORT_TO_STARSTAR;
Nguyễn Thái Ngọc Duy327f2f32012-10-15 06:25:54164 t_ch = *++text;
165 }
Nguyễn Thái Ngọc Duy9b3497c2013-01-01 02:44:05166 return WM_ABORT_ALL;
Nguyễn Thái Ngọc Duy327f2f32012-10-15 06:25:54167 case '[':
168 p_ch = *++p;
169#ifdef NEGATE_CLASS2
170 if (p_ch == NEGATE_CLASS2)
171 p_ch = NEGATE_CLASS;
172#endif
Nguyễn Thái Ngọc Duy9b3497c2013-01-01 02:44:05173 /* Assign literal 1/0 because of "matched" comparison. */
174 negated = p_ch == NEGATE_CLASS ? 1 : 0;
Nguyễn Thái Ngọc Duyb6a3d332013-01-01 02:44:04175 if (negated) {
Nguyễn Thái Ngọc Duy327f2f32012-10-15 06:25:54176 /* Inverted character class. */
177 p_ch = *++p;
178 }
179 prev_ch = 0;
Nguyễn Thái Ngọc Duy9b3497c2013-01-01 02:44:05180 matched = 0;
Nguyễn Thái Ngọc Duy327f2f32012-10-15 06:25:54181 do {
182 if (!p_ch)
Nguyễn Thái Ngọc Duy9b3497c2013-01-01 02:44:05183 return WM_ABORT_ALL;
Nguyễn Thái Ngọc Duy327f2f32012-10-15 06:25:54184 if (p_ch == '\\') {
185 p_ch = *++p;
186 if (!p_ch)
Nguyễn Thái Ngọc Duy9b3497c2013-01-01 02:44:05187 return WM_ABORT_ALL;
Nguyễn Thái Ngọc Duy327f2f32012-10-15 06:25:54188 if (t_ch == p_ch)
Nguyễn Thái Ngọc Duy9b3497c2013-01-01 02:44:05189 matched = 1;
Nguyễn Thái Ngọc Duy327f2f32012-10-15 06:25:54190 } else if (p_ch == '-' && prev_ch && p[1] && p[1] != ']') {
191 p_ch = *++p;
192 if (p_ch == '\\') {
193 p_ch = *++p;
194 if (!p_ch)
Nguyễn Thái Ngọc Duy9b3497c2013-01-01 02:44:05195 return WM_ABORT_ALL;
Nguyễn Thái Ngọc Duy327f2f32012-10-15 06:25:54196 }
197 if (t_ch <= p_ch && t_ch >= prev_ch)
Nguyễn Thái Ngọc Duy9b3497c2013-01-01 02:44:05198 matched = 1;
Anthony Ramineb79c0c32013-05-30 10:19:10199 else if ((flags & WM_CASEFOLD) && ISLOWER(t_ch)) {
200 uchar t_ch_upper = toupper(t_ch);
201 if (t_ch_upper <= p_ch && t_ch_upper >= prev_ch)
202 matched = 1;
203 }
Nguyễn Thái Ngọc Duy327f2f32012-10-15 06:25:54204 p_ch = 0; /* This makes "prev_ch" get set to 0. */
205 } else if (p_ch == '[' && p[1] == ':') {
206 const uchar *s;
207 int i;
208 for (s = p += 2; (p_ch = *p) && p_ch != ']'; p++) {} /*SHARED ITERATOR*/
209 if (!p_ch)
Nguyễn Thái Ngọc Duy9b3497c2013-01-01 02:44:05210 return WM_ABORT_ALL;
Nguyễn Thái Ngọc Duy327f2f32012-10-15 06:25:54211 i = p - s - 1;
212 if (i < 0 || p[-1] != ':') {
213 /* Didn't find ":]", so treat like a normal set. */
214 p = s - 2;
215 p_ch = '[';
216 if (t_ch == p_ch)
Nguyễn Thái Ngọc Duy9b3497c2013-01-01 02:44:05217 matched = 1;
Nguyễn Thái Ngọc Duy327f2f32012-10-15 06:25:54218 continue;
219 }
220 if (CC_EQ(s,i, "alnum")) {
221 if (ISALNUM(t_ch))
Nguyễn Thái Ngọc Duy9b3497c2013-01-01 02:44:05222 matched = 1;
Nguyễn Thái Ngọc Duy327f2f32012-10-15 06:25:54223 } else if (CC_EQ(s,i, "alpha")) {
224 if (ISALPHA(t_ch))
Nguyễn Thái Ngọc Duy9b3497c2013-01-01 02:44:05225 matched = 1;
Nguyễn Thái Ngọc Duy327f2f32012-10-15 06:25:54226 } else if (CC_EQ(s,i, "blank")) {
227 if (ISBLANK(t_ch))
Nguyễn Thái Ngọc Duy9b3497c2013-01-01 02:44:05228 matched = 1;
