quote.c

   1 #include "cache.h"
   2 #include "quote.h"
   3 #include "argv-array.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 (!isalpha(*p) && !isdigit(*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 static 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 || !isspace(*src))
 157                                 return NULL;
 158                         do {
 159                                 c = *++src;
 160                         } while (isspace(c));
 161                         *dst = 0;
 162                         *next = src;
 163                         return arg;
 164                 }
 165         }
 166 }
 167
 168 char *sq_dequote(char *arg)
 169 {
 170         return sq_dequote_step(arg, NULL);
 171 }
 172
 173 static int sq_dequote_to_argv_internal(char *arg,
 174                                        const char ***argv, int *nr, int *alloc,
 175                                        struct argv_array *array)
 176 {
 177         char *next = arg;
 178
 179         if (!*arg)
 180                 return 0;
 181         do {
 182                 char *dequoted = sq_dequote_step(next, &next);
 183                 if (!dequoted)
 184                         return -1;
 185                 if (argv) {
 186                         ALLOC_GROW(*argv, *nr + 1, *alloc);
 187                         (*argv)[(*nr)++] = dequoted;
 188                 }
 189                 if (array)
 190                         argv_array_push(array, dequoted);
 191         } while (next);
 192
 193         return 0;
 194 }
 195
 196 int sq_dequote_to_argv(char *arg, const char ***argv, int *nr, int *alloc)
 197 {
 198         return sq_dequote_to_argv_internal(arg, argv, nr, alloc, NULL);
 199 }
 200
 201 int sq_dequote_to_argv_array(char *arg, struct argv_array *array)
 202 {
 203         return sq_dequote_to_argv_internal(arg, NULL, NULL, NULL, array);
 204 }
 205
 206 /* 1 means: quote as octal
 207  * 0 means: quote as octal if (quote_path_fully)
 208  * -1 means: never quote
 209  * c: quote as "\\c"
 210  */
 211 #define X8(x)   x, x, x, x, x, x, x, x
 212 #define X16(x)  X8(x), X8(x)
 213 static signed char const sq_lookup[256] = {
 214         /*           0    1    2    3    4    5    6    7 */
 215         /* 0x00 */   1,   1,   1,   1,   1,   1,   1, 'a',
 216         /* 0x08 */ 'b', 't', 'n', 'v', 'f', 'r',   1,   1,
 217         /* 0x10 */ X16(1),
 218         /* 0x20 */  -1,  -1, '"',  -1,  -1,  -1,  -1,  -1,
 219         /* 0x28 */ X16(-1), X16(-1), X16(-1),
 220         /* 0x58 */  -1,  -1,  -1,  -1,'\\',  -1,  -1,  -1,
 221         /* 0x60 */ X16(-1), X8(-1),
 222         /* 0x78 */  -1,  -1,  -1,  -1,  -1,  -1,  -1,   1,
 223         /* 0x80 */ /* set to 0 */
 224 };
 225
 226 static inline int sq_must_quote(char c)
 227 {
 228         return sq_lookup[(unsigned char)c] + quote_path_fully > 0;
 229 }
 230
 231 /* returns the longest prefix not needing a quote up to maxlen if positive.
 232    This stops at the first \0 because it's marked as a character needing an
 233    escape */
 234 static size_t next_quote_pos(const char *s, ssize_t maxlen)
 235 {
 236         size_t len;
 237         if (maxlen < 0) {
 238                 for (len = 0; !sq_must_quote(s[len]); len++);
 239         } else {
 240                 for (len = 0; len < maxlen && !sq_must_quote(s[len]); len++);
 241         }
 242         return len;
 243 }
 244
 245 /*
 246  * C-style name quoting.
 247  *
 248  * (1) if sb and fp are both NULL, inspect the input name and counts the
 249  *     number of bytes that are needed to hold c_style quoted version of name,
 250  *     counting the double quotes around it but not terminating NUL, and
 251  *     returns it.
 252  *     However, if name does not need c_style quoting, it returns 0.
 253  *
 254  * (2) if sb or fp are not NULL, it emits the c_style quoted version
 255  *     of name, enclosed with double quotes if asked and needed only.
 256  *     Return value is the same as in (1).
