quote.c

   1 #include "cache.h"
   2 #include "alloc.h"
   3 #include "quote.h"
   4 #include "strvec.h"
   5
   6 int quote_path_fully = 1;
   7
   8 static inline int need_bs_quote(char c)
   9 {
  10         return (c == '\'' || c == '!');
  11 }
  12
  13 /* Help to copy the thing properly quoted for the shell safety.
  14  * any single quote is replaced with '\'', any exclamation point
  15  * is replaced with '\!', and the whole thing is enclosed in a
  16  * single quote pair.
  17  *
  18  * E.g.
  19  *  original     sq_quote     result
  20  *  name     ==> name      ==> 'name'
  21  *  a b      ==> a b       ==> 'a b'
  22  *  a'b      ==> a'\''b    ==> 'a'\''b'
  23  *  a!b      ==> a'\!'b    ==> 'a'\!'b'
  24  */
  25 void sq_quote_buf(struct strbuf *dst, const char *src)
  26 {
  27         char *to_free = NULL;
  28
  29         if (dst->buf == src)
  30                 to_free = strbuf_detach(dst, NULL);
  31
  32         strbuf_addch(dst, '\'');
  33         while (*src) {
  34                 size_t len = strcspn(src, "'!");
  35                 strbuf_add(dst, src, len);
  36                 src += len;
  37                 while (need_bs_quote(*src)) {
  38                         strbuf_addstr(dst, "'\\");
  39                         strbuf_addch(dst, *src++);
  40                         strbuf_addch(dst, '\'');
  41                 }
  42         }
  43         strbuf_addch(dst, '\'');
  44         free(to_free);
  45 }
  46
  47 void sq_quote_buf_pretty(struct strbuf *dst, const char *src)
  48 {
  49         static const char ok_punct[] = "+,-./:=@_^";
  50         const char *p;
  51
  52         /* Avoid losing a zero-length string by adding '' */
  53         if (!*src) {
  54                 strbuf_addstr(dst, "''");
  55                 return;
  56         }
  57
  58         for (p = src; *p; p++) {
  59                 if (!isalnum(*p) && !strchr(ok_punct, *p)) {
  60                         sq_quote_buf(dst, src);
  61                         return;
  62                 }
  63         }
  64
  65         /* if we get here, we did not need quoting */
  66         strbuf_addstr(dst, src);
  67 }
  68
  69 void sq_quotef(struct strbuf *dst, const char *fmt, ...)
  70 {
  71         struct strbuf src = STRBUF_INIT;
  72
  73         va_list ap;
  74         va_start(ap, fmt);
  75         strbuf_vaddf(&src, fmt, ap);
  76         va_end(ap);
  77
  78         sq_quote_buf(dst, src.buf);
  79         strbuf_release(&src);
  80 }
  81
  82 void sq_quote_argv(struct strbuf *dst, const char **argv)
  83 {
  84         int i;
  85
  86         /* Copy into destination buffer. */
  87         strbuf_grow(dst, 255);
  88         for (i = 0; argv[i]; ++i) {
  89                 strbuf_addch(dst, ' ');
  90                 sq_quote_buf(dst, argv[i]);
  91         }
  92 }
  93
  94 /*
  95  * Legacy function to append each argv value, quoted as necessasry,
  96  * with whitespace before each value.  This results in a leading
  97  * space in the result.
  98  */
  99 void sq_quote_argv_pretty(struct strbuf *dst, const char **argv)
 100 {
 101         if (argv[0])
 102                 strbuf_addch(dst, ' ');
 103         sq_append_quote_argv_pretty(dst, argv);
 104 }
 105
 106 /*
 107  * Append each argv value, quoted as necessary, with whitespace between them.
