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