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