|
| 1 | +/* |
| 2 | + * Copyright (C) the libgit2 contributors. All rights reserved. |
| 3 | + * |
| 4 | + * This file is part of libgit2, distributed under the GNU GPL v2 with |
| 5 | + * a Linking Exception. For full terms see the included COPYING file. |
| 6 | + * |
| 7 | + * Do shell-style pattern matching for ?, \, [], and * characters. |
| 8 | + * It is 8bit clean. |
| 9 | + * |
| 10 | + * Written by Rich $alz, mirror!rs, Wed Nov 26 19:03:17 EST 1986. |
| 11 | + * Rich $alz is now <rsalz@bbn.com>. |
| 12 | + * |
| 13 | + * Modified by Wayne Davison to special-case '/' matching, to make '**' |
| 14 | + * work differently than '*', and to fix the character-class code. |
| 15 | + * |
| 16 | + * Imported from git.git. |
| 17 | + */ |
| 18 | + |
| 19 | +#include <ctype.h> |
| 20 | +#include <string.h> |
| 21 | +#include "wildmatch.h" |
| 22 | + |
| 23 | +#define GIT_SPACE 0x01 |
| 24 | +#define GIT_DIGIT 0x02 |
| 25 | +#define GIT_ALPHA 0x04 |
| 26 | +#define GIT_GLOB_SPECIAL 0x08 |
| 27 | +#define GIT_REGEX_SPECIAL 0x10 |
| 28 | +#define GIT_PATHSPEC_MAGIC 0x20 |
| 29 | +#define GIT_CNTRL 0x40 |
| 30 | +#define GIT_PUNCT 0x80 |
| 31 | + |
| 32 | +enum { |
| 33 | + S = GIT_SPACE, |
| 34 | + A = GIT_ALPHA, |
| 35 | + D = GIT_DIGIT, |
| 36 | + G = GIT_GLOB_SPECIAL, /* *, ?, [, \\ */ |
| 37 | + R = GIT_REGEX_SPECIAL, /* $, (, ), +, ., ^, {, | */ |
| 38 | + P = GIT_PATHSPEC_MAGIC, /* other non-alnum, except for ] and } */ |
| 39 | + X = GIT_CNTRL, |
| 40 | + U = GIT_PUNCT, |
| 41 | + Z = GIT_CNTRL | GIT_SPACE |
| 42 | +}; |
| 43 | + |
| 44 | +static const unsigned char sane_ctype[256] = { |
| 45 | + X, X, X, X, X, X, X, X, X, Z, Z, X, X, Z, X, X, /* 0.. 15 */ |
| 46 | + X, X, X, X, X, X, X, X, X, X, X, X, X, X, X, X, /* 16.. 31 */ |
| 47 | + S, P, P, P, R, P, P, P, R, R, G, R, P, P, R, P, /* 32.. 47 */ |
| 48 | + D, D, D, D, D, D, D, D, D, D, P, P, P, P, P, G, /* 48.. 63 */ |
| 49 | + P, A, A, A, A, A, A, A, A, A, A, A, A, A, A, A, /* 64.. 79 */ |
| 50 | + A, A, A, A, A, A, A, A, A, A, A, G, G, U, R, P, /* 80.. 95 */ |
| 51 | + P, A, A, A, A, A, A, A, A, A, A, A, A, A, A, A, /* 96..111 */ |
| 52 | + A, A, A, A, A, A, A, A, A, A, A, R, R, U, P, X, /* 112..127 */ |
| 53 | + /* Nothing in the 128.. range */ |
| 54 | +}; |
| 55 | + |
| 56 | +#define sane_istest(x,mask) ((sane_ctype[(unsigned char)(x)] & (mask)) != 0) |
| 57 | +#define is_glob_special(x) sane_istest(x,GIT_GLOB_SPECIAL) |
| 58 | + |
| 59 | +typedef unsigned char uchar; |
| 60 | + |
| 61 | +/* What character marks an inverted character class? */ |
| 62 | +#define NEGATE_CLASS '!' |
| 63 | +#define NEGATE_CLASS2 '^' |
| 64 | + |
| 65 | +#define CC_EQ(class, len, litmatch) ((len) == sizeof (litmatch)-1 \ |
| 66 | + && *(class) == *(litmatch) \ |
| 67 | + && strncmp((char*)class, litmatch, len) == 0) |
| 68 | + |
| 69 | +#if defined STDC_HEADERS || !defined isascii |
| 70 | +# define ISASCII(c) 1 |
| 71 | +#else |
| 72 | +# define ISASCII(c) isascii(c) |
| 73 | +#endif |
| 74 | + |
| 75 | +#ifdef isblank |
