apply.c

   1 /*
   2  * apply.c
   3  *
   4  * Copyright (C) Linus Torvalds, 2005
   5  *
   6  * This applies patches on top of some (arbitrary) version of the SCM.
   7  *
   8  * NOTE! It does all its work in the index file, and only cares about
   9  * the files in the working directory if you tell it to "merge" the
  10  * patch apply.
  11  *
  12  * Even when merging it always takes the source from the index, and
  13  * uses the working tree as a "branch" for a 3-way merge.
  14  */
  15 #include <ctype.h>
  16
  17 #include "cache.h"
  18
  19 // We default to the merge behaviour, since that's what most people would
  20 // expect.
  21 //
  22 //  --check turns on checking that the working tree matches the
  23 //    files that are being modified, but doesn't apply the patch
  24 //  --stat does just a diffstat, and doesn't actually apply
  25 //  --show-files shows the directory changes
  26 //
  27 static int merge_patch = 1;
  28 static int check_index = 0;
  29 static int write_index = 0;
  30 static int diffstat = 0;
  31 static int summary = 0;
  32 static int check = 0;
  33 static int apply = 1;
  34 static int show_files = 0;
  35 static const char apply_usage[] = "git-apply [--stat] [--summary] [--check] [--show-files] <patch>";
  36
  37 /*
  38  * For "diff-stat" like behaviour, we keep track of the biggest change
  39  * we've seen, and the longest filename. That allows us to do simple
  40  * scaling.
  41  */
  42 static int max_change, max_len;
  43
  44 /*
  45  * Various "current state", notably line numbers and what
  46  * file (and how) we're patching right now.. The "is_xxxx"
  47  * things are flags, where -1 means "don't know yet".
  48  */
  49 static int linenr = 1;
  50
  51 struct fragment {
  52         unsigned long oldpos, oldlines;
  53         unsigned long newpos, newlines;
  54         const char *patch;
  55         int size;
  56         struct fragment *next;
  57 };
  58
  59 struct patch {
  60         char *new_name, *old_name, *def_name;
  61         unsigned int old_mode, new_mode;
  62         int is_rename, is_copy, is_new, is_delete;
  63         int lines_added, lines_deleted;
  64         int score;
  65         struct fragment *fragments;
  66         char *result;
  67         unsigned long resultsize;
  68         struct patch *next;
  69 };
  70
  71 #define CHUNKSIZE (8192)
  72 #define SLOP (16)
  73
  74 static void *read_patch_file(int fd, unsigned long *sizep)
  75 {
  76         unsigned long size = 0, alloc = CHUNKSIZE;
  77         void *buffer = xmalloc(alloc);
  78
  79         for (;;) {
  80                 int nr = alloc - size;
  81                 if (nr < 1024) {
  82                         alloc += CHUNKSIZE;
  83                         buffer = xrealloc(buffer, alloc);
  84                         nr = alloc - size;
  85                 }
  86                 nr = read(fd, buffer + size, nr);
  87                 if (!nr)
  88                         break;
  89                 if (nr < 0) {
  90                         if (errno == EAGAIN)
  91                                 continue;
  92                         die("git-apply: read returned %s", strerror(errno));
  93                 }
  94                 size += nr;
  95         }
  96         *sizep = size;
  97
  98         /*
  99          * Make sure that we have some slop in the buffer
 100          * so that we can do speculative "memcmp" etc, and
 101          * see to it that it is NUL-filled.
 102          */
 103         if (alloc < size + SLOP)
 104                 buffer = xrealloc(buffer, size + SLOP);
 105         memset(buffer + size, 0, SLOP);
 106         return buffer;
 107 }
 108
 109 static unsigned long linelen(const char *buffer, unsigned long size)
 110 {
 111         unsigned long len = 0;
 112         while (size--) {
 113                 len++;
 114                 if (*buffer++ == '\n')
 115                         break;
 116         }
 117         return len;
 118 }
 119
 120 static int is_dev_null(const char *str)
 121 {
 122         return !memcmp("/dev/null", str, 9) && isspace(str[9]);
 123 }
 124
 125 #define TERM_SPACE      1
 126 #define TERM_TAB        2
 127
 128 static int name_terminate(const char *name, int namelen, int c, int terminate)
 129 {
 130         if (c == ' ' && !(terminate & TERM_SPACE))
 131                 return 0;
 132         if (c == '\t' && !(terminate & TERM_TAB))
 133                 return 0;
 134
 135         return 1;
 136 }
 137
 138 static char * find_name(const char *line, char *def, int p_value, int terminate)
 139 {
 140         int len;
 141         const char *start = line;
 142         char *name;
 143
 144         for (;;) {
 145                 char c = *line;
 146
 147                 if (isspace(c)) {
 148                         if (c == '\n')
 149                                 break;
 150                         if (name_terminate(start, line-start, c, terminate))
 151                                 break;
 152                 }
 153                 line++;
 154                 if (c == '/' && !--p_value)
 155                         start = line;
 156         }
 157         if (!start)
 158                 return def;
 159         len = line - start;
 160         if (!len)
 161                 return def;
 162
 163         /*
 164          * Generally we prefer the shorter name, especially
 165          * if the other one is just a variation of that with
 166          * something else tacked on to the end (ie "file.orig"
 167          * or "file~").
 168          */
 169         if (def) {
 170                 int deflen = strlen(def);
 171                 if (deflen < len && !strncmp(start, def, deflen))
 172                         return def;
 173         }
 174
 175         name = xmalloc(len + 1);
 176         memcpy(name, start, len);
 177         name[len] = 0;
 178         free(def);
 179         return name;
 180 }
 181
 182 /*
 183  * Get the name etc info from the --/+++ lines of a traditional patch header
 184  *
 185  * NOTE! This hardcodes "-p1" behaviour in filename detection.
 186  *
 187  * FIXME! The end-of-filename heuristics are kind of screwy. For existing
 188  * files, we can happily check the index for a match, but for creating a
 189  * new file we should try to match whatever "patch" does. I have no idea.
