Fossil

Check-in [fa0ea2c3fa]
Login

Check-in [fa0ea2c3fa]

Many hyperlinks are disabled.
Use anonymous login to enable hyperlinks.

Overview
Comment:Redesign the control-artifact and manifest parser to run faster.
Downloads: Tarball | ZIP archive
Timelines: family | ancestors | descendants | both | experimental
Files: files | file ages | folders
SHA1: fa0ea2c3fad489fc035804b4e76be38fd0e8b5ac
User & Date: drh 2010-10-25 13:52:20.000
Context
2010-10-25
17:21
Fix bugs in the handling of deleted and added files on delta manifests. ... (check-in: 05b152aedf user: drh tags: experimental)
13:52
Redesign the control-artifact and manifest parser to run faster. ... (check-in: fa0ea2c3fa user: drh tags: experimental)
11:12
Bug fix in the manifest parser. ... (check-in: d10f5e9fee user: drh tags: experimental)
Changes
Unified Diff Ignore Whitespace Patch
Changes to src/blob.c.
︙︙
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

  while( i<n && !fossil_isspace(aData[i]) ){ i++; }
  blob_extract(pFrom, i-pFrom->iCursor, pTo);
  while( i<n && fossil_isspace(aData[i]) ){ i++; }
  pFrom->iCursor = i;
  return pTo->nUsed;
}

/*
** This is a variation on blob_token() used by the manifest parser.
** The input pFrom is a line of text that is separated into tokens
** by single space characters.  This routine extracts the next token
** into pTo, leaves pFrom pointing at the start of the token beyond the
** the extracted token, and returns the length of the extracted token.
*/
int fast_token(Blob *pFrom, Blob *pTo){
  char *aData = pFrom->aData;
  int n = pFrom->nUsed;
  int i = pFrom->iCursor;
  int x;
  if( i>=n ){
    blob_zero(pTo);
    return 0;
  }
  pTo->aData = &aData[i];
  while( i<n && aData[i]!=' ' ){ i++; }
  if( i>pFrom->iCursor && aData[i-1]=='\n' ){ i--; }
  x = i - pFrom->iCursor;
  pTo->nUsed = x;
  pTo->nAlloc = x;
  pTo->iCursor = 0;
  pTo->xRealloc = blobReallocStatic;
  pFrom->iCursor = i+1;
  return x;
}

/*
** Extract everything from the current cursor to the end of the blob
** into a new blob.  The new blob is an ephemerial reference to the
** original blob.  The cursor of the original blob is unchanged.
*/
int blob_tail(Blob *pFrom, Blob *pTo){
  int iCursor = pFrom->iCursor;








<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<








474
475
476
477
478
479
480




























481
482
483
484
485
486
487

  while( i<n && !fossil_isspace(aData[i]) ){ i++; }
  blob_extract(pFrom, i-pFrom->iCursor, pTo);
  while( i<n && fossil_isspace(aData[i]) ){ i++; }
  pFrom->iCursor = i;
  return pTo->nUsed;
}





























/*
** Extract everything from the current cursor to the end of the blob
** into a new blob.  The new blob is an ephemerial reference to the
** original blob.  The cursor of the original blob is unchanged.
*/
int blob_tail(Blob *pFrom, Blob *pTo){
  int iCursor = pFrom->iCursor;
︙︙
Changes to src/branch.c.
︙︙
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102

  zDate = date_in_standard_format(zDateOvrd ? zDateOvrd : "now");
  zDate[10] = 'T';
  blob_appendf(&branch, "D %s\n", zDate);

  /* Copy all of the content from the parent into the branch */
  for(i=0; i<pParent->nFile; ++i){
    blob_appendf(&branch, "F %F", pParent->aFile[i].zName);
    if( pParent->aFile[i].zUuid[0] ){
      blob_appendf(&branch, " %s", pParent->aFile[i].zUuid);
      if( pParent->aFile[i].zPerm[0] ){
        blob_appendf(&branch, " %s", pParent->aFile[i].zPerm);
      }
    }
    blob_append(&branch, "\n", 1);
  }








|








88
89
90
91
92
93
94
95
96
97
98
99
100
101
102

  zDate = date_in_standard_format(zDateOvrd ? zDateOvrd : "now");
  zDate[10] = 'T';
  blob_appendf(&branch, "D %s\n", zDate);

  /* Copy all of the content from the parent into the branch */
  for(i=0; i<pParent->nFile; ++i){
    blob_appendf(&branch, "F %F", pParent->aFile[i].zName);
    if( pParent->aFile[i].zUuid ){
      blob_appendf(&branch, " %s", pParent->aFile[i].zUuid);
      if( pParent->aFile[i].zPerm[0] ){
        blob_appendf(&branch, " %s", pParent->aFile[i].zPerm);
      }
    }
    blob_append(&branch, "\n", 1);
  }
︙︙
Changes to src/manifest.c.
︙︙
182
183
184
185
186
187
188




























































































189
190
191
192
193
194
195

  memset(&manifestCache, 0, sizeof(manifestCache));
}

#ifdef FOSSIL_DONT_VERIFY_MANIFEST_MD5SUM
# define md5sum_init(X)
# define md5sum_step_text(X,Y)
#endif





























































































/*
** Parse a blob into a Manifest object.  The Manifest object
** takes over the input blob and will free it when the
** Manifest object is freed.  Zeros are inserted into the blob
** as string terminators so that blob should not be used again.
**








>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>








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

  memset(&manifestCache, 0, sizeof(manifestCache));
}

#ifdef FOSSIL_DONT_VERIFY_MANIFEST_MD5SUM
# define md5sum_init(X)
# define md5sum_step_text(X,Y)
#endif

