quote.con commit Merge branch 'jj/log-doc' (a2cb44c)
   1#include "cache.h"
   2#include "quote.h"
   3#include "argv-array.h"
   4
   5int quote_path_fully = 1;
   6
   7/* Help to copy the thing properly quoted for the shell safety.
   8 * any single quote is replaced with '\'', any exclamation point
   9 * is replaced with '\!', and the whole thing is enclosed in a
  10 *
  11 * E.g.
  12 *  original     sq_quote     result
  13 *  name     ==> name      ==> 'name'
  14 *  a b      ==> a b       ==> 'a b'
  15 *  a'b      ==> a'\''b    ==> 'a'\''b'
  16 *  a!b      ==> a'\!'b    ==> 'a'\!'b'
  17 */
  18static inline int need_bs_quote(char c)
  19{
  20        return (c == '\'' || c == '!');
  21}
  22
  23void sq_quote_buf(struct strbuf *dst, const char *src)
  24{
  25        char *to_free = NULL;
  26
  27        if (dst->buf == src)
  28                to_free = strbuf_detach(dst, NULL);
  29
  30        strbuf_addch(dst, '\'');
  31        while (*src) {
  32                size_t len = strcspn(src, "'!");
  33                strbuf_add(dst, src, len);
  34                src += len;
  35                while (need_bs_quote(*src)) {
  36                        strbuf_addstr(dst, "'\\");
  37                        strbuf_addch(dst, *src++);
  38                        strbuf_addch(dst, '\'');
  39                }
  40        }
  41        strbuf_addch(dst, '\'');
  42        free(to_free);
  43}
  44
  45void sq_quote_argv(struct strbuf *dst, const char** argv, size_t maxlen)
  46{
  47        int i;
  48
  49        /* Copy into destination buffer. */
  50        strbuf_grow(dst, 255);
  51        for (i = 0; argv[i]; ++i) {
  52                strbuf_addch(dst, ' ');
  53                sq_quote_buf(dst, argv[i]);
  54                if (maxlen && dst->len > maxlen)
  55                        die("Too many or long arguments");
  56        }
  57}
  58
  59static char *sq_dequote_step(char *arg, char **next)
  60{
  61        char *dst = arg;
  62        char *src = arg;
  63        char c;
  64
  65        if (*src != '\'')
  66                return NULL;
  67        for (;;) {
  68                c = *++src;
  69                if (!c)
  70                        return NULL;
  71                if (c != '\'') {
  72                        *dst++ = c;
  73                        continue;
  74                }
  75                /* We stepped out of sq */
  76                switch (*++src) {
  77                case '\0':
  78                        *dst = 0;
  79                        if (next)
  80                                *next = NULL;
  81                        return arg;
  82                case '\\':
  83                        c = *++src;
  84                        if (need_bs_quote(c) && *++src == '\'') {
  85                                *dst++ = c;
  86                                continue;
  87                        }
  88                /* Fallthrough */
  89                default:
  90                        if (!next || !isspace(*src))
  91                                return NULL;
  92                        do {
  93                                c = *++src;
  94                        } while (isspace(c));
  95                        *dst = 0;
  96                        *next = src;
  97                        return arg;
  98                }
  99        }
 100}
 101
 102char *sq_dequote(char *arg)
 103{
 104        return sq_dequote_step(arg, NULL);
 105}
 106
 107static int sq_dequote_to_argv_internal(char *arg,
 108                                       const char ***argv, int *nr, int *alloc,
 109                                       struct argv_array *array)
 110{
 111        char *next = arg;
 112
 113        if (!*arg)
 114                return 0;
 115        do {
 116                char *dequoted = sq_dequote_step(next, &next);
 117                if (!dequoted)
 118                        return -1;
 119                if (argv) {
 120                        ALLOC_GROW(*argv, *nr + 1, *alloc);
 121                        (*argv)[(*nr)++] = dequoted;
 122                }
 123                if (array)
 124                        argv_array_push(array, dequoted);
 125        } while (next);
 126
 127        return 0;
 128}
 129
 130int sq_dequote_to_argv(char *arg, const char ***argv, int *nr, int *alloc)
 131{
 132        return sq_dequote_to_argv_internal(arg, argv, nr, alloc, NULL);
 133}
 134
 135int sq_dequote_to_argv_array(char *arg, struct argv_array *array)
 136{
 137        return sq_dequote_to_argv_internal(arg, NULL, NULL, NULL, array);
 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 || (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
 246size_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
 251void 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
 267void 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
 277void 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
 298void write_name_quoted_relative(const char *name, const char *prefix,
 299                                FILE *fp, int terminator)
 300{
 301        struct strbuf sb = STRBUF_INIT;
 302
 303        name = relative_path(name, prefix, &sb);
 304        write_name_quoted(name, fp, terminator);
 305
 306        strbuf_release(&sb);
 307}
 308
 309/* quote path as relative to the given prefix */
 310char *quote_path_relative(const char *in, const char *prefix,
 311                          struct strbuf *out)
 312{
 313        struct strbuf sb = STRBUF_INIT;
 314        const char *rel = relative_path(in, prefix, &sb);
 315        strbuf_reset(out);
 316        quote_c_style_counted(rel, strlen(rel), out, NULL, 0);
 317        strbuf_release(&sb);
 318
 319        return out->buf;
 320}
 321
 322/*
 323 * C-style name unquoting.
