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_quotef(struct strbuf *dst, const char *fmt, ...)
  47 {
  48         struct strbuf src = STRBUF_INIT;
  49
  50         va_list ap;
  51         va_start(ap, fmt);
  52         strbuf_vaddf(&src, fmt, ap);
  53         va_end(ap);
  54
  55         sq_quote_buf(dst, src.buf);
  56         strbuf_release(&src);
  57 }
  58
  59 void sq_quote_argv(struct strbuf *dst, const char** argv, size_t maxlen)
  60 {
  61         int i;
  62
  63         /* Copy into destination buffer. */
  64         strbuf_grow(dst, 255);
  65         for (i = 0; argv[i]; ++i) {
  66                 strbuf_addch(dst, ' ');
  67                 sq_quote_buf(dst, argv[i]);
  68                 if (maxlen && dst->len > maxlen)
  69                         die("Too many or long arguments");
  70         }
  71 }
  72
  73 static char *sq_dequote_step(char *arg, char **next)
  74 {
  75         char *dst = arg;
  76         char *src = arg;
  77         char c;
  78
  79         if (*src != '\'')
  80                 return NULL;
  81         for (;;) {
  82                 c = *++src;
  83                 if (!c)
  84                         return NULL;
  85                 if (c != '\'') {
  86                         *dst++ = c;
  87                         continue;
  88                 }
  89                 /* We stepped out of sq */
  90                 switch (*++src) {
  91                 case '\0':
  92                         *dst = 0;
  93                         if (next)
  94                                 *next = NULL;
  95                         return arg;
  96                 case '\\':
  97                         c = *++src;
  98                         if (need_bs_quote(c) && *++src == '\'') {
  99                                 *dst++ = c;
 100                                 continue;
 101                         }
 102                 /* Fallthrough */
 103                 default:
 104                         if (!next || !isspace(*src))
 105                                 return NULL;
 106                         do {
 107                                 c = *++src;
 108                         } while (isspace(c));
 109                         *dst = 0;
 110                         *next = src;
 111                         return arg;
 112                 }
 113         }
 114 }
 115
 116 char *sq_dequote(char *arg)
 117 {
 118         return sq_dequote_step(arg, NULL);
 119 }
 120
 121 static int sq_dequote_to_argv_internal(char *arg,
 122                                        const char ***argv, int *nr, int *alloc,
 123                                        struct argv_array *array)
 124 {
 125         char *next = arg;
 126
 127         if (!*arg)
 128                 return 0;
 129         do {
 130                 char *dequoted = sq_dequote_step(next, &next);
 131                 if (!dequoted)
 132                         return -1;
 133                 if (argv) {
 134                         ALLOC_GROW(*argv, *nr + 1, *alloc);
 135                         (*argv)[(*nr)++] = dequoted;
 136                 }
 137                 if (array)
 138                         argv_array_push(array, dequoted);
 139         } while (next);
 140
 141         return 0;
 142 }
 143
 144 int sq_dequote_to_argv(char *arg, const char ***argv, int *nr, int *alloc)
 145 {
 146         return sq_dequote_to_argv_internal(arg, argv, nr, alloc, NULL);
 147 }
 148
 149 int sq_dequote_to_argv_array(char *arg, struct argv_array *array)
 150 {
 151         return sq_dequote_to_argv_internal(arg, NULL, NULL, NULL, array);
 152 }
 153
 154 /* 1 means: quote as octal
 155  * 0 means: quote as octal if (quote_path_fully)
 156  * -1 means: never quote
 157  * c: quote as "\\c"
 158  */
 159 #define X8(x)   x, x, x, x, x, x, x, x
 160 #define X16(x)  X8(x), X8(x)
 161 static signed char const sq_lookup[256] = {
 162         /*           0    1    2    3    4    5    6    7 */
 163         /* 0x00 */   1,   1,   1,   1,   1,   1,   1, 'a',
 164         /* 0x08 */ 'b', 't', 'n', 'v', 'f', 'r',   1,   1,
 165         /* 0x10 */ X16(1),
 166         /* 0x20 */  -1,  -1, '"',  -1,  -1,  -1,  -1,  -1,
 167         /* 0x28 */ X16(-1), X16(-1), X16(-1),
 168         /* 0x58 */  -1,  -1,  -1,  -1,'\\',  -1,  -1,  -1,
 169         /* 0x60 */ X16(-1), X8(-1),
 170         /* 0x78 */  -1,  -1,  -1,  -1,  -1,  -1,  -1,   1,
 171         /* 0x80 */ /* set to 0 */
 172 };
 173
 174 static inline int sq_must_quote(char c)
 175 {
 176         return sq_lookup[(unsigned char)c] + quote_path_fully > 0;
 177 }
 178
 179 /* returns the longest prefix not needing a quote up to maxlen if positive.
