xdiff / xpatience.con commit merge-recursive: expose merge options for builtin merge (635a7bb)
   1/*
   2 *  LibXDiff by Davide Libenzi ( File Differential Library )
   3 *  Copyright (C) 2003-2009 Davide Libenzi, Johannes E. Schindelin
   4 *
   5 *  This library is free software; you can redistribute it and/or
   6 *  modify it under the terms of the GNU Lesser General Public
   7 *  License as published by the Free Software Foundation; either
   8 *  version 2.1 of the License, or (at your option) any later version.
   9 *
  10 *  This library is distributed in the hope that it will be useful,
  11 *  but WITHOUT ANY WARRANTY; without even the implied warranty of
  12 *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  13 *  Lesser General Public License for more details.
  14 *
  15 *  You should have received a copy of the GNU Lesser General Public
  16 *  License along with this library; if not, write to the Free Software
  17 *  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
  18 *
  19 *  Davide Libenzi <davidel@xmailserver.org>
  20 *
  21 */
  22#include "xinclude.h"
  23#include "xtypes.h"
  24#include "xdiff.h"
  25
  26/*
  27 * The basic idea of patience diff is to find lines that are unique in
  28 * both files.  These are intuitively the ones that we want to see as
  29 * common lines.
  30 *
  31 * The maximal ordered sequence of such line pairs (where ordered means
  32 * that the order in the sequence agrees with the order of the lines in
  33 * both files) naturally defines an initial set of common lines.
  34 *
  35 * Now, the algorithm tries to extend the set of common lines by growing
  36 * the line ranges where the files have identical lines.
  37 *
  38 * Between those common lines, the patience diff algorithm is applied
  39 * recursively, until no unique line pairs can be found; these line ranges
  40 * are handled by the well-known Myers algorithm.
  41 */
  42
  43#define NON_UNIQUE ULONG_MAX
  44
  45/*
  46 * This is a hash mapping from line hash to line numbers in the first and
  47 * second file.
  48 */
  49struct hashmap {
  50        int nr, alloc;
  51        struct entry {
  52                unsigned long hash;
  53                /*
  54                 * 0 = unused entry, 1 = first line, 2 = second, etc.
  55                 * line2 is NON_UNIQUE if the line is not unique
  56                 * in either the first or the second file.
  57                 */
  58                unsigned long line1, line2;
  59                /*
  60                 * "next" & "previous" are used for the longest common
  61                 * sequence;
  62                 * initially, "next" reflects only the order in file1.
  63                 */
  64                struct entry *next, *previous;
  65        } *entries, *first, *last;
  66        /* were common records found? */
  67        unsigned long has_matches;
  68        mmfile_t *file1, *file2;
  69        xdfenv_t *env;
  70        xpparam_t const *xpp;
  71};
  72
  73/* The argument "pass" is 1 for the first file, 2 for the second. */
  74static void insert_record(int line, struct hashmap *map, int pass)
  75{
  76        xrecord_t **records = pass == 1 ?
  77                map->env->xdf1.recs : map->env->xdf2.recs;
  78        xrecord_t *record = records[line - 1], *other;
  79        /*
  80         * After xdl_prepare_env() (or more precisely, due to
  81         * xdl_classify_record()), the "ha" member of the records (AKA lines)
  82         * is _not_ the hash anymore, but a linearized version of it.  In
  83         * other words, the "ha" member is guaranteed to start with 0 and
  84         * the second record's ha can only be 0 or 1, etc.
  85         *
  86         * So we multiply ha by 2 in the hope that the hashing was
  87         * "unique enough".
  88         */
  89        int index = (int)((record->ha << 1) % map->alloc);
  90
  91        while (map->entries[index].line1) {
  92                other = map->env->xdf1.recs[map->entries[index].line1 - 1];
  93                if (map->entries[index].hash != record->ha ||
  94                                !xdl_recmatch(record->ptr, record->size,
  95                                        other->ptr, other->size,
  96                                        map->xpp->flags)) {
  97                        if (++index >= map->alloc)
  98                                index = 0;
  99                        continue;
 100                }
 101                if (pass == 2)
 102                        map->has_matches = 1;
 103                if (pass == 1 || map->entries[index].line2)
 104                        map->entries[index].line2 = NON_UNIQUE;
 105                else
 106                        map->entries[index].line2 = line;
 107                return;
 108        }
 109        if (pass == 2)
 110                return;
 111        map->entries[index].line1 = line;
 112        map->entries[index].hash = record->ha;
 113        if (!map->first)
 114                map->first = map->entries + index;
 115        if (map->last) {
 116                map->last->next = map->entries + index;
 117                map->entries[index].previous = map->last;
 118        }
 119        map->last = map->entries + index;
 120        map->nr++;
 121}
 122
 123/*
 124 * This function has to be called for each recursion into the inter-hunk
 125 * parts, as previously non-unique lines can become unique when being
 126 * restricted to a smaller part of the files.
