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