Fossil

merge3.c at tip
Login

merge3.c at tip

File src/merge3.c from the latest check-in


     1
     2
     3
     4
     5
     6
     7
     8
     9
    10
    11
    12
    13
    14
    15
    16
    17
    18
    19
    20
    21
    22
    23
    24
    25
    26
    27
    28
    29
    30
    31
    32
    33
    34
    35
    36
    37
    38
    39
    40
    41
    42
    43
    44
    45
    46
    47
    48
    49
    50
    51
    52
    53
    54
    55
    56
    57
    58
    59
    60
    61
    62
    63
    64
    65
    66
    67
    68
    69
    70
    71
    72
    73
    74
    75
    76
    77
    78
    79
    80
    81
    82
    83
    84
    85
    86
    87
    88
    89
    90
    91
    92
    93
    94
    95
    96
    97
    98
    99
   100
   101
   102
   103
   104
   105
   106
   107
   108
   109
   110
   111
   112
   113
   114
   115
   116
   117
   118
   119
   120
   121
   122
   123
   124
   125
   126
   127
   128
   129
   130
   131
   132
   133
   134
   135
   136
   137
   138
   139
   140
   141
   142
   143
   144
   145
   146
   147
   148
   149
   150
   151
   152
   153
   154
   155
   156
   157
   158
   159
   160
   161
   162
   163
   164
   165
   166
   167
   168
   169
   170
   171
   172
   173
   174
   175
   176
   177
   178
   179
   180
   181
   182
   183
   184
   185
   186
   187
   188
   189
   190
   191
   192
   193
   194
   195
   196
   197
   198
   199
   200
   201
   202
   203
   204
   205
   206
   207
   208
   209
   210
   211
   212
   213
   214
   215
   216
   217
   218
   219
   220
   221
   222
   223
   224
   225
   226
   227
   228
   229
   230
   231
   232
   233
   234
   235
   236
   237
   238
   239
   240
   241
   242
   243
   244
   245
   246
   247
   248
   249
   250
   251
   252
   253
   254
   255
   256
   257
   258
   259
   260
   261
   262
   263
   264
   265
   266
   267
   268
   269
   270
   271
   272
   273
   274
   275
   276
   277
   278
   279
   280
   281
   282
   283
   284
   285
   286
   287
   288
   289
   290
   291
   292
   293
   294
   295
   296
   297
   298
   299
   300
   301
   302
   303
   304
   305
   306
   307
   308
   309
   310
   311
   312
   313
   314
   315
   316
   317
   318
   319
   320
   321
   322
   323
   324
   325
   326
   327
   328
   329
   330
   331
   332
   333
   334
   335
   336
   337
   338
   339
   340
   341
   342
   343
   344
   345
   346
   347
   348
   349
   350
   351
   352
   353
   354
   355
   356
   357
   358
   359
   360
   361
   362
   363
   364
   365
   366
   367
   368
   369
   370
   371
   372
   373
   374
   375
   376
   377
   378
   379
   380
   381
   382
   383
   384
   385
   386
   387
   388
   389
   390
   391
   392
   393
   394
   395
   396
   397
   398
   399
   400
   401
   402
   403
   404
   405
   406
   407
   408
   409
   410
   411
   412
   413
   414
   415
   416
   417
   418
   419
   420
   421
   422
   423
   424
   425
   426
   427
   428
   429
   430
   431
   432
   433
   434
   435
   436
   437
   438
   439
   440
   441
   442
   443
   444
   445
   446
   447
   448
   449
   450
   451
   452
   453
   454
   455
   456
   457
   458
   459
   460
   461
   462
   463
   464
   465
   466
   467
   468
   469
   470
   471
   472
   473
   474
   475
   476
   477
   478
   479
   480
   481
   482
   483
   484
   485
   486
   487
   488
   489
   490
   491
   492
   493
   494
   495
   496
   497
   498
   499
   500
   501
   502
   503
   504
   505
   506
   507
   508
   509
   510
   511
   512
   513
   514
   515
   516
   517
   518
   519
   520
   521
   522
   523
   524
   525
   526
   527
   528
   529
   530
   531
   532
   533
   534
   535
   536
   537
   538
   539
   540
   541
   542
   543
   544
   545
   546
   547
   548
   549
   550
   551
   552
   553
   554
   555
   556
   557
   558
   559
   560
   561
   562
   563
   564
   565
   566
   567
   568
   569
   570
   571
   572
   573
   574
   575
   576
   577
   578
   579
   580
   581
   582
   583
   584
   585
   586
   587
   588
   589
   590
   591
   592
   593
   594
   595
   596
/*
** Copyright (c) 2007 D. Richard Hipp
**
** This program is free software; you can redistribute it and/or
** modify it under the terms of the Simplified BSD License (also
** known as the "2-Clause License" or "FreeBSD License".)

