| 1 | // SPDX-License-Identifier: GPL-2.0-only |
| 2 | // |
| 3 | // Traverse the source tree, parsing all .gitignore files, and print file paths |
| 4 | // that are ignored by git. |
| 5 | // The output is suitable to the --exclude-from option of tar. |
| 6 | // This is useful until the --exclude-vcs-ignores option gets working correctly. |
| 7 | // |
| 8 | // Copyright (C) 2023 Masahiro Yamada <masahiroy@kernel.org> |
| 9 | // (a lot of code imported from GIT) |
| 10 | |
| 11 | #include <assert.h> |
| 12 | #include <dirent.h> |
| 13 | #include <errno.h> |
| 14 | #include <fcntl.h> |
| 15 | #include <getopt.h> |
| 16 | #include <stdarg.h> |
| 17 | #include <stdbool.h> |
| 18 | #include <stdio.h> |
| 19 | #include <stdlib.h> |
| 20 | #include <string.h> |
| 21 | #include <sys/stat.h> |
| 22 | #include <sys/types.h> |
| 23 | #include <unistd.h> |
| 24 | |
| 25 | // Imported from commit 23c56f7bd5f1667f8b793d796bf30e39545920f6 in GIT |
| 26 | // |
| 27 | //---------------------------(IMPORT FROM GIT BEGIN)--------------------------- |
| 28 | |
| 29 | // Copied from environment.c |
| 30 | |
| 31 | static bool ignore_case; |
| 32 | |
| 33 | // Copied from git-compat-util.h |
| 34 | |
| 35 | /* Sane ctype - no locale, and works with signed chars */ |
| 36 | #undef isascii |
| 37 | #undef isspace |
| 38 | #undef isdigit |
| 39 | #undef isalpha |
| 40 | #undef isalnum |
| 41 | #undef isprint |
| 42 | #undef islower |
| 43 | #undef isupper |
| 44 | #undef tolower |
| 45 | #undef toupper |
| 46 | #undef iscntrl |
| 47 | #undef ispunct |
| 48 | #undef isxdigit |
| 49 | |
| 50 | static const unsigned char sane_ctype[256]; |
| 51 | #define GIT_SPACE 0x01 |
| 52 | #define GIT_DIGIT 0x02 |
| 53 | #define GIT_ALPHA 0x04 |
| 54 | #define GIT_GLOB_SPECIAL 0x08 |
| 55 | #define GIT_REGEX_SPECIAL 0x10 |
| 56 | #define GIT_PATHSPEC_MAGIC 0x20 |
| 57 | #define GIT_CNTRL 0x40 |
| 58 | #define GIT_PUNCT 0x80 |
| 59 | #define sane_istest(x,mask) ((sane_ctype[(unsigned char)(x)] & (mask)) != 0) |
| 60 | #define isascii(x) (((x) & ~0x7f) == 0) |
| 61 | #define isspace(x) sane_istest(x,GIT_SPACE) |
| 62 | #define isdigit(x) sane_istest(x,GIT_DIGIT) |
| 63 | #define isalpha(x) sane_istest(x,GIT_ALPHA) |
| 64 | #define isalnum(x) sane_istest(x,GIT_ALPHA | GIT_DIGIT) |
| 65 | #define isprint(x) ((x) >= 0x20 && (x) <= 0x7e) |
| 66 | #define islower(x) sane_iscase(x, 1) |
| 67 | #define isupper(x) sane_iscase(x, 0) |
| 68 | #define is_glob_special(x) sane_istest(x,GIT_GLOB_SPECIAL) |
| 69 | #define iscntrl(x) (sane_istest(x,GIT_CNTRL)) |
| 70 | #define ispunct(x) sane_istest(x, GIT_PUNCT | GIT_REGEX_SPECIAL | \ |
| 71 | GIT_GLOB_SPECIAL | GIT_PATHSPEC_MAGIC) |
| 72 | #define isxdigit(x) (hexval_table[(unsigned char)(x)] != -1) |
| 73 | #define tolower(x) sane_case((unsigned char)(x), 0x20) |
| 74 | #define toupper(x) sane_case((unsigned char)(x), 0) |
| 75 | |
| 76 | static inline int sane_case(int x, int high) |
| 77 | { |
| 78 | if (sane_istest(x, GIT_ALPHA)) |
| 79 | x = (x & ~0x20) | high; |
| 80 | return x; |
| 81 | } |
| 82 | |
| 83 | static inline int sane_iscase(int x, int is_lower) |
| 84 | { |
| 85 | if (!sane_istest(x, GIT_ALPHA)) |
| 86 | return 0; |
| 87 | |
| 88 | if (is_lower) |
| 89 | return (x & 0x20) != 0; |
| 90 | else |
| 91 | return (x & 0x20) == 0; |
| 92 | } |
| 93 | |
| 94 | // Copied from ctype.c |
| 95 | |
| 96 | enum { |
| 97 | S = GIT_SPACE, |
| 98 | A = GIT_ALPHA, |
| 99 | D = GIT_DIGIT, |
| 100 | G = GIT_GLOB_SPECIAL, /* *, ?, [, \\ */ |
| 101 | R = GIT_REGEX_SPECIAL, /* $, (, ), +, ., ^, {, | */ |
| 102 | P = GIT_PATHSPEC_MAGIC, /* other non-alnum, except for ] and } */ |
| 103 | X = GIT_CNTRL, |
| 104 | U = GIT_PUNCT, |
| 105 | Z = GIT_CNTRL | GIT_SPACE |
| 106 | }; |
| 107 | |
| 108 | static const unsigned char sane_ctype[256] = { |
| 109 | X, X, X, X, X, X, X, X, X, Z, Z, X, X, Z, X, X, /* 0.. 15 */ |
| 110 | X, X, X, X, X, X, X, X, X, X, X, X, X, X, X, X, /* 16.. 31 */ |
| 111 | S, P, P, P, R, P, P, P, R, R, G, R, P, P, R, P, /* 32.. 47 */ |
| 112 | D, D, D, D, D, D, D, D, D, D, P, P, P, P, P, G, /* 48.. 63 */ |
| 113 | P, A, A, A, A, A, A, A, A, A, A, A, A, A, A, A, /* 64.. 79 */ |
| 114 | A, A, A, A, A, A, A, A, A, A, A, G, G, U, R, P, /* 80.. 95 */ |
| 115 | P, A, A, A, A, A, A, A, A, A, A, A, A, A, A, A, /* 96..111 */ |