 190  */
 191 static void parse_traditional_patch(const char *first, const char *second, struct patch *patch)
 192 {
 193         int p_value = 1;
 194         char *name;
 195
 196         first += 4;     // skip "--- "
 197         second += 4;    // skip "+++ "
 198         if (is_dev_null(first)) {
 199                 patch->is_new = 1;
 200                 patch->is_delete = 0;
 201                 name = find_name(second, NULL, p_value, TERM_SPACE | TERM_TAB);
 202                 patch->new_name = name;
 203         } else if (is_dev_null(second)) {
 204                 patch->is_new = 0;
 205                 patch->is_delete = 1;
 206                 name = find_name(first, NULL, p_value, TERM_SPACE | TERM_TAB);
 207                 patch->old_name = name;
 208         } else {
 209                 name = find_name(first, NULL, p_value, TERM_SPACE | TERM_TAB);
 210                 name = find_name(second, name, p_value, TERM_SPACE | TERM_TAB);
 211                 patch->old_name = patch->new_name = name;
 212         }
 213         if (!name)
 214                 die("unable to find filename in patch at line %d", linenr);
 215 }
 216
 217 static int gitdiff_hdrend(const char *line, struct patch *patch)
 218 {
 219         return -1;
 220 }
 221
 222 /*
 223  * We're anal about diff header consistency, to make
 224  * sure that we don't end up having strange ambiguous
 225  * patches floating around.
 226  *
 227  * As a result, gitdiff_{old|new}name() will check
 228  * their names against any previous information, just
 229  * to make sure..
 230  */
 231 static char *gitdiff_verify_name(const char *line, int isnull, char *orig_name, const char *oldnew)
 232 {
 233         int len;
 234         const char *name;
 235
 236         if (!orig_name && !isnull)
 237                 return find_name(line, NULL, 1, 0);
 238
 239         name = "/dev/null";
 240         len = 9;
 241         if (orig_name) {
 242                 name = orig_name;
 243                 len = strlen(name);
 244                 if (isnull)
 245                         die("git-apply: bad git-diff - expected /dev/null, got %s on line %d", name, linenr);
 246         }
 247
 248         if (*name == '/')
 249                 goto absolute_path;
 250
 251         for (;;) {
 252                 char c = *line++;
 253                 if (c == '\n')
 254                         break;
 255                 if (c != '/')
 256                         continue;
 257 absolute_path:
 258                 if (memcmp(line, name, len) || line[len] != '\n')
 259                         break;
 260                 return orig_name;
 261         }
 262         die("git-apply: bad git-diff - inconsistent %s filename on line %d", oldnew, linenr);
 263         return NULL;
 264 }
 265
 266 static int gitdiff_oldname(const char *line, struct patch *patch)
 267 {
 268         patch->old_name = gitdiff_verify_name(line, patch->is_new, patch->old_name, "old");
 269         return 0;
 270 }
 271
 272 static int gitdiff_newname(const char *line, struct patch *patch)
 273 {
 274         patch->new_name = gitdiff_verify_name(line, patch->is_delete, patch->new_name, "new");
 275         return 0;
 276 }
 277
 278 static int gitdiff_oldmode(const char *line, struct patch *patch)
 279 {
 280         patch->old_mode = strtoul(line, NULL, 8);
 281         return 0;
 282 }
 283
 284 static int gitdiff_newmode(const char *line, struct patch *patch)
 285 {
 286         patch->new_mode = strtoul(line, NULL, 8);
 287         return 0;
 288 }
 289
 290 static int gitdiff_delete(const char *line, struct patch *patch)
 291 {
 292         patch->is_delete = 1;
 293         patch->old_name = patch->def_name;
 294         return gitdiff_oldmode(line, patch);
 295 }
 296
 297 static int gitdiff_newfile(const char *line, struct patch *patch)
 298 {
 299         patch->is_new = 1;
 300         patch->new_name = patch->def_name;
 301         return gitdiff_newmode(line, patch);
 302 }
 303
 304 static int gitdiff_copysrc(const char *line, struct patch *patch)
 305 {
 306         patch->is_copy = 1;
 307         patch->old_name = find_name(line, NULL, 0, 0);
 308         return 0;
 309 }
 310
 311 static int gitdiff_copydst(const char *line, struct patch *patch)
 312 {
 313         patch->is_copy = 1;
 314         patch->new_name = find_name(line, NULL, 0, 0);
 315         return 0;
 316 }
 317
 318 static int gitdiff_renamesrc(const char *line, struct patch *patch)
 319 {
 320         patch->is_rename = 1;
 321         patch->old_name = find_name(line, NULL, 0, 0);
 322         return 0;
 323 }
 324
 325 static int gitdiff_renamedst(const char *line, struct patch *patch)
 326 {
 327         patch->is_rename = 1;
 328         patch->new_name = find_name(line, NULL, 0, 0);
 329         return 0;
 330 }
 331
 332 static int gitdiff_similarity(const char *line, struct patch *patch)
 333 {
 334         if ((patch->score = strtoul(line, NULL, 10)) == ULONG_MAX)
 335                 patch->score = 0;
 336         return 0;
 337 }
 338
 339 static int gitdiff_dissimilarity(const char *line, struct patch *patch)
 340 {
 341         if ((patch->score = strtoul(line, NULL, 10)) == ULONG_MAX)
 342                 patch->score = 0;
 343         return 0;
 344 }
 345
 346 /*
 347  * This is normal for a diff that doesn't change anything: we'll fall through
 348  * into the next diff. Tell the parser to break out.
 349  */
 350 static int gitdiff_unrecognized(const char *line, struct patch *patch)
 351 {
 352         return -1;
 353 }
 354
 355 static char *git_header_name(char *line)
 356 {
 357         int len;
 358         char *name, *second;
 359
 360         /*
 361          * Find the first '/'
 362          */
 363         name = line;
 364         for (;;) {
 365                 char c = *name++;
 366                 if (c == '\n')
 367                         return NULL;
 368                 if (c == '/')
 369                         break;
 370         }
 371
 372         /*
 373          * We don't accept absolute paths (/dev/null) as possibly valid
 374          */
 375         if (name == line+1)
 376                 return NULL;
 377
 378         /*
 379          * Accept a name only if it shows up twice, exactly the same
 380          * form.
