quote.con commit general improvements (43abf13)
   1#include "cache.h"
   2#include "quote.h"
   3#include "argv-array.h"
   4
   5int quote_path_fully = 1;
   6
   7static 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 */
  24void 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
  46void 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
  62void 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
  75void 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/*
  88 * Legacy function to append each argv value, quoted as necessasry,
  89 * with whitespace before each value.  This results in a leading
  90 * space in the result.
  91 */
  92void sq_quote_argv_pretty(struct strbuf *dst, const char **argv)
  93{
  94        if (argv[0])
  95                strbuf_addch(dst, ' ');
  96        sq_append_quote_argv_pretty(dst, argv);
  97}
  98
  99/*
 100 * Append each argv value, quoted as necessary, with whitespace between them.
 101 */
 102void sq_append_quote_argv_pretty(struct strbuf *dst, const char **argv)
 103{
 104        int i;
 105
 106        for (i = 0; argv[i]; i++) {
 107                if (i > 0)
 108                        strbuf_addch(dst, ' ');
 109                sq_quote_buf_pretty(dst, argv[i]);
 110        }
 111}
 112
 113static char *sq_dequote_step(char *arg, char **next)
 114{
 115        char *dst = arg;
 116        char *src = arg;
 117        char c;
 118
 119        if (*src != '\'')
 120                return NULL;
 121        for (;;) {
 122                c = *++src;
 123                if (!c)
 124                        return NULL;
 125                if (c != '\'') {
 126                        *dst++ = c;
 127                        continue;
 128                }
 129                /* We stepped out of sq */
 130                switch (*++src) {
 131                case '\0':
 132                        *dst = 0;
 133                        if (next)
 134                                *next = NULL;
 135                        return arg;
 136                case '\\':
 137                        /*
 138                         * Allow backslashed characters outside of
 139                         * single-quotes only if they need escaping,
 140                         * and only if we resume the single-quoted part
 141                         * afterward.
 142                         */
 143                        if (need_bs_quote(src[1]) && src[2] == '\'') {
 144                                *dst++ = src[1];
 145                                src += 2;
 146                                continue;
 147                        }
 148                /* Fallthrough */
 149                default:
 150                        if (!next || !isspace(*src))
 151                                return NULL;
 152                        do {
 153                                c = *++src;
 154                        } while (isspace(c));
 155                        *dst = 0;
 156                        *next = src;
 157                        return arg;
 158                }
 159        }
 160}
 161
 162char *sq_dequote(char *arg)
 163{
 164        return sq_dequote_step(arg, NULL);
 165}
 166
 167static int sq_dequote_to_argv_internal(char *arg,
 168                                       const char ***argv, int *nr, int *alloc,
 169                                       struct argv_array *array)
 170{
 171        char *next = arg;
 172
 173        if (!*arg)
 174                return 0;
 175        do {
 176                char *dequoted = sq_dequote_step(next, &next);
 177                if (!dequoted)
 178                        return -1;
 179                if (argv) {
 180                        ALLOC_GROW(*argv, *nr + 1, *alloc);
 181                        (*argv)[(*nr)++] = dequoted;
 182                }
 183                if (array)
 184                        argv_array_push(array, dequoted);
 185        } while (next);
 186
 187        return 0;
 188}
 189
 190int sq_dequote_to_argv(char *arg, const char ***argv, int *nr, int *alloc)
 191{
 192        return sq_dequote_to_argv_internal(arg, argv, nr, alloc, NULL);
