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;
 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         strbuf_init(&nbuf, 0);
 303         if (strbuf_read(&nbuf, async.out, len) < 0) {
 304                 error("read from external filter %s failed", cmd);
 305                 ret = 0;
 306         }
 307         if (close(async.out)) {
 308                 error("read from external filter %s failed", cmd);
 309                 ret = 0;
 310         }
 311         if (finish_async(&async)) {
 312                 error("external filter %s failed", cmd);
 313                 ret = 0;
 314         }
 315
 316         if (ret) {
 317                 strbuf_swap(dst, &nbuf);
 318         }
 319         strbuf_release(&nbuf);
 320         return ret;
 321 }
 322
 323 static struct convert_driver {
 324         const char *name;
 325         struct convert_driver *next;
 326         char *smudge;
 327         char *clean;
 328 } *user_convert, **user_convert_tail;
 329
 330 static int read_convert_config(const char *var, const char *value)
 331 {
 332         const char *ep, *name;
 333         int namelen;
 334         struct convert_driver *drv;
 335
 336         /*
 337          * External conversion drivers are configured using
 338          * "filter.<name>.variable".
 339          */
 340         if (prefixcmp(var, "filter.") || (ep = strrchr(var, '.')) == var + 6)
 341                 return 0;
 342         name = var + 7;
 343         namelen = ep - name;
 344         for (drv = user_convert; drv; drv = drv->next)
 345                 if (!strncmp(drv->name, name, namelen) && !drv->name[namelen])
 346                         break;
 347         if (!drv) {
 348                 drv = xcalloc(1, sizeof(struct convert_driver));
 349                 drv->name = xmemdupz(name, namelen);
 350                 *user_convert_tail = drv;
 351                 user_convert_tail = &(drv->next);
 352         }
 353
 354         ep++;
 355
 356         /*
 357          * filter.<name>.smudge and filter.<name>.clean specifies
 358          * the command line:
 359          *
 360          *      command-line
 361          *
 362          * The command-line will not be interpolated in any way.
 363          */
 364
 365         if (!strcmp("smudge", ep)) {
 366                 if (!value)
 367                         return config_error_nonbool(var);
 368                 drv->smudge = strdup(value);
 369                 return 0;
 370         }
 371
 372         if (!strcmp("clean", ep)) {
 373                 if (!value)
 374                         return config_error_nonbool(var);
 375                 drv->clean = strdup(value);
 376                 return 0;
 377         }
 378         return 0;
 379 }
 380
 381 static void setup_convert_check(struct git_attr_check *check)
 382 {
 383         static struct git_attr *attr_crlf;
 384         static struct git_attr *attr_ident;
 385         static struct git_attr *attr_filter;
 386
 387         if (!attr_crlf) {
 388                 attr_crlf = git_attr("crlf", 4);
 389                 attr_ident = git_attr("ident", 5);
 390                 attr_filter = git_attr("filter", 6);
 391                 user_convert_tail = &user_convert;
 392                 git_config(read_convert_config);
 393         }
 394         check[0].attr = attr_crlf;
 395         check[1].attr = attr_ident;
 396         check[2].attr = attr_filter;
 397 }
 398
 399 static int count_ident(const char *cp, unsigned long size)
 400 {
 401         /*
 402          * "$Id: 0000000000000000000000000000000000000000 $" <=> "$Id$"
 403          */
 404         int cnt = 0;
 405         char ch;
 406
 407         while (size) {
 408                 ch = *cp++;
 409                 size--;
 410                 if (ch != '$')
 411                         continue;
 412                 if (size < 3)
 413                         break;
 414                 if (memcmp("Id", cp, 2))
 415                         continue;
 416                 ch = cp[2];
 417                 cp += 3;
 418                 size -= 3;
 419                 if (ch == '$')
 420                         cnt++; /* $Id$ */
 421                 if (ch != ':')
 422                         continue;
 423
 424                 /*
 425                  * "$Id: ... "; scan up to the closing dollar sign and discard.
 426                  */
 427                 while (size) {
 428                         ch = *cp++;
 429                         size--;
 430                         if (ch == '$') {
 431                                 cnt++;
 432                                 break;
 433                         }
 434                 }
 435         }
 436         return cnt;
 437 }
 438
 439 static int ident_to_git(const char *path, const char *src, size_t len,
 440                         struct strbuf *buf, int ident)
 441 {
 442         char *dst, *dollar;
 443
 444         if (!ident || !count_ident(src, len))
 445                 return 0;
 446
 447         /* only grow if not in place */
 448         if (strbuf_avail(buf) + buf->len < len)
 449                 strbuf_grow(buf, len - buf->len);
 450         dst = buf->buf;
 451         for (;;) {
 452                 dollar = memchr(src, '$', len);
 453                 if (!dollar)
 454                         break;
 455                 memcpy(dst, src, dollar + 1 - src);
 456                 dst += dollar + 1 - src;
 457                 len -= dollar + 1 - src;
 458                 src  = dollar + 1;
 459
 460                 if (len > 3 && !memcmp(src, "Id:", 3)) {
 461                         dollar = memchr(src + 3, '$', len - 3);
 462                         if (!dollar)
 463                                 break;
 464                         memcpy(dst, "Id$", 3);
 465                         dst += 3;
 466                         len -= dollar + 1 - src;
 467                         src  = dollar + 1;
 468                 }
 469         }
 470         memcpy(dst, src, len);
 471         strbuf_setlen(buf, dst + len - buf->buf);
 472         return 1;
 473 }
 474
 475 static int ident_to_worktree(const char *path, const char *src, size_t len,