/*
** Remove the PGP signature from the artifact, if there is one.
*/
static void remove_pgp_signature(char **pz, int *pn){
  char *z = *pz;
  int n = *pn;
  int i;
  if( memcmp(z, "-----BEGIN PGP SIGNED MESSAGE-----", 34)!=0 ) return;
  for(i=34; i<n && (z[i-1]!='\n' || z[i-2]!='\n'); i++){}
  if( i>=n ) return;
  z += i;
  n -= i;
  *pz = z;
  for(i=n-1; i>=0; i--){
    if( z[i]=='\n' && memcmp(&z[i],"\n-----BEGIN PGP SIGNATURE-", 25)==0 ){
      n = i+1;
      break;
    }
  }
  *pn = n;
  return;
}

/*
** Verify the Z-card checksum on the artifact, if there is such a
** checksum.  Return 0 if there is no Z-card.  Return 1 if the Z-card
** exists and is correct.  Return 2 if the Z-card exists and has the wrong
** value.
**
**   0123456789 123456789 123456789 123456789 
**   Z aea84f4f863865a8d59d0384e4d2a41c
*/
static int verify_z_card(const char *z, int n){
  if( n<35 ) return 0;
  if( z[n-35]!='Z' || z[n-34]!=' ' ) return 0;
  md5sum_init();
  md5sum_step_text(z, n-35);
  if( memcmp(&z[n-33], md5sum_finish(0), 32)==0 ){
    return 1;
  }else{
    return 2;
  }
}

/*
** A structure used for rapid parsing of the Manifest file
*/
typedef struct ManifestText ManifestText;
struct ManifestText {
  char *z;           /* The first character of the next token */
  char *zEnd;        /* One character beyond the end of the manifest */
  int atEol;         /* True if z points to the start of a new line */
};

/*
** Return a pointer to the next token.  The token is zero-terminated.
** Return NULL if there are no more tokens on the current line.
*/
static char *next_token(ManifestText *p, int *pLen){
  char *z;
  char *zStart;
  int c;
  if( p->atEol ) return 0;
  zStart = z = p->z;
  while( (c=(*z))!=' ' && c!='\n' ){ z++; }
  *z = 0;
  p->z = &z[1];
  p->atEol = c=='\n';
  if( pLen ) *pLen = z - zStart;
  return zStart;
}

/*
** Return the card-type for the next card.  Or, return 0 if there are no
** more cards or if we are not at the end of the current card.
*/
static char next_card(ManifestText *p){
  char c;
  if( !p->atEol || p->z>=p->zEnd ) return 0;
  c = p->z[0];
  if( p->z[1]==' ' ){
    p->z += 2;
    p->atEol = 0;
  }else if( p->z[1]=='\n' ){
    p->z += 2;
    p->atEol = 1;
  }else{
    c = 0;
  }
  return c;
}

/*
** Parse a blob into a Manifest object.  The Manifest object
** takes over the input blob and will free it when the
** Manifest object is freed.  Zeros are inserted into the blob
** as string terminators so that blob should not be used again.
**
︙︙
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

** Each card is divided into tokens by a single space character.
** The first token is a single upper-case letter which is the card type.
** The card type determines the other parameters to the card.
** Cards must occur in lexicographical order.
*/
static Manifest *manifest_parse(Blob *pContent, int rid){
  Manifest *p;
  int seenHeader = 0;
  int seenZ = 0;
  int i, lineNo=0;
  Blob line, a1, a2, a3, a4;

  char cPrevType = 0;






  /* Every control artifact ends with a '\n' character.  Exit early
  ** if that is not the case for this artifact. */


  i = blob_size(pContent);




  if( i<=0 || blob_buffer(pContent)[i-1]!='\n' ){













    blob_reset(pContent);
    return 0;
  }



  p = fossil_malloc( sizeof(*p) );
  memset(p, 0, sizeof(*p));
  memcpy(&p->content, pContent, sizeof(p->content));
  p->rid = rid;
  blob_zero(pContent);
  pContent = &p->content;

  blob_zero(&a1);
  blob_zero(&a2);

  blob_zero(&a3);
  md5sum_init();
  while( blob_line(pContent, &line) ){
    char *z = blob_buffer(&line);

    lineNo++;
    if( z[0]=='-' ){
      if( strncmp(z, "-----BEGIN PGP ", 15)!=0 ){
        goto manifest_syntax_error;
      }
      if( seenHeader ){
        break;
      }
      while( blob_line(pContent, &line)>2 ){}
      if( blob_line(pContent, &line)==0 ) break;
      z = blob_buffer(&line);
    }
    if( z[0]<cPrevType ){
      /* Lines of a manifest must occur in lexicographical order */
      goto manifest_syntax_error;
    }
    cPrevType = z[0];
    seenHeader = 1;
    if( blob_size(&line)>2 && z[1]!=' ' ) goto manifest_syntax_error;
    line.iCursor = 2;
    switch( z[0] ){
      /*
      **     A <filename> <target> ?<source>?
      **
      ** Identifies an attachment to either a wiki page or a ticket.
      ** <source> is the artifact that is the attachment.  <source>
      ** is omitted to delete an attachment.  <target> is the name of
      ** a wiki page or ticket to which that attachment is connected.
      */
      case 'A': {
        char *zName, *zTarget, *zSrc;
        md5sum_step_text(blob_buffer(&line), blob_size(&line));
        if( fast_token(&line, &a1)==0 ) goto manifest_syntax_error;