 180    This stops at the first \0 because it's marked as a character needing an
 181    escape */
 182 static size_t next_quote_pos(const char *s, ssize_t maxlen)
 183 {
 184         size_t len;
 185         if (maxlen < 0) {
 186                 for (len = 0; !sq_must_quote(s[len]); len++);
 187         } else {
 188                 for (len = 0; len < maxlen && !sq_must_quote(s[len]); len++);
 189         }
 190         return len;
 191 }
 192
 193 /*
 194  * C-style name quoting.
 195  *
 196  * (1) if sb and fp are both NULL, inspect the input name and counts the
 197  *     number of bytes that are needed to hold c_style quoted version of name,
 198  *     counting the double quotes around it but not terminating NUL, and
 199  *     returns it.
 200  *     However, if name does not need c_style quoting, it returns 0.
 201  *
 202  * (2) if sb or fp are not NULL, it emits the c_style quoted version
 203  *     of name, enclosed with double quotes if asked and needed only.
 204  *     Return value is the same as in (1).
 205  */
 206 static size_t quote_c_style_counted(const char *name, ssize_t maxlen,
 207                                     struct strbuf *sb, FILE *fp, int no_dq)
 208 {
 209 #undef EMIT
 210 #define EMIT(c)                                 \
 211         do {                                        \
 212                 if (sb) strbuf_addch(sb, (c));          \
 213                 if (fp) fputc((c), fp);                 \
 214                 count++;                                \
 215         } while (0)
 216 #define EMITBUF(s, l)                           \
 217         do {                                        \
 218                 if (sb) strbuf_add(sb, (s), (l));       \
 219                 if (fp) fwrite((s), (l), 1, fp);        \
 220                 count += (l);                           \
 221         } while (0)
 222
 223         size_t len, count = 0;
 224         const char *p = name;
 225
 226         for (;;) {
 227                 int ch;
 228
 229                 len = next_quote_pos(p, maxlen);
 230                 if (len == maxlen || (maxlen < 0 && !p[len]))
 231                         break;
 232
 233                 if (!no_dq && p == name)
 234                         EMIT('"');
 235
 236                 EMITBUF(p, len);
 237                 EMIT('\\');
 238                 p += len;
 239                 ch = (unsigned char)*p++;
 240                 if (maxlen >= 0)
 241                         maxlen -= len + 1;
 242                 if (sq_lookup[ch] >= ' ') {
 243                         EMIT(sq_lookup[ch]);
 244                 } else {
 245                         EMIT(((ch >> 6) & 03) + '0');
 246                         EMIT(((ch >> 3) & 07) + '0');
 247                         EMIT(((ch >> 0) & 07) + '0');
 248                 }
 249         }
 250
 251         EMITBUF(p, len);
 252         if (p == name)   /* no ending quote needed */
 253                 return 0;
 254
 255         if (!no_dq)
 256                 EMIT('"');
 257         return count;
 258 }
 259
 260 size_t quote_c_style(const char *name, struct strbuf *sb, FILE *fp, int nodq)
 261 {
 262         return quote_c_style_counted(name, -1, sb, fp, nodq);
 263 }
 264
 265 void quote_two_c_style(struct strbuf *sb, const char *prefix, const char *path, int nodq)
 266 {
 267         if (quote_c_style(prefix, NULL, NULL, 0) ||
 268             quote_c_style(path, NULL, NULL, 0)) {
 269                 if (!nodq)
 270                         strbuf_addch(sb, '"');
 271                 quote_c_style(prefix, sb, NULL, 1);
 272                 quote_c_style(path, sb, NULL, 1);
 273                 if (!nodq)
 274                         strbuf_addch(sb, '"');
 275         } else {
 276                 strbuf_addstr(sb, prefix);
