~vcs-imports/squirrelmail/trunk

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
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
<?php

/**
 * strings.php
 *
 * This code provides various string manipulation functions that are
 * used by the rest of the SquirrelMail code.
 *
 * @copyright 1999-2015 The SquirrelMail Project Team
 * @license http://opensource.org/licenses/gpl-license.php GNU Public License
 * @version $Id$
 * @package squirrelmail
 */

/**
 * Appends citation markers to the string.
 * Also appends a trailing space.
 *
 * @author Justus Pendleton
 * @param string $str The string to append to
 * @param int $citeLevel the number of markers to append
 * @return null
 * @since 1.5.1
 */
function sqMakeCite (&$str, $citeLevel) {
    for ($i = 0; $i < $citeLevel; $i++) {
        $str .= '>';
    }
    if ($citeLevel != 0) {
        $str .= ' ';
    }
}

/**
 * Create a newline in the string, adding citation
 * markers to the newline as necessary.
 *
 * @author Justus Pendleton
 * @param string $str the string to make a newline in
 * @param int $citeLevel the citation level the newline is at
 * @param int $column starting column of the newline
 * @return null
 * @since 1.5.1
 */
function sqMakeNewLine (&$str, $citeLevel, &$column) {
    $str .= "\n";
    $column = 0;
    if ($citeLevel > 0) {
        sqMakeCite ($str, $citeLevel);
        $column = $citeLevel + 1;
    } else {
        $column = 0;
    }
}

/**
 * Checks for spaces in strings - only used if PHP doesn't have native ctype support
 *
 * You might be able to rewrite the function by adding short evaluation form.
 *
 * possible problems:
 *  - iso-2022-xx charsets  - hex 20 might be part of other symbol. I might
 * be wrong. 0x20 is not used in iso-2022-jp. I haven't checked iso-2022-kr
 * and iso-2022-cn mappings.
 *
 *  - no-break space (&nbsp;) - it is 8bit symbol, that depends on charset.
 * there are at least three different charset groups that have nbsp in
 * different places.
 *
 * I don't see any charset/nbsp options in php ctype either.
 *
 * @param string $string tested string
 * @return bool true when only whitespace symbols are present in test string
 * @since 1.5.1
 */
function sm_ctype_space($string) {
    if ( preg_match('/^[\x09-\x0D]|^\x20/', $string) || $string=='') {
        return true;
    } else {
        return false;
    }
}

/**
 * Wraps text at $wrap characters.  While sqWordWrap takes
 * a single line of text and wraps it, this function works
 * on the entire corpus at once, this allows it to be a little
 * bit smarter and when and how to wrap.
 *
 * @author Justus Pendleton
 * @param string $body the entire body of text
 * @param int $wrap the maximum line length
 * @return string the wrapped text
 * @since 1.5.1
 */
function &sqBodyWrap (&$body, $wrap) {
    //check for ctype support, and fake it if it doesn't exist
    if (!function_exists('ctype_space')) {
        function ctype_space ($string) {
            return sm_ctype_space($string);
        }
    }

    // the newly wrapped text
    $outString = '';
    // current column since the last newline in the outstring
    $outStringCol = 0;
    $length = sq_strlen($body);
    // where we are in the original string
    $pos = 0;
    // the number of >>> citation markers we are currently at
    $citeLevel = 0;

    // the main loop, whenever we start a newline of input text
    // we start from here
    while ($pos < $length) {
       // we're at the beginning of a line, get the new cite level
       $newCiteLevel = 0;

       while (($pos < $length) && (sq_substr($body,$pos,1) == '>')) {
           $newCiteLevel++;
           $pos++;

           // skip over any spaces interleaved among the cite markers
           while (($pos < $length) && (sq_substr($body,$pos,1) == ' ')) {

               $pos++;

           }
           if ($pos >= $length) {
               break;
           }
       }

       // special case: if this is a blank line then maintain it
       // (i.e. try to preserve original paragraph breaks)
       // unless they occur at the very beginning of the text
       if ((sq_substr($body,$pos,1) == "\n" ) && (sq_strlen($outString) != 0)) {
           $outStringLast = $outString{sq_strlen($outString) - 1};
           if ($outStringLast != "\n") {
               $outString .= "\n";
           }
           sqMakeCite ($outString, $newCiteLevel);
           $outString .= "\n";
           $pos++;
           $outStringCol = 0;
           continue;
       }

       // if the cite level has changed, then start a new line
       // with the new cite level.
       if (($citeLevel != $newCiteLevel) && ($pos > ($newCiteLevel + 1)) && ($outStringCol != 0)) {
           sqMakeNewLine ($outString, 0, $outStringCol);
       }