 381          */
 382         for (len = 0 ; ; len++) {
 383                 char c = name[len];
 384
 385                 switch (c) {
 386                 default:
 387                         continue;
 388                 case '\n':
 389                         break;
 390                 case '\t': case ' ':
 391                         second = name+len;
 392                         for (;;) {
 393                                 char c = *second++;
 394                                 if (c == '\n')
 395                                         return NULL;
 396                                 if (c == '/')
 397                                         break;
 398                         }
 399                         if (second[len] == '\n' && !memcmp(name, second, len)) {
 400                                 char *ret = xmalloc(len + 1);
 401                                 memcpy(ret, name, len);
 402                                 ret[len] = 0;
 403                                 return ret;
 404                         }
 405                 }
 406         }
 407         return NULL;
 408 }
 409
 410 /* Verify that we recognize the lines following a git header */
 411 static int parse_git_header(char *line, int len, unsigned int size, struct patch *patch)
 412 {
 413         unsigned long offset;
 414
 415         /* A git diff has explicit new/delete information, so we don't guess */
 416         patch->is_new = 0;
 417         patch->is_delete = 0;
 418
 419         /*
 420          * Some things may not have the old name in the
 421          * rest of the headers anywhere (pure mode changes,
 422          * or removing or adding empty files), so we get
 423          * the default name from the header.
 424          */
 425         patch->def_name = git_header_name(line + strlen("diff --git "));
 426
 427         line += len;
 428         size -= len;
 429         linenr++;
 430         for (offset = len ; size > 0 ; offset += len, size -= len, line += len, linenr++) {
 431                 static const struct opentry {
 432                         const char *str;
 433                         int (*fn)(const char *, struct patch *);
 434                 } optable[] = {
 435                         { "@@ -", gitdiff_hdrend },
 436                         { "--- ", gitdiff_oldname },
 437                         { "+++ ", gitdiff_newname },
 438                         { "old mode ", gitdiff_oldmode },
 439                         { "new mode ", gitdiff_newmode },
 440                         { "deleted file mode ", gitdiff_delete },
 441                         { "new file mode ", gitdiff_newfile },
 442                         { "copy from ", gitdiff_copysrc },
 443                         { "copy to ", gitdiff_copydst },
 444                         { "rename old ", gitdiff_renamesrc },
 445                         { "rename new ", gitdiff_renamedst },
 446                         { "rename from ", gitdiff_renamesrc },
 447                         { "rename to ", gitdiff_renamedst },
 448                         { "similarity index ", gitdiff_similarity },
 449                         { "dissimilarity index ", gitdiff_dissimilarity },
 450                         { "", gitdiff_unrecognized },
 451                 };
 452                 int i;
 453
 454                 len = linelen(line, size);
 455                 if (!len || line[len-1] != '\n')
 456                         break;
 457                 for (i = 0; i < sizeof(optable) / sizeof(optable[0]); i++) {
 458                         const struct opentry *p = optable + i;
 459                         int oplen = strlen(p->str);
 460                         if (len < oplen || memcmp(p->str, line, oplen))
 461                                 continue;
 462                         if (p->fn(line + oplen, patch) < 0)
 463                                 return offset;
 464                         break;
 465                 }
 466         }
 467
 468         return offset;
 469 }
 470
 471 static int parse_num(const char *line, unsigned long *p)
 472 {
 473         char *ptr;
 474
 475         if (!isdigit(*line))
 476                 return 0;
 477         *p = strtoul(line, &ptr, 10);
 478         return ptr - line;
 479 }
 480
 481 static int parse_range(const char *line, int len, int offset, const char *expect,
 482                         unsigned long *p1, unsigned long *p2)
 483 {
 484         int digits, ex;
 485
 486         if (offset < 0 || offset >= len)
 487                 return -1;
 488         line += offset;
 489         len -= offset;
 490
 491         digits = parse_num(line, p1);
 492         if (!digits)
 493                 return -1;
 494
 495         offset += digits;
 496         line += digits;
 497         len -= digits;
 498
 499         *p2 = *p1;
 500         if (*line == ',') {
 501                 digits = parse_num(line+1, p2);
 502                 if (!digits)
 503                         return -1;
 504
 505                 offset += digits+1;
 506                 line += digits+1;
 507                 len -= digits+1;
 508         }
 509
 510         ex = strlen(expect);
 511         if (ex > len)
 512                 return -1;
 513         if (memcmp(line, expect, ex))
 514                 return -1;
 515
 516         return offset + ex;
 517 }
 518
 519 /*
 520  * Parse a unified diff fragment header of the
 521  * form "@@ -a,b +c,d @@"
 522  */
 523 static int parse_fragment_header(char *line, int len, struct fragment *fragment)
 524 {
 525         int offset;
 526
 527         if (!len || line[len-1] != '\n')
 528                 return -1;
 529
 530         /* Figure out the number of lines in a fragment */
 531         offset = parse_range(line, len, 4, " +", &fragment->oldpos, &fragment->oldlines);
 532         offset = parse_range(line, len, offset, " @@", &fragment->newpos, &fragment->newlines);
 533
 534         return offset;
 535 }
 536
 537 static int find_header(char *line, unsigned long size, int *hdrsize, struct patch *patch)
 538 {
 539         unsigned long offset, len;
 540
 541         patch->is_rename = patch->is_copy = 0;
 542         patch->is_new = patch->is_delete = -1;
 543         patch->old_mode = patch->new_mode = 0;
 544         patch->old_name = patch->new_name = NULL;
 545         for (offset = 0; size > 0; offset += len, size -= len, line += len, linenr++) {
 546                 unsigned long nextlen;
 547
 548                 len = linelen(line, size);
 549                 if (!len)
 550                         break;
 551
 552                 /* Testing this early allows us to take a few shortcuts.. */
 553                 if (len < 6)
 554                         continue;
 555
 556                 /*
 557                  * Make sure we don't find any unconnected patch fragmants.
