quote.c

   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 static int sq_dequote_to_argv_internal(char *arg,
 175                                        const char ***argv, int *nr, int *alloc,
 176                                        struct strvec *array)
 177 {
 178         char *next = arg;
 179
 180         if (!*arg)
 181                 return 0;
 182         do {
 183                 char *dequoted = sq_dequote_step(next, &next);
 184                 if (!dequoted)
 185                         return -1;
 186                 if (next) {
 187                         char c;
 188                         if (!isspace(*next))
 189                                 return -1;
 190                         do {
 191                                 c = *++next;
 192                         } while (isspace(c));
 193                 }
 194                 if (argv) {
 195                         ALLOC_GROW(*argv, *nr + 1, *alloc);
 196                         (*argv)[(*nr)++] = dequoted;
 197                 }
 198                 if (array)
 199                         strvec_push(array, dequoted);
 200         } while (next);
 201
 202         return 0;
 203 }
 204
 205 int sq_dequote_to_argv(char *arg, const char ***argv, int *nr, int *alloc)
 206 {
 207         return sq_dequote_to_argv_internal(arg, argv, nr, alloc, NULL);
 208 }
 209
 210 int sq_dequote_to_strvec(char *arg, struct strvec *array)
 211 {
 212         return sq_dequote_to_argv_internal(arg, NULL, NULL, NULL, array);
 213 }
 214
 215 /* 1 means: quote as octal
 216  * 0 means: quote as octal if (quote_path_fully)
 217  * -1 means: never quote
 218  * c: quote as "\\c"
 219  */
 220 #define X8(x)   x, x, x, x, x, x, x, x
 221 #define X16(x)  X8(x), X8(x)
 222 static signed char const cq_lookup[256] = {
 223         /*           0    1    2    3    4    5    6    7 */
 224         /* 0x00 */   1,   1,   1,   1,   1,   1,   1, 'a',
 225         /* 0x08 */ 'b', 't', 'n', 'v', 'f', 'r',   1,   1,
 226         /* 0x10 */ X16(1),
 227         /* 0x20 */  -1,  -1, '"',  -1,  -1,  -1,  -1,  -1,
 228         /* 0x28 */ X16(-1), X16(-1), X16(-1),
 229         /* 0x58 */  -1,  -1,  -1,  -1,'\\',  -1,  -1,  -1,
 230         /* 0x60 */ X16(-1), X8(-1),
 231         /* 0x78 */  -1,  -1,  -1,  -1,  -1,  -1,  -1,   1,
 232         /* 0x80 */ /* set to 0 */
 233 };
 234
 235 static inline int cq_must_quote(char c)
 236 {
 237         return cq_lookup[(unsigned char)c] + quote_path_fully > 0;
 238 }
 239
 240 /* returns the longest prefix not needing a quote up to maxlen if positive.
 241    This stops at the first \0 because it's marked as a character needing an
 242    escape */
 243 static size_t next_quote_pos(const char *s, ssize_t maxlen)
 244 {
 245         size_t len;
 246         if (maxlen < 0) {
 247                 for (len = 0; !cq_must_quote(s[len]); len++);
 248         } else {
 249                 for (len = 0; len < maxlen && !cq_must_quote(s[len]); len++);
 250         }
 251         return len;
 252 }
 253
 254 /*
 255  * C-style name quoting.
 256  *
 257  * (1) if sb and fp are both NULL, inspect the input name and counts the
 258  *     number of bytes that are needed to hold c_style quoted version of name,
 259  *     counting the double quotes around it but not terminating NUL, and
 260  *     returns it.
 261  *     However, if name does not need c_style quoting, it returns 0.
 262  *
 263  * (2) if sb or fp are not NULL, it emits the c_style quoted version
 264  *     of name, enclosed with double quotes if asked and needed only.
 265  *     Return value is the same as in (1).