 257  */
 258 static size_t quote_c_style_counted(const char *name, ssize_t maxlen,
 259                                     struct strbuf *sb, FILE *fp, int no_dq)
 260 {
 261 #undef EMIT
 262 #define EMIT(c)                                 \
 263         do {                                        \
 264                 if (sb) strbuf_addch(sb, (c));          \
 265                 if (fp) fputc((c), fp);                 \
 266                 count++;                                \
 267         } while (0)
 268 #define EMITBUF(s, l)                           \
 269         do {                                        \
 270                 if (sb) strbuf_add(sb, (s), (l));       \
 271                 if (fp) fwrite((s), (l), 1, fp);        \
 272                 count += (l);                           \
 273         } while (0)
 274
 275         size_t len, count = 0;
 276         const char *p = name;
 277
 278         for (;;) {
 279                 int ch;
 280
 281                 len = next_quote_pos(p, maxlen);
 282                 if (len == maxlen || (maxlen < 0 && !p[len]))
 283                         break;
 284
 285                 if (!no_dq && p == name)
 286                         EMIT('"');
 287
 288                 EMITBUF(p, len);
 289                 EMIT('\\');
 290                 p += len;
 291                 ch = (unsigned char)*p++;
 292                 if (maxlen >= 0)
 293                         maxlen -= len + 1;
 294                 if (sq_lookup[ch] >= ' ') {
 295                         EMIT(sq_lookup[ch]);
 296                 } else {
 297                         EMIT(((ch >> 6) & 03) + '0');
 298                         EMIT(((ch >> 3) & 07) + '0');
 299                         EMIT(((ch >> 0) & 07) + '0');
 300                 }
 301         }
 302
 303         EMITBUF(p, len);
 304         if (p == name)   /* no ending quote needed */
 305                 return 0;
 306
 307         if (!no_dq)
 308                 EMIT('"');
 309         return count;
 310 }
 311
 312 size_t quote_c_style(const char *name, struct strbuf *sb, FILE *fp, int nodq)
 313 {
 314         return quote_c_style_counted(name, -1, sb, fp, nodq);
 315 }
 316
 317 void quote_two_c_style(struct strbuf *sb, const char *prefix, const char *path, int nodq)
 318 {
 319         if (quote_c_style(prefix, NULL, NULL, 0) ||
 320             quote_c_style(path, NULL, NULL, 0)) {
 321                 if (!nodq)
 322                         strbuf_addch(sb, '"');
 323                 quote_c_style(prefix, sb, NULL, 1);
 324                 quote_c_style(path, sb, NULL, 1);
 325                 if (!nodq)
 326                         strbuf_addch(sb, '"');
 327         } else {
 328                 strbuf_addstr(sb, prefix);
 329                 strbuf_addstr(sb, path);
 330         }
 331 }
 332
 333 void write_name_quoted(const char *name, FILE *fp, int terminator)
 334 {
 335         if (terminator) {
 336                 quote_c_style(name, NULL, fp, 0);
 337         } else {
 338                 fputs(name, fp);
 339         }
 340         fputc(terminator, fp);
 341 }
 342
 343 void write_name_quoted_relative(const char *name, const char *prefix,
 344                                 FILE *fp, int terminator)
 345 {
 346         struct strbuf sb = STRBUF_INIT;
 347
 348         name = relative_path(name, prefix, &sb);
 349         write_name_quoted(name, fp, terminator);
 350
 351         strbuf_release(&sb);
 352 }
 353
 354 /* quote path as relative to the given prefix */
 355 char *quote_path_relative(const char *in, const char *prefix,
 356                           struct strbuf *out)
 357 {
 358         struct strbuf sb = STRBUF_INIT;
 359         const char *rel = relative_path(in, prefix, &sb);
 360         strbuf_reset(out);
 361         quote_c_style_counted(rel, strlen(rel), out, NULL, 0);
 362         strbuf_release(&sb);
 363
 364         return out->buf;
 365 }
 366
 367 /*
 368  * C-style name unquoting.
 369  *
 370  * Quoted should point at the opening double quote.
 371  * + Returns 0 if it was able to unquote the string properly, and appends the
 372  *   result in the strbuf `sb'.
 373  * + Returns -1 in case of error, and doesn't touch the strbuf. Though note
 374  *   that this function will allocate memory in the strbuf, so calling
 375  *   strbuf_release is mandatory whichever result unquote_c_style returns.
 376  *
 377  * Updates endp pointer to point at one past the ending double quote if given.
