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
  63 /*
  64  * The same heuristics as diff.c::mmfile_is_binary()
  65  */
  66 static int is_binary(unsigned long size, struct text_stat *stats)
  67 {
  68
  69         if ((stats->printable >> 7) < stats->nonprintable)
  70                 return 1;
  71         /*
  72          * Other heuristics? Average line length might be relevant,
  73          * as might LF vs CR vs CRLF counts..
  74          *
  75          * NOTE! It might be normal to have a low ratio of CRLF to LF
  76          * (somebody starts with a LF-only file and edits it with an editor
  77          * that adds CRLF only to lines that are added..). But do  we
  78          * want to support CR-only? Probably not.
  79          */
  80         return 0;
  81 }
  82
  83 static int crlf_to_git(const char *path, const char *src, size_t len,
  84                        struct strbuf *buf, int action)
  85 {
  86         struct text_stat stats;
  87         char *dst;
  88
  89         if ((action == CRLF_BINARY) || !auto_crlf || !len)
  90                 return 0;
  91
  92         gather_stats(src, len, &stats);
  93         /* No CR? Nothing to convert, regardless. */
  94         if (!stats.cr)
  95                 return 0;
  96
  97         if (action == CRLF_GUESS) {
  98                 /*
  99                  * We're currently not going to even try to convert stuff
 100                  * that has bare CR characters. Does anybody do that crazy
 101                  * stuff?
 102                  */
 103                 if (stats.cr != stats.crlf)
 104                         return 0;
 105
 106                 /*
 107                  * And add some heuristics for binary vs text, of course...
 108                  */
 109                 if (is_binary(len, &stats))
 110                         return 0;
 111         }
 112
 113         strbuf_grow(buf, len);
 114         dst = buf->buf;
 115         if (action == CRLF_GUESS) {
 116                 /*
 117                  * If we guessed, we already know we rejected a file with
 118                  * lone CR, and we can strip a CR without looking at what
 119                  * follow it.
 120                  */
 121                 do {
 122                         unsigned char c = *src++;
 123                         if (c != '\r')
 124                                 *dst++ = c;
 125                 } while (--len);
 126         } else {
 127                 do {
 128                         unsigned char c = *src++;
 129                         if (! (c == '\r' && (1 < len && *src == '\n')))
 130                                 *dst++ = c;
 131                 } while (--len);
 132         }
 133         strbuf_setlen(buf, dst - buf->buf);
 134         return 1;
 135 }
 136
 137 static int crlf_to_worktree(const char *path, const char *src, size_t len,
 138                             struct strbuf *buf, int action)
 139 {
 140         char *to_free = NULL;
 141         struct text_stat stats;
 142
 143         if ((action == CRLF_BINARY) || (action == CRLF_INPUT) ||
 144             auto_crlf <= 0)
 145                 return 0;
 146
 147         if (!len)
 148                 return 0;
 149
 150         gather_stats(src, len, &stats);
 151
 152         /* No LF? Nothing to convert, regardless. */
 153         if (!stats.lf)
 154                 return 0;
 155
 156         /* Was it already in CRLF format? */
 157         if (stats.lf == stats.crlf)
 158                 return 0;
 159
 160         if (action == CRLF_GUESS) {
 161                 /* If we have any bare CR characters, we're not going to touch it */
 162                 if (stats.cr != stats.crlf)
 163                         return 0;
 164
 165                 if (is_binary(len, &stats))
 166                         return 0;
 167         }
 168
 169         /* are we "faking" in place editing ? */
 170         if (src == buf->buf)
 171                 to_free = strbuf_detach(buf);
 172
 173         strbuf_grow(buf, len + stats.lf - stats.crlf);
 174         for (;;) {
 175                 const char *nl = memchr(src, '\n', len);
 176                 if (!nl)
 177                         break;
 178                 if (nl > src && nl[-1] == '\r') {
 179                         strbuf_add(buf, src, nl + 1 - src);
 180                 } else {
 181                         strbuf_add(buf, src, nl - src);
 182                         strbuf_addstr(buf, "\r\n");
 183                 }
 184                 len -= nl + 1 - src;
 185                 src  = nl + 1;
 186         }
 187         strbuf_add(buf, src, len);
 188
 189         free(to_free);
 190         return 1;
 191 }
 192
 193 static int filter_buffer(const char *path, const char *src,
 194                          unsigned long size, const char *cmd)
 195 {
 196         /*
 197          * Spawn cmd and feed the buffer contents through its stdin.