       $citeLevel = $newCiteLevel;

       // prepend the quote level if necessary
       if ($outStringCol == 0) {
           sqMakeCite ($outString, $citeLevel);
           // if we added a citation then move the column
           // out by citelevel + 1 (the cite markers + the space)
           $outStringCol = $citeLevel + ($citeLevel ? 1 : 0);
       } else if ($outStringCol > $citeLevel) {
           // not a cite and we're not at the beginning of a line
           // in the output.  add a space to separate the new text
           // from previous text.
           $outString .= ' ';
           $outStringCol++;
       }

       // find the next newline -- we don't want to go further than that
       $nextNewline = sq_strpos ($body, "\n", $pos);
       if ($nextNewline === FALSE) {
           $nextNewline = $length;
       }

       // Don't wrap unquoted lines at all.  For now the textarea
       // will work fine for this.  Maybe revisit this later though
       // (for completeness more than anything else, I think)
       if ($citeLevel == 0) {
           $outString .= sq_substr ($body, $pos, ($nextNewline - $pos));
           $outStringCol = $nextNewline - $pos;
           if ($nextNewline != $length) {
               sqMakeNewLine ($outString, 0, $outStringCol);
           }
           $pos = $nextNewline + 1;
           continue;
       }
       /**
        * Set this to false to stop appending short strings to previous lines
        */
       $smartwrap = true;
       // inner loop, (obviously) handles wrapping up to
       // the next newline
       while ($pos < $nextNewline) {
           // skip over initial spaces
           while (($pos < $nextNewline) && (ctype_space (sq_substr($body,$pos,1)))) {
               $pos++;
           }
           // if this is a short line then just append it and continue outer loop
           if (($outStringCol + $nextNewline - $pos) <= ($wrap - $citeLevel - 1) ) {
               // if this is the final line in the input string then include
               // any trailing newlines
               //      echo substr($body,$pos,$wrap). "<br />";
               if (($nextNewline + 1 == $length) && (sq_substr($body,$nextNewline,1) == "\n")) {
                   $nextNewline++;
               }

               // trim trailing spaces
               $lastRealChar = $nextNewline;
               while (($lastRealChar > $pos && $lastRealChar < $length) && (ctype_space (sq_substr($body,$lastRealChar,1)))) {
                   $lastRealChar--;
               }
               // decide if appending the short string is what we want
               if (($nextNewline < $length && sq_substr($body,$nextNewline,1) == "\n") &&
                     isset($lastRealChar)) {
                   $mypos = $pos;
                   //check the first word:
                   while (($mypos < $length) && (sq_substr($body,$mypos,1) == '>')) {
                       $mypos++;
                       // skip over any spaces interleaved among the cite markers
                       while (($mypos < $length) && (sq_substr($body,$mypos,1) == ' ')) {
                           $mypos++;
                       }
                   }
/*
                     $ldnspacecnt = 0;
                     if ($mypos == $nextNewline+1) {
                        while (($mypos < $length) && ($body{$mypos} == ' ')) {
                         $ldnspacecnt++;
                        }
                     }
*/

                   $firstword = sq_substr($body,$mypos,sq_strpos($body,' ',$mypos) - $mypos);
                   //if ($dowrap || $ldnspacecnt > 1 || ($firstword && (
                   if (!$smartwrap || $firstword && (
                                        $firstword{0} == '-' ||
                                        $firstword{0} == '+' ||
                                        $firstword{0} == '*' ||
                                        sq_substr($firstword,0,1) == sq_strtoupper(sq_substr($firstword,0,1)) ||
                                        strpos($firstword,':'))) {
                        $outString .= sq_substr($body,$pos,($lastRealChar - $pos+1));
                        $outStringCol += ($lastRealChar - $pos);
                        sqMakeNewLine($outString,$citeLevel,$outStringCol);
                        $nextNewline++;
                        $pos = $nextNewline;
                        $outStringCol--;
                        continue;
                   }

               }

               $outString .= sq_substr ($body, $pos, ($lastRealChar - $pos + 1));
               $outStringCol += ($lastRealChar - $pos);
               $pos = $nextNewline + 1;
               continue;
           }

           $eol = $pos + $wrap - $citeLevel - $outStringCol;
           // eol is the tentative end of line.
           // look backwards for there for a whitespace to break at.
           // if it's already less than our current position then
           // our current line is already too long, break immediately
           // and restart outer loop
           if ($eol <= $pos) {
               sqMakeNewLine ($outString, $citeLevel, $outStringCol);
               continue;
           }

