Fossil

Artifact [03104c70]
Login

Artifact [03104c70]

Artifact 03104c701ed9ff5faab31e902b7c182858145cd758317f2039027594ec4b2c1b:


     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
   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
   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
   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
   758
   759
   760
   761
   762
   763
   764
   765
   766
   767
   768
   769
   770
   771
   772
   773
   774
   775
   776
   777
   778
   779
   780
   781
   782
   783
   784
   785
   786
   787
   788
   789
   790
   791
   792
   793
   794
   795
   796
   797
   798
   799
   800
   801
   802
   803
   804
   805
   806
   807
   808
   809
   810
   811
   812
   813
   814
   815
   816
   817
   818
   819
   820
   821
   822
   823
   824
   825
   826
   827
   828
   829
   830
   831
   832
   833
   834
   835
   836
   837
   838
   839
   840
   841
   842
   843
   844
   845
   846
   847
   848
   849
   850
   851
   852
   853
   854
   855
   856
   857
   858
   859
   860
   861
   862
   863
   864
   865
   866
   867
   868
   869
   870
   871
   872
   873
   874
   875
   876
   877
   878
   879
   880
   881
   882
   883
   884
   885
   886
   887
   888
   889
   890
   891
   892
   893
   894
   895
   896
   897
   898
   899
   900
   901
   902
   903
   904
   905
   906
   907
   908
   909
   910
   911
   912
   913
   914
   915
   916
   917
   918
   919
   920
   921
   922
   923
   924
   925
   926
   927
   928
   929
   930
   931
   932
   933
   934
   935
   936
   937
   938
   939
   940
   941
   942
   943
   944
   945
   946
   947
   948
   949
   950
   951
   952
   953
   954
   955
   956
   957
   958
   959
   960
   961
   962
   963
   964
   965
   966
   967
   968
   969
   970
   971
   972
   973
   974
   975
   976
   977
   978
   979
   980
   981
   982
   983
   984
   985
   986
   987
   988
   989
   990
   991
   992
   993
   994
   995
   996
   997
   998
   999
  1000
  1001
  1002
  1003
  1004
  1005
  1006
  1007
  1008
  1009
  1010
  1011
  1012
  1013
  1014
  1015
  1016
  1017
  1018
  1019
  1020
  1021
  1022
  1023
  1024
  1025
  1026
  1027
  1028
  1029
  1030
  1031
  1032
  1033
  1034
  1035
  1036
  1037
  1038
  1039
  1040
  1041
  1042
  1043
  1044
  1045
  1046
  1047
  1048
  1049
  1050
  1051
  1052
  1053
  1054
  1055
  1056
  1057
  1058
  1059
  1060
  1061
  1062
  1063
  1064
  1065
  1066
  1067
  1068
  1069
  1070
  1071
  1072
  1073
  1074
  1075
  1076
  1077
  1078
  1079
  1080
  1081
  1082
  1083
  1084
  1085
  1086
  1087
  1088
  1089
  1090
  1091
  1092
  1093
  1094
  1095
  1096
  1097
  1098
  1099
  1100
  1101
  1102
  1103
  1104
  1105
  1106
  1107
  1108
  1109
  1110
  1111
  1112
  1113
  1114
  1115
  1116
  1117
  1118
  1119
  1120
  1121
  1122
  1123
  1124
  1125
  1126
  1127
  1128
  1129
  1130
  1131
  1132
  1133
  1134
  1135
  1136
  1137
  1138
  1139
  1140
  1141
  1142
  1143
  1144
  1145
  1146
  1147
  1148
  1149
  1150
  1151
  1152
  1153
  1154
  1155
  1156
  1157
  1158
  1159
  1160
  1161
  1162
  1163
  1164
  1165
  1166
  1167
  1168
  1169
  1170
  1171
  1172
  1173
  1174
  1175
  1176
  1177
  1178
  1179
  1180
  1181
  1182
  1183
  1184
  1185
  1186
  1187
  1188
  1189
  1190
  1191
  1192
  1193
  1194
  1195
  1196
  1197
  1198
  1199
  1200
  1201
  1202
  1203
  1204
  1205
  1206
  1207
  1208
  1209
  1210
  1211
  1212
  1213
  1214
  1215
  1216
  1217
  1218
  1219
  1220
  1221
  1222
  1223
  1224
  1225
  1226
  1227
  1228
  1229
  1230
  1231
  1232
  1233
  1234
  1235
  1236
  1237
  1238
  1239
  1240
  1241
  1242
  1243
  1244
  1245
  1246
  1247
  1248
  1249
  1250
  1251
  1252
  1253
  1254
  1255
  1256
  1257
  1258
  1259
  1260
  1261
  1262
  1263
  1264
  1265
  1266
  1267
  1268
  1269
  1270
  1271
  1272
  1273
  1274
  1275
  1276
  1277
  1278
  1279
  1280
  1281
  1282
  1283
  1284
  1285
  1286
  1287
  1288
  1289
  1290
  1291
  1292
  1293
  1294
  1295
  1296
  1297
  1298
  1299
  1300
  1301
  1302
  1303
  1304
  1305
  1306
  1307
  1308
  1309
  1310
  1311
  1312
  1313
  1314
  1315
  1316
  1317
  1318
  1319
  1320
  1321
  1322
  1323
  1324
  1325
  1326
  1327
  1328
  1329
  1330
  1331
  1332
  1333
  1334
  1335
  1336
  1337
  1338
  1339
  1340
  1341
  1342
  1343
  1344
  1345
  1346
  1347
  1348
  1349
  1350
  1351
  1352
  1353
  1354
  1355
  1356
  1357
  1358
  1359
  1360
  1361
  1362
  1363
  1364
  1365
  1366
  1367
  1368
  1369
  1370
/*
** Copyright (c) 2006 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/
**
*******************************************************************************
**
** A Blob is a variable-length containers for arbitrary string
** or binary data.
*/
#include "config.h"
#if defined(FOSSIL_ENABLE_MINIZ)
#  define MINIZ_HEADER_FILE_ONLY
#  include "miniz.c"
#else
#  include <zlib.h>
#endif
#include "blob.h"
#if defined(_WIN32)
#include <fcntl.h>
#include <io.h>
#endif

