| 1 | #define DISABLE_SIGN_COMPARE_WARNINGS |
| 2 | |
| 3 | #include "git-compat-util.h" |
| 4 | #include "path.h" |
| 5 | #include "quote.h" |
| 6 | #include "strbuf.h" |
| 7 | #include "strvec.h" |
| 8 | |
| 9 | int quote_path_fully = 1; |
| 10 | |
| 11 | static inline int need_bs_quote(char c) |
| 12 | { |
| 13 | return (c == '\'' || c == '!'); |
| 14 | } |
| 15 | |
| 16 | /* Help to copy the thing properly quoted for the shell safety. |
| 17 | * any single quote is replaced with '\'', any exclamation point |
| 18 | * is replaced with '\!', and the whole thing is enclosed in a |
| 19 | * single quote pair. |
| 20 | * |
| 21 | * E.g. |
| 22 | * original sq_quote result |
| 23 | * name ==> name ==> 'name' |
| 24 | * a b ==> a b ==> 'a b' |
| 25 | * a'b ==> a'\''b ==> 'a'\''b' |
| 26 | * a!b ==> a'\!'b ==> 'a'\!'b' |
| 27 | */ |
| 28 | void sq_quote_buf(struct strbuf *dst, const char *src) |
| 29 | { |
| 30 | char *to_free = NULL; |
| 31 | |
| 32 | if (dst->buf == src) |
| 33 | to_free = strbuf_detach(dst, NULL); |
| 34 | |
| 35 | strbuf_addch(dst, '\''); |
| 36 | while (*src) { |
| 37 | size_t len = strcspn(src, "'!"); |
| 38 | strbuf_add(dst, src, len); |
| 39 | src += len; |
| 40 | while (need_bs_quote(*src)) { |
| 41 | strbuf_addstr(dst, "'\\"); |
| 42 | strbuf_addch(dst, *src++); |
| 43 | strbuf_addch(dst, '\''); |
| 44 | } |
| 45 | } |
| 46 | strbuf_addch(dst, '\''); |
| 47 | free(to_free); |
| 48 | } |
| 49 | |
| 50 | void sq_quote_buf_pretty(struct strbuf *dst, const char *src) |
| 51 | { |
| 52 | static const char ok_punct[] = "+,-./:=@_^"; |
| 53 | const char *p; |
| 54 | |
| 55 | /* Avoid losing a zero-length string by adding '' */ |
| 56 | if (!*src) { |
| 57 | strbuf_addstr(dst, "''"); |
| 58 | return; |
| 59 | } |
| 60 | |
| 61 | for (p = src; *p; p++) { |
| 62 | if (!isalnum(*p) && !strchr(ok_punct, *p)) { |
| 63 | sq_quote_buf(dst, src); |
| 64 | return; |
| 65 | } |
| 66 | } |
| 67 | |
| 68 | /* if we get here, we did not need quoting */ |
| 69 | strbuf_addstr(dst, src); |
| 70 | } |
| 71 | |
| 72 | void sq_quotef(struct strbuf *dst, const char *fmt, ...) |
| 73 | { |
| 74 | struct strbuf src = STRBUF_INIT; |
| 75 | |
| 76 | va_list ap; |
| 77 | va_start(ap, fmt); |
| 78 | strbuf_vaddf(&src, fmt, ap); |
| 79 | va_end(ap); |
| 80 | |
| 81 | sq_quote_buf(dst, src.buf); |
| 82 | strbuf_release(&src); |
| 83 | } |
| 84 | |
| 85 | void sq_quote_argv(struct strbuf *dst, const char **argv) |
| 86 | { |
| 87 | int i; |
| 88 | |
| 89 | /* Copy into destination buffer. */ |
| 90 | strbuf_grow(dst, 255); |
| 91 | for (i = 0; argv[i]; ++i) { |
| 92 | strbuf_addch(dst, ' '); |
| 93 | sq_quote_buf(dst, argv[i]); |
| 94 | } |
| 95 | } |
| 96 | |
| 97 | /* |
| 98 | * Legacy function to append each argv value, quoted as necessasry, |