 277                 strbuf_addstr(sb, path);
 278         }
 279 }
 280
 281 void write_name_quoted(const char *name, FILE *fp, int terminator)
 282 {
 283         if (terminator) {
 284                 quote_c_style(name, NULL, fp, 0);
 285         } else {
 286                 fputs(name, fp);
 287         }
 288         fputc(terminator, fp);
 289 }
 290
 291 void write_name_quoted_relative(const char *name, const char *prefix,
 292                                 FILE *fp, int terminator)
 293 {
 294         struct strbuf sb = STRBUF_INIT;
 295
 296         name = relative_path(name, prefix, &sb);
 297         write_name_quoted(name, fp, terminator);
 298
 299         strbuf_release(&sb);
 300 }
 301
 302 /* quote path as relative to the given prefix */
 303 char *quote_path_relative(const char *in, const char *prefix,
 304                           struct strbuf *out)
 305 {
 306         struct strbuf sb = STRBUF_INIT;
 307         const char *rel = relative_path(in, prefix, &sb);
 308         strbuf_reset(out);
 309         quote_c_style_counted(rel, strlen(rel), out, NULL, 0);
 310         strbuf_release(&sb);
 311
 312         return out->buf;
 313 }
 314
 315 /*
 316  * C-style name unquoting.
 317  *
 318  * Quoted should point at the opening double quote.
 319  * + Returns 0 if it was able to unquote the string properly, and appends the
 320  *   result in the strbuf `sb'.
 321  * + Returns -1 in case of error, and doesn't touch the strbuf. Though note
 322  *   that this function will allocate memory in the strbuf, so calling
 323  *   strbuf_release is mandatory whichever result unquote_c_style returns.
 324  *
 325  * Updates endp pointer to point at one past the ending double quote if given.
 326  */
 327 int unquote_c_style(struct strbuf *sb, const char *quoted, const char **endp)
 328 {
 329         size_t oldlen = sb->len, len;
 330         int ch, ac;
 331
 332         if (*quoted++ != '"')
 333                 return -1;
 334
 335         for (;;) {
 336                 len = strcspn(quoted, "\"\\");
 337                 strbuf_add(sb, quoted, len);
 338                 quoted += len;
 339
 340                 switch (*quoted++) {
 341                   case '"':
 342                         if (endp)
 343                                 *endp = quoted;
 344                         return 0;
 345                   case '\\':
 346                         break;
 347                   default:
 348                         goto error;
 349                 }
 350
 351                 switch ((ch = *quoted++)) {
 352                 case 'a': ch = '\a'; break;
 353                 case 'b': ch = '\b'; break;
 354                 case 'f': ch = '\f'; break;
 355                 case 'n': ch = '\n'; break;
 356                 case 'r': ch = '\r'; break;
 357                 case 't': ch = '\t'; break;
 358                 case 'v': ch = '\v'; break;
 359
 360                 case '\\': case '"':
 361                         break; /* verbatim */
 362
 363                 /* octal values with first digit over 4 overflow */
 364                 case '0': case '1': case '2': case '3':
 365                                         ac = ((ch - '0') << 6);
 366                         if ((ch = *quoted++) < '0' || '7' < ch)
 367                                 goto error;
 368                                         ac |= ((ch - '0') << 3);
 369                         if ((ch = *quoted++) < '0' || '7' < ch)
 370                                 goto error;
 371                                         ac |= (ch - '0');
 372                                         ch = ac;
 373                                         break;
 374                                 default:
 375                         goto error;
 376                         }