#if INTERFACE
/*
** A Blob can hold a string or a binary object of arbitrary size.  The
** size changes as necessary.
*/
struct Blob {
  unsigned int nUsed;            /* Number of bytes used in aData[] */
  unsigned int nAlloc;           /* Number of bytes allocated for aData[] */
  unsigned int iCursor;          /* Next character of input to parse */
  unsigned int blobFlags;        /* One or more BLOBFLAG_* bits */
  char *aData;                   /* Where the information is stored */
  void (*xRealloc)(Blob*, unsigned int); /* Function to reallocate the buffer */
};

/*
** Allowed values for Blob.blobFlags
*/
#define BLOBFLAG_NotSQL  0x0001      /* Non-SQL text */

/*
** The current size of a Blob
*/
#define blob_size(X)  ((X)->nUsed)

/*
** The buffer holding the blob data
*/
#define blob_buffer(X)  ((X)->aData)

/*
** Seek whence parameter values
*/
#define BLOB_SEEK_SET 1
#define BLOB_SEEK_CUR 2
#define BLOB_SEEK_END 3

#endif /* INTERFACE */

/*
** Make sure a blob is initialized
*/
#define blob_is_init(x) \
  assert((x)->xRealloc==blobReallocMalloc || (x)->xRealloc==blobReallocStatic)

/*
** Make sure a blob does not contain malloced memory.
**
** This might fail if we are unlucky and x is uninitialized.  For that
** reason it should only be used locally for debugging.  Leave it turned
** off for production.
*/
#if 0  /* Enable for debugging only */
#define assert_blob_is_reset(x) assert(blob_is_reset(x))
#else
#define assert_blob_is_reset(x)
#endif



/*
** We find that the built-in isspace() function does not work for
** some international character sets.  So here is a substitute.
*/
int fossil_isspace(char c){
  return c==' ' || (c<='\r' && c>='\t');
}

/*
** Other replacements for ctype.h functions.
*/
int fossil_islower(char c){ return c>='a' && c<='z'; }
int fossil_isupper(char c){ return c>='A' && c<='Z'; }
int fossil_isdigit(char c){ return c>='0' && c<='9'; }
int fossil_tolower(char c){
  return fossil_isupper(c) ? c - 'A' + 'a' : c;
}
int fossil_toupper(char c){
  return fossil_islower(c) ? c - 'a' + 'A' : c;
}
int fossil_isalpha(char c){
  return (c>='a' && c<='z') || (c>='A' && c<='Z');
}
int fossil_isalnum(char c){
  return (c>='a' && c<='z') || (c>='A' && c<='Z') || (c>='0' && c<='9');
}

/* Return true if and only if the entire string consists of only
** alphanumeric characters.
*/
int fossil_no_strange_characters(const char *z){
  while( z && (fossil_isalnum(z[0]) || z[0]=='_' || z[0]=='-') ) z++;
  return z[0]==0;
}


/*
** COMMAND: test-isspace
**
** Verify that the fossil_isspace() routine is working correctly by
** testing it on all possible inputs.
*/
void isspace_cmd(void){
  int i;
  for(i=0; i<=255; i++){
    if( i==' ' || i=='\n' || i=='\t' || i=='\v'
        || i=='\f' || i=='\r' ){
      assert( fossil_isspace((char)i) );
    }else{
      assert( !fossil_isspace((char)i) );
    }
  }
  fossil_print("All 256 characters OK\n");
}

/*
** This routine is called if a blob operation fails because we
** have run out of memory.
*/
static void blob_panic(void){
  static const char zErrMsg[] = "out of memory\n";
  fputs(zErrMsg, stderr);
  fossil_exit(1);
}

/*
** A reallocation function that assumes that aData came from malloc().
** This function attempts to resize the buffer of the blob to hold
** newSize bytes.
**
** No attempt is made to recover from an out-of-memory error.
** If an OOM error occurs, an error message is printed on stderr
** and the program exits.
*/
void blobReallocMalloc(Blob *pBlob, unsigned int newSize){
  if( newSize==0 ){
    free(pBlob->aData);
    pBlob->aData = 0;
    pBlob->nAlloc = 0;
    pBlob->nUsed = 0;
    pBlob->iCursor = 0;
    pBlob->blobFlags = 0;
  }else if( newSize>pBlob->nAlloc || newSize<pBlob->nAlloc-4000 ){
    char *pNew = fossil_realloc(pBlob->aData, newSize);
    pBlob->aData = pNew;
    pBlob->nAlloc = newSize;
    if( pBlob->nUsed>pBlob->nAlloc ){
      pBlob->nUsed = pBlob->nAlloc;
    }
  }
}

/*
** An initializer for Blobs
*/
#if INTERFACE
#define BLOB_INITIALIZER  {0,0,0,0,0,blobReallocMalloc}
#endif
const Blob empty_blob = BLOB_INITIALIZER;

/*
** A reallocation function for when the initial string is in unmanaged
** space.  Copy the string to memory obtained from malloc().
*/
static void blobReallocStatic(Blob *pBlob, unsigned int newSize){
  if( newSize==0 ){
    *pBlob = empty_blob;
  }else{
    char *pNew = fossil_malloc( newSize );
    if( pBlob->nUsed>newSize ) pBlob->nUsed = newSize;
    memcpy(pNew, pBlob->aData, pBlob->nUsed);
    pBlob->aData = pNew;
    pBlob->xRealloc = blobReallocMalloc;
    pBlob->nAlloc = newSize;
  }
}

/*
** Reset a blob to be an empty container.
*/
void blob_reset(Blob *pBlob){
  blob_is_init(pBlob);
  pBlob->xRealloc(pBlob, 0);
}


/*
** Return true if the blob has been zeroed - in other words if it contains
** no malloced memory.  This only works reliably if the blob has been
** initialized - it can return a false negative on an uninitialized blob.
*/
int blob_is_reset(Blob *pBlob){
  if( pBlob==0 ) return 1;
  if( pBlob->nUsed ) return 0;
  if( pBlob->xRealloc==blobReallocMalloc && pBlob->nAlloc ) return 0;
  return 1;
}