           // start looking backwards for whitespace to break at.
           $breakPoint = $eol;
           while (($breakPoint > $pos) && (! ctype_space (sq_substr($body,$breakPoint,1)))) {
               $breakPoint--;
           }

           // if we didn't find a breakpoint by looking backward then we
           // need to figure out what to do about that
           if ($breakPoint == $pos) {
               // if we are not at the beginning then end this line
               // and start a new loop
               if ($outStringCol > ($citeLevel + 1)) {
                   sqMakeNewLine ($outString, $citeLevel, $outStringCol);
                   continue;
               } else {
                   // just hard break here.  most likely we are breaking
                   // a really long URL.  could also try searching
                   // forward for a break point, which is what Mozilla
                   // does.  don't bother for now.
                   $breakPoint = $eol;
               }
           }

           // special case: maybe we should have wrapped last
           // time.  if the first breakpoint here makes the
           // current line too long and there is already text on
           // the current line, break and loop again if at
           // beginning of current line, don't force break
           $SLOP = 6;
           if ((($outStringCol + ($breakPoint - $pos)) > ($wrap + $SLOP)) && ($outStringCol > ($citeLevel + 1))) {
               sqMakeNewLine ($outString, $citeLevel, $outStringCol);
               continue;
           }

           // skip newlines or whitespace at the beginning of the string
           $substring = sq_substr ($body, $pos, ($breakPoint - $pos));
           $substring = rtrim ($substring); // do rtrim and ctype_space have the same ideas about whitespace?
           $outString .= $substring;
           $outStringCol += sq_strlen ($substring);
           // advance past the whitespace which caused the wrap
           $pos = $breakPoint;
           while (($pos < $length) && (ctype_space (sq_substr($body,$pos,1)))) {
               $pos++;
           }
           if ($pos < $length) {
               sqMakeNewLine ($outString, $citeLevel, $outStringCol);
           }
       }
    }

    return $outString;
}

/**
 * Wraps text at $wrap characters
 *
 * Has a problem with special HTML characters, so call this before
 * you do character translation.
 *
 * Specifically, &amp;#039; comes up as 5 characters instead of 1.
 * This should not add newlines to the end of lines.
 *
 * @param string $line the line of text to wrap, by ref
 * @param int $wrap the maximum line lenth
 * @param string $charset name of charset used in $line string. Available since v.1.5.1.
 * @return void
 * @since 1.0
 */
function sqWordWrap(&$line, $wrap, $charset='') {
    global $languages, $squirrelmail_language;

    // Use custom wrapping function, if translation provides it
    if (isset($languages[$squirrelmail_language]['XTRA_CODE']) &&
        function_exists($languages[$squirrelmail_language]['XTRA_CODE'] . '_wordwrap')) {
        if (mb_detect_encoding($line) != 'ASCII') {
            $line = call_user_func($languages[$squirrelmail_language]['XTRA_CODE'] . '_wordwrap', $line, $wrap);
            return;
        }
    }

    preg_match('/^([\t >]*)([^\t >].*)?$/', $line, $regs);
    $beginning_spaces = $regs[1];
    if (isset($regs[2])) {
        $words = explode(' ', $regs[2]);
    } else {
        $words = array();
    }

    $i = 0;
    $line = $beginning_spaces;

    while ($i < count($words)) {
        /* Force one word to be on a line (minimum) */
        $line .= $words[$i];
        $line_len = strlen($beginning_spaces) + sq_strlen($words[$i],$charset) + 2;
        if (isset($words[$i + 1]))
            $line_len += sq_strlen($words[$i + 1],$charset);
        $i ++;

        /* Add more words (as long as they fit) */
        while ($line_len < $wrap && $i < count($words)) {
            $line .= ' ' . $words[$i];
            $i++;
            if (isset($words[$i]))
                $line_len += sq_strlen($words[$i],$charset) + 1;
            else
                $line_len += 1;
        }

        /* Skip spaces if they are the first thing on a continued line */
        while (!isset($words[$i]) && $i < count($words)) {
            $i ++;
        }

        /* Go to the next line if we have more to process */
        if ($i < count($words)) {
            $line .= "\n";
        }
    }
}

