convert.c

   1 #include "cache.h"
   2 #include "attr.h"
   3 #include "run-command.h"
   4
   5 /*
   6  * convert.c - convert a file when checking it out and checking it in.
   7  *
   8  * This should use the pathname to decide on whether it wants to do some
   9  * more interesting conversions (automatic gzip/unzip, general format
  10  * conversions etc etc), but by default it just does automatic CRLF<->LF
  11  * translation when the "auto_crlf" option is set.
  12  */
  13
  14 #define CRLF_GUESS      (-1)
  15 #define CRLF_BINARY     0
  16 #define CRLF_TEXT       1
  17 #define CRLF_INPUT      2
  18
  19 struct text_stat {
  20         /* NUL, CR, LF and CRLF counts */
  21         unsigned nul, cr, lf, crlf;
  22
  23         /* These are just approximations! */
  24         unsigned printable, nonprintable;
  25 };
  26
  27 static void gather_stats(const char *buf, unsigned long size, struct text_stat *stats)
  28 {
  29         unsigned long i;
  30
  31         memset(stats, 0, sizeof(*stats));
  32
  33         for (i = 0; i < size; i++) {
  34                 unsigned char c = buf[i];
  35                 if (c == '\r') {
  36                         stats->cr++;
  37                         if (i+1 < size && buf[i+1] == '\n')
  38                                 stats->crlf++;
  39                         continue;
  40                 }
  41                 if (c == '\n') {
  42                         stats->lf++;
  43                         continue;
  44                 }
  45                 if (c == 127)
  46                         /* DEL */
  47                         stats->nonprintable++;
  48                 else if (c < 32) {
  49                         switch (c) {
  50                                 /* BS, HT, ESC and FF */
  51                         case '\b': case '\t': case '\033': case '\014':
  52                                 stats->printable++;
  53                                 break;
  54                         case 0:
  55                                 stats->nul++;
  56                                 /* fall through */
  57                         default:
  58                                 stats->nonprintable++;
  59                         }
  60                 }
  61                 else
  62                         stats->printable++;
  63         }
  64
  65         /* If file ends with EOF then don't count this EOF as non-printable. */
  66         if (size >= 1 && buf[size-1] == '\032')
  67                 stats->nonprintable--;
  68 }
  69
  70 /*
  71  * The same heuristics as diff.c::mmfile_is_binary()
  72  */
  73 static int is_binary(unsigned long size, struct text_stat *stats)
  74 {
  75
  76         if (stats->nul)
  77                 return 1;
  78         if ((stats->printable >> 7) < stats->nonprintable)
  79                 return 1;
  80         /*
  81          * Other heuristics? Average line length might be relevant,
  82          * as might LF vs CR vs CRLF counts..
  83          *
  84          * NOTE! It might be normal to have a low ratio of CRLF to LF
  85          * (somebody starts with a LF-only file and edits it with an editor
  86          * that adds CRLF only to lines that are added..). But do  we
  87          * want to support CR-only? Probably not.
  88          */
  89         return 0;
  90 }
  91
  92 static void check_safe_crlf(const char *path, int action,
  93                             struct text_stat *stats, enum safe_crlf checksafe)
  94 {
  95         if (!checksafe)
  96                 return;
  97
  98         if (action == CRLF_INPUT || auto_crlf <= 0) {
  99                 /*
 100                  * CRLFs would not be restored by checkout:
 101                  * check if we'd remove CRLFs
 102                  */
 103                 if (stats->crlf) {
 104                         if (checksafe == SAFE_CRLF_WARN)
 105                                 warning("CRLF will be replaced by LF in %s.", path);
 106                         else /* i.e. SAFE_CRLF_FAIL */
 107                                 die("CRLF would be replaced by LF in %s.", path);
 108                 }
 109         } else if (auto_crlf > 0) {
 110                 /*
 111                  * CRLFs would be added by checkout:
 112                  * check if we have "naked" LFs
 113                  */
 114                 if (stats->lf != stats->crlf) {
 115                         if (checksafe == SAFE_CRLF_WARN)
 116                                 warning("LF will be replaced by CRLF in %s", path);
 117                         else /* i.e. SAFE_CRLF_FAIL */
 118                                 die("LF would be replaced by CRLF in %s", path);
 119                 }
 120         }
 121 }
 122
 123 static int crlf_to_git(const char *path, const char *src, size_t len,
 124                        struct strbuf *buf, int action, enum safe_crlf checksafe)
 125 {
 126         struct text_stat stats;
 127         char *dst;
 128
 129         if ((action == CRLF_BINARY) || !auto_crlf || !len)
 130                 return 0;
 131
 132         gather_stats(src, len, &stats);
 133
 134         if (action == CRLF_GUESS) {
 135                 /*
 136                  * We're currently not going to even try to convert stuff
 137                  * that has bare CR characters. Does anybody do that crazy
 138                  * stuff?