 558                  * That's a sign that we didn't find a header, and that a
 559                  * patch has become corrupted/broken up.
 560                  */
 561                 if (!memcmp("@@ -", line, 4)) {
 562                         struct fragment dummy;
 563                         if (parse_fragment_header(line, len, &dummy) < 0)
 564                                 continue;
 565                         error("patch fragment without header at line %d: %.*s", linenr, len-1, line);
 566                 }
 567
 568                 if (size < len + 6)
 569                         break;
 570
 571                 /*
 572                  * Git patch? It might not have a real patch, just a rename
 573                  * or mode change, so we handle that specially
 574                  */
 575                 if (!memcmp("diff --git ", line, 11)) {
 576                         int git_hdr_len = parse_git_header(line, len, size, patch);
 577                         if (git_hdr_len <= len)
 578                                 continue;
 579                         if (!patch->old_name && !patch->new_name) {
 580                                 if (!patch->def_name)
 581                                         die("git diff header lacks filename information (line %d)", linenr);
 582                                 patch->old_name = patch->new_name = patch->def_name;
 583                         }
 584                         *hdrsize = git_hdr_len;
 585                         return offset;
 586                 }
 587
 588                 /** --- followed by +++ ? */
 589                 if (memcmp("--- ", line,  4) || memcmp("+++ ", line + len, 4))
 590                         continue;
 591
 592                 /*
 593                  * We only accept unified patches, so we want it to
 594                  * at least have "@@ -a,b +c,d @@\n", which is 14 chars
 595                  * minimum
 596                  */
 597                 nextlen = linelen(line + len, size - len);
 598                 if (size < nextlen + 14 || memcmp("@@ -", line + len + nextlen, 4))
 599                         continue;
 600
 601                 /* Ok, we'll consider it a patch */
 602                 parse_traditional_patch(line, line+len, patch);
 603                 *hdrsize = len + nextlen;
 604                 linenr += 2;
 605                 return offset;
 606         }
 607         return -1;
 608 }
 609
 610 /*
 611  * Parse a unified diff. Note that this really needs
 612  * to parse each fragment separately, since the only
 613  * way to know the difference between a "---" that is
 614  * part of a patch, and a "---" that starts the next
 615  * patch is to look at the line counts..
 616  */
 617 static int parse_fragment(char *line, unsigned long size, struct patch *patch, struct fragment *fragment)
 618 {
 619         int added, deleted;
 620         int len = linelen(line, size), offset;
 621         unsigned long oldlines, newlines;
 622
 623         offset = parse_fragment_header(line, len, fragment);
 624         if (offset < 0)
 625                 return -1;
 626         oldlines = fragment->oldlines;
 627         newlines = fragment->newlines;
 628
 629         if (patch->is_new < 0) {
 630                 patch->is_new =  !oldlines;
 631                 if (!oldlines)
 632                         patch->old_name = NULL;
 633         }
 634         if (patch->is_delete < 0) {
 635                 patch->is_delete = !newlines;
 636                 if (!newlines)
 637                         patch->new_name = NULL;
 638         }
 639
 640         if (patch->is_new != !oldlines)
 641                 return error("new file depends on old contents");
 642         if (patch->is_delete != !newlines) {
 643                 if (newlines)
 644                         return error("deleted file still has contents");
 645                 fprintf(stderr, "** warning: file %s becomes empty but is not deleted\n", patch->new_name);
 646         }
 647
 648         /* Parse the thing.. */
 649         line += len;
 650         size -= len;
 651         linenr++;
 652         added = deleted = 0;
 653         for (offset = len; size > 0; offset += len, size -= len, line += len, linenr++) {
 654                 if (!oldlines && !newlines)
 655                         break;
 656                 len = linelen(line, size);
 657                 if (!len || line[len-1] != '\n')
 658                         return -1;
 659                 switch (*line) {
 660                 default:
 661                         return -1;
 662                 case ' ':
 663                         oldlines--;
 664                         newlines--;
 665                         break;
 666                 case '-':
 667                         deleted++;
 668                         oldlines--;
 669                         break;
 670                 case '+':
 671                         added++;
 672                         newlines--;
 673                         break;
 674                 /* We allow "\ No newline at end of file" */
 675                 case '\\':
 676                         if (len < 12 || memcmp(line, "\\ No newline", 12))
 677                                 return -1;
 678                         break;
 679                 }
 680         }
 681         patch->lines_added += added;
 682         patch->lines_deleted += deleted;
 683         return offset;
 684 }
 685
 686 static int parse_single_patch(char *line, unsigned long size, struct patch *patch)
 687 {
 688         unsigned long offset = 0;
 689         struct fragment **fragp = &patch->fragments;
 690
 691         while (size > 4 && !memcmp(line, "@@ -", 4)) {
 692                 struct fragment *fragment;
 693                 int len;
 694
 695                 fragment = xmalloc(sizeof(*fragment));
 696                 memset(fragment, 0, sizeof(*fragment));
 697                 len = parse_fragment(line, size, patch, fragment);
 698                 if (len <= 0)
 699                         die("corrupt patch at line %d", linenr);
 700
 701                 fragment->patch = line;
 702                 fragment->size = len;
 703
 704                 *fragp = fragment;
 705                 fragp = &fragment->next;
 706
 707                 offset += len;
 708                 line += len;
 709                 size -= len;
 710         }
 711         return offset;
 712 }
 713
 714 static int parse_chunk(char *buffer, unsigned long size, struct patch *patch)
 715 {
 716         int hdrsize, patchsize;
 717         int offset = find_header(buffer, size, &hdrsize, patch);
 718
 719         if (offset < 0)
 720                 return offset;
 721
 722         patchsize = parse_single_patch(buffer + offset + hdrsize, size - offset - hdrsize, patch);
 723
 724         return offset + hdrsize + patchsize;
 725 }
 726
 727 static const char pluses[] = "++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++";
 728 static const char minuses[]= "----------------------------------------------------------------------";
 729
 730 static void show_stats(struct patch *patch)
 731 {
 732         char *name = patch->new_name;
 733         int len, max, add, del, total;
 734
 735         if (!name)
 736                 name = patch->old_name;
 737
 738         /*
 739          * "scale" the filename
 740          */
 741         len = strlen(name);
 742         max = max_len;
 743         if (max > 50)
 744                 max = 50;
 745         if (len > max)
 746                 name += len - max;
 747         len = max;
 748
 749         /*
 750          * scale the add/delete
 751          */
 752         max = max_change;
 753         if (max + len > 70)
 754                 max = 70 - len;
 755
 756         add = patch->lines_added;
 757         del = patch->lines_deleted;
 758         total = add + del;
 759
 760         if (max_change > 0) {
 761                 total = (total * max + max_change / 2) / max_change;
 762                 add = (add * max + max_change / 2) / max_change;
 763                 del = total - add;
 764         }
 765         printf(" %-*s |%5d %.*s%.*s\n",
 766                 len, name, patch->lines_added + patch->lines_deleted,
 767                 add, pluses, del, minuses);
 768 }
 769
 770 static int read_old_data(struct stat *st, const char *path, void *buf, unsigned long size)
 771 {
 772         int fd;
 773         unsigned long got;
 774
 775         switch (st->st_mode & S_IFMT) {
 776         case S_IFLNK:
 777                 return readlink(path, buf, size);
 778         case S_IFREG:
 779                 fd = open(path, O_RDONLY);
 780                 if (fd < 0)
 781                         return error("unable to open %s", path);
 782                 got = 0;
 783                 for (;;) {
 784                         int ret = read(fd, buf + got, size - got);
 785                         if (ret < 0) {
 786                                 if (errno == EAGAIN)
 787                                         continue;
 788                                 break;
 789                         }
 790                         if (!ret)
 791                                 break;
 792                         got += ret;
 793                 }
 794                 close(fd);
 795                 return got;
 796
 797         default:
 798                 return -1;
 799         }
 800 }
 801
 802 static int find_offset(const char *buf, unsigned long size, const char *fragment, unsigned long fragsize, int line)
 803 {
 804         int i;
 805         unsigned long start, backwards, forwards;
 806
 807         if (fragsize > size)
 808                 return -1;
 809
 810         start = 0;
 811         if (line > 1) {
 812                 unsigned long offset = 0;
 813                 i = line-1;
 814                 while (offset + fragsize <= size) {
 815                         if (buf[offset++] == '\n') {
 816                                 start = offset;
 817                                 if (!--i)
 818                                         break;
 819                         }
 820                 }
 821         }
 822
 823         /* Exact line number? */
 824         if (!memcmp(buf + start, fragment, fragsize))
 825                 return start;
 826
 827         /*
 828          * There's probably some smart way to do this, but I'll leave
 829          * that to the smart and beautiful people. I'm simple and stupid.
 830          */
 831         backwards = start;
 832         forwards = start;
 833         for (i = 0; ; i++) {
 834                 unsigned long try;
 835                 int n;
 836
 837                 /* "backward" */
 838                 if (i & 1) {
 839                         if (!backwards) {
 840                                 if (forwards + fragsize > size)
 841                                         break;
 842                                 continue;
 843                         }
 844                         do {
 845                                 --backwards;
 846                         } while (backwards && buf[backwards-1] != '\n');
 847                         try = backwards;
 848                 } else {
 849                         while (forwards + fragsize <= size) {
 850                                 if (buf[forwards++] == '\n')
 851                                         break;
 852                         }
 853                         try = forwards;
 854                 }
 855
 856                 if (try + fragsize > size)
 857                         continue;
 858                 if (memcmp(buf + try, fragment, fragsize))
 859                         continue;
 860                 n = (i >> 1)+1;
 861                 if (i & 1)
 862                         n = -n;
 863                 return try;
 864         }
 865
 866         /*
 867          * We should start searching forward and backward.
 868          */
 869         return -1;
 870 }
 871
 872 struct buffer_desc {
 873         char *buffer;
 874         unsigned long size;
 875         unsigned long alloc;
 876 };
 877
 878 static int apply_one_fragment(struct buffer_desc *desc, struct fragment *frag)
 879 {
 880         char *buf = desc->buffer;
 881         const char *patch = frag->patch;
 882         int offset, size = frag->size;
 883         char *old = xmalloc(size);
 884         char *new = xmalloc(size);
 885         int oldsize = 0, newsize = 0;
 886
 887         while (size > 0) {
 888                 int len = linelen(patch, size);
 889                 int plen;
 890
 891                 if (!len)
 892                         break;
 893
 894                 /*
 895                  * "plen" is how much of the line we should use for
 896                  * the actual patch data. Normally we just remove the
 897                  * first character on the line, but if the line is
 898                  * followed by "\ No newline", then we also remove the
 899                  * last one (which is the newline, of course).