 108  */
 109 void sq_append_quote_argv_pretty(struct strbuf *dst, const char **argv)
 110 {
 111         int i;
 112
 113         for (i = 0; argv[i]; i++) {
 114                 if (i > 0)
 115                         strbuf_addch(dst, ' ');
 116                 sq_quote_buf_pretty(dst, argv[i]);
 117         }
 118 }
 119
 120 char *sq_dequote_step(char *arg, char **next)
 121 {
 122         char *dst = arg;
 123         char *src = arg;
 124         char c;
 125
 126         if (*src != '\'')
 127                 return NULL;
 128         for (;;) {
 129                 c = *++src;
 130                 if (!c)
 131                         return NULL;
 132                 if (c != '\'') {
 133                         *dst++ = c;
 134                         continue;
 135                 }
 136                 /* We stepped out of sq */
 137                 switch (*++src) {
 138                 case '\0':
 139                         *dst = 0;
 140                         if (next)
 141                                 *next = NULL;
 142                         return arg;
 143                 case '\\':
 144                         /*
 145                          * Allow backslashed characters outside of
 146                          * single-quotes only if they need escaping,
 147                          * and only if we resume the single-quoted part
 148                          * afterward.
 149                          */
 150                         if (need_bs_quote(src[1]) && src[2] == '\'') {
 151                                 *dst++ = src[1];
 152                                 src += 2;
 153                                 continue;
 154                         }
 155                 /* Fallthrough */
 156                 default:
 157                         if (!next)
 158                                 return NULL;
 159                         *dst = 0;
 160                         *next = src;
 161                         return arg;
 162                 }
 163         }
 164 }
 165
 166 char *sq_dequote(char *arg)
 167 {
 168         return sq_dequote_step(arg, NULL);
 169 }
 170
 171 static int sq_dequote_to_argv_internal(char *arg,
 172                                        const char ***argv, int *nr, int *alloc,
 173                                        struct strvec *array)
 174 {
 175         char *next = arg;
 176
 177         if (!*arg)
 178                 return 0;
 179         do {
 180                 char *dequoted = sq_dequote_step(next, &next);
 181                 if (!dequoted)
 182                         return -1;
 183                 if (next) {
 184                         char c;
 185                         if (!isspace(*next))
 186                                 return -1;
 187                         do {
 188                                 c = *++next;
 189                         } while (isspace(c));
 190                 }
 191                 if (argv) {
 192                         ALLOC_GROW(*argv, *nr + 1, *alloc);
 193                         (*argv)[(*nr)++] = dequoted;
 194                 }
 195                 if (array)
 196                         strvec_push(array, dequoted);
 197         } while (next);
 198
 199         return 0;
 200 }
 201
 202 int sq_dequote_to_argv(char *arg, const char ***argv, int *nr, int *alloc)
 203 {
 204         return sq_dequote_to_argv_internal(arg, argv, nr, alloc, NULL);
 205 }
 206
 207 int sq_dequote_to_strvec(char *arg, struct strvec *array)
 208 {
 209         return sq_dequote_to_argv_internal(arg, NULL, NULL, NULL, array);
 210 }
 211
 212 /* 1 means: quote as octal
 213  * 0 means: quote as octal if (quote_path_fully)
 214  * -1 means: never quote
 215  * c: quote as "\\c"
 216  */
 217 #define X8(x)   x, x, x, x, x, x, x, x
 218 #define X16(x)  X8(x), X8(x)
 219 static signed char const cq_lookup[256] = {
 220         /*           0    1    2    3    4    5    6    7 */
 221         /* 0x00 */   1,   1,   1,   1,   1,   1,   1, 'a',
 222         /* 0x08 */ 'b', 't', 'n', 'v', 'f', 'r',   1,   1,
 223         /* 0x10 */ X16(1),
 224         /* 0x20 */  -1,  -1, '"',  -1,  -1,  -1,  -1,  -1,
 225         /* 0x28 */ X16(-1), X16(-1), X16(-1),
 226         /* 0x58 */  -1,  -1,  -1,  -1,'\\',  -1,  -1,  -1,
 227         /* 0x60 */ X16(-1), X8(-1),
 228         /* 0x78 */  -1,  -1,  -1,  -1,  -1,  -1,  -1,   1,
 229         /* 0x80 */ /* set to 0 */
 230 };
 231
 232 static inline int cq_must_quote(char c)
 233 {
 234         return cq_lookup[(unsigned char)c] + quote_path_fully > 0;
 235 }
 236
 237 /* returns the longest prefix not needing a quote up to maxlen if positive.
