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 #include <fnmatch.h>
  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         /* If a fragment ends with an incomplete line, we failed to include
 683          * it in the above loop because we hit oldlines == newlines == 0
 684          * before seeing it.
 685          */
 686         if (12 < size && !memcmp(line, "\\ No newline", 12))
 687                 offset += linelen(line, size);
 688
 689         patch->lines_added += added;
 690         patch->lines_deleted += deleted;
 691         return offset;
 692 }
 693
 694 static int parse_single_patch(char *line, unsigned long size, struct patch *patch)
 695 {
 696         unsigned long offset = 0;
 697         struct fragment **fragp = &patch->fragments;
 698
 699         while (size > 4 && !memcmp(line, "@@ -", 4)) {
 700                 struct fragment *fragment;
 701                 int len;
 702
 703                 fragment = xmalloc(sizeof(*fragment));
 704                 memset(fragment, 0, sizeof(*fragment));
 705                 len = parse_fragment(line, size, patch, fragment);
 706                 if (len <= 0)
 707                         die("corrupt patch at line %d", linenr);
 708
 709                 fragment->patch = line;
 710                 fragment->size = len;
 711
 712                 *fragp = fragment;
 713                 fragp = &fragment->next;
 714
 715                 offset += len;
 716                 line += len;
 717                 size -= len;
 718         }
 719         return offset;
 720 }
 721
 722 static int parse_chunk(char *buffer, unsigned long size, struct patch *patch)
 723 {
 724         int hdrsize, patchsize;
 725         int offset = find_header(buffer, size, &hdrsize, patch);
 726
 727         if (offset < 0)
 728                 return offset;
 729
 730         patchsize = parse_single_patch(buffer + offset + hdrsize, size - offset - hdrsize, patch);
 731
 732         return offset + hdrsize + patchsize;
 733 }
 734
 735 static const char pluses[] = "++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++";
 736 static const char minuses[]= "----------------------------------------------------------------------";
 737
 738 static void show_stats(struct patch *patch)
 739 {
 740         char *name = patch->new_name;
 741         int len, max, add, del, total;
 742
 743         if (!name)
 744                 name = patch->old_name;
 745
 746         /*
 747          * "scale" the filename
 748          */
 749         len = strlen(name);
 750         max = max_len;
 751         if (max > 50)
 752                 max = 50;
 753         if (len > max)
 754                 name += len - max;
 755         len = max;
 756
 757         /*
 758          * scale the add/delete
 759          */
 760         max = max_change;
 761         if (max + len > 70)
 762                 max = 70 - len;
 763
 764         add = patch->lines_added;
 765         del = patch->lines_deleted;
 766         total = add + del;
 767
 768         if (max_change > 0) {
 769                 total = (total * max + max_change / 2) / max_change;
 770                 add = (add * max + max_change / 2) / max_change;
 771                 del = total - add;
 772         }
 773         printf(" %-*s |%5d %.*s%.*s\n",
 774                 len, name, patch->lines_added + patch->lines_deleted,
 775                 add, pluses, del, minuses);
 776 }
 777
 778 static int read_old_data(struct stat *st, const char *path, void *buf, unsigned long size)
 779 {
 780         int fd;
 781         unsigned long got;
 782
 783         switch (st->st_mode & S_IFMT) {
 784         case S_IFLNK:
 785                 return readlink(path, buf, size);
 786         case S_IFREG:
 787                 fd = open(path, O_RDONLY);
 788                 if (fd < 0)
 789                         return error("unable to open %s", path);
 790                 got = 0;
 791                 for (;;) {
 792                         int ret = read(fd, buf + got, size - got);
 793                         if (ret < 0) {
 794                                 if (errno == EAGAIN)
 795                                         continue;
 796                                 break;
 797                         }
 798                         if (!ret)
 799                                 break;
 800                         got += ret;
 801                 }
 802                 close(fd);
 803                 return got;
 804
 805         default:
 806                 return -1;
 807         }
 808 }
 809
 810 static int find_offset(const char *buf, unsigned long size, const char *fragment, unsigned long fragsize, int line)
 811 {
 812         int i;
 813         unsigned long start, backwards, forwards;
 814
 815         if (fragsize > size)
 816                 return -1;
 817
 818         start = 0;
 819         if (line > 1) {
 820                 unsigned long offset = 0;
 821                 i = line-1;
 822                 while (offset + fragsize <= size) {
 823                         if (buf[offset++] == '\n') {
 824                                 start = offset;
 825                                 if (!--i)
 826                                         break;
 827                         }
 828                 }
 829         }
 830
 831         /* Exact line number? */
 832         if (!memcmp(buf + start, fragment, fragsize))
 833                 return start;
 834
 835         /*
 836          * There's probably some smart way to do this, but I'll leave
 837          * that to the smart and beautiful people. I'm simple and stupid.
