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         /* CR, LF and CRLF counts */
  21         unsigned 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                         default:
  55                                 stats->nonprintable++;
  56                         }
  57                 }
  58                 else
  59                         stats->printable++;
  60         }
  61
  62         // If file ends with EOF then don't count this EOF as non-printable
  63         if ( size >= 1 && buf[size-1] == '\032' )
  64                 stats->nonprintable--;
  65 }
  66
  67 /*
  68  * The same heuristics as diff.c::mmfile_is_binary()
  69  */
  70 static int is_binary(unsigned long size, struct text_stat *stats)
  71 {
  72
  73         if ((stats->printable >> 7) < stats->nonprintable)
  74                 return 1;
  75         /*
  76          * Other heuristics? Average line length might be relevant,
  77          * as might LF vs CR vs CRLF counts..
  78          *
  79          * NOTE! It might be normal to have a low ratio of CRLF to LF
  80          * (somebody starts with a LF-only file and edits it with an editor
  81          * that adds CRLF only to lines that are added..). But do  we
  82          * want to support CR-only? Probably not.
  83          */
  84         return 0;
  85 }
  86
  87 static char *crlf_to_git(const char *path, const char *src, unsigned long *sizep, int action)
  88 {
  89         char *buffer, *dst;
  90         unsigned long size, nsize;
  91         struct text_stat stats;
  92
  93         if ((action == CRLF_BINARY) || !auto_crlf)
  94                 return NULL;
  95
  96         size = *sizep;
  97         if (!size)
  98                 return NULL;
  99
 100         gather_stats(src, size, &stats);
 101
 102         /* No CR? Nothing to convert, regardless. */
 103         if (!stats.cr)
 104                 return NULL;
 105
 106         if (action == CRLF_GUESS) {
 107                 /*
 108                  * We're currently not going to even try to convert stuff
 109                  * that has bare CR characters. Does anybody do that crazy
 110                  * stuff?
 111                  */
 112                 if (stats.cr != stats.crlf)
 113                         return NULL;
 114
 115                 /*
 116                  * And add some heuristics for binary vs text, of course...
 117                  */
 118                 if (is_binary(size, &stats))
 119                         return NULL;
 120         }
 121
 122         /*
 123          * Ok, allocate a new buffer, fill it in, and return it
 124          * to let the caller know that we switched buffers.
 125          */
 126         nsize = size - stats.crlf;
 127         buffer = xmalloc(nsize);
 128         *sizep = nsize;
 129
 130         dst = buffer;
 131         if (action == CRLF_GUESS) {
 132                 /*
 133                  * If we guessed, we already know we rejected a file with
 134                  * lone CR, and we can strip a CR without looking at what
 135                  * follow it.
 136                  */
 137                 do {
 138                         unsigned char c = *src++;
 139                         if (c != '\r')
 140                                 *dst++ = c;
 141                 } while (--size);
 142         } else {
 143                 do {
 144                         unsigned char c = *src++;
 145                         if (! (c == '\r' && (1 < size && *src == '\n')))
 146                                 *dst++ = c;
 147                 } while (--size);
 148         }
 149
 150         return buffer;
 151 }
 152
 153 static char *crlf_to_worktree(const char *path, const char *src, unsigned long *sizep, int action)
 154 {
 155         char *buffer, *dst;
 156         unsigned long size, nsize;
 157         struct text_stat stats;
 158         unsigned char last;
 159
 160         if ((action == CRLF_BINARY) || (action == CRLF_INPUT) ||
 161             auto_crlf <= 0)
 162                 return NULL;
 163
 164         size = *sizep;
 165         if (!size)
 166                 return NULL;
 167
 168         gather_stats(src, size, &stats);
 169
 170         /* No LF? Nothing to convert, regardless. */
 171         if (!stats.lf)
 172                 return NULL;
 173
 174         /* Was it already in CRLF format? */
 175         if (stats.lf == stats.crlf)
 176                 return NULL;
 177
 178         if (action == CRLF_GUESS) {
 179                 /* If we have any bare CR characters, we're not going to touch it */
 180                 if (stats.cr != stats.crlf)
 181                         return NULL;
 182
 183                 if (is_binary(size, &stats))
 184                         return NULL;
 185         }
 186
 187         /*
 188          * Ok, allocate a new buffer, fill it in, and return it
 189          * to let the caller know that we switched buffers.