| 76 | +# define ISBLANK(c) (ISASCII(c) && isblank(c)) |
| 77 | +#else |
| 78 | +# define ISBLANK(c) ((c) == ' ' || (c) == '\t') |
| 79 | +#endif |
| 80 | + |
| 81 | +#ifdef isgraph |
| 82 | +# define ISGRAPH(c) (ISASCII(c) && isgraph(c)) |
| 83 | +#else |
| 84 | +# define ISGRAPH(c) (ISASCII(c) && isprint(c) && !isspace(c)) |
| 85 | +#endif |
| 86 | + |
| 87 | +#define ISPRINT(c) (ISASCII(c) && isprint(c)) |
| 88 | +#define ISDIGIT(c) (ISASCII(c) && isdigit(c)) |
| 89 | +#define ISALNUM(c) (ISASCII(c) && isalnum(c)) |
| 90 | +#define ISALPHA(c) (ISASCII(c) && isalpha(c)) |
| 91 | +#define ISCNTRL(c) (ISASCII(c) && iscntrl(c)) |
| 92 | +#define ISLOWER(c) (ISASCII(c) && islower(c)) |
| 93 | +#define ISPUNCT(c) (ISASCII(c) && ispunct(c)) |
| 94 | +#define ISSPACE(c) (ISASCII(c) && isspace(c)) |
| 95 | +#define ISUPPER(c) (ISASCII(c) && isupper(c)) |
| 96 | +#define ISXDIGIT(c) (ISASCII(c) && isxdigit(c)) |
| 97 | + |
| 98 | +/* Match pattern "p" against "text" */ |
| 99 | +static int dowild(const uchar *p, const uchar *text, unsigned int flags) |
| 100 | +{ |
| 101 | + uchar p_ch; |
| 102 | + const uchar *pattern = p; |
| 103 | + |
| 104 | + for ( ; (p_ch = *p) != '\0'; text++, p++) { |
| 105 | + int matched, match_slash, negated; |
| 106 | + uchar t_ch, prev_ch; |
| 107 | + if ((t_ch = *text) == '\0' && p_ch != '*') |
| 108 | + return WM_ABORT_ALL; |
| 109 | + if ((flags & WM_CASEFOLD) && ISUPPER(t_ch)) |
| 110 | + t_ch = tolower(t_ch); |
| 111 | + if ((flags & WM_CASEFOLD) && ISUPPER(p_ch)) |
| 112 | + p_ch = tolower(p_ch); |
| 113 | + switch (p_ch) { |
| 114 | + case '\\': |
| 115 | + /* Literal match with following character. Note that the test |
| 116 | + * in "default" handles the p[1] == '\0' failure case. */ |
| 117 | + p_ch = *++p; |
| 118 | + /* FALLTHROUGH */ |
| 119 | + default: |
| 120 | + if (t_ch != p_ch) |
| 121 | + return WM_NOMATCH; |
| 122 | + continue; |
| 123 | + case '?': |
| 124 | + /* Match anything but '/'. */ |
| 125 | + if ((flags & WM_PATHNAME) && t_ch == '/') |
| 126 | + return WM_NOMATCH; |
| 127 | + continue; |
| 128 | + case '*': |
| 129 | + if (*++p == '*') { |
| 130 | + const uchar *prev_p = p - 2; |
| 131 | + while (*++p == '*') {} |
| 132 | + if (!(flags & WM_PATHNAME)) |
| 133 | + /* without WM_PATHNAME, '*' == '**' */ |
| 134 | + match_slash = 1; |
| 135 | + else if ((prev_p < pattern || *prev_p == '/') && |
| 136 | + (*p == '\0' || *p == '/' || |
| 137 | + (p[0] == '\\' && p[1] == '/'))) { |
| 138 | + /* |
| 139 | + * Assuming we already match 'foo/' and are at |
| 140 | + * <star star slash>, just assume it matches |
| 141 | + * nothing and go ahead match the rest of the |
| 142 | + * pattern with the remaining string. This |
| 143 | + * helps make foo/<*><*>/bar (<> because |
| 144 | + * otherwise it breaks C comment syntax) match |
| 145 | + * both foo/bar and foo/a/bar. |
| 146 | + */ |
| 147 | + if (p[0] == '/' && |
| 148 | + dowild(p + 1, text, flags) == WM_MATCH) |
| 149 | + return WM_MATCH; |
| 150 | + match_slash = 1; |
| 151 | + } else /* WM_PATHNAME is set */ |
