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