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