| 152 | + match_slash = 0; |
| 153 | + } else |
| 154 | + /* without WM_PATHNAME, '*' == '**' */ |
| 155 | + match_slash = flags & WM_PATHNAME ? 0 : 1; |
| 156 | + if (*p == '\0') { |
| 157 | + /* Trailing "**" matches everything. Trailing "*" matches |
| 158 | + * only if there are no more slash characters. */ |
| 159 | + if (!match_slash) { |
| 160 | + if (strchr((char*)text, '/') != NULL) |
| 161 | + return WM_NOMATCH; |
| 162 | + } |
| 163 | + return WM_MATCH; |
| 164 | + } else if (!match_slash && *p == '/') { |
| 165 | + /* |
| 166 | + * _one_ asterisk followed by a slash |
| 167 | + * with WM_PATHNAME matches the next |
| 168 | + * directory |
| 169 | + */ |
| 170 | + const char *slash = strchr((char*)text, '/'); |
| 171 | + if (!slash) |
| 172 | + return WM_NOMATCH; |
| 173 | + text = (const uchar*)slash; |
| 174 | + /* the slash is consumed by the top-level for loop */ |
| 175 | + break; |
| 176 | + } |
| 177 | + while (1) { |
| 178 | + if (t_ch == '\0') |
| 179 | + break; |
| 180 | + /* |
| 181 | + * Try to advance faster when an asterisk is |
| 182 | + * followed by a literal. We know in this case |
| 183 | + * that the string before the literal |
| 184 | + * must belong to "*". |
| 185 | + * If match_slash is false, do not look past |
| 186 | + * the first slash as it cannot belong to '*'. |
| 187 | + */ |
| 188 | + if (!is_glob_special(*p)) { |
| 189 | + p_ch = *p; |
| 190 | + if ((flags & WM_CASEFOLD) && ISUPPER(p_ch)) |
| 191 | + p_ch = tolower(p_ch); |
| 192 | + while ((t_ch = *text) != '\0' && |
| 193 | + (match_slash || t_ch != '/')) { |
| 194 | + if ((flags & WM_CASEFOLD) && ISUPPER(t_ch)) |
| 195 | + t_ch = tolower(t_ch); |
| 196 | + if (t_ch == p_ch) |
| 197 | + break; |
| 198 | + text++; |
| 199 | + } |
| 200 | + if (t_ch != p_ch) |
| 201 | + return WM_NOMATCH; |
| 202 | + } |
| 203 | + if ((matched = dowild(p, text, flags)) != WM_NOMATCH) { |
| 204 | + if (!match_slash || matched != WM_ABORT_TO_STARSTAR) |
| 205 | + return matched; |
| 206 | + } else if (!match_slash && t_ch == '/') |
| 207 | + return WM_ABORT_TO_STARSTAR; |
| 208 | + t_ch = *++text; |
| 209 | + } |
| 210 | + return WM_ABORT_ALL; |
| 211 | + case '[': |
| 212 | + p_ch = *++p; |
| 213 | +#ifdef NEGATE_CLASS2 |
| 214 | + if (p_ch == NEGATE_CLASS2) |
| 215 | + p_ch = NEGATE_CLASS; |
| 216 | +#endif |
| 217 | + /* Assign literal 1/0 because of "matched" comparison. */ |
| 218 | + negated = p_ch == NEGATE_CLASS ? 1 : 0; |
| 219 | + if (negated) { |
| 220 | + /* Inverted character class. */ |
| 221 | + p_ch = *++p; |
| 222 | + } |
| 223 | + prev_ch = 0; |
| 224 | + matched = 0; |
| 225 | + do { |
| 226 | + if (!p_ch) |
| 227 | + return WM_ABORT_ALL; |
| 228 | + if (p_ch == '\\') { |
| 229 | + p_ch = *++p; |
| 230 | + if (!p_ch) |
| 231 | + return WM_ABORT_ALL; |
| 232 | + if (t_ch == p_ch) |
| 233 | + matched = 1; |
| 234 | + } else if (p_ch == '-' && prev_ch && p[1] && p[1] != ']') { |
| 235 | + p_ch = *++p; |
| 236 | + if (p_ch == '\\') { |
| 237 | + p_ch = *++p; |
