tree-diff.c

   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  */
  22 static 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 */
  42 static 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  */
  80 static 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
 128 static 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
 144 static int diff_tree(struct tree_desc *t1, struct tree_desc *t2,
 145                      const char *base_str, struct diff_options *opt)
 146 {
 147         struct strbuf base;
 148         int baselen = strlen(base_str);
 149
 150         /* Enable recursion indefinitely */
 151         opt->pathspec.recursive = DIFF_OPT_TST(opt, RECURSIVE);
 152
 153         strbuf_init(&base, PATH_MAX);
 154         strbuf_add(&base, base_str, baselen);
 155
 156         for (;;) {
 157                 int cmp;
 158
 159                 if (diff_can_quit_early(opt))
 160                         break;
 161                 if (opt->pathspec.nr) {
 162                         skip_uninteresting(t1, &base, opt);
 163                         skip_uninteresting(t2, &base, opt);
 164                 }
 165                 if (!t1->size && !t2->size)
 166                         break;
 167
 168                 cmp = tree_entry_pathcmp(t1, t2);
 169
 170                 /* t1 = t2 */
 171                 if (cmp == 0) {
 172                         if (DIFF_OPT_TST(opt, FIND_COPIES_HARDER) ||
 173                             hashcmp(t1->entry.sha1, t2->entry.sha1) ||
 174                             (t1->entry.mode != t2->entry.mode))
 175                                 show_path(&base, opt, t1, t2);
 176
 177                         update_tree_entry(t1);
 178                         update_tree_entry(t2);
 179                 }
 180
 181                 /* t1 < t2 */
 182                 else if (cmp < 0) {
 183                         show_path(&base, opt, t1, /*t2=*/NULL);
 184                         update_tree_entry(t1);
 185                 }
 186
 187                 /* t1 > t2 */
 188                 else {
 189                         show_path(&base, opt, /*t1=*/NULL, t2);
 190                         update_tree_entry(t2);
 191                 }
 192         }
 193
 194         strbuf_release(&base);
 195         return 0;
 196 }
 197
 198 /*
 199  * Does it look like the resulting diff might be due to a rename?
 200  *  - single entry
 201  *  - not a valid previous file
 202  */
 203 static inline int diff_might_be_rename(void)
 204 {
 205         return diff_queued_diff.nr == 1 &&
 206                 !DIFF_FILE_VALID(diff_queued_diff.queue[0]->one);
 207 }
 208
 209 static void try_to_follow_renames(struct tree_desc *t1, struct tree_desc *t2, const char *base, struct diff_options *opt)
 210 {
 211         struct diff_options diff_opts;
 212         struct diff_queue_struct *q = &diff_queued_diff;
 213         struct diff_filepair *choice;
 214         int i;
 215
 216         /*
 217          * follow-rename code is very specific, we need exactly one
 218          * path. Magic that matches more than one path is not
 219          * supported.
 220          */
 221         GUARD_PATHSPEC(&opt->pathspec, PATHSPEC_FROMTOP | PATHSPEC_LITERAL);
 222 #if 0
 223         /*
 224          * We should reject wildcards as well. Unfortunately we
 225          * haven't got a reliable way to detect that 'foo\*bar' in
 226          * fact has no wildcards. nowildcard_len is merely a hint for
 227          * optimization. Let it slip for now until wildmatch is taught
 228          * about dry-run mode and returns wildcard info.
 229          */
 230         if (opt->pathspec.has_wildcard)
 231                 die("BUG:%s:%d: wildcards are not supported",
 232                     __FILE__, __LINE__);
 233 #endif
 234
 235         /* Remove the file creation entry from the diff queue, and remember it */
 236         choice = q->queue[0];
 237         q->nr = 0;
 238
 239         diff_setup(&diff_opts);
 240         DIFF_OPT_SET(&diff_opts, RECURSIVE);
 241         DIFF_OPT_SET(&diff_opts, FIND_COPIES_HARDER);
 242         diff_opts.output_format = DIFF_FORMAT_NO_OUTPUT;
 243         diff_opts.single_follow = opt->pathspec.items[0].match;
 244         diff_opts.break_opt = opt->break_opt;
 245         diff_opts.rename_score = opt->rename_score;
 246         diff_setup_done(&diff_opts);
 247         diff_tree(t1, t2, base, &diff_opts);
 248         diffcore_std(&diff_opts);
 249         free_pathspec(&diff_opts.pathspec);
 250
 251         /* Go through the new set of filepairing, and see if we find a more interesting one */
 252         opt->found_follow = 0;
 253         for (i = 0; i < q->nr; i++) {
 254                 struct diff_filepair *p = q->queue[i];
 255
 256                 /*
 257                  * Found a source? Not only do we use that for the new
 258                  * diff_queued_diff, we will also use that as the path in
 259                  * the future!
 260                  */
 261                 if ((p->status == 'R' || p->status == 'C') &&
 262                     !strcmp(p->two->path, opt->pathspec.items[0].match)) {
 263                         const char *path[2];
 264
 265                         /* Switch the file-pairs around */
 266                         q->queue[i] = choice;
 267                         choice = p;
 268
 269                         /* Update the path we use from now on.. */
 270                         path[0] = p->one->path;
 271                         path[1] = NULL;
 272                         free_pathspec(&opt->pathspec);
 273                         parse_pathspec(&opt->pathspec,
 274                                        PATHSPEC_ALL_MAGIC & ~PATHSPEC_LITERAL,
 275                                        PATHSPEC_LITERAL_PATH, "", path);
 276
 277                         /*
 278                          * The caller expects us to return a set of vanilla
 279                          * filepairs to let a later call to diffcore_std()
 280                          * it makes to sort the renames out (among other
 281                          * things), but we already have found renames
 282                          * ourselves; signal diffcore_std() not to muck with
 283                          * rename information.
 284                          */
 285                         opt->found_follow = 1;
 286                         break;
 287                 }
 288         }
 289
 290         /*
 291          * Then, discard all the non-relevant file pairs...
 292          */
 293         for (i = 0; i < q->nr; i++) {
 294                 struct diff_filepair *p = q->queue[i];
 295                 diff_free_filepair(p);
 296         }
 297
 298         /*
 299          * .. and re-instate the one we want (which might be either the
 300          * original one, or the rename/copy we found)
 301          */
 302         q->queue[0] = choice;
 303         q->nr = 1;
 304 }
 305
 306 int diff_tree_sha1(const unsigned char *old, const unsigned char *new, const char *base, struct diff_options *opt)
 307 {
 308         void *tree1, *tree2;
 309         struct tree_desc t1, t2;
 310         unsigned long size1, size2;
 311         int retval;
 312
 313         tree1 = fill_tree_descriptor(&t1, old);
 314         tree2 = fill_tree_descriptor(&t2, new);
 315         size1 = t1.size;
 316         size2 = t2.size;
 317         retval = diff_tree(&t1, &t2, base, opt);
 318         if (!*base && DIFF_OPT_TST(opt, FOLLOW_RENAMES) && diff_might_be_rename()) {
 319                 init_tree_desc(&t1, tree1, size1);
 320                 init_tree_desc(&t2, tree2, size2);
 321                 try_to_follow_renames(&t1, &t2, base, opt);
 322         }
 323         free(tree1);
 324         free(tree2);
 325         return retval;
 326 }
 327
 328 int diff_root_tree_sha1(const unsigned char *new, const char *base, struct diff_options *opt)
 329 {
 330         return diff_tree_sha1(NULL, new, base, opt);
 331 }