 238    This stops at the first \0 because it's marked as a character needing an
 239    escape */
 240 static size_t next_quote_pos(const char *s, ssize_t maxlen)
 241 {
 242         size_t len;
 243         if (maxlen < 0) {
 244                 for (len = 0; !cq_must_quote(s[len]); len++);
 245         } else {
 246                 for (len = 0; len < maxlen && !cq_must_quote(s[len]); len++);
 247         }
 248         return len;
 249 }
 250
 251 /*
 252  * C-style name quoting.
 253  *
 254  * (1) if sb and fp are both NULL, inspect the input name and counts the
 255  *     number of bytes that are needed to hold c_style quoted version of name,
 256  *     counting the double quotes around it but not terminating NUL, and
 257  *     returns it.
 258  *     However, if name does not need c_style quoting, it returns 0.
 259  *
 260  * (2) if sb or fp are not NULL, it emits the c_style quoted version
 261  *     of name, enclosed with double quotes if asked and needed only.
 262  *     Return value is the same as in (1).
 263  */
 264 static size_t quote_c_style_counted(const char *name, ssize_t maxlen,
 265                                     struct strbuf *sb, FILE *fp, unsigned flags)
 266 {
 267 #undef EMIT
 268 #define EMIT(c)                                 \
 269         do {                                        \
 270                 if (sb) strbuf_addch(sb, (c));          \
 271                 if (fp) fputc((c), fp);                 \
 272                 count++;                                \
 273         } while (0)
 274 #define EMITBUF(s, l)                           \
 275         do {                                        \
 276                 if (sb) strbuf_add(sb, (s), (l));       \
 277                 if (fp) fwrite((s), (l), 1, fp);        \
 278                 count += (l);                           \
 279         } while (0)
 280
 281         int no_dq = !!(flags & CQUOTE_NODQ);
 282         size_t len, count = 0;
 283         const char *p = name;
 284
 285         for (;;) {
 286                 int ch;
 287
 288                 len = next_quote_pos(p, maxlen);
 289                 if (len == maxlen || (maxlen < 0 && !p[len]))
 290                         break;
 291
 292                 if (!no_dq && p == name)
 293                         EMIT('"');
 294
 295                 EMITBUF(p, len);
 296                 EMIT('\\');
 297                 p += len;
 298                 ch = (unsigned char)*p++;
 299                 if (maxlen >= 0)
 300                         maxlen -= len + 1;
 301                 if (cq_lookup[ch] >= ' ') {
 302                         EMIT(cq_lookup[ch]);
 303                 } else {
 304                         EMIT(((ch >> 6) & 03) + '0');
 305                         EMIT(((ch >> 3) & 07) + '0');
 306                         EMIT(((ch >> 0) & 07) + '0');
 307                 }
 308         }
 309
 310         EMITBUF(p, len);
 311         if (p == name)   /* no ending quote needed */
 312                 return 0;
 313
 314         if (!no_dq)
 315                 EMIT('"');
 316         return count;
 317 }
 318
 319 size_t quote_c_style(const char *name, struct strbuf *sb, FILE *fp, unsigned flags)
 320 {
 321         return quote_c_style_counted(name, -1, sb, fp, flags);
 322 }
 323
 324 void quote_two_c_style(struct strbuf *sb, const char *prefix, const char *path,
 325                        unsigned flags)
 326 {
 327         int nodq = !!(flags & CQUOTE_NODQ);
 328         if (quote_c_style(prefix, NULL, NULL, 0) ||
 329             quote_c_style(path, NULL, NULL, 0)) {
 330                 if (!nodq)
 331                         strbuf_addch(sb, '"');
 332                 quote_c_style(prefix, sb, NULL, CQUOTE_NODQ);
 333                 quote_c_style(path, sb, NULL, CQUOTE_NODQ);
 334                 if (!nodq)
 335                         strbuf_addch(sb, '"');
 336         } else {
 337                 strbuf_addstr(sb, prefix);
 338                 strbuf_addstr(sb, path);
 339         }
 340 }
 341
 342 void write_name_quoted(const char *name, FILE *fp, int terminator)
 343 {
 344         if (terminator) {
 345                 quote_c_style(name, NULL, fp, 0);
 346         } else {
