tree-diff.con commit tree-diff: rework diff_tree interface to be sha1 based (52894e7)
   1/*
   2 * Helper functions for tree diff generation
   3 */
   4#include "cache.h"
   5#include "diff.h"
   6#include "diffcore.h"
   7#include "tree.h"
   8
   9/*
  10 * Compare two tree entries, taking into account only path/S_ISDIR(mode),
  11 * but not their sha1's.
  12 *
  13 * NOTE files and directories *always* compare differently, even when having
  14 *      the same name - thanks to base_name_compare().
  15 *
  16 * NOTE empty (=invalid) descriptor(s) take part in comparison as +infty,
  17 *      so that they sort *after* valid tree entries.
  18 *
  19 *      Due to this convention, if trees are scanned in sorted order, all
  20 *      non-empty descriptors will be processed first.
  21 */
  22static int tree_entry_pathcmp(struct tree_desc *t1, struct tree_desc *t2)
  23{
  24        struct name_entry *e1, *e2;
  25        int cmp;
  26
  27        /* empty descriptors sort after valid tree entries */
  28        if (!t1->size)
  29                return t2->size ? 1 : 0;
  30        else if (!t2->size)
  31                return -1;
  32
  33        e1 = &t1->entry;
  34        e2 = &t2->entry;
  35        cmp = base_name_compare(e1->path, tree_entry_len(e1), e1->mode,
  36                                e2->path, tree_entry_len(e2), e2->mode);
  37        return cmp;
  38}
  39
  40
  41/* convert path, t1/t2 -> opt->diff_*() callbacks */
  42static void emit_diff(struct diff_options *opt, struct strbuf *path,
  43                      struct tree_desc *t1, struct tree_desc *t2)
  44{
  45        unsigned int mode1 = t1 ? t1->entry.mode : 0;
  46        unsigned int mode2 = t2 ? t2->entry.mode : 0;
  47
  48        if (mode1 && mode2) {
  49                opt->change(opt, mode1, mode2, t1->entry.sha1, t2->entry.sha1,
  50                        1, 1, path->buf, 0, 0);
  51        }
  52        else {
  53                const unsigned char *sha1;
  54                unsigned int mode;
  55                int addremove;
  56
  57                if (mode2) {
  58                        addremove = '+';
  59                        sha1 = t2->entry.sha1;
  60                        mode = mode2;
  61                } else {
  62                        addremove = '-';
  63                        sha1 = t1->entry.sha1;
  64                        mode = mode1;
  65                }
  66
  67                opt->add_remove(opt, addremove, mode, sha1, 1, path->buf, 0);
  68        }
  69}
  70
  71
  72/* new path should be added to diff
  73 *
  74 * 3 cases on how/when it should be called and behaves:
  75 *
  76 *      !t1,  t2        -> path added, parent lacks it
  77 *       t1, !t2        -> path removed from parent
  78 *       t1,  t2        -> path modified
  79 */
  80static void show_path(struct strbuf *base, struct diff_options *opt,
  81                      struct tree_desc *t1, struct tree_desc *t2)
  82{
  83        unsigned mode;
  84        const char *path;
  85        int pathlen;
  86        int old_baselen = base->len;
  87        int isdir, recurse = 0, emitthis = 1;
  88
  89        /* at least something has to be valid */
  90        assert(t1 || t2);
  91
  92        if (t2) {
  93                /* path present in resulting tree */
  94                tree_entry_extract(t2, &path, &mode);
  95                pathlen = tree_entry_len(&t2->entry);
  96                isdir = S_ISDIR(mode);
  97        } else {
  98                /*
  99                 * a path was removed - take path from parent. Also take
 100                 * mode from parent, to decide on recursion.