** This program is distributed in the hope that it will be useful,
** but without any warranty; without even the implied warranty of
** merchantability or fitness for a particular purpose.
**
** Author contact information:
**   drh@hwaci.com
**   http://www.hwaci.com/drh/
**
*******************************************************************************
**
** This module implements a 3-way merge
*/
#include "config.h"
#include "merge3.h"

#if 0
#define DEBUG(X)  X
#define ISDEBUG 1
#else
#define DEBUG(X)
#define ISDEBUG 0
#endif

/* The minimum of two integers */
#ifndef min
#  define min(A,B)  (A<B?A:B)
#endif

/*
** Compare N lines of text from pV1 and pV2.  If the lines
** are the same, return true.  Return false if one or more of the N
** lines are different.
**
** The cursors on both pV1 and pV2 is unchanged by this comparison.
*/
static int sameLines(Blob *pV1, Blob *pV2, int N){
  char *z1, *z2;
  int i;
  char c;

  if( N==0 ) return 1;
  z1 = &blob_buffer(pV1)[blob_tell(pV1)];
  z2 = &blob_buffer(pV2)[blob_tell(pV2)];
  for(i=0; (c=z1[i])==z2[i]; i++){
    if( c=='\n' || c==0 ){
      N--;
      if( N==0 || c==0 ) return 1;
    }
  }
  return 0;
}

/*
** Look at the next edit triple in both aC1 and aC2.  (An "edit triple" is
** three integers describing the number of copies, deletes, and inserts in
** moving from the original to the edited copy of the file.) If the three
** integers of the edit triples describe an identical edit, then return 1.
** If the edits are different, return 0.
*/
static int sameEdit(
  int *aC1,      /* Array of edit integers for file 1 */
  int *aC2,      /* Array of edit integers for file 2 */
  Blob *pV1,     /* Text of file 1 */
  Blob *pV2      /* Text of file 2 */
){
  if( aC1[0]!=aC2[0] ) return 0;
  if( aC1[1]!=aC2[1] ) return 0;
  if( aC1[2]!=aC2[2] ) return 0;
  if( sameLines(pV1, pV2, aC1[2]) ) return 1;
  return 0;
}

/*
** The aC[] array contains triples of integers.  Within each triple, the
** elements are:
**
**   (0)  The number of lines to copy
**   (1)  The number of lines to delete
**   (2)  The number of liens to insert
**
** Suppose we want to advance over sz lines of the original file.  This routine
** returns true if that advance would land us on a copy operation.  It
** returns false if the advance would end on a delete.
*/
static int ends_at_CPY(int *aC, int sz){
  while( sz>0 && (aC[0]>0 || aC[1]>0 || aC[2]>0) ){
    if( aC[0]>=sz ) return 1;
    sz -= aC[0];
    if( aC[1]>sz ) return 0;
    sz -= aC[1];
    aC += 3;
  }
  return 1;
}