/*
** Initialize a blob to a string or byte-array constant of a specified length.
** Any prior data in the blob is discarded.
*/
void blob_init(Blob *pBlob, const char *zData, int size){
  assert_blob_is_reset(pBlob);
  if( zData==0 ){
    *pBlob = empty_blob;
  }else{
    if( size<=0 ) size = strlen(zData);
    pBlob->nUsed = pBlob->nAlloc = size;
    pBlob->aData = (char*)zData;
    pBlob->iCursor = 0;
    pBlob->blobFlags = 0;
    pBlob->xRealloc = blobReallocStatic;
  }
}

/*
** Initialize a blob to a nul-terminated string.
** Any prior data in the blob is discarded.
*/
void blob_set(Blob *pBlob, const char *zStr){
  blob_init(pBlob, zStr, -1);
}

/*
** Initialize a blob to a nul-terminated string obtained from fossil_malloc().
** The blob will take responsibility for freeing the string.
*/
void blob_set_dynamic(Blob *pBlob, char *zStr){
  blob_init(pBlob, zStr, -1);
  pBlob->xRealloc = blobReallocMalloc;
}

/*
** Initialize a blob to an empty string.
*/
void blob_zero(Blob *pBlob){
  static const char zEmpty[] = "";
  assert_blob_is_reset(pBlob);
  pBlob->nUsed = 0;
  pBlob->nAlloc = 1;
  pBlob->aData = (char*)zEmpty;
  pBlob->iCursor = 0;
  pBlob->blobFlags = 0;
  pBlob->xRealloc = blobReallocStatic;
}

/*
** Append text or data to the end of a blob.
*/
void blob_append(Blob *pBlob, const char *aData, int nData){
  sqlite3_int64 nNew;
  assert( aData!=0 || nData==0 );
  blob_is_init(pBlob);
  if( nData<0 ) nData = strlen(aData);
  if( nData==0 ) return;
  nNew = pBlob->nUsed;
  nNew += nData;
  if( nNew >= pBlob->nAlloc ){
    nNew += pBlob->nAlloc;
    nNew += 100;
    if( nNew>=0x7fff0000 ){
      blob_panic();
    }
    pBlob->xRealloc(pBlob, (int)nNew);
    if( pBlob->nUsed + nData >= pBlob->nAlloc ){
      blob_panic();
    }
  }
  memcpy(&pBlob->aData[pBlob->nUsed], aData, nData);
  pBlob->nUsed += nData;
  pBlob->aData[pBlob->nUsed] = 0;   /* Blobs are always nul-terminated */
}

/*
** Append a single character to the blob
*/
void blob_append_char(Blob *pBlob, char c){
  if( pBlob->nUsed+1 >= pBlob->nAlloc ){
    pBlob->xRealloc(pBlob, pBlob->nUsed + pBlob->nAlloc + 100);
    if( pBlob->nUsed + 1 >= pBlob->nAlloc ){
      blob_panic();
    }
  }
  pBlob->aData[pBlob->nUsed++] = c;    
}

/*
** Copy a blob
*/
void blob_copy(Blob *pTo, Blob *pFrom){
  blob_is_init(pFrom);
  blob_zero(pTo);
  blob_append(pTo, blob_buffer(pFrom), blob_size(pFrom));
}

/*
** Return a pointer to a null-terminated string for a blob.
*/
char *blob_str(Blob *p){
  blob_is_init(p);
  if( p->nUsed==0 ){
    blob_append_char(p, 0); /* NOTE: Changes nUsed. */
    p->nUsed = 0;
  }
  if( p->aData[p->nUsed]!=0 ){
    blob_materialize(p);
  }
  return p->aData;
}

/*
** Return a pointer to a null-terminated string for a blob that has
** been created using blob_append_sql() and not blob_appendf().  If
** text was ever added using blob_appendf() then throw an error.
*/
char *blob_sql_text(Blob *p){
  blob_is_init(p);
  if( (p->blobFlags & BLOBFLAG_NotSQL) ){
    fossil_panic("use of blob_appendf() to construct SQL text");
  }
  return blob_str(p);
}


/*
** Return a pointer to a null-terminated string for a blob.
**
** WARNING:  If the blob is ephemeral, it might cause a '\000'
** character to be inserted into the middle of the parent blob.
** Example:  Suppose p is a token extracted from some larger
** blob pBig using blob_token().  If you call this routine on p,
** then a '\000' character will be inserted in the middle of
** pBig in order to cause p to be nul-terminated.  If pBig
** should not be modified, then use blob_str() instead of this
** routine.  blob_str() will make a copy of the p if necessary
** to avoid modifying pBig.
*/
char *blob_terminate(Blob *p){
  blob_is_init(p);
  if( p->nUsed==0 ) return "";
  p->aData[p->nUsed] = 0;
  return p->aData;
}

/*
** Compare two blobs.  Return negative, zero, or positive if the first
** blob is less then, equal to, or greater than the second.
*/
int blob_compare(Blob *pA, Blob *pB){
  int szA, szB, sz, rc;
  blob_is_init(pA);
  blob_is_init(pB);
  szA = blob_size(pA);
  szB = blob_size(pB);
  sz = szA<szB ? szA : szB;
  rc = memcmp(blob_buffer(pA), blob_buffer(pB), sz);
  if( rc==0 ){
    rc = szA - szB;
  }
  return rc;
}

/*
** Compare two blobs in constant time and return zero if they are equal.
** Constant time comparison only applies for blobs of the same length.
** If lengths are different, immediately returns 1.
*/
int blob_constant_time_cmp(Blob *pA, Blob *pB){
  int szA, szB, i;
  unsigned char *buf1, *buf2;
  unsigned char rc = 0;

  blob_is_init(pA);
  blob_is_init(pB);
  szA = blob_size(pA);
  szB = blob_size(pB);
  if( szA!=szB || szA==0 ) return 1;

  buf1 = (unsigned char*)blob_buffer(pA);
  buf2 = (unsigned char*)blob_buffer(pB);

  for( i=0; i<szA; i++ ){
    rc = rc | (buf1[i] ^ buf2[i]);
  }

  return rc;
}

/*
** Compare a blob to a string.  Return TRUE if they are equal.
*/
int blob_eq_str(Blob *pBlob, const char *z, int n){
  Blob t;
  blob_is_init(pBlob);
  if( n<=0 ) n = strlen(z);
  t.aData = (char*)z;
  t.nUsed = n;
  t.xRealloc = blobReallocStatic;
  return blob_compare(pBlob, &t)==0;
}