 377                 strbuf_addch(sb, ch);
 378                 }
 379
 380   error:
 381         strbuf_setlen(sb, oldlen);
 382         return -1;
 383 }
 384
 385 /* quoting as a string literal for other languages */
 386
 387 void perl_quote_buf(struct strbuf *sb, const char *src)
 388 {
 389         const char sq = '\'';
 390         const char bq = '\\';
 391         char c;
 392
 393         strbuf_addch(sb, sq);
 394         while ((c = *src++)) {
 395                 if (c == sq || c == bq)
 396                         strbuf_addch(sb, bq);
 397                 strbuf_addch(sb, c);
 398         }
 399         strbuf_addch(sb, sq);
 400 }
 401
 402 void python_quote_buf(struct strbuf *sb, const char *src)
 403 {
 404         const char sq = '\'';
 405         const char bq = '\\';
 406         const char nl = '\n';
 407         char c;
 408
 409         strbuf_addch(sb, sq);
 410         while ((c = *src++)) {
 411                 if (c == nl) {
 412                         strbuf_addch(sb, bq);
 413                         strbuf_addch(sb, 'n');
 414                         continue;
 415                 }
 416                 if (c == sq || c == bq)
 417                         strbuf_addch(sb, bq);
 418                 strbuf_addch(sb, c);
 419         }
 420         strbuf_addch(sb, sq);
 421 }
 422
 423 void tcl_quote_buf(struct strbuf *sb, const char *src)
 424 {
 425         char c;
 426
 427         strbuf_addch(sb, '"');
 428         while ((c = *src++)) {
 429                 switch (c) {
 430                 case '[': case ']':
 431                 case '{': case '}':
 432                 case '$': case '\\': case '"':
 433                         strbuf_addch(sb, '\\');
 434                 default:
 435                         strbuf_addch(sb, c);
 436                         break;
 437                 case '\f':
 438                         strbuf_addstr(sb, "\\f");
 439                         break;
 440                 case '\r':
 441                         strbuf_addstr(sb, "\\r");
 442                         break;
 443                 case '\n':
 444                         strbuf_addstr(sb, "\\n");
 445                         break;
 446                 case '\t':
 447                         strbuf_addstr(sb, "\\t");
 448                         break;
 449                 case '\v':
 450                         strbuf_addstr(sb, "\\v");
 451                         break;
 452                 }
 453         }
 454         strbuf_addch(sb, '"');
 455 }
 456
 457 void basic_regex_quote_buf(struct strbuf *sb, const char *src)
 458 {
 459         char c;
 460
 461         if (*src == '^') {
 462                 /* only beginning '^' is special and needs quoting */
 463                 strbuf_addch(sb, '\\');
 464                 strbuf_addch(sb, *src++);
 465         }
 466         if (*src == '*')
 467                 /* beginning '*' is not special, no quoting */
 468                 strbuf_addch(sb, *src++);
 469
 470         while ((c = *src++)) {
 471                 switch (c) {
 472                 case '[':
 473                 case '.':
 474                 case '\\':
 475                 case '*':
 476                         strbuf_addch(sb, '\\');
 477                         strbuf_addch(sb, c);
 478                         break;
 479
 480                 case '$':
 481                         /* only the end '$' is special and needs quoting */
 482                         if (*src == '\0')
 483                                 strbuf_addch(sb, '\\');
 484                         strbuf_addch(sb, c);
 485                         break;
 486
 487                 default:
 488                         strbuf_addch(sb, c);
 489                         break;
 490                 }
 491         }
 492 }