/*
** pSrc contains an edited file where aC[] describes the edit.  Part of
** pSrc has already been output.  This routine outputs additional lines
** of pSrc - lines that correspond to the next sz lines of the original
** unedited file.
**
** Note that sz counts the number of lines of text in the original file.
** But text is output from the edited file.  So the number of lines transfer
** to pOut might be different from sz.  Fewer lines appear in pOut if there
** are deletes.  More lines appear if there are inserts.
**
** The aC[] array is updated and the new index into aC[] is returned.
*/
static int output_one_side(
  Blob *pOut,     /* Write to this blob */
  Blob *pSrc,     /* The edited file that is to be copied to pOut */
  int *aC,        /* Array of integer triples describing the edit */
  int i,          /* Index in aC[] of current location in pSrc */
  int sz,         /* Number of lines in unedited source to output */
  int *pLn        /* Line number counter */
){
  while( sz>0 ){
    if( aC[i]==0 && aC[i+1]==0 && aC[i+2]==0 ) break;
    if( aC[i]>=sz ){
      blob_copy_lines(pOut, pSrc, sz);  *pLn += sz;
      aC[i] -= sz;
      break;
    }
    blob_copy_lines(pOut, pSrc, aC[i]);      *pLn += aC[i];
    blob_copy_lines(pOut, pSrc, aC[i+2]);    *pLn += aC[i+2];
    sz -= aC[i] + aC[i+1];
    i += 3;
  }
  return i;
}

/*
** Text of boundary markers for merge conflicts.
*/
static const char *const mergeMarker[] = {
 /*123456789 123456789 123456789 123456789 123456789 123456789 123456789*/
  "<<<<<<< BEGIN MERGE CONFLICT: local copy shown first <<<<<<<<<<<<",
  "||||||| COMMON ANCESTOR content follows |||||||||||||||||||||||||",
  "======= MERGED IN content follows ===============================",
  ">>>>>>> END MERGE CONFLICT >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>"
};

/*
** Return true if the input blob contains any CR/LF pairs on the first
** ten lines. This should be enough to detect files that use mainly CR/LF
** line endings without causing a performance impact for LF only files.
*/
int contains_crlf(Blob *p){
  int i;
  int j = 0;
  const int maxL = 10;             /* Max lines to check */
  const char *z = blob_buffer(p);
  int n = blob_size(p)+1;
  for(i=1; i<n; ){
    if( z[i-1]=='\r' && z[i]=='\n' ) return 1;
    while( i<n && z[i]!='\n' ){ i++; }
    j++;
    if( j>maxL ) return 0;
  }
  return 0;
}

/*
** Ensure that the text in pBlob ends with a new line.
** If useCrLf is true adds "\r\n" otherwise '\n'.
*/
void ensure_line_end(Blob *pBlob, int useCrLf){
  if( pBlob->nUsed<=0 ) return;
  if( pBlob->aData[pBlob->nUsed-1]!='\n' ){
    if( useCrLf ) blob_append_char(pBlob, '\r');
    blob_append_char(pBlob, '\n');
  }
}

/*
** Write out one of the four merge-marks.
*/
void append_merge_mark(Blob *pOut, int iMark, int ln, int useCrLf){
  ensure_line_end(pOut, useCrLf);
  blob_append(pOut, mergeMarker[iMark], -1);
  if( ln>0 ) blob_appendf(pOut, " (line %d)", ln);
  ensure_line_end(pOut, useCrLf);
}

/*
** Do a three-way merge.  Initialize pOut to contain the result.
**
** The merge is an edit against pV2.  Both pV1 and pV2 have a
** common origin at pPivot.  Apply the changes of pPivot ==> pV1
** to pV2.
**
** The return is 0 upon complete success. If any input file is binary,
** -1 is returned and pOut is unmodified.  If there are merge
** conflicts, the merge proceeds as best as it can and the number
** of conflicts is returns
*/
static int blob_merge(Blob *pPivot, Blob *pV1, Blob *pV2, Blob *pOut){
  int *aC1;              /* Changes from pPivot to pV1 */
  int *aC2;              /* Changes from pPivot to pV2 */
  int i1, i2;            /* Index into aC1[] and aC2[] */
  int nCpy, nDel, nIns;  /* Number of lines to copy, delete, or insert */
  int limit1, limit2;    /* Sizes of aC1[] and aC2[] */
  int nConflict = 0;     /* Number of merge conflicts seen so far */
  int useCrLf = 0;
  int ln1, ln2, lnPivot; /* Line numbers for all files */
  DiffConfig DCfg;