| 238 | + if (!p_ch) |
| 239 | + return WM_ABORT_ALL; |
| 240 | + } |
| 241 | + if (t_ch <= p_ch && t_ch >= prev_ch) |
| 242 | + matched = 1; |
| 243 | + else if ((flags & WM_CASEFOLD) && ISLOWER(t_ch)) { |
| 244 | + uchar t_ch_upper = toupper(t_ch); |
| 245 | + if (t_ch_upper <= p_ch && t_ch_upper >= prev_ch) |
| 246 | + matched = 1; |
| 247 | + } |
| 248 | + p_ch = 0; /* This makes "prev_ch" get set to 0. */ |
| 249 | + } else if (p_ch == '[' && p[1] == ':') { |
| 250 | + const uchar *s; |
| 251 | + int i; |
| 252 | + for (s = p += 2; (p_ch = *p) && p_ch != ']'; p++) {} /*SHARED ITERATOR*/ |
| 253 | + if (!p_ch) |
| 254 | + return WM_ABORT_ALL; |
| 255 | + i = (int)(p - s - 1); |
| 256 | + if (i < 0 || p[-1] != ':') { |
| 257 | + /* Didn't find ":]", so treat like a normal set. */ |
| 258 | + p = s - 2; |
| 259 | + p_ch = '['; |
| 260 | + if (t_ch == p_ch) |
| 261 | + matched = 1; |
| 262 | + continue; |
| 263 | + } |
| 264 | + if (CC_EQ(s,i, "alnum")) { |
| 265 | + if (ISALNUM(t_ch)) |
| 266 | + matched = 1; |
| 267 | + } else if (CC_EQ(s,i, "alpha")) { |
| 268 | + if (ISALPHA(t_ch)) |
| 269 | + matched = 1; |
| 270 | + } else if (CC_EQ(s,i, "blank")) { |
| 271 | + if (ISBLANK(t_ch)) |
| 272 | + matched = 1; |
| 273 | + } else if (CC_EQ(s,i, "cntrl")) { |
| 274 | + if (ISCNTRL(t_ch)) |
| 275 | + matched = 1; |
| 276 | + } else if (CC_EQ(s,i, "digit")) { |
| 277 | + if (ISDIGIT(t_ch)) |
| 278 | + matched = 1; |
| 279 | + } else if (CC_EQ(s,i, "graph")) { |
| 280 | + if (ISGRAPH(t_ch)) |
| 281 | + matched = 1; |
| 282 | + } else if (CC_EQ(s,i, "lower")) { |
| 283 | + if (ISLOWER(t_ch)) |
| 284 | + matched = 1; |
| 285 | + } else if (CC_EQ(s,i, "print")) { |
| 286 | + if (ISPRINT(t_ch)) |
| 287 | + matched = 1; |
| 288 | + } else if (CC_EQ(s,i, "punct")) { |
| 289 | + if (ISPUNCT(t_ch)) |
| 290 | + matched = 1; |
| 291 | + } else if (CC_EQ(s,i, "space")) { |
| 292 | + if (ISSPACE(t_ch)) |
| 293 | + matched = 1; |
| 294 | + } else if (CC_EQ(s,i, "upper")) { |
| 295 | + if (ISUPPER(t_ch)) |
| 296 | + matched = 1; |
| 297 | + else if ((flags & WM_CASEFOLD) && ISLOWER(t_ch)) |
| 298 | + matched = 1; |
| 299 | + } else if (CC_EQ(s,i, "xdigit")) { |
| 300 | + if (ISXDIGIT(t_ch)) |
| 301 | + matched = 1; |
| 302 | + } else /* malformed [:class:] string */ |
| 303 | + return WM_ABORT_ALL; |
| 304 | + p_ch = 0; /* This makes "prev_ch" get set to 0. */ |
| 305 | + } else if (t_ch == p_ch) |
| 306 | + matched = 1; |
| 307 | + } while (prev_ch = p_ch, (p_ch = *++p) != ']'); |
| 308 | + if (matched == negated || |
| 309 | + ((flags & WM_PATHNAME) && t_ch == '/')) |
| 310 | + return WM_NOMATCH; |
| 311 | + continue; |
| 312 | + } |
| 313 | + } |
| 314 | + |
| 315 | + return *text ? WM_NOMATCH : WM_MATCH; |
| 316 | +} |
| 317 | + |
| 318 | +/* Match the "pattern" against the "text" string. */ |
| 319 | +int wildmatch(const char *pattern, const char *text, unsigned int flags) |
| 320 | +{ |
| 321 | + return dowild((const uchar*)pattern, (const uchar*)text, flags); |
| 322 | +} |
0 commit comments