/*
** This macro compares a blob against a string constant.  We use the sizeof()
** operator on the string constant twice, so it really does need to be a
** string literal or character array - not a character pointer.
*/
#if INTERFACE
# define blob_eq(B,S) \
     ((B)->nUsed==sizeof(S)-1 && memcmp((B)->aData,S,sizeof(S)-1)==0)
#endif


/*
** Attempt to resize a blob so that its internal buffer is
** nByte in size.  The blob is truncated if necessary.
*/
void blob_resize(Blob *pBlob, unsigned int newSize){
  pBlob->xRealloc(pBlob, newSize+1);
  pBlob->nUsed = newSize;
  pBlob->aData[newSize] = 0;
}

/*
** Make sure a blob is nul-terminated and is not a pointer to unmanaged
** space.  Return a pointer to the data.
*/
char *blob_materialize(Blob *pBlob){
  blob_resize(pBlob, pBlob->nUsed);
  return pBlob->aData;
}


/*
** Call dehttpize on a blob.  This causes an ephemeral blob to be
** materialized.
*/
void blob_dehttpize(Blob *pBlob){
  blob_materialize(pBlob);
  pBlob->nUsed = dehttpize(pBlob->aData);
}

/*
** Extract N bytes from blob pFrom and use it to initialize blob pTo.
** Return the actual number of bytes extracted.
**
** After this call completes, pTo will be an ephemeral blob.
*/
int blob_extract(Blob *pFrom, int N, Blob *pTo){
  blob_is_init(pFrom);
  assert_blob_is_reset(pTo);
  if( pFrom->iCursor + N > pFrom->nUsed ){
    N = pFrom->nUsed - pFrom->iCursor;
    if( N<=0 ){
      blob_zero(pTo);
      return 0;
    }
  }
  pTo->nUsed = N;
  pTo->nAlloc = N;
  pTo->aData = &pFrom->aData[pFrom->iCursor];
  pTo->iCursor = 0;
  pTo->xRealloc = blobReallocStatic;
  pFrom->iCursor += N;
  return N;
}

/*
** Rewind the cursor on a blob back to the beginning.
*/
void blob_rewind(Blob *p){
  p->iCursor = 0;
}

/*
** Truncate a blob back to zero length
*/
void blob_truncate(Blob *p, int sz){
  if( sz>=0 && sz<p->nUsed ) p->nUsed = sz;
}

/*
** Seek the cursor in a blob to the indicated offset.
*/
int blob_seek(Blob *p, int offset, int whence){
  if( whence==BLOB_SEEK_SET ){
    p->iCursor = offset;
  }else if( whence==BLOB_SEEK_CUR ){
    p->iCursor += offset;
  }else if( whence==BLOB_SEEK_END ){
    p->iCursor = p->nUsed + offset - 1;
  }
  if( p->iCursor>p->nUsed ){
    p->iCursor = p->nUsed;
  }
  return p->iCursor;
}

/*
** Return the current offset into the blob
*/
int blob_tell(Blob *p){
  return p->iCursor;
}

/*
** Extract a single line of text from pFrom beginning at the current
** cursor location and use that line of text to initialize pTo.
** pTo will include the terminating \n.  Return the number of bytes
** in the line including the \n at the end.  0 is returned at
** end-of-file.
**
** The cursor of pFrom is left pointing at the first byte past the
** \n that terminated the line.
**
** pTo will be an ephermeral blob.  If pFrom changes, it might alter
** pTo as well.
*/
int blob_line(Blob *pFrom, Blob *pTo){
  char *aData = pFrom->aData;
  int n = pFrom->nUsed;
  int i = pFrom->iCursor;

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

/*
** Trim whitespace off of the end of a blob.  Return the number
** of characters remaining.
**
** All this does is reduce the length counter.  This routine does
** not insert a new zero terminator.
*/
int blob_trim(Blob *p){
  char *z = p->aData;
  int n = p->nUsed;
  while( n>0 && fossil_isspace(z[n-1]) ){ n--; }
  p->nUsed = n;
  return n;
}

/*
** Extract a single token from pFrom and use it to initialize pTo.
** Return the number of bytes in the token.  If no token is found,
** return 0.
**
** A token consists of one or more non-space characters.  Leading
** whitespace is ignored.
**
** The cursor of pFrom is left pointing at the first character past
** the end of the token.
**
** pTo will be an ephermeral blob.  If pFrom changes, it might alter
** pTo as well.
*/
int blob_token(Blob *pFrom, Blob *pTo){
  char *aData = pFrom->aData;
  int n = pFrom->nUsed;
  int i = pFrom->iCursor;
  while( i<n && fossil_isspace(aData[i]) ){ i++; }
  pFrom->iCursor = i;
  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 a single SQL token from pFrom and use it to initialize pTo.
** Return the number of bytes in the token.  If no token is found,
** return 0.
**
** An SQL token consists of one or more non-space characters.  If the
** first character is ' then the token is terminated by a matching '
** (ignoring double '') or by the end of the string
**
** The cursor of pFrom is left pointing at the first character past
** the end of the token.
**
** pTo will be an ephermeral blob.  If pFrom changes, it might alter
** pTo as well.
*/
int blob_sqltoken(Blob *pFrom, Blob *pTo){
  char *aData = pFrom->aData;
  int n = pFrom->nUsed;
  int i = pFrom->iCursor;
  while( i<n && fossil_isspace(aData[i]) ){ i++; }
  pFrom->iCursor = i;
  if( aData[i]=='\'' ){
    i++;
    while( i<n ){
      if( aData[i]=='\'' ){
        if( aData[++i]!='\'' ) break;
      }
      i++;
    }
  }else{
    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;
  blob_extract(pFrom, pFrom->nUsed-pFrom->iCursor, pTo);
  pFrom->iCursor = iCursor;
  return pTo->nUsed;
}