 347                 fputs(name, fp);
 348         }
 349         fputc(terminator, fp);
 350 }
 351
 352 void write_name_quoted_relative(const char *name, const char *prefix,
 353                                 FILE *fp, int terminator)
 354 {
 355         struct strbuf sb = STRBUF_INIT;
 356
 357         name = relative_path(name, prefix, &sb);
 358         write_name_quoted(name, fp, terminator);
 359
 360         strbuf_release(&sb);
 361 }
 362
 363 /* quote path as relative to the given prefix */
 364 char *quote_path(const char *in, const char *prefix, struct strbuf *out, unsigned flags)
 365 {
 366         struct strbuf sb = STRBUF_INIT;
 367         const char *rel = relative_path(in, prefix, &sb);
 368         int force_dq = ((flags & QUOTE_PATH_QUOTE_SP) && strchr(rel, ' '));
 369
 370         strbuf_reset(out);
 371
 372         /*
 373          * If the caller wants us to enclose the output in a dq-pair
 374          * whether quote_c_style_counted() needs to, we do it ourselves
 375          * and tell quote_c_style_counted() not to.
 376          */
 377         if (force_dq)
 378                 strbuf_addch(out, '"');
 379         quote_c_style_counted(rel, strlen(rel), out, NULL,
 380                               force_dq ? CQUOTE_NODQ : 0);
 381         if (force_dq)
 382                 strbuf_addch(out, '"');
 383         strbuf_release(&sb);
 384
 385         return out->buf;
 386 }
 387
 388 /*
 389  * C-style name unquoting.
 390  *
 391  * Quoted should point at the opening double quote.
 392  * + Returns 0 if it was able to unquote the string properly, and appends the
 393  *   result in the strbuf `sb'.
 394  * + Returns -1 in case of error, and doesn't touch the strbuf. Though note
 395  *   that this function will allocate memory in the strbuf, so calling
 396  *   strbuf_release is mandatory whichever result unquote_c_style returns.
 397  *
 398  * Updates endp pointer to point at one past the ending double quote if given.
 399  */
 400 int unquote_c_style(struct strbuf *sb, const char *quoted, const char **endp)
 401 {
 402         size_t oldlen = sb->len, len;
 403         int ch, ac;
 404
 405         if (*quoted++ != '"')
 406                 return -1;
 407
 408         for (;;) {
 409                 len = strcspn(quoted, "\"\\");
 410                 strbuf_add(sb, quoted, len);
 411                 quoted += len;
 412
 413                 switch (*quoted++) {
 414                   case '"':
 415                         if (endp)
 416                                 *endp = quoted;
 417                         return 0;
 418                   case '\\':
 419                         break;
 420                   default:
 421                         goto error;
 422                 }
 423
 424                 switch ((ch = *quoted++)) {
 425                 case 'a': ch = '\a'; break;
 426                 case 'b': ch = '\b'; break;
 427                 case 'f': ch = '\f'; break;
 428                 case 'n': ch = '\n'; break;
 429                 case 'r': ch = '\r'; break;
 430                 case 't': ch = '\t'; break;
 431                 case 'v': ch = '\v'; break;
 432
 433                 case '\\': case '"':
 434                         break; /* verbatim */
 435
 436                 /* octal values with first digit over 4 overflow */
 437                 case '0': case '1': case '2': case '3':
 438                                         ac = ((ch - '0') << 6);
 439                         if ((ch = *quoted++) < '0' || '7' < ch)
 440                                 goto error;
 441                                         ac |= ((ch - '0') << 3);
 442                         if ((ch = *quoted++) < '0' || '7' < ch)
 443                                 goto error;
 444                                         ac |= (ch - '0');
 445                                         ch = ac;
 446                                         break;
 447                                 default:
 448                         goto error;
 449                         }
 450                 strbuf_addch(sb, ch);
 451                 }
 452
 453   error:
 454         strbuf_setlen(sb, oldlen);
 455         return -1;
 456 }
 457