        if( fast_token(&line, &a2)==0 ) goto manifest_syntax_error;
        if( p->zAttachName!=0 ) goto manifest_syntax_error;
        zName = blob_terminate(&a1);
        zTarget = blob_terminate(&a2);
        fast_token(&line, &a3);
        zSrc = blob_terminate(&a3);
        defossilize(zName);
        if( !file_is_simple_pathname(zName) ){
          goto manifest_syntax_error;
        }
        defossilize(zTarget);
        if( (blob_size(&a2)!=UUID_SIZE || !validate16(zTarget, UUID_SIZE))
           && !wiki_name_is_wellformed((const unsigned char *)zTarget) ){
          goto manifest_syntax_error;
        }
        if( blob_size(&a3)>0
         && (blob_size(&a3)!=UUID_SIZE || !validate16(zSrc, UUID_SIZE)) ){
          goto manifest_syntax_error;
        }
        p->zAttachName = (char*)file_tail(zName);
        p->zAttachSrc = zSrc;
        p->zAttachTarget = zTarget;
        break;
      }

      /*
      **    B <uuid>
      **
      ** A B-line gives the UUID for the baselinen of a delta-manifest.
      */
      case 'B': {
        char *zBaseline;
        if( p->zBaseline ) goto manifest_syntax_error;
        md5sum_step_text(blob_buffer(&line), blob_size(&line));

        if( fast_token(&line, &a1)==0 ) goto manifest_syntax_error;
        zBaseline = blob_terminate(&a1);
        if( blob_size(&a1)!=UUID_SIZE ) goto manifest_syntax_error;
        if( !validate16(zBaseline, UUID_SIZE) ) goto manifest_syntax_error;
        p->zBaseline = zBaseline;
        break;
      }


      /*
      **     C <comment>
      **
      ** Comment text is fossil-encoded.  There may be no more than
      ** one C line.  C lines are required for manifests and are
      ** disallowed on all other control files.
      */
      case 'C': {
        md5sum_step_text(blob_buffer(&line), blob_size(&line));
        if( p->zComment!=0 ) goto manifest_syntax_error;
        if( fast_token(&line, &a1)==0 ) goto manifest_syntax_error;
        if( fast_token(&line, &a2)!=0 ) goto manifest_syntax_error;
        p->zComment = blob_terminate(&a1);
        defossilize(p->zComment);
        break;
      }

      /*
      **     D <timestamp>
      **
      ** The timestamp should be ISO 8601.   YYYY-MM-DDtHH:MM:SS
      ** There can be no more than 1 D line.  D lines are required
      ** for all control files except for clusters.
      */
      case 'D': {
        char *zDate;
        md5sum_step_text(blob_buffer(&line), blob_size(&line));
        if( p->rDate!=0.0 ) goto manifest_syntax_error;
        if( fast_token(&line, &a1)==0 ) goto manifest_syntax_error;
        if( fast_token(&line, &a2)!=0 ) goto manifest_syntax_error;
        zDate = blob_terminate(&a1);
        p->rDate = db_double(0.0, "SELECT julianday(%Q)", zDate);

        break;
      }

      /*
      **     E <timestamp> <uuid>
      **
      ** An "event" card that contains the timestamp of the event in the 
      ** format YYYY-MM-DDtHH:MM:SS and a unique identifier for the event.
      ** The event timestamp is distinct from the D timestamp.  The D
      ** timestamp is when the artifact was created whereas the E timestamp
      ** is when the specific event is said to occur.
      */
      case 'E': {
        char *zEDate;
        md5sum_step_text(blob_buffer(&line), blob_size(&line));
        if( p->rEventDate!=0.0 ) goto manifest_syntax_error;
        if( fast_token(&line, &a1)==0 ) goto manifest_syntax_error;
        if( fast_token(&line, &a2)==0 ) goto manifest_syntax_error;
        if( fast_token(&line, &a3)!=0 ) goto manifest_syntax_error;
        zEDate = blob_terminate(&a1);
        p->rEventDate = db_double(0.0, "SELECT julianday(%Q)", zEDate);
        if( p->rEventDate<=0.0 ) goto manifest_syntax_error;

        if( blob_size(&a2)!=UUID_SIZE ) goto manifest_syntax_error;
        p->zEventId = blob_terminate(&a2);
        if( !validate16(p->zEventId, UUID_SIZE) ) goto manifest_syntax_error;
        break;
      }

      /*
      **     F <filename> ?<uuid>? ?<permissions>? ?<old-name>?
      **
      ** Identifies a file in a manifest.  Multiple F lines are
      ** allowed in a manifest.  F lines are not allowed in any
      ** other control file.  The filename and old-name are fossil-encoded.
      */
      case 'F': {
        char *zName, *zUuid, *zPerm, *zPriorName;
        md5sum_step_text(blob_buffer(&line), blob_size(&line));
        if( fast_token(&line, &a1)==0 ) goto manifest_syntax_error;
        zName = blob_terminate(&a1);
        fast_token(&line, &a2);
        zUuid = blob_terminate(&a2);
        if( p->zBaseline==0 || zUuid[0]!=0 ){
          if( blob_size(&a2)!=UUID_SIZE ) goto manifest_syntax_error;
          if( !validate16(zUuid, UUID_SIZE) ) goto manifest_syntax_error;
        }
        fast_token(&line, &a3);
        zPerm = blob_terminate(&a3);
        defossilize(zName);
        if( !file_is_simple_pathname(zName) ){
          goto manifest_syntax_error;
        }





        fast_token(&line, &a4);
        zPriorName = blob_terminate(&a4);
        if( zPriorName[0] ){
          defossilize(zPriorName);
          if( !file_is_simple_pathname(zPriorName) ){
            goto manifest_syntax_error;
          }
        }else{
          zPriorName = 0;
        }
        if( p->nFile>=p->nFileAlloc ){
          p->nFileAlloc = p->nFileAlloc*2 + 10;
          p->aFile = fossil_realloc(p->aFile, 
                                    p->nFileAlloc*sizeof(p->aFile[0]) );
        }
        i = p->nFile++;