| 116 | A, A, A, A, A, A, A, A, A, A, A, R, R, U, P, X, /* 112..127 */ |
| 117 | /* Nothing in the 128.. range */ |
| 118 | }; |
| 119 | |
| 120 | // Copied from hex.c |
| 121 | |
| 122 | static const signed char hexval_table[256] = { |
| 123 | -1, -1, -1, -1, -1, -1, -1, -1, /* 00-07 */ |
| 124 | -1, -1, -1, -1, -1, -1, -1, -1, /* 08-0f */ |
| 125 | -1, -1, -1, -1, -1, -1, -1, -1, /* 10-17 */ |
| 126 | -1, -1, -1, -1, -1, -1, -1, -1, /* 18-1f */ |
| 127 | -1, -1, -1, -1, -1, -1, -1, -1, /* 20-27 */ |
| 128 | -1, -1, -1, -1, -1, -1, -1, -1, /* 28-2f */ |
| 129 | 0, 1, 2, 3, 4, 5, 6, 7, /* 30-37 */ |
| 130 | 8, 9, -1, -1, -1, -1, -1, -1, /* 38-3f */ |
| 131 | -1, 10, 11, 12, 13, 14, 15, -1, /* 40-47 */ |
| 132 | -1, -1, -1, -1, -1, -1, -1, -1, /* 48-4f */ |
| 133 | -1, -1, -1, -1, -1, -1, -1, -1, /* 50-57 */ |
| 134 | -1, -1, -1, -1, -1, -1, -1, -1, /* 58-5f */ |
| 135 | -1, 10, 11, 12, 13, 14, 15, -1, /* 60-67 */ |
| 136 | -1, -1, -1, -1, -1, -1, -1, -1, /* 68-67 */ |
| 137 | -1, -1, -1, -1, -1, -1, -1, -1, /* 70-77 */ |
| 138 | -1, -1, -1, -1, -1, -1, -1, -1, /* 78-7f */ |
| 139 | -1, -1, -1, -1, -1, -1, -1, -1, /* 80-87 */ |
| 140 | -1, -1, -1, -1, -1, -1, -1, -1, /* 88-8f */ |
| 141 | -1, -1, -1, -1, -1, -1, -1, -1, /* 90-97 */ |
| 142 | -1, -1, -1, -1, -1, -1, -1, -1, /* 98-9f */ |
| 143 | -1, -1, -1, -1, -1, -1, -1, -1, /* a0-a7 */ |
| 144 | -1, -1, -1, -1, -1, -1, -1, -1, /* a8-af */ |
| 145 | -1, -1, -1, -1, -1, -1, -1, -1, /* b0-b7 */ |
| 146 | -1, -1, -1, -1, -1, -1, -1, -1, /* b8-bf */ |
| 147 | -1, -1, -1, -1, -1, -1, -1, -1, /* c0-c7 */ |
| 148 | -1, -1, -1, -1, -1, -1, -1, -1, /* c8-cf */ |
| 149 | -1, -1, -1, -1, -1, -1, -1, -1, /* d0-d7 */ |
| 150 | -1, -1, -1, -1, -1, -1, -1, -1, /* d8-df */ |
| 151 | -1, -1, -1, -1, -1, -1, -1, -1, /* e0-e7 */ |
| 152 | -1, -1, -1, -1, -1, -1, -1, -1, /* e8-ef */ |
| 153 | -1, -1, -1, -1, -1, -1, -1, -1, /* f0-f7 */ |
| 154 | -1, -1, -1, -1, -1, -1, -1, -1, /* f8-ff */ |
| 155 | }; |
| 156 | |
| 157 | // Copied from wildmatch.h |
| 158 | |
| 159 | #define WM_CASEFOLD 1 |
| 160 | #define WM_PATHNAME 2 |
| 161 | |
| 162 | #define WM_NOMATCH 1 |
| 163 | #define WM_MATCH 0 |
| 164 | #define WM_ABORT_ALL -1 |
| 165 | #define WM_ABORT_TO_STARSTAR -2 |
| 166 | |
| 167 | // Copied from wildmatch.c |
| 168 | |
| 169 | typedef unsigned char uchar; |
| 170 | |
| 171 | // local modification: remove NEGATE_CLASS(2) |
| 172 | |
| 173 | #define CC_EQ(class, len, litmatch) ((len) == sizeof (litmatch)-1 \ |
| 174 | && *(class) == *(litmatch) \ |
| 175 | && strncmp((char*)class, litmatch, len) == 0) |
| 176 | |
| 177 | // local modification: simpilify macros |
| 178 | #define ISBLANK(c) ((c) == ' ' || (c) == '\t') |
| 179 | #define ISGRAPH(c) (isprint(c) && !isspace(c)) |
| 180 | #define ISPRINT(c) isprint(c) |
| 181 | #define ISDIGIT(c) isdigit(c) |
| 182 | #define ISALNUM(c) isalnum(c) |
| 183 | #define ISALPHA(c) isalpha(c) |
| 184 | #define ISCNTRL(c) iscntrl(c) |
| 185 | #define ISLOWER(c) islower(c) |
| 186 | #define ISPUNCT(c) ispunct(c) |
| 187 | #define ISSPACE(c) isspace(c) |
| 188 | #define ISUPPER(c) isupper(c) |
| 189 | #define ISXDIGIT(c) isxdigit(c) |
| 190 | |
| 191 | /* Match pattern "p" against "text" */ |
| 192 | static int dowild(const uchar *p, const uchar *text, unsigned int flags) |
| 193 | { |
| 194 | uchar p_ch; |
| 195 | const uchar *pattern = p; |
| 196 | |
| 197 | for ( ; (p_ch = *p) != '\0'; text++, p++) { |
| 198 | int matched, match_slash, negated; |
| 199 | uchar t_ch, prev_ch; |
| 200 | if ((t_ch = *text) == '\0' && p_ch != '*') |
| 201 | return WM_ABORT_ALL; |
| 202 | if ((flags & WM_CASEFOLD) && ISUPPER(t_ch)) |
| 203 | t_ch = tolower(t_ch); |
| 204 | if ((flags & WM_CASEFOLD) && ISUPPER(p_ch)) |
| 205 | p_ch = tolower(p_ch); |
| 206 | switch (p_ch) { |
| 207 | case '\\': |
| 208 | /* Literal match with following character. Note that the test |
| 209 | * in "default" handles the p[1] == '\0' failure case. */ |
| 210 | p_ch = *++p; |
| 211 | /* FALLTHROUGH */ |
| 212 | default: |
| 213 | if (t_ch != p_ch) |
| 214 | return WM_NOMATCH; |
| 215 | continue; |
| 216 | case '?': |
| 217 | /* Match anything but '/'. */ |
| 218 | if ((flags & WM_PATHNAME) && t_ch == '/') |
| 219 | return WM_NOMATCH; |
| 220 | continue; |
| 221 | case '*': |
| 222 | if (*++p == '*') { |
