quote.c

   1 #include "cache.h"
   2 #include "quote.h"
   3
   4 int quote_path_fully = 1;
   5
   6 /* Help to copy the thing properly quoted for the shell safety.
   7  * any single quote is replaced with '\'', any exclamation point
   8  * is replaced with '\!', and the whole thing is enclosed in a
   9  *
  10  * E.g.
  11  *  original     sq_quote     result
  12  *  name     ==> name      ==> 'name'
  13  *  a b      ==> a b       ==> 'a b'
  14  *  a'b      ==> a'\''b    ==> 'a'\''b'
  15  *  a!b      ==> a'\!'b    ==> 'a'\!'b'
  16  */
  17 static inline int need_bs_quote(char c)
  18 {
  19         return (c == '\'' || c == '!');
  20 }
  21
  22 void sq_quote_buf(struct strbuf *dst, const char *src)
  23 {
  24         char *to_free = NULL;
  25
  26         if (dst->buf == src)
  27                 to_free = strbuf_detach(dst, NULL);
  28
  29         strbuf_addch(dst, '\'');
  30         while (*src) {
  31                 size_t len = strcspn(src, "'!");
  32                 strbuf_add(dst, src, len);
  33                 src += len;
  34                 while (need_bs_quote(*src)) {
  35                         strbuf_addstr(dst, "'\\");
  36                         strbuf_addch(dst, *src++);
  37                         strbuf_addch(dst, '\'');
  38                 }
  39         }
  40         strbuf_addch(dst, '\'');
  41         free(to_free);
  42 }
  43
  44 void sq_quote_print(FILE *stream, const char *src)
  45 {
  46         char c;
  47
  48         fputc('\'', stream);
  49         while ((c = *src++)) {
  50                 if (need_bs_quote(c)) {
  51                         fputs("'\\", stream);
  52                         fputc(c, stream);
  53                         fputc('\'', stream);
  54                 } else {
  55                         fputc(c, stream);
  56                 }
  57         }
  58         fputc('\'', stream);
  59 }
  60
  61 void sq_quote_argv(struct strbuf *dst, const char** argv, size_t maxlen)
  62 {
  63         int i;
  64
  65         /* Copy into destination buffer. */
  66         strbuf_grow(dst, 255);
  67         for (i = 0; argv[i]; ++i) {
  68                 strbuf_addch(dst, ' ');
  69                 sq_quote_buf(dst, argv[i]);
  70                 if (maxlen && dst->len > maxlen)
  71                         die("Too many or long arguments");
  72         }
  73 }
  74
  75 static char *sq_dequote_step(char *arg, char **next)
  76 {
  77         char *dst = arg;
  78         char *src = arg;
  79         char c;
  80
  81         if (*src != '\'')
  82                 return NULL;
  83         for (;;) {
  84                 c = *++src;
  85                 if (!c)
  86                         return NULL;
  87                 if (c != '\'') {
  88                         *dst++ = c;
  89                         continue;
  90                 }
  91                 /* We stepped out of sq */
  92                 switch (*++src) {
  93                 case '\0':
  94                         *dst = 0;
  95                         if (next)
  96                                 *next = NULL;
  97                         return arg;
  98                 case '\\':
  99                         c = *++src;
 100                         if (need_bs_quote(c) && *++src == '\'') {
 101                                 *dst++ = c;
 102                                 continue;
 103                         }
 104                 /* Fallthrough */
 105                 default:
 106                         if (!next || !isspace(*src))
 107                                 return NULL;
 108                         do {
 109                                 c = *++src;
 110                         } while (isspace(c));
 111                         *dst = 0;
 112                         *next = src;
 113                         return arg;
 114                 }
 115         }
 116 }
 117
 118 char *sq_dequote(char *arg)
 119 {
 120         return sq_dequote_step(arg, NULL);
 121 }
 122
 123 int sq_dequote_to_argv(char *arg, const char ***argv, int *nr, int *alloc)
 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                 ALLOC_GROW(*argv, *nr + 1, *alloc);
 134                 (*argv)[(*nr)++] = dequoted;
 135         } while (next);
 136
 137         return 0;
 138 }
 139
 140 /* 1 means: quote as octal
 141  * 0 means: quote as octal if (quote_path_fully)
 142  * -1 means: never quote
 143  * c: quote as "\\c"
 144  */
 145 #define X8(x)   x, x, x, x, x, x, x, x
 146 #define X16(x)  X8(x), X8(x)
 147 static signed char const sq_lookup[256] = {
 148         /*           0    1    2    3    4    5    6    7 */
 149         /* 0x00 */   1,   1,   1,   1,   1,   1,   1, 'a',
 150         /* 0x08 */ 'b', 't', 'n', 'v', 'f', 'r',   1,   1,
 151         /* 0x10 */ X16(1),
 152         /* 0x20 */  -1,  -1, '"',  -1,  -1,  -1,  -1,  -1,
 153         /* 0x28 */ X16(-1), X16(-1), X16(-1),
 154         /* 0x58 */  -1,  -1,  -1,  -1,'\\',  -1,  -1,  -1,
 155         /* 0x60 */ X16(-1), X8(-1),
 156         /* 0x78 */  -1,  -1,  -1,  -1,  -1,  -1,  -1,   1,
 157         /* 0x80 */ /* set to 0 */
 158 };
 159
 160 static inline int sq_must_quote(char c)
 161 {
 162         return sq_lookup[(unsigned char)c] + quote_path_fully > 0;
 163 }
 164
 165 /* returns the longest prefix not needing a quote up to maxlen if positive.