 139                  */
 140                 if (stats.cr != stats.crlf)
 141                         return 0;
 142
 143                 /*
 144                  * And add some heuristics for binary vs text, of course...
 145                  */
 146                 if (is_binary(len, &stats))
 147                         return 0;
 148         }
 149
 150         check_safe_crlf(path, action, &stats, checksafe);
 151
 152         /* Optimization: No CR? Nothing to convert, regardless. */
 153         if (!stats.cr)
 154                 return 0;
 155
 156         /* only grow if not in place */
 157         if (strbuf_avail(buf) + buf->len < len)
 158                 strbuf_grow(buf, len - buf->len);
 159         dst = buf->buf;
 160         if (action == CRLF_GUESS) {
 161                 /*
 162                  * If we guessed, we already know we rejected a file with
 163                  * lone CR, and we can strip a CR without looking at what
 164                  * follow it.
 165                  */
 166                 do {
 167                         unsigned char c = *src++;
 168                         if (c != '\r')
 169                                 *dst++ = c;
 170                 } while (--len);
 171         } else {
 172                 do {
 173                         unsigned char c = *src++;
 174                         if (! (c == '\r' && (1 < len && *src == '\n')))
 175                                 *dst++ = c;
 176                 } while (--len);
 177         }
 178         strbuf_setlen(buf, dst - buf->buf);
 179         return 1;
 180 }
 181
 182 static int crlf_to_worktree(const char *path, const char *src, size_t len,
 183                             struct strbuf *buf, int action)
 184 {
 185         char *to_free = NULL;
 186         struct text_stat stats;
 187
 188         if ((action == CRLF_BINARY) || (action == CRLF_INPUT) ||
 189             auto_crlf <= 0)
 190                 return 0;
 191
 192         if (!len)
 193                 return 0;
 194
 195         gather_stats(src, len, &stats);
 196
 197         /* No LF? Nothing to convert, regardless. */
 198         if (!stats.lf)
 199                 return 0;
 200
 201         /* Was it already in CRLF format? */
 202         if (stats.lf == stats.crlf)
 203                 return 0;
 204
 205         if (action == CRLF_GUESS) {
 206                 /* If we have any bare CR characters, we're not going to touch it */
 207                 if (stats.cr != stats.crlf)
 208                         return 0;
 209
 210                 if (is_binary(len, &stats))
 211                         return 0;
 212         }
 213
 214         /* are we "faking" in place editing ? */
 215         if (src == buf->buf)
 216                 to_free = strbuf_detach(buf, NULL);
 217
 218         strbuf_grow(buf, len + stats.lf - stats.crlf);
 219         for (;;) {
 220                 const char *nl = memchr(src, '\n', len);
 221                 if (!nl)
 222                         break;
 223                 if (nl > src && nl[-1] == '\r') {
 224                         strbuf_add(buf, src, nl + 1 - src);
 225                 } else {
 226                         strbuf_add(buf, src, nl - src);
 227                         strbuf_addstr(buf, "\r\n");
 228                 }
 229                 len -= nl + 1 - src;
 230                 src  = nl + 1;
 231         }
 232         strbuf_add(buf, src, len);
 233
 234         free(to_free);
 235         return 1;
 236 }
 237
 238 struct filter_params {
 239         const char *src;
 240         unsigned long size;
 241         const char *cmd;
 242 };
 243
 244 static int filter_buffer(int fd, void *data)
 245 {
 246         /*
 247          * Spawn cmd and feed the buffer contents through its stdin.