<


<
>

>
>
>
>
>


|
>
>
|
>
>
>
>
|
>
>
>
>
>
>
>
>
>
>
>
>
>




>
>







|
<
>
|
|
<
|
>

<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
|










|
|
>
>
|

<
<
<
<





|



<
|














<

<
>
|
<
|
|
<












<

|
|
<












<
<
|
<
<
<
|
>













<
<
|
<
<
<
<
|

>
|
<












|
<
<
<
|
<
<
|
<
<
<
<




>
>
>
>
>
|
|
|




<
<








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

** Each card is divided into tokens by a single space character.
** The first token is a single upper-case letter which is the card type.
** The card type determines the other parameters to the card.
** Cards must occur in lexicographical order.
*/
static Manifest *manifest_parse(Blob *pContent, int rid){
  Manifest *p;

  int seenZ = 0;
  int i, lineNo=0;

  ManifestText x;
  char cPrevType = 0;
  char cType;
  char *z;
  int n;
  char *zUuid;
  int sz;

  /* Every control artifact ends with a '\n' character.  Exit early
  ** if that is not the case for this artifact.
  */
  z = blob_buffer(pContent);
  n = blob_size(pContent);
  if( n<=0 || z[n-1]!='\n' ){
    blob_reset(pContent);
    return 0;
  }

  /* Strip off the PGP signature if there is one.  Then verify the
  ** Z-card.
  */
  remove_pgp_signature(&z, &n);
  if( verify_z_card(z, n)==0 ){
    blob_reset(pContent);
    return 0;
  }

  /* Verify that the first few characters of the artifact look like
  ** a control artifact.
  */
  if( n<10 || z[0]<'A' || z[0]>'Z' || z[1]!=' ' ){
    blob_reset(pContent);
    return 0;
  }

  /* Allocate a Manifest object to hold the parsed control artifact.
  */
  p = fossil_malloc( sizeof(*p) );
  memset(p, 0, sizeof(*p));
  memcpy(&p->content, pContent, sizeof(p->content));
  p->rid = rid;
  blob_zero(pContent);
  pContent = &p->content;

  /* Begin parsing, card by card.

  */
  x.z = z;
  x.zEnd = &z[n];

  x.atEol = 1;
  while( (cType = next_card(&x))!=0 && cType>=cPrevType ){
    lineNo++;



















    switch( cType ){
      /*
      **     A <filename> <target> ?<source>?
      **
      ** Identifies an attachment to either a wiki page or a ticket.
      ** <source> is the artifact that is the attachment.  <source>
      ** is omitted to delete an attachment.  <target> is the name of
      ** a wiki page or ticket to which that attachment is connected.
      */
      case 'A': {
        char *zName, *zTarget, *zSrc;
        int nTarget, nSrc;
        zName = next_token(&x, 0);
        zTarget = next_token(&x, &nTarget);
        zSrc = next_token(&x, &nSrc);
        if( zName==0 || zTarget==0 ) goto manifest_syntax_error;      
        if( p->zAttachName!=0 ) goto manifest_syntax_error;




        defossilize(zName);
        if( !file_is_simple_pathname(zName) ){
          goto manifest_syntax_error;
        }
        defossilize(zTarget);
        if( (nTarget!=UUID_SIZE || !validate16(zTarget, UUID_SIZE))
           && !wiki_name_is_wellformed((const unsigned char *)zTarget) ){
          goto manifest_syntax_error;
        }

        if( zSrc && (nSrc!=UUID_SIZE || !validate16(zSrc, UUID_SIZE)) ){
          goto manifest_syntax_error;
        }
        p->zAttachName = (char*)file_tail(zName);
        p->zAttachSrc = zSrc;
        p->zAttachTarget = zTarget;
        break;
      }

      /*
      **    B <uuid>
      **
      ** A B-line gives the UUID for the baselinen of a delta-manifest.
      */
      case 'B': {

        if( p->zBaseline ) goto manifest_syntax_error;

        p->zBaseline = next_token(&x, &sz);
        if( p->zBaseline==0 ) goto manifest_syntax_error;

        if( sz!=UUID_SIZE ) goto manifest_syntax_error;
        if( !validate16(p->zBaseline, UUID_SIZE) ) goto manifest_syntax_error;

        break;
      }


      /*
      **     C <comment>
      **
      ** Comment text is fossil-encoded.  There may be no more than
      ** one C line.  C lines are required for manifests and are
      ** disallowed on all other control files.
      */
      case 'C': {

        if( p->zComment!=0 ) goto manifest_syntax_error;
        p->zComment = next_token(&x, 0);
        if( p->zComment==0 ) goto manifest_syntax_error;

        defossilize(p->zComment);
        break;
      }

      /*
      **     D <timestamp>
      **
      ** The timestamp should be ISO 8601.   YYYY-MM-DDtHH:MM:SS
      ** There can be no more than 1 D line.  D lines are required
      ** for all control files except for clusters.
      */
      case 'D': {


        if( p->rDate>0.0 ) goto manifest_syntax_error;



        p->rDate = db_double(0.0, "SELECT julianday(%Q)", next_token(&x,0));
        if( p->rDate<=0.0 ) goto manifest_syntax_error;
        break;
      }