 166    This stops at the first \0 because it's marked as a character needing an
 167    escape */
 168 static size_t next_quote_pos(const char *s, ssize_t maxlen)
 169 {
 170         size_t len;
 171         if (maxlen < 0) {
 172                 for (len = 0; !sq_must_quote(s[len]); len++);
 173         } else {
 174                 for (len = 0; len < maxlen && !sq_must_quote(s[len]); len++);
 175         }
 176         return len;
 177 }
 178
 179 /*
 180  * C-style name quoting.
 181  *
 182  * (1) if sb and fp are both NULL, inspect the input name and counts the
 183  *     number of bytes that are needed to hold c_style quoted version of name,
 184  *     counting the double quotes around it but not terminating NUL, and
 185  *     returns it.
 186  *     However, if name does not need c_style quoting, it returns 0.
 187  *
 188  * (2) if sb or fp are not NULL, it emits the c_style quoted version
 189  *     of name, enclosed with double quotes if asked and needed only.
 190  *     Return value is the same as in (1).
 191  */
 192 static size_t quote_c_style_counted(const char *name, ssize_t maxlen,
 193                                     struct strbuf *sb, FILE *fp, int no_dq)
 194 {
 195 #undef EMIT
 196 #define EMIT(c)                                 \
 197         do {                                        \
 198                 if (sb) strbuf_addch(sb, (c));          \
 199                 if (fp) fputc((c), fp);                 \
 200                 count++;                                \
 201         } while (0)
 202 #define EMITBUF(s, l)                           \
 203         do {                                        \
 204                 if (sb) strbuf_add(sb, (s), (l));       \
 205                 if (fp) fwrite((s), (l), 1, fp);        \
 206                 count += (l);                           \
 207         } while (0)
 208
 209         size_t len, count = 0;
 210         const char *p = name;
 211
 212         for (;;) {
 213                 int ch;
 214
 215                 len = next_quote_pos(p, maxlen);
 216                 if (len == maxlen || (maxlen < 0 && !p[len]))
 217                         break;
 218
 219                 if (!no_dq && p == name)
 220                         EMIT('"');
 221
 222                 EMITBUF(p, len);
 223                 EMIT('\\');
 224                 p += len;
 225                 ch = (unsigned char)*p++;
 226                 if (maxlen >= 0)
 227                         maxlen -= len + 1;
 228                 if (sq_lookup[ch] >= ' ') {
 229                         EMIT(sq_lookup[ch]);
 230                 } else {
 231                         EMIT(((ch >> 6) & 03) + '0');
 232                         EMIT(((ch >> 3) & 07) + '0');
 233                         EMIT(((ch >> 0) & 07) + '0');
 234                 }
 235         }
 236
 237         EMITBUF(p, len);
 238         if (p == name)   /* no ending quote needed */
 239                 return 0;
 240
 241         if (!no_dq)
 242                 EMIT('"');
 243         return count;
 244 }
 245
 246 size_t quote_c_style(const char *name, struct strbuf *sb, FILE *fp, int nodq)
 247 {
 248         return quote_c_style_counted(name, -1, sb, fp, nodq);
 249 }
 250
 251 void quote_two_c_style(struct strbuf *sb, const char *prefix, const char *path, int nodq)
 252 {
 253         if (quote_c_style(prefix, NULL, NULL, 0) ||
 254             quote_c_style(path, NULL, NULL, 0)) {
 255                 if (!nodq)
 256                         strbuf_addch(sb, '"');
 257                 quote_c_style(prefix, sb, NULL, 1);
 258                 quote_c_style(path, sb, NULL, 1);
 259                 if (!nodq)
 260                         strbuf_addch(sb, '"');
 261         } else {
