quote.c

   1 #include "git-compat-util.h"
   2 #include "path.h"
   3 #include "quote.h"
   4 #include "strbuf.h"
   5 #include "strvec.h"
   6
   7 int quote_path_fully = 1;
   8
   9 static inline int need_bs_quote(char c)
  10 {
  11         return (c == '\'' || c == '!');
  12 }
  13
  14 /* Help to copy the thing properly quoted for the shell safety.
  15  * any single quote is replaced with '\'', any exclamation point
  16  * is replaced with '\!', and the whole thing is enclosed in a
  17  * single quote pair.
  18  *
  19  * E.g.
  20  *  original     sq_quote     result
  21  *  name     ==> name      ==> 'name'
  22  *  a b      ==> a b       ==> 'a b'
  23  *  a'b      ==> a'\''b    ==> 'a'\''b'
  24  *  a!b      ==> a'\!'b    ==> 'a'\!'b'
  25  */
  26 void sq_quote_buf(struct strbuf *dst, const char *src)
  27 {
  28         char *to_free = NULL;
  29
  30         if (dst->buf == src)
  31                 to_free = strbuf_detach(dst, NULL);
  32
  33         strbuf_addch(dst, '\'');
  34         while (*src) {
  35                 size_t len = strcspn(src, "'!");
  36                 strbuf_add(dst, src, len);
  37                 src += len;
  38                 while (need_bs_quote(*src)) {
  39                         strbuf_addstr(dst, "'\\");
  40                         strbuf_addch(dst, *src++);
  41                         strbuf_addch(dst, '\'');
  42                 }
  43         }
  44         strbuf_addch(dst, '\'');
  45         free(to_free);
  46 }
  47
  48 void sq_quote_buf_pretty(struct strbuf *dst, const char *src)
  49 {
  50         static const char ok_punct[] = "+,-./:=@_^";
  51         const char *p;
  52
  53         /* Avoid losing a zero-length string by adding '' */
  54         if (!*src) {
  55                 strbuf_addstr(dst, "''");
  56                 return;
  57         }
  58
  59         for (p = src; *p; p++) {
  60                 if (!isalnum(*p) && !strchr(ok_punct, *p)) {
  61                         sq_quote_buf(dst, src);
  62                         return;
  63                 }
  64         }
  65
  66         /* if we get here, we did not need quoting */
  67         strbuf_addstr(dst, src);
  68 }
  69
  70 void sq_quotef(struct strbuf *dst, const char *fmt, ...)
  71 {
  72         struct strbuf src = STRBUF_INIT;
  73
  74         va_list ap;
  75         va_start(ap, fmt);
  76         strbuf_vaddf(&src, fmt, ap);
  77         va_end(ap);
  78
  79         sq_quote_buf(dst, src.buf);
  80         strbuf_release(&src);
  81 }
  82
  83 void sq_quote_argv(struct strbuf *dst, const char **argv)
  84 {
  85         int i;
  86
  87         /* Copy into destination buffer. */
  88         strbuf_grow(dst, 255);
  89         for (i = 0; argv[i]; ++i) {
  90                 strbuf_addch(dst, ' ');
  91                 sq_quote_buf(dst, argv[i]);
  92         }
  93 }
  94
  95 /*
  96  * Legacy function to append each argv value, quoted as necessasry,
  97  * with whitespace before each value.  This results in a leading
  98  * space in the result.
  99  */
 100 void sq_quote_argv_pretty(struct strbuf *dst, const char **argv)
 101 {
 102         if (argv[0])
 103                 strbuf_addch(dst, ' ');
 104         sq_append_quote_argv_pretty(dst, argv);
 105 }
 106
 107 /*
 108  * Append each argv value, quoted as necessary, with whitespace between them.