| 223 | const uchar *prev_p = p - 2; |
| 224 | while (*++p == '*') {} |
| 225 | if (!(flags & WM_PATHNAME)) |
| 226 | /* without WM_PATHNAME, '*' == '**' */ |
| 227 | match_slash = 1; |
| 228 | else if ((prev_p < pattern || *prev_p == '/') && |
| 229 | (*p == '\0' || *p == '/' || |
| 230 | (p[0] == '\\' && p[1] == '/'))) { |
| 231 | /* |
| 232 | * Assuming we already match 'foo/' and are at |
| 233 | * <star star slash>, just assume it matches |
| 234 | * nothing and go ahead match the rest of the |
| 235 | * pattern with the remaining string. This |
| 236 | * helps make foo/<*><*>/bar (<> because |
| 237 | * otherwise it breaks C comment syntax) match |
| 238 | * both foo/bar and foo/a/bar. |
| 239 | */ |
| 240 | if (p[0] == '/' && |
| 241 | dowild(p + 1, text, flags) == WM_MATCH) |
| 242 | return WM_MATCH; |
| 243 | match_slash = 1; |
| 244 | } else /* WM_PATHNAME is set */ |
| 245 | match_slash = 0; |
| 246 | } else |
| 247 | /* without WM_PATHNAME, '*' == '**' */ |
| 248 | match_slash = flags & WM_PATHNAME ? 0 : 1; |
| 249 | if (*p == '\0') { |
| 250 | /* Trailing "**" matches everything. Trailing "*" matches |
| 251 | * only if there are no more slash characters. */ |
| 252 | if (!match_slash) { |
| 253 | if (strchr((char *)text, '/')) |
| 254 | return WM_NOMATCH; |
| 255 | } |
| 256 | return WM_MATCH; |
| 257 | } else if (!match_slash && *p == '/') { |
| 258 | /* |
| 259 | * _one_ asterisk followed by a slash |
| 260 | * with WM_PATHNAME matches the next |
| 261 | * directory |
| 262 | */ |
| 263 | const char *slash = strchr((char*)text, '/'); |
| 264 | if (!slash) |
| 265 | return WM_NOMATCH; |
| 266 | text = (const uchar*)slash; |
| 267 | /* the slash is consumed by the top-level for loop */ |
| 268 | break; |
| 269 | } |
| 270 | while (1) { |
| 271 | if (t_ch == '\0') |
| 272 | break; |
| 273 | /* |
| 274 | * Try to advance faster when an asterisk is |
| 275 | * followed by a literal. We know in this case |
| 276 | * that the string before the literal |
| 277 | * must belong to "*". |
| 278 | * If match_slash is false, do not look past |
| 279 | * the first slash as it cannot belong to '*'. |
| 280 | */ |
| 281 | if (!is_glob_special(*p)) { |
| 282 | p_ch = *p; |
| 283 | if ((flags & WM_CASEFOLD) && ISUPPER(p_ch)) |
| 284 | p_ch = tolower(p_ch); |
| 285 | while ((t_ch = *text) != '\0' && |
| 286 | (match_slash || t_ch != '/')) { |
| 287 | if ((flags & WM_CASEFOLD) && ISUPPER(t_ch)) |
| 288 | t_ch = tolower(t_ch); |
| 289 | if (t_ch == p_ch) |
| 290 | break; |
| 291 | text++; |
| 292 | } |
| 293 | if (t_ch != p_ch) |
| 294 | return WM_NOMATCH; |
| 295 | } |
| 296 | if ((matched = dowild(p, text, flags)) != WM_NOMATCH) { |
| 297 | if (!match_slash || matched != WM_ABORT_TO_STARSTAR) |
| 298 | return matched; |
| 299 | } else if (!match_slash && t_ch == '/') |
| 300 | return WM_ABORT_TO_STARSTAR; |
| 301 | t_ch = *++text; |
| 302 | } |
| 303 | return WM_ABORT_ALL; |
| 304 | case '[': |
| 305 | p_ch = *++p; |
| 306 | if (p_ch == '^') |
| 307 | p_ch = '!'; |
| 308 | /* Assign literal 1/0 because of "matched" comparison. */ |
| 309 | negated = p_ch == '!' ? 1 : 0; |
| 310 | if (negated) { |
| 311 | /* Inverted character class. */ |
| 312 | p_ch = *++p; |
| 313 | } |
| 314 | prev_ch = 0; |
| 315 | matched = 0; |
| 316 | do { |
| 317 | if (!p_ch) |
| 318 | return WM_ABORT_ALL; |
| 319 | if (p_ch == '\\') { |
| 320 | p_ch = *++p; |
| 321 | if (!p_ch) |
| 322 | return WM_ABORT_ALL; |
| 323 | if (t_ch == p_ch) |
| 324 | matched = 1; |
| 325 | } else if (p_ch == '-' && prev_ch && p[1] && p[1] != ']') { |
| 326 | p_ch = *++p; |
| 327 | if (p_ch == '\\') { |
| 328 | p_ch = *++p; |
| 329 | if (!p_ch) |
| 330 | return WM_ABORT_ALL; |
| 331 | } |
| 332 | if (t_ch <= p_ch && t_ch >= prev_ch) |
| 333 | matched = 1; |
| 334 | else if ((flags & WM_CASEFOLD) && ISLOWER(t_ch)) { |
| 335 | uchar t_ch_upper = toupper(t_ch); |
| 336 | if (t_ch_upper <= p_ch && t_ch_upper >= prev_ch) |
| 337 | matched = 1; |
| 338 | } |
| 339 | p_ch = 0; /* This makes "prev_ch" get set to 0. */ |
| 340 | } else if (p_ch == '[' && p[1] == ':') { |
| 341 | const uchar *s; |
| 342 | int i; |
| 343 | for (s = p += 2; (p_ch = *p) && p_ch != ']'; p++) {} /*SHARED ITERATOR*/ |
| 344 | if (!p_ch) |
| 345 | return WM_ABORT_ALL; |
| 346 | i = p - s - 1; |
| 347 | if (i < 0 || p[-1] != ':') { |