 198          */
 199         struct child_process child_process;
 200         int pipe_feed[2];
 201         int write_err, status;
 202
 203         memset(&child_process, 0, sizeof(child_process));
 204
 205         if (pipe(pipe_feed) < 0) {
 206                 error("cannot create pipe to run external filter %s", cmd);
 207                 return 1;
 208         }
 209
 210         child_process.pid = fork();
 211         if (child_process.pid < 0) {
 212                 error("cannot fork to run external filter %s", cmd);
 213                 close(pipe_feed[0]);
 214                 close(pipe_feed[1]);
 215                 return 1;
 216         }
 217         if (!child_process.pid) {
 218                 dup2(pipe_feed[0], 0);
 219                 close(pipe_feed[0]);
 220                 close(pipe_feed[1]);
 221                 execlp("sh", "sh", "-c", cmd, NULL);
 222                 return 1;
 223         }
 224         close(pipe_feed[0]);
 225
 226         write_err = (write_in_full(pipe_feed[1], src, size) < 0);
 227         if (close(pipe_feed[1]))
 228                 write_err = 1;
 229         if (write_err)
 230                 error("cannot feed the input to external filter %s", cmd);
 231
 232         status = finish_command(&child_process);
 233         if (status)
 234                 error("external filter %s failed %d", cmd, -status);
 235         return (write_err || status);
 236 }
 237
 238 static int apply_filter(const char *path, const char *src, size_t len,
 239                         struct strbuf *dst, const char *cmd)
 240 {
 241         /*
 242          * Create a pipeline to have the command filter the buffer's
 243          * contents.
 244          *
 245          * (child --> cmd) --> us
 246          */
 247         int pipe_feed[2];
 248         int status, ret = 1;
 249         struct child_process child_process;
 250         struct strbuf nbuf;
 251
 252         if (!cmd)
 253                 return 0;
 254
 255         memset(&child_process, 0, sizeof(child_process));
 256
 257         if (pipe(pipe_feed) < 0) {
 258                 error("cannot create pipe to run external filter %s", cmd);
 259                 return 0;
 260         }
 261
 262         fflush(NULL);
 263         child_process.pid = fork();
 264         if (child_process.pid < 0) {
 265                 error("cannot fork to run external filter %s", cmd);
 266                 close(pipe_feed[0]);
 267                 close(pipe_feed[1]);
 268                 return 0;
 269         }
 270         if (!child_process.pid) {
 271                 dup2(pipe_feed[1], 1);
 272                 close(pipe_feed[0]);
 273                 close(pipe_feed[1]);
 274                 exit(filter_buffer(path, src, len, cmd));
 275         }
 276         close(pipe_feed[1]);
 277
 278         strbuf_init(&nbuf, 0);
 279         if (strbuf_read(&nbuf, pipe_feed[0], len) < 0) {
 280                 error("read from external filter %s failed", cmd);
 281                 ret = 0;
 282         }
 283         if (close(pipe_feed[0])) {
 284                 ret = error("read from external filter %s failed", cmd);
 285                 ret = 0;
 286         }
 287         status = finish_command(&child_process);
 288         if (status) {
 289                 ret = error("external filter %s failed %d", cmd, -status);
 290                 ret = 0;
 291         }
 292
 293         if (ret) {
 294                 *dst = nbuf;
 295         } else {
 296                 strbuf_release(&nbuf);
 297         }
 298         return ret;
 299 }
 300
 301 static struct convert_driver {
 302         const char *name;
 303         struct convert_driver *next;
 304         char *smudge;
 305         char *clean;
 306 } *user_convert, **user_convert_tail;
 307
 308 static int read_convert_config(const char *var, const char *value)
 309 {
 310         const char *ep, *name;
 311         int namelen;
 312         struct convert_driver *drv;
 313
 314         /*
 315          * External conversion drivers are configured using
 316          * "filter.<name>.variable".
 317          */
 318         if (prefixcmp(var, "filter.") || (ep = strrchr(var, '.')) == var + 6)
 319                 return 0;
 320         name = var + 7;
 321         namelen = ep - name;
 322         for (drv = user_convert; drv; drv = drv->next)
 323                 if (!strncmp(drv->name, name, namelen) && !drv->name[namelen])
 324                         break;
 325         if (!drv) {
 326                 drv = xcalloc(1, sizeof(struct convert_driver));
 327                 drv->name = xmemdupz(name, namelen);
 328                 *user_convert_tail = drv;
 329                 user_convert_tail = &(drv->next);
 330         }
 331
 332         ep++;
 333
 334         /*
 335          * filter.<name>.smudge and filter.<name>.clean specifies
 336          * the command line:
 337          *
 338          *      command-line
 339          *
 340          * The command-line will not be interpolated in any way.