 262                 strbuf_addstr(sb, prefix);
 263                 strbuf_addstr(sb, path);
 264         }
 265 }
 266
 267 void write_name_quoted(const char *name, FILE *fp, int terminator)
 268 {
 269         if (terminator) {
 270                 quote_c_style(name, NULL, fp, 0);
 271         } else {
 272                 fputs(name, fp);
 273         }
 274         fputc(terminator, fp);
 275 }
 276
 277 void write_name_quotedpfx(const char *pfx, size_t pfxlen,
 278                           const char *name, FILE *fp, int terminator)
 279 {
 280         int needquote = 0;
 281
 282         if (terminator) {
 283                 needquote = next_quote_pos(pfx, pfxlen) < pfxlen
 284                         || name[next_quote_pos(name, -1)];
 285         }
 286         if (needquote) {
 287                 fputc('"', fp);
 288                 quote_c_style_counted(pfx, pfxlen, NULL, fp, 1);
 289                 quote_c_style(name, NULL, fp, 1);
 290                 fputc('"', fp);
 291         } else {
 292                 fwrite(pfx, pfxlen, 1, fp);
 293                 fputs(name, fp);
 294         }
 295         fputc(terminator, fp);
 296 }
 297
 298 static const char *path_relative(const char *in, int len,
 299                                  struct strbuf *sb, const char *prefix,
 300                                  int prefix_len);
 301
 302 void write_name_quoted_relative(const char *name, size_t len,
 303                                 const char *prefix, size_t prefix_len,
 304                                 FILE *fp, int terminator)
 305 {
 306         struct strbuf sb = STRBUF_INIT;
 307
 308         name = path_relative(name, len, &sb, prefix, prefix_len);
 309         write_name_quoted(name, fp, terminator);
 310
 311         strbuf_release(&sb);
 312 }
 313
 314 /*
 315  * Give path as relative to prefix.
 316  *
 317  * The strbuf may or may not be used, so do not assume it contains the
 318  * returned path.
 319  */
 320 static const char *path_relative(const char *in, int len,
 321                                  struct strbuf *sb, const char *prefix,
 322                                  int prefix_len)
 323 {
 324         int off, i;
 325
 326         if (len < 0)
 327                 len = strlen(in);
 328         if (prefix_len < 0) {
 329                 if (prefix)
 330                         prefix_len = strlen(prefix);
 331                 else
 332                         prefix_len = 0;
 333         }
 334
 335         off = 0;
 336         i = 0;
 337         while (i < prefix_len && i < len && prefix[i] == in[i]) {
 338                 if (prefix[i] == '/')
 339                         off = i + 1;
 340                 i++;
 341         }
 342         in += off;
 343         len -= off;
 344
 345         if (i >= prefix_len)
 346                 return in;
 347
 348         strbuf_reset(sb);
 349         strbuf_grow(sb, len);
 350
 351         while (i < prefix_len) {
 352                 if (prefix[i] == '/')
 353                         strbuf_addstr(sb, "../");
 354                 i++;
 355         }
 356         strbuf_add(sb, in, len);
 357
 358         return sb->buf;
 359 }
 360
 361 /* quote path as relative to the given prefix */
 362 char *quote_path_relative(const char *in, int len,
 363                           struct strbuf *out, const char *prefix)
 364 {
 365         struct strbuf sb = STRBUF_INIT;
 366         const char *rel = path_relative(in, len, &sb, prefix, -1);
 367         strbuf_reset(out);
 368         quote_c_style_counted(rel, strlen(rel), out, NULL, 0);
 369         strbuf_release(&sb);
 370
 371         if (!out->len)
 372                 strbuf_addstr(out, "./");
 373
 374         return out->buf;
 375 }
 376
 377 /*
 378  * C-style name unquoting.