 190          */
 191         nsize = size + stats.lf - stats.crlf;
 192         buffer = xmalloc(nsize);
 193         *sizep = nsize;
 194         last = 0;
 195
 196         dst = buffer;
 197         do {
 198                 unsigned char c = *src++;
 199                 if (c == '\n' && last != '\r')
 200                         *dst++ = '\r';
 201                 *dst++ = c;
 202                 last = c;
 203         } while (--size);
 204
 205         return buffer;
 206 }
 207
 208 static int filter_buffer(const char *path, const char *src,
 209                          unsigned long size, const char *cmd)
 210 {
 211         /*
 212          * Spawn cmd and feed the buffer contents through its stdin.
 213          */
 214         struct child_process child_process;
 215         int pipe_feed[2];
 216         int write_err, status;
 217
 218         memset(&child_process, 0, sizeof(child_process));
 219
 220         if (pipe(pipe_feed) < 0) {
 221                 error("cannot create pipe to run external filter %s", cmd);
 222                 return 1;
 223         }
 224
 225         child_process.pid = fork();
 226         if (child_process.pid < 0) {
 227                 error("cannot fork to run external filter %s", cmd);
 228                 close(pipe_feed[0]);
 229                 close(pipe_feed[1]);
 230                 return 1;
 231         }
 232         if (!child_process.pid) {
 233                 dup2(pipe_feed[0], 0);
 234                 close(pipe_feed[0]);
 235                 close(pipe_feed[1]);
 236                 execlp("sh", "sh", "-c", cmd, NULL);
 237                 return 1;
 238         }
 239         close(pipe_feed[0]);
 240
 241         write_err = (write_in_full(pipe_feed[1], src, size) < 0);
 242         if (close(pipe_feed[1]))
 243                 write_err = 1;
 244         if (write_err)
 245                 error("cannot feed the input to external filter %s", cmd);
 246
 247         status = finish_command(&child_process);
 248         if (status)
 249                 error("external filter %s failed %d", cmd, -status);
 250         return (write_err || status);
 251 }
 252
 253 static char *apply_filter(const char *path, const char *src,
 254                           unsigned long *sizep, const char *cmd)
 255 {
 256         /*
 257          * Create a pipeline to have the command filter the buffer's
 258          * contents.
 259          *
 260          * (child --> cmd) --> us
 261          */
 262         const int SLOP = 4096;
 263         int pipe_feed[2];
 264         int status;
 265         char *dst;
 266         unsigned long dstsize, dstalloc;
 267         struct child_process child_process;
 268
 269         if (!cmd)
 270                 return NULL;
 271
 272         memset(&child_process, 0, sizeof(child_process));
 273
 274         if (pipe(pipe_feed) < 0) {
 275                 error("cannot create pipe to run external filter %s", cmd);
 276                 return NULL;
 277         }
 278
 279         fflush(NULL);
 280         child_process.pid = fork();
 281         if (child_process.pid < 0) {
 282                 error("cannot fork to run external filter %s", cmd);
 283                 close(pipe_feed[0]);
 284                 close(pipe_feed[1]);
 285                 return NULL;
 286         }
 287         if (!child_process.pid) {
 288                 dup2(pipe_feed[1], 1);
 289                 close(pipe_feed[0]);
 290                 close(pipe_feed[1]);
 291                 exit(filter_buffer(path, src, *sizep, cmd));
 292         }
 293         close(pipe_feed[1]);
 294
 295         dstalloc = *sizep;
 296         dst = xmalloc(dstalloc);
 297         dstsize = 0;
 298
 299         while (1) {
 300                 ssize_t numread = xread(pipe_feed[0], dst + dstsize,
 301                                         dstalloc - dstsize);
 302
 303                 if (numread <= 0) {
 304                         if (!numread)
 305                                 break;
 306                         error("read from external filter %s failed", cmd);
 307                         free(dst);
 308                         dst = NULL;
 309                         break;
 310                 }
 311                 dstsize += numread;
 312                 if (dstalloc <= dstsize + SLOP) {
 313                         dstalloc = dstsize + SLOP;
 314                         dst = xrealloc(dst, dstalloc);
 315                 }
 316         }
 317         if (close(pipe_feed[0])) {
 318                 error("read from external filter %s failed", cmd);
 319                 free(dst);
 320                 dst = NULL;
 321         }
 322
 323         status = finish_command(&child_process);
 324         if (status) {
 325                 error("external filter %s failed %d", cmd, -status);
 326                 free(dst);
 327                 dst = NULL;
 328         }
 329
 330         if (dst)
 331                 *sizep = dstsize;
 332         return dst;
 333 }
 334
 335 static struct convert_driver {
 336         const char *name;
 337         struct convert_driver *next;
 338         char *smudge;
 339         char *clean;
 340 } *user_convert, **user_convert_tail;
 341
 342 static int read_convert_config(const char *var, const char *value)
 343 {
 344         const char *ep, *name;
 345         int namelen;
 346         struct convert_driver *drv;
 347
 348         /*
 349          * External conversion drivers are configured using
 350          * "filter.<name>.variable".