      /*
      **     E <timestamp> <uuid>
      **
      ** An "event" card that contains the timestamp of the event in the 
      ** format YYYY-MM-DDtHH:MM:SS and a unique identifier for the event.
      ** The event timestamp is distinct from the D timestamp.  The D
      ** timestamp is when the artifact was created whereas the E timestamp
      ** is when the specific event is said to occur.
      */
      case 'E': {


        if( p->rEventDate>0.0 ) goto manifest_syntax_error;




        p->rEventDate = db_double(0.0,"SELECT julianday(%Q)", next_token(&x,0));
        if( p->rEventDate<=0.0 ) goto manifest_syntax_error;
        p->zEventId = next_token(&x, &sz);
        if( sz!=UUID_SIZE ) goto manifest_syntax_error;

        if( !validate16(p->zEventId, UUID_SIZE) ) goto manifest_syntax_error;
        break;
      }

      /*
      **     F <filename> ?<uuid>? ?<permissions>? ?<old-name>?
      **
      ** Identifies a file in a manifest.  Multiple F lines are
      ** allowed in a manifest.  F lines are not allowed in any
      ** other control file.  The filename and old-name are fossil-encoded.
      */
      case 'F': {
        char *zName, *zPerm, *zPriorName;



        zName = next_token(&x,0);


        if( zName==0 ) goto manifest_syntax_error;




        defossilize(zName);
        if( !file_is_simple_pathname(zName) ){
          goto manifest_syntax_error;
        }
        zUuid = next_token(&x, &sz);
        if( p->zBaseline==0 || zUuid!=0 ){
          if( sz!=UUID_SIZE ) goto manifest_syntax_error;
          if( !validate16(zUuid, UUID_SIZE) ) goto manifest_syntax_error;
        }
        zPerm = next_token(&x,0);
        zPriorName = next_token(&x,0);
        if( zPriorName ){
          defossilize(zPriorName);
          if( !file_is_simple_pathname(zPriorName) ){
            goto manifest_syntax_error;
          }


        }
        if( p->nFile>=p->nFileAlloc ){
          p->nFileAlloc = p->nFileAlloc*2 + 10;
          p->aFile = fossil_realloc(p->aFile, 
                                    p->nFileAlloc*sizeof(p->aFile[0]) );
        }
        i = p->nFile++;
︙︙
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454

      ** Specifies a name value pair for ticket.  If the first character
      ** of <name> is "+" then the <value> is appended to any preexisting
      ** value.  If <value> is omitted then it is understood to be an
      ** empty string.
      */
      case 'J': {
        char *zName, *zValue;
        md5sum_step_text(blob_buffer(&line), blob_size(&line));
        if( fast_token(&line, &a1)==0 ) goto manifest_syntax_error;
        fast_token(&line, &a2);
        if( fast_token(&line, &a3)!=0 ) goto manifest_syntax_error;
        zName = blob_terminate(&a1);
        zValue = blob_terminate(&a2);
        defossilize(zValue);
        if( p->nField>=p->nFieldAlloc ){
          p->nFieldAlloc = p->nFieldAlloc*2 + 10;
          p->aField = fossil_realloc(p->aField,
                               p->nFieldAlloc*sizeof(p->aField[0]) );
        }
        i = p->nField++;








<
|
|
|
<
|








509
510
511
512
513
514
515

516
517
518

519
520
521
522
523
524
525
526

      ** Specifies a name value pair for ticket.  If the first character
      ** of <name> is "+" then the <value> is appended to any preexisting
      ** value.  If <value> is omitted then it is understood to be an
      ** empty string.
      */
      case 'J': {
        char *zName, *zValue;

        zName = next_token(&x,0);
        zValue = next_token(&x,0);
        if( zName==0 ) goto manifest_syntax_error;

        if( zValue==0 ) zValue = "";
        defossilize(zValue);
        if( p->nField>=p->nFieldAlloc ){
          p->nFieldAlloc = p->nFieldAlloc*2 + 10;
          p->aField = fossil_realloc(p->aField,
                               p->nFieldAlloc*sizeof(p->aField[0]) );
        }
        i = p->nField++;
︙︙
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

      /*
      **    K <uuid>
      **
      ** A K-line gives the UUID for the ticket which this control file
      ** is amending.
      */
      case 'K': {
        char *zUuid;
        md5sum_step_text(blob_buffer(&line), blob_size(&line));
        if( fast_token(&line, &a1)==0 ) goto manifest_syntax_error;
        zUuid = blob_terminate(&a1);
        if( blob_size(&a1)!=UUID_SIZE ) goto manifest_syntax_error;
        if( !validate16(zUuid, UUID_SIZE) ) goto manifest_syntax_error;
        if( p->zTicketUuid!=0 ) goto manifest_syntax_error;
        p->zTicketUuid = zUuid;
        break;
      }

      /*
      **     L <wikititle>
      **
      ** The wiki page title is fossil-encoded.  There may be no more than
      ** one L line.
      */
      case 'L': {
        md5sum_step_text(blob_buffer(&line), blob_size(&line));
        if( p->zWikiTitle!=0 ) goto manifest_syntax_error;
        if( fast_token(&line, &a1)==0 ) goto manifest_syntax_error;
        if( fast_token(&line, &a2)!=0 ) goto manifest_syntax_error;
        p->zWikiTitle = blob_terminate(&a1);
        defossilize(p->zWikiTitle);
        if( !wiki_name_is_wellformed((const unsigned char *)p->zWikiTitle) ){
          goto manifest_syntax_error;
        }
        break;
      }