 109  */
 110 void sq_append_quote_argv_pretty(struct strbuf *dst, const char **argv)
 111 {
 112         int i;
 113
 114         for (i = 0; argv[i]; i++) {
 115                 if (i > 0)
 116                         strbuf_addch(dst, ' ');
 117                 sq_quote_buf_pretty(dst, argv[i]);
 118         }
 119 }
 120
 121 char *sq_dequote_step(char *arg, char **next)
 122 {
 123         char *dst = arg;
 124         char *src = arg;
 125         char c;
 126
 127         if (*src != '\'')
 128                 return NULL;
 129         for (;;) {
 130                 c = *++src;
 131                 if (!c)
 132                         return NULL;
 133                 if (c != '\'') {
 134                         *dst++ = c;
 135                         continue;
 136                 }
 137                 /* We stepped out of sq */
 138                 switch (*++src) {
 139                 case '\0':
 140                         *dst = 0;
 141                         if (next)
 142                                 *next = NULL;
 143                         return arg;
 144                 case '\\':
 145                         /*
 146                          * Allow backslashed characters outside of
 147                          * single-quotes only if they need escaping,
 148                          * and only if we resume the single-quoted part
 149                          * afterward.
 150                          */
 151                         if (need_bs_quote(src[1]) && src[2] == '\'') {
 152                                 *dst++ = src[1];
 153                                 src += 2;
 154                                 continue;
 155                         }
 156                 /* Fallthrough */
 157                 default:
 158                         if (!next)
 159                                 return NULL;
 160                         *dst = 0;
 161                         *next = src;
 162                         return arg;
 163                 }
 164         }
 165 }
 166
 167 char *sq_dequote(char *arg)
 168 {
 169         return sq_dequote_step(arg, NULL);
 170 }
 171
 172 static int sq_dequote_to_argv_internal(char *arg,
 173                                        const char ***argv, int *nr, int *alloc,
 174                                        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                 if (argv) {
 193                         ALLOC_GROW(*argv, *nr + 1, *alloc);
 194                         (*argv)[(*nr)++] = dequoted;
 195                 }
 196                 if (array)
 197                         strvec_push(array, dequoted);
 198         } while (next);
 199
 200         return 0;
 201 }
 202
 203 int sq_dequote_to_argv(char *arg, const char ***argv, int *nr, int *alloc)
 204 {
 205         return sq_dequote_to_argv_internal(arg, argv, nr, alloc, NULL);
 206 }
 207
 208 int sq_dequote_to_strvec(char *arg, struct strvec *array)
 209 {
 210         return sq_dequote_to_argv_internal(arg, NULL, NULL, NULL, array);
 211 }
 212
 213 /* 1 means: quote as octal
 214  * 0 means: quote as octal if (quote_path_fully)
 215  * -1 means: never quote
 216  * c: quote as "\\c"
 217  */
 218 #define X8(x)   x, x, x, x, x, x, x, x
 219 #define X16(x)  X8(x), X8(x)
 220 static signed char const cq_lookup[256] = {
 221         /*           0    1    2    3    4    5    6    7 */
 222         /* 0x00 */   1,   1,   1,   1,   1,   1,   1, 'a',
 223         /* 0x08 */ 'b', 't', 'n', 'v', 'f', 'r',   1,   1,
 224         /* 0x10 */ X16(1),
 225         /* 0x20 */  -1,  -1, '"',  -1,  -1,  -1,  -1,  -1,
 226         /* 0x28 */ X16(-1), X16(-1), X16(-1),
 227         /* 0x58 */  -1,  -1,  -1,  -1,'\\',  -1,  -1,  -1,
 228         /* 0x60 */ X16(-1), X8(-1),
 229         /* 0x78 */  -1,  -1,  -1,  -1,  -1,  -1,  -1,   1,
 230         /* 0x80 */ /* set to 0 */
 231 };
 232
 233 static inline int cq_must_quote(char c)
 234 {
 235         return cq_lookup[(unsigned char)c] + quote_path_fully > 0;
 236 }
 237
 238 /* returns the longest prefix not needing a quote up to maxlen if positive.
 239    This stops at the first \0 because it's marked as a character needing an
 240    escape */
 241 static size_t next_quote_pos(const char *s, ssize_t maxlen)
 242 {
 243         size_t len;
 244         if (maxlen < 0) {
 245                 for (len = 0; !cq_must_quote(s[len]); len++);
 246         } else {
 247                 for (len = 0; len < maxlen && !cq_must_quote(s[len]); len++);
 248         }
 249         return len;
 250 }
 251
 252 /*
 253  * C-style name quoting.
 254  *
 255  * (1) if sb and fp are both NULL, inspect the input name and counts the
 256  *     number of bytes that are needed to hold c_style quoted version of name,
 257  *     counting the double quotes around it but not terminating NUL, and
 258  *     returns it.
 259  *     However, if name does not need c_style quoting, it returns 0.
 260  *
 261  * (2) if sb or fp are not NULL, it emits the c_style quoted version
 262  *     of name, enclosed with double quotes if asked and needed only.
 263  *     Return value is the same as in (1).
 264  */
 265 static size_t quote_c_style_counted(const char *name, ssize_t maxlen,
 266                                     struct strbuf *sb, FILE *fp, unsigned flags)
 267 {
 268 #undef EMIT
 269 #define EMIT(c)                                 \
 270         do {                                        \
 271                 if (sb) strbuf_addch(sb, (c));          \
 272                 if (fp) fputc((c), fp);                 \
 273                 count++;                                \
 274         } while (0)
 275 #define EMITBUF(s, l)                           \
 276         do {                                        \
 277                 if (sb) strbuf_add(sb, (s), (l));       \
 278                 if (fp) fwrite((s), (l), 1, fp);        \
 279                 count += (l);                           \
 280         } while (0)
 281
 282         int no_dq = !!(flags & CQUOTE_NODQ);
 283         size_t len, count = 0;
 284         const char *p = name;
 285
 286         for (;;) {
 287                 int ch;
 288
 289                 len = next_quote_pos(p, maxlen);
 290                 if (len == maxlen || (maxlen < 0 && !p[len]))
 291                         break;
 292
 293                 if (!no_dq && p == name)
 294                         EMIT('"');
 295
 296                 EMITBUF(p, len);
 297                 EMIT('\\');
 298                 p += len;
 299                 ch = (unsigned char)*p++;
 300                 if (maxlen >= 0)
 301                         maxlen -= len + 1;
 302                 if (cq_lookup[ch] >= ' ') {
 303                         EMIT(cq_lookup[ch]);
 304                 } else {
 305                         EMIT(((ch >> 6) & 03) + '0');
 306                         EMIT(((ch >> 3) & 07) + '0');
 307                         EMIT(((ch >> 0) & 07) + '0');
 308                 }
 309         }
 310
 311         EMITBUF(p, len);
 312         if (p == name)   /* no ending quote needed */
 313                 return 0;
 314
 315         if (!no_dq)
 316                 EMIT('"');
 317         return count;
 318 }
 319
 320 size_t quote_c_style(const char *name, struct strbuf *sb, FILE *fp, unsigned flags)
 321 {
 322         return quote_c_style_counted(name, -1, sb, fp, flags);
 323 }
 324
 325 void quote_two_c_style(struct strbuf *sb, const char *prefix, const char *path,
 326                        unsigned flags)
 327 {
 328         int nodq = !!(flags & CQUOTE_NODQ);
 329         if (quote_c_style(prefix, NULL, NULL, 0) ||
 330             quote_c_style(path, NULL, NULL, 0)) {
 331                 if (!nodq)
 332                         strbuf_addch(sb, '"');
 333                 quote_c_style(prefix, sb, NULL, CQUOTE_NODQ);
 334                 quote_c_style(path, sb, NULL, CQUOTE_NODQ);
 335                 if (!nodq)
 336                         strbuf_addch(sb, '"');
 337         } else {
 338                 strbuf_addstr(sb, prefix);
 339                 strbuf_addstr(sb, path);
 340         }
 341 }
 342
 343 void write_name_quoted(const char *name, FILE *fp, int terminator)
 344 {
 345         if (terminator) {
 346                 quote_c_style(name, NULL, fp, 0);
 347         } else {
 348                 fputs(name, fp);
 349         }
 350         fputc(terminator, fp);
 351 }
 352
 353 void write_name_quoted_relative(const char *name, const char *prefix,
 354                                 FILE *fp, int terminator)
 355 {
 356         struct strbuf sb = STRBUF_INIT;
 357
 358         name = relative_path(name, prefix, &sb);
 359         write_name_quoted(name, fp, terminator);
 360
 361         strbuf_release(&sb);
 362 }
 363
 364 /* quote path as relative to the given prefix */
 365 char *quote_path(const char *in, const char *prefix, struct strbuf *out, unsigned flags)
 366 {
 367         struct strbuf sb = STRBUF_INIT;
 368         const char *rel = relative_path(in, prefix, &sb);
 369         int force_dq = ((flags & QUOTE_PATH_QUOTE_SP) && strchr(rel, ' '));
 370
 371         strbuf_reset(out);
 372
 373         /*
 374          * If the caller wants us to enclose the output in a dq-pair
 375          * whether quote_c_style_counted() needs to, we do it ourselves
 376          * and tell quote_c_style_counted() not to.
 377          */
 378         if (force_dq)
 379                 strbuf_addch(out, '"');
 380         quote_c_style_counted(rel, strlen(rel), out, NULL,
 381                               force_dq ? CQUOTE_NODQ : 0);
 382         if (force_dq)
 383                 strbuf_addch(out, '"');
 384         strbuf_release(&sb);
 385
 386         return out->buf;
 387 }
 388
 389 /*
 390  * C-style name unquoting.
 391  *
 392  * Quoted should point at the opening double quote.
 393  * + Returns 0 if it was able to unquote the string properly, and appends the
 394  *   result in the strbuf `sb'.
 395  * + Returns -1 in case of error, and doesn't touch the strbuf. Though note
 396  *   that this function will allocate memory in the strbuf, so calling
 397  *   strbuf_release is mandatory whichever result unquote_c_style returns.
 398  *
 399  * Updates endp pointer to point at one past the ending double quote if given.
 400  */
 401 int unquote_c_style(struct strbuf *sb, const char *quoted, const char **endp)
 402 {
 403         size_t oldlen = sb->len, len;
 404         int ch, ac;
 405
 406         if (*quoted++ != '"')
 407                 return -1;
 408
 409         for (;;) {
 410                 len = strcspn(quoted, "\"\\");
 411                 strbuf_add(sb, quoted, len);
 412                 quoted += len;
 413
 414                 switch (*quoted++) {
 415                   case '"':
 416                         if (endp)
 417                                 *endp = quoted;
 418                         return 0;
 419                   case '\\':
 420                         break;
 421                   default:
 422                         goto error;
 423                 }
 424
 425                 switch ((ch = *quoted++)) {
 426                 case 'a': ch = '\a'; break;
 427                 case 'b': ch = '\b'; break;
 428                 case 'f': ch = '\f'; break;
 429                 case 'n': ch = '\n'; break;
 430                 case 'r': ch = '\r'; break;
 431                 case 't': ch = '\t'; break;
 432                 case 'v': ch = '\v'; break;
 433
 434                 case '\\': case '"':
 435                         break; /* verbatim */
 436
 437                 /* octal values with first digit over 4 overflow */
 438                 case '0': case '1': case '2': case '3':
 439                                         ac = ((ch - '0') << 6);
 440                         if ((ch = *quoted++) < '0' || '7' < ch)
 441                                 goto error;
 442                                         ac |= ((ch - '0') << 3);
 443                         if ((ch = *quoted++) < '0' || '7' < ch)
 444                                 goto error;
 445                                         ac |= (ch - '0');
 446                                         ch = ac;
 447                                         break;
 448                                 default:
 449                         goto error;
 450                         }
 451                 strbuf_addch(sb, ch);
 452                 }
 453
 454   error:
 455         strbuf_setlen(sb, oldlen);
 456         return -1;
 457 }
 458