| 99 | * with whitespace before each value. This results in a leading |
| 100 | * space in the result. |
| 101 | */ |
| 102 | void sq_quote_argv_pretty(struct strbuf *dst, const char **argv) |
| 103 | { |
| 104 | if (argv[0]) |
| 105 | strbuf_addch(dst, ' '); |
| 106 | sq_append_quote_argv_pretty(dst, argv); |
| 107 | } |
| 108 | |
| 109 | /* |
| 110 | * Append each argv value, quoted as necessary, with whitespace between them. |
| 111 | */ |
| 112 | void sq_append_quote_argv_pretty(struct strbuf *dst, const char **argv) |
| 113 | { |
| 114 | int i; |
| 115 | |
| 116 | for (i = 0; argv[i]; i++) { |
| 117 | if (i > 0) |
| 118 | strbuf_addch(dst, ' '); |
| 119 | sq_quote_buf_pretty(dst, argv[i]); |
| 120 | } |
| 121 | } |
| 122 | |
| 123 | char *sq_dequote_step(char *arg, char **next) |
| 124 | { |
| 125 | char *dst = arg; |
| 126 | char *src = arg; |
| 127 | char c; |
| 128 | |
| 129 | if (*src != '\'') |
| 130 | return NULL; |
| 131 | for (;;) { |
| 132 | c = *++src; |
| 133 | if (!c) |
| 134 | return NULL; |
| 135 | if (c != '\'') { |
| 136 | *dst++ = c; |
| 137 | continue; |
| 138 | } |
| 139 | /* We stepped out of sq */ |
| 140 | switch (*++src) { |
| 141 | case '\0': |
| 142 | *dst = 0; |
| 143 | if (next) |
| 144 | *next = NULL; |
| 145 | return arg; |
| 146 | case '\\': |
| 147 | /* |
| 148 | * Allow backslashed characters outside of |
| 149 | * single-quotes only if they need escaping, |
| 150 | * and only if we resume the single-quoted part |
| 151 | * afterward. |
| 152 | */ |
| 153 | if (need_bs_quote(src[1]) && src[2] == '\'') { |
| 154 | *dst++ = src[1]; |
| 155 | src += 2; |
| 156 | continue; |
| 157 | } |
| 158 | /* Fallthrough */ |
| 159 | default: |
| 160 | if (!next) |
| 161 | return NULL; |
| 162 | *dst = 0; |
| 163 | *next = src; |
| 164 | return arg; |
| 165 | } |
| 166 | } |
| 167 | } |
| 168 | |
| 169 | char *sq_dequote(char *arg) |
| 170 | { |
| 171 | return sq_dequote_step(arg, NULL); |
| 172 | } |
| 173 | |
| 174 | int sq_dequote_to_strvec(char *arg, struct strvec *array) |
| 175 | { |
| 176 | char *next = arg; |
| 177 | |
| 178 | if (!*arg) |
| 179 | return 0; |
| 180 | do { |
| 181 | char *dequoted = sq_dequote_step(next, &next); |
| 182 | if (!dequoted) |
| 183 | return -1; |
| 184 | if (next) { |
| 185 | char c; |
| 186 | if (!isspace(*next)) |
| 187 | return -1; |
| 188 | do { |
| 189 | c = *++next; |
| 190 | } while (isspace(c)); |
| 191 | } |
| 192 | strvec_push(array, dequoted); |
| 193 | } while (next); |
| 194 | |
| 195 | return 0; |
| 196 | } |
| 197 | |
| 198 | /* 1 means: quote as octal |
| 199 | * 0 means: quote as octal if (quote_path_fully) |
| 200 | * -1 means: never quote |
| 201 | * c: quote as "\\c" |
| 202 | */ |
| 203 | #define X8(x) x, x, x, x, x, x, x, x |
| 204 | #define X16(x) X8(x), X8(x) |
| 205 | static signed char const cq_lookup[256] = { |
| 206 | /* 0 1 2 3 4 5 6 7 */ |
| 207 | /* 0x00 */ 1, 1, 1, 1, 1, 1, 1, 'a', |
| 208 | /* 0x08 */ 'b', 't', 'n', 'v', 'f', 'r', 1, 1, |
| 209 | /* 0x10 */ X16(1), |
| 210 | /* 0x20 */ -1, -1, '"', -1, -1, -1, -1, -1, |
| 211 | /* 0x28 */ X16(-1), X16(-1), X16(-1), |
| 212 | /* 0x58 */ -1, -1, -1, -1,'\\', -1, -1, -1, |
| 213 | /* 0x60 */ X16(-1), X8(-1), |
| 214 | /* 0x78 */ -1, -1, -1, -1, -1, -1, -1, 1, |
| 215 | /* 0x80 */ /* set to 0 */ |
| 216 | }; |
| 217 | |
| 218 | static inline int cq_must_quote(char c) |
| 219 | { |
| 220 | return cq_lookup[(unsigned char)c] + quote_path_fully > 0; |
| 221 | } |
| 222 | |
| 223 | /* returns the longest prefix not needing a quote up to maxlen if positive. |
| 224 | This stops at the first \0 because it's marked as a character needing an |
| 225 | escape */ |
| 226 | static size_t next_quote_pos(const char *s, ssize_t maxlen) |
| 227 | { |
| 228 | size_t len; |
| 229 | if (maxlen < 0) { |
| 230 | for (len = 0; !cq_must_quote(s[len]); len++); |
| 231 | } else { |
| 232 | for (len = 0; len < maxlen && !cq_must_quote(s[len]); len++); |
| 233 | } |
| 234 | return len; |
| 235 | } |
| 236 | |
| 237 | /* |
| 238 | * C-style name quoting. |
| 239 | * |
| 240 | * (1) if sb and fp are both NULL, inspect the input name and counts the |
| 241 | * number of bytes that are needed to hold c_style quoted version of name, |
| 242 | * counting the double quotes around it but not terminating NUL, and |
| 243 | * returns it. |
| 244 | * However, if name does not need c_style quoting, it returns 0. |
| 245 | * |
| 246 | * (2) if sb or fp are not NULL, it emits the c_style quoted version |
| 247 | * of name, enclosed with double quotes if asked and needed only. |
| 248 | * Return value is the same as in (1). |
| 249 | */ |
| 250 | static size_t quote_c_style_counted(const char *name, ssize_t maxlen, |
| 251 | struct strbuf *sb, FILE *fp, unsigned flags) |
| 252 | { |
| 253 | #undef EMIT |
| 254 | #define EMIT(c) \ |
| 255 | do { \ |
| 256 | if (sb) strbuf_addch(sb, (c)); \ |
| 257 | if (fp) fputc((c), fp); \ |
| 258 | count++; \ |
| 259 | } while (0) |
| 260 | #define EMITBUF(s, l) \ |
| 261 | do { \ |
| 262 | if (sb) strbuf_add(sb, (s), (l)); \ |
| 263 | if (fp) fwrite((s), (l), 1, fp); \ |
| 264 | count += (l); \ |
| 265 | } while (0) |
| 266 | |
| 267 | int no_dq = !!(flags & CQUOTE_NODQ); |
| 268 | size_t len, count = 0; |
| 269 | const char *p = name; |
| 270 | |
| 271 | for (;;) { |
| 272 | int ch; |
| 273 | |
| 274 | len = next_quote_pos(p, maxlen); |
| 275 | if (len == maxlen || (maxlen < 0 && !p[len])) |
| 276 | break; |
| 277 | |
| 278 | if (!no_dq && p == name) |
| 279 | EMIT('"'); |
| 280 | |
| 281 | EMITBUF(p, len); |
| 282 | EMIT('\\'); |
| 283 | p += len; |
| 284 | ch = (unsigned char)*p++; |
| 285 | if (maxlen >= 0) |
| 286 | maxlen -= len + 1; |
| 287 | if (cq_lookup[ch] >= ' ') { |
| 288 | EMIT(cq_lookup[ch]); |
| 289 | } else { |
| 290 | EMIT(((ch >> 6) & 03) + '0'); |
| 291 | EMIT(((ch >> 3) & 07) + '0'); |
| 292 | EMIT(((ch >> 0) & 07) + '0'); |
| 293 | } |
| 294 | } |
| 295 | |
| 296 | EMITBUF(p, len); |
| 297 | if (p == name) /* no ending quote needed */ |
| 298 | return 0; |
| 299 | |
| 300 | if (!no_dq) |
| 301 | EMIT('"'); |
| 302 | return count; |
| 303 | } |
| 304 | |
| 305 | size_t quote_c_style(const char *name, struct strbuf *sb, FILE *fp, unsigned flags) |
| 306 | { |
| 307 | return quote_c_style_counted(name, -1, sb, fp, flags); |
| 308 | } |
| 309 | |
| 310 | void quote_two_c_style(struct strbuf *sb, const char *prefix, const char *path, |
| 311 | unsigned flags) |
| 312 | { |
| 313 | int nodq = !!(flags & CQUOTE_NODQ); |
| 314 | if (quote_c_style(prefix, NULL, NULL, 0) || |
| 315 | quote_c_style(path, NULL, NULL, 0)) { |
| 316 | if (!nodq) |
| 317 | strbuf_addch(sb, '"'); |
| 318 | quote_c_style(prefix, sb, NULL, CQUOTE_NODQ); |
| 319 | quote_c_style(path, sb, NULL, CQUOTE_NODQ); |
| 320 | if (!nodq) |
| 321 | strbuf_addch(sb, '"'); |
| 322 | } else { |
| 323 | strbuf_addstr(sb, prefix); |
| 324 | strbuf_addstr(sb, path); |
| 325 | } |
| 326 | } |
| 327 | |
| 328 | void write_name_quoted(const char *name, FILE *fp, int terminator) |
| 329 | { |
| 330 | if (terminator) { |
| 331 | quote_c_style(name, NULL, fp, 0); |
| 332 | } else { |
| 333 | fputs(name, fp); |
| 334 | } |
| 335 | fputc(terminator, fp); |
| 336 | } |
| 337 | |
| 338 | void write_name_quoted_relative(const char *name, const char *prefix, |
| 339 | FILE *fp, int terminator) |
| 340 | { |
| 341 | struct strbuf sb = STRBUF_INIT; |
| 342 | |
| 343 | name = relative_path(name, prefix, &sb); |
| 344 | write_name_quoted(name, fp, terminator); |
| 345 | |
| 346 | strbuf_release(&sb); |
| 347 | } |
| 348 | |
| 349 | /* quote path as relative to the given prefix */ |
| 350 | char *quote_path(const char *in, const char *prefix, struct strbuf *out, unsigned flags) |
| 351 | { |
| 352 | struct strbuf sb = STRBUF_INIT; |
| 353 | const char *rel = relative_path(in, prefix, &sb); |
| 354 | int force_dq = ((flags & QUOTE_PATH_QUOTE_SP) && strchr(rel, ' ')); |
| 355 | |
| 356 | strbuf_reset(out); |
| 357 | |
| 358 | /* |
| 359 | * If the caller wants us to enclose the output in a dq-pair |
| 360 | * whether quote_c_style_counted() needs to, we do it ourselves |
| 361 | * and tell quote_c_style_counted() not to. |
| 362 | */ |
| 363 | if (force_dq) |
| 364 | strbuf_addch(out, '"'); |
| 365 | quote_c_style_counted(rel, strlen(rel), out, NULL, |
| 366 | force_dq ? CQUOTE_NODQ : 0); |
| 367 | if (force_dq) |
| 368 | strbuf_addch(out, '"'); |
| 369 | strbuf_release(&sb); |
| 370 | |
| 371 | return out->buf; |
| 372 | } |
| 373 | |
| 374 | /* |
| 375 | * C-style name unquoting. |
| 376 | * |
| 377 | * Quoted should point at the opening double quote. |
| 378 | * + Returns 0 if it was able to unquote the string properly, and appends the |
| 379 | * result in the strbuf `sb'. |
| 380 | * + Returns -1 in case of error, and doesn't touch the strbuf. Though note |
| 381 | * that this function will allocate memory in the strbuf, so calling |
| 382 | * strbuf_release is mandatory whichever result unquote_c_style returns. |
| 383 | * |
| 384 | * Updates endp pointer to point at one past the ending double quote if given. |
| 385 | */ |
| 386 | int unquote_c_style(struct strbuf *sb, const char *quoted, const char **endp) |
| 387 | { |
| 388 | size_t oldlen = sb->len, len; |
| 389 | int ch, ac; |
| 390 | |
| 391 | if (*quoted++ != '"') |
| 392 | return -1; |
| 393 | |
| 394 | for (;;) { |
| 395 | len = strcspn(quoted, "\"\\"); |
| 396 | strbuf_add(sb, quoted, len); |
| 397 | quoted += len; |
| 398 | |
| 399 | switch (*quoted++) { |
| 400 | case '"': |
| 401 | if (endp) |
| 402 | *endp = quoted; |
| 403 | return 0; |
| 404 | case '\\': |
| 405 | break; |
| 406 | default: |
| 407 | goto error; |
| 408 | } |
| 409 | |
| 410 | switch ((ch = *quoted++)) { |
| 411 | case 'a': ch = '\a'; break; |
| 412 | case 'b': ch = '\b'; break; |
| 413 | case 'f': ch = '\f'; break; |
| 414 | case 'n': ch = '\n'; break; |
| 415 | case 'r': ch = '\r'; break; |
| 416 | case 't': ch = '\t'; break; |
| 417 | case 'v': ch = '\v'; break; |
| 418 | |
| 419 | case '\\': case '"': |
| 420 | break; /* verbatim */ |
| 421 | |
| 422 | /* octal values with first digit over 4 overflow */ |
| 423 | case '0': case '1': case '2': case '3': |
| 424 | ac = ((ch - '0') << 6); |
| 425 | if ((ch = *quoted++) < '0' || '7' < ch) |
| 426 | goto error; |
| 427 | ac |= ((ch - '0') << 3); |
| 428 | if ((ch = *quoted++) < '0' || '7' < ch) |
| 429 | goto error; |
| 430 | ac |= (ch - '0'); |
| 431 | ch = ac; |
| 432 | break; |
| 433 | default: |
| 434 | goto error; |
| 435 | } |
| 436 | strbuf_addch(sb, ch); |
| 437 | } |
| 438 | |
| 439 | error: |
| 440 | strbuf_setlen(sb, oldlen); |
| 441 | return -1; |
| 442 | } |
| 443 | |
| 444 | /* quoting as a string literal for other languages */ |
| 445 | |
| 446 | void perl_quote_buf(struct strbuf *sb, const char *src) |
| 447 | { |
| 448 | const char sq = '\''; |
| 449 | const char bq = '\\'; |
| 450 | char c; |
| 451 | |
| 452 | strbuf_addch(sb, sq); |
| 453 | while ((c = *src++)) { |
| 454 | if (c == sq || c == bq) |
| 455 | strbuf_addch(sb, bq); |
| 456 | strbuf_addch(sb, c); |
| 457 | } |
| 458 | strbuf_addch(sb, sq); |
| 459 | } |
| 460 | |
| 461 | void perl_quote_buf_with_len(struct strbuf *sb, const char *src, size_t len) |
| 462 | { |
| 463 | const char sq = '\''; |
| 464 | const char bq = '\\'; |
| 465 | const char *c = src; |
| 466 | const char *end = src + len; |
| 467 | |
| 468 | strbuf_addch(sb, sq); |
| 469 | while (c != end) { |
| 470 | if (*c == sq || *c == bq) |
| 471 | strbuf_addch(sb, bq); |
| 472 | strbuf_addch(sb, *c); |
| 473 | c++; |
| 474 | } |
| 475 | strbuf_addch(sb, sq); |
| 476 | } |
| 477 | |
| 478 | void python_quote_buf(struct strbuf *sb, const char *src) |
| 479 | { |
| 480 | const char sq = '\''; |
| 481 | const char bq = '\\'; |
| 482 | const char nl = '\n'; |
| 483 | char c; |
| 484 | |
| 485 | strbuf_addch(sb, sq); |
| 486 | while ((c = *src++)) { |
| 487 | if (c == nl) { |
| 488 | strbuf_addch(sb, bq); |
| 489 | strbuf_addch(sb, 'n'); |
| 490 | continue; |
| 491 | } |
| 492 | if (c == sq || c == bq) |
| 493 | strbuf_addch(sb, bq); |
| 494 | strbuf_addch(sb, c); |
| 495 | } |
| 496 | strbuf_addch(sb, sq); |
| 497 | } |
| 498 | |
| 499 | void tcl_quote_buf(struct strbuf *sb, const char *src) |
| 500 | { |
| 501 | char c; |
| 502 | |
| 503 | strbuf_addch(sb, '"'); |
| 504 | while ((c = *src++)) { |
| 505 | switch (c) { |
| 506 | case '[': case ']': |
| 507 | case '{': case '}': |
| 508 | case '$': case '\\': case '"': |
| 509 | strbuf_addch(sb, '\\'); |
| 510 | /* fallthrough */ |
| 511 | default: |
| 512 | strbuf_addch(sb, c); |
| 513 | break; |
| 514 | case '\f': |
| 515 | strbuf_addstr(sb, "\\f"); |
| 516 | break; |
| 517 | case '\r': |
| 518 | strbuf_addstr(sb, "\\r"); |
| 519 | break; |
| 520 | case '\n': |
| 521 | strbuf_addstr(sb, "\\n"); |
| 522 | break; |
| 523 | case '\t': |
| 524 | strbuf_addstr(sb, "\\t"); |
| 525 | break; |
| 526 | case '\v': |
| 527 | strbuf_addstr(sb, "\\v"); |
| 528 | break; |
| 529 | } |
| 530 | } |
| 531 | strbuf_addch(sb, '"'); |
| 532 | } |
| 533 | |
| 534 | void basic_regex_quote_buf(struct strbuf *sb, const char *src) |
| 535 | { |
| 536 | char c; |
| 537 | |
| 538 | if (*src == '^') { |
| 539 | /* only beginning '^' is special and needs quoting */ |
| 540 | strbuf_addch(sb, '\\'); |
| 541 | strbuf_addch(sb, *src++); |
| 542 | } |
| 543 | if (*src == '*') |
| 544 | /* beginning '*' is not special, no quoting */ |
| 545 | strbuf_addch(sb, *src++); |
| 546 | |
| 547 | while ((c = *src++)) { |
| 548 | switch (c) { |
| 549 | case '[': |
| 550 | case '.': |
| 551 | case '\\': |
| 552 | case '*': |
| 553 | strbuf_addch(sb, '\\'); |
| 554 | strbuf_addch(sb, c); |
| 555 | break; |
| 556 | |
| 557 | case '$': |
| 558 | /* only the end '$' is special and needs quoting */ |
| 559 | if (*src == '\0') |
| 560 | strbuf_addch(sb, '\\'); |
| 561 | strbuf_addch(sb, c); |
| 562 | break; |
| 563 | |
| 564 | default: |
| 565 | strbuf_addch(sb, c); |
| 566 | break; |
| 567 | } |
| 568 | } |
| 569 | } |