 341          */
 342
 343         if (!strcmp("smudge", ep)) {
 344                 if (!value)
 345                         return error("%s: lacks value", var);
 346                 drv->smudge = strdup(value);
 347                 return 0;
 348         }
 349
 350         if (!strcmp("clean", ep)) {
 351                 if (!value)
 352                         return error("%s: lacks value", var);
 353                 drv->clean = strdup(value);
 354                 return 0;
 355         }
 356         return 0;
 357 }
 358
 359 static void setup_convert_check(struct git_attr_check *check)
 360 {
 361         static struct git_attr *attr_crlf;
 362         static struct git_attr *attr_ident;
 363         static struct git_attr *attr_filter;
 364
 365         if (!attr_crlf) {
 366                 attr_crlf = git_attr("crlf", 4);
 367                 attr_ident = git_attr("ident", 5);
 368                 attr_filter = git_attr("filter", 6);
 369                 user_convert_tail = &user_convert;
 370                 git_config(read_convert_config);
 371         }
 372         check[0].attr = attr_crlf;
 373         check[1].attr = attr_ident;
 374         check[2].attr = attr_filter;
 375 }
 376
 377 static int count_ident(const char *cp, unsigned long size)
 378 {
 379         /*
 380          * "$Id: 0000000000000000000000000000000000000000 $" <=> "$Id$"
 381          */
 382         int cnt = 0;
 383         char ch;
 384
 385         while (size) {
 386                 ch = *cp++;
 387                 size--;
 388                 if (ch != '$')
 389                         continue;
 390                 if (size < 3)
 391                         break;
 392                 if (memcmp("Id", cp, 2))
 393                         continue;
 394                 ch = cp[2];
 395                 cp += 3;
 396                 size -= 3;
 397                 if (ch == '$')
 398                         cnt++; /* $Id$ */
 399                 if (ch != ':')
 400                         continue;
 401
 402                 /*
 403                  * "$Id: ... "; scan up to the closing dollar sign and discard.
 404                  */
 405                 while (size) {
 406                         ch = *cp++;
 407                         size--;
 408                         if (ch == '$') {
 409                                 cnt++;
 410                                 break;
 411                         }
 412                 }
 413         }
 414         return cnt;
 415 }
 416
 417 static int ident_to_git(const char *path, const char *src, size_t len,
 418                         struct strbuf *buf, int ident)
 419 {
 420         char *dst, *dollar;
 421
 422         if (!ident || !count_ident(src, len))
 423                 return 0;
 424
 425         strbuf_grow(buf, len);
 426         dst = buf->buf;
 427         for (;;) {
 428                 dollar = memchr(src, '$', len);
 429                 if (!dollar)
 430                         break;
 431                 memcpy(dst, src, dollar + 1 - src);
 432                 dst += dollar + 1 - src;
 433                 len -= dollar + 1 - src;
 434                 src  = dollar + 1;
 435
 436                 if (len > 3 && !memcmp(src, "Id:", 3)) {
 437                         dollar = memchr(src + 3, '$', len - 3);
 438                         if (!dollar)
 439                                 break;
 440                         memcpy(dst, "Id$", 3);
 441                         dst += 3;
 442                         len -= dollar + 1 - src;
 443                         src  = dollar + 1;
 444                 }
 445         }
 446         memcpy(dst, src, len);
 447         strbuf_setlen(buf, dst + len - buf->buf);
 448         return 1;
 449 }
 450
 451 static int ident_to_worktree(const char *path, const char *src, size_t len,
 452                              struct strbuf *buf, int ident)
 453 {
 454         unsigned char sha1[20];
 455         char *to_free = NULL, *dollar;
 456         int cnt;
 457
 458         if (!ident)
 459                 return 0;
 460
 461         cnt = count_ident(src, len);
 462         if (!cnt)
 463                 return 0;
 464
 465         /* are we "faking" in place editing ? */
 466         if (src == buf->buf)
 467                 to_free = strbuf_detach(buf);
 468         hash_sha1_file(src, len, "blob", sha1);
 469
 470         strbuf_grow(buf, len + cnt * 43);
 471         for (;;) {
 472                 /* step 1: run to the next '$' */
 473                 dollar = memchr(src, '$', len);
 474                 if (!dollar)
 475                         break;
 476                 strbuf_add(buf, src, dollar + 1 - src);
 477                 len -= dollar + 1 - src;
 478                 src  = dollar + 1;
 479
 480                 /* step 2: does it looks like a bit like Id:xxx$ or Id$ ? */
 481                 if (len < 3 || memcmp("Id", src, 2))
 482                         continue;