/*
** Copy N lines of text from pFrom into pTo.  The copy begins at the
** current cursor position of pIn.  The pIn cursor is left pointing
** at the first character past the last \n copied.
**
** If pTo==NULL then this routine simply skips over N lines.
*/
void blob_copy_lines(Blob *pTo, Blob *pFrom, int N){
  char *z = pFrom->aData;
  int i = pFrom->iCursor;
  int n = pFrom->nUsed;
  int cnt = 0;

  if( N==0 ) return;
  while( i<n ){
    if( z[i]=='\n' ){
      cnt++;
      if( cnt==N ){
        i++;
        break;
      }
    }
    i++;
  }
  if( pTo ){
    blob_append(pTo, &pFrom->aData[pFrom->iCursor], i - pFrom->iCursor);
  }
  pFrom->iCursor = i;
}

/*
** Ensure that the text in pBlob ends with '\n'
*/
void blob_add_final_newline(Blob *pBlob){
  if( pBlob->nUsed<=0 ) return;
  if( pBlob->aData[pBlob->nUsed-1]!='\n' ){
    blob_append_char(pBlob, '\n');
  }
}

/*
** Return true if the blob contains a valid base16 identifier artifact hash.
**
** The value returned is actually one of HNAME_SHA1 OR HNAME_K256 if the
** hash is valid.  Both of these are non-zero and therefore "true".
** If the hash is not valid, then HNAME_ERROR is returned, which is zero or
** false.
*/
int blob_is_hname(Blob *pBlob){
  return hname_validate(blob_buffer(pBlob), blob_size(pBlob));
}

/*
** Return true if the blob contains a valid filename
*/
int blob_is_filename(Blob *pBlob){
  return file_is_simple_pathname(blob_str(pBlob), 1);
}

/*
** Return true if the blob contains a valid 32-bit integer.  Store
** the integer value in *pValue.
*/
int blob_is_int(Blob *pBlob, int *pValue){
  const char *z = blob_buffer(pBlob);
  int i, n, c, v;
  n = blob_size(pBlob);
  v = 0;
  for(i=0; i<n && (c = z[i])!=0 && c>='0' && c<='9'; i++){
    v = v*10 + c - '0';
  }
  if( i==n ){
    *pValue = v;
    return 1;
  }else{
    return 0;
  }
}

/*
** Return true if the blob contains a valid 64-bit integer.  Store
** the integer value in *pValue.
*/
int blob_is_int64(Blob *pBlob, sqlite3_int64 *pValue){
  const char *z = blob_buffer(pBlob);
  int i, n, c;
  sqlite3_int64 v;
  n = blob_size(pBlob);
  v = 0;
  for(i=0; i<n && (c = z[i])!=0 && c>='0' && c<='9'; i++){
    v = v*10 + c - '0';
  }
  if( i==n ){
    *pValue = v;
    return 1;
  }else{
    return 0;
  }
}

/*
** Zero or reset an array of Blobs.
*/
void blobarray_zero(Blob *aBlob, int n){
  int i;
  for(i=0; i<n; i++) blob_zero(&aBlob[i]);
}
void blobarray_reset(Blob *aBlob, int n){
  int i;
  for(i=0; i<n; i++) blob_reset(&aBlob[i]);
}

/*
** Parse a blob into space-separated tokens.  Store each token in
** an element of the blobarray aToken[].  aToken[] is nToken elements in
** size.  Return the number of tokens seen.
*/
int blob_tokenize(Blob *pIn, Blob *aToken, int nToken){
  int i;
  for(i=0; i<nToken && blob_token(pIn, &aToken[i]); i++){}
  return i;
}

/*
** Do printf-style string rendering and append the results to a blob.
**
** The blob_appendf() version sets the BLOBFLAG_NotSQL bit in Blob.blobFlags
** whereas blob_append_sql() does not.
*/
void blob_appendf(Blob *pBlob, const char *zFormat, ...){
  if( pBlob ){
    va_list ap;
    va_start(ap, zFormat);
    vxprintf(pBlob, zFormat, ap);
    va_end(ap);
    pBlob->blobFlags |= BLOBFLAG_NotSQL;
  }
}
void blob_append_sql(Blob *pBlob, const char *zFormat, ...){
  if( pBlob ){
    va_list ap;
    va_start(ap, zFormat);
    vxprintf(pBlob, zFormat, ap);
    va_end(ap);
  }
}
void blob_vappendf(Blob *pBlob, const char *zFormat, va_list ap){
  if( pBlob ) vxprintf(pBlob, zFormat, ap);
}

/*
** Initialize a blob to the data on an input channel.  Return
** the number of bytes read into the blob.  Any prior content
** of the blob is discarded, not freed.
*/
int blob_read_from_channel(Blob *pBlob, FILE *in, int nToRead){
  size_t n;
  blob_zero(pBlob);
  if( nToRead<0 ){
    char zBuf[10000];
    while( !feof(in) ){
      n = fread(zBuf, 1, sizeof(zBuf), in);
      if( n>0 ){
        blob_append(pBlob, zBuf, n);
      }
    }
  }else{
    blob_resize(pBlob, nToRead);
    n = fread(blob_buffer(pBlob), 1, nToRead, in);
    blob_resize(pBlob, n);
  }
  return blob_size(pBlob);
}

/*
** Initialize a blob to be the content of a file.  If the filename
** is blank or "-" then read from standard input.
**
** If zFilename is a symbolic link, behavior depends on the eFType
** parameter:
**
**    *  If eFType is ExtFILE or allow-symlinks is OFF, then the
**       pBlob is initialized to the *content* of the object to which
**       the zFilename symlink points.
**
**    *  If eFType is RepoFILE and allow-symlinks is ON, then the
**       pBlob is initialized to the *name* of the object to which
**       the zFilename symlink points.
**
** Any prior content of the blob is discarded, not freed.
**
** Return the number of bytes read. Calls fossil_fatal() on error (i.e.
** it exit()s and does not return).
*/
sqlite3_int64 blob_read_from_file(
  Blob *pBlob,               /* The blob to be initialized */
  const char *zFilename,     /* Extract content from this file */
  int eFType                 /* ExtFILE or RepoFILE - see above */
){
  sqlite3_int64 size, got;
  FILE *in;
  if( zFilename==0 || zFilename[0]==0
        || (zFilename[0]=='-' && zFilename[1]==0) ){
    return blob_read_from_channel(pBlob, stdin, -1);
  }
  if( file_islink(zFilename) ){
    return blob_read_link(pBlob, zFilename);
  }
  size = file_size(zFilename, eFType);
  blob_zero(pBlob);
  if( size<0 ){
    fossil_fatal("no such file: %s", zFilename);
  }
  if( size==0 ){
    return 0;
  }
  blob_resize(pBlob, size);
  in = fossil_fopen(zFilename, "rb");
  if( in==0 ){
    fossil_fatal("cannot open %s for reading", zFilename);
  }
  got = fread(blob_buffer(pBlob), 1, size, in);
  fclose(in);
  if( got<size ){
    blob_resize(pBlob, got);
  }
  return got;
}