Nguyễn Thái Ngọc Duy327f2f32012-10-15 06:25:54229 } else if (CC_EQ(s,i, "cntrl")) {
230 if (ISCNTRL(t_ch))
Nguyễn Thái Ngọc Duy9b3497c2013-01-01 02:44:05231 matched = 1;
Nguyễn Thái Ngọc Duy327f2f32012-10-15 06:25:54232 } else if (CC_EQ(s,i, "digit")) {
233 if (ISDIGIT(t_ch))
Nguyễn Thái Ngọc Duy9b3497c2013-01-01 02:44:05234 matched = 1;
Nguyễn Thái Ngọc Duy327f2f32012-10-15 06:25:54235 } else if (CC_EQ(s,i, "graph")) {
236 if (ISGRAPH(t_ch))
Nguyễn Thái Ngọc Duy9b3497c2013-01-01 02:44:05237 matched = 1;
Nguyễn Thái Ngọc Duy327f2f32012-10-15 06:25:54238 } else if (CC_EQ(s,i, "lower")) {
239 if (ISLOWER(t_ch))
Nguyễn Thái Ngọc Duy9b3497c2013-01-01 02:44:05240 matched = 1;
Nguyễn Thái Ngọc Duy327f2f32012-10-15 06:25:54241 } else if (CC_EQ(s,i, "print")) {
242 if (ISPRINT(t_ch))
Nguyễn Thái Ngọc Duy9b3497c2013-01-01 02:44:05243 matched = 1;
Nguyễn Thái Ngọc Duy327f2f32012-10-15 06:25:54244 } else if (CC_EQ(s,i, "punct")) {
245 if (ISPUNCT(t_ch))
Nguyễn Thái Ngọc Duy9b3497c2013-01-01 02:44:05246 matched = 1;
Nguyễn Thái Ngọc Duy327f2f32012-10-15 06:25:54247 } else if (CC_EQ(s,i, "space")) {
248 if (ISSPACE(t_ch))
Nguyễn Thái Ngọc Duy9b3497c2013-01-01 02:44:05249 matched = 1;
Nguyễn Thái Ngọc Duy327f2f32012-10-15 06:25:54250 } else if (CC_EQ(s,i, "upper")) {
251 if (ISUPPER(t_ch))
Nguyễn Thái Ngọc Duy9b3497c2013-01-01 02:44:05252 matched = 1;
Anthony Ramineb79c0c32013-05-30 10:19:10253 else if ((flags & WM_CASEFOLD) && ISLOWER(t_ch))
254 matched = 1;
Nguyễn Thái Ngọc Duy327f2f32012-10-15 06:25:54255 } else if (CC_EQ(s,i, "xdigit")) {
256 if (ISXDIGIT(t_ch))
Nguyễn Thái Ngọc Duy9b3497c2013-01-01 02:44:05257 matched = 1;
Nguyễn Thái Ngọc Duy327f2f32012-10-15 06:25:54258 } else /* malformed [:class:] string */
Nguyễn Thái Ngọc Duy9b3497c2013-01-01 02:44:05259 return WM_ABORT_ALL;
Nguyễn Thái Ngọc Duy327f2f32012-10-15 06:25:54260 p_ch = 0; /* This makes "prev_ch" get set to 0. */
261 } else if (t_ch == p_ch)
Nguyễn Thái Ngọc Duy9b3497c2013-01-01 02:44:05262 matched = 1;
Nguyễn Thái Ngọc Duy327f2f32012-10-15 06:25:54263 } while (prev_ch = p_ch, (p_ch = *++p) != ']');
Nguyễn Thái Ngọc Duyc41244e2013-01-01 02:44:07264 if (matched == negated ||
265 ((flags & WM_PATHNAME) && t_ch == '/'))
Nguyễn Thái Ngọc Duy9b3497c2013-01-01 02:44:05266 return WM_NOMATCH;
Nguyễn Thái Ngọc Duy327f2f32012-10-15 06:25:54267 continue;
268 }
Nguyễn Thái Ngọc Duy5230f602012-10-15 06:25:52269 }
Nguyễn Thái Ngọc Duy5230f602012-10-15 06:25:52270
Nguyễn Thái Ngọc Duy9b3497c2013-01-01 02:44:05271 return *text ? WM_NOMATCH : WM_MATCH;
Nguyễn Thái Ngọc Duy5230f602012-10-15 06:25:52272}
273
274/* Match the "pattern" against the "text" string. */
Ævar Arnfjörð Bjarmason55d34262017-06-22 21:38:08275int wildmatch(const char *pattern, const char *text, unsigned int flags)
Nguyễn Thái Ngọc Duy5230f602012-10-15 06:25:52276{
Nguyễn Thái Ngọc Duy0c528162013-01-01 02:44:06277 return dowild((const uchar*)pattern, (const uchar*)text, flags);
Nguyễn Thái Ngọc Duy5230f602012-10-15 06:25:52278}