  blob_zero(pOut);         /* Merge results stored in pOut */

  /* If both pV1 and pV2 start with a UTF-8 byte-order-mark (BOM),
  ** keep it in the output. This should be secure enough not to cause
  ** unintended changes to the merged file and consistent with what
  ** users are using in their source files.
  */
  if( starts_with_utf8_bom(pV1, 0) && starts_with_utf8_bom(pV2, 0) ){
    blob_append(pOut, (char*)get_utf8_bom(0), -1);
  }

  /* Check once to see if both pV1 and pV2 contains CR/LF endings.
  ** If true, CR/LF pair will be used later to append the
  ** boundary markers for merge conflicts.
  */
  if( contains_crlf(pV1) && contains_crlf(pV2) ){
    useCrLf = 1;
  }

  /* Compute the edits that occur from pPivot => pV1 (into aC1)
  ** and pPivot => pV2 (into aC2).  Each of the aC1 and aC2 arrays is
  ** an array of integer triples.  Within each triple, the first integer
  ** is the number of lines of text to copy directly from the pivot,
  ** the second integer is the number of lines of text to omit from the
  ** pivot, and the third integer is the number of lines of text that are
  ** inserted.  The edit array ends with a triple of 0,0,0.
  */
  diff_config_init(&DCfg, 0);
  aC1 = text_diff(pPivot, pV1, 0, &DCfg);
  aC2 = text_diff(pPivot, pV2, 0, &DCfg);
  if( aC1==0 || aC2==0 ){
    free(aC1);
    free(aC2);
    return -1;
  }

  blob_rewind(pV1);        /* Rewind inputs:  Needed to reconstruct output */
  blob_rewind(pV2);
  blob_rewind(pPivot);

  /* Determine the length of the aC1[] and aC2[] change vectors */
  for(i1=0; aC1[i1] || aC1[i1+1] || aC1[i1+2]; i1+=3){}
  limit1 = i1;
  for(i2=0; aC2[i2] || aC2[i2+1] || aC2[i2+2]; i2+=3){}
  limit2 = i2;

  DEBUG(
    for(i1=0; i1<limit1; i1+=3){
      printf("c1: %4d %4d %4d\n", aC1[i1], aC1[i1+1], aC1[i1+2]);
    }
    for(i2=0; i2<limit2; i2+=3){
      printf("c2: %4d %4d %4d\n", aC2[i2], aC2[i2+1], aC2[i2+2]);
    }
  )

  /* Loop over the two edit vectors and use them to compute merged text
  ** which is written into pOut.  i1 and i2 are multiples of 3 which are
  ** indices into aC1[] and aC2[] to the edit triple currently being
  ** processed
  */
  i1 = i2 = 0;
  ln1 = ln2 = lnPivot = 1;
  while( i1<limit1 && i2<limit2 ){
    DEBUG( printf("%d: %2d %2d %2d   %d: %2d %2d %2d\n",
           i1/3, aC1[i1], aC1[i1+1], aC1[i1+2],
           i2/3, aC2[i2], aC2[i2+1], aC2[i2+2]); )

