diff-lib.con commit Merge branch 'jk/maint-diffstat-overflow' (bc32d34)
   1/*
   2 * Copyright (C) 2005 Junio C Hamano
   3 */
   4#include "cache.h"
   5#include "quote.h"
   6#include "commit.h"
   7#include "diff.h"
   8#include "diffcore.h"
   9#include "revision.h"
  10#include "cache-tree.h"
  11#include "unpack-trees.h"
  12#include "refs.h"
  13#include "submodule.h"
  14
  15/*
  16 * diff-files
  17 */
  18
  19/*
  20 * Has the work tree entity been removed?
  21 *
  22 * Return 1 if it was removed from the work tree, 0 if an entity to be
  23 * compared with the cache entry ce still exists (the latter includes
  24 * the case where a directory that is not a submodule repository
  25 * exists for ce that is a submodule -- it is a submodule that is not
  26 * checked out).  Return negative for an error.
  27 */
  28static int check_removed(const struct cache_entry *ce, struct stat *st)
  29{
  30        if (lstat(ce->name, st) < 0) {
  31                if (errno != ENOENT && errno != ENOTDIR)
  32                        return -1;
  33                return 1;
  34        }
  35        if (has_symlink_leading_path(ce->name, ce_namelen(ce)))
  36                return 1;
  37        if (S_ISDIR(st->st_mode)) {
  38                unsigned char sub[20];
  39
  40                /*
  41                 * If ce is already a gitlink, we can have a plain
  42                 * directory (i.e. the submodule is not checked out),
  43                 * or a checked out submodule.  Either case this is not
  44                 * a case where something was removed from the work tree,
  45                 * so we will return 0.
  46                 *
  47                 * Otherwise, if the directory is not a submodule
  48                 * repository, that means ce which was a blob turned into
  49                 * a directory --- the blob was removed!
  50                 */
  51                if (!S_ISGITLINK(ce->ce_mode) &&
  52                    resolve_gitlink_ref(ce->name, "HEAD", sub))
  53                        return 1;
  54        }
  55        return 0;
  56}
  57
  58/*
  59 * Has a file changed or has a submodule new commits or a dirty work tree?
  60 *
  61 * Return 1 when changes are detected, 0 otherwise. If the DIRTY_SUBMODULES
  62 * option is set, the caller does not only want to know if a submodule is
  63 * modified at all but wants to know all the conditions that are met (new
  64 * commits, untracked content and/or modified content).
  65 */
  66static int match_stat_with_submodule(struct diff_options *diffopt,
  67                                      struct cache_entry *ce, struct stat *st,
  68                                      unsigned ce_option, unsigned *dirty_submodule)
  69{
  70        int changed = ce_match_stat(ce, st, ce_option);
  71        if (S_ISGITLINK(ce->ce_mode)
  72            && !DIFF_OPT_TST(diffopt, IGNORE_SUBMODULES)
  73            && (!changed || DIFF_OPT_TST(diffopt, DIRTY_SUBMODULES))) {
  74                *dirty_submodule = is_submodule_modified(ce->name, DIFF_OPT_TST(diffopt, IGNORE_UNTRACKED_IN_SUBMODULES));
  75        }
  76        return changed;
  77}
  78
  79int run_diff_files(struct rev_info *revs, unsigned int option)
  80{
  81        int entries, i;
  82        int diff_unmerged_stage = revs->max_count;
  83        int silent_on_removed = option & DIFF_SILENT_ON_REMOVED;
  84        unsigned ce_option = ((option & DIFF_RACY_IS_MODIFIED)
  85                              ? CE_MATCH_RACY_IS_DIRTY : 0);
  86
  87        diff_set_mnemonic_prefix(&revs->diffopt, "i/", "w/");
  88
  89        if (diff_unmerged_stage < 0)
  90                diff_unmerged_stage = 2;
  91        entries = active_nr;
  92        for (i = 0; i < entries; i++) {
  93                struct stat st;
  94                unsigned int oldmode, newmode;
  95                struct cache_entry *ce = active_cache[i];
  96                int changed;
  97                unsigned dirty_submodule = 0;
  98
  99                if (DIFF_OPT_TST(&revs->diffopt, QUICK) &&
 100                        DIFF_OPT_TST(&revs->diffopt, HAS_CHANGES))
 101                        break;
 102
 103                if (!ce_path_match(ce, revs->prune_data))
 104                        continue;
 105
 106                if (ce_stage(ce)) {
 107                        struct combine_diff_path *dpath;
 108                        int num_compare_stages = 0;
 109                        size_t path_len;
 110
 111                        path_len = ce_namelen(ce);
 112
 113                        dpath = xmalloc(combine_diff_path_size(5, path_len));
 114                        dpath->path = (char *) &(dpath->parent[5]);
 115
 116                        dpath->next = NULL;
 117                        dpath->len = path_len;
 118                        memcpy(dpath->path, ce->name, path_len);
 119                        dpath->path[path_len] = '\0';
 120                        hashclr(dpath->sha1);
 121                        memset(&(dpath->parent[0]), 0,