 379  *
 380  * Quoted should point at the opening double quote.
 381  * + Returns 0 if it was able to unquote the string properly, and appends the
 382  *   result in the strbuf `sb'.
 383  * + Returns -1 in case of error, and doesn't touch the strbuf. Though note
 384  *   that this function will allocate memory in the strbuf, so calling
 385  *   strbuf_release is mandatory whichever result unquote_c_style returns.
 386  *
 387  * Updates endp pointer to point at one past the ending double quote if given.
 388  */
 389 int unquote_c_style(struct strbuf *sb, const char *quoted, const char **endp)
 390 {
 391         size_t oldlen = sb->len, len;
 392         int ch, ac;
 393
 394         if (*quoted++ != '"')
 395                 return -1;
 396
 397         for (;;) {
 398                 len = strcspn(quoted, "\"\\");
 399                 strbuf_add(sb, quoted, len);
 400                 quoted += len;
 401
 402                 switch (*quoted++) {
 403                   case '"':
 404                         if (endp)
 405                                 *endp = quoted;
 406                         return 0;
 407                   case '\\':
 408                         break;
 409                   default:
 410                         goto error;
 411                 }
 412
 413                 switch ((ch = *quoted++)) {
 414                 case 'a': ch = '\a'; break;
 415                 case 'b': ch = '\b'; break;
 416                 case 'f': ch = '\f'; break;
 417                 case 'n': ch = '\n'; break;
 418                 case 'r': ch = '\r'; break;
 419                 case 't': ch = '\t'; break;
 420                 case 'v': ch = '\v'; break;
 421
 422                 case '\\': case '"':
 423                         break; /* verbatim */
 424
 425                 /* octal values with first digit over 4 overflow */
 426                 case '0': case '1': case '2': case '3':
 427                                         ac = ((ch - '0') << 6);
 428                         if ((ch = *quoted++) < '0' || '7' < ch)
 429                                 goto error;
 430                                         ac |= ((ch - '0') << 3);
 431                         if ((ch = *quoted++) < '0' || '7' < ch)
 432                                 goto error;
 433                                         ac |= (ch - '0');
 434                                         ch = ac;
 435                                         break;
 436                                 default:
 437                         goto error;
 438                         }
 439                 strbuf_addch(sb, ch);
 440                 }
 441
 442   error:
 443         strbuf_setlen(sb, oldlen);
 444         return -1;
 445 }
 446
 447 /* quoting as a string literal for other languages */
 448
 449 void perl_quote_print(FILE *stream, const char *src)
 450 {
 451         const char sq = '\'';
 452         const char bq = '\\';
 453         char c;
 454
 455         fputc(sq, stream);
 456         while ((c = *src++)) {
 457                 if (c == sq || c == bq)
 458                         fputc(bq, stream);
 459                 fputc(c, stream);
 460         }
 461         fputc(sq, stream);
 462 }
 463
 464 void python_quote_print(FILE *stream, const char *src)
 465 {
 466         const char sq = '\'';
 467         const char bq = '\\';
 468         const char nl = '\n';
 469         char c;
 470
 471         fputc(sq, stream);
 472         while ((c = *src++)) {
 473                 if (c == nl) {
 474                         fputc(bq, stream);
 475                         fputc('n', stream);
 476                         continue;
 477                 }
 478                 if (c == sq || c == bq)
 479                         fputc(bq, stream);
 480                 fputc(c, stream);
 481         }
 482         fputc(sq, stream);
 483 }
 484
 485 void tcl_quote_print(FILE *stream, const char *src)
 486 {
 487         char c;
 488
 489         fputc('"', stream);
 490         while ((c = *src++)) {
 491                 switch (c) {
 492                 case '[': case ']':
 493                 case '{': case '}':
 494                 case '$': case '\\': case '"':
 495                         fputc('\\', stream);
 496                 default:
 497                         fputc(c, stream);
 498                         break;
 499                 case '\f':
 500                         fputs("\\f", stream);
 501                         break;
 502                 case '\r':
 503                         fputs("\\r", stream);
 504                         break;
 505                 case '\n':
 506                         fputs("\\n", stream);
 507                         break;
 508                 case '\t':
 509                         fputs("\\t", stream);
 510                         break;
 511                 case '\v':
 512                         fputs("\\v", stream);
 513                         break;
 514                 }
 515         }
 516         fputc('"', stream);
 517 }