quote.con commit fmt-merge-msg: refactor merge title formatting (403994e)
   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 || (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
 298/* quote path as relative to the given prefix */
 299char *quote_path_relative(const char *in, int len,
 300                          struct strbuf *out, const char *prefix)
 301{
 302        int needquote;
 303
 304        if (len < 0)
 305                len = strlen(in);
 306
 307        /* "../" prefix itself does not need quoting, but "in" might. */
 308        needquote = next_quote_pos(in, len) < len;
 309        strbuf_setlen(out, 0);
 310        strbuf_grow(out, len);
 311
 312        if (needquote)
 313                strbuf_addch(out, '"');
 314        if (prefix) {
 315                int off = 0;
 316                while (prefix[off] && off < len && prefix[off] == in[off])
 317                        if (prefix[off] == '/') {
 318                                prefix += off + 1;
 319                                in += off + 1;
 320                                len -= off + 1;
 321                                off = 0;
 322                        } else
 323                                off++;
 324
 325                for (; *prefix; prefix++)
 326                        if (*prefix == '/')
 327                                strbuf_addstr(out, "../");
 328        }
 329
 330        quote_c_style_counted (in, len, out, NULL, 1);
 331
 332        if (needquote)
 333                strbuf_addch(out, '"');
 334        if (!out->len)
 335                strbuf_addstr(out, "./");
 336
 337        return out->buf;
 338}
 339
 340/*
 341 * C-style name unquoting.
 342 *
 343 * Quoted should point at the opening double quote.
 344 * + Returns 0 if it was able to unquote the string properly, and appends the
 345 *   result in the strbuf `sb'.
 346 * + Returns -1 in case of error, and doesn't touch the strbuf. Though note
 347 *   that this function will allocate memory in the strbuf, so calling
 348 *   strbuf_release is mandatory whichever result unquote_c_style returns.
 349 *
 350 * Updates endp pointer to point at one past the ending double quote if given.
 351 */
 352int unquote_c_style(struct strbuf *sb, const char *quoted, const char **endp)
 353{
 354        size_t oldlen = sb->len, len;
 355        int ch, ac;
 356
 357        if (*quoted++ != '"')
 358                return -1;
 359
 360        for (;;) {
 361                len = strcspn(quoted, "\"\\");
 362                strbuf_add(sb, quoted, len);
 363                quoted += len;
 364
 365                switch (*quoted++) {
 366                  case '"':
 367                        if (endp)
 368                                *endp = quoted;
 369                        return 0;
 370                  case '\\':
 371                        break;
 372                  default:
 373                        goto error;
 374                }
 375
 376                switch ((ch = *quoted++)) {
 377                case 'a': ch = '\a'; break;
 378                case 'b': ch = '\b'; break;
 379                case 'f': ch = '\f'; break;
 380                case 'n': ch = '\n'; break;
 381                case 'r': ch = '\r'; break;
 382                case 't': ch = '\t'; break;
 383                case 'v': ch = '\v'; break;
 384
 385                case '\\': case '"':
 386                        break; /* verbatim */
 387
 388                /* octal values with first digit over 4 overflow */
 389                case '0': case '1': case '2': case '3':
 390                                        ac = ((ch - '0') << 6);
 391                        if ((ch = *quoted++) < '0' || '7' < ch)
 392                                goto error;
 393                                        ac |= ((ch - '0') << 3);
 394                        if ((ch = *quoted++) < '0' || '7' < ch)
 395                                goto error;
 396                                        ac |= (ch - '0');
 397                                        ch = ac;
 398                                        break;
 399                                default:
 400                        goto error;
 401                        }
 402                strbuf_addch(sb, ch);
 403                }
 404
 405  error:
 406        strbuf_setlen(sb, oldlen);
 407        return -1;
 408}
 409
 410/* quoting as a string literal for other languages */
 411
 412void perl_quote_print(FILE *stream, const char *src)
 413{
 414        const char sq = '\'';
 415        const char bq = '\\';
 416        char c;
 417
 418        fputc(sq, stream);
 419        while ((c = *src++)) {
 420                if (c == sq || c == bq)
 421                        fputc(bq, stream);
 422                fputc(c, stream);
 423        }
 424        fputc(sq, stream);
 425}
 426
 427void python_quote_print(FILE *stream, const char *src)
 428{
 429        const char sq = '\'';
 430        const char bq = '\\';
 431        const char nl = '\n';
 432        char c;
 433
 434        fputc(sq, stream);
 435        while ((c = *src++)) {
 436                if (c == nl) {
 437                        fputc(bq, stream);
 438                        fputc('n', stream);
 439                        continue;
 440                }
 441                if (c == sq || c == bq)
 442                        fputc(bq, stream);
 443                fputc(c, stream);
 444        }
 445        fputc(sq, stream);
 446}
 447
 448void tcl_quote_print(FILE *stream, const char *src)
 449{
 450        char c;
 451
 452        fputc('"', stream);
 453        while ((c = *src++)) {
 454                switch (c) {
 455                case '[': case ']':
 456                case '{': case '}':
 457                case '$': case '\\': case '"':
 458                        fputc('\\', stream);
 459                default:
 460                        fputc(c, stream);
 461                        break;
 462                case '\f':
 463                        fputs("\\f", stream);
 464                        break;
 465                case '\r':
 466                        fputs("\\r", stream);
 467                        break;
 468                case '\n':
 469                        fputs("\\n", stream);
 470                        break;
 471                case '\t':
 472                        fputs("\\t", stream);
 473                        break;
 474                case '\v':
 475                        fputs("\\v", stream);
 476                        break;
 477                }
 478        }
 479        fputc('"', stream);
 480}