    if( aC1[i1]>0 && aC2[i2]>0 ){
      /* Output text that is unchanged in both V1 and V2 */
      nCpy = min(aC1[i1], aC2[i2]);
      DEBUG( printf("COPY %d\n", nCpy); )
      blob_copy_lines(pOut, pPivot, nCpy); lnPivot += nCpy;
      blob_copy_lines(0, pV1, nCpy);       ln1 += nCpy;
      blob_copy_lines(0, pV2, nCpy);       ln2 += nCpy;
      aC1[i1] -= nCpy;
      aC2[i2] -= nCpy;
    }else
    if( aC1[i1] >= aC2[i2+1] && aC1[i1]>0 && aC2[i2+1]+aC2[i2+2]>0 ){
      /* Output edits to V2 that occurs within unchanged regions of V1 */
      nDel = aC2[i2+1];
      nIns = aC2[i2+2];
      DEBUG( printf("EDIT -%d+%d left\n", nDel, nIns); )
      blob_copy_lines(0, pPivot, nDel);    lnPivot += nDel;
      blob_copy_lines(0, pV1, nDel);       ln1 += nDel;
      blob_copy_lines(pOut, pV2, nIns);    ln2 += nIns;
      aC1[i1] -= nDel;
      i2 += 3;
    }else
    if( aC2[i2] >= aC1[i1+1] && aC2[i2]>0 && aC1[i1+1]+aC1[i1+2]>0 ){
      /* Output edits to V1 that occur within unchanged regions of V2 */
      nDel = aC1[i1+1];
      nIns = aC1[i1+2];
      DEBUG( printf("EDIT -%d+%d right\n", nDel, nIns); )
      blob_copy_lines(0, pPivot, nDel);    lnPivot += nDel;
      blob_copy_lines(0, pV2, nDel);       ln2 += nDel;
      blob_copy_lines(pOut, pV1, nIns);    ln1 += nIns;
      aC2[i2] -= nDel;
      i1 += 3;
    }else
    if( sameEdit(&aC1[i1], &aC2[i2], pV1, pV2) ){
      /* Output edits that are identical in both V1 and V2. */
      assert( aC1[i1]==0 );
      nDel = aC1[i1+1];
      nIns = aC1[i1+2];
      DEBUG( printf("EDIT -%d+%d both\n", nDel, nIns); )
      blob_copy_lines(0, pPivot, nDel);    lnPivot += nDel;
      blob_copy_lines(pOut, pV1, nIns);    ln1 += nIns;
      blob_copy_lines(0, pV2, nIns);       ln2 += nIns;
      i1 += 3;
      i2 += 3;
    }else
    {
      /* We have found a region where different edits to V1 and V2 overlap.
      ** This is a merge conflict.  Find the size of the conflict, then
      ** output both possible edits separated by distinctive marks.
      */
      int sz = 1;    /* Size of the conflict in lines */
      nConflict++;
      while( !ends_at_CPY(&aC1[i1], sz) || !ends_at_CPY(&aC2[i2], sz) ){
        sz++;
      }
      DEBUG( printf("CONFLICT %d\n", sz); )

      append_merge_mark(pOut, 0, ln1, useCrLf);
      i1 = output_one_side(pOut, pV1, aC1, i1, sz, &ln1);

      append_merge_mark(pOut, 1, lnPivot, useCrLf);
      blob_copy_lines(pOut, pPivot, sz);   lnPivot += sz;

      append_merge_mark(pOut, 2, ln2, useCrLf);
      i2 = output_one_side(pOut, pV2, aC2, i2, sz, &ln2);

      append_merge_mark(pOut, 3, -1, useCrLf);
   }

    /* If we are finished with an edit triple, advance to the next
    ** triple.
    */
    if( i1<limit1 && aC1[i1]==0 && aC1[i1+1]==0 && aC1[i1+2]==0 ) i1+=3;
    if( i2<limit2 && aC2[i2]==0 && aC2[i2+1]==0 && aC2[i2+2]==0 ) i2+=3;
  }

  /* When one of the two edit vectors reaches its end, there might still
  ** be an insert in the other edit vector.  Output this remaining
  ** insert.
  */
  DEBUG( printf("%d: %2d %2d %2d   %d: %2d %2d %2d\n",
         i1/3, aC1[i1], aC1[i1+1], aC1[i1+2],
         i2/3, aC2[i2], aC2[i2+1], aC2[i2+2]); )
  if( i1<limit1 && aC1[i1+2]>0 ){
    DEBUG( printf("INSERT +%d left\n", aC1[i1+2]); )
    blob_copy_lines(pOut, pV1, aC1[i1+2]);
  }else if( i2<limit2 && aC2[i2+2]>0 ){
    DEBUG( printf("INSERT +%d right\n", aC2[i2+2]); )
    blob_copy_lines(pOut, pV2, aC2[i2+2]);
  }