      /*
      **    M <uuid>
      **
      ** An M-line identifies another artifact by its UUID.  M-lines
      ** occur in clusters only.
      */
      case 'M': {
        char *zUuid;
        md5sum_step_text(blob_buffer(&line), blob_size(&line));
        if( fast_token(&line, &a1)==0 ) goto manifest_syntax_error;
        zUuid = blob_terminate(&a1);
        if( blob_size(&a1)!=UUID_SIZE ) goto manifest_syntax_error;
        if( !validate16(zUuid, UUID_SIZE) ) goto manifest_syntax_error;
        if( p->nCChild>=p->nCChildAlloc ){
          p->nCChildAlloc = p->nCChildAlloc*2 + 10;
          p->azCChild = fossil_realloc(p->azCChild
                                 , p->nCChildAlloc*sizeof(p->azCChild[0]) );
        }
        i = p->nCChild++;








<
<
|
|
|
|
<
<










<

|
|
<














|
<
|
<
|








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

      /*
      **    K <uuid>
      **
      ** A K-line gives the UUID for the ticket which this control file
      ** is amending.
      */
      case 'K': {


        if( p->zTicketUuid!=0 ) goto manifest_syntax_error;
        p->zTicketUuid = next_token(&x, &sz);
        if( sz!=UUID_SIZE ) goto manifest_syntax_error;
        if( !validate16(p->zTicketUuid, UUID_SIZE) ) goto manifest_syntax_error;


        break;
      }

      /*
      **     L <wikititle>
      **
      ** The wiki page title is fossil-encoded.  There may be no more than
      ** one L line.
      */
      case 'L': {

        if( p->zWikiTitle!=0 ) goto manifest_syntax_error;
        p->zWikiTitle = next_token(&x,0);
        if( p->zWikiTitle==0 ) goto manifest_syntax_error;

        defossilize(p->zWikiTitle);
        if( !wiki_name_is_wellformed((const unsigned char *)p->zWikiTitle) ){
          goto manifest_syntax_error;
        }
        break;
      }

      /*
      **    M <uuid>
      **
      ** An M-line identifies another artifact by its UUID.  M-lines
      ** occur in clusters only.
      */
      case 'M': {
        zUuid = next_token(&x, &sz);

        if( zUuid==0 ) goto manifest_syntax_error;

        if( sz!=UUID_SIZE ) goto manifest_syntax_error;
        if( !validate16(zUuid, UUID_SIZE) ) goto manifest_syntax_error;
        if( p->nCChild>=p->nCChildAlloc ){
          p->nCChildAlloc = p->nCChildAlloc*2 + 10;
          p->azCChild = fossil_realloc(p->azCChild
                                 , p->nCChildAlloc*sizeof(p->azCChild[0]) );
        }
        i = p->nCChild++;
︙︙
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
597
598
599
600
601
602
603
604
605
606
607
608
609
610

      **     P <uuid> ...
      **
      ** Specify one or more other artifacts where are the parents of
      ** this artifact.  The first parent is the primary parent.  All
      ** others are parents by merge.
      */
      case 'P': {
        md5sum_step_text(blob_buffer(&line), blob_size(&line));
        while( fast_token(&line, &a1) ){
          char *zUuid;
          if( blob_size(&a1)!=UUID_SIZE ) goto manifest_syntax_error;
          zUuid = blob_terminate(&a1);
          if( !validate16(zUuid, UUID_SIZE) ) goto manifest_syntax_error;
          if( p->nParent>=p->nParentAlloc ){
            p->nParentAlloc = p->nParentAlloc*2 + 5;
            p->azParent = fossil_realloc(p->azParent,
                               p->nParentAlloc*sizeof(char*));
          }
          i = p->nParent++;
          p->azParent[i] = zUuid;
        }
        break;
      }

      /*
      **     R <md5sum>
      **
      ** Specify the MD5 checksum over the name and content of all files
      ** in the manifest.
      */
      case 'R': {
        md5sum_step_text(blob_buffer(&line), blob_size(&line));
        if( p->zRepoCksum!=0 ) goto manifest_syntax_error;
        if( fast_token(&line, &a1)==0 ) goto manifest_syntax_error;
        if( fast_token(&line, &a2)!=0 ) goto manifest_syntax_error;
        if( blob_size(&a1)!=32 ) goto manifest_syntax_error;
        p->zRepoCksum = blob_terminate(&a1);
        if( !validate16(p->zRepoCksum, 32) ) goto manifest_syntax_error;
        break;
      }

      /*
      **    T (+|*|-)<tagname> <uuid> ?<value>?
      **
      ** Create or cancel a tag or property.  The tagname is fossil-encoded.
      ** The first character of the name must be either "+" to create a
      ** singleton tag, "*" to create a propagating tag, or "-" to create
      ** anti-tag that undoes a prior "+" or blocks propagation of of
      ** a "*".
      **
      ** The tag is applied to <uuid>.  If <uuid> is "*" then the tag is
      ** applied to the current manifest.  If <value> is provided then 
      ** the tag is really a property with the given value.
      **
      ** Tags are not allowed in clusters.  Multiple T lines are allowed.
      */
      case 'T': {
        char *zName, *zUuid, *zValue;
        md5sum_step_text(blob_buffer(&line), blob_size(&line));
        if( fast_token(&line, &a1)==0 ){
          goto manifest_syntax_error;
        }
        if( fast_token(&line, &a2)==0 ){
          goto manifest_syntax_error;
        }
        zName = blob_terminate(&a1);
        zUuid = blob_terminate(&a2);
        if( fast_token(&line, &a3)==0 ){
          zValue = 0;
        }else{
          zValue = blob_terminate(&a3);
          defossilize(zValue);
        }
        if( blob_size(&a2)==UUID_SIZE && validate16(zUuid, UUID_SIZE) ){
          /* A valid uuid */
        }else if( blob_size(&a2)==1 && zUuid[0]=='*' ){
          zUuid = 0;
        }else{
          goto manifest_syntax_error;
        }
        defossilize(zName);
        if( zName[0]!='-' && zName[0]!='+' && zName[0]!='*' ){
          goto manifest_syntax_error;








