quote.con commit Merge branch 'jc/branch-d' (630724c)
   1#include "cache.h"
   2#include "quote.h"
   3
   4int 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 */
  17static inline int need_bs_quote(char c)
  18{
  19        return (c == '\'' || c == '!');
  20}
  21
  22void 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
  44void 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
  61void 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
  75static 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
 118char *sq_dequote(char *arg)
 119{
 120        return sq_dequote_step(arg, NULL);
 121}
 122
 123int 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)
 147static 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
 160static 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 */
 168static 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 */
 192static 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 || !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 (sq_lookup[ch] >= ' ') {
 227                        EMIT(sq_lookup[ch]);
 228                } else {
 229                        EMIT(((ch >> 6) & 03) + '0');
 230                        EMIT(((ch >> 3) & 07) + '0');
 231                        EMIT(((ch >> 0) & 07) + '0');
 232                }
 233        }
 234
 235        EMITBUF(p, len);
 236        if (p == name)   /* no ending quote needed */
 237                return 0;
 238
 239        if (!no_dq)
 240                EMIT('"');
 241        return count;
 242}
 243
 244size_t quote_c_style(const char *name, struct strbuf *sb, FILE *fp, int nodq)
 245{
 246        return quote_c_style_counted(name, -1, sb, fp, nodq);
 247}
 248
 249void quote_two_c_style(struct strbuf *sb, const char *prefix, const char *path, int nodq)
 250{
 251        if (quote_c_style(prefix, NULL, NULL, 0) ||
 252            quote_c_style(path, NULL, NULL, 0)) {
 253                if (!nodq)
 254                        strbuf_addch(sb, '"');
 255                quote_c_style(prefix, sb, NULL, 1);
 256                quote_c_style(path, sb, NULL, 1);
 257                if (!nodq)
 258                        strbuf_addch(sb, '"');
 259        } else {
 260                strbuf_addstr(sb, prefix);
 261                strbuf_addstr(sb, path);
 262        }
 263}
 264
 265void write_name_quoted(const char *name, FILE *fp, int terminator)
 266{
 267        if (terminator) {
 268                quote_c_style(name, NULL, fp, 0);
 269        } else {
 270                fputs(name, fp);
 271        }
 272        fputc(terminator, fp);
 273}
 274
 275void write_name_quotedpfx(const char *pfx, size_t pfxlen,
 276                          const char *name, FILE *fp, int terminator)
 277{
 278        int needquote = 0;
 279
 280        if (terminator) {
 281                needquote = next_quote_pos(pfx, pfxlen) < pfxlen
 282                        || name[next_quote_pos(name, -1)];
 283        }
 284        if (needquote) {
 285                fputc('"', fp);
 286                quote_c_style_counted(pfx, pfxlen, NULL, fp, 1);
 287                quote_c_style(name, NULL, fp, 1);
 288                fputc('"', fp);
 289        } else {
 290                fwrite(pfx, pfxlen, 1, fp);
 291                fputs(name, fp);
 292        }
 293        fputc(terminator, fp);
 294}
 295
 296/* quote path as relative to the given prefix */
 297char *quote_path_relative(const char *in, int len,
 298                          struct strbuf *out, const char *prefix)
 299{
 300        int needquote;
 301
 302        if (len < 0)
 303                len = strlen(in);
 304
 305        /* "../" prefix itself does not need quoting, but "in" might. */
 306        needquote = next_quote_pos(in, len) < len;
 307        strbuf_setlen(out, 0);
 308        strbuf_grow(out, len);
 309
 310        if (needquote)
 311                strbuf_addch(out, '"');
 312        if (prefix) {
 313                int off = 0;
 314                while (prefix[off] && off < len && prefix[off] == in[off])
 315                        if (prefix[off] == '/') {
 316                                prefix += off + 1;
 317                                in += off + 1;
 318                                len -= off + 1;
 319                                off = 0;
 320                        } else
 321                                off++;
 322
 323                for (; *prefix; prefix++)
 324                        if (*prefix == '/')
 325                                strbuf_addstr(out, "../");
 326        }
 327
 328        quote_c_style_counted (in, len, out, NULL, 1);
 329
 330        if (needquote)
 331                strbuf_addch(out, '"');
 332        if (!out->len)
 333                strbuf_addstr(out, "./");
 334
 335        return out->buf;
 336}
 337
 338/*
 339 * C-style name unquoting.