 266  */
 267 static size_t quote_c_style_counted(const char *name, ssize_t maxlen,
 268                                     struct strbuf *sb, FILE *fp, unsigned flags)
 269 {
 270 #undef EMIT
 271 #define EMIT(c)                                 \
 272         do {                                        \
 273                 if (sb) strbuf_addch(sb, (c));          \
 274                 if (fp) fputc((c), fp);                 \
 275                 count++;                                \
 276         } while (0)
 277 #define EMITBUF(s, l)                           \
 278         do {                                        \
 279                 if (sb) strbuf_add(sb, (s), (l));       \
 280                 if (fp) fwrite((s), (l), 1, fp);        \
 281                 count += (l);                           \
 282         } while (0)
 283
 284         int no_dq = !!(flags & CQUOTE_NODQ);
 285         size_t len, count = 0;
 286         const char *p = name;
 287
 288         for (;;) {
 289                 int ch;
 290
 291                 len = next_quote_pos(p, maxlen);
 292                 if (len == maxlen || (maxlen < 0 && !p[len]))
 293                         break;
 294
 295                 if (!no_dq && p == name)
 296                         EMIT('"');
 297
 298                 EMITBUF(p, len);
 299                 EMIT('\\');
 300                 p += len;
 301                 ch = (unsigned char)*p++;
 302                 if (maxlen >= 0)
 303                         maxlen -= len + 1;
 304                 if (cq_lookup[ch] >= ' ') {
 305                         EMIT(cq_lookup[ch]);
 306                 } else {
 307                         EMIT(((ch >> 6) & 03) + '0');
 308                         EMIT(((ch >> 3) & 07) + '0');
 309                         EMIT(((ch >> 0) & 07) + '0');
 310                 }
 311         }
 312
 313         EMITBUF(p, len);
 314         if (p == name)   /* no ending quote needed */
 315                 return 0;
 316
 317         if (!no_dq)
 318                 EMIT('"');
 319         return count;
 320 }
 321
 322 size_t quote_c_style(const char *name, struct strbuf *sb, FILE *fp, unsigned flags)
 323 {
 324         return quote_c_style_counted(name, -1, sb, fp, flags);
 325 }
 326
 327 void quote_two_c_style(struct strbuf *sb, const char *prefix, const char *path,
 328                        unsigned flags)
 329 {
 330         int nodq = !!(flags & CQUOTE_NODQ);
 331         if (quote_c_style(prefix, NULL, NULL, 0) ||
 332             quote_c_style(path, NULL, NULL, 0)) {
 333                 if (!nodq)
 334                         strbuf_addch(sb, '"');
 335                 quote_c_style(prefix, sb, NULL, CQUOTE_NODQ);
 336                 quote_c_style(path, sb, NULL, CQUOTE_NODQ);
 337                 if (!nodq)
 338                         strbuf_addch(sb, '"');
 339         } else {
 340                 strbuf_addstr(sb, prefix);
 341                 strbuf_addstr(sb, path);
 342         }
 343 }
 344
 345 void write_name_quoted(const char *name, FILE *fp, int terminator)
 346 {
 347         if (terminator) {
 348                 quote_c_style(name, NULL, fp, 0);
 349         } else {
 350                 fputs(name, fp);
 351         }
 352         fputc(terminator, fp);
 353 }
 354
 355 void write_name_quoted_relative(const char *name, const char *prefix,
 356                                 FILE *fp, int terminator)
 357 {
 358         struct strbuf sb = STRBUF_INIT;
 359
 360         name = relative_path(name, prefix, &sb);
 361         write_name_quoted(name, fp, terminator);
 362
 363         strbuf_release(&sb);
 364 }
 365
 366 /* quote path as relative to the given prefix */
 367 char *quote_path(const char *in, const char *prefix, struct strbuf *out, unsigned flags)
 368 {
 369         struct strbuf sb = STRBUF_INIT;
 370         const char *rel = relative_path(in, prefix, &sb);
 371         int force_dq = ((flags & QUOTE_PATH_QUOTE_SP) && strchr(rel, ' '));
 372
 373         strbuf_reset(out);
 374
 375         /*
 376          * If the caller wants us to enclose the output in a dq-pair
 377          * whether quote_c_style_counted() needs to, we do it ourselves
 378          * and tell quote_c_style_counted() not to.
 379          */
 380         if (force_dq)
 381                 strbuf_addch(out, '"');
 382         quote_c_style_counted(rel, strlen(rel), out, NULL,
 383                               force_dq ? CQUOTE_NODQ : 0);
 384         if (force_dq)
 385                 strbuf_addch(out, '"');
 386         strbuf_release(&sb);
 387
 388         return out->buf;
 389 }
 390
 391 /*
 392  * C-style name unquoting.
 393  *
 394  * Quoted should point at the opening double quote.
 395  * + Returns 0 if it was able to unquote the string properly, and appends the
 396  *   result in the strbuf `sb'.
 397  * + Returns -1 in case of error, and doesn't touch the strbuf. Though note
 398  *   that this function will allocate memory in the strbuf, so calling
 399  *   strbuf_release is mandatory whichever result unquote_c_style returns.
 400  *
 401  * Updates endp pointer to point at one past the ending double quote if given.
