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[] =
  36 "git-apply [--no-merge] [--stat] [--summary] [--check] [--index] [--apply] [--show-files] <patch>...";
  37
  38 /*
  39  * For "diff-stat" like behaviour, we keep track of the biggest change
  40  * we've seen, and the longest filename. That allows us to do simple
  41  * scaling.
  42  */
  43 static int max_change, max_len;
  44
  45 /*
  46  * Various "current state", notably line numbers and what
  47  * file (and how) we're patching right now.. The "is_xxxx"
  48  * things are flags, where -1 means "don't know yet".
  49  */
  50 static int linenr = 1;
  51
  52 struct fragment {
  53         unsigned long oldpos, oldlines;
  54         unsigned long newpos, newlines;
  55         const char *patch;
  56         int size;
  57         struct fragment *next;
  58 };
  59
  60 struct patch {
  61         char *new_name, *old_name, *def_name;
  62         unsigned int old_mode, new_mode;
  63         int is_rename, is_copy, is_new, is_delete;
  64         int lines_added, lines_deleted;
  65         int score;
  66         struct fragment *fragments;
  67         char *result;
  68         unsigned long resultsize;
  69         struct patch *next;
  70 };
  71
  72 #define CHUNKSIZE (8192)
  73 #define SLOP (16)
  74
  75 static void *read_patch_file(int fd, unsigned long *sizep)
  76 {
  77         unsigned long size = 0, alloc = CHUNKSIZE;
  78         void *buffer = xmalloc(alloc);
  79
  80         for (;;) {
  81                 int nr = alloc - size;
  82                 if (nr < 1024) {
  83                         alloc += CHUNKSIZE;
  84                         buffer = xrealloc(buffer, alloc);
  85                         nr = alloc - size;
  86                 }
  87                 nr = read(fd, buffer + size, nr);
  88                 if (!nr)
  89                         break;
  90                 if (nr < 0) {
  91                         if (errno == EAGAIN)
  92                                 continue;
  93                         die("git-apply: read returned %s", strerror(errno));
  94                 }
  95                 size += nr;
  96         }
  97         *sizep = size;
  98
  99         /*
 100          * Make sure that we have some slop in the buffer
 101          * so that we can do speculative "memcmp" etc, and
 102          * see to it that it is NUL-filled.
 103          */
 104         if (alloc < size + SLOP)
 105                 buffer = xrealloc(buffer, size + SLOP);
 106         memset(buffer + size, 0, SLOP);
 107         return buffer;
 108 }
 109
 110 static unsigned long linelen(const char *buffer, unsigned long size)
 111 {
 112         unsigned long len = 0;
 113         while (size--) {
 114                 len++;
 115                 if (*buffer++ == '\n')
 116                         break;
 117         }
 118         return len;
 119 }
 120
 121 static int is_dev_null(const char *str)
 122 {
 123         return !memcmp("/dev/null", str, 9) && isspace(str[9]);
 124 }
 125
 126 #define TERM_SPACE      1
 127 #define TERM_TAB        2
 128
 129 static int name_terminate(const char *name, int namelen, int c, int terminate)
 130 {
 131         if (c == ' ' && !(terminate & TERM_SPACE))
 132                 return 0;
 133         if (c == '\t' && !(terminate & TERM_TAB))
 134                 return 0;
 135
 136         return 1;
 137 }
 138
 139 static char * find_name(const char *line, char *def, int p_value, int terminate)
 140 {
 141         int len;
 142         const char *start = line;
 143         char *name;
 144
 145         for (;;) {
 146                 char c = *line;
 147
 148                 if (isspace(c)) {
 149                         if (c == '\n')
 150                                 break;
 151                         if (name_terminate(start, line-start, c, terminate))
 152                                 break;
 153                 }
 154                 line++;
 155                 if (c == '/' && !--p_value)
 156                         start = line;
 157         }
 158         if (!start)
 159                 return def;
 160         len = line - start;
 161         if (!len)
 162                 return def;
 163
 164         /*
 165          * Generally we prefer the shorter name, especially
 166          * if the other one is just a variation of that with
 167          * something else tacked on to the end (ie "file.orig"
 168          * or "file~").
 169          */
 170         if (def) {
 171                 int deflen = strlen(def);
 172                 if (deflen < len && !strncmp(start, def, deflen))
 173                         return def;
 174         }
 175
 176         name = xmalloc(len + 1);
 177         memcpy(name, start, len);
 178         name[len] = 0;
 179         free(def);
 180         return name;
 181 }
 182
 183 /*
 184  * Get the name etc info from the --/+++ lines of a traditional patch header
 185  *
 186  * NOTE! This hardcodes "-p1" behaviour in filename detection.
 187  *
 188  * FIXME! The end-of-filename heuristics are kind of screwy. For existing
 189  * files, we can happily check the index for a match, but for creating a
 190  * new file we should try to match whatever "patch" does. I have no idea.
 191  */
 192 static void parse_traditional_patch(const char *first, const char *second, struct patch *patch)
 193 {
 194         int p_value = 1;
 195         char *name;
 196
 197         first += 4;     // skip "--- "
 198         second += 4;    // skip "+++ "
 199         if (is_dev_null(first)) {
 200                 patch->is_new = 1;
 201                 patch->is_delete = 0;
 202                 name = find_name(second, NULL, p_value, TERM_SPACE | TERM_TAB);
 203                 patch->new_name = name;
 204         } else if (is_dev_null(second)) {
 205                 patch->is_new = 0;
 206                 patch->is_delete = 1;
 207                 name = find_name(first, NULL, p_value, TERM_SPACE | TERM_TAB);
 208                 patch->old_name = name;
 209         } else {
 210                 name = find_name(first, NULL, p_value, TERM_SPACE | TERM_TAB);
 211                 name = find_name(second, name, p_value, TERM_SPACE | TERM_TAB);
 212                 patch->old_name = patch->new_name = name;
 213         }
 214         if (!name)
 215                 die("unable to find filename in patch at line %d", linenr);
 216 }
 217
 218 static int gitdiff_hdrend(const char *line, struct patch *patch)
 219 {
 220         return -1;
 221 }
 222
 223 /*
 224  * We're anal about diff header consistency, to make
 225  * sure that we don't end up having strange ambiguous
 226  * patches floating around.
 227  *
 228  * As a result, gitdiff_{old|new}name() will check
 229  * their names against any previous information, just
 230  * to make sure..
 231  */
 232 static char *gitdiff_verify_name(const char *line, int isnull, char *orig_name, const char *oldnew)
 233 {
 234         int len;
 235         const char *name;
 236
 237         if (!orig_name && !isnull)
 238                 return find_name(line, NULL, 1, 0);
 239
 240         name = "/dev/null";
 241         len = 9;
 242         if (orig_name) {
 243                 name = orig_name;
 244                 len = strlen(name);
 245                 if (isnull)
 246                         die("git-apply: bad git-diff - expected /dev/null, got %s on line %d", name, linenr);
 247         }
 248
 249         if (*name == '/')
 250                 goto absolute_path;
 251
 252         for (;;) {
 253                 char c = *line++;
 254                 if (c == '\n')
 255                         break;
 256                 if (c != '/')
 257                         continue;
 258 absolute_path:
 259                 if (memcmp(line, name, len) || line[len] != '\n')
 260                         break;
 261                 return orig_name;
 262         }
 263         die("git-apply: bad git-diff - inconsistent %s filename on line %d", oldnew, linenr);
 264         return NULL;
 265 }
 266
 267 static int gitdiff_oldname(const char *line, struct patch *patch)
 268 {
 269         patch->old_name = gitdiff_verify_name(line, patch->is_new, patch->old_name, "old");
 270         return 0;
 271 }
 272
 273 static int gitdiff_newname(const char *line, struct patch *patch)
 274 {
 275         patch->new_name = gitdiff_verify_name(line, patch->is_delete, patch->new_name, "new");
 276         return 0;
 277 }
 278
 279 static int gitdiff_oldmode(const char *line, struct patch *patch)
 280 {
 281         patch->old_mode = strtoul(line, NULL, 8);
 282         return 0;
 283 }
 284
 285 static int gitdiff_newmode(const char *line, struct patch *patch)
 286 {
 287         patch->new_mode = strtoul(line, NULL, 8);
 288         return 0;
 289 }
 290
 291 static int gitdiff_delete(const char *line, struct patch *patch)
 292 {
 293         patch->is_delete = 1;
 294         patch->old_name = patch->def_name;
 295         return gitdiff_oldmode(line, patch);
 296 }
 297
 298 static int gitdiff_newfile(const char *line, struct patch *patch)
 299 {
 300         patch->is_new = 1;
 301         patch->new_name = patch->def_name;
 302         return gitdiff_newmode(line, patch);
 303 }
 304
 305 static int gitdiff_copysrc(const char *line, struct patch *patch)
 306 {
 307         patch->is_copy = 1;
 308         patch->old_name = find_name(line, NULL, 0, 0);
 309         return 0;
 310 }
 311
 312 static int gitdiff_copydst(const char *line, struct patch *patch)
 313 {
 314         patch->is_copy = 1;
 315         patch->new_name = find_name(line, NULL, 0, 0);
 316         return 0;
 317 }
 318
 319 static int gitdiff_renamesrc(const char *line, struct patch *patch)
 320 {
 321         patch->is_rename = 1;
 322         patch->old_name = find_name(line, NULL, 0, 0);
 323         return 0;
 324 }
 325
 326 static int gitdiff_renamedst(const char *line, struct patch *patch)
 327 {
 328         patch->is_rename = 1;
 329         patch->new_name = find_name(line, NULL, 0, 0);
 330         return 0;
 331 }
 332
 333 static int gitdiff_similarity(const char *line, struct patch *patch)
 334 {
 335         if ((patch->score = strtoul(line, NULL, 10)) == ULONG_MAX)
 336                 patch->score = 0;
 337         return 0;
 338 }
 339
 340 static int gitdiff_dissimilarity(const char *line, struct patch *patch)
 341 {
 342         if ((patch->score = strtoul(line, NULL, 10)) == ULONG_MAX)
 343                 patch->score = 0;
 344         return 0;
 345 }
 346
 347 /*
 348  * This is normal for a diff that doesn't change anything: we'll fall through
 349  * into the next diff. Tell the parser to break out.
 350  */
 351 static int gitdiff_unrecognized(const char *line, struct patch *patch)
 352 {
 353         return -1;
 354 }
 355
 356 static char *git_header_name(char *line)
 357 {
 358         int len;
 359         char *name, *second;
 360
 361         /*
 362          * Find the first '/'
 363          */
 364         name = line;
 365         for (;;) {
 366                 char c = *name++;
 367                 if (c == '\n')
 368                         return NULL;
 369                 if (c == '/')
 370                         break;
 371         }
 372
 373         /*
 374          * We don't accept absolute paths (/dev/null) as possibly valid
 375          */
 376         if (name == line+1)
 377                 return NULL;
 378
 379         /*
 380          * Accept a name only if it shows up twice, exactly the same
 381          * form.
 382          */
 383         for (len = 0 ; ; len++) {
 384                 char c = name[len];
 385
 386                 switch (c) {
 387                 default:
 388                         continue;
 389                 case '\n':
 390                         break;
 391                 case '\t': case ' ':
 392                         second = name+len;
 393                         for (;;) {
 394                                 char c = *second++;
 395                                 if (c == '\n')
 396                                         return NULL;
 397                                 if (c == '/')
 398                                         break;
 399                         }
 400                         if (second[len] == '\n' && !memcmp(name, second, len)) {
 401                                 char *ret = xmalloc(len + 1);
 402                                 memcpy(ret, name, len);
 403                                 ret[len] = 0;
 404                                 return ret;
 405                         }
 406                 }
 407         }
 408         return NULL;
 409 }
 410
 411 /* Verify that we recognize the lines following a git header */
 412 static int parse_git_header(char *line, int len, unsigned int size, struct patch *patch)
 413 {
 414         unsigned long offset;
 415
 416         /* A git diff has explicit new/delete information, so we don't guess */
 417         patch->is_new = 0;
 418         patch->is_delete = 0;
 419
 420         /*
 421          * Some things may not have the old name in the
 422          * rest of the headers anywhere (pure mode changes,
 423          * or removing or adding empty files), so we get
 424          * the default name from the header.
 425          */
 426         patch->def_name = git_header_name(line + strlen("diff --git "));
 427
 428         line += len;
 429         size -= len;
 430         linenr++;
 431         for (offset = len ; size > 0 ; offset += len, size -= len, line += len, linenr++) {
 432                 static const struct opentry {
 433                         const char *str;
 434                         int (*fn)(const char *, struct patch *);
 435                 } optable[] = {
 436                         { "@@ -", gitdiff_hdrend },
 437                         { "--- ", gitdiff_oldname },
 438                         { "+++ ", gitdiff_newname },
 439                         { "old mode ", gitdiff_oldmode },
 440                         { "new mode ", gitdiff_newmode },
 441                         { "deleted file mode ", gitdiff_delete },
 442                         { "new file mode ", gitdiff_newfile },
 443                         { "copy from ", gitdiff_copysrc },
 444                         { "copy to ", gitdiff_copydst },
 445                         { "rename old ", gitdiff_renamesrc },
 446                         { "rename new ", gitdiff_renamedst },
 447                         { "rename from ", gitdiff_renamesrc },
 448                         { "rename to ", gitdiff_renamedst },
 449                         { "similarity index ", gitdiff_similarity },
 450                         { "dissimilarity index ", gitdiff_dissimilarity },
 451                         { "", gitdiff_unrecognized },
 452                 };
 453                 int i;
 454
 455                 len = linelen(line, size);
 456                 if (!len || line[len-1] != '\n')
 457                         break;
 458                 for (i = 0; i < sizeof(optable) / sizeof(optable[0]); i++) {
 459                         const struct opentry *p = optable + i;
 460                         int oplen = strlen(p->str);
 461                         if (len < oplen || memcmp(p->str, line, oplen))
 462                                 continue;
 463                         if (p->fn(line + oplen, patch) < 0)
 464                                 return offset;
 465                         break;
 466                 }
 467         }
 468
 469         return offset;
 470 }
 471
 472 static int parse_num(const char *line, unsigned long *p)
 473 {
 474         char *ptr;
 475
 476         if (!isdigit(*line))
 477                 return 0;
 478         *p = strtoul(line, &ptr, 10);
 479         return ptr - line;
 480 }
 481
 482 static int parse_range(const char *line, int len, int offset, const char *expect,
 483                         unsigned long *p1, unsigned long *p2)
 484 {
 485         int digits, ex;
 486
 487         if (offset < 0 || offset >= len)
 488                 return -1;
 489         line += offset;
 490         len -= offset;
 491
 492         digits = parse_num(line, p1);
 493         if (!digits)
 494                 return -1;
 495
 496         offset += digits;
 497         line += digits;
 498         len -= digits;
 499
 500         *p2 = *p1;
 501         if (*line == ',') {
 502                 digits = parse_num(line+1, p2);
 503                 if (!digits)
 504                         return -1;
 505
 506                 offset += digits+1;
 507                 line += digits+1;
 508                 len -= digits+1;
 509         }
 510
 511         ex = strlen(expect);
 512         if (ex > len)
 513                 return -1;
 514         if (memcmp(line, expect, ex))
 515                 return -1;
 516
 517         return offset + ex;
 518 }
 519
 520 /*
 521  * Parse a unified diff fragment header of the
 522  * form "@@ -a,b +c,d @@"
 523  */
 524 static int parse_fragment_header(char *line, int len, struct fragment *fragment)
 525 {
 526         int offset;
 527
 528         if (!len || line[len-1] != '\n')
 529                 return -1;
 530
 531         /* Figure out the number of lines in a fragment */
 532         offset = parse_range(line, len, 4, " +", &fragment->oldpos, &fragment->oldlines);
 533         offset = parse_range(line, len, offset, " @@", &fragment->newpos, &fragment->newlines);
 534
 535         return offset;
 536 }
 537
 538 static int find_header(char *line, unsigned long size, int *hdrsize, struct patch *patch)
 539 {
 540         unsigned long offset, len;
 541
 542         patch->is_rename = patch->is_copy = 0;
 543         patch->is_new = patch->is_delete = -1;
 544         patch->old_mode = patch->new_mode = 0;
 545         patch->old_name = patch->new_name = NULL;
 546         for (offset = 0; size > 0; offset += len, size -= len, line += len, linenr++) {
 547                 unsigned long nextlen;
 548
 549                 len = linelen(line, size);
 550                 if (!len)
 551                         break;
 552
 553                 /* Testing this early allows us to take a few shortcuts.. */
 554                 if (len < 6)
 555                         continue;
 556
 557                 /*
 558                  * Make sure we don't find any unconnected patch fragmants.
 559                  * That's a sign that we didn't find a header, and that a
 560                  * patch has become corrupted/broken up.
 561                  */
 562                 if (!memcmp("@@ -", line, 4)) {
 563                         struct fragment dummy;
 564                         if (parse_fragment_header(line, len, &dummy) < 0)
 565                                 continue;
 566                         error("patch fragment without header at line %d: %.*s", linenr, len-1, line);
 567                 }
 568
 569                 if (size < len + 6)
 570                         break;
 571
 572                 /*
 573                  * Git patch? It might not have a real patch, just a rename
 574                  * or mode change, so we handle that specially
 575                  */
 576                 if (!memcmp("diff --git ", line, 11)) {
 577                         int git_hdr_len = parse_git_header(line, len, size, patch);
 578                         if (git_hdr_len <= len)
 579                                 continue;
 580                         if (!patch->old_name && !patch->new_name) {
 581                                 if (!patch->def_name)
 582                                         die("git diff header lacks filename information (line %d)", linenr);
 583                                 patch->old_name = patch->new_name = patch->def_name;
 584                         }
 585                         *hdrsize = git_hdr_len;
 586                         return offset;
 587                 }
 588
 589                 /** --- followed by +++ ? */
 590                 if (memcmp("--- ", line,  4) || memcmp("+++ ", line + len, 4))
 591                         continue;
 592
 593                 /*
 594                  * We only accept unified patches, so we want it to
 595                  * at least have "@@ -a,b +c,d @@\n", which is 14 chars
 596                  * minimum
 597                  */
 598                 nextlen = linelen(line + len, size - len);
 599                 if (size < nextlen + 14 || memcmp("@@ -", line + len + nextlen, 4))
 600                         continue;
 601
 602                 /* Ok, we'll consider it a patch */
 603                 parse_traditional_patch(line, line+len, patch);
 604                 *hdrsize = len + nextlen;
 605                 linenr += 2;
 606                 return offset;
 607         }
 608         return -1;
 609 }
 610
 611 /*
 612  * Parse a unified diff. Note that this really needs
 613  * to parse each fragment separately, since the only
 614  * way to know the difference between a "---" that is
 615  * part of a patch, and a "---" that starts the next
 616  * patch is to look at the line counts..
 617  */
 618 static int parse_fragment(char *line, unsigned long size, struct patch *patch, struct fragment *fragment)
 619 {
 620         int added, deleted;
 621         int len = linelen(line, size), offset;
 622         unsigned long oldlines, newlines;
 623
 624         offset = parse_fragment_header(line, len, fragment);
 625         if (offset < 0)
 626                 return -1;
 627         oldlines = fragment->oldlines;
 628         newlines = fragment->newlines;
 629
 630         if (patch->is_new < 0) {
 631                 patch->is_new =  !oldlines;
 632                 if (!oldlines)
 633                         patch->old_name = NULL;
 634         }
 635         if (patch->is_delete < 0) {
 636                 patch->is_delete = !newlines;
 637                 if (!newlines)
 638                         patch->new_name = NULL;
 639         }
 640
 641         if (patch->is_new != !oldlines)
 642                 return error("new file depends on old contents");
 643         if (patch->is_delete != !newlines) {
 644                 if (newlines)
 645                         return error("deleted file still has contents");
 646                 fprintf(stderr, "** warning: file %s becomes empty but is not deleted\n", patch->new_name);
 647         }
 648
 649         /* Parse the thing.. */
 650         line += len;
 651         size -= len;
 652         linenr++;
 653         added = deleted = 0;
 654         for (offset = len; size > 0; offset += len, size -= len, line += len, linenr++) {
 655                 if (!oldlines && !newlines)
 656                         break;
 657                 len = linelen(line, size);
 658                 if (!len || line[len-1] != '\n')
 659                         return -1;
 660                 switch (*line) {
 661                 default:
 662                         return -1;
 663                 case ' ':
 664                         oldlines--;
 665                         newlines--;
 666                         break;
 667                 case '-':
 668                         deleted++;
 669                         oldlines--;
 670                         break;
 671                 case '+':
 672                         added++;
 673                         newlines--;
 674                         break;
 675                 /* We allow "\ No newline at end of file" */
 676                 case '\\':
 677                         if (len < 12 || memcmp(line, "\\ No newline", 12))
 678                                 return -1;
 679                         break;
 680                 }
 681         }
 682         patch->lines_added += added;
 683         patch->lines_deleted += deleted;
 684         return offset;
 685 }
 686
 687 static int parse_single_patch(char *line, unsigned long size, struct patch *patch)
 688 {
 689         unsigned long offset = 0;
 690         struct fragment **fragp = &patch->fragments;
 691
 692         while (size > 4 && !memcmp(line, "@@ -", 4)) {
 693                 struct fragment *fragment;
 694                 int len;
 695
 696                 fragment = xmalloc(sizeof(*fragment));
 697                 memset(fragment, 0, sizeof(*fragment));
 698                 len = parse_fragment(line, size, patch, fragment);
 699                 if (len <= 0)
 700                         die("corrupt patch at line %d", linenr);
 701
 702                 fragment->patch = line;
 703                 fragment->size = len;
 704
 705                 *fragp = fragment;
 706                 fragp = &fragment->next;
 707
 708                 offset += len;
 709                 line += len;
 710                 size -= len;
 711         }
 712         return offset;
 713 }
 714
 715 static int parse_chunk(char *buffer, unsigned long size, struct patch *patch)
 716 {
 717         int hdrsize, patchsize;
 718         int offset = find_header(buffer, size, &hdrsize, patch);
 719
 720         if (offset < 0)
 721                 return offset;
 722
 723         patchsize = parse_single_patch(buffer + offset + hdrsize, size - offset - hdrsize, patch);
 724
 725         return offset + hdrsize + patchsize;
 726 }
 727
 728 static const char pluses[] = "++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++";
 729 static const char minuses[]= "----------------------------------------------------------------------";
 730
 731 static void show_stats(struct patch *patch)
 732 {
 733         char *name = patch->new_name;
 734         int len, max, add, del, total;
 735
 736         if (!name)
 737                 name = patch->old_name;
 738
 739         /*
 740          * "scale" the filename
 741          */
 742         len = strlen(name);
 743         max = max_len;
 744         if (max > 50)
 745                 max = 50;
 746         if (len > max)
 747                 name += len - max;
 748         len = max;
 749
 750         /*
 751          * scale the add/delete
 752          */
 753         max = max_change;
 754         if (max + len > 70)
 755                 max = 70 - len;
 756
 757         add = patch->lines_added;
 758         del = patch->lines_deleted;
 759         total = add + del;
 760
 761         if (max_change > 0) {
 762                 total = (total * max + max_change / 2) / max_change;
 763                 add = (add * max + max_change / 2) / max_change;
 764                 del = total - add;
 765         }
 766         printf(" %-*s |%5d %.*s%.*s\n",
 767                 len, name, patch->lines_added + patch->lines_deleted,
 768                 add, pluses, del, minuses);
 769 }
 770
 771 static int read_old_data(struct stat *st, const char *path, void *buf, unsigned long size)
 772 {
 773         int fd;
 774         unsigned long got;
 775
 776         switch (st->st_mode & S_IFMT) {
 777         case S_IFLNK:
 778                 return readlink(path, buf, size);
 779         case S_IFREG:
 780                 fd = open(path, O_RDONLY);
 781                 if (fd < 0)
 782                         return error("unable to open %s", path);
 783                 got = 0;
 784                 for (;;) {
 785                         int ret = read(fd, buf + got, size - got);
 786                         if (ret < 0) {
 787                                 if (errno == EAGAIN)
 788                                         continue;
 789                                 break;
 790                         }
 791                         if (!ret)
 792                                 break;
 793                         got += ret;
 794                 }
 795                 close(fd);
 796                 return got;
 797
 798         default:
 799                 return -1;
 800         }
 801 }
 802
 803 static int find_offset(const char *buf, unsigned long size, const char *fragment, unsigned long fragsize, int line)
 804 {
 805         int i;
 806         unsigned long start, backwards, forwards;
 807
 808         if (fragsize > size)
 809                 return -1;
 810
 811         start = 0;
 812         if (line > 1) {
 813                 unsigned long offset = 0;
 814                 i = line-1;
 815                 while (offset + fragsize <= size) {
 816                         if (buf[offset++] == '\n') {
 817                                 start = offset;
 818                                 if (!--i)
 819                                         break;
 820                         }
 821                 }
 822         }
 823
 824         /* Exact line number? */
 825         if (!memcmp(buf + start, fragment, fragsize))
 826                 return start;
 827
 828         /*
 829          * There's probably some smart way to do this, but I'll leave
 830          * that to the smart and beautiful people. I'm simple and stupid.
 831          */
 832         backwards = start;
 833         forwards = start;
 834         for (i = 0; ; i++) {
 835                 unsigned long try;
 836                 int n;
 837
 838                 /* "backward" */
 839                 if (i & 1) {
 840                         if (!backwards) {
 841                                 if (forwards + fragsize > size)
 842                                         break;
 843                                 continue;
 844                         }
 845                         do {
 846                                 --backwards;
 847                         } while (backwards && buf[backwards-1] != '\n');
 848                         try = backwards;
 849                 } else {
 850                         while (forwards + fragsize <= size) {
 851                                 if (buf[forwards++] == '\n')
 852                                         break;
 853                         }
 854                         try = forwards;
 855                 }
 856
 857                 if (try + fragsize > size)
 858                         continue;
 859                 if (memcmp(buf + try, fragment, fragsize))
 860                         continue;
 861                 n = (i >> 1)+1;
 862                 if (i & 1)
 863                         n = -n;
 864                 return try;
 865         }
 866
 867         /*
 868          * We should start searching forward and backward.
 869          */
 870         return -1;
 871 }
 872
 873 struct buffer_desc {
 874         char *buffer;
 875         unsigned long size;
 876         unsigned long alloc;
 877 };
 878
 879 static int apply_one_fragment(struct buffer_desc *desc, struct fragment *frag)
 880 {
 881         char *buf = desc->buffer;
 882         const char *patch = frag->patch;
 883         int offset, size = frag->size;
 884         char *old = xmalloc(size);
 885         char *new = xmalloc(size);
 886         int oldsize = 0, newsize = 0;
 887
 888         while (size > 0) {
 889                 int len = linelen(patch, size);
 890                 int plen;
 891
 892                 if (!len)
 893                         break;
 894
 895                 /*
 896                  * "plen" is how much of the line we should use for
 897                  * the actual patch data. Normally we just remove the
 898                  * first character on the line, but if the line is
 899                  * followed by "\ No newline", then we also remove the
 900                  * last one (which is the newline, of course).
 901                  */
 902                 plen = len-1;
 903                 if (len > size && patch[len] == '\\')
 904                         plen--;
 905                 switch (*patch) {
 906                 case ' ':
 907                 case '-':
 908                         memcpy(old + oldsize, patch + 1, plen);
 909                         oldsize += plen;
 910                         if (*patch == '-')
 911                                 break;
 912                 /* Fall-through for ' ' */
 913                 case '+':
 914                         memcpy(new + newsize, patch + 1, plen);
 915                         newsize += plen;
 916                         break;
 917                 case '@': case '\\':
 918                         /* Ignore it, we already handled it */
 919                         break;
 920                 default:
 921                         return -1;
 922                 }
 923                 patch += len;
 924                 size -= len;
 925         }
 926
 927         offset = find_offset(buf, desc->size, old, oldsize, frag->newpos);
 928         if (offset >= 0) {
 929                 int diff = newsize - oldsize;
 930                 unsigned long size = desc->size + diff;
 931                 unsigned long alloc = desc->alloc;
 932
 933                 if (size > alloc) {
 934                         alloc = size + 8192;
 935                         desc->alloc = alloc;
 936                         buf = xrealloc(buf, alloc);
 937                         desc->buffer = buf;
 938                 }
 939                 desc->size = size;
 940                 memmove(buf + offset + newsize, buf + offset + oldsize, size - offset - newsize);
 941                 memcpy(buf + offset, new, newsize);
 942                 offset = 0;
 943         }
 944
 945         free(old);
 946         free(new);
 947         return offset;
 948 }
 949
 950 static int apply_fragments(struct buffer_desc *desc, struct patch *patch)
 951 {
 952         struct fragment *frag = patch->fragments;
 953
 954         while (frag) {
 955                 if (apply_one_fragment(desc, frag) < 0)
 956                         return error("patch failed: %s:%d", patch->old_name, frag->oldpos);
 957                 frag = frag->next;
 958         }
 959         return 0;
 960 }
 961
 962 static int apply_data(struct patch *patch, struct stat *st)
 963 {
 964         char *buf;
 965         unsigned long size, alloc;
 966         struct buffer_desc desc;
 967
 968         size = 0;
 969         alloc = 0;
 970         buf = NULL;
 971         if (patch->old_name) {
 972                 size = st->st_size;
 973                 alloc = size + 8192;
 974                 buf = xmalloc(alloc);
 975                 if (read_old_data(st, patch->old_name, buf, alloc) != size)
 976                         return error("read of %s failed", patch->old_name);
 977         }
 978
 979         desc.size = size;
 980         desc.alloc = alloc;
 981         desc.buffer = buf;
 982         if (apply_fragments(&desc, patch) < 0)
 983                 return -1;
 984         patch->result = desc.buffer;
 985         patch->resultsize = desc.size;
 986
 987         if (patch->is_delete && patch->resultsize)
 988                 return error("removal patch leaves file contents");
 989
 990         return 0;
 991 }
 992
 993 static int check_patch(struct patch *patch)
 994 {
 995         struct stat st;
 996         const char *old_name = patch->old_name;
 997         const char *new_name = patch->new_name;
 998
 999         if (old_name) {
1000                 int changed;
1001
1002                 if (lstat(old_name, &st) < 0)
1003                         return error("%s: %s", old_name, strerror(errno));
1004                 if (check_index) {
1005                         int pos = cache_name_pos(old_name, strlen(old_name));
1006                         if (pos < 0)
1007                                 return error("%s: does not exist in index", old_name);
1008                         changed = ce_match_stat(active_cache[pos], &st);
1009                         if (changed)
1010                                 return error("%s: does not match index", old_name);
1011                 }
1012                 if (patch->is_new < 0)
1013                         patch->is_new = 0;
1014                 st.st_mode = ntohl(create_ce_mode(st.st_mode));
1015                 if (!patch->old_mode)
1016                         patch->old_mode = st.st_mode;
1017                 if ((st.st_mode ^ patch->old_mode) & S_IFMT)
1018                         return error("%s: wrong type", old_name);
1019                 if (st.st_mode != patch->old_mode)
1020                         fprintf(stderr, "warning: %s has type %o, expected %o\n",
1021                                 old_name, st.st_mode, patch->old_mode);
1022         }
1023
1024         if (new_name && (patch->is_new | patch->is_rename | patch->is_copy)) {
1025                 if (check_index && cache_name_pos(new_name, strlen(new_name)) >= 0)
1026                         return error("%s: already exists in index", new_name);
1027                 if (!lstat(new_name, &st))
1028                         return error("%s: already exists in working directory", new_name);
1029                 if (errno != ENOENT)
1030                         return error("%s: %s", new_name, strerror(errno));
1031                 if (!patch->new_mode)
1032                         patch->new_mode = S_IFREG | 0644;
1033         }
1034
1035         if (new_name && old_name) {
1036                 int same = !strcmp(old_name, new_name);
1037                 if (!patch->new_mode)
1038                         patch->new_mode = patch->old_mode;
1039                 if ((patch->old_mode ^ patch->new_mode) & S_IFMT)
1040                         return error("new mode (%o) of %s does not match old mode (%o)%s%s",
1041                                 patch->new_mode, new_name, patch->old_mode,
1042                                 same ? "" : " of ", same ? "" : old_name);
1043         }
1044
1045         if (apply_data(patch, &st) < 0)
1046                 return error("%s: patch does not apply", old_name);
1047         return 0;
1048 }
1049
1050 static int check_patch_list(struct patch *patch)
1051 {
1052         int error = 0;
1053
1054         for (;patch ; patch = patch->next)
1055                 error |= check_patch(patch);
1056         return error;
1057 }
1058
1059 static void show_file(int c, unsigned int mode, const char *name)
1060 {
1061         printf("%c %o %s\n", c, mode, name);
1062 }
1063
1064 static void show_file_list(struct patch *patch)
1065 {
1066         for (;patch ; patch = patch->next) {
1067                 if (patch->is_rename) {
1068                         show_file('-', patch->old_mode, patch->old_name);
1069                         show_file('+', patch->new_mode, patch->new_name);
1070                         continue;
1071                 }
1072                 if (patch->is_copy || patch->is_new) {
1073                         show_file('+', patch->new_mode, patch->new_name);
1074                         continue;
1075                 }
1076                 if (patch->is_delete) {
1077                         show_file('-', patch->old_mode, patch->old_name);
1078                         continue;
1079                 }
1080                 if (patch->old_mode && patch->new_mode && patch->old_mode != patch->new_mode) {
1081                         printf("M %o:%o %s\n", patch->old_mode, patch->new_mode, patch->old_name);
1082                         continue;
1083                 }
1084                 printf("M %o %s\n", patch->old_mode, patch->old_name);
1085         }
1086 }
1087
1088 static void stat_patch_list(struct patch *patch)
1089 {
1090         int files, adds, dels;
1091
1092         for (files = adds = dels = 0 ; patch ; patch = patch->next) {
1093                 files++;
1094                 adds += patch->lines_added;
1095                 dels += patch->lines_deleted;
1096                 show_stats(patch);
1097         }
1098
1099         printf(" %d files changed, %d insertions(+), %d deletions(-)\n", files, adds, dels);
1100 }
1101
1102 static void show_file_mode_name(const char *newdelete, unsigned int mode, const char *name)
1103 {
1104         if (mode)
1105                 printf(" %s mode %06o %s\n", newdelete, mode, name);
1106         else
1107                 printf(" %s %s\n", newdelete, name);
1108 }
1109
1110 static void show_mode_change(struct patch *p, int show_name)
1111 {
1112         if (p->old_mode && p->new_mode && p->old_mode != p->new_mode) {
1113                 if (show_name)
1114                         printf(" mode change %06o => %06o %s\n",
1115                                p->old_mode, p->new_mode, p->new_name);
1116                 else
1117                         printf(" mode change %06o => %06o\n",
1118                                p->old_mode, p->new_mode);
1119         }
1120 }
1121
1122 static void show_rename_copy(struct patch *p)
1123 {
1124         const char *renamecopy = p->is_rename ? "rename" : "copy";
1125         const char *old, *new;
1126
1127         /* Find common prefix */
1128         old = p->old_name;
1129         new = p->new_name;
1130         while (1) {
1131                 const char *slash_old, *slash_new;
1132                 slash_old = strchr(old, '/');
1133                 slash_new = strchr(new, '/');
1134                 if (!slash_old ||
1135                     !slash_new ||
1136                     slash_old - old != slash_new - new ||
1137                     memcmp(old, new, slash_new - new))
1138                         break;
1139                 old = slash_old + 1;
1140                 new = slash_new + 1;
1141         }
1142         /* p->old_name thru old is the common prefix, and old and new
1143          * through the end of names are renames
1144          */
1145         if (old != p->old_name)
1146                 printf(" %s %.*s{%s => %s} (%d%%)\n", renamecopy,
1147                        (int)(old - p->old_name), p->old_name,
1148                        old, new, p->score);
1149         else
1150                 printf(" %s %s => %s (%d%%)\n", renamecopy,
1151                        p->old_name, p->new_name, p->score);
1152         show_mode_change(p, 0);
1153 }
1154
1155 static void summary_patch_list(struct patch *patch)
1156 {
1157         struct patch *p;
1158
1159         for (p = patch; p; p = p->next) {
1160                 if (p->is_new)
1161                         show_file_mode_name("create", p->new_mode, p->new_name);
1162                 else if (p->is_delete)
1163                         show_file_mode_name("delete", p->old_mode, p->old_name);
1164                 else {
1165                         if (p->is_rename || p->is_copy)
1166                                 show_rename_copy(p);
1167                         else {
1168                                 if (p->score) {
1169                                         printf(" rewrite %s (%d%%)\n",
1170                                                p->new_name, p->score);
1171                                         show_mode_change(p, 0);
1172                                 }
1173                                 else
1174                                         show_mode_change(p, 1);
1175                         }
1176                 }
1177         }
1178 }
1179
1180 static void patch_stats(struct patch *patch)
1181 {
1182         int lines = patch->lines_added + patch->lines_deleted;
1183
1184         if (lines > max_change)
1185                 max_change = lines;
1186         if (patch->old_name) {
1187                 int len = strlen(patch->old_name);
1188                 if (len > max_len)
1189                         max_len = len;
1190         }
1191         if (patch->new_name) {
1192                 int len = strlen(patch->new_name);
1193                 if (len > max_len)
1194                         max_len = len;
1195         }
1196 }
1197
1198 static void remove_file(struct patch *patch)
1199 {
1200         if (write_index) {
1201                 if (remove_file_from_cache(patch->old_name) < 0)
1202                         die("unable to remove %s from index", patch->old_name);
1203         }
1204         unlink(patch->old_name);
1205 }
1206
1207 static void add_index_file(const char *path, unsigned mode, void *buf, unsigned long size)
1208 {
1209         struct stat st;
1210         struct cache_entry *ce;
1211         int namelen = strlen(path);
1212         unsigned ce_size = cache_entry_size(namelen);
1213
1214         if (!write_index)
1215                 return;
1216
1217         ce = xmalloc(ce_size);
1218         memset(ce, 0, ce_size);
1219         memcpy(ce->name, path, namelen);
1220         ce->ce_mode = create_ce_mode(mode);
1221         ce->ce_flags = htons(namelen);
1222         if (lstat(path, &st) < 0)
1223                 die("unable to stat newly created file %s", path);
1224         fill_stat_cache_info(ce, &st);
1225         if (write_sha1_file(buf, size, "blob", ce->sha1) < 0)
1226                 die("unable to create backing store for newly created file %s", path);
1227         if (add_cache_entry(ce, ADD_CACHE_OK_TO_ADD) < 0)
1228                 die("unable to add cache entry for %s", path);
1229 }
1230
1231 static void create_subdirectories(const char *path)
1232 {
1233         int len = strlen(path);
1234         char *buf = xmalloc(len + 1);
1235         const char *slash = path;
1236
1237         while ((slash = strchr(slash+1, '/')) != NULL) {
1238                 len = slash - path;
1239                 memcpy(buf, path, len);
1240                 buf[len] = 0;
1241                 if (mkdir(buf, 0777) < 0) {
1242                         if (errno != EEXIST)
1243                                 break;
1244                 }
1245         }
1246         free(buf);
1247 }
1248
1249 static int try_create_file(const char *path, unsigned int mode, const char *buf, unsigned long size)
1250 {
1251         int fd;
1252
1253         if (S_ISLNK(mode))
1254                 return symlink(buf, path);
1255         fd = open(path, O_CREAT | O_EXCL | O_WRONLY | O_TRUNC, (mode & 0100) ? 0777 : 0666);
1256         if (fd < 0)
1257                 return -1;
1258         while (size) {
1259                 int written = write(fd, buf, size);
1260                 if (written < 0) {
1261                         if (errno == EINTR || errno == EAGAIN)
1262                                 continue;
1263                         die("writing file %s: %s", path, strerror(errno));
1264                 }
1265                 if (!written)
1266                         die("out of space writing file %s", path);
1267                 buf += written;
1268                 size -= written;
1269         }
1270         if (close(fd) < 0)
1271                 die("closing file %s: %s", path, strerror(errno));
1272         return 0;
1273 }
1274
1275 /*
1276  * We optimistically assume that the directories exist,
1277  * which is true 99% of the time anyway. If they don't,
1278  * we create them and try again.
1279  */
1280 static void create_one_file(const char *path, unsigned mode, const char *buf, unsigned long size)
1281 {
1282         if (!try_create_file(path, mode, buf, size))
1283                 return;
1284
1285         if (errno == ENOENT) {
1286                 create_subdirectories(path);
1287                 if (!try_create_file(path, mode, buf, size))
1288                         return;
1289         }
1290
1291         if (errno == EEXIST) {
1292                 unsigned int nr = getpid();
1293
1294                 for (;;) {
1295                         const char *newpath;
1296                         newpath = mkpath("%s~%u", path, nr);
1297                         if (!try_create_file(newpath, mode, buf, size)) {
1298                                 if (!rename(newpath, path))
1299                                         return;
1300                                 unlink(newpath);
1301                                 break;
1302                         }
1303                         if (errno != EEXIST)
1304                                 break;
1305                 }
1306         }
1307         die("unable to write file %s mode %o", path, mode);
1308 }
1309
1310 static void create_file(struct patch *patch)
1311 {
1312         const char *path = patch->new_name;
1313         unsigned mode = patch->new_mode;
1314         unsigned long size = patch->resultsize;
1315         char *buf = patch->result;
1316
1317         if (!mode)
1318                 mode = S_IFREG | 0644;
1319         create_one_file(path, mode, buf, size);
1320         add_index_file(path, mode, buf, size);
1321 }
1322
1323 static void write_out_one_result(struct patch *patch)
1324 {
1325         if (patch->is_delete > 0) {
1326                 remove_file(patch);
1327                 return;
1328         }
1329         if (patch->is_new > 0 || patch->is_copy) {
1330                 create_file(patch);
1331                 return;
1332         }
1333         /*
1334          * Rename or modification boils down to the same
1335          * thing: remove the old, write the new
1336          */
1337         remove_file(patch);
1338         create_file(patch);
1339 }
1340
1341 static void write_out_results(struct patch *list)
1342 {
1343         if (!list)
1344                 die("No changes");
1345
1346         while (list) {
1347                 write_out_one_result(list);
1348                 list = list->next;
1349         }
1350 }
1351
1352 static struct cache_file cache_file;
1353
1354 static int apply_patch(int fd)
1355 {
1356         int newfd;
1357         unsigned long offset, size;
1358         char *buffer = read_patch_file(fd, &size);
1359         struct patch *list = NULL, **listp = &list;
1360
1361         if (!buffer)
1362                 return -1;
1363         offset = 0;
1364         while (size > 0) {
1365                 struct patch *patch;
1366                 int nr;
1367
1368                 patch = xmalloc(sizeof(*patch));
1369                 memset(patch, 0, sizeof(*patch));
1370                 nr = parse_chunk(buffer + offset, size, patch);
1371                 if (nr < 0)
1372                         break;
1373                 patch_stats(patch);
1374                 *listp = patch;
1375                 listp = &patch->next;
1376                 offset += nr;
1377                 size -= nr;
1378         }
1379
1380         newfd = -1;
1381         write_index = check_index && apply;
1382         if (write_index)
1383                 newfd = hold_index_file_for_update(&cache_file, get_index_file());
1384         if (check_index) {
1385                 if (read_cache() < 0)
1386                         die("unable to read index file");
1387         }
1388
1389         if ((check || apply) && check_patch_list(list) < 0)
1390                 exit(1);
1391
1392         if (apply)
1393                 write_out_results(list);
1394
1395         if (write_index) {
1396                 if (write_cache(newfd, active_cache, active_nr) ||
1397                     commit_index_file(&cache_file))
1398                         die("Unable to write new cachefile");
1399         }
1400
1401         if (show_files)
1402                 show_file_list(list);
1403
1404         if (diffstat)
1405                 stat_patch_list(list);
1406
1407         if (summary)
1408                 summary_patch_list(list);
1409
1410         free(buffer);
1411         return 0;
1412 }
1413
1414 int main(int argc, char **argv)
1415 {
1416         int i;
1417         int read_stdin = 1;
1418
1419         for (i = 1; i < argc; i++) {
1420                 const char *arg = argv[i];
1421                 int fd;
1422
1423                 if (!strcmp(arg, "-")) {
1424                         apply_patch(0);
1425                         read_stdin = 0;
1426                         continue;
1427                 }
1428                 /* NEEDSWORK: this does not do anything at this moment. */
1429                 if (!strcmp(arg, "--no-merge")) {
1430                         merge_patch = 0;
1431                         continue;
1432                 }
1433                 if (!strcmp(arg, "--stat")) {
1434                         apply = 0;
1435                         diffstat = 1;
1436                         continue;
1437                 }
1438                 if (!strcmp(arg, "--summary")) {
1439                         apply = 0;
1440                         summary = 1;
1441                         continue;
1442                 }
1443                 if (!strcmp(arg, "--check")) {
1444                         apply = 0;
1445                         check = 1;
1446                         continue;
1447                 }
1448                 if (!strcmp(arg, "--index")) {
1449                         check_index = 1;
1450                         continue;
1451                 }
1452                 if (!strcmp(arg, "--apply")) {
1453                         apply = 1;
1454                         continue;
1455                 }
1456                 if (!strcmp(arg, "--show-files")) {
1457                         show_files = 1;
1458                         continue;
1459                 }
1460                 fd = open(arg, O_RDONLY);
1461                 if (fd < 0)
1462                         usage(apply_usage);
1463                 read_stdin = 0;
1464                 apply_patch(fd);
1465                 close(fd);
1466         }
1467         if (read_stdin)
1468                 apply_patch(0);
1469         return 0;
1470 }