 127 *
 128 * It is assumed that env has been prepared using xdl_prepare().
 129 */
 130static int fill_hashmap(mmfile_t *file1, mmfile_t *file2,
 131                xpparam_t const *xpp, xdfenv_t *env,
 132                struct hashmap *result,
 133                int line1, int count1, int line2, int count2)
 134{
 135        result->file1 = file1;
 136        result->file2 = file2;
 137        result->xpp = xpp;
 138        result->env = env;
 139
 140        /* We know exactly how large we want the hash map */
 141        result->alloc = count1 * 2;
 142        result->entries = (struct entry *)
 143                xdl_malloc(result->alloc * sizeof(struct entry));
 144        if (!result->entries)
 145                return -1;
 146        memset(result->entries, 0, result->alloc * sizeof(struct entry));
 147
 148        /* First, fill with entries from the first file */
 149        while (count1--)
 150                insert_record(line1++, result, 1);
 151
 152        /* Then search for matches in the second file */
 153        while (count2--)
 154                insert_record(line2++, result, 2);
 155
 156        return 0;
 157}
 158
 159/*
 160 * Find the longest sequence with a smaller last element (meaning a smaller
 161 * line2, as we construct the sequence with entries ordered by line1).
 162 */
 163static int binary_search(struct entry **sequence, int longest,
 164                struct entry *entry)
 165{
 166        int left = -1, right = longest;
 167
 168        while (left + 1 < right) {
 169                int middle = (left + right) / 2;
 170                /* by construction, no two entries can be equal */
 171                if (sequence[middle]->line2 > entry->line2)
 172                        right = middle;
 173                else
 174                        left = middle;
 175        }
 176        /* return the index in "sequence", _not_ the sequence length */
 177        return left;
 178}
 179
 180/*
 181 * The idea is to start with the list of common unique lines sorted by
 182 * the order in file1.  For each of these pairs, the longest (partial)
 183 * sequence whose last element's line2 is smaller is determined.
 184 *
 185 * For efficiency, the sequences are kept in a list containing exactly one
 186 * item per sequence length: the sequence with the smallest last
 187 * element (in terms of line2).
 188 */
 189static struct entry *find_longest_common_sequence(struct hashmap *map)
 190{
 191        struct entry **sequence = xdl_malloc(map->nr * sizeof(struct entry *));
 192        int longest = 0, i;
 193        struct entry *entry;
 194
 195        for (entry = map->first; entry; entry = entry->next) {
 196                if (!entry->line2 || entry->line2 == NON_UNIQUE)
 197                        continue;
 198                i = binary_search(sequence, longest, entry);
 199                entry->previous = i < 0 ? NULL : sequence[i];
 200                sequence[++i] = entry;
 201                if (i == longest)
 202                        longest++;
 203        }
 204
 205        /* No common unique lines were found */
 206        if (!longest) {
 207                xdl_free(sequence);
 208                return NULL;
 209        }
 210
 211        /* Iterate starting at the last element, adjusting the "next" members */
 212        entry = sequence[longest - 1];
 213        entry->next = NULL;
 214        while (entry->previous) {
 215                entry->previous->next = entry;
 216                entry = entry->previous;
 217        }
 218        xdl_free(sequence);
 219        return entry;
 220}
 221
 222static int match(struct hashmap *map, int line1, int line2)
 223{
 224        xrecord_t *record1 = map->env->xdf1.recs[line1 - 1];
 225        xrecord_t *record2 = map->env->xdf2.recs[line2 - 1];
 226        return xdl_recmatch(record1->ptr, record1->size,
 227                record2->ptr, record2->size, map->xpp->flags);
 228}
 229
 230static int patience_diff(mmfile_t *file1, mmfile_t *file2,
 231                xpparam_t const *xpp, xdfenv_t *env,
 232                int line1, int count1, int line2, int count2);
 233
 234static int walk_common_sequence(struct hashmap *map, struct entry *first,
 235                int line1, int count1, int line2, int count2)
 236{
 237        int end1 = line1 + count1, end2 = line2 + count2;
 238        int next1, next2;
 239
 240        for (;;) {
 241                /* Try to grow the line ranges of common lines */
 242                if (first) {
 243                        next1 = first->line1;
 244                        next2 = first->line2;
 245                        while (next1 > line1 && next2 > line2 &&
 246                                        match(map, next1 - 1, next2 - 1)) {
 247                                next1--;
 248                                next2--;
 249                        }
 250                } else {
 251                        next1 = end1;
 252                        next2 = end2;
 253                }
 254                while (line1 < next1 && line2 < next2 &&
 255                                match(map, line1, line2)) {
 256                        line1++;
 257                        line2++;
 258                }
 259
 260                /* Recurse */
 261                if (next1 > line1 || next2 > line2) {
 262                        struct hashmap submap;
 263
 264                        memset(&submap, 0, sizeof(submap));
 265                        if (patience_diff(map->file1, map->file2,
 266                                        map->xpp, map->env,
 267                                        line1, next1 - line1,
 268                                        line2, next2 - line2))
 269                                return -1;
 270                }
 271
 272                if (!first)
 273                        return 0;
 274
 275                while (first->next &&
 276                                first->next->line1 == first->line1 + 1 &&
 277                                first->next->line2 == first->line2 + 1)
 278                        first = first->next;
 279
 280                line1 = first->line1 + 1;
 281                line2 = first->line2 + 1;
 282
 283                first = first->next;
 284        }
 285}
 286
 287static int fall_back_to_classic_diff(struct hashmap *map,
 288                int line1, int count1, int line2, int count2)
 289{
 290        /*
 291         * This probably does not work outside Git, since
 292         * we have a very simple mmfile structure.