 351          */
 352         if (prefixcmp(var, "filter.") || (ep = strrchr(var, '.')) == var + 6)
 353                 return 0;
 354         name = var + 7;
 355         namelen = ep - name;
 356         for (drv = user_convert; drv; drv = drv->next)
 357                 if (!strncmp(drv->name, name, namelen) && !drv->name[namelen])
 358                         break;
 359         if (!drv) {
 360                 char *namebuf;
 361                 drv = xcalloc(1, sizeof(struct convert_driver));
 362                 namebuf = xmalloc(namelen + 1);
 363                 memcpy(namebuf, name, namelen);
 364                 namebuf[namelen] = 0;
 365                 drv->name = namebuf;
 366                 drv->next = NULL;
 367                 *user_convert_tail = drv;
 368                 user_convert_tail = &(drv->next);
 369         }
 370
 371         ep++;
 372
 373         /*
 374          * filter.<name>.smudge and filter.<name>.clean specifies
 375          * the command line:
 376          *
 377          *      command-line
 378          *
 379          * The command-line will not be interpolated in any way.
 380          */
 381
 382         if (!strcmp("smudge", ep)) {
 383                 if (!value)
 384                         return error("%s: lacks value", var);
 385                 drv->smudge = strdup(value);
 386                 return 0;
 387         }
 388
 389         if (!strcmp("clean", ep)) {
 390                 if (!value)
 391                         return error("%s: lacks value", var);
 392                 drv->clean = strdup(value);
 393                 return 0;
 394         }
 395         return 0;
 396 }
 397
 398 static void setup_convert_check(struct git_attr_check *check)
 399 {
 400         static struct git_attr *attr_crlf;
 401         static struct git_attr *attr_ident;
 402         static struct git_attr *attr_filter;
 403
 404         if (!attr_crlf) {
 405                 attr_crlf = git_attr("crlf", 4);
 406                 attr_ident = git_attr("ident", 5);
 407                 attr_filter = git_attr("filter", 6);
 408                 user_convert_tail = &user_convert;
 409                 git_config(read_convert_config);
 410         }
 411         check[0].attr = attr_crlf;
 412         check[1].attr = attr_ident;
 413         check[2].attr = attr_filter;
 414 }
 415
 416 static int count_ident(const char *cp, unsigned long size)
 417 {
 418         /*
 419          * "$Id: 0000000000000000000000000000000000000000 $" <=> "$Id$"
 420          */
 421         int cnt = 0;
 422         char ch;
 423
 424         while (size) {
 425                 ch = *cp++;
 426                 size--;
 427                 if (ch != '$')
 428                         continue;
 429                 if (size < 3)
 430                         break;
 431                 if (memcmp("Id", cp, 2))
 432                         continue;
 433                 ch = cp[2];
 434                 cp += 3;
 435                 size -= 3;
 436                 if (ch == '$')
 437                         cnt++; /* $Id$ */
 438                 if (ch != ':')
 439                         continue;
 440
 441                 /*
 442                  * "$Id: ... "; scan up to the closing dollar sign and discard.