 324 *
 325 * Quoted should point at the opening double quote.
 326 * + Returns 0 if it was able to unquote the string properly, and appends the
 327 *   result in the strbuf `sb'.
 328 * + Returns -1 in case of error, and doesn't touch the strbuf. Though note
 329 *   that this function will allocate memory in the strbuf, so calling
 330 *   strbuf_release is mandatory whichever result unquote_c_style returns.
 331 *
 332 * Updates endp pointer to point at one past the ending double quote if given.
 333 */
 334int unquote_c_style(struct strbuf *sb, const char *quoted, const char **endp)
 335{
 336        size_t oldlen = sb->len, len;
 337        int ch, ac;
 338
 339        if (*quoted++ != '"')
 340                return -1;
 341
 342        for (;;) {
 343                len = strcspn(quoted, "\"\\");
 344                strbuf_add(sb, quoted, len);
 345                quoted += len;
 346
 347                switch (*quoted++) {
 348                  case '"':
 349                        if (endp)
 350                                *endp = quoted;
 351                        return 0;
 352                  case '\\':
 353                        break;
 354                  default:
 355                        goto error;
 356                }
 357
 358                switch ((ch = *quoted++)) {
 359                case 'a': ch = '\a'; break;
 360                case 'b': ch = '\b'; break;
 361                case 'f': ch = '\f'; break;
 362                case 'n': ch = '\n'; break;
 363                case 'r': ch = '\r'; break;
 364                case 't': ch = '\t'; break;
 365                case 'v': ch = '\v'; break;
 366
 367                case '\\': case '"':
 368                        break; /* verbatim */
 369
 370                /* octal values with first digit over 4 overflow */
 371                case '0': case '1': case '2': case '3':
 372                                        ac = ((ch - '0') << 6);
 373                        if ((ch = *quoted++) < '0' || '7' < ch)
 374                                goto error;
 375                                        ac |= ((ch - '0') << 3);
 376                        if ((ch = *quoted++) < '0' || '7' < ch)
 377                                goto error;
 378                                        ac |= (ch - '0');
 379                                        ch = ac;
 380                                        break;
 381                                default:
 382                        goto error;
 383                        }
 384                strbuf_addch(sb, ch);
 385                }
 386
 387  error:
 388        strbuf_setlen(sb, oldlen);
 389        return -1;
 390}
 391
 392/* quoting as a string literal for other languages */
 393
 394void perl_quote_buf(struct strbuf *sb, const char *src)
 395{
 396        const char sq = '\'';
 397        const char bq = '\\';
 398        char c;
 399
 400        strbuf_addch(sb, sq);
 401        while ((c = *src++)) {
 402                if (c == sq || c == bq)
 403                        strbuf_addch(sb, bq);
 404                strbuf_addch(sb, c);
 405        }
 406        strbuf_addch(sb, sq);
 407}
 408
 409void python_quote_buf(struct strbuf *sb, const char *src)
 410{
 411        const char sq = '\'';
 412        const char bq = '\\';
 413        const char nl = '\n';
 414        char c;
 415
 416        strbuf_addch(sb, sq);
 417        while ((c = *src++)) {
 418                if (c == nl) {
 419                        strbuf_addch(sb, bq);
 420                        strbuf_addch(sb, 'n');
 421                        continue;
 422                }
 423                if (c == sq || c == bq)
 424                        strbuf_addch(sb, bq);
 425                strbuf_addch(sb, c);
 426        }
 427        strbuf_addch(sb, sq);
 428}
 429
 430void tcl_quote_buf(struct strbuf *sb, const char *src)
 431{
 432        char c;
 433
 434        strbuf_addch(sb, '"');
 435        while ((c = *src++)) {
 436                switch (c) {
 437                case '[': case ']':
 438                case '{': case '}':
 439                case '$': case '\\': case '"':
 440                        strbuf_addch(sb, '\\');
 441                default:
 442                        strbuf_addch(sb, c);
 443                        break;
 444                case '\f':
 445                        strbuf_addstr(sb, "\\f");
 446                        break;
 447                case '\r':
 448                        strbuf_addstr(sb, "\\r");
 449                        break;
 450                case '\n':
 451                        strbuf_addstr(sb, "\\n");
 452                        break;
 453                case '\t':
 454                        strbuf_addstr(sb, "\\t");
 455                        break;
 456                case '\v':
 457                        strbuf_addstr(sb, "\\v");
 458                        break;
 459                }
 460        }
 461        strbuf_addch(sb, '"');
 462}