 378  */
 379 int unquote_c_style(struct strbuf *sb, const char *quoted, const char **endp)
 380 {
 381         size_t oldlen = sb->len, len;
 382         int ch, ac;
 383
 384         if (*quoted++ != '"')
 385                 return -1;
 386
 387         for (;;) {
 388                 len = strcspn(quoted, "\"\\");
 389                 strbuf_add(sb, quoted, len);
 390                 quoted += len;
 391
 392                 switch (*quoted++) {
 393                   case '"':
 394                         if (endp)
 395                                 *endp = quoted;
 396                         return 0;
 397                   case '\\':
 398                         break;
 399                   default:
 400                         goto error;
 401                 }
 402
 403                 switch ((ch = *quoted++)) {
 404                 case 'a': ch = '\a'; break;
 405                 case 'b': ch = '\b'; break;
 406                 case 'f': ch = '\f'; break;
 407                 case 'n': ch = '\n'; break;
 408                 case 'r': ch = '\r'; break;
 409                 case 't': ch = '\t'; break;
 410                 case 'v': ch = '\v'; break;
 411
 412                 case '\\': case '"':
 413                         break; /* verbatim */
 414
 415                 /* octal values with first digit over 4 overflow */
 416                 case '0': case '1': case '2': case '3':
 417                                         ac = ((ch - '0') << 6);
 418                         if ((ch = *quoted++) < '0' || '7' < ch)
 419                                 goto error;
 420                                         ac |= ((ch - '0') << 3);
 421                         if ((ch = *quoted++) < '0' || '7' < ch)
 422                                 goto error;
 423                                         ac |= (ch - '0');
 424                                         ch = ac;
 425                                         break;
 426                                 default:
 427                         goto error;
 428                         }
 429                 strbuf_addch(sb, ch);
 430                 }
 431
 432   error:
 433         strbuf_setlen(sb, oldlen);
 434         return -1;
 435 }
 436
 437 /* quoting as a string literal for other languages */
 438
 439 void perl_quote_buf(struct strbuf *sb, const char *src)
 440 {
 441         const char sq = '\'';
 442         const char bq = '\\';
 443         char c;
 444
 445         strbuf_addch(sb, sq);
 446         while ((c = *src++)) {
 447                 if (c == sq || c == bq)
 448                         strbuf_addch(sb, bq);
 449                 strbuf_addch(sb, c);
 450         }
 451         strbuf_addch(sb, sq);
 452 }
 453
 454 void python_quote_buf(struct strbuf *sb, const char *src)
 455 {
 456         const char sq = '\'';
 457         const char bq = '\\';
 458         const char nl = '\n';
 459         char c;
 460
 461         strbuf_addch(sb, sq);
 462         while ((c = *src++)) {
 463                 if (c == nl) {
 464                         strbuf_addch(sb, bq);
 465                         strbuf_addch(sb, 'n');
 466                         continue;
 467                 }
 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 tcl_quote_buf(struct strbuf *sb, const char *src)
 476 {
 477         char c;
 478
 479         strbuf_addch(sb, '"');
 480         while ((c = *src++)) {
 481                 switch (c) {
 482                 case '[': case ']':
 483                 case '{': case '}':
 484                 case '$': case '\\': case '"':
 485                         strbuf_addch(sb, '\\');
 486                         /* fallthrough */
 487                 default:
 488                         strbuf_addch(sb, c);
 489                         break;
 490                 case '\f':
 491                         strbuf_addstr(sb, "\\f");
 492                         break;
 493                 case '\r':
 494                         strbuf_addstr(sb, "\\r");
 495                         break;
 496                 case '\n':
 497                         strbuf_addstr(sb, "\\n");
 498                         break;
 499                 case '\t':
 500                         strbuf_addstr(sb, "\\t");
 501                         break;
 502                 case '\v':
 503                         strbuf_addstr(sb, "\\v");
 504                         break;
 505                 }
 506         }
 507         strbuf_addch(sb, '"');
 508 }
 509
 510 void basic_regex_quote_buf(struct strbuf *sb, const char *src)
 511 {
 512         char c;
 513
 514         if (*src == '^') {
 515                 /* only beginning '^' is special and needs quoting */
 516                 strbuf_addch(sb, '\\');
 517                 strbuf_addch(sb, *src++);
 518         }
 519         if (*src == '*')
 520                 /* beginning '*' is not special, no quoting */
 521                 strbuf_addch(sb, *src++);
 522
 523         while ((c = *src++)) {
 524                 switch (c) {
 525                 case '[':
 526                 case '.':
 527                 case '\\':
 528                 case '*':
 529                         strbuf_addch(sb, '\\');
 530                         strbuf_addch(sb, c);
 531                         break;
 532
 533                 case '$':
 534                         /* only the end '$' is special and needs quoting */
 535                         if (*src == '\0')
 536                                 strbuf_addch(sb, '\\');
 537                         strbuf_addch(sb, c);
 538                         break;
 539
 540                 default:
 541                         strbuf_addch(sb, c);
 542                         break;
 543                 }
 544         }
 545 }