 193}
 194
 195int sq_dequote_to_argv_array(char *arg, struct argv_array *array)
 196{
 197        return sq_dequote_to_argv_internal(arg, NULL, NULL, NULL, array);
 198}
 199
 200/* 1 means: quote as octal
 201 * 0 means: quote as octal if (quote_path_fully)
 202 * -1 means: never quote
 203 * c: quote as "\\c"
 204 */
 205#define X8(x)   x, x, x, x, x, x, x, x
 206#define X16(x)  X8(x), X8(x)
 207static signed char const sq_lookup[256] = {
 208        /*           0    1    2    3    4    5    6    7 */
 209        /* 0x00 */   1,   1,   1,   1,   1,   1,   1, 'a',
 210        /* 0x08 */ 'b', 't', 'n', 'v', 'f', 'r',   1,   1,
 211        /* 0x10 */ X16(1),
 212        /* 0x20 */  -1,  -1, '"',  -1,  -1,  -1,  -1,  -1,
 213        /* 0x28 */ X16(-1), X16(-1), X16(-1),
 214        /* 0x58 */  -1,  -1,  -1,  -1,'\\',  -1,  -1,  -1,
 215        /* 0x60 */ X16(-1), X8(-1),
 216        /* 0x78 */  -1,  -1,  -1,  -1,  -1,  -1,  -1,   1,
 217        /* 0x80 */ /* set to 0 */
 218};
 219
 220static inline int sq_must_quote(char c)
 221{
 222        return sq_lookup[(unsigned char)c] + quote_path_fully > 0;
 223}
 224
 225/* returns the longest prefix not needing a quote up to maxlen if positive.
 226   This stops at the first \0 because it's marked as a character needing an
 227   escape */
 228static size_t next_quote_pos(const char *s, ssize_t maxlen)
 229{
 230        size_t len;
 231        if (maxlen < 0) {
 232                for (len = 0; !sq_must_quote(s[len]); len++);
 233        } else {
 234                for (len = 0; len < maxlen && !sq_must_quote(s[len]); len++);
 235        }
 236        return len;
 237}
 238
 239/*
 240 * C-style name quoting.
 241 *
 242 * (1) if sb and fp are both NULL, inspect the input name and counts the
 243 *     number of bytes that are needed to hold c_style quoted version of name,
 244 *     counting the double quotes around it but not terminating NUL, and
 245 *     returns it.
 246 *     However, if name does not need c_style quoting, it returns 0.
 247 *
 248 * (2) if sb or fp are not NULL, it emits the c_style quoted version
 249 *     of name, enclosed with double quotes if asked and needed only.
 250 *     Return value is the same as in (1).
 251 */
 252static size_t quote_c_style_counted(const char *name, ssize_t maxlen,
 253                                    struct strbuf *sb, FILE *fp, int no_dq)
 254{
 255#undef EMIT
 256#define EMIT(c)                                 \
 257        do {                                        \
 258                if (sb) strbuf_addch(sb, (c));          \
 259                if (fp) fputc((c), fp);                 \
 260                count++;                                \
 261        } while (0)
 262#define EMITBUF(s, l)                           \
 263        do {                                        \
 264                if (sb) strbuf_add(sb, (s), (l));       \
 265                if (fp) fwrite((s), (l), 1, fp);        \
 266                count += (l);                           \
 267        } while (0)
 268
 269        size_t len, count = 0;
 270        const char *p = name;
 271
 272        for (;;) {
 273                int ch;
 274
 275                len = next_quote_pos(p, maxlen);
 276                if (len == maxlen || (maxlen < 0 && !p[len]))
 277                        break;
 278
 279                if (!no_dq && p == name)
 280                        EMIT('"');
 281
 282                EMITBUF(p, len);
 283                EMIT('\\');
 284                p += len;
 285                ch = (unsigned char)*p++;
 286                if (maxlen >= 0)
 287                        maxlen -= len + 1;
 288                if (sq_lookup[ch] >= ' ') {
 289                        EMIT(sq_lookup[ch]);
 290                } else {
 291                        EMIT(((ch >> 6) & 03) + '0');
 292                        EMIT(((ch >> 3) & 07) + '0');
 293                        EMIT(((ch >> 0) & 07) + '0');
 294                }
 295        }
 296
 297        EMITBUF(p, len);
 298        if (p == name)   /* no ending quote needed */
 299                return 0;
 300