 122                               sizeof(struct combine_diff_parent)*5);
 123
 124                        changed = check_removed(ce, &st);
 125                        if (!changed)
 126                                dpath->mode = ce_mode_from_stat(ce, st.st_mode);
 127                        else {
 128                                if (changed < 0) {
 129                                        perror(ce->name);
 130                                        continue;
 131                                }
 132                                if (silent_on_removed)
 133                                        continue;
 134                        }
 135
 136                        while (i < entries) {
 137                                struct cache_entry *nce = active_cache[i];
 138                                int stage;
 139
 140                                if (strcmp(ce->name, nce->name))
 141                                        break;
 142
 143                                /* Stage #2 (ours) is the first parent,
 144                                 * stage #3 (theirs) is the second.
 145                                 */
 146                                stage = ce_stage(nce);
 147                                if (2 <= stage) {
 148                                        int mode = nce->ce_mode;
 149                                        num_compare_stages++;
 150                                        hashcpy(dpath->parent[stage-2].sha1, nce->sha1);
 151                                        dpath->parent[stage-2].mode = ce_mode_from_stat(nce, mode);
 152                                        dpath->parent[stage-2].status =
 153                                                DIFF_STATUS_MODIFIED;
 154                                }
 155
 156                                /* diff against the proper unmerged stage */
 157                                if (stage == diff_unmerged_stage)
 158                                        ce = nce;
 159                                i++;
 160                        }
 161                        /*
 162                         * Compensate for loop update
 163                         */
 164                        i--;
 165
 166                        if (revs->combine_merges && num_compare_stages == 2) {
 167                                show_combined_diff(dpath, 2,
 168                                                   revs->dense_combined_merges,
 169                                                   revs);
 170                                free(dpath);
 171                                continue;
 172                        }
 173                        free(dpath);
 174                        dpath = NULL;
 175
 176                        /*
 177                         * Show the diff for the 'ce' if we found the one
 178                         * from the desired stage.
 179                         */
 180                        diff_unmerge(&revs->diffopt, ce->name, 0, null_sha1);
 181                        if (ce_stage(ce) != diff_unmerged_stage)
 182                                continue;
 183                }
 184
 185                if (ce_uptodate(ce) || ce_skip_worktree(ce))
 186                        continue;
 187
 188                /* If CE_VALID is set, don't look at workdir for file removal */
 189                changed = (ce->ce_flags & CE_VALID) ? 0 : check_removed(ce, &st);
 190                if (changed) {
 191                        if (changed < 0) {
 192                                perror(ce->name);
 193                                continue;
 194                        }
 195                        if (silent_on_removed)
 196                                continue;
 197                        diff_addremove(&revs->diffopt, '-', ce->ce_mode,
 198                                       ce->sha1, ce->name, 0);
 199                        continue;
 200                }
 201                changed = match_stat_with_submodule(&revs->diffopt, ce, &st,
 202                                                    ce_option, &dirty_submodule);
 203                if (!changed && !dirty_submodule) {
 204                        ce_mark_uptodate(ce);
 205                        if (!DIFF_OPT_TST(&revs->diffopt, FIND_COPIES_HARDER))
 206                                continue;
 207                }
 208                oldmode = ce->ce_mode;
 209                newmode = ce_mode_from_stat(ce, st.st_mode);
 210                diff_change(&revs->diffopt, oldmode, newmode,
 211                            ce->sha1, (changed ? null_sha1 : ce->sha1),
 212                            ce->name, 0, dirty_submodule);
 213
 214        }
 215        diffcore_std(&revs->diffopt);
 216        diff_flush(&revs->diffopt);
 217        return 0;
 218}
 219
 220/*
 221 * diff-index
 222 */
 223
 224/* A file entry went away or appeared */
 225static void diff_index_show_file(struct rev_info *revs,
 226                                 const char *prefix,
 227                                 struct cache_entry *ce,