| 348 | /* Didn't find ":]", so treat like a normal set. */ |
| 349 | p = s - 2; |
| 350 | p_ch = '['; |
| 351 | if (t_ch == p_ch) |
| 352 | matched = 1; |
| 353 | continue; |
| 354 | } |
| 355 | if (CC_EQ(s,i, "alnum")) { |
| 356 | if (ISALNUM(t_ch)) |
| 357 | matched = 1; |
| 358 | } else if (CC_EQ(s,i, "alpha")) { |
| 359 | if (ISALPHA(t_ch)) |
| 360 | matched = 1; |
| 361 | } else if (CC_EQ(s,i, "blank")) { |
| 362 | if (ISBLANK(t_ch)) |
| 363 | matched = 1; |
| 364 | } else if (CC_EQ(s,i, "cntrl")) { |
| 365 | if (ISCNTRL(t_ch)) |
| 366 | matched = 1; |
| 367 | } else if (CC_EQ(s,i, "digit")) { |
| 368 | if (ISDIGIT(t_ch)) |
| 369 | matched = 1; |
| 370 | } else if (CC_EQ(s,i, "graph")) { |
| 371 | if (ISGRAPH(t_ch)) |
| 372 | matched = 1; |
| 373 | } else if (CC_EQ(s,i, "lower")) { |
| 374 | if (ISLOWER(t_ch)) |
| 375 | matched = 1; |
| 376 | } else if (CC_EQ(s,i, "print")) { |
| 377 | if (ISPRINT(t_ch)) |
| 378 | matched = 1; |
| 379 | } else if (CC_EQ(s,i, "punct")) { |
| 380 | if (ISPUNCT(t_ch)) |
| 381 | matched = 1; |
| 382 | } else if (CC_EQ(s,i, "space")) { |
| 383 | if (ISSPACE(t_ch)) |
| 384 | matched = 1; |
| 385 | } else if (CC_EQ(s,i, "upper")) { |
| 386 | if (ISUPPER(t_ch)) |
| 387 | matched = 1; |
| 388 | else if ((flags & WM_CASEFOLD) && ISLOWER(t_ch)) |
| 389 | matched = 1; |
| 390 | } else if (CC_EQ(s,i, "xdigit")) { |
| 391 | if (ISXDIGIT(t_ch)) |
| 392 | matched = 1; |
| 393 | } else /* malformed [:class:] string */ |
| 394 | return WM_ABORT_ALL; |
| 395 | p_ch = 0; /* This makes "prev_ch" get set to 0. */ |
| 396 | } else if (t_ch == p_ch) |
| 397 | matched = 1; |
| 398 | } while (prev_ch = p_ch, (p_ch = *++p) != ']'); |
| 399 | if (matched == negated || |
| 400 | ((flags & WM_PATHNAME) && t_ch == '/')) |
| 401 | return WM_NOMATCH; |
| 402 | continue; |
| 403 | } |
| 404 | } |
| 405 | |
| 406 | return *text ? WM_NOMATCH : WM_MATCH; |
| 407 | } |
| 408 | |
| 409 | /* Match the "pattern" against the "text" string. */ |
| 410 | static int wildmatch(const char *pattern, const char *text, unsigned int flags) |
| 411 | { |
| 412 | // local modification: move WM_CASEFOLD here |
| 413 | if (ignore_case) |
| 414 | flags |= WM_CASEFOLD; |
| 415 | |
| 416 | return dowild((const uchar*)pattern, (const uchar*)text, flags); |
| 417 | } |
| 418 | |
| 419 | // Copied from dir.h |
| 420 | |
| 421 | #define PATTERN_FLAG_NODIR 1 |
| 422 | #define PATTERN_FLAG_ENDSWITH 4 |
| 423 | #define PATTERN_FLAG_MUSTBEDIR 8 |
| 424 | #define PATTERN_FLAG_NEGATIVE 16 |
| 425 | |
| 426 | // Copied from dir.c |
| 427 | |
| 428 | static int fspathncmp(const char *a, const char *b, size_t count) |
| 429 | { |
| 430 | return ignore_case ? strncasecmp(a, b, count) : strncmp(a, b, count); |
| 431 | } |
| 432 | |
| 433 | static int simple_length(const char *match) |
| 434 | { |
| 435 | int len = -1; |
| 436 | |
| 437 | for (;;) { |
| 438 | unsigned char c = *match++; |
| 439 | len++; |
| 440 | if (c == '\0' || is_glob_special(c)) |
| 441 | return len; |
| 442 | } |
| 443 | } |
| 444 | |
| 445 | static int no_wildcard(const char *string) |
| 446 | { |
| 447 | return string[simple_length(string)] == '\0'; |
| 448 | } |
| 449 | |
| 450 | static void parse_path_pattern(const char **pattern, |
| 451 | int *patternlen, |
| 452 | unsigned *flags, |
| 453 | int *nowildcardlen) |
| 454 | { |
| 455 | const char *p = *pattern; |
| 456 | size_t i, len; |
| 457 | |
| 458 | *flags = 0; |
| 459 | if (*p == '!') { |
| 460 | *flags |= PATTERN_FLAG_NEGATIVE; |
| 461 | p++; |
| 462 | } |
| 463 | len = strlen(p); |
| 464 | if (len && p[len - 1] == '/') { |
| 465 | len--; |
| 466 | *flags |= PATTERN_FLAG_MUSTBEDIR; |
| 467 | } |
| 468 | for (i = 0; i < len; i++) { |
| 469 | if (p[i] == '/') |
| 470 | break; |
| 471 | } |
| 472 | if (i == len) |
| 473 | *flags |= PATTERN_FLAG_NODIR; |
| 474 | *nowildcardlen = simple_length(p); |
| 475 | /* |
| 476 | * we should have excluded the trailing slash from 'p' too, |
| 477 | * but that's one more allocation. Instead just make sure |
| 478 | * nowildcardlen does not exceed real patternlen |
| 479 | */ |
| 480 | if (*nowildcardlen > len) |
| 481 | *nowildcardlen = len; |
| 482 | if (*p == '*' && no_wildcard(p + 1)) |
| 483 | *flags |= PATTERN_FLAG_ENDSWITH; |
| 484 | *pattern = p; |
| 485 | *patternlen = len; |
