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