/*
** Reads symlink destination path and puts int into blob.
** Any prior content of the blob is discarded, not freed.
**
** Returns length of destination path.
**
** On windows, zeros blob and returns 0.
*/
int blob_read_link(Blob *pBlob, const char *zFilename){
#if !defined(_WIN32)
  char zBuf[1024];
  ssize_t len = readlink(zFilename, zBuf, 1023);
  if( len < 0 ){
    fossil_fatal("cannot read symbolic link %s", zFilename);
  }
  zBuf[len] = 0;   /* null-terminate */
  blob_zero(pBlob);
  blob_appendf(pBlob, "%s", zBuf);
  return len;
#else
  blob_zero(pBlob);
  return 0;
#endif
}

/*
** Write the content of a blob into a file.
**
** If the filename is blank or "-" then write to standard output.
**
** This routine always assumes ExtFILE.  If zFilename is a symbolic link
** then the content is written into the object that symbolic link points
** to, not into the symbolic link itself.  This is true regardless of
** the allow-symlinks setting.
**
** Return the number of bytes written.
*/
int blob_write_to_file(Blob *pBlob, const char *zFilename){
  FILE *out;
  int nWrote;

  if( zFilename[0]==0 || (zFilename[0]=='-' && zFilename[1]==0) ){
    blob_is_init(pBlob);
#if defined(_WIN32)
    nWrote = fossil_utf8_to_console(blob_buffer(pBlob), blob_size(pBlob), 0);
    if( nWrote>=0 ) return nWrote;
    fflush(stdout);
    _setmode(_fileno(stdout), _O_BINARY);
#endif
    nWrote = fwrite(blob_buffer(pBlob), 1, blob_size(pBlob), stdout);
#if defined(_WIN32)
    fflush(stdout);
    _setmode(_fileno(stdout), _O_TEXT);
#endif
  }else{
    file_mkfolder(zFilename, ExtFILE, 1, 0);
    out = fossil_fopen(zFilename, "wb");
    if( out==0 ){
#if _WIN32
      const char *zReserved = file_is_win_reserved(zFilename);
      if( zReserved ){
        fossil_fatal("cannot open \"%s\" because \"%s\" is "
             "a reserved name on Windows", zFilename, zReserved);
      }
#endif
      fossil_fatal_recursive("unable to open file \"%s\" for writing",
                             zFilename);
      return 0;
    }
    blob_is_init(pBlob);
    nWrote = fwrite(blob_buffer(pBlob), 1, blob_size(pBlob), out);
    fclose(out);
    if( nWrote!=blob_size(pBlob) ){
      fossil_fatal_recursive("short write: %d of %d bytes to %s", nWrote,
         blob_size(pBlob), zFilename);
    }
  }
  return nWrote;
}

/*
** Compress a blob pIn.  Store the result in pOut.  It is ok for pIn and
** pOut to be the same blob.
**
** pOut must either be the same as pIn or else uninitialized.
*/
void blob_compress(Blob *pIn, Blob *pOut){
  unsigned int nIn = blob_size(pIn);
  unsigned int nOut = 13 + nIn + (nIn+999)/1000;
  unsigned long int nOut2;
  unsigned char *outBuf;
  Blob temp;
  blob_zero(&temp);
  blob_resize(&temp, nOut+4);
  outBuf = (unsigned char*)blob_buffer(&temp);
  outBuf[0] = nIn>>24 & 0xff;
  outBuf[1] = nIn>>16 & 0xff;
  outBuf[2] = nIn>>8 & 0xff;
  outBuf[3] = nIn & 0xff;
  nOut2 = (long int)nOut;
  compress(&outBuf[4], &nOut2,
           (unsigned char*)blob_buffer(pIn), blob_size(pIn));
  if( pOut==pIn ) blob_reset(pOut);
  assert_blob_is_reset(pOut);
  *pOut = temp;
  blob_resize(pOut, nOut2+4);
}

/*
** COMMAND: test-compress
**
** Usage: %fossil test-compress INPUTFILE OUTPUTFILE
**
** Run compression on INPUTFILE and write the result into OUTPUTFILE.
**
** This is used to test and debug the blob_compress() routine.
*/
void compress_cmd(void){
  Blob f;
  if( g.argc!=4 ) usage("INPUTFILE OUTPUTFILE");
  blob_read_from_file(&f, g.argv[2], ExtFILE);
  blob_compress(&f, &f);
  blob_write_to_file(&f, g.argv[3]);
}

/*
** Compress the concatenation of a blobs pIn1 and pIn2.  Store the result
** in pOut.
**
** pOut must be either uninitialized or must be the same as either pIn1 or
** pIn2.
*/
void blob_compress2(Blob *pIn1, Blob *pIn2, Blob *pOut){
  unsigned int nIn = blob_size(pIn1) + blob_size(pIn2);
  unsigned int nOut = 13 + nIn + (nIn+999)/1000;
  unsigned char *outBuf;
  z_stream stream;
  Blob temp;
  blob_zero(&temp);
  blob_resize(&temp, nOut+4);
  outBuf = (unsigned char*)blob_buffer(&temp);
  outBuf[0] = nIn>>24 & 0xff;
  outBuf[1] = nIn>>16 & 0xff;
  outBuf[2] = nIn>>8 & 0xff;
  outBuf[3] = nIn & 0xff;
  stream.zalloc = (alloc_func)0;
  stream.zfree = (free_func)0;
  stream.opaque = 0;
  stream.avail_out = nOut;
  stream.next_out = &outBuf[4];
  deflateInit(&stream, 9);
  stream.avail_in = blob_size(pIn1);
  stream.next_in = (unsigned char*)blob_buffer(pIn1);
  deflate(&stream, 0);
  stream.avail_in = blob_size(pIn2);
  stream.next_in = (unsigned char*)blob_buffer(pIn2);
  deflate(&stream, 0);
  deflate(&stream, Z_FINISH);
  blob_resize(&temp, stream.total_out + 4);
  deflateEnd(&stream);
  if( pOut==pIn1 ) blob_reset(pOut);
  if( pOut==pIn2 ) blob_reset(pOut);
  assert_blob_is_reset(pOut);
  *pOut = temp;
}