 248          */
 249         struct child_process child_process;
 250         struct filter_params *params = (struct filter_params *)data;
 251         int write_err, status;
 252         const char *argv[] = { "sh", "-c", params->cmd, NULL };
 253
 254         memset(&child_process, 0, sizeof(child_process));
 255         child_process.argv = argv;
 256         child_process.in = -1;
 257         child_process.out = fd;
 258
 259         if (start_command(&child_process))
 260                 return error("cannot fork to run external filter %s", params->cmd);
 261
 262         write_err = (write_in_full(child_process.in, params->src, params->size) < 0);
 263         if (close(child_process.in))
 264                 write_err = 1;
 265         if (write_err)
 266                 error("cannot feed the input to external filter %s", params->cmd);
 267
 268         status = finish_command(&child_process);
 269         if (status)
 270                 error("external filter %s failed %d", params->cmd, status);
 271         return (write_err || status);
 272 }
 273
 274 static int apply_filter(const char *path, const char *src, size_t len,
 275                         struct strbuf *dst, const char *cmd)
 276 {
 277         /*
 278          * Create a pipeline to have the command filter the buffer's
 279          * contents.
 280          *
 281          * (child --> cmd) --> us
 282          */
 283         int ret = 1;
 284         struct strbuf nbuf = STRBUF_INIT;
 285         struct async async;
 286         struct filter_params params;
 287
 288         if (!cmd)
 289                 return 0;
 290
 291         memset(&async, 0, sizeof(async));
 292         async.proc = filter_buffer;
 293         async.data = &params;
 294         params.src = src;
 295         params.size = len;
 296         params.cmd = cmd;
 297
 298         fflush(NULL);
 299         if (start_async(&async))
 300                 return 0;       /* error was already reported */
 301
 302         if (strbuf_read(&nbuf, async.out, len) < 0) {
 303                 error("read from external filter %s failed", cmd);
 304                 ret = 0;
 305         }
 306         if (close(async.out)) {
 307                 error("read from external filter %s failed", cmd);
 308                 ret = 0;
 309         }
 310         if (finish_async(&async)) {
 311                 error("external filter %s failed", cmd);
 312                 ret = 0;
 313         }
 314
 315         if (ret) {
 316                 strbuf_swap(dst, &nbuf);
 317         }
 318         strbuf_release(&nbuf);
 319         return ret;
 320 }
 321
 322 static struct convert_driver {
 323         const char *name;
 324         struct convert_driver *next;
 325         const char *smudge;
 326         const char *clean;
 327 } *user_convert, **user_convert_tail;
 328
 329 static int read_convert_config(const char *var, const char *value, void *cb)
 330 {
 331         const char *ep, *name;
 332         int namelen;
 333         struct convert_driver *drv;
 334
 335         /*
 336          * External conversion drivers are configured using
 337          * "filter.<name>.variable".
 338          */
 339         if (prefixcmp(var, "filter.") || (ep = strrchr(var, '.')) == var + 6)
 340                 return 0;
 341         name = var + 7;
 342         namelen = ep - name;
 343         for (drv = user_convert; drv; drv = drv->next)
 344                 if (!strncmp(drv->name, name, namelen) && !drv->name[namelen])
 345                         break;
 346         if (!drv) {
 347                 drv = xcalloc(1, sizeof(struct convert_driver));
 348                 drv->name = xmemdupz(name, namelen);
 349                 *user_convert_tail = drv;
 350                 user_convert_tail = &(drv->next);
 351         }
 352
 353         ep++;
 354
 355         /*
 356          * filter.<name>.smudge and filter.<name>.clean specifies
 357          * the command line:
 358          *
 359          *      command-line
 360          *
 361          * The command-line will not be interpolated in any way.