 458 /* quoting as a string literal for other languages */
 459
 460 void perl_quote_buf(struct strbuf *sb, const char *src)
 461 {
 462         const char sq = '\'';
 463         const char bq = '\\';
 464         char c;
 465
 466         strbuf_addch(sb, sq);
 467         while ((c = *src++)) {
 468                 if (c == sq || c == bq)
 469                         strbuf_addch(sb, bq);
 470                 strbuf_addch(sb, c);
 471         }
 472         strbuf_addch(sb, sq);
 473 }
 474
 475 void perl_quote_buf_with_len(struct strbuf *sb, const char *src, size_t len)
 476 {
 477         const char sq = '\'';
 478         const char bq = '\\';
 479         const char *c = src;
 480         const char *end = src + len;
 481
 482         strbuf_addch(sb, sq);
 483         while (c != end) {
 484                 if (*c == sq || *c == bq)
 485                         strbuf_addch(sb, bq);
 486                 strbuf_addch(sb, *c);
 487                 c++;
 488         }
 489         strbuf_addch(sb, sq);
 490 }
 491
 492 void python_quote_buf(struct strbuf *sb, const char *src)
 493 {
 494         const char sq = '\'';
 495         const char bq = '\\';
 496         const char nl = '\n';
 497         char c;
 498
 499         strbuf_addch(sb, sq);
 500         while ((c = *src++)) {
 501                 if (c == nl) {
 502                         strbuf_addch(sb, bq);
 503                         strbuf_addch(sb, 'n');
 504                         continue;
 505                 }
 506                 if (c == sq || c == bq)
 507                         strbuf_addch(sb, bq);
 508                 strbuf_addch(sb, c);
 509         }
 510         strbuf_addch(sb, sq);
 511 }
 512
 513 void tcl_quote_buf(struct strbuf *sb, const char *src)
 514 {
 515         char c;
 516
 517         strbuf_addch(sb, '"');
 518         while ((c = *src++)) {
 519                 switch (c) {
 520                 case '[': case ']':
 521                 case '{': case '}':
 522                 case '$': case '\\': case '"':
 523                         strbuf_addch(sb, '\\');
 524                         /* fallthrough */
 525                 default:
 526                         strbuf_addch(sb, c);
 527                         break;
 528                 case '\f':
 529                         strbuf_addstr(sb, "\\f");
 530                         break;
 531                 case '\r':
 532                         strbuf_addstr(sb, "\\r");
 533                         break;
 534                 case '\n':
 535                         strbuf_addstr(sb, "\\n");
 536                         break;
 537                 case '\t':
 538                         strbuf_addstr(sb, "\\t");
 539                         break;
 540                 case '\v':
 541                         strbuf_addstr(sb, "\\v");
 542                         break;
 543                 }
 544         }
 545         strbuf_addch(sb, '"');
 546 }
 547
 548 void basic_regex_quote_buf(struct strbuf *sb, const char *src)
 549 {
 550         char c;
 551
 552         if (*src == '^') {
 553                 /* only beginning '^' is special and needs quoting */
 554                 strbuf_addch(sb, '\\');
 555                 strbuf_addch(sb, *src++);
 556         }
 557         if (*src == '*')
 558                 /* beginning '*' is not special, no quoting */
 559                 strbuf_addch(sb, *src++);
 560
 561         while ((c = *src++)) {
 562                 switch (c) {
 563                 case '[':
 564                 case '.':
 565                 case '\\':
 566                 case '*':
 567                         strbuf_addch(sb, '\\');
 568                         strbuf_addch(sb, c);
 569                         break;
 570
 571                 case '$':
 572                         /* only the end '$' is special and needs quoting */
 573                         if (*src == '\0')
 574                                 strbuf_addch(sb, '\\');
 575                         strbuf_addch(sb, c);
 576                         break;
 577
 578                 default:
 579                         strbuf_addch(sb, c);
 580                         break;
 581                 }
 582         }
 583 }