/*
** COMMAND: test-compress-2
**
** Usage: %fossil test-compress-2 IN1 IN2 OUT
**
** Read files IN1 and IN2, concatenate the content, compress the
** content, then write results into OUT.
**
** This is used to test and debug the blob_compress2() routine.
*/
void compress2_cmd(void){
  Blob f1, f2;
  if( g.argc!=5 ) usage("INPUTFILE1 INPUTFILE2 OUTPUTFILE");
  blob_read_from_file(&f1, g.argv[2], ExtFILE);
  blob_read_from_file(&f2, g.argv[3], ExtFILE);
  blob_compress2(&f1, &f2, &f1);
  blob_write_to_file(&f1, g.argv[4]);
}

/*
** Uncompress blob pIn and store the result in pOut.  It is ok for pIn and
** pOut to be the same blob.
**
** pOut must be either uninitialized or the same as pIn.
*/
int blob_uncompress(Blob *pIn, Blob *pOut){
  unsigned int nOut;
  unsigned char *inBuf;
  unsigned int nIn = blob_size(pIn);
  Blob temp;
  int rc;
  unsigned long int nOut2;
  if( nIn<=4 ){
    return 0;
  }
  inBuf = (unsigned char*)blob_buffer(pIn);
  nOut = (inBuf[0]<<24) + (inBuf[1]<<16) + (inBuf[2]<<8) + inBuf[3];
  blob_zero(&temp);
  blob_resize(&temp, nOut+1);
  nOut2 = (long int)nOut;
  rc = uncompress((unsigned char*)blob_buffer(&temp), &nOut2,
                  &inBuf[4], nIn - 4);
  if( rc!=Z_OK ){
    blob_reset(&temp);
    return 1;
  }
  blob_resize(&temp, nOut2);
  if( pOut==pIn ) blob_reset(pOut);
  assert_blob_is_reset(pOut);
  *pOut = temp;
  return 0;
}

/*
** COMMAND: test-uncompress
**
** Usage: %fossil test-uncompress IN OUT
**
** Read the content of file IN, uncompress that content, and write the
** result into OUT.  This command is intended for testing of the
** blob_compress() function.
*/
void uncompress_cmd(void){
  Blob f;
  if( g.argc!=4 ) usage("INPUTFILE OUTPUTFILE");
  blob_read_from_file(&f, g.argv[2], ExtFILE);
  blob_uncompress(&f, &f);
  blob_write_to_file(&f, g.argv[3]);
}

/*
** COMMAND: test-cycle-compress
**
** Compress and uncompress each file named on the command line.
** Verify that the original content is recovered.
*/
void test_cycle_compress(void){
  int i;
  Blob b1, b2, b3;
  for(i=2; i<g.argc; i++){
    blob_read_from_file(&b1, g.argv[i], ExtFILE);
    blob_compress(&b1, &b2);
    blob_uncompress(&b2, &b3);
    if( blob_compare(&b1, &b3) ){
      fossil_fatal("compress/uncompress cycle failed for %s", g.argv[i]);
    }
    blob_reset(&b1);
    blob_reset(&b2);
    blob_reset(&b3);
  }
  fossil_print("ok\n");
}

#if defined(_WIN32) || defined(__CYGWIN__)
/*
** Convert every \n character in the given blob into \r\n.
*/
void blob_add_cr(Blob *p){
  char *z = p->aData;
  int j   = p->nUsed;
  int i, n;
  for(i=n=0; i<j; i++){
    if( z[i]=='\n' ) n++;
  }
  j += n;
  if( j>=p->nAlloc ){
    blob_resize(p, j);
    z = p->aData;
  }
  p->nUsed = j;
  z[j] = 0;
  while( j>i ){
    if( (z[--j] = z[--i]) =='\n' ){
      z[--j] = '\r';
    }
  }
}
#endif

/*
** Remove every \r character from the given blob, replacing each one with
** a \n character if it was not already part of a \r\n pair.
*/
void blob_to_lf_only(Blob *p){
  int i, j;
  char *z = blob_materialize(p);
  for(i=j=0; z[i]; i++){
    if( z[i]!='\r' ) z[j++] = z[i];
    else if( z[i+1]!='\n' ) z[j++] = '\n';
  }
  z[j] = 0;
  p->nUsed = j;
}

/*
** Convert blob from cp1252 to UTF-8. As cp1252 is a superset
** of iso8859-1, this is useful on UNIX as well.
**
** This table contains the character translations for 0x80..0xA0.
*/

static const unsigned short cp1252[32] = {
  0x20ac,   0x81, 0x201A, 0x0192, 0x201E, 0x2026, 0x2020, 0x2021,
  0x02C6, 0x2030, 0x0160, 0x2039, 0x0152,   0x8D, 0x017D,   0x8F,
    0x90, 0x2018, 0x2019, 0x201C, 0x201D, 0x2022, 0x2013, 0x2014,
   0x2DC, 0x2122, 0x0161, 0x203A, 0x0153,   0x9D, 0x017E, 0x0178
};