 900                  */
 901                 plen = len-1;
 902                 if (len > size && patch[len] == '\\')
 903                         plen--;
 904                 switch (*patch) {
 905                 case ' ':
 906                 case '-':
 907                         memcpy(old + oldsize, patch + 1, plen);
 908                         oldsize += plen;
 909                         if (*patch == '-')
 910                                 break;
 911                 /* Fall-through for ' ' */
 912                 case '+':
 913                         memcpy(new + newsize, patch + 1, plen);
 914                         newsize += plen;
 915                         break;
 916                 case '@': case '\\':
 917                         /* Ignore it, we already handled it */
 918                         break;
 919                 default:
 920                         return -1;
 921                 }
 922                 patch += len;
 923                 size -= len;
 924         }
 925
 926         offset = find_offset(buf, desc->size, old, oldsize, frag->newpos);
 927         if (offset >= 0) {
 928                 int diff = newsize - oldsize;
 929                 unsigned long size = desc->size + diff;
 930                 unsigned long alloc = desc->alloc;
 931
 932                 if (size > alloc) {
 933                         alloc = size + 8192;
 934                         desc->alloc = alloc;
 935                         buf = xrealloc(buf, alloc);
 936                         desc->buffer = buf;
 937                 }
 938                 desc->size = size;
 939                 memmove(buf + offset + newsize, buf + offset + oldsize, size - offset - newsize);
 940                 memcpy(buf + offset, new, newsize);
 941                 offset = 0;
 942         }
 943
 944         free(old);
 945         free(new);
 946         return offset;
 947 }
 948
 949 static int apply_fragments(struct buffer_desc *desc, struct patch *patch)
 950 {
 951         struct fragment *frag = patch->fragments;
 952
 953         while (frag) {
 954                 if (apply_one_fragment(desc, frag) < 0)
 955                         return error("patch failed: %s:%d", patch->old_name, frag->oldpos);
 956                 frag = frag->next;
 957         }
 958         return 0;
 959 }
 960
 961 static int apply_data(struct patch *patch, struct stat *st)
 962 {
 963         char *buf;
 964         unsigned long size, alloc;
 965         struct buffer_desc desc;
 966
 967         size = 0;
 968         alloc = 0;
 969         buf = NULL;
 970         if (patch->old_name) {
 971                 size = st->st_size;
 972                 alloc = size + 8192;
 973                 buf = xmalloc(alloc);
 974                 if (read_old_data(st, patch->old_name, buf, alloc) != size)
 975                         return error("read of %s failed", patch->old_name);
 976         }
 977
 978         desc.size = size;
 979         desc.alloc = alloc;
 980         desc.buffer = buf;
 981         if (apply_fragments(&desc, patch) < 0)
 982                 return -1;
 983         patch->result = desc.buffer;
 984         patch->resultsize = desc.size;
 985
 986         if (patch->is_delete && patch->resultsize)
 987                 return error("removal patch leaves file contents");
 988
 989         return 0;
 990 }
 991
 992 static int check_patch(struct patch *patch)
 993 {
 994         struct stat st;
 995         const char *old_name = patch->old_name;
 996         const char *new_name = patch->new_name;
 997
 998         if (old_name) {
 999                 int changed;
1000
1001                 if (lstat(old_name, &st) < 0)
1002                         return error("%s: %s", old_name, strerror(errno));
1003                 if (check_index) {
1004                         int pos = cache_name_pos(old_name, strlen(old_name));
1005                         if (pos < 0)
1006                                 return error("%s: does not exist in index", old_name);
1007                         changed = ce_match_stat(active_cache[pos], &st);
1008                         if (changed)
1009                                 return error("%s: does not match index", old_name);
1010                 }
1011                 if (patch->is_new < 0)
1012                         patch->is_new = 0;
1013                 st.st_mode = ntohl(create_ce_mode(st.st_mode));
1014                 if (!patch->old_mode)
1015                         patch->old_mode = st.st_mode;
1016                 if ((st.st_mode ^ patch->old_mode) & S_IFMT)
1017                         return error("%s: wrong type", old_name);
1018                 if (st.st_mode != patch->old_mode)
1019                         fprintf(stderr, "warning: %s has type %o, expected %o\n",
1020                                 old_name, st.st_mode, patch->old_mode);
1021         }
1022
1023         if (new_name && (patch->is_new | patch->is_rename | patch->is_copy)) {
1024                 if (check_index && cache_name_pos(new_name, strlen(new_name)) >= 0)
1025                         return error("%s: already exists in index", new_name);
1026                 if (!lstat(new_name, &st))
1027                         return error("%s: already exists in working directory", new_name);
1028                 if (errno != ENOENT)
1029                         return error("%s: %s", new_name, strerror(errno));
1030                 if (!patch->new_mode)
1031                         patch->new_mode = S_IFREG | 0644;
1032         }
1033
1034         if (new_name && old_name) {
1035                 int same = !strcmp(old_name, new_name);
1036                 if (!patch->new_mode)
1037                         patch->new_mode = patch->old_mode;
1038                 if ((patch->old_mode ^ patch->new_mode) & S_IFMT)
1039                         return error("new mode (%o) of %s does not match old mode (%o)%s%s",
1040                                 patch->new_mode, new_name, patch->old_mode,
1041                                 same ? "" : " of ", same ? "" : old_name);
1042         }
1043
1044         if (apply_data(patch, &st) < 0)
1045                 return error("%s: patch does not apply", old_name);
1046         return 0;
1047 }
1048
1049 static int check_patch_list(struct patch *patch)
1050 {
1051         int error = 0;
1052
1053         for (;patch ; patch = patch->next)
1054                 error |= check_patch(patch);
1055         return error;
1056 }
1057
1058 static void show_file(int c, unsigned int mode, const char *name)
1059 {
1060         printf("%c %o %s\n", c, mode, name);
1061 }
1062
1063 static void show_file_list(struct patch *patch)
1064 {
1065         for (;patch ; patch = patch->next) {
1066                 if (patch->is_rename) {
1067                         show_file('-', patch->old_mode, patch->old_name);
1068                         show_file('+', patch->new_mode, patch->new_name);
1069                         continue;
1070                 }
1071                 if (patch->is_copy || patch->is_new) {
1072                         show_file('+', patch->new_mode, patch->new_name);
1073                         continue;
1074                 }
1075                 if (patch->is_delete) {
1076                         show_file('-', patch->old_mode, patch->old_name);
1077                         continue;
1078                 }
1079                 if (patch->old_mode && patch->new_mode && patch->old_mode != patch->new_mode) {
1080                         printf("M %o:%o %s\n", patch->old_mode, patch->new_mode, patch->old_name);
1081                         continue;
1082                 }
1083                 printf("M %o %s\n", patch->old_mode, patch->old_name);
1084         }
1085 }
1086
1087 static void stat_patch_list(struct patch *patch)
1088 {
1089         int files, adds, dels;
1090
1091         for (files = adds = dels = 0 ; patch ; patch = patch->next) {
1092                 files++;
1093                 adds += patch->lines_added;
1094                 dels += patch->lines_deleted;
1095                 show_stats(patch);
1096         }
1097
1098         printf(" %d files changed, %d insertions(+), %d deletions(-)\n", files, adds, dels);
1099 }
1100
1101 static void show_file_mode_name(const char *newdelete, unsigned int mode, const char *name)
1102 {
1103         if (mode)
1104                 printf(" %s mode %06o %s\n", newdelete, mode, name);
1105         else
1106                 printf(" %s %s\n", newdelete, name);
1107 }
1108
1109 static void show_mode_change(struct patch *p, int show_name)
1110 {
1111         if (p->old_mode && p->new_mode && p->old_mode != p->new_mode) {
1112                 if (show_name)
1113                         printf(" mode change %06o => %06o %s\n",
1114                                p->old_mode, p->new_mode, p->new_name);
1115                 else
1116                         printf(" mode change %06o => %06o\n",
1117                                p->old_mode, p->new_mode);
1118         }
1119 }
1120
1121 static void show_rename_copy(struct patch *p)
1122 {
1123         const char *renamecopy = p->is_rename ? "rename" : "copy";
1124         const char *old, *new;
1125
1126         /* Find common prefix */
1127         old = p->old_name;
1128         new = p->new_name;
1129         while (1) {
1130                 const char *slash_old, *slash_new;
1131                 slash_old = strchr(old, '/');
1132                 slash_new = strchr(new, '/');
1133                 if (!slash_old ||
1134                     !slash_new ||
1135                     slash_old - old != slash_new - new ||
1136                     memcmp(old, new, slash_new - new))
1137                         break;
1138                 old = slash_old + 1;
1139                 new = slash_new + 1;
1140         }
1141         /* p->old_name thru old is the common prefix, and old and new
1142          * through the end of names are renames
1143          */
1144         if (old != p->old_name)
1145                 printf(" %s %.*s{%s => %s} (%d%%)\n", renamecopy,
1146                        old - p->old_name, p->old_name,
1147                        old, new, p->score);
1148         else
1149                 printf(" %s %s => %s (%d%%)\n", renamecopy,
1150                        p->old_name, p->new_name, p->score);
1151         show_mode_change(p, 0);
1152 }
1153
1154 static void summary_patch_list(struct patch *patch)
1155 {
1156         struct patch *p;
1157
1158         for (p = patch; p; p = p->next) {
1159                 if (p->is_new)
1160                         show_file_mode_name("create", p->new_mode, p->new_name);
1161                 else if (p->is_delete)
1162                         show_file_mode_name("delete", p->old_mode, p->old_name);
1163                 else {
1164                         if (p->is_rename || p->is_copy)
1165                                 show_rename_copy(p);
1166                         else {
1167                                 if (p->score) {
1168                                         printf(" rewrite %s (%d%%)\n",
1169                                                p->new_name, p->score);
1170                                         show_mode_change(p, 0);
1171                                 }
1172                                 else
1173                                         show_mode_change(p, 1);
1174                         }
1175                 }
1176         }
1177 }
1178
1179 static void patch_stats(struct patch *patch)
1180 {
1181         int lines = patch->lines_added + patch->lines_deleted;
1182
1183         if (lines > max_change)
1184                 max_change = lines;
1185         if (patch->old_name) {
1186                 int len = strlen(patch->old_name);
1187                 if (len > max_len)
1188                         max_len = len;
1189         }
1190         if (patch->new_name) {
1191                 int len = strlen(patch->new_name);
1192                 if (len > max_len)
1193                         max_len = len;
1194         }
1195 }
1196
1197 static void remove_file(struct patch *patch)
1198 {
1199         if (write_index) {
1200                 if (remove_file_from_cache(patch->old_name) < 0)
1201                         die("unable to remove %s from index", patch->old_name);
1202         }
1203         unlink(patch->old_name);
1204 }
1205
1206 static void add_index_file(const char *path, unsigned mode, void *buf, unsigned long size)
1207 {
1208         struct stat st;
1209         struct cache_entry *ce;
1210         int namelen = strlen(path);
1211         unsigned ce_size = cache_entry_size(namelen);
1212
1213         if (!write_index)
1214                 return;
1215
1216         ce = xmalloc(ce_size);
1217         memset(ce, 0, ce_size);
1218         memcpy(ce->name, path, namelen);
1219         ce->ce_mode = create_ce_mode(mode);
1220         ce->ce_flags = htons(namelen);
1221         if (lstat(path, &st) < 0)
1222                 die("unable to stat newly created file %s", path);
1223         fill_stat_cache_info(ce, &st);
1224         if (write_sha1_file(buf, size, "blob", ce->sha1) < 0)
1225                 die("unable to create backing store for newly created file %s", path);
1226         if (add_cache_entry(ce, ADD_CACHE_OK_TO_ADD) < 0)
1227                 die("unable to add cache entry for %s", path);
1228 }
1229
1230 static void create_subdirectories(const char *path)
1231 {
1232         int len = strlen(path);
1233         char *buf = xmalloc(len + 1);
1234         const char *slash = path;
1235
1236         while ((slash = strchr(slash+1, '/')) != NULL) {
1237                 len = slash - path;
1238                 memcpy(buf, path, len);
1239                 buf[len] = 0;
1240                 if (mkdir(buf, 0755) < 0) {
1241                         if (errno != EEXIST)
1242                                 break;
1243                 }
1244         }
1245         free(buf);
1246 }
1247
1248 /*
1249  * We optimistically assume that the directories exist,
1250  * which is true 99% of the time anyway. If they don't,
1251  * we create them and try again.