 838          */
 839         backwards = start;
 840         forwards = start;
 841         for (i = 0; ; i++) {
 842                 unsigned long try;
 843                 int n;
 844
 845                 /* "backward" */
 846                 if (i & 1) {
 847                         if (!backwards) {
 848                                 if (forwards + fragsize > size)
 849                                         break;
 850                                 continue;
 851                         }
 852                         do {
 853                                 --backwards;
 854                         } while (backwards && buf[backwards-1] != '\n');
 855                         try = backwards;
 856                 } else {
 857                         while (forwards + fragsize <= size) {
 858                                 if (buf[forwards++] == '\n')
 859                                         break;
 860                         }
 861                         try = forwards;
 862                 }
 863
 864                 if (try + fragsize > size)
 865                         continue;
 866                 if (memcmp(buf + try, fragment, fragsize))
 867                         continue;
 868                 n = (i >> 1)+1;
 869                 if (i & 1)
 870                         n = -n;
 871                 return try;
 872         }
 873
 874         /*
 875          * We should start searching forward and backward.
 876          */
 877         return -1;
 878 }
 879
 880 struct buffer_desc {
 881         char *buffer;
 882         unsigned long size;
 883         unsigned long alloc;
 884 };
 885
 886 static int apply_one_fragment(struct buffer_desc *desc, struct fragment *frag)
 887 {
 888         char *buf = desc->buffer;
 889         const char *patch = frag->patch;
 890         int offset, size = frag->size;
 891         char *old = xmalloc(size);
 892         char *new = xmalloc(size);
 893         int oldsize = 0, newsize = 0;
 894
 895         while (size > 0) {
 896                 int len = linelen(patch, size);
 897                 int plen;
 898
 899                 if (!len)
 900                         break;
 901
 902                 /*
 903                  * "plen" is how much of the line we should use for
 904                  * the actual patch data. Normally we just remove the
 905                  * first character on the line, but if the line is
 906                  * followed by "\ No newline", then we also remove the
 907                  * last one (which is the newline, of course).