 443                  */
 444                 while (size) {
 445                         ch = *cp++;
 446                         size--;
 447                         if (ch == '$') {
 448                                 cnt++;
 449                                 break;
 450                         }
 451                 }
 452         }
 453         return cnt;
 454 }
 455
 456 static char *ident_to_git(const char *path, const char *src, unsigned long *sizep, int ident)
 457 {
 458         int cnt;
 459         unsigned long size;
 460         char *dst, *buf;
 461
 462         if (!ident)
 463                 return NULL;
 464         size = *sizep;
 465         cnt = count_ident(src, size);
 466         if (!cnt)
 467                 return NULL;
 468         buf = xmalloc(size);
 469
 470         for (dst = buf; size; size--) {
 471                 char ch = *src++;
 472                 *dst++ = ch;
 473                 if ((ch == '$') && (3 <= size) &&
 474                     !memcmp("Id:", src, 3)) {
 475                         unsigned long rem = size - 3;
 476                         const char *cp = src + 3;
 477                         do {
 478                                 ch = *cp++;
 479                                 if (ch == '$')
 480                                         break;
 481                                 rem--;
 482                         } while (rem);
 483                         if (!rem)
 484                                 continue;
 485                         memcpy(dst, "Id$", 3);
 486                         dst += 3;
 487                         size -= (cp - src);
 488                         src = cp;
 489                 }
 490         }
 491
 492         *sizep = dst - buf;
 493         return buf;
 494 }
 495
 496 static char *ident_to_worktree(const char *path, const char *src, unsigned long *sizep, int ident)
 497 {
 498         int cnt;
 499         unsigned long size;
 500         char *dst, *buf;
 501         unsigned char sha1[20];
 502
 503         if (!ident)
 504                 return NULL;
 505
 506         size = *sizep;
 507         cnt = count_ident(src, size);
 508         if (!cnt)
 509                 return NULL;
 510
 511         hash_sha1_file(src, size, "blob", sha1);
 512         buf = xmalloc(size + cnt * 43);
 513
 514         for (dst = buf; size; size--) {
 515                 const char *cp;
 516                 /* Fetch next source character, move the pointer on */
 517                 char ch = *src++;
 518                 /* Copy the current character to the destination */
 519                 *dst++ = ch;
 520                 /* If the current character is "$" or there are less than three
 521                  * remaining bytes or the two bytes following this one are not
 522                  * "Id", then simply read the next character */
 523                 if ((ch != '$') || (size < 3) || memcmp("Id", src, 2))
 524                         continue;
 525                 /*
 526                  * Here when
 527                  *  - There are more than 2 bytes remaining
 528                  *  - The current three bytes are "$Id"
 529                  * with
 530                  *  - ch == "$"
 531                  *  - src[0] == "I"
 532                  */
 533
 534                 /*
 535                  * It's possible that an expanded Id has crept its way into the
 536                  * repository, we cope with that by stripping the expansion out
 537                  */
 538                 if (src[2] == ':') {
 539                         /* Expanded keywords have "$Id:" at the front */
 540
 541                         /* discard up to but not including the closing $ */
 542                         unsigned long rem = size - 3;
 543                         /* Point at first byte after the ":" */
 544                         cp = src + 3;
 545                         /*
 546                          * Throw away characters until either
 547                          *  - we reach a "$"
 548                          *  - we run out of bytes (rem == 0)
 549                          */
 550                         do {
 551                                 ch = *cp;
 552                                 if (ch == '$')
 553                                         break;
 554                                 cp++;
 555                                 rem--;
 556                         } while (rem);
 557                         /* If the above finished because it ran out of characters, then
 558                          * this is an incomplete keyword, so don't run the expansion */
 559                         if (!rem)
 560                                 continue;
 561                 } else if (src[2] == '$')
 562                         cp = src + 2;
 563                 else
 564                         /* Anything other than "$Id:XXX$" or $Id$ and we skip the
 565                          * expansion */
