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