 228                                 const unsigned char *sha1, unsigned int mode,
 229                                 unsigned dirty_submodule)
 230{
 231        diff_addremove(&revs->diffopt, prefix[0], mode,
 232                       sha1, ce->name, dirty_submodule);
 233}
 234
 235static int get_stat_data(struct cache_entry *ce,
 236                         const unsigned char **sha1p,
 237                         unsigned int *modep,
 238                         int cached, int match_missing,
 239                         unsigned *dirty_submodule, struct diff_options *diffopt)
 240{
 241        const unsigned char *sha1 = ce->sha1;
 242        unsigned int mode = ce->ce_mode;
 243
 244        if (!cached && !ce_uptodate(ce)) {
 245                int changed;
 246                struct stat st;
 247                changed = check_removed(ce, &st);
 248                if (changed < 0)
 249                        return -1;
 250                else if (changed) {
 251                        if (match_missing) {
 252                                *sha1p = sha1;
 253                                *modep = mode;
 254                                return 0;
 255                        }
 256                        return -1;
 257                }
 258                changed = match_stat_with_submodule(diffopt, ce, &st,
 259                                                    0, dirty_submodule);
 260                if (changed) {
 261                        mode = ce_mode_from_stat(ce, st.st_mode);
 262                        sha1 = null_sha1;
 263                }
 264        }
 265
 266        *sha1p = sha1;
 267        *modep = mode;
 268        return 0;
 269}
 270
 271static void show_new_file(struct rev_info *revs,
 272                          struct cache_entry *new,
 273                          int cached, int match_missing)
 274{
 275        const unsigned char *sha1;
 276        unsigned int mode;
 277        unsigned dirty_submodule = 0;
 278
 279        /*
 280         * New file in the index: it might actually be different in
 281         * the working copy.
 282         */
 283        if (get_stat_data(new, &sha1, &mode, cached, match_missing,
 284            &dirty_submodule, &revs->diffopt) < 0)
 285                return;
 286
 287        diff_index_show_file(revs, "+", new, sha1, mode, dirty_submodule);
 288}
 289
 290static int show_modified(struct rev_info *revs,
 291                         struct cache_entry *old,
 292                         struct cache_entry *new,
 293                         int report_missing,
 294                         int cached, int match_missing)
 295{
 296        unsigned int mode, oldmode;
 297        const unsigned char *sha1;
 298        unsigned dirty_submodule = 0;
 299
 300        if (get_stat_data(new, &sha1, &mode, cached, match_missing,
 301                          &dirty_submodule, &revs->diffopt) < 0) {
 302                if (report_missing)
 303                        diff_index_show_file(revs, "-", old,
 304                                             old->sha1, old->ce_mode, 0);
 305                return -1;
 306        }
 307
 308        if (revs->combine_merges && !cached &&
 309            (hashcmp(sha1, old->sha1) || hashcmp(old->sha1, new->sha1))) {
 310                struct combine_diff_path *p;
 311                int pathlen = ce_namelen(new);
 312
 313                p = xmalloc(combine_diff_path_size(2, pathlen));
 314                p->path = (char *) &p->parent[2];
 315                p->next = NULL;
 316                p->len = pathlen;
 317                memcpy(p->path, new->name, pathlen);
 318                p->path[pathlen] = 0;
 319                p->mode = mode;
 320                hashclr(p->sha1);
 321                memset(p->parent, 0, 2 * sizeof(struct combine_diff_parent));
 322                p->parent[0].status = DIFF_STATUS_MODIFIED;
 323                p->parent[0].mode = new->ce_mode;
 324                hashcpy(p->parent[0].sha1, new->sha1);
 325                p->parent[1].status = DIFF_STATUS_MODIFIED;
 326                p->parent[1].mode = old->ce_mode;
 327                hashcpy(p->parent[1].sha1, old->sha1);
 328                show_combined_diff(p, 2, revs->dense_combined_merges, revs);
 329                free(p);
 330                return 0;
 331        }
 332
 333        oldmode = old->ce_mode;
 334        if (mode == oldmode && !hashcmp(sha1, old->sha1) && !dirty_submodule &&
 335            !DIFF_OPT_TST(&revs->diffopt, FIND_COPIES_HARDER))
 336                return 0;
 337
 338        diff_change(&revs->diffopt, oldmode, mode,
 339                    old->sha1, sha1, old->name, 0, dirty_submodule);
 340        return 0;
 341}
 342
 343/*
 344 * This gets a mix of an existing index and a tree, one pathname entry
 345 * at a time. The index entry may be a single stage-0 one, but it could
 346 * also be multiple unmerged entries (in which case idx_pos/idx_nr will
 347 * give you the position and number of entries in the index).