  free(aC1);
  free(aC2);
  return nConflict;
}

/*
** Return true if the input string contains a merge marker on a line by
** itself.
*/
int contains_merge_marker(Blob *p){
  int i, j;
  int len = (int)strlen(mergeMarker[0]);
  const char *z = blob_buffer(p);
  int n = blob_size(p) - len + 1;
  assert( len==(int)strlen(mergeMarker[1]) );
  assert( len==(int)strlen(mergeMarker[2]) );
  assert( len==(int)strlen(mergeMarker[3]) );
  assert( count(mergeMarker)==4 );
  for(i=0; i<n; ){
    for(j=0; j<4; j++){
      if( (memcmp(&z[i], mergeMarker[j], len)==0) ){
        return 1;
      }
    }
    while( i<n && z[i]!='\n' ){ i++; }
    while( i<n && (z[i]=='\n' || z[i]=='\r') ){ i++; }
  }
  return 0;
}

/*
** Return true if the named file contains an unresolved merge marker line.
*/
int file_contains_merge_marker(const char *zFullpath){
  Blob file;
  int rc;
  blob_read_from_file(&file, zFullpath, ExtFILE);
  rc = contains_merge_marker(&file);
  blob_reset(&file);
  return rc;
}

/*
** COMMAND: 3-way-merge*
**
** Usage: %fossil 3-way-merge BASELINE V1 V2 MERGED
**
** Inputs are files BASELINE, V1, and V2.  The file MERGED is generated
** as output.
**
** BASELINE is a common ancestor of two files V1 and V2 that have diverging
** edits.  The generated output file MERGED is the combination of all
** changes in both V1 and V2.
**
** This command has no effect on the Fossil repository.  It is a utility
** command made available for the convenience of users.  This command can
** be used, for example, to help import changes from an upstream project.
**
** Suppose an upstream project has a file named "Xup.c" which is imported
** with modifications to the local project as "Xlocal.c".  Suppose further
** that the "Xbase.c" is an exact copy of the last imported "Xup.c".
** Then to import the latest "Xup.c" while preserving all the local changes:
**
**      fossil 3-way-merge Xbase.c Xlocal.c Xup.c Xlocal.c
**      cp Xup.c Xbase.c
**      # Verify that everything still works
**      fossil commit
**
*/
void delta_3waymerge_cmd(void){
  Blob pivot, v1, v2, merged;
  int nConflict;

  /* We should be done with options.. */
  verify_all_options();

  if( g.argc!=6 ){
    usage("PIVOT V1 V2 MERGED");
  }
  if( blob_read_from_file(&pivot, g.argv[2], ExtFILE)<0 ){
    fossil_fatal("cannot read %s", g.argv[2]);
  }
  if( blob_read_from_file(&v1, g.argv[3], ExtFILE)<0 ){
    fossil_fatal("cannot read %s", g.argv[3]);
  }
  if( blob_read_from_file(&v2, g.argv[4], ExtFILE)<0 ){
    fossil_fatal("cannot read %s", g.argv[4]);
  }
  nConflict = blob_merge(&pivot, &v1, &v2, &merged);
  if( blob_write_to_file(&merged, g.argv[5])<(int)blob_size(&merged) ){
    fossil_fatal("cannot write %s", g.argv[4]);
  }
  blob_reset(&pivot);
  blob_reset(&v1);
  blob_reset(&v2);
  blob_reset(&merged);
  if( nConflict>0 ) fossil_warning("WARNING: %d merge conflicts", nConflict);
}

/*
** aSubst is an array of string pairs.  The first element of each pair is
** a string that begins with %.  The second element is a replacement for that
** string.
**
** This routine makes a copy of zInput into memory obtained from malloc and
** performance all applicable substitutions on that string.
*/
char *string_subst(const char *zInput, int nSubst, const char **azSubst){
  Blob x;
  int i, j;
  blob_zero(&x);
  while( zInput[0] ){
    for(i=0; zInput[i] && zInput[i]!='%'; i++){}
    if( i>0 ){
      blob_append(&x, zInput, i);
      zInput += i;
    }
    if( zInput[0]==0 ) break;
    for(j=0; j<nSubst; j+=2){
      int n = strlen(azSubst[j]);
      if( strncmp(zInput, azSubst[j], n)==0 ){
        blob_append(&x, azSubst[j+1], -1);
        zInput += n;
        break;
      }
    }
    if( j>=nSubst ){
      blob_append(&x, "%", 1);
      zInput++;
    }
  }
  return blob_str(&x);
}