 340 *
 341 * Quoted should point at the opening double quote.
 342 * + Returns 0 if it was able to unquote the string properly, and appends the
 343 *   result in the strbuf `sb'.
 344 * + Returns -1 in case of error, and doesn't touch the strbuf. Though note
 345 *   that this function will allocate memory in the strbuf, so calling
 346 *   strbuf_release is mandatory whichever result unquote_c_style returns.
 347 *
 348 * Updates endp pointer to point at one past the ending double quote if given.
 349 */
 350int unquote_c_style(struct strbuf *sb, const char *quoted, const char **endp)
 351{
 352        size_t oldlen = sb->len, len;
 353        int ch, ac;
 354
 355        if (*quoted++ != '"')
 356                return -1;
 357
 358        for (;;) {
 359                len = strcspn(quoted, "\"\\");
 360                strbuf_add(sb, quoted, len);
 361                quoted += len;
 362
 363                switch (*quoted++) {
 364                  case '"':
 365                        if (endp)
 366                                *endp = quoted;
 367                        return 0;
 368                  case '\\':
 369                        break;
 370                  default:
 371                        goto error;
 372                }
 373
 374                switch ((ch = *quoted++)) {
 375                case 'a': ch = '\a'; break;
 376                case 'b': ch = '\b'; break;
 377                case 'f': ch = '\f'; break;
 378                case 'n': ch = '\n'; break;
 379                case 'r': ch = '\r'; break;
 380                case 't': ch = '\t'; break;
 381                case 'v': ch = '\v'; break;
 382
 383                case '\\': case '"':
 384                        break; /* verbatim */
 385
 386                /* octal values with first digit over 4 overflow */
 387                case '0': case '1': case '2': case '3':
 388                                        ac = ((ch - '0') << 6);
 389                        if ((ch = *quoted++) < '0' || '7' < ch)
 390                                goto error;
 391                                        ac |= ((ch - '0') << 3);
 392                        if ((ch = *quoted++) < '0' || '7' < ch)
 393                                goto error;
 394                                        ac |= (ch - '0');
 395                                        ch = ac;
 396                                        break;
 397                                default:
 398                        goto error;
 399                        }
 400                strbuf_addch(sb, ch);
 401                }
 402
 403  error:
 404        strbuf_setlen(sb, oldlen);
 405        return -1;
 406}
 407
 408/* quoting as a string literal for other languages */
 409
 410void perl_quote_print(FILE *stream, const char *src)
 411{
 412        const char sq = '\'';
 413        const char bq = '\\';
 414        char c;
 415
 416        fputc(sq, stream);
 417        while ((c = *src++)) {
 418                if (c == sq || c == bq)
 419                        fputc(bq, stream);
 420                fputc(c, stream);
 421        }
 422        fputc(sq, stream);
 423}
 424
 425void python_quote_print(FILE *stream, const char *src)
 426{
 427        const char sq = '\'';
 428        const char bq = '\\';
 429        const char nl = '\n';
 430        char c;
 431
 432        fputc(sq, stream);
 433        while ((c = *src++)) {
 434                if (c == nl) {
 435                        fputc(bq, stream);
 436                        fputc('n', stream);
 437                        continue;
 438                }
 439                if (c == sq || c == bq)
 440                        fputc(bq, stream);
 441                fputc(c, stream);
 442        }
 443        fputc(sq, stream);
 444}
 445
 446void tcl_quote_print(FILE *stream, const char *src)
 447{
 448        char c;
 449
 450        fputc('"', stream);
 451        while ((c = *src++)) {
 452                switch (c) {
 453                case '[': case ']':
 454                case '{': case '}':
 455                case '$': case '\\': case '"':
 456                        fputc('\\', stream);
 457                default:
 458                        fputc(c, stream);
 459                        break;
 460                case '\f':
 461                        fputs("\\f", stream);
 462                        break;
 463                case '\r':
 464                        fputs("\\r", stream);
 465                        break;
 466                case '\n':
 467                        fputs("\\n", stream);
 468                        break;
 469                case '\t':
 470                        fputs("\\t", stream);
 471                        break;
 472                case '\v':
 473                        fputs("\\v", stream);
 474                        break;
 475                }
 476        }
 477        fputc('"', stream);
 478}