 476                              struct strbuf *buf, int ident)
 477 {
 478         unsigned char sha1[20];
 479         char *to_free = NULL, *dollar;
 480         int cnt;
 481
 482         if (!ident)
 483                 return 0;
 484
 485         cnt = count_ident(src, len);
 486         if (!cnt)
 487                 return 0;
 488
 489         /* are we "faking" in place editing ? */
 490         if (src == buf->buf)
 491                 to_free = strbuf_detach(buf, NULL);
 492         hash_sha1_file(src, len, "blob", sha1);
 493
 494         strbuf_grow(buf, len + cnt * 43);
 495         for (;;) {
 496                 /* step 1: run to the next '$' */
 497                 dollar = memchr(src, '$', len);
 498                 if (!dollar)
 499                         break;
 500                 strbuf_add(buf, src, dollar + 1 - src);
 501                 len -= dollar + 1 - src;
 502                 src  = dollar + 1;
 503
 504                 /* step 2: does it looks like a bit like Id:xxx$ or Id$ ? */
 505                 if (len < 3 || memcmp("Id", src, 2))
 506                         continue;
 507
 508                 /* step 3: skip over Id$ or Id:xxxxx$ */
 509                 if (src[2] == '$') {
 510                         src += 3;
 511                         len -= 3;
 512                 } else if (src[2] == ':') {
 513                         /*
 514                          * It's possible that an expanded Id has crept its way into the
 515                          * repository, we cope with that by stripping the expansion out
 516                          */
 517                         dollar = memchr(src + 3, '$', len - 3);
 518                         if (!dollar) {
 519                                 /* incomplete keyword, no more '$', so just quit the loop */
 520                                 break;
 521                         }
 522
 523                         len -= dollar + 1 - src;
 524                         src  = dollar + 1;
 525                 } else {
 526                         /* it wasn't a "Id$" or "Id:xxxx$" */
 527                         continue;
 528                 }
 529
 530                 /* step 4: substitute */
 531                 strbuf_addstr(buf, "Id: ");
 532                 strbuf_add(buf, sha1_to_hex(sha1), 40);
 533                 strbuf_addstr(buf, " $");
 534         }
 535         strbuf_add(buf, src, len);
 536
 537         free(to_free);
 538         return 1;
 539 }
 540
 541 static int git_path_check_crlf(const char *path, struct git_attr_check *check)
 542 {
 543         const char *value = check->value;
 544
 545         if (ATTR_TRUE(value))
 546                 return CRLF_TEXT;
 547         else if (ATTR_FALSE(value))
 548                 return CRLF_BINARY;
 549         else if (ATTR_UNSET(value))
 550                 ;
 551         else if (!strcmp(value, "input"))
 552                 return CRLF_INPUT;
 553         return CRLF_GUESS;
 554 }
 555
 556 static struct convert_driver *git_path_check_convert(const char *path,
 557                                              struct git_attr_check *check)
 558 {
 559         const char *value = check->value;
 560         struct convert_driver *drv;
 561
 562         if (ATTR_TRUE(value) || ATTR_FALSE(value) || ATTR_UNSET(value))
 563                 return NULL;
 564         for (drv = user_convert; drv; drv = drv->next)
 565                 if (!strcmp(value, drv->name))
 566                         return drv;
 567         return NULL;
 568 }
 569
 570 static int git_path_check_ident(const char *path, struct git_attr_check *check)
 571 {
 572         const char *value = check->value;
 573
 574         return !!ATTR_TRUE(value);
 575 }
 576
 577 int convert_to_git(const char *path, const char *src, size_t len,
 578                    struct strbuf *dst, enum safe_crlf checksafe)
 579 {
 580         struct git_attr_check check[3];
 581         int crlf = CRLF_GUESS;
 582         int ident = 0, ret = 0;
 583         char *filter = NULL;
 584
 585         setup_convert_check(check);
 586         if (!git_checkattr(path, ARRAY_SIZE(check), check)) {
 587                 struct convert_driver *drv;
 588                 crlf = git_path_check_crlf(path, check + 0);
 589                 ident = git_path_check_ident(path, check + 1);
 590                 drv = git_path_check_convert(path, check + 2);
 591                 if (drv && drv->clean)
 592                         filter = drv->clean;
 593         }
 594
 595         ret |= apply_filter(path, src, len, dst, filter);
 596         if (ret) {
 597                 src = dst->buf;
 598                 len = dst->len;
 599         }
 600         ret |= crlf_to_git(path, src, len, dst, crlf, checksafe);
 601         if (ret) {
 602                 src = dst->buf;
 603                 len = dst->len;
 604         }
 605         return ret | ident_to_git(path, src, len, dst, ident);
 606 }
 607
 608 int convert_to_working_tree(const char *path, const char *src, size_t len, struct strbuf *dst)
 609 {
 610         struct git_attr_check check[3];
 611         int crlf = CRLF_GUESS;
 612         int ident = 0, ret = 0;
 613         char *filter = NULL;
 614
 615         setup_convert_check(check);
 616         if (!git_checkattr(path, ARRAY_SIZE(check), check)) {
 617                 struct convert_driver *drv;
 618                 crlf = git_path_check_crlf(path, check + 0);
 619                 ident = git_path_check_ident(path, check + 1);
 620                 drv = git_path_check_convert(path, check + 2);
 621                 if (drv && drv->smudge)
 622                         filter = drv->smudge;
 623         }
 624
 625         ret |= ident_to_worktree(path, src, len, dst, ident);
 626         if (ret) {
 627                 src = dst->buf;
 628                 len = dst->len;
 629         }
 630         ret |= crlf_to_worktree(path, src, len, dst, crlf);
 631         if (ret) {
 632                 src = dst->buf;
 633                 len = dst->len;
 634         }
 635         return ret | apply_filter(path, src, len, dst, filter);
 636 }