 293         *
 294         * Note: ideally, we would reuse the prepared environment, but
 295         * the libxdiff interface does not (yet) allow for diffing only
 296         * ranges of lines instead of the whole files.
 297         */
 298        mmfile_t subfile1, subfile2;
 299        xpparam_t xpp;
 300        xdfenv_t env;
 301
 302        subfile1.ptr = (char *)map->env->xdf1.recs[line1 - 1]->ptr;
 303        subfile1.size = map->env->xdf1.recs[line1 + count1 - 2]->ptr +
 304                map->env->xdf1.recs[line1 + count1 - 2]->size - subfile1.ptr;
 305        subfile2.ptr = (char *)map->env->xdf2.recs[line2 - 1]->ptr;
 306        subfile2.size = map->env->xdf2.recs[line2 + count2 - 2]->ptr +
 307                map->env->xdf2.recs[line2 + count2 - 2]->size - subfile2.ptr;
 308        xpp.flags = map->xpp->flags & ~XDF_PATIENCE_DIFF;
 309        if (xdl_do_diff(&subfile1, &subfile2, &xpp, &env) < 0)
 310                return -1;
 311
 312        memcpy(map->env->xdf1.rchg + line1 - 1, env.xdf1.rchg, count1);
 313        memcpy(map->env->xdf2.rchg + line2 - 1, env.xdf2.rchg, count2);
 314
 315        xdl_free_env(&env);
 316
 317        return 0;
 318}
 319
 320/*
 321 * Recursively find the longest common sequence of unique lines,
 322 * and if none was found, ask xdl_do_diff() to do the job.
 323 *
 324 * This function assumes that env was prepared with xdl_prepare_env().
 325 */
 326static int patience_diff(mmfile_t *file1, mmfile_t *file2,
 327                xpparam_t const *xpp, xdfenv_t *env,
 328                int line1, int count1, int line2, int count2)
 329{
 330        struct hashmap map;
 331        struct entry *first;
 332        int result = 0;
 333
 334        /* trivial case: one side is empty */
 335        if (!count1) {
 336                while(count2--)
 337                        env->xdf2.rchg[line2++ - 1] = 1;
 338                return 0;
 339        } else if (!count2) {
 340                while(count1--)
 341                        env->xdf1.rchg[line1++ - 1] = 1;
 342                return 0;
 343        }
 344
 345        memset(&map, 0, sizeof(map));
 346        if (fill_hashmap(file1, file2, xpp, env, &map,
 347                        line1, count1, line2, count2))
 348                return -1;
 349
 350        /* are there any matching lines at all? */
 351        if (!map.has_matches) {
 352                while(count1--)
 353                        env->xdf1.rchg[line1++ - 1] = 1;
 354                while(count2--)
 355                        env->xdf2.rchg[line2++ - 1] = 1;
 356                xdl_free(map.entries);
 357                return 0;
 358        }
 359
 360        first = find_longest_common_sequence(&map);
 361        if (first)
 362                result = walk_common_sequence(&map, first,
 363                        line1, count1, line2, count2);
 364        else
 365                result = fall_back_to_classic_diff(&map,
 366                        line1, count1, line2, count2);
 367
 368        xdl_free(map.entries);
 369        return result;
 370}
 371
 372int xdl_do_patience_diff(mmfile_t *file1, mmfile_t *file2,
 373                xpparam_t const *xpp, xdfenv_t *env)
 374{
 375        if (xdl_prepare_env(file1, file2, xpp, env) < 0)
 376                return -1;
 377
 378        /* environment is cleaned up in xdl_diff() */
 379        return patience_diff(file1, file2, xpp, env,
 380                        1, env->xdf1.nrec, 1, env->xdf2.nrec);
 381}