#if INTERFACE
/*
** Flags to the 3-way merger
*/
#define MERGE_DRYRUN  0x0001
/*
** The MERGE_KEEP_FILES flag specifies that merge_3way() should retain
** its temporary files on error. By default they are removed after the
** merge, regardless of success or failure.
*/
#define MERGE_KEEP_FILES 0x0002
#endif


/*
** This routine is a wrapper around blob_merge() with the following
** enhancements:
**
**    (1) If the merge-command is defined, then use the external merging
**        program specified instead of the built-in blob-merge to do the
**        merging.  Panic if the external merger fails.
**        ** Not currently implemented **
**
**    (2) If gmerge-command is defined and there are merge conflicts in
**        blob_merge() then invoke the external graphical merger to resolve
**        the conflicts.
**
**    (3) If a merge conflict occurs and gmerge-command is not defined,
**        then write the pivot, original, and merge-in files to the
**        filesystem.
*/
int merge_3way(
  Blob *pPivot,       /* Common ancestor (older) */
  const char *zV1,    /* Name of file for version merging into (mine) */
  Blob *pV2,          /* Version merging from (yours) */
  Blob *pOut,         /* Output written here */
  unsigned mergeFlags /* Flags that control operation */
){
  Blob v1;            /* Content of zV1 */
  int rc;             /* Return code of subroutines and this routine */
  const char *zGMerge;   /* Name of the gmerge command */

  blob_read_from_file(&v1, zV1, ExtFILE);
  rc = blob_merge(pPivot, &v1, pV2, pOut);
  zGMerge = rc<=0 ? 0 : db_get("gmerge-command", 0);
  if( (mergeFlags & MERGE_DRYRUN)==0
      && ((zGMerge!=0 && zGMerge[0]!=0)
          || (rc!=0 && (mergeFlags & MERGE_KEEP_FILES)!=0)) ){
    char *zPivot;       /* Name of the pivot file */
    char *zOrig;        /* Name of the original content file */
    char *zOther;       /* Name of the merge file */

    zPivot = file_newname(zV1, "baseline", 1);
    blob_write_to_file(pPivot, zPivot);
    zOrig = file_newname(zV1, "original", 1);
    blob_write_to_file(&v1, zOrig);
    zOther = file_newname(zV1, "merge", 1);
    blob_write_to_file(pV2, zOther);
    if( rc>0 ){
      if( zGMerge && zGMerge[0] ){
        char *zOut;     /* Temporary output file */
        char *zCmd;     /* Command to invoke */
        const char *azSubst[8];  /* Strings to be substituted */
        zOut = file_newname(zV1, "output", 1);
        azSubst[0] = "%baseline";  azSubst[1] = zPivot;
        azSubst[2] = "%original";  azSubst[3] = zOrig;
        azSubst[4] = "%merge";     azSubst[5] = zOther;
        azSubst[6] = "%output";    azSubst[7] = zOut;
        zCmd = string_subst(zGMerge, 8, azSubst);
        printf("%s\n", zCmd); fflush(stdout);
        fossil_system(zCmd);
        if( file_size(zOut, RepoFILE)>=0 ){
          blob_read_from_file(pOut, zOut, ExtFILE);
          file_delete(zOut);
        }
        fossil_free(zCmd);
        fossil_free(zOut);
      }
    }
    if( (mergeFlags & MERGE_KEEP_FILES)==0 ){
      file_delete(zPivot);
      file_delete(zOrig);
      file_delete(zOther);
    }
    fossil_free(zPivot);
    fossil_free(zOrig);
    fossil_free(zOther);
  }
  blob_reset(&v1);
  return rc;
}