 908                  */
 909                 plen = len-1;
 910                 if (len < size && patch[len] == '\\')
 911                         plen--;
 912                 switch (*patch) {
 913                 case ' ':
 914                 case '-':
 915                         memcpy(old + oldsize, patch + 1, plen);
 916                         oldsize += plen;
 917                         if (*patch == '-')
 918                                 break;
 919                 /* Fall-through for ' ' */
 920                 case '+':
 921                         memcpy(new + newsize, patch + 1, plen);
 922                         newsize += plen;
 923                         break;
 924                 case '@': case '\\':
 925                         /* Ignore it, we already handled it */
 926                         break;
 927                 default:
 928                         return -1;
 929                 }
 930                 patch += len;
 931                 size -= len;
 932         }
 933
 934         offset = find_offset(buf, desc->size, old, oldsize, frag->newpos);
 935         if (offset >= 0) {
 936                 int diff = newsize - oldsize;
 937                 unsigned long size = desc->size + diff;
 938                 unsigned long alloc = desc->alloc;
 939
 940                 if (size > alloc) {
 941                         alloc = size + 8192;
 942                         desc->alloc = alloc;
 943                         buf = xrealloc(buf, alloc);
 944                         desc->buffer = buf;
 945                 }
 946                 desc->size = size;
 947                 memmove(buf + offset + newsize, buf + offset + oldsize, size - offset - newsize);
 948                 memcpy(buf + offset, new, newsize);
 949                 offset = 0;
 950         }
 951
 952         free(old);
 953         free(new);
 954         return offset;
 955 }
 956
 957 static int apply_fragments(struct buffer_desc *desc, struct patch *patch)
 958 {
 959         struct fragment *frag = patch->fragments;
 960
 961         while (frag) {
 962                 if (apply_one_fragment(desc, frag) < 0)
 963                         return error("patch failed: %s:%d", patch->old_name, frag->oldpos);
 964                 frag = frag->next;
 965         }
 966         return 0;
 967 }
 968
 969 static int apply_data(struct patch *patch, struct stat *st)
 970 {
 971         char *buf;
 972         unsigned long size, alloc;
 973         struct buffer_desc desc;
 974
 975         size = 0;
 976         alloc = 0;
 977         buf = NULL;
 978         if (patch->old_name) {
 979                 size = st->st_size;
 980                 alloc = size + 8192;
 981                 buf = xmalloc(alloc);
 982                 if (read_old_data(st, patch->old_name, buf, alloc) != size)
 983                         return error("read of %s failed", patch->old_name);
 984         }
 985
 986         desc.size = size;
 987         desc.alloc = alloc;
 988         desc.buffer = buf;
 989         if (apply_fragments(&desc, patch) < 0)
 990                 return -1;
 991         patch->result = desc.buffer;
 992         patch->resultsize = desc.size;
 993
 994         if (patch->is_delete && patch->resultsize)
 995                 return error("removal patch leaves file contents");
 996
 997         return 0;
 998 }
 999
1000 static int check_patch(struct patch *patch)
1001 {
1002         struct stat st;
1003         const char *old_name = patch->old_name;
1004         const char *new_name = patch->new_name;
1005
1006         if (old_name) {
1007                 int changed;
1008
1009                 if (lstat(old_name, &st) < 0)
1010                         return error("%s: %s", old_name, strerror(errno));
1011                 if (check_index) {
1012                         int pos = cache_name_pos(old_name, strlen(old_name));
1013                         if (pos < 0)
1014                                 return error("%s: does not exist in index", old_name);
1015                         changed = ce_match_stat(active_cache[pos], &st);
1016                         if (changed)
1017                                 return error("%s: does not match index", old_name);
1018                 }
1019                 if (patch->is_new < 0)
1020                         patch->is_new = 0;
1021                 st.st_mode = ntohl(create_ce_mode(st.st_mode));
1022                 if (!patch->old_mode)
1023                         patch->old_mode = st.st_mode;
1024                 if ((st.st_mode ^ patch->old_mode) & S_IFMT)
1025                         return error("%s: wrong type", old_name);
1026                 if (st.st_mode != patch->old_mode)
1027                         fprintf(stderr, "warning: %s has type %o, expected %o\n",
1028                                 old_name, st.st_mode, patch->old_mode);
1029         }
1030
1031         if (new_name && (patch->is_new | patch->is_rename | patch->is_copy)) {
1032                 if (check_index && cache_name_pos(new_name, strlen(new_name)) >= 0)