<
|
<
|
<



















<

|
<
|
<




















|
<
|
|
<
|
|
<
<
<
<
|
<
<
|
<
|

|








592
593
594
595
596
597
598

599

600

601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619

620
621

622

623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643

644
645

646
647




648


649

650
651
652
653
654
655
656
657
658
659

      **     P <uuid> ...
      **
      ** Specify one or more other artifacts where are the parents of
      ** this artifact.  The first parent is the primary parent.  All
      ** others are parents by merge.
      */
      case 'P': {

        while( (zUuid = next_token(&x, &sz))!=0 ){

          if( sz!=UUID_SIZE ) goto manifest_syntax_error;

          if( !validate16(zUuid, UUID_SIZE) ) goto manifest_syntax_error;
          if( p->nParent>=p->nParentAlloc ){
            p->nParentAlloc = p->nParentAlloc*2 + 5;
            p->azParent = fossil_realloc(p->azParent,
                               p->nParentAlloc*sizeof(char*));
          }
          i = p->nParent++;
          p->azParent[i] = zUuid;
        }
        break;
      }

      /*
      **     R <md5sum>
      **
      ** Specify the MD5 checksum over the name and content of all files
      ** in the manifest.
      */
      case 'R': {

        if( p->zRepoCksum!=0 ) goto manifest_syntax_error;
        p->zRepoCksum = next_token(&x, &sz);

        if( sz!=32 ) goto manifest_syntax_error;

        if( !validate16(p->zRepoCksum, 32) ) goto manifest_syntax_error;
        break;
      }

      /*
      **    T (+|*|-)<tagname> <uuid> ?<value>?
      **
      ** Create or cancel a tag or property.  The tagname is fossil-encoded.
      ** The first character of the name must be either "+" to create a
      ** singleton tag, "*" to create a propagating tag, or "-" to create
      ** anti-tag that undoes a prior "+" or blocks propagation of of
      ** a "*".
      **
      ** The tag is applied to <uuid>.  If <uuid> is "*" then the tag is
      ** applied to the current manifest.  If <value> is provided then 
      ** the tag is really a property with the given value.
      **
      ** Tags are not allowed in clusters.  Multiple T lines are allowed.
      */
      case 'T': {
        char *zName, *zValue;

        zName = next_token(&x, 0);
        if( zName==0 ) goto manifest_syntax_error;

        zUuid = next_token(&x, &sz);
        if( zUuid==0 ) goto manifest_syntax_error;




        zValue = next_token(&x, 0);


        if( zValue ) defossilize(zValue);

        if( sz==UUID_SIZE && validate16(zUuid, UUID_SIZE) ){
          /* A valid uuid */
        }else if( sz==1 && zUuid[0]=='*' ){
          zUuid = 0;
        }else{
          goto manifest_syntax_error;
        }
        defossilize(zName);
        if( zName[0]!='-' && zName[0]!='+' && zName[0]!='*' ){
          goto manifest_syntax_error;
︙︙
631
632
633
634
635
636
637
638
639
640

641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657

658
659
660
661
662
663



664
665
666
667
668
669
670
671
672
673

674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719

      **     U ?<login>?
      **
      ** Identify the user who created this control file by their
      ** login.  Only one U line is allowed.  Prohibited in clusters.
      ** If the user name is omitted, take that to be "anonymous".
      */
      case 'U': {
        md5sum_step_text(blob_buffer(&line), blob_size(&line));
        if( p->zUser!=0 ) goto manifest_syntax_error;
        if( fast_token(&line, &a1)==0 ){

          p->zUser = "anonymous";
        }else{
          p->zUser = blob_terminate(&a1);
          defossilize(p->zUser);
        }
        if( fast_token(&line, &a2)!=0 ) goto manifest_syntax_error;
        break;
      }

      /*
      **     W <size>
      **
      ** The next <size> bytes of the file contain the text of the wiki
      ** page.  There is always an extra \n before the start of the next
      ** record.
      */
      case 'W': {

        int size;
        Blob wiki;
        md5sum_step_text(blob_buffer(&line), blob_size(&line));
        if( fast_token(&line, &a1)==0 ) goto manifest_syntax_error;
        if( fast_token(&line, &a2)!=0 ) goto manifest_syntax_error;
        if( !blob_is_int(&a1, &size) ) goto manifest_syntax_error;



        if( size<0 ) goto manifest_syntax_error;
        if( p->zWiki!=0 ) goto manifest_syntax_error;
        blob_zero(&wiki);
        if( blob_extract(pContent, size+1, &wiki)!=size+1 ){
          goto manifest_syntax_error;
        }
        p->zWiki = blob_buffer(&wiki);
        md5sum_step_text(p->zWiki, size+1);
        if( p->zWiki[size]!='\n' ) goto manifest_syntax_error;
        p->zWiki[size] = 0;