 402  */
 403 int unquote_c_style(struct strbuf *sb, const char *quoted, const char **endp)
 404 {
 405         size_t oldlen = sb->len, len;
 406         int ch, ac;
 407
 408         if (*quoted++ != '"')
 409                 return -1;
 410
 411         for (;;) {
 412                 len = strcspn(quoted, "\"\\");
 413                 strbuf_add(sb, quoted, len);
 414                 quoted += len;
 415
 416                 switch (*quoted++) {
 417                   case '"':
 418                         if (endp)
 419                                 *endp = quoted;
 420                         return 0;
 421                   case '\\':
 422                         break;
 423                   default:
 424                         goto error;
 425                 }
 426
 427                 switch ((ch = *quoted++)) {
 428                 case 'a': ch = '\a'; break;
 429                 case 'b': ch = '\b'; break;
 430                 case 'f': ch = '\f'; break;
 431                 case 'n': ch = '\n'; break;
 432                 case 'r': ch = '\r'; break;
 433                 case 't': ch = '\t'; break;
 434                 case 'v': ch = '\v'; break;
 435
 436                 case '\\': case '"':
 437                         break; /* verbatim */
 438
 439                 /* octal values with first digit over 4 overflow */
 440                 case '0': case '1': case '2': case '3':
 441                                         ac = ((ch - '0') << 6);
 442                         if ((ch = *quoted++) < '0' || '7' < ch)
 443                                 goto error;
 444                                         ac |= ((ch - '0') << 3);
 445                         if ((ch = *quoted++) < '0' || '7' < ch)
 446                                 goto error;
 447                                         ac |= (ch - '0');
 448                                         ch = ac;
 449                                         break;
 450                                 default:
 451                         goto error;
 452                         }
 453                 strbuf_addch(sb, ch);
 454                 }
 455
 456   error:
 457         strbuf_setlen(sb, oldlen);
 458         return -1;
 459 }
 460
 461 /* quoting as a string literal for other languages */
 462
 463 void perl_quote_buf(struct strbuf *sb, const char *src)
 464 {
 465         const char sq = '\'';
 466         const char bq = '\\';
 467         char c;
 468
 469         strbuf_addch(sb, sq);
 470         while ((c = *src++)) {
 471                 if (c == sq || c == bq)
 472                         strbuf_addch(sb, bq);
 473                 strbuf_addch(sb, c);
 474         }
 475         strbuf_addch(sb, sq);
 476 }
 477
 478 void perl_quote_buf_with_len(struct strbuf *sb, const char *src, size_t len)
 479 {
 480         const char sq = '\'';
 481         const char bq = '\\';
 482         const char *c = src;
 483         const char *end = src + len;
 484
 485         strbuf_addch(sb, sq);
 486         while (c != end) {
 487                 if (*c == sq || *c == bq)
 488                         strbuf_addch(sb, bq);
 489                 strbuf_addch(sb, *c);
 490                 c++;
 491         }
 492         strbuf_addch(sb, sq);
 493 }
 494
 495 void python_quote_buf(struct strbuf *sb, const char *src)
 496 {
 497         const char sq = '\'';
 498         const char bq = '\\';
 499         const char nl = '\n';
 500         char c;
 501
 502         strbuf_addch(sb, sq);
 503         while ((c = *src++)) {
 504                 if (c == nl) {
 505                         strbuf_addch(sb, bq);
 506                         strbuf_addch(sb, 'n');
 507                         continue;
 508                 }
 509                 if (c == sq || c == bq)
 510                         strbuf_addch(sb, bq);
 511                 strbuf_addch(sb, c);
 512         }
 513         strbuf_addch(sb, sq);
 514 }
 515
 516 void tcl_quote_buf(struct strbuf *sb, const char *src)
 517 {
 518         char c;
 519
 520         strbuf_addch(sb, '"');
 521         while ((c = *src++)) {
 522                 switch (c) {
 523                 case '[': case ']':
 524                 case '{': case '}':
 525                 case '$': case '\\': case '"':
 526                         strbuf_addch(sb, '\\');
 527                         /* fallthrough */
 528                 default:
 529                         strbuf_addch(sb, c);
 530                         break;
 531                 case '\f':
 532                         strbuf_addstr(sb, "\\f");
 533                         break;
 534                 case '\r':
 535                         strbuf_addstr(sb, "\\r");
 536                         break;
 537                 case '\n':
 538                         strbuf_addstr(sb, "\\n");
 539                         break;
 540                 case '\t':
 541                         strbuf_addstr(sb, "\\t");
 542                         break;
 543                 case '\v':
 544                         strbuf_addstr(sb, "\\v");
 545                         break;
 546                 }
 547         }
 548         strbuf_addch(sb, '"');
 549 }
 550
 551 void basic_regex_quote_buf(struct strbuf *sb, const char *src)
 552 {
 553         char c;
 554
 555         if (*src == '^') {
 556                 /* only beginning '^' is special and needs quoting */
 557                 strbuf_addch(sb, '\\');
 558                 strbuf_addch(sb, *src++);
 559         }
 560         if (*src == '*')
 561                 /* beginning '*' is not special, no quoting */
 562                 strbuf_addch(sb, *src++);
 563
 564         while ((c = *src++)) {
 565                 switch (c) {
 566                 case '[':
 567                 case '.':
 568                 case '\\':
 569                 case '*':
 570                         strbuf_addch(sb, '\\');
 571                         strbuf_addch(sb, c);
 572                         break;
 573
 574                 case '$':
 575                         /* only the end '$' is special and needs quoting */
 576                         if (*src == '\0')
 577                                 strbuf_addch(sb, '\\');
 578                         strbuf_addch(sb, c);
 579                         break;
 580
 581                 default:
 582                         strbuf_addch(sb, c);
 583                         break;
 584                 }
 585         }
 586 }