 101                 */
 102                tree_entry_extract(t1, &path, &mode);
 103                pathlen = tree_entry_len(&t1->entry);
 104
 105                isdir = S_ISDIR(mode);
 106                mode = 0;
 107        }
 108
 109        if (DIFF_OPT_TST(opt, RECURSIVE) && isdir) {
 110                recurse = 1;
 111                emitthis = DIFF_OPT_TST(opt, TREE_IN_RECURSIVE);
 112        }
 113
 114        strbuf_add(base, path, pathlen);
 115
 116        if (emitthis)
 117                emit_diff(opt, base, t1, t2);
 118
 119        if (recurse) {
 120                strbuf_addch(base, '/');
 121                diff_tree_sha1(t1 ? t1->entry.sha1 : NULL,
 122                               t2 ? t2->entry.sha1 : NULL, base->buf, opt);
 123        }
 124
 125        strbuf_setlen(base, old_baselen);
 126}
 127
 128static void skip_uninteresting(struct tree_desc *t, struct strbuf *base,
 129                               struct diff_options *opt)
 130{
 131        enum interesting match;
 132
 133        while (t->size) {
 134                match = tree_entry_interesting(&t->entry, base, 0, &opt->pathspec);
 135                if (match) {
 136                        if (match == all_entries_not_interesting)
 137                                t->size = 0;
 138                        break;
 139                }
 140                update_tree_entry(t);
 141        }
 142}
 143
 144static int ll_diff_tree_sha1(const unsigned char *old, const unsigned char *new,
 145                             const char *base_str, struct diff_options *opt)
 146{
 147        struct tree_desc t1, t2;
 148        void *t1tree, *t2tree;
 149        struct strbuf base;
 150        int baselen = strlen(base_str);
 151
 152        t1tree = fill_tree_descriptor(&t1, old);
 153        t2tree = fill_tree_descriptor(&t2, new);
 154
 155        /* Enable recursion indefinitely */
 156        opt->pathspec.recursive = DIFF_OPT_TST(opt, RECURSIVE);
 157
 158        strbuf_init(&base, PATH_MAX);
 159        strbuf_add(&base, base_str, baselen);
 160
 161        for (;;) {
 162                int cmp;
 163
 164                if (diff_can_quit_early(opt))
 165                        break;
 166                if (opt->pathspec.nr) {
 167                        skip_uninteresting(&t1, &base, opt);
 168                        skip_uninteresting(&t2, &base, opt);
 169                }
 170                if (!t1.size && !t2.size)
 171                        break;
 172
 173                cmp = tree_entry_pathcmp(&t1, &t2);
 174
 175                /* t1 = t2 */
 176                if (cmp == 0) {
 177                        if (DIFF_OPT_TST(opt, FIND_COPIES_HARDER) ||
 178                            hashcmp(t1.entry.sha1, t2.entry.sha1) ||
 179                            (t1.entry.mode != t2.entry.mode))
 180                                show_path(&base, opt, &t1, &t2);
 181
 182                        update_tree_entry(&t1);
 183                        update_tree_entry(&t2);
 184                }
 185
 186                /* t1 < t2 */
 187                else if (cmp < 0) {
 188                        show_path(&base, opt, &t1, /*t2=*/NULL);
 189                        update_tree_entry(&t1);
 190                }
 191
 192                /* t1 > t2 */
 193                else {
 194                        show_path(&base, opt, /*t1=*/NULL, &t2);
 195                        update_tree_entry(&t2);
 196                }
 197        }
 198
 199        strbuf_release(&base);
 200        free(t2tree);
 201        free(t1tree);
 202        return 0;
 203}
 204
 205/*
 206 * Does it look like the resulting diff might be due to a rename?
 207 *  - single entry
 208 *  - not a valid previous file
 209 */
 210static inline int diff_might_be_rename(void)
 211{
 212        return diff_queued_diff.nr == 1 &&
 213                !DIFF_FILE_VALID(diff_queued_diff.queue[0]->one);
 214}
 215
 216static void try_to_follow_renames(const unsigned char *old, const unsigned char *new, const char *base, struct diff_options *opt)
 217{
 218        struct diff_options diff_opts;
 219        struct diff_queue_struct *q = &diff_queued_diff;
 220        struct diff_filepair *choice;
 221        int i;
 222
 223        /*
 224         * follow-rename code is very specific, we need exactly one
 225         * path. Magic that matches more than one path is not
 226         * supported.