/**
 * Does the opposite of sqWordWrap()
 * @param string $body the text to un-wordwrap
 * @return void
 * @since 1.0
 */
function sqUnWordWrap(&$body) {
    global $squirrelmail_language;

    if ($squirrelmail_language == 'ja_JP') {
        return;
    }

    $lines = explode("\n", $body);
    $body = '';
    $PreviousSpaces = '';
    $cnt = count($lines);
    for ($i = 0; $i < $cnt; $i ++) {
        preg_match("/^([\t >]*)([^\t >].*)?$/", $lines[$i], $regs);
        $CurrentSpaces = $regs[1];
        if (isset($regs[2])) {
            $CurrentRest = $regs[2];
        } else {
            $CurrentRest = '';
        }

        if ($i == 0) {
            $PreviousSpaces = $CurrentSpaces;
            $body = $lines[$i];
        } else if (($PreviousSpaces == $CurrentSpaces) /* Do the beginnings match */
                   && (strlen($lines[$i - 1]) > 65)    /* Over 65 characters long */
                   && strlen($CurrentRest)) {          /* and there's a line to continue with */
            $body .= ' ' . $CurrentRest;
        } else {
            $body .= "\n" . $lines[$i];
            $PreviousSpaces = $CurrentSpaces;
        }
    }
    $body .= "\n";
}

/**
 * If $haystack is a full mailbox name and $needle is the mailbox
 * separator character, returns the last part of the mailbox name.
 *
 * @param string haystack full mailbox name to search
 * @param string needle the mailbox separator character
 * @return string the last part of the mailbox name
 * @since 1.0
 */
function readShortMailboxName($haystack, $needle) {

    if ($needle == '') {
        $elem = $haystack;
    } else {
        $parts = explode($needle, $haystack);
        $elem = array_pop($parts);
        while ($elem == '' && count($parts)) {
            $elem = array_pop($parts);
        }
    }
    return( $elem );
}


/**
 * get_location
 *
 * Determines the location to forward to, relative to your server.
 * This is used in HTTP Location: redirects.
 *
 * If set, it uses $config_location_base as the first part of the URL,
 * specifically, the protocol, hostname and port parts. The path is
 * always autodetected.
 *
 * @return string the base url for this SquirrelMail installation
 * @since 1.0
 */
function get_location () {

    global $imap_server_type, $config_location_base,
           $is_secure_connection, $sq_ignore_http_x_forwarded_headers;

    /* Get the path, handle virtual directories */
    if(strpos(php_self(), '?')) {
        $path = substr(php_self(), 0, strpos(php_self(), '?'));
    } else {
        $path = php_self();
    }
    $path = substr($path, 0, strrpos($path, '/'));

    // proto+host+port are already set in config:
    if ( !empty($config_location_base) ) {
        return $config_location_base . $path ;
    }
    // we computed it before, get it from the session:
    if ( sqgetGlobalVar('sq_base_url', $full_url, SQ_SESSION) ) {
      return $full_url . $path;
    }
    // else: autodetect

    /* Check if this is a HTTPS or regular HTTP request. */
    $proto = 'http://';
    if ($is_secure_connection)
        $proto = 'https://';

    /* Get the hostname from the Host header or server config. */
    if ($sq_ignore_http_x_forwarded_headers
     || !sqgetGlobalVar('HTTP_X_FORWARDED_HOST', $host, SQ_SERVER)
     || empty($host)) {
        if ( !sqgetGlobalVar('HTTP_HOST', $host, SQ_SERVER) || empty($host) ) {
            if ( !sqgetGlobalVar('SERVER_NAME', $host, SQ_SERVER) || empty($host) ) {
                $host = '';
            }
        }
    }

    $port = '';
    if (! strstr($host, ':')) {
        // Note: HTTP_X_FORWARDED_PROTO could be sent from the client and
        //       therefore possibly spoofed/hackable.  Thus, SquirrelMail
        //       ignores such headers by default.  The administrator
        //       can tell SM to use such header values by setting
        //       $sq_ignore_http_x_forwarded_headers to boolean FALSE
        //       in config/config.php or by using config/conf.pl.
        global $sq_ignore_http_x_forwarded_headers;
        if ($sq_ignore_http_x_forwarded_headers
         || !sqgetGlobalVar('HTTP_X_FORWARDED_PROTO', $forwarded_proto, SQ_SERVER))
            $forwarded_proto = '';
        if (sqgetGlobalVar('SERVER_PORT', $server_port, SQ_SERVER)) {
            if (($server_port != 80 && $proto == 'http://') ||
                ($server_port != 443 && $proto == 'https://' &&
                 strcasecmp($forwarded_proto, 'https') !== 0)) {
                $port = sprintf(':%d', $server_port);
            }
        }
    }

    /* this is a workaround for the weird macosx caching that
     * causes Apache to return 16080 as the port number, which causes
     * SM to bail */

    if ($imap_server_type == 'macosx' && $port == ':16080') {
        $port = '';
    }