| 486 | } |
| 487 | |
| 488 | static void trim_trailing_spaces(char *buf) |
| 489 | { |
| 490 | char *p, *last_space = NULL; |
| 491 | |
| 492 | for (p = buf; *p; p++) |
| 493 | switch (*p) { |
| 494 | case ' ': |
| 495 | if (!last_space) |
| 496 | last_space = p; |
| 497 | break; |
| 498 | case '\\': |
| 499 | p++; |
| 500 | if (!*p) |
| 501 | return; |
| 502 | /* fallthrough */ |
| 503 | default: |
| 504 | last_space = NULL; |
| 505 | } |
| 506 | |
| 507 | if (last_space) |
| 508 | *last_space = '\0'; |
| 509 | } |
| 510 | |
| 511 | static int match_basename(const char *basename, int basenamelen, |
| 512 | const char *pattern, int prefix, int patternlen, |
| 513 | unsigned flags) |
| 514 | { |
| 515 | if (prefix == patternlen) { |
| 516 | if (patternlen == basenamelen && |
| 517 | !fspathncmp(pattern, basename, basenamelen)) |
| 518 | return 1; |
| 519 | } else if (flags & PATTERN_FLAG_ENDSWITH) { |
| 520 | /* "*literal" matching against "fooliteral" */ |
| 521 | if (patternlen - 1 <= basenamelen && |
| 522 | !fspathncmp(pattern + 1, |
| 523 | basename + basenamelen - (patternlen - 1), |
| 524 | patternlen - 1)) |
| 525 | return 1; |
| 526 | } else { |
| 527 | // local modification: call wildmatch() directly |
| 528 | if (!wildmatch(pattern, basename, flags)) |
| 529 | return 1; |
| 530 | } |
| 531 | return 0; |
| 532 | } |
| 533 | |
| 534 | static int match_pathname(const char *pathname, int pathlen, |
| 535 | const char *base, int baselen, |
| 536 | const char *pattern, int prefix, int patternlen) |
| 537 | { |
| 538 | // local modification: remove local variables |
| 539 | |
| 540 | /* |
| 541 | * match with FNM_PATHNAME; the pattern has base implicitly |
| 542 | * in front of it. |
| 543 | */ |
| 544 | if (*pattern == '/') { |
| 545 | pattern++; |
| 546 | patternlen--; |
| 547 | prefix--; |
| 548 | } |
| 549 | |
| 550 | /* |
| 551 | * baselen does not count the trailing slash. base[] may or |
| 552 | * may not end with a trailing slash though. |
| 553 | */ |
| 554 | if (pathlen < baselen + 1 || |
| 555 | (baselen && pathname[baselen] != '/') || |
| 556 | fspathncmp(pathname, base, baselen)) |
| 557 | return 0; |
| 558 | |
| 559 | // local modification: simplified because always baselen > 0 |
| 560 | pathname += baselen + 1; |
| 561 | pathlen -= baselen + 1; |
| 562 | |
| 563 | if (prefix) { |
| 564 | /* |
| 565 | * if the non-wildcard part is longer than the |
| 566 | * remaining pathname, surely it cannot match. |
| 567 | */ |
| 568 | if (prefix > pathlen) |
| 569 | return 0; |
| 570 | |
| 571 | if (fspathncmp(pattern, pathname, prefix)) |
| 572 | return 0; |
| 573 | pattern += prefix; |
| 574 | patternlen -= prefix; |
| 575 | pathname += prefix; |
| 576 | pathlen -= prefix; |
| 577 | |
| 578 | /* |
| 579 | * If the whole pattern did not have a wildcard, |
| 580 | * then our prefix match is all we need; we |
| 581 | * do not need to call fnmatch at all. |
| 582 | */ |
| 583 | if (!patternlen && !pathlen) |
| 584 | return 1; |
| 585 | } |
| 586 | |
| 587 | // local modification: call wildmatch() directly |
| 588 | return !wildmatch(pattern, pathname, WM_PATHNAME); |
| 589 | } |
| 590 | |
| 591 | // Copied from git/utf8.c |
| 592 | |
| 593 | static const char utf8_bom[] = "\357\273\277"; |
| 594 | |
| 595 | //----------------------------(IMPORT FROM GIT END)---------------------------- |
| 596 | |
| 597 | struct pattern { |
| 598 | unsigned int flags; |
| 599 | int nowildcardlen; |
| 600 | int patternlen; |
| 601 | int dirlen; |
| 602 | char pattern[]; |
| 603 | }; |
| 604 | |
| 605 | static struct pattern **pattern_list; |
| 606 | static int nr_patterns, alloced_patterns; |
| 607 | |
| 608 | // Remember the number of patterns at each directory level |
| 609 | static int *nr_patterns_at; |
| 610 | // Track the current/max directory level; |
| 611 | static int depth, max_depth; |
| 612 | static bool debug_on; |
| 613 | static FILE *out_fp, *stat_fp; |
| 614 | static char *prefix = ""; |
| 615 | static char *progname; |
| 616 | |
| 617 | static void __attribute__((noreturn)) perror_exit(const char *s) |
| 618 | { |
| 619 | perror(s); |
| 620 | |
| 621 | exit(EXIT_FAILURE); |
| 622 | } |
| 623 | |
| 624 | static void __attribute__((noreturn)) error_exit(const char *fmt, ...) |
| 625 | { |
| 626 | va_list args; |
| 627 | |
| 628 | fprintf(stderr, "%s: error: ", progname); |
| 629 | |
| 630 | va_start(args, fmt); |
| 631 | vfprintf(stderr, fmt, args); |
| 632 | va_end(args); |
| 633 | |
| 634 | exit(EXIT_FAILURE); |
| 635 | } |
| 636 | |
| 637 | static void debug(const char *fmt, ...) |
| 638 | { |
| 639 | va_list args; |
| 640 | int i; |
| 641 | |
| 642 | if (!debug_on) |
| 643 | return; |
| 644 | |
| 645 | fprintf(stderr, "[DEBUG] "); |
| 646 | |
| 647 | for (i = 0; i < depth * 2; i++) |
| 648 | fputc(' ', stderr); |
| 649 | |
| 650 | va_start(args, fmt); |
| 651 | vfprintf(stderr, fmt, args); |
| 652 | va_end(args); |
| 653 | } |
| 654 | |
| 655 | static void *xrealloc(void *ptr, size_t size) |
| 656 | { |
| 657 | ptr = realloc(ptr, size); |
| 658 | if (!ptr) |
| 659 | perror_exit(progname); |
| 660 | |
| 661 | return ptr; |
| 662 | } |
| 663 | |
| 664 | static void *xmalloc(size_t size) |
| 665 | { |
| 666 | return xrealloc(NULL, size); |
| 667 | } |
| 668 | |
| 669 | // similar to last_matching_pattern_from_list() in GIT |
| 670 | static bool is_ignored(const char *path, int pathlen, int dirlen, bool is_dir) |
| 671 | { |
| 672 | int i; |
| 673 | |
| 674 | // Search in the reverse order because the last matching pattern wins. |
| 675 | for (i = nr_patterns - 1; i >= 0; i--) { |
| 676 | struct pattern *p = pattern_list[i]; |
| 677 | unsigned int flags = p->flags; |
| 678 | const char *gitignore_dir = p->pattern + p->patternlen + 1; |
| 679 | bool ignored; |
| 680 | |
| 681 | if ((flags & PATTERN_FLAG_MUSTBEDIR) && !is_dir) |
| 682 | continue; |
| 683 | |
| 684 | if (flags & PATTERN_FLAG_NODIR) { |
| 685 | if (!match_basename(path + dirlen + 1, |
| 686 | pathlen - dirlen - 1, |
| 687 | p->pattern, |
| 688 | p->nowildcardlen, |
| 689 | p->patternlen, |
| 690 | p->flags)) |
| 691 | continue; |
| 692 | } else { |
| 693 | if (!match_pathname(path, pathlen, |
| 694 | gitignore_dir, p->dirlen, |
| 695 | p->pattern, |
| 696 | p->nowildcardlen, |
| 697 | p->patternlen)) |
| 698 | continue; |
| 699 | } |
| 700 | |
| 701 | debug("%s: matches %s%s%s (%s/.gitignore)\n", path, |
| 702 | flags & PATTERN_FLAG_NEGATIVE ? "!" : "", p->pattern, |
| 703 | flags & PATTERN_FLAG_MUSTBEDIR ? "/" : "", |
| 704 | gitignore_dir); |
| 705 | |
| 706 | ignored = (flags & PATTERN_FLAG_NEGATIVE) == 0; |
| 707 | if (ignored) |
| 708 | debug("Ignore: %s\n", path); |
| 709 | |
| 710 | return ignored; |
| 711 | } |
| 712 | |
| 713 | debug("%s: no match\n", path); |
| 714 | |
| 715 | return false; |
| 716 | } |
| 717 | |
| 718 | static void add_pattern(const char *string, const char *dir, int dirlen) |
| 719 | { |
| 720 | struct pattern *p; |
| 721 | int patternlen, nowildcardlen; |
| 722 | unsigned int flags; |
| 723 | |
| 724 | parse_path_pattern(&string, &patternlen, &flags, &nowildcardlen); |
| 725 | |
| 726 | if (patternlen == 0) |
| 727 | return; |
| 728 | |
| 729 | p = xmalloc(sizeof(*p) + patternlen + dirlen + 2); |
| 730 | |
| 731 | memcpy(p->pattern, string, patternlen); |
| 732 | p->pattern[patternlen] = 0; |
| 733 | memcpy(p->pattern + patternlen + 1, dir, dirlen); |
| 734 | p->pattern[patternlen + 1 + dirlen] = 0; |
| 735 | |
| 736 | p->patternlen = patternlen; |
| 737 | p->nowildcardlen = nowildcardlen; |
| 738 | p->dirlen = dirlen; |
| 739 | p->flags = flags; |
| 740 | |
| 741 | debug("Add pattern: %s%s%s\n", |
| 742 | flags & PATTERN_FLAG_NEGATIVE ? "!" : "", p->pattern, |
| 743 | flags & PATTERN_FLAG_MUSTBEDIR ? "/" : ""); |
| 744 | |
| 745 | if (nr_patterns >= alloced_patterns) { |
| 746 | alloced_patterns += 128; |
| 747 | pattern_list = xrealloc(pattern_list, |
| 748 | sizeof(*pattern_list) * alloced_patterns); |
| 749 | } |
| 750 | |
| 751 | pattern_list[nr_patterns++] = p; |
| 752 | } |
| 753 | |
| 754 | // similar to add_patterns_from_buffer() in GIT |
| 755 | static void add_patterns_from_gitignore(const char *dir, int dirlen) |
| 756 | { |
| 757 | struct stat st; |
| 758 | char path[PATH_MAX], *buf, *entry; |
| 759 | size_t size; |
| 760 | int fd, pathlen, i; |
| 761 | |
| 762 | pathlen = snprintf(path, sizeof(path), "%s/.gitignore", dir); |
| 763 | if (pathlen >= sizeof(path)) |
| 764 | error_exit("%s: too long path was truncated\n", path); |
| 765 | |
| 766 | fd = open(path, O_RDONLY | O_NOFOLLOW); |
| 767 | if (fd < 0) { |
| 768 | if (errno != ENOENT) |
| 769 | return perror_exit(path); |
| 770 | return; |
| 771 | } |
| 772 | |
| 773 | if (fstat(fd, &st) < 0) |
| 774 | perror_exit(path); |
| 775 | |
| 776 | size = st.st_size; |
| 777 | |
| 778 | buf = xmalloc(size + 1); |
| 779 | if (read(fd, buf, st.st_size) != st.st_size) |
| 780 | perror_exit(path); |
| 781 | |
| 782 | buf[st.st_size] = '\n'; |
| 783 | if (close(fd)) |
| 784 | perror_exit(path); |
| 785 | |
| 786 | debug("Parse %s\n", path); |
| 787 | |
| 788 | entry = buf; |
| 789 | |
| 790 | // skip utf8 bom |
| 791 | if (!strncmp(entry, utf8_bom, strlen(utf8_bom))) |
| 792 | entry += strlen(utf8_bom); |
| 793 | |
| 794 | for (i = entry - buf; i < size; i++) { |
| 795 | if (buf[i] == '\n') { |
| 796 | if (entry != buf + i && entry[0] != '#') { |
| 797 | buf[i - (i && buf[i-1] == '\r')] = 0; |
| 798 | trim_trailing_spaces(entry); |
| 799 | add_pattern(entry, dir, dirlen); |
| 800 | } |
| 801 | entry = buf + i + 1; |
| 802 | } |
| 803 | } |
| 804 | |
| 805 | free(buf); |
| 806 | } |
| 807 | |
| 808 | // Save the current number of patterns and increment the depth |
| 809 | static void increment_depth(void) |
| 810 | { |
| 811 | if (depth >= max_depth) { |
| 812 | max_depth += 1; |
| 813 | nr_patterns_at = xrealloc(nr_patterns_at, |
| 814 | sizeof(*nr_patterns_at) * max_depth); |
| 815 | } |
| 816 | |
| 817 | nr_patterns_at[depth] = nr_patterns; |
| 818 | depth++; |
| 819 | } |
| 820 | |
| 821 | // Decrement the depth, and free up the patterns of this directory level. |
| 822 | static void decrement_depth(void) |
| 823 | { |
| 824 | depth--; |
| 825 | assert(depth >= 0); |
| 826 | |
| 827 | while (nr_patterns > nr_patterns_at[depth]) |
| 828 | free(pattern_list[--nr_patterns]); |
| 829 | } |
| 830 | |
| 831 | static void print_path(const char *path) |
| 832 | { |
| 833 | // The path always starts with "./" |
| 834 | assert(strlen(path) >= 2); |
| 835 | |
| 836 | // Replace the root directory with a preferred prefix. |
| 837 | // This is useful for the tar command. |
| 838 | fprintf(out_fp, "%s%s\n", prefix, path + 2); |
| 839 | } |
| 840 | |
| 841 | static void print_stat(const char *path, struct stat *st) |
| 842 | { |
| 843 | if (!stat_fp) |
| 844 | return; |
| 845 | |
| 846 | if (!S_ISREG(st->st_mode) && !S_ISLNK(st->st_mode)) |
| 847 | return; |
| 848 | |
| 849 | assert(strlen(path) >= 2); |
| 850 | |
| 851 | fprintf(stat_fp, "%c %9ld %10ld %s\n", |
| 852 | S_ISLNK(st->st_mode) ? 'l' : '-', |
| 853 | st->st_size, st->st_mtim.tv_sec, path + 2); |
| 854 | } |
| 855 | |
| 856 | // Traverse the entire directory tree, parsing .gitignore files. |
| 857 | // Print file paths that are not tracked by git. |
| 858 | // |
| 859 | // Return true if all files under the directory are ignored, false otherwise. |
| 860 | static bool traverse_directory(const char *dir, int dirlen) |
| 861 | { |
| 862 | bool all_ignored = true; |
| 863 | DIR *dirp; |
| 864 | |
| 865 | debug("Enter[%d]: %s\n", depth, dir); |
| 866 | increment_depth(); |
| 867 | |
| 868 | add_patterns_from_gitignore(dir, dirlen); |
| 869 | |
| 870 | dirp = opendir(dir); |
| 871 | if (!dirp) |
| 872 | perror_exit(dir); |
| 873 | |
| 874 | while (1) { |
| 875 | struct dirent *d; |
| 876 | struct stat st; |
| 877 | char path[PATH_MAX]; |
| 878 | int pathlen; |
| 879 | bool ignored; |
| 880 | |
| 881 | errno = 0; |
| 882 | d = readdir(dirp); |
| 883 | if (!d) { |
| 884 | if (errno) |
| 885 | perror_exit(dir); |
| 886 | break; |
| 887 | } |
| 888 | |
| 889 | if (!strcmp(d->d_name, "..") || !strcmp(d->d_name, ".")) |
| 890 | continue; |
| 891 | |
| 892 | pathlen = snprintf(path, sizeof(path), "%s/%s", dir, d->d_name); |
| 893 | if (pathlen >= sizeof(path)) |
| 894 | error_exit("%s: too long path was truncated\n", path); |
| 895 | |
| 896 | if (lstat(path, &st) < 0) |
| 897 | perror_exit(path); |
| 898 | |
| 899 | if ((!S_ISREG(st.st_mode) && !S_ISDIR(st.st_mode) && !S_ISLNK(st.st_mode)) || |
| 900 | is_ignored(path, pathlen, dirlen, S_ISDIR(st.st_mode))) { |
| 901 | ignored = true; |
| 902 | } else { |
| 903 | if (S_ISDIR(st.st_mode) && !S_ISLNK(st.st_mode)) |
| 904 | // If all the files in a directory are ignored, |
| 905 | // let's ignore that directory as well. This |
| 906 | // will avoid empty directories in the tarball. |
| 907 | ignored = traverse_directory(path, pathlen); |