 362          */
 363
 364         if (!strcmp("smudge", ep))
 365                 return git_config_string(&drv->smudge, var, value);
 366
 367         if (!strcmp("clean", ep))
 368                 return git_config_string(&drv->clean, var, value);
 369
 370         return 0;
 371 }
 372
 373 static void setup_convert_check(struct git_attr_check *check)
 374 {
 375         static struct git_attr *attr_crlf;
 376         static struct git_attr *attr_ident;
 377         static struct git_attr *attr_filter;
 378
 379         if (!attr_crlf) {
 380                 attr_crlf = git_attr("crlf", 4);
 381                 attr_ident = git_attr("ident", 5);
 382                 attr_filter = git_attr("filter", 6);
 383                 user_convert_tail = &user_convert;
 384                 git_config(read_convert_config, NULL);
 385         }
 386         check[0].attr = attr_crlf;
 387         check[1].attr = attr_ident;
 388         check[2].attr = attr_filter;
 389 }
 390
 391 static int count_ident(const char *cp, unsigned long size)
 392 {
 393         /*
 394          * "$Id: 0000000000000000000000000000000000000000 $" <=> "$Id$"
 395          */
 396         int cnt = 0;
 397         char ch;
 398
 399         while (size) {
 400                 ch = *cp++;
 401                 size--;
 402                 if (ch != '$')
 403                         continue;
 404                 if (size < 3)
 405                         break;
 406                 if (memcmp("Id", cp, 2))
 407                         continue;
 408                 ch = cp[2];
 409                 cp += 3;
 410                 size -= 3;
 411                 if (ch == '$')
 412                         cnt++; /* $Id$ */
 413                 if (ch != ':')
 414                         continue;
 415
 416                 /*
 417                  * "$Id: ... "; scan up to the closing dollar sign and discard.
 418                  */
 419                 while (size) {
 420                         ch = *cp++;
 421                         size--;
 422                         if (ch == '$') {
 423                                 cnt++;
 424                                 break;
 425                         }
 426                 }
 427         }
 428         return cnt;
 429 }
 430
 431 static int ident_to_git(const char *path, const char *src, size_t len,
 432                         struct strbuf *buf, int ident)
 433 {
 434         char *dst, *dollar;
 435
 436         if (!ident || !count_ident(src, len))
 437                 return 0;
 438
 439         /* only grow if not in place */
 440         if (strbuf_avail(buf) + buf->len < len)
 441                 strbuf_grow(buf, len - buf->len);
 442         dst = buf->buf;
 443         for (;;) {
 444                 dollar = memchr(src, '$', len);
 445                 if (!dollar)
 446                         break;
 447                 memcpy(dst, src, dollar + 1 - src);
 448                 dst += dollar + 1 - src;
 449                 len -= dollar + 1 - src;
 450                 src  = dollar + 1;
 451
 452                 if (len > 3 && !memcmp(src, "Id:", 3)) {
 453                         dollar = memchr(src + 3, '$', len - 3);
 454                         if (!dollar)
 455                                 break;
 456                         memcpy(dst, "Id$", 3);
 457                         dst += 3;
 458                         len -= dollar + 1 - src;
 459                         src  = dollar + 1;
 460                 }
 461         }
 462         memcpy(dst, src, len);
 463         strbuf_setlen(buf, dst + len - buf->buf);
 464         return 1;
 465 }
 466
 467 static int ident_to_worktree(const char *path, const char *src, size_t len,
 468                              struct strbuf *buf, int ident)
 469 {
 470         unsigned char sha1[20];
 471         char *to_free = NULL, *dollar;
 472         int cnt;
 473
 474         if (!ident)
 475                 return 0;
 476
 477         cnt = count_ident(src, len);
 478         if (!cnt)
 479                 return 0;
 480
 481         /* are we "faking" in place editing ? */
 482         if (src == buf->buf)
 483                 to_free = strbuf_detach(buf, NULL);
 484         hash_sha1_file(src, len, "blob", sha1);
 485
 486         strbuf_grow(buf, len + cnt * 43);
 487         for (;;) {
 488                 /* step 1: run to the next '$' */
 489                 dollar = memchr(src, '$', len);
 490                 if (!dollar)
 491                         break;
 492                 strbuf_add(buf, src, dollar + 1 - src);
 493                 len -= dollar + 1 - src;
 494                 src  = dollar + 1;
 495
 496                 /* step 2: does it looks like a bit like Id:xxx$ or Id$ ? */
 497                 if (len < 3 || memcmp("Id", src, 2))
 498                         continue;
 499
 500                 /* step 3: skip over Id$ or Id:xxxxx$ */