 348 */
 349static void do_oneway_diff(struct unpack_trees_options *o,
 350        struct cache_entry *idx,
 351        struct cache_entry *tree)
 352{
 353        struct rev_info *revs = o->unpack_data;
 354        int match_missing, cached;
 355
 356        /* if the entry is not checked out, don't examine work tree */
 357        cached = o->index_only ||
 358                (idx && ((idx->ce_flags & CE_VALID) || ce_skip_worktree(idx)));
 359        /*
 360         * Backward compatibility wart - "diff-index -m" does
 361         * not mean "do not ignore merges", but "match_missing".
 362         *
 363         * But with the revision flag parsing, that's found in
 364         * "!revs->ignore_merges".
 365         */
 366        match_missing = !revs->ignore_merges;
 367
 368        if (cached && idx && ce_stage(idx)) {
 369                diff_unmerge(&revs->diffopt, idx->name, idx->ce_mode,
 370                             idx->sha1);
 371                return;
 372        }
 373
 374        /*
 375         * Something added to the tree?
 376         */
 377        if (!tree) {
 378                show_new_file(revs, idx, cached, match_missing);
 379                return;
 380        }
 381
 382        /*
 383         * Something removed from the tree?
 384         */
 385        if (!idx) {
 386                diff_index_show_file(revs, "-", tree, tree->sha1, tree->ce_mode, 0);
 387                return;
 388        }
 389
 390        /* Show difference between old and new */
 391        show_modified(revs, tree, idx, 1, cached, match_missing);
 392}
 393
 394/*
 395 * The unpack_trees() interface is designed for merging, so
 396 * the different source entries are designed primarily for
 397 * the source trees, with the old index being really mainly
 398 * used for being replaced by the result.
 399 *
 400 * For diffing, the index is more important, and we only have a
 401 * single tree.
 402 *
 403 * We're supposed to advance o->pos to skip what we have already processed.
 404 *
 405 * This wrapper makes it all more readable, and takes care of all
 406 * the fairly complex unpack_trees() semantic requirements, including
 407 * the skipping, the path matching, the type conflict cases etc.
 408 */
 409static int oneway_diff(struct cache_entry **src, struct unpack_trees_options *o)
 410{
 411        struct cache_entry *idx = src[0];
 412        struct cache_entry *tree = src[1];
 413        struct rev_info *revs = o->unpack_data;
 414
 415        /*
 416         * Unpack-trees generates a DF/conflict entry if
 417         * there was a directory in the index and a tree
 418         * in the tree. From a diff standpoint, that's a
 419         * delete of the tree and a create of the file.