1252  */
1253 static int create_regular_file(const char *path, unsigned int mode)
1254 {
1255         int ret = open(path, O_WRONLY | O_CREAT | O_TRUNC, mode);
1256
1257         if (ret < 0 && errno == ENOENT) {
1258                 create_subdirectories(path);
1259                 ret = open(path, O_WRONLY | O_CREAT | O_TRUNC, mode);
1260         }
1261         return ret;
1262 }
1263
1264 static int create_symlink(const char *buf, const char *path)
1265 {
1266         int ret = symlink(buf, path);
1267
1268         if (ret < 0 && errno == ENOENT) {
1269                 create_subdirectories(path);
1270                 ret = symlink(buf, path);
1271         }
1272         return ret;
1273 }
1274
1275 static void create_file(struct patch *patch)
1276 {
1277         const char *path = patch->new_name;
1278         unsigned mode = patch->new_mode;
1279         unsigned long size = patch->resultsize;
1280         char *buf = patch->result;
1281
1282         if (!mode)
1283                 mode = S_IFREG | 0644;
1284         if (S_ISREG(mode)) {
1285                 int fd;
1286                 mode = (mode & 0100) ? 0777 : 0666;
1287                 fd = create_regular_file(path, mode);
1288                 if (fd < 0)
1289                         die("unable to create file %s (%s)", path, strerror(errno));
1290                 if (write(fd, buf, size) != size)
1291                         die("unable to write file %s", path);
1292                 close(fd);
1293                 add_index_file(path, mode, buf, size);
1294                 return;
1295         }
1296         if (S_ISLNK(mode)) {
1297                 if (size && buf[size-1] == '\n')
1298                         size--;
1299                 buf[size] = 0;
1300                 if (create_symlink(buf, path) < 0)
1301                         die("unable to write symlink %s", path);
1302                 add_index_file(path, mode, buf, size);
1303                 return;
1304         }
1305         die("unable to write file mode %o", mode);
1306 }
1307
1308 static void write_out_one_result(struct patch *patch)
1309 {
1310         if (patch->is_delete > 0) {
1311                 remove_file(patch);
1312                 return;
1313         }
1314         if (patch->is_new > 0 || patch->is_copy) {
1315                 create_file(patch);
1316                 return;
1317         }
1318         /*
1319          * Rename or modification boils down to the same
1320          * thing: remove the old, write the new
1321          */
1322         remove_file(patch);
1323         create_file(patch);
1324 }
1325
1326 static void write_out_results(struct patch *list)
1327 {
1328         if (!list)
1329                 die("No changes");
1330
1331         while (list) {
1332                 write_out_one_result(list);
1333                 list = list->next;
1334         }
1335 }
1336
1337 static struct cache_file cache_file;
1338
1339 static int apply_patch(int fd)
1340 {
1341         int newfd;
1342         unsigned long offset, size;
1343         char *buffer = read_patch_file(fd, &size);
1344         struct patch *list = NULL, **listp = &list;
1345
1346         if (!buffer)
1347                 return -1;
1348         offset = 0;
1349         while (size > 0) {
1350                 struct patch *patch;
1351                 int nr;
1352
1353                 patch = xmalloc(sizeof(*patch));
1354                 memset(patch, 0, sizeof(*patch));
1355                 nr = parse_chunk(buffer + offset, size, patch);
1356                 if (nr < 0)
1357                         break;
1358                 patch_stats(patch);
1359                 *listp = patch;
1360                 listp = &patch->next;
1361                 offset += nr;
1362                 size -= nr;
1363         }
1364
1365         newfd = -1;
1366         write_index = check_index && apply;
1367         if (write_index)
1368                 newfd = hold_index_file_for_update(&cache_file, get_index_file());
1369         if (check_index) {
1370                 if (read_cache() < 0)
1371                         die("unable to read index file");
1372         }
1373
1374         if ((check || apply) && check_patch_list(list) < 0)
1375                 exit(1);
1376
1377         if (apply)
1378                 write_out_results(list);
1379
1380         if (write_index) {
1381                 if (write_cache(newfd, active_cache, active_nr) ||
1382                     commit_index_file(&cache_file))
1383                         die("Unable to write new cachefile");
1384         }
1385
1386         if (show_files)
1387                 show_file_list(list);
1388
1389         if (diffstat)
1390                 stat_patch_list(list);
1391
1392         if (summary)
1393                 summary_patch_list(list);
1394
1395         free(buffer);
1396         return 0;
1397 }
1398
1399 int main(int argc, char **argv)
1400 {
1401         int i;
1402         int read_stdin = 1;
1403
1404         for (i = 1; i < argc; i++) {
1405                 const char *arg = argv[i];
1406                 int fd;
1407
1408                 if (!strcmp(arg, "-")) {
1409                         apply_patch(0);
1410                         read_stdin = 0;
1411                         continue;
1412                 }
1413                 if (!strcmp(arg, "--no-merge")) {
1414                         merge_patch = 0;
1415                         continue;
1416                 }
1417                 if (!strcmp(arg, "--stat")) {
1418                         apply = 0;
1419                         diffstat = 1;
1420                         continue;
1421                 }
1422                 if (!strcmp(arg, "--summary")) {
1423                         apply = 0;
1424                         summary = 1;
1425                         continue;
1426                 }
1427                 if (!strcmp(arg, "--check")) {
1428                         apply = 0;
1429                         check = 1;
1430                         continue;
1431                 }
1432                 if (!strcmp(arg, "--index")) {
1433                         check_index = 1;
1434                         continue;
1435                 }
1436                 if (!strcmp(arg, "--apply")) {
1437                         apply = 1;
1438                         continue;
1439                 }
1440                 if (!strcmp(arg, "--show-files")) {
1441                         show_files = 1;
1442                         continue;
1443                 }
1444                 fd = open(arg, O_RDONLY);
1445                 if (fd < 0)
1446                         usage(apply_usage);
1447                 read_stdin = 0;
1448                 apply_patch(fd);
1449                 close(fd);
1450         }
1451         if (read_stdin)
1452                 apply_patch(0);
1453         return 0;
1454 }