| 908 | else |
| 909 | ignored = false; |
| 910 | } |
| 911 | |
| 912 | if (ignored) { |
| 913 | print_path(path); |
| 914 | } else { |
| 915 | print_stat(path, &st); |
| 916 | all_ignored = false; |
| 917 | } |
| 918 | } |
| 919 | |
| 920 | if (closedir(dirp)) |
| 921 | perror_exit(dir); |
| 922 | |
| 923 | decrement_depth(); |
| 924 | debug("Leave[%d]: %s\n", depth, dir); |
| 925 | |
| 926 | return all_ignored; |
| 927 | } |
| 928 | |
| 929 | static void usage(void) |
| 930 | { |
| 931 | fprintf(stderr, |
| 932 | "usage: %s [options]\n" |
| 933 | "\n" |
| 934 | "Show files that are ignored by git\n" |
| 935 | "\n" |
| 936 | "options:\n" |
| 937 | " -d, --debug print debug messages to stderr\n" |
| 938 | " -e, --exclude PATTERN add the given exclude pattern\n" |
| 939 | " -h, --help show this help message and exit\n" |
| 940 | " -i, --ignore-case Ignore case differences between the patterns and the files\n" |
| 941 | " -o, --output FILE output the ignored files to a file (default: '-', i.e. stdout)\n" |
| 942 | " -p, --prefix PREFIX prefix added to each path (default: empty string)\n" |
| 943 | " -r, --rootdir DIR root of the source tree (default: current working directory)\n" |
| 944 | " -s, --stat FILE output the file stat of non-ignored files to a file\n", |
| 945 | progname); |
| 946 | } |
| 947 | |
| 948 | static void open_output(const char *pathname, FILE **fp) |
| 949 | { |
| 950 | if (strcmp(pathname, "-")) { |
| 951 | *fp = fopen(pathname, "w"); |
| 952 | if (!*fp) |
| 953 | perror_exit(pathname); |
| 954 | } else { |
| 955 | *fp = stdout; |
| 956 | } |
| 957 | } |
| 958 | |
| 959 | static void close_output(const char *pathname, FILE *fp) |
| 960 | { |
| 961 | fflush(fp); |
| 962 | |
| 963 | if (ferror(fp)) |
| 964 | error_exit("not all data was written to the output\n"); |
| 965 | |
| 966 | if (fclose(fp)) |
| 967 | perror_exit(pathname); |
| 968 | } |
| 969 | |
| 970 | int main(int argc, char *argv[]) |
| 971 | { |
| 972 | const char *output = "-"; |
| 973 | const char *rootdir = "."; |
| 974 | const char *stat = NULL; |
| 975 | |
| 976 | progname = strrchr(argv[0], '/'); |
| 977 | if (progname) |
| 978 | progname++; |
| 979 | else |
| 980 | progname = argv[0]; |
| 981 | |
| 982 | while (1) { |
| 983 | static struct option long_options[] = { |
| 984 | {"debug", no_argument, NULL, 'd'}, |
| 985 | {"help", no_argument, NULL, 'h'}, |
| 986 | {"ignore-case", no_argument, NULL, 'i'}, |
| 987 | {"output", required_argument, NULL, 'o'}, |
| 988 | {"prefix", required_argument, NULL, 'p'}, |
| 989 | {"rootdir", required_argument, NULL, 'r'}, |
| 990 | {"stat", required_argument, NULL, 's'}, |
| 991 | {"exclude", required_argument, NULL, 'x'}, |
| 992 | {}, |
| 993 | }; |
| 994 | |
| 995 | int c = getopt_long(argc, argv, "dhino:p:r:s:x:", long_options, NULL); |
| 996 | |
| 997 | if (c == -1) |
| 998 | break; |
| 999 | |
| 1000 | switch (c) { |
| 1001 | case 'd': |
| 1002 | debug_on = true; |
| 1003 | break; |
| 1004 | case 'h': |
| 1005 | usage(); |
| 1006 | exit(0); |
| 1007 | case 'i': |
| 1008 | ignore_case = true; |
| 1009 | break; |
| 1010 | case 'o': |
| 1011 | output = optarg; |
| 1012 | break; |
| 1013 | case 'p': |
| 1014 | prefix = optarg; |
| 1015 | break; |
| 1016 | case 'r': |
| 1017 | rootdir = optarg; |
| 1018 | break; |
| 1019 | case 's': |
| 1020 | stat = optarg; |
| 1021 | break; |
| 1022 | case 'x': |
| 1023 | add_pattern(optarg, ".", strlen(".")); |
| 1024 | break; |
| 1025 | case '?': |
| 1026 | usage(); |
| 1027 | /* fallthrough */ |
| 1028 | default: |
| 1029 | exit(EXIT_FAILURE); |
| 1030 | } |
| 1031 | } |
| 1032 | |
| 1033 | open_output(output, &out_fp); |
| 1034 | if (stat && stat[0]) |
| 1035 | open_output(stat, &stat_fp); |
| 1036 | |
| 1037 | if (chdir(rootdir)) |
| 1038 | perror_exit(rootdir); |
| 1039 | |
| 1040 | add_pattern(".git/", ".", strlen(".")); |
| 1041 | |
| 1042 | if (traverse_directory(".", strlen("."))) |
| 1043 | print_path("./"); |
| 1044 | |
| 1045 | assert(depth == 0); |
| 1046 | |
| 1047 | while (nr_patterns > 0) |
| 1048 | free(pattern_list[--nr_patterns]); |
| 1049 | free(pattern_list); |
| 1050 | free(nr_patterns_at); |
| 1051 | |
| 1052 | close_output(output, out_fp); |
| 1053 | if (stat_fp) |
| 1054 | close_output(stat, stat_fp); |
| 1055 | |
| 1056 | return 0; |
| 1057 | } |