    /* Fallback is to omit the server name and use a relative */
    /* URI, although this is not RFC 2616 compliant.          */
    $full_url = ($host ? $proto . $host . $port : '');
    sqsession_register($full_url, 'sq_base_url');
    return $full_url . $path;
}


/**
 * Get Message List URI
 *
 * @param string $mailbox      Current mailbox name (unencoded/raw)
 * @param string $startMessage The mailbox page offset
 * @param string $what         Any current search parameters (OPTIONAL; 
 *                             default empty string)
 *
 * @return string The message list URI
 *
 * @since 1.5.2
 *
 */
function get_message_list_uri($mailbox, $startMessage, $what='') {

    global $base_uri;

    $urlMailbox = urlencode($mailbox);

    $list_xtra = "?where=read_body.php&what=$what&mailbox=" . $urlMailbox.
                 "&startMessage=$startMessage";

    return $base_uri .'src/right_main.php'. $list_xtra;
}


/**
 * Encrypts password
 *
 * These functions are used to encrypt the password before it is
 * stored in a cookie. The encryption key is generated by
 * OneTimePadCreate();
 *
 * @param string $string the (password)string to encrypt
 * @param string $epad the encryption key
 * @return string the base64-encoded encrypted password
 * @since 1.0
 */
function OneTimePadEncrypt ($string, $epad) {
    $pad = base64_decode($epad);

    if (strlen($pad)>0) {
        // make sure that pad is longer than string
        while (strlen($string)>strlen($pad)) {
            $pad.=$pad;
        }
    } else {
        // FIXME: what should we do when $epad is not base64 encoded or empty.
    }

    $encrypted = '';
    for ($i = 0; $i < strlen ($string); $i++) {
        $encrypted .= chr (ord($string[$i]) ^ ord($pad[$i]));
    }

    return base64_encode($encrypted);
}

/**
 * Decrypts a password from the cookie
 *
 * Decrypts a password from the cookie, encrypted by OneTimePadEncrypt.
 * This uses the encryption key that is stored in the session.
 *
 * @param string $string the string to decrypt
 * @param string $epad the encryption key from the session
 * @return string the decrypted password
 * @since 1.0
 */
function OneTimePadDecrypt ($string, $epad) {
    $pad = base64_decode($epad);

    if (strlen($pad)>0) {
        // make sure that pad is longer than string
        while (strlen($string)>strlen($pad)) {
            $pad.=$pad;
        }
    } else {
        // FIXME: what should we do when $epad is not base64 encoded or empty.
    }

    $encrypted = base64_decode ($string);
    $decrypted = '';
    for ($i = 0; $i < strlen ($encrypted); $i++) {
        $decrypted .= chr (ord($encrypted[$i]) ^ ord($pad[$i]));
    }

    return $decrypted;
}

/**
 * Creates encryption key
 *
 * Creates an encryption key for encrypting the password stored in the cookie.
 * The encryption key itself is stored in the session.
 *
 * Pad must be longer or equal to encoded string length in 1.4.4/1.5.0 and older.
 * @param int $length optional, length of the string to generate
 * @return string the encryption key
 * @since 1.0
 */
function OneTimePadCreate ($length=100) {
    $pad = '';
    for ($i = 0; $i < $length; $i++) {
        $pad .= chr(mt_rand(0,255));
    }

    return base64_encode($pad);
}

/**
  * Returns a string showing a byte size figure in
  * a more easily digested (readable) format
  *
  * @param int $bytes the size in bytes
  *
  * @return string The size in human readable format
  *
  * @since 1.0
  *
  */
function show_readable_size($bytes) {
    $bytes /= 1024;
    $type = _("KiB");

    if ($bytes / 1024 > 1) {
        $bytes /= 1024;
        $type = _("MiB");
    }

    if ($bytes < 10) {
        $bytes *= 10;
        settype($bytes, 'integer');
        $bytes /= 10;
    } else {
        settype($bytes, 'integer');
    }

    global $nbsp;
    return $bytes . $nbsp . $type;
}

/**
 * Generates a random string from the character set you pass in
 *
 * @param int $size the length of the string to generate
 * @param string $chars a string containing the characters to use
 * @param int $flags a flag to add a specific set to the characters to use:
 *     Flags:
 *       1 = add lowercase a-z to $chars
 *       2 = add uppercase A-Z to $chars
 *       4 = add numbers 0-9 to $chars
 * @return string the random string
 * @since 1.0
 */
function GenerateRandomString($size, $chars, $flags = 0) {
    if ($flags & 0x1) {
        $chars .= 'abcdefghijklmