quote.c

   1 #include "cache.h"
   2 #include "quote.h"
   3 #include "strvec.h"
   4
   5 int quote_path_fully = 1;
   6
   7 static inline int need_bs_quote(char c)
   8 {
   9         return (c == '\'' || c == '!');
  10 }
  11
  12 /* Help to copy the thing properly quoted for the shell safety.
  13  * any single quote is replaced with '\'', any exclamation point
  14  * is replaced with '\!', and the whole thing is enclosed in a
  15  * single quote pair.
  16  *
  17  * E.g.
  18  *  original     sq_quote     result
  19  *  name     ==> name      ==> 'name'
  20  *  a b      ==> a b       ==> 'a b'
  21  *  a'b      ==> a'\''b    ==> 'a'\''b'
  22  *  a!b      ==> a'\!'b    ==> 'a'\!'b'
  23  */
  24 void sq_quote_buf(struct strbuf *dst, const char *src)
  25 {
  26         char *to_free = NULL;
  27
  28         if (dst->buf == src)
  29                 to_free = strbuf_detach(dst, NULL);
  30
  31         strbuf_addch(dst, '\'');
  32         while (*src) {
  33                 size_t len = strcspn(src, "'!");
  34                 strbuf_add(dst, src, len);
  35                 src += len;
  36                 while (need_bs_quote(*src)) {
  37                         strbuf_addstr(dst, "'\\");
  38                         strbuf_addch(dst, *src++);
  39                         strbuf_addch(dst, '\'');
  40                 }
  41         }
  42         strbuf_addch(dst, '\'');
  43         free(to_free);
  44 }
  45
  46 void sq_quote_buf_pretty(struct strbuf *dst, const char *src)
  47 {
  48         static const char ok_punct[] = "+,-./:=@_^";
  49         const char *p;
  50
  51         /* Avoid losing a zero-length string by adding '' */
  52         if (!*src) {
  53                 strbuf_addstr(dst, "''");
  54                 return;
  55         }
  56
  57         for (p = src; *p; p++) {
  58                 if (!isalnum(*p) && !strchr(ok_punct, *p)) {
  59                         sq_quote_buf(dst, src);
  60                         return;
  61                 }
  62         }
  63
  64         /* if we get here, we did not need quoting */
  65         strbuf_addstr(dst, src);
  66 }
  67
  68 void sq_quotef(struct strbuf *dst, const char *fmt, ...)
  69 {
  70         struct strbuf src = STRBUF_INIT;
  71
  72         va_list ap;
  73         va_start(ap, fmt);
  74         strbuf_vaddf(&src, fmt, ap);
  75         va_end(ap);
  76
  77         sq_quote_buf(dst, src.buf);
  78         strbuf_release(&src);
  79 }
  80
  81 void sq_quote_argv(struct strbuf *dst, const char **argv)
  82 {
  83         int i;
  84
  85         /* Copy into destination buffer. */
  86         strbuf_grow(dst, 255);
  87         for (i = 0; argv[i]; ++i) {
  88                 strbuf_addch(dst, ' ');
  89                 sq_quote_buf(dst, argv[i]);
  90         }
  91 }
  92
  93 /*
  94  * Legacy function to append each argv value, quoted as necessasry,
  95  * with whitespace before each value.  This results in a leading
  96  * space in the result.
  97  */
  98 void sq_quote_argv_pretty(struct strbuf *dst, const char **argv)
  99 {
 100         if (argv[0])
 101                 strbuf_addch(dst, ' ');
 102         sq_append_quote_argv_pretty(dst, argv);
 103 }
 104
 105 /*
 106  * Append each argv value, quoted as necessary, with whitespace between them.
 107  */
 108 void sq_append_quote_argv_pretty(struct strbuf *dst, const char **argv)
 109 {
 110         int i;
 111
 112         for (i = 0; argv[i]; i++) {
 113                 if (i > 0)
 114                         strbuf_addch(dst, ' ');
 115                 sq_quote_buf_pretty(dst, argv[i]);
 116         }
 117 }
 118
 119 char *sq_dequote_step(char *arg, char **next)
 120 {
 121         char *dst = arg;
 122         char *src = arg;
 123         char c;
 124
 125         if (*src != '\'')
 126                 return NULL;
 127         for (;;) {
 128                 c = *++src;
 129                 if (!c)
 130                         return NULL;
 131                 if (c != '\'') {
 132                         *dst++ = c;
 133                         continue;
 134                 }
 135                 /* We stepped out of sq */
 136                 switch (*++src) {
 137                 case '\0':
 138                         *dst = 0;
 139                         if (next)
 140                                 *next = NULL;
 141                         return arg;
 142                 case '\\':
 143                         /*
 144                          * Allow backslashed characters outside of
 145                          * single-quotes only if they need escaping,
 146                          * and only if we resume the single-quoted part
 147                          * afterward.
 148                          */
 149                         if (need_bs_quote(src[1]) && src[2] == '\'') {
 150                                 *dst++ = src[1];
 151                                 src += 2;
 152                                 continue;
 153                         }
 154                 /* Fallthrough */
 155                 default:
 156                         if (!next)
 157                                 return NULL;
 158                         *dst = 0;
 159                         *next = src;
 160                         return arg;
 161                 }
 162         }
 163 }
 164
 165 char *sq_dequote(char *arg)
 166 {
 167         return sq_dequote_step(arg, NULL);
 168 }
 169
 170 static int sq_dequote_to_argv_internal(char *arg,
 171                                        const char ***argv, int *nr, int *alloc,
 172                                        struct strvec *array)
 173 {
 174         char *next = arg;
 175
 176         if (!*arg)
 177                 return 0;
 178         do {
 179                 char *dequoted = sq_dequote_step(next, &next);
 180                 if (!dequoted)
 181                         return -1;
 182                 if (next) {
 183                         char c;
 184                         if (!isspace(*next))
 185                                 return -1;
 186                         do {
 187                                 c = *++next;
 188                         } while (isspace(c));
 189                 }
 190                 if (argv) {
 191                         ALLOC_GROW(*argv, *nr + 1, *alloc);
 192                         (*argv)[(*nr)++] = dequoted;
 193                 }
 194                 if (array)
 195                         strvec_push(array, dequoted);
 196         } while (next);
 197
 198         return 0;
 199 }
 200
 201 int sq_dequote_to_argv(char *arg, const char ***argv, int *nr, int *alloc)
 202 {
 203         return sq_dequote_to_argv_internal(arg, argv, nr, alloc, NULL);
 204 }
 205
 206 int sq_dequote_to_strvec(char *arg, struct strvec *array)
 207 {
 208         return sq_dequote_to_argv_internal(arg, NULL, NULL, NULL, array);
 209 }
 210
 211 /* 1 means: quote as octal
 212  * 0 means: quote as octal if (quote_path_fully)
 213  * -1 means: never quote
 214  * c: quote as "\\c"
 215  */
 216 #define X8(x)   x, x, x, x, x, x, x, x
 217 #define X16(x)  X8(x), X8(x)
 218 static signed char const cq_lookup[256] = {
 219         /*           0    1    2    3    4    5    6    7 */
 220         /* 0x00 */   1,   1,   1,   1,   1,   1,   1, 'a',
 221         /* 0x08 */ 'b', 't', 'n', 'v', 'f', 'r',   1,   1,
 222         /* 0x10 */ X16(1),
 223         /* 0x20 */  -1,  -1, '"',  -1,  -1,  -1,  -1,  -1,
 224         /* 0x28 */ X16(-1), X16(-1), X16(-1),
 225         /* 0x58 */  -1,  -1,  -1,  -1,'\\',  -1,  -1,  -1,
 226         /* 0x60 */ X16(-1), X8(-1),
 227         /* 0x78 */  -1,  -1,  -1,  -1,  -1,  -1,  -1,   1,
 228         /* 0x80 */ /* set to 0 */
 229 };
 230
 231 static inline int cq_must_quote(char c)
 232 {
 233         return cq_lookup[(unsigned char)c] + quote_path_fully > 0;
 234 }
 235
 236 /* returns the longest prefix not needing a quote up to maxlen if positive.
 237    This stops at the first \0 because it's marked as a character needing an
 238    escape */
 239 static size_t next_quote_pos(const char *s, ssize_t maxlen)
 240 {
 241         size_t len;
 242         if (maxlen < 0) {
 243                 for (len = 0; !cq_must_quote(s[len]); len++);
 244         } else {
 245                 for (len = 0; len < maxlen && !cq_must_quote(s[len]); len++);
 246         }
 247         return len;
 248 }
 249
 250 /*
 251  * C-style name quoting.
 252  *
 253  * (1) if sb and fp are both NULL, inspect the input name and counts the
 254  *     number of bytes that are needed to hold c_style quoted version of name,
 255  *     counting the double quotes around it but not terminating NUL, and
 256  *     returns it.
 257  *     However, if name does not need c_style quoting, it returns 0.
 258  *
 259  * (2) if sb or fp are not NULL, it emits the c_style quoted version
 260  *     of name, enclosed with double quotes if asked and needed only.
 261  *     Return value is the same as in (1).
 262  */
 263 static size_t quote_c_style_counted(const char *name, ssize_t maxlen,
 264                                     struct strbuf *sb, FILE *fp, unsigned flags)
 265 {
 266 #undef EMIT
 267 #define EMIT(c)                                 \
 268         do {                                        \
 269                 if (sb) strbuf_addch(sb, (c));          \
 270                 if (fp) fputc((c), fp);                 \
 271                 count++;                                \
 272         } while (0)
 273 #define EMITBUF(s, l)                           \
 274         do {                                        \
 275                 if (sb) strbuf_add(sb, (s), (l));       \
 276                 if (fp) fwrite((s), (l), 1, fp);        \
 277                 count += (l);                           \
 278         } while (0)
 279
 280         int no_dq = !!(flags & CQUOTE_NODQ);
 281         size_t len, count = 0;
 282         const char *p = name;
 283
 284         for (;;) {
 285                 int ch;
 286
 287                 len = next_quote_pos(p, maxlen);
 288                 if (len == maxlen || (maxlen < 0 && !p[len]))
 289                         break;
 290
 291                 if (!no_dq && p == name)
 292                         EMIT('"');
 293
 294                 EMITBUF(p, len);
 295                 EMIT('\\');
 296                 p += len;
 297                 ch = (unsigned char)*p++;
 298                 if (maxlen >= 0)
 299                         maxlen -= len + 1;
 300                 if (cq_lookup[ch] >= ' ') {
 301                         EMIT(cq_lookup[ch]);
 302                 } else {
 303                         EMIT(((ch >> 6) & 03) + '0');
 304                         EMIT(((ch >> 3) & 07) + '0');
 305                         EMIT(((ch >> 0) & 07) + '0');
 306                 }
 307         }
 308
 309         EMITBUF(p, len);
 310         if (p == name)   /* no ending quote needed */
 311                 return 0;
 312
 313         if (!no_dq)
 314                 EMIT('"');
 315         return count;
 316 }
 317
 318 size_t quote_c_style(const char *name, struct strbuf *sb, FILE *fp, unsigned flags)
 319 {
 320         return quote_c_style_counted(name, -1, sb, fp, flags);
 321 }
 322
 323 void quote_two_c_style(struct strbuf *sb, const char *prefix, const char *path,
 324                        unsigned flags)
 325 {
 326         int nodq = !!(flags & CQUOTE_NODQ);
 327         if (quote_c_style(prefix, NULL, NULL, 0) ||
 328             quote_c_style(path, NULL, NULL, 0)) {
 329                 if (!nodq)
 330                         strbuf_addch(sb, '"');
 331                 quote_c_style(prefix, sb, NULL, CQUOTE_NODQ);
 332                 quote_c_style(path, sb, NULL, CQUOTE_NODQ);
 333                 if (!nodq)
 334                         strbuf_addch(sb, '"');
 335         } else {
 336                 strbuf_addstr(sb, prefix);
 337                 strbuf_addstr(sb, path);
 338         }
 339 }
 340
 341 void write_name_quoted(const char *name, FILE *fp, int terminator)
 342 {
 343         if (terminator) {
 344                 quote_c_style(name, NULL, fp, 0);
 345         } else {
 346                 fputs(name, fp);
 347         }
 348         fputc(terminator, fp);
 349 }
 350
 351 void write_name_quoted_relative(const char *name, const char *prefix,
 352                                 FILE *fp, int terminator)
 353 {
 354         struct strbuf sb = STRBUF_INIT;
 355
 356         name = relative_path(name, prefix, &sb);
 357         write_name_quoted(name, fp, terminator);
 358
 359         strbuf_release(&sb);
 360 }
 361
 362 /* quote path as relative to the given prefix */
 363 char *quote_path(const char *in, const char *prefix, struct strbuf *out, unsigned flags)
 364 {
 365         struct strbuf sb = STRBUF_INIT;
 366         const char *rel = relative_path(in, prefix, &sb);
 367         int force_dq = ((flags & QUOTE_PATH_QUOTE_SP) && strchr(rel, ' '));
 368
 369         strbuf_reset(out);
 370
 371         /*
 372          * If the caller wants us to enclose the output in a dq-pair
 373          * whether quote_c_style_counted() needs to, we do it ourselves
 374          * and tell quote_c_style_counted() not to.
 375          */
 376         if (force_dq)
 377                 strbuf_addch(out, '"');
 378         quote_c_style_counted(rel, strlen(rel), out, NULL,
 379                               force_dq ? CQUOTE_NODQ : 0);
 380         if (force_dq)
 381                 strbuf_addch(out, '"');
 382         strbuf_release(&sb);
 383
 384         return out->buf;
 385 }
 386
 387 /*
 388  * C-style name unquoting.
 389  *
 390  * Quoted should point at the opening double quote.
 391  * + Returns 0 if it was able to unquote the string properly, and appends the
 392  *   result in the strbuf `sb'.
 393  * + Returns -1 in case of error, and doesn't touch the strbuf. Though note
 394  *   that this function will allocate memory in the strbuf, so calling
 395  *   strbuf_release is mandatory whichever result unquote_c_style returns.
 396  *
 397  * Updates endp pointer to point at one past the ending double quote if given.
 398  */
 399 int unquote_c_style(struct strbuf *sb, const char *quoted, const char **endp)
 400 {
 401         size_t oldlen = sb->len, len;
 402         int ch, ac;
 403
 404         if (*quoted++ != '"')
 405                 return -1;
 406
 407         for (;;) {
 408                 len = strcspn(quoted, "\"\\");
 409                 strbuf_add(sb, quoted, len);
 410                 quoted += len;
 411
 412                 switch (*quoted++) {
 413                   case '"':
 414                         if (endp)
 415                                 *endp = quoted;
 416                         return 0;
 417                   case '\\':
 418                         break;
 419                   default:
 420                         goto error;
 421                 }
 422
 423                 switch ((ch = *quoted++)) {
 424                 case 'a': ch = '\a'; break;
 425                 case 'b': ch = '\b'; break;
 426                 case 'f': ch = '\f'; break;
 427                 case 'n': ch = '\n'; break;
 428                 case 'r': ch = '\r'; break;
 429                 case 't': ch = '\t'; break;
 430                 case 'v': ch = '\v'; break;
 431
 432                 case '\\': case '"':
 433                         break; /* verbatim */
 434
 435                 /* octal values with first digit over 4 overflow */
 436                 case '0': case '1': case '2': case '3':
 437                                         ac = ((ch - '0') << 6);
 438                         if ((ch = *quoted++) < '0' || '7' < ch)
 439                                 goto error;
 440                                         ac |= ((ch - '0') << 3);
 441                         if ((ch = *quoted++) < '0' || '7' < ch)
 442                                 goto error;
 443                                         ac |= (ch - '0');
 444                                         ch = ac;
 445                                         break;
 446                                 default:
 447                         goto error;
 448                         }
 449                 strbuf_addch(sb, ch);
 450                 }
 451
 452   error:
 453         strbuf_setlen(sb, oldlen);
 454         return -1;
 455 }
 456
 457 /* quoting as a string literal for other languages */
 458
 459 void perl_quote_buf(struct strbuf *sb, const char *src)
 460 {
 461         const char sq = '\'';
 462         const char bq = '\\';
 463         char c;
 464
 465         strbuf_addch(sb, sq);
 466         while ((c = *src++)) {
 467                 if (c == sq || c == bq)
 468                         strbuf_addch(sb, bq);
 469                 strbuf_addch(sb, c);
 470         }
 471         strbuf_addch(sb, sq);
 472 }
 473
 474 void perl_quote_buf_with_len(struct strbuf *sb, const char *src, size_t len)
 475 {
 476         const char sq = '\'';
 477         const char bq = '\\';
 478         const char *c = src;
 479         const char *end = src + len;
 480
 481         strbuf_addch(sb, sq);
 482         while (c != end) {
 483                 if (*c == sq || *c == bq)
 484                         strbuf_addch(sb, bq);
 485                 strbuf_addch(sb, *c);
 486                 c++;
 487         }
 488         strbuf_addch(sb, sq);
 489 }
 490
 491 void python_quote_buf(struct strbuf *sb, const char *src)
 492 {
 493         const char sq = '\'';
 494         const char bq = '\\';
 495         const char nl = '\n';
 496         char c;
 497
 498         strbuf_addch(sb, sq);
 499         while ((c = *src++)) {
 500                 if (c == nl) {
 501                         strbuf_addch(sb, bq);
 502                         strbuf_addch(sb, 'n');
 503                         continue;
 504                 }
 505                 if (c == sq || c == bq)
 506                         strbuf_addch(sb, bq);
 507                 strbuf_addch(sb, c);
 508         }
 509         strbuf_addch(sb, sq);
 510 }
 511
 512 void tcl_quote_buf(struct strbuf *sb, const char *src)
 513 {
 514         char c;
 515
 516         strbuf_addch(sb, '"');
 517         while ((c = *src++)) {
 518                 switch (c) {
 519                 case '[': case ']':
 520                 case '{': case '}':
 521                 case '$': case '\\': case '"':
 522                         strbuf_addch(sb, '\\');
 523                         /* fallthrough */
 524                 default:
 525                         strbuf_addch(sb, c);
 526                         break;
 527                 case '\f':
 528                         strbuf_addstr(sb, "\\f");
 529                         break;
 530                 case '\r':
 531                         strbuf_addstr(sb, "\\r");
 532                         break;
 533                 case '\n':
 534                         strbuf_addstr(sb, "\\n");
 535                         break;
 536                 case '\t':
 537                         strbuf_addstr(sb, "\\t");
 538                         break;
 539                 case '\v':
 540                         strbuf_addstr(sb, "\\v");
 541                         break;
 542                 }
 543         }
 544         strbuf_addch(sb, '"');
 545 }
 546
 547 void basic_regex_quote_buf(struct strbuf *sb, const char *src)
 548 {
 549         char c;
 550
 551         if (*src == '^') {
 552                 /* only beginning '^' is special and needs quoting */
 553                 strbuf_addch(sb, '\\');
 554                 strbuf_addch(sb, *src++);
 555         }
 556         if (*src == '*')
 557                 /* beginning '*' is not special, no quoting */
 558                 strbuf_addch(sb, *src++);
 559
 560         while ((c = *src++)) {
 561                 switch (c) {
 562                 case '[':
 563                 case '.':
 564                 case '\\':
 565                 case '*':
 566                         strbuf_addch(sb, '\\');
 567                         strbuf_addch(sb, c);
 568                         break;
 569
 570                 case '$':
 571                         /* only the end '$' is special and needs quoting */
 572                         if (*src == '\0')
 573                                 strbuf_addch(sb, '\\');
 574                         strbuf_addch(sb, c);
 575                         break;
 576
 577                 default:
 578                         strbuf_addch(sb, c);
 579                         break;
 580                 }
 581         }
 582 }