 420         */
 421        if (tree == o->df_conflict_entry)
 422                tree = NULL;
 423
 424        if (ce_path_match(idx ? idx : tree, revs->prune_data))
 425                do_oneway_diff(o, idx, tree);
 426
 427        return 0;
 428}
 429
 430int run_diff_index(struct rev_info *revs, int cached)
 431{
 432        struct object *ent;
 433        struct tree *tree;
 434        const char *tree_name;
 435        struct unpack_trees_options opts;
 436        struct tree_desc t;
 437
 438        ent = revs->pending.objects[0].item;
 439        tree_name = revs->pending.objects[0].name;
 440        tree = parse_tree_indirect(ent->sha1);
 441        if (!tree)
 442                return error("bad tree object %s", tree_name);
 443
 444        memset(&opts, 0, sizeof(opts));
 445        opts.head_idx = 1;
 446        opts.index_only = cached;
 447        opts.diff_index_cached = (cached &&
 448                                  !DIFF_OPT_TST(&revs->diffopt, FIND_COPIES_HARDER));
 449        opts.merge = 1;
 450        opts.fn = oneway_diff;
 451        opts.unpack_data = revs;
 452        opts.src_index = &the_index;
 453        opts.dst_index = NULL;
 454
 455        init_tree_desc(&t, tree->buffer, tree->size);
 456        if (unpack_trees(1, &t, &opts))
 457                exit(128);
 458
 459        diff_set_mnemonic_prefix(&revs->diffopt, "c/", cached ? "i/" : "w/");
 460        diffcore_fix_diff_index(&revs->diffopt);
 461        diffcore_std(&revs->diffopt);
 462        diff_flush(&revs->diffopt);
 463        return 0;
 464}
 465
 466int do_diff_cache(const unsigned char *tree_sha1, struct diff_options *opt)
 467{
 468        struct tree *tree;
 469        struct rev_info revs;
 470        int i;
 471        struct cache_entry **dst;
 472        struct cache_entry *last = NULL;
 473        struct unpack_trees_options opts;
 474        struct tree_desc t;
 475
 476        /*
 477         * This is used by git-blame to run diff-cache internally;
 478         * it potentially needs to repeatedly run this, so we will
 479         * start by removing the higher order entries the last round
 480         * left behind.
 481         */
 482        dst = active_cache;
 483        for (i = 0; i < active_nr; i++) {
 484                struct cache_entry *ce = active_cache[i];
 485                if (ce_stage(ce)) {
 486                        if (last && !strcmp(ce->name, last->name))
 487                                continue;
 488                        cache_tree_invalidate_path(active_cache_tree,
 489                                                   ce->name);
 490                        last = ce;
 491                        ce->ce_flags |= CE_REMOVE;
 492                }
 493                *dst++ = ce;
 494        }
 495        active_nr = dst - active_cache;
 496
 497        init_revisions(&revs, NULL);
 498        revs.prune_data = opt->paths;
 499        tree = parse_tree_indirect(tree_sha1);
 500        if (!tree)
 501                die("bad tree object %s", sha1_to_hex(tree_sha1));
 502
 503        memset(&opts, 0, sizeof(opts));
 504        opts.head_idx = 1;
 505        opts.index_only = 1;
 506        opts.diff_index_cached = !DIFF_OPT_TST(opt, FIND_COPIES_HARDER);
 507        opts.merge = 1;
 508        opts.fn = oneway_diff;
 509        opts.unpack_data = &revs;
 510        opts.src_index = &the_index;
 511        opts.dst_index = &the_index;
 512
 513        init_tree_desc(&t, tree->buffer, tree->size);
 514        if (unpack_trees(1, &t, &opts))
 515                exit(128);
 516        return 0;
 517}
 518
 519int index_differs_from(const char *def, int diff_flags)
 520{
 521        struct rev_info rev;
 522        struct setup_revision_opt opt;
 523
 524        init_revisions(&rev, NULL);
 525        memset(&opt, 0, sizeof(opt));
 526        opt.def = def;
 527        setup_revisions(0, NULL, &rev, &opt);
 528        DIFF_OPT_SET(&rev.diffopt, QUICK);
 529        DIFF_OPT_SET(&rev.diffopt, EXIT_WITH_STATUS);
 530        rev.diffopt.flags |= diff_flags;
 531        run_diff_index(&rev, 1);
 532        if (rev.pending.alloc)
 533                free(rev.pending.objects);
 534        return (DIFF_OPT_TST(&rev.diffopt, HAS_CHANGES) != 0);
 535}