 501                 if (src[2] == '$') {
 502                         src += 3;
 503                         len -= 3;
 504                 } else if (src[2] == ':') {
 505                         /*
 506                          * It's possible that an expanded Id has crept its way into the
 507                          * repository, we cope with that by stripping the expansion out
 508                          */
 509                         dollar = memchr(src + 3, '$', len - 3);
 510                         if (!dollar) {
 511                                 /* incomplete keyword, no more '$', so just quit the loop */
 512                                 break;
 513                         }
 514
 515                         len -= dollar + 1 - src;
 516                         src  = dollar + 1;
 517                 } else {
 518                         /* it wasn't a "Id$" or "Id:xxxx$" */
 519                         continue;
 520                 }
 521
 522                 /* step 4: substitute */
 523                 strbuf_addstr(buf, "Id: ");
 524                 strbuf_add(buf, sha1_to_hex(sha1), 40);
 525                 strbuf_addstr(buf, " $");
 526         }
 527         strbuf_add(buf, src, len);
 528
 529         free(to_free);
 530         return 1;
 531 }
 532
 533 static int git_path_check_crlf(const char *path, struct git_attr_check *check)
 534 {
 535         const char *value = check->value;
 536
 537         if (ATTR_TRUE(value))
 538                 return CRLF_TEXT;
 539         else if (ATTR_FALSE(value))
 540                 return CRLF_BINARY;
 541         else if (ATTR_UNSET(value))
 542                 ;
 543         else if (!strcmp(value, "input"))
 544                 return CRLF_INPUT;
 545         return CRLF_GUESS;
 546 }
 547
 548 static struct convert_driver *git_path_check_convert(const char *path,
 549                                              struct git_attr_check *check)
 550 {
 551         const char *value = check->value;
 552         struct convert_driver *drv;
 553
 554         if (ATTR_TRUE(value) || ATTR_FALSE(value) || ATTR_UNSET(value))
 555                 return NULL;
 556         for (drv = user_convert; drv; drv = drv->next)
 557                 if (!strcmp(value, drv->name))
 558                         return drv;
 559         return NULL;
 560 }
 561
 562 static int git_path_check_ident(const char *path, struct git_attr_check *check)
 563 {
 564         const char *value = check->value;
 565
 566         return !!ATTR_TRUE(value);
 567 }
 568
 569 int convert_to_git(const char *path, const char *src, size_t len,
 570                    struct strbuf *dst, enum safe_crlf checksafe)
 571 {
 572         struct git_attr_check check[3];
 573         int crlf = CRLF_GUESS;
 574         int ident = 0, ret = 0;
 575         const char *filter = NULL;
 576
 577         setup_convert_check(check);
 578         if (!git_checkattr(path, ARRAY_SIZE(check), check)) {
 579                 struct convert_driver *drv;
 580                 crlf = git_path_check_crlf(path, check + 0);
 581                 ident = git_path_check_ident(path, check + 1);
 582                 drv = git_path_check_convert(path, check + 2);
 583                 if (drv && drv->clean)
 584                         filter = drv->clean;
 585         }
 586
 587         ret |= apply_filter(path, src, len, dst, filter);
 588         if (ret) {
 589                 src = dst->buf;
 590                 len = dst->len;
 591         }
 592         ret |= crlf_to_git(path, src, len, dst, crlf, checksafe);
 593         if (ret) {
 594                 src = dst->buf;
 595                 len = dst->len;
 596         }
 597         return ret | ident_to_git(path, src, len, dst, ident);
 598 }
 599
 600 int convert_to_working_tree(const char *path, const char *src, size_t len, struct strbuf *dst)
 601 {
 602         struct git_attr_check check[3];
 603         int crlf = CRLF_GUESS;
 604         int ident = 0, ret = 0;
 605         const char *filter = NULL;
 606
 607         setup_convert_check(check);
 608         if (!git_checkattr(path, ARRAY_SIZE(check), check)) {
 609                 struct convert_driver *drv;
 610                 crlf = git_path_check_crlf(path, check + 0);
 611                 ident = git_path_check_ident(path, check + 1);
 612                 drv = git_path_check_convert(path, check + 2);
 613                 if (drv && drv->smudge)
 614                         filter = drv->smudge;
 615         }
 616
 617         ret |= ident_to_worktree(path, src, len, dst, ident);
 618         if (ret) {
 619                 src = dst->buf;
 620                 len = dst->len;
 621         }
 622         ret |= crlf_to_worktree(path, src, len, dst, crlf);
 623         if (ret) {
 624                 src = dst->buf;
 625                 len = dst->len;
 626         }
 627         return ret | apply_filter(path, src, len, dst, filter);
 628 }