 227         */
 228        GUARD_PATHSPEC(&opt->pathspec, PATHSPEC_FROMTOP | PATHSPEC_LITERAL);
 229#if 0
 230        /*
 231         * We should reject wildcards as well. Unfortunately we
 232         * haven't got a reliable way to detect that 'foo\*bar' in
 233         * fact has no wildcards. nowildcard_len is merely a hint for
 234         * optimization. Let it slip for now until wildmatch is taught
 235         * about dry-run mode and returns wildcard info.
 236         */
 237        if (opt->pathspec.has_wildcard)
 238                die("BUG:%s:%d: wildcards are not supported",
 239                    __FILE__, __LINE__);
 240#endif
 241
 242        /* Remove the file creation entry from the diff queue, and remember it */
 243        choice = q->queue[0];
 244        q->nr = 0;
 245
 246        diff_setup(&diff_opts);
 247        DIFF_OPT_SET(&diff_opts, RECURSIVE);
 248        DIFF_OPT_SET(&diff_opts, FIND_COPIES_HARDER);
 249        diff_opts.output_format = DIFF_FORMAT_NO_OUTPUT;
 250        diff_opts.single_follow = opt->pathspec.items[0].match;
 251        diff_opts.break_opt = opt->break_opt;
 252        diff_opts.rename_score = opt->rename_score;
 253        diff_setup_done(&diff_opts);
 254        ll_diff_tree_sha1(old, new, base, &diff_opts);
 255        diffcore_std(&diff_opts);
 256        free_pathspec(&diff_opts.pathspec);
 257
 258        /* Go through the new set of filepairing, and see if we find a more interesting one */
 259        opt->found_follow = 0;
 260        for (i = 0; i < q->nr; i++) {
 261                struct diff_filepair *p = q->queue[i];
 262
 263                /*
 264                 * Found a source? Not only do we use that for the new
 265                 * diff_queued_diff, we will also use that as the path in
 266                 * the future!
 267                 */
 268                if ((p->status == 'R' || p->status == 'C') &&
 269                    !strcmp(p->two->path, opt->pathspec.items[0].match)) {
 270                        const char *path[2];
 271
 272                        /* Switch the file-pairs around */
 273                        q->queue[i] = choice;
 274                        choice = p;
 275
 276                        /* Update the path we use from now on.. */
 277                        path[0] = p->one->path;
 278                        path[1] = NULL;
 279                        free_pathspec(&opt->pathspec);
 280                        parse_pathspec(&opt->pathspec,
 281                                       PATHSPEC_ALL_MAGIC & ~PATHSPEC_LITERAL,
 282                                       PATHSPEC_LITERAL_PATH, "", path);
 283
 284                        /*
 285                         * The caller expects us to return a set of vanilla
 286                         * filepairs to let a later call to diffcore_std()
 287                         * it makes to sort the renames out (among other
 288                         * things), but we already have found renames
 289                         * ourselves; signal diffcore_std() not to muck with
 290                         * rename information.
 291                         */
 292                        opt->found_follow = 1;
 293                        break;
 294                }
 295        }
 296
 297        /*
 298         * Then, discard all the non-relevant file pairs...
 299         */
 300        for (i = 0; i < q->nr; i++) {
 301                struct diff_filepair *p = q->queue[i];
 302                diff_free_filepair(p);
 303        }
 304
 305        /*
 306         * .. and re-instate the one we want (which might be either the
 307         * original one, or the rename/copy we found)
 308         */
 309        q->queue[0] = choice;
 310        q->nr = 1;
 311}
 312
 313int diff_tree_sha1(const unsigned char *old, const unsigned char *new, const char *base, struct diff_options *opt)
 314{
 315        int retval;
 316
 317        retval = ll_diff_tree_sha1(old, new, base, opt);
 318        if (!*base && DIFF_OPT_TST(opt, FOLLOW_RENAMES) && diff_might_be_rename())
 319                try_to_follow_renames(old, new, base, opt);
 320
 321        return retval;
 322}
 323
 324int diff_root_tree_sha1(const unsigned char *new, const char *base, struct diff_options *opt)
 325{
 326        return diff_tree_sha1(NULL, new, base, opt);
 327}