 566                         continue;
 567
 568                 /* cp is now pointing at the last $ of the keyword */
 569
 570                 memcpy(dst, "Id: ", 4);
 571                 dst += 4;
 572                 memcpy(dst, sha1_to_hex(sha1), 40);
 573                 dst += 40;
 574                 *dst++ = ' ';
 575
 576                 /* Adjust for the characters we've discarded */
 577                 size -= (cp - src);
 578                 src = cp;
 579
 580                 /* Copy the final "$" */
 581                 *dst++ = *src++;
 582                 size--;
 583         }
 584
 585         *sizep = dst - buf;
 586         return buf;
 587 }
 588
 589 static int git_path_check_crlf(const char *path, struct git_attr_check *check)
 590 {
 591         const char *value = check->value;
 592
 593         if (ATTR_TRUE(value))
 594                 return CRLF_TEXT;
 595         else if (ATTR_FALSE(value))
 596                 return CRLF_BINARY;
 597         else if (ATTR_UNSET(value))
 598                 ;
 599         else if (!strcmp(value, "input"))
 600                 return CRLF_INPUT;
 601         return CRLF_GUESS;
 602 }
 603
 604 static struct convert_driver *git_path_check_convert(const char *path,
 605                                              struct git_attr_check *check)
 606 {
 607         const char *value = check->value;
 608         struct convert_driver *drv;
 609
 610         if (ATTR_TRUE(value) || ATTR_FALSE(value) || ATTR_UNSET(value))
 611                 return NULL;
 612         for (drv = user_convert; drv; drv = drv->next)
 613                 if (!strcmp(value, drv->name))
 614                         return drv;
 615         return NULL;
 616 }
 617
 618 static int git_path_check_ident(const char *path, struct git_attr_check *check)
 619 {
 620         const char *value = check->value;
 621
 622         return !!ATTR_TRUE(value);
 623 }
 624
 625 char *convert_to_git(const char *path, const char *src, unsigned long *sizep)
 626 {
 627         struct git_attr_check check[3];
 628         int crlf = CRLF_GUESS;
 629         int ident = 0;
 630         char *filter = NULL;
 631         char *buf, *buf2;
 632
 633         setup_convert_check(check);
 634         if (!git_checkattr(path, ARRAY_SIZE(check), check)) {
 635                 struct convert_driver *drv;
 636                 crlf = git_path_check_crlf(path, check + 0);
 637                 ident = git_path_check_ident(path, check + 1);
 638                 drv = git_path_check_convert(path, check + 2);
 639                 if (drv && drv->clean)
 640                         filter = drv->clean;
 641         }
 642
 643         buf = apply_filter(path, src, sizep, filter);
 644
 645         buf2 = crlf_to_git(path, buf ? buf : src, sizep, crlf);
 646         if (buf2) {
 647                 free(buf);
 648                 buf = buf2;
 649         }
 650
 651         buf2 = ident_to_git(path, buf ? buf : src, sizep, ident);
 652         if (buf2) {
 653                 free(buf);
 654                 buf = buf2;
 655         }
 656
 657         return buf;
 658 }
 659
 660 char *convert_to_working_tree(const char *path, const char *src, unsigned long *sizep)
 661 {
 662         struct git_attr_check check[3];
 663         int crlf = CRLF_GUESS;
 664         int ident = 0;
 665         char *filter = NULL;
 666         char *buf, *buf2;
 667
 668         setup_convert_check(check);
 669         if (!git_checkattr(path, ARRAY_SIZE(check), check)) {
 670                 struct convert_driver *drv;
 671                 crlf = git_path_check_crlf(path, check + 0);
 672                 ident = git_path_check_ident(path, check + 1);
 673                 drv = git_path_check_convert(path, check + 2);
 674                 if (drv && drv->smudge)
 675                         filter = drv->smudge;
 676         }
 677
 678         buf = ident_to_worktree(path, src, sizep, ident);
 679
 680         buf2 = crlf_to_worktree(path, buf ? buf : src, sizep, crlf);
 681         if (buf2) {
 682                 free(buf);
 683                 buf = buf2;
 684         }
 685
 686         buf2 = apply_filter(path, buf ? buf : src, sizep, filter);
 687         if (buf2) {
 688                 free(buf);
 689                 buf = buf2;
 690         }
 691
 692         return buf;
 693 }
 694
 695 void *convert_sha1_file(const char *path, const unsigned char *sha1,
 696                         unsigned int mode, enum object_type *type,
 697                         unsigned long *size)
 698 {
 699         void *buffer = read_sha1_file(sha1, type, size);
 700         if (S_ISREG(mode) && buffer) {
 701                 void *converted = convert_to_working_tree(path, buffer, size);
 702                 if (converted) {
 703                         free(buffer);
 704                         buffer = converted;
 705                 }
 706         }
 707         return buffer;
 708 }