 301        if (!no_dq)
 302                EMIT('"');
 303        return count;
 304}
 305
 306size_t quote_c_style(const char *name, struct strbuf *sb, FILE *fp, int nodq)
 307{
 308        return quote_c_style_counted(name, -1, sb, fp, nodq);
 309}
 310
 311void quote_two_c_style(struct strbuf *sb, const char *prefix, const char *path, int nodq)
 312{
 313        if (quote_c_style(prefix, NULL, NULL, 0) ||
 314            quote_c_style(path, NULL, NULL, 0)) {
 315                if (!nodq)
 316                        strbuf_addch(sb, '"');
 317                quote_c_style(prefix, sb, NULL, 1);
 318                quote_c_style(path, sb, NULL, 1);
 319                if (!nodq)
 320                        strbuf_addch(sb, '"');
 321        } else {
 322                strbuf_addstr(sb, prefix);
 323                strbuf_addstr(sb, path);
 324        }
 325}
 326
 327void write_name_quoted(const char *name, FILE *fp, int terminator)
 328{
 329        if (terminator) {
 330                quote_c_style(name, NULL, fp, 0);
 331        } else {
 332                fputs(name, fp);
 333        }
 334        fputc(terminator, fp);
 335}
 336
 337void write_name_quoted_relative(const char *name, const char *prefix,
 338                                FILE *fp, int terminator)
 339{
 340        struct strbuf sb = STRBUF_INIT;
 341
 342        name = relative_path(name, prefix, &sb);
 343        write_name_quoted(name, fp, terminator);
 344
 345        strbuf_release(&sb);
 346}
 347
 348/* quote path as relative to the given prefix */
 349char *quote_path_relative(const char *in, const char *prefix,
 350                          struct strbuf *out)
 351{
 352        struct strbuf sb = STRBUF_INIT;
 353        const char *rel = relative_path(in, prefix, &sb);
 354        strbuf_reset(out);
 355        quote_c_style_counted(rel, strlen(rel), out, NULL, 0);
 356        strbuf_release(&sb);
 357
 358        return out->buf;
 359}
 360
 361/*
 362 * C-style name unquoting.
 363 *
 364 * Quoted should point at the opening double quote.
 365 * + Returns 0 if it was able to unquote the string properly, and appends the
 366 *   result in the strbuf `sb'.
 367 * + Returns -1 in case of error, and doesn't touch the strbuf. Though note
 368 *   that this function will allocate memory in the strbuf, so calling
 369 *   strbuf_release is mandatory whichever result unquote_c_style returns.
 370 *
 371 * Updates endp pointer to point at one past the ending double quote if given.
 372 */
 373int unquote_c_style(struct strbuf *sb, const char *quoted, const char **endp)
 374{
 375        size_t oldlen = sb->len, len;
 376        int ch, ac;
 377
 378        if (*quoted++ != '"')
 379                return -1;
 380
 381        for (;;) {
 382                len = strcspn(quoted, "\"\\");
 383                strbuf_add(sb, quoted, len);
 384                quoted += len;
 385
 386                switch (*quoted++) {
 387                  case '"':
 388                        if (endp)
 389                                *endp = quoted;
 390                        return 0;
 391                  case '\\':
 392                        break;
 393                  default:
 394                        goto error;
 395                }
 396
 397                switch ((ch = *quoted++)) {
 398                case 'a': ch = '\a'; break;
 399                case 'b': ch = '\b'; break;
 400                case 'f': ch = '\f'; break;
 401                case 'n': ch = '\n'; break;
 402                case 'r': ch = '\r'; break;
 403                case 't': ch = '\t'; break;
 404                case 'v': ch = '\v'; break;
 405
 406                case '\\': case '"':
 407                        break; /* verbatim */
 408
 409                /* octal values with first digit over 4 overflow */
 410                case '0': case '1': case '2': case '3':
 411                                        ac = ((ch - '0') << 6);
 412                        if ((ch = *quoted++) < '0' || '7' < ch)
 413                                goto error;
 414                                        ac |= ((ch - '0') << 3);