        break;
      }


      /*
      **     Z <md5sum>
      **
      ** MD5 checksum on this control file.  The checksum is over all
      ** lines (other than PGP-signature lines) prior to the current
      ** line.  This must be the last record.
      **
      ** This card is required for all control file types except for
      ** Manifest.  It is not required for manifest only for historical
      ** compatibility reasons.
      */
      case 'Z': {
#ifndef FOSSIL_DONT_VERIFY_MANIFEST_MD5SUM
        int rc;
        Blob hash;
#endif
        if( fast_token(&line, &a1)==0 ) goto manifest_syntax_error;
        if( fast_token(&line, &a2)!=0 ) goto manifest_syntax_error;
        if( blob_size(&a1)!=32 ) goto manifest_syntax_error;
        if( !validate16(blob_buffer(&a1), 32) ) goto manifest_syntax_error;
#ifndef FOSSIL_DONT_VERIFY_MANIFEST_MD5SUM
        md5sum_finish(&hash);
        rc = blob_compare(&hash, &a1);
        blob_reset(&hash);
        if( rc!=0 ) goto manifest_syntax_error;
#endif
        seenZ = 1;
        break;
      }
      default: {
        goto manifest_syntax_error;
      }
    }
  }
  if( !seenHeader ) goto manifest_syntax_error;

  if( p->nFile>0 || p->zRepoCksum!=0 || p->zBaseline ){
    if( p->nCChild>0 ) goto manifest_syntax_error;
    if( p->rDate<=0.0 ) goto manifest_syntax_error;
    if( p->nField>0 ) goto manifest_syntax_error;
    if( p->zTicketUuid ) goto manifest_syntax_error;
    if( p->zWiki ) goto manifest_syntax_error;








<

|
>


<


<











>
|

<
|
|
|
>
>
>



<
|
<
|
|
|
|
>
















<
<
<
<
|
<
|
|
<
<
<
<
<
<








|








680
681
682
683
684
685
686

687
688
689
690
691

692
693

694
695
696
697
698
699
700
701
702
703
704
705
706
707

708
709
710
711
712
713
714
715
716

717

718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738




739

740
741






742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757

      **     U ?<login>?
      **
      ** Identify the user who created this control file by their
      ** login.  Only one U line is allowed.  Prohibited in clusters.
      ** If the user name is omitted, take that to be "anonymous".
      */
      case 'U': {

        if( p->zUser!=0 ) goto manifest_syntax_error;
        p->zUser = next_token(&x, 0);
        if( p->zUser==0 ){
          p->zUser = "anonymous";
        }else{

          defossilize(p->zUser);
        }

        break;
      }

      /*
      **     W <size>
      **
      ** The next <size> bytes of the file contain the text of the wiki
      ** page.  There is always an extra \n before the start of the next
      ** record.
      */
      case 'W': {
        char *zSize;
        int size, c;
        Blob wiki;

        zSize = next_token(&x, 0);
        if( zSize==0 ) goto manifest_syntax_error;
        if( x.atEol==0 ) goto manifest_syntax_error;
        for(size=0; (c = zSize[0])>='0' && c<='9'; zSize++){
           size = size*10 + c - '0';
        }
        if( size<0 ) goto manifest_syntax_error;
        if( p->zWiki!=0 ) goto manifest_syntax_error;
        blob_zero(&wiki);

        if( (&x.z[size+1])>=x.zEnd ) goto manifest_syntax_error;

        p->zWiki = x.z;
        x.z += size;
        if( x.z[0]!='\n' ) goto manifest_syntax_error;
        x.z[0] = 0;
        x.z++;
        break;
      }


      /*
      **     Z <md5sum>
      **
      ** MD5 checksum on this control file.  The checksum is over all
      ** lines (other than PGP-signature lines) prior to the current
      ** line.  This must be the last record.
      **
      ** This card is required for all control file types except for
      ** Manifest.  It is not required for manifest only for historical
      ** compatibility reasons.
      */
      case 'Z': {




        zUuid = next_token(&x, &sz);

        if( sz!=32 ) goto manifest_syntax_error;
        if( !validate16(zUuid, 32) ) goto manifest_syntax_error;






        seenZ = 1;
        break;
      }
      default: {
        goto manifest_syntax_error;
      }
    }
  }
  if( x.z<x.zEnd ) goto manifest_syntax_error;

  if( p->nFile>0 || p->zRepoCksum!=0 || p->zBaseline ){
    if( p->nCChild>0 ) goto manifest_syntax_error;
    if( p->rDate<=0.0 ) goto manifest_syntax_error;
    if( p->nField>0 ) goto manifest_syntax_error;
    if( p->zTicketUuid ) goto manifest_syntax_error;
    if( p->zWiki ) goto manifest_syntax_error;
︙︙
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953

        pOut = &pB->aFile[pB->iFile++];
        break;
      }else if( cmp>0 ){
        /* The next delta entry comes before the next baseline
        ** entry so return the delta entry */
        pOut = &p->aFile[p->iFile++];
        break;
      }else if( p->aFile[p->iFile].zUuid[0] ){
        /* The next delta entry is a replacement for the next baseline
        ** entry.  Skip the baseline entry and return the delta entry */
        pB->iFile++;
        pOut = &p->aFile[p->iFile++];
        break;
      }else{
        /* The next delta entry is a delete of the next baseline








|








977
978
979
980
981
982
983
984
985
986
987
988
989
990
991

        pOut = &pB->aFile[pB->iFile++];
        break;
      }else if( cmp>0 ){
        /* The next delta entry comes before the next baseline
        ** entry so return the delta entry */
        pOut = &p->aFile[p->iFile++];
        break;
      }else if( p->aFile[p->iFile].zUuid ){
        /* The next delta entry is a replacement for the next baseline
        ** entry.  Skip the baseline entry and return the delta entry */
        pB->iFile++;
        pOut = &p->aFile[p->iFile++];
        break;
      }else{
        /* The next delta entry is a delete of the next baseline
︙︙