 459 /* quoting as a string literal for other languages */
 460
 461 void perl_quote_buf(struct strbuf *sb, const char *src)
 462 {
 463         const char sq = '\'';
 464         const char bq = '\\';
 465         char c;
 466
 467         strbuf_addch(sb, sq);
 468         while ((c = *src++)) {
 469                 if (c == sq || c == bq)
 470                         strbuf_addch(sb, bq);
 471                 strbuf_addch(sb, c);
 472         }
 473         strbuf_addch(sb, sq);
 474 }
 475
 476 void perl_quote_buf_with_len(struct strbuf *sb, const char *src, size_t len)
 477 {
 478         const char sq = '\'';
 479         const char bq = '\\';
 480         const char *c = src;
 481         const char *end = src + len;
 482
 483         strbuf_addch(sb, sq);
 484         while (c != end) {
 485                 if (*c == sq || *c == bq)
 486                         strbuf_addch(sb, bq);
 487                 strbuf_addch(sb, *c);
 488                 c++;
 489         }
 490         strbuf_addch(sb, sq);
 491 }
 492
 493 void python_quote_buf(struct strbuf *sb, const char *src)
 494 {
 495         const char sq = '\'';
 496         const char bq = '\\';
 497         const char nl = '\n';
 498         char c;
 499
 500         strbuf_addch(sb, sq);
 501         while ((c = *src++)) {
 502                 if (c == nl) {
 503                         strbuf_addch(sb, bq);
 504                         strbuf_addch(sb, 'n');
 505                         continue;
 506                 }
 507                 if (c == sq || c == bq)
 508                         strbuf_addch(sb, bq);
 509                 strbuf_addch(sb, c);
 510         }
 511         strbuf_addch(sb, sq);
 512 }
 513
 514 void tcl_quote_buf(struct strbuf *sb, const char *src)
 515 {
 516         char c;
 517
 518         strbuf_addch(sb, '"');
 519         while ((c = *src++)) {
 520                 switch (c) {
 521                 case '[': case ']':
 522                 case '{': case '}':
 523                 case '$': case '\\': case '"':
 524                         strbuf_addch(sb, '\\');
 525                         /* fallthrough */
 526                 default:
 527                         strbuf_addch(sb, c);
 528                         break;
 529                 case '\f':
 530                         strbuf_addstr(sb, "\\f");
 531                         break;
 532                 case '\r':
 533                         strbuf_addstr(sb, "\\r");
 534                         break;
 535                 case '\n':
 536                         strbuf_addstr(sb, "\\n");
 537                         break;
 538                 case '\t':
 539                         strbuf_addstr(sb, "\\t");
 540                         break;
 541                 case '\v':
 542                         strbuf_addstr(sb, "\\v");
 543                         break;
 544                 }
 545         }
 546         strbuf_addch(sb, '"');
 547 }
 548
 549 void basic_regex_quote_buf(struct strbuf *sb, const char *src)
 550 {
 551         char c;
 552
 553         if (*src == '^') {
 554                 /* only beginning '^' is special and needs quoting */
 555                 strbuf_addch(sb, '\\');
 556                 strbuf_addch(sb, *src++);
 557         }
 558         if (*src == '*')
 559                 /* beginning '*' is not special, no quoting */
 560                 strbuf_addch(sb, *src++);
 561
 562         while ((c = *src++)) {
 563                 switch (c) {
 564                 case '[':
 565                 case '.':
 566                 case '\\':
 567                 case '*':
 568                         strbuf_addch(sb, '\\');
 569                         strbuf_addch(sb, c);
 570                         break;
 571
 572                 case '$':
 573                         /* only the end '$' is special and needs quoting */
 574                         if (*src == '\0')
 575                                 strbuf_addch(sb, '\\');
 576                         strbuf_addch(sb, c);
 577                         break;
 578
 579                 default:
 580                         strbuf_addch(sb, c);
 581                         break;
 582                 }
 583         }
 584 }