 483
 484                 /* step 3: skip over Id$ or Id:xxxxx$ */
 485                 if (src[2] == '$') {
 486                         src += 3;
 487                         len -= 3;
 488                 } else if (src[2] == ':') {
 489                         /*
 490                          * It's possible that an expanded Id has crept its way into the
 491                          * repository, we cope with that by stripping the expansion out
 492                          */
 493                         dollar = memchr(src + 3, '$', len - 3);
 494                         if (!dollar) {
 495                                 /* incomplete keyword, no more '$', so just quit the loop */
 496                                 break;
 497                         }
 498
 499                         len -= dollar + 1 - src;
 500                         src  = dollar + 1;
 501                 } else {
 502                         /* it wasn't a "Id$" or "Id:xxxx$" */
 503                         continue;
 504                 }
 505
 506                 /* step 4: substitute */
 507                 strbuf_addstr(buf, "Id: ");
 508                 strbuf_add(buf, sha1_to_hex(sha1), 40);
 509                 strbuf_addstr(buf, " $");
 510         }
 511         strbuf_add(buf, src, len);
 512
 513         free(to_free);
 514         return 1;
 515 }
 516
 517 static int git_path_check_crlf(const char *path, struct git_attr_check *check)
 518 {
 519         const char *value = check->value;
 520
 521         if (ATTR_TRUE(value))
 522                 return CRLF_TEXT;
 523         else if (ATTR_FALSE(value))
 524                 return CRLF_BINARY;
 525         else if (ATTR_UNSET(value))
 526                 ;
 527         else if (!strcmp(value, "input"))
 528                 return CRLF_INPUT;
 529         return CRLF_GUESS;
 530 }
 531
 532 static struct convert_driver *git_path_check_convert(const char *path,
 533                                              struct git_attr_check *check)
 534 {
 535         const char *value = check->value;
 536         struct convert_driver *drv;
 537
 538         if (ATTR_TRUE(value) || ATTR_FALSE(value) || ATTR_UNSET(value))
 539                 return NULL;
 540         for (drv = user_convert; drv; drv = drv->next)
 541                 if (!strcmp(value, drv->name))
 542                         return drv;
 543         return NULL;
 544 }
 545
 546 static int git_path_check_ident(const char *path, struct git_attr_check *check)
 547 {
 548         const char *value = check->value;
 549
 550         return !!ATTR_TRUE(value);
 551 }
 552
 553 int convert_to_git(const char *path, const char *src, size_t len, struct strbuf *dst)
 554 {
 555         struct git_attr_check check[3];
 556         int crlf = CRLF_GUESS;
 557         int ident = 0, ret = 0;
 558         char *filter = NULL;
 559
 560         setup_convert_check(check);
 561         if (!git_checkattr(path, ARRAY_SIZE(check), check)) {
 562                 struct convert_driver *drv;
 563                 crlf = git_path_check_crlf(path, check + 0);
 564                 ident = git_path_check_ident(path, check + 1);
 565                 drv = git_path_check_convert(path, check + 2);
 566                 if (drv && drv->clean)
 567                         filter = drv->clean;
 568         }
 569
 570         ret |= apply_filter(path, src, len, dst, filter);
 571         if (ret) {
 572                 src = dst->buf;
 573                 len = dst->len;
 574         }
 575         ret |= crlf_to_git(path, src, len, dst, crlf);
 576         if (ret) {
 577                 src = dst->buf;
 578                 len = dst->len;
 579         }
 580         return ret | ident_to_git(path, src, len, dst, ident);
 581 }
 582
 583 int convert_to_working_tree(const char *path, const char *src, size_t len, struct strbuf *dst)
 584 {
 585         struct git_attr_check check[3];
 586         int crlf = CRLF_GUESS;
 587         int ident = 0, ret = 0;
 588         char *filter = NULL;
 589
 590         setup_convert_check(check);
 591         if (!git_checkattr(path, ARRAY_SIZE(check), check)) {
 592                 struct convert_driver *drv;
 593                 crlf = git_path_check_crlf(path, check + 0);
 594                 ident = git_path_check_ident(path, check + 1);
 595                 drv = git_path_check_convert(path, check + 2);
 596                 if (drv && drv->smudge)
 597                         filter = drv->smudge;
 598         }
 599
 600         ret |= ident_to_worktree(path, src, len, dst, ident);
 601         if (ret) {
 602                 src = dst->buf;
 603                 len = dst->len;
 604         }
 605         ret |= crlf_to_worktree(path, src, len, dst, crlf);
 606         if (ret) {
 607                 src = dst->buf;
 608                 len = dst->len;
 609         }
 610         return ret | apply_filter(path, src, len, dst, filter);
 611 }