void blob_cp1252_to_utf8(Blob *p){
  unsigned char *z = (unsigned char *)p->aData;
  int j   = p->nUsed;
  int i, n;
  for(i=n=0; i<j; i++){
    if( z[i]>=0x80 ){
      if( (z[i]<0xa0) && (cp1252[z[i]&0x1f]>=0x800) ){
        n++;
      }
      n++;
    }
  }
  j += n;
  if( j>=p->nAlloc ){
    blob_resize(p, j);
    z = (unsigned char *)p->aData;
  }
  p->nUsed = j;
  z[j] = 0;
  while( j>i ){
    if( z[--i]>=0x80 ){
      if( z[i]<0xa0 ){
        unsigned short sym = cp1252[z[i]&0x1f];
        if( sym>=0x800 ){
          z[--j] = 0x80 | (sym&0x3f);
          z[--j] = 0x80 | ((sym>>6)&0x3f);
          z[--j] = 0xe0 | (sym>>12);
        }else{
          z[--j] = 0x80 | (sym&0x3f);
          z[--j] = 0xc0 | (sym>>6);
        }
      }else{
        z[--j] = 0x80 | (z[i]&0x3f);
        z[--j] = 0xC0 | (z[i]>>6);
      }
    }else{
      z[--j] = z[i];
    }
  }
}

/*
** pBlob is a shell command under construction.  This routine safely
** appends argument zIn.
**
** The argument is escaped if it contains white space or other characters
** that need to be escaped for the shell.  If zIn contains characters
** that cannot be safely escaped, then throw a fatal error.
**
** The argument is expected to a filename of some kinds.  As shell commands
** commonly have command-line options that begin with "-" and since we
** do not want an attacker to be able to invoke these switches using
** filenames that begin with "-", if zIn begins with "-", prepend
** an additional "./".
*/
void blob_append_escaped_arg(Blob *pBlob, const char *zIn){
  int i;
  char c;
  int needEscape = 0;
  int n = blob_size(pBlob);
  char *z = blob_buffer(pBlob);
#if defined(_WIN32)
  const char cQuote = '"';    /* Use "..." quoting on windows */
#else
  const char cQuote = '\'';   /* Use '...' quoting on unix */
#endif

  for(i=0; (c = zIn[i])!=0; i++){
    if( c==cQuote || c=='\\' || c<' ' || c==';' || c=='*' || c=='?' || c=='[') {
      Blob bad;
      blob_token(pBlob, &bad);
      fossil_fatal("the [%s] argument to the \"%s\" command contains "
                   "a character (ascii 0x%02x) that is a security risk",
                   zIn, blob_str(&bad), c);
    }
    if( !needEscape && !fossil_isalnum(c) && c!='/' && c!='.' && c!='_' ){
      needEscape = 1;
    }
  }
  if( n>0 && !fossil_isspace(z[n-1]) ){
    blob_append_char(pBlob, ' ');
  }
  if( needEscape ) blob_append_char(pBlob, cQuote);
  if( zIn[0]=='-' ) blob_append(pBlob, "./", 2);
  blob_append(pBlob, zIn, -1);
  if( needEscape ) blob_append_char(pBlob, cQuote);
}

/*
** A read(2)-like impl for the Blob class. Reads (copies) up to nLen
** bytes from pIn, starting at position pIn->iCursor, and copies them
** to pDest (which must be valid memory at least nLen bytes long).
**
** Returns the number of bytes read/copied, which may be less than
** nLen (if end-of-blob is encountered).
**
** Updates pIn's cursor.
**
** Returns 0 if pIn contains no data.
*/
unsigned int blob_read(Blob *pIn, void * pDest, unsigned int nLen ){
  if( !pIn->aData || (pIn->iCursor >= pIn->nUsed) ){
    return 0;
  } else if( (pIn->iCursor + nLen) > (unsigned int)pIn->nUsed ){
    nLen = (unsigned int) (pIn->nUsed - pIn->iCursor);
  }
  assert( pIn->nUsed > pIn->iCursor );
  assert( (pIn->iCursor+nLen)  <= pIn->nUsed );
  if( nLen ){
    memcpy( pDest, pIn->aData, nLen );
    pIn->iCursor += nLen;
  }
  return nLen;
}

/*
** Swaps the contents of the given blobs. Results
** are unspecified if either value is NULL or both
** point to the same blob.
*/
void blob_swap( Blob *pLeft, Blob *pRight ){
  Blob swap = *pLeft;
  *pLeft = *pRight;
  *pRight = swap;
}

/*
** Strip a possible byte-order-mark (BOM) from the blob. On Windows, if there
** is either no BOM at all or an (le/be) UTF-16 BOM, a conversion to UTF-8 is
** done.  If useMbcs is false and there is no BOM, the input string is assumed
** to be UTF-8 already, so no conversion is done.
*/
void blob_to_utf8_no_bom(Blob *pBlob, int useMbcs){
  char *zUtf8;
  int bomSize = 0;
  int bomReverse = 0;
  if( starts_with_utf8_bom(pBlob, &bomSize) ){
    struct Blob temp;
    zUtf8 = blob_str(pBlob) + bomSize;
    blob_zero(&temp);
    blob_append(&temp, zUtf8, -1);
    blob_swap(pBlob, &temp);
    blob_reset(&temp);
  }else if( starts_with_utf16_bom(pBlob, &bomSize, &bomReverse) ){
    zUtf8 = blob_buffer(pBlob);
    if( bomReverse ){
      /* Found BOM, but with reversed bytes */
      unsigned int i = blob_size(pBlob);
      while( i>0 ){
        /* swap bytes of unicode representation */
        char zTemp = zUtf8[--i];
        zUtf8[i] = zUtf8[i-1];
        zUtf8[--i] = zTemp;
      }
    }
    /* Make sure the blob contains two terminating 0-bytes */
    blob_append_char(pBlob, 0);
    zUtf8 = blob_str(pBlob) + bomSize;
    zUtf8 = fossil_unicode_to_utf8(zUtf8);
    blob_set_dynamic(pBlob, zUtf8);
  }else if( useMbcs && invalid_utf8(pBlob) ){
#if defined(_WIN32) || defined(__CYGWIN__)
    zUtf8 = fossil_mbcs_to_utf8(blob_str(pBlob));
    blob_reset(pBlob);
    blob_append(pBlob, zUtf8, -1);
    fossil_mbcs_free(zUtf8);
#else
    blob_cp1252_to_utf8(pBlob);
#endif /* _WIN32 */
  }
}