 415                        if ((ch = *quoted++) < '0' || '7' < ch)
 416                                goto error;
 417                                        ac |= (ch - '0');
 418                                        ch = ac;
 419                                        break;
 420                                default:
 421                        goto error;
 422                        }
 423                strbuf_addch(sb, ch);
 424                }
 425
 426  error:
 427        strbuf_setlen(sb, oldlen);
 428        return -1;
 429}
 430
 431/* quoting as a string literal for other languages */
 432
 433void perl_quote_buf(struct strbuf *sb, const char *src)
 434{
 435        const char sq = '\'';
 436        const char bq = '\\';
 437        char c;
 438
 439        strbuf_addch(sb, sq);
 440        while ((c = *src++)) {
 441                if (c == sq || c == bq)
 442                        strbuf_addch(sb, bq);
 443                strbuf_addch(sb, c);
 444        }
 445        strbuf_addch(sb, sq);
 446}
 447
 448void python_quote_buf(struct strbuf *sb, const char *src)
 449{
 450        const char sq = '\'';
 451        const char bq = '\\';
 452        const char nl = '\n';
 453        char c;
 454
 455        strbuf_addch(sb, sq);
 456        while ((c = *src++)) {
 457                if (c == nl) {
 458                        strbuf_addch(sb, bq);
 459                        strbuf_addch(sb, 'n');
 460                        continue;
 461                }
 462                if (c == sq || c == bq)
 463                        strbuf_addch(sb, bq);
 464                strbuf_addch(sb, c);
 465        }
 466        strbuf_addch(sb, sq);
 467}
 468
 469void tcl_quote_buf(struct strbuf *sb, const char *src)
 470{
 471        char c;
 472
 473        strbuf_addch(sb, '"');
 474        while ((c = *src++)) {
 475                switch (c) {
 476                case '[': case ']':
 477                case '{': case '}':
 478                case '$': case '\\': case '"':
 479                        strbuf_addch(sb, '\\');
 480                        /* fallthrough */
 481                default:
 482                        strbuf_addch(sb, c);
 483                        break;
 484                case '\f':
 485                        strbuf_addstr(sb, "\\f");
 486                        break;
 487                case '\r':
 488                        strbuf_addstr(sb, "\\r");
 489                        break;
 490                case '\n':
 491                        strbuf_addstr(sb, "\\n");
 492                        break;
 493                case '\t':
 494                        strbuf_addstr(sb, "\\t");
 495                        break;
 496                case '\v':
 497                        strbuf_addstr(sb, "\\v");
 498                        break;
 499                }
 500        }
 501        strbuf_addch(sb, '"');
 502}
 503
 504void basic_regex_quote_buf(struct strbuf *sb, const char *src)
 505{
 506        char c;
 507
 508        if (*src == '^') {
 509                /* only beginning '^' is special and needs quoting */
 510                strbuf_addch(sb, '\\');
 511                strbuf_addch(sb, *src++);
 512        }
 513        if (*src == '*')
 514                /* beginning '*' is not special, no quoting */
 515                strbuf_addch(sb, *src++);
 516
 517        while ((c = *src++)) {
 518                switch (c) {
 519                case '[':
 520                case '.':
 521                case '\\':
 522                case '*':
 523                        strbuf_addch(sb, '\\');
 524                        strbuf_addch(sb, c);
 525                        break;
 526
 527                case '$':
 528                        /* only the end '$' is special and needs quoting */
 529                        if (*src == '\0')
 530                                strbuf_addch(sb, '\\');
 531                        strbuf_addch(sb, c);
 532                        break;
 533
 534                default:
 535                        strbuf_addch(sb, c);
 536                        break;
 537                }
 538        }
 539}