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