1033                         return error("%s: already exists in index", new_name);
1034                 if (!lstat(new_name, &st))
1035                         return error("%s: already exists in working directory", new_name);
1036                 if (errno != ENOENT)
1037                         return error("%s: %s", new_name, strerror(errno));
1038                 if (!patch->new_mode)
1039                         patch->new_mode = S_IFREG | 0644;
1040         }
1041
1042         if (new_name && old_name) {
1043                 int same = !strcmp(old_name, new_name);
1044                 if (!patch->new_mode)
1045                         patch->new_mode = patch->old_mode;
1046                 if ((patch->old_mode ^ patch->new_mode) & S_IFMT)
1047                         return error("new mode (%o) of %s does not match old mode (%o)%s%s",
1048                                 patch->new_mode, new_name, patch->old_mode,
1049                                 same ? "" : " of ", same ? "" : old_name);
1050         }
1051
1052         if (apply_data(patch, &st) < 0)
1053                 return error("%s: patch does not apply", old_name);
1054         return 0;
1055 }
1056
1057 static int check_patch_list(struct patch *patch)
1058 {
1059         int error = 0;
1060
1061         for (;patch ; patch = patch->next)
1062                 error |= check_patch(patch);
1063         return error;
1064 }
1065
1066 static void show_file(int c, unsigned int mode, const char *name)
1067 {
1068         printf("%c %o %s\n", c, mode, name);
1069 }
1070
1071 static void show_file_list(struct patch *patch)
1072 {
1073         for (;patch ; patch = patch->next) {
1074                 if (patch->is_rename) {
1075                         show_file('-', patch->old_mode, patch->old_name);
1076                         show_file('+', patch->new_mode, patch->new_name);
1077                         continue;
1078                 }
1079                 if (patch->is_copy || patch->is_new) {
1080                         show_file('+', patch->new_mode, patch->new_name);
1081                         continue;
1082                 }
1083                 if (patch->is_delete) {
1084                         show_file('-', patch->old_mode, patch->old_name);
1085                         continue;
1086                 }
1087                 if (patch->old_mode && patch->new_mode && patch->old_mode != patch->new_mode) {
1088                         printf("M %o:%o %s\n", patch->old_mode, patch->new_mode, patch->old_name);
1089                         continue;
1090                 }
1091                 printf("M %o %s\n", patch->old_mode, patch->old_name);
1092         }
1093 }
1094
1095 static void stat_patch_list(struct patch *patch)
1096 {
1097         int files, adds, dels;
1098
1099         for (files = adds = dels = 0 ; patch ; patch = patch->next) {
1100                 files++;
1101                 adds += patch->lines_added;
1102                 dels += patch->lines_deleted;
1103                 show_stats(patch);
1104         }
1105
1106         printf(" %d files changed, %d insertions(+), %d deletions(-)\n", files, adds, dels);
1107 }
1108
1109 static void show_file_mode_name(const char *newdelete, unsigned int mode, const char *name)
1110 {
1111         if (mode)
1112                 printf(" %s mode %06o %s\n", newdelete, mode, name);
1113         else
1114                 printf(" %s %s\n", newdelete, name);
1115 }
1116
1117 static void show_mode_change(struct patch *p, int show_name)
1118 {
1119         if (p->old_mode && p->new_mode && p->old_mode != p->new_mode) {
1120                 if (show_name)
1121                         printf(" mode change %06o => %06o %s\n",
1122                                p->old_mode, p->new_mode, p->new_name);
1123                 else
1124                         printf(" mode change %06o => %06o\n",
1125                                p->old_mode, p->new_mode);
1126         }
1127 }
1128
1129 static void show_rename_copy(struct patch *p)
1130 {
1131         const char *renamecopy = p->is_rename ? "rename" : "copy";
1132         const char *old, *new;
1133
1134         /* Find common prefix */
1135         old = p->old_name;
1136         new = p->new_name;
1137         while (1) {
1138                 const char *slash_old, *slash_new;
1139                 slash_old = strchr(old, '/');
1140                 slash_new = strchr(new, '/');
1141                 if (!slash_old ||
1142                     !slash_new ||
1143                     slash_old - old != slash_new - new ||
1144                     memcmp(old, new, slash_new - new))
1145                         break;
1146                 old = slash_old + 1;
1147                 new = slash_new + 1;
1148         }
1149         /* p->old_name thru old is the common prefix, and old and new
1150          * through the end of names are renames
1151          */
1152         if (old != p->old_name)
1153                 printf(" %s %.*s{%s => %s} (%d%%)\n", renamecopy,
1154                        (int)(old - p->old_name), p->old_name,
1155                        old, new, p->score);
1156         else
1157                 printf(" %s %s => %s (%d%%)\n", renamecopy,
1158                        p->old_name, p->new_name, p->score);
1159         show_mode_change(p, 0);
1160 }
1161
1162 static void summary_patch_list(struct patch *patch)
1163 {
1164         struct patch *p;
1165
1166         for (p = patch; p; p = p->next) {
1167                 if (p->is_new)
1168                         show_file_mode_name("create", p->new_mode, p->new_name);
1169                 else if (p->is_delete)
1170                         show_file_mode_name("delete", p->old_mode, p->old_name);
1171                 else {
1172                         if (p->is_rename || p->is_copy)
1173                                 show_rename_copy(p);
1174                         else {
1175                                 if (p->score) {
1176                                         printf(" rewrite %s (%d%%)\n",
1177                                                p->new_name, p->score);
1178                                         show_mode_change(p, 0);
1179                                 }
1180                                 else
1181                                         show_mode_change(p, 1);
1182                         }
1183                 }
1184         }
1185 }
1186
1187 static void patch_stats(struct patch *patch)
1188 {
1189         int lines = patch->lines_added + patch->lines_deleted;
1190
1191         if (lines > max_change)
1192                 max_change = lines;
1193         if (patch->old_name) {
1194                 int len = strlen(patch->old_name);
1195                 if (len > max_len)
1196                         max_len = len;
1197         }
1198         if (patch->new_name) {
1199                 int len = strlen(patch->new_name);
1200                 if (len > max_len)
1201                         max_len = len;
1202         }
1203 }
1204
1205 static void remove_file(struct patch *patch)
1206 {
1207         if (write_index) {
1208                 if (remove_file_from_cache(patch->old_name) < 0)
1209                         die("unable to remove %s from index", patch->old_name);
1210         }
1211         unlink(patch->old_name);
1212 }
1213
1214 static void add_index_file(const char *path, unsigned mode, void *buf, unsigned long size)
1215 {
1216         struct stat st;
1217         struct cache_entry *ce;
1218         int namelen = strlen(path);
1219         unsigned ce_size = cache_entry_size(namelen);
1220
1221         if (!write_index)
1222                 return;
1223
1224         ce = xmalloc(ce_size);
1225         memset(ce, 0, ce_size);
1226         memcpy(ce->name, path, namelen);
1227         ce->ce_mode = create_ce_mode(mode);
1228         ce->ce_flags = htons(namelen);
1229         if (lstat(path, &st) < 0)
1230                 die("unable to stat newly created file %s", path);
1231         fill_stat_cache_info(ce, &st);
1232         if (write_sha1_file(buf, size, "blob", ce->sha1) < 0)
1233                 die("unable to create backing store for newly created file %s", path);
1234         if (add_cache_entry(ce, ADD_CACHE_OK_TO_ADD) < 0)
1235                 die("unable to add cache entry for %s", path);
1236 }
1237
1238 static void create_subdirectories(const char *path)
1239 {
1240         int len = strlen(path);
1241         char *buf = xmalloc(len + 1);
1242         const char *slash = path;
1243
1244         while ((slash = strchr(slash+1, '/')) != NULL) {
1245                 len = slash - path;
1246                 memcpy(buf, path, len);
1247                 buf[len] = 0;
1248                 if (mkdir(buf, 0777) < 0) {
1249                         if (errno != EEXIST)
1250                                 break;
1251                 }
1252         }
1253         free(buf);
1254 }
1255
1256 static int try_create_file(const char *path, unsigned int mode, const char *buf, unsigned long size)
1257 {
1258         int fd;
1259
1260         if (S_ISLNK(mode))
1261                 return symlink(buf, path);
1262         fd = open(path, O_CREAT | O_EXCL | O_WRONLY | O_TRUNC, (mode & 0100) ? 0777 : 0666);
1263         if (fd < 0)
1264                 return -1;
1265         while (size) {
1266                 int written = write(fd, buf, size);
1267                 if (written < 0) {
1268                         if (errno == EINTR || errno == EAGAIN)
1269                                 continue;
1270                         die("writing file %s: %s", path, strerror(errno));
1271                 }
1272                 if (!written)
1273                         die("out of space writing file %s", path);
1274                 buf += written;
1275                 size -= written;
1276         }
1277         if (close(fd) < 0)
1278                 die("closing file %s: %s", path, strerror(errno));
1279         return 0;
1280 }
1281
1282 /*
1283  * We optimistically assume that the directories exist,
1284  * which is true 99% of the time anyway. If they don't,
1285  * we create them and try again.
1286  */
1287 static void create_one_file(const char *path, unsigned mode, const char *buf, unsigned long size)
1288 {
1289         if (!try_create_file(path, mode, buf, size))
1290                 return;
1291
1292         if (errno == ENOENT) {
1293                 create_subdirectories(path);
1294                 if (!try_create_file(path, mode, buf, size))
1295                         return;
1296         }
1297
1298         if (errno == EEXIST) {
1299                 unsigned int nr = getpid();
1300
1301                 for (;;) {
1302                         const char *newpath;
1303                         newpath = mkpath("%s~%u", path, nr);
1304                         if (!try_create_file(newpath, mode, buf, size)) {
1305                                 if (!rename(newpath, path))
1306                                         return;
1307                                 unlink(newpath);
1308                                 break;
1309                         }
1310                         if (errno != EEXIST)
1311                                 break;
1312                 }
1313         }
1314         die("unable to write file %s mode %o", path, mode);
1315 }
1316
1317 static void create_file(struct patch *patch)
1318 {
1319         const char *path = patch->new_name;
1320         unsigned mode = patch->new_mode;
1321         unsigned long size = patch->resultsize;
1322         char *buf = patch->result;
1323
1324         if (!mode)
1325                 mode = S_IFREG | 0644;
1326         create_one_file(path, mode, buf, size);
1327         add_index_file(path, mode, buf, size);
1328 }
1329
1330 static void write_out_one_result(struct patch *patch)
1331 {
1332         if (patch->is_delete > 0) {
1333                 remove_file(patch);
1334                 return;
1335         }
1336         if (patch->is_new > 0 || patch->is_copy) {
1337                 create_file(patch);
1338                 return;
1339         }
1340         /*
1341          * Rename or modification boils down to the same
1342          * thing: remove the old, write the new
1343          */
1344         remove_file(patch);
1345         create_file(patch);
1346 }
1347
1348 static void write_out_results(struct patch *list, int skipped_patch)
1349 {
1350         if (!list && !skipped_patch)
1351                 die("No changes");
1352
1353         while (list) {
1354                 write_out_one_result(list);
1355                 list = list->next;
1356         }
1357 }
1358
1359 static struct cache_file cache_file;
1360
1361 static struct excludes {
1362         struct excludes *next;
1363         const char *path;
1364 } *excludes;
1365
1366 static int use_patch(struct patch *p)
1367 {
1368         const char *pathname = p->new_name ? : p->old_name;
1369         struct excludes *x = excludes;
1370         while (x) {
1371                 if (fnmatch(x->path, pathname, 0) == 0)
1372                         return 0;
1373                 x = x->next;
1374         }
1375         return 1;
1376 }
1377
1378 static int apply_patch(int fd)
1379 {
1380         int newfd;
1381         unsigned long offset, size;
1382         char *buffer = read_patch_file(fd, &size);
1383         struct patch *list = NULL, **listp = &list;
1384         int skipped_patch = 0;
1385
1386         if (!buffer)
1387                 return -1;
1388         offset = 0;
1389         while (size > 0) {
1390                 struct patch *patch;
1391                 int nr;
1392
1393                 patch = xmalloc(sizeof(*patch));
1394                 memset(patch, 0, sizeof(*patch));
1395                 nr = parse_chunk(buffer + offset, size, patch);
1396                 if (nr < 0)
1397                         break;
1398                 if (use_patch(patch)) {
1399                         patch_stats(patch);
1400                         *listp = patch;
1401                         listp = &patch->next;
1402                 } else {
1403                         /* perhaps free it a bit better? */
1404                         free(patch);
1405                         skipped_patch++;
1406                 }
1407                 offset += nr;
1408                 size -= nr;
1409         }
1410
1411         newfd = -1;
1412         write_index = check_index && apply;
1413         if (write_index)
1414                 newfd = hold_index_file_for_update(&cache_file, get_index_file());
1415         if (check_index) {
1416                 if (read_cache() < 0)
1417                         die("unable to read index file");
1418         }
1419
1420         if ((check || apply) && check_patch_list(list) < 0)
1421                 exit(1);
1422
1423         if (apply)
1424                 write_out_results(list, skipped_patch);
1425
1426         if (write_index) {
1427                 if (write_cache(newfd, active_cache, active_nr) ||
1428                     commit_index_file(&cache_file))
1429                         die("Unable to write new cachefile");
1430         }
1431
1432         if (show_files)
1433                 show_file_list(list);
1434
1435         if (diffstat)
1436                 stat_patch_list(list);
1437
1438         if (summary)
1439                 summary_patch_list(list);
1440
1441         free(buffer);
1442         return 0;
1443 }
1444
1445 int main(int argc, char **argv)
1446 {
1447         int i;
1448         int read_stdin = 1;
1449
1450         for (i = 1; i < argc; i++) {
1451                 const char *arg = argv[i];
1452                 int fd;
1453
1454                 if (!strcmp(arg, "-")) {
1455                         apply_patch(0);
1456                         read_stdin = 0;
1457                         continue;
1458                 }
1459                 if (!strncmp(arg, "--exclude=", 10)) {
1460                         struct excludes *x = xmalloc(sizeof(*x));
1461                         x->path = arg + 10;
1462                         x->next = excludes;
1463                         excludes = x;
1464                         continue;
1465                 }
1466                 /* NEEDSWORK: this does not do anything at this moment. */
1467                 if (!strcmp(arg, "--no-merge")) {
1468                         merge_patch = 0;
1469                         continue;
1470                 }
1471                 if (!strcmp(arg, "--stat")) {
1472                         apply = 0;
1473                         diffstat = 1;
1474                         continue;
1475                 }
1476                 if (!strcmp(arg, "--summary")) {
1477                         apply = 0;
1478                         summary = 1;
1479                         continue;
1480                 }
1481                 if (!strcmp(arg, "--check")) {
1482                         apply = 0;
1483                         check = 1;
1484                         continue;
1485                 }
1486                 if (!strcmp(arg, "--index")) {
1487                         check_index = 1;
1488                         continue;
1489                 }
1490                 if (!strcmp(arg, "--apply")) {
1491                         apply = 1;
1492                         continue;
1493                 }
1494                 if (!strcmp(arg, "--show-files")) {
1495                         show_files = 1;
1496                         continue;
1497                 }
1498                 fd = open(arg, O_RDONLY);
1499                 if (fd < 0)
1500                         usage(apply_usage);
1501                 read_stdin = 0;
1502                 apply_patch(fd);
1503                 close(fd);
1504         }
1505         if (read_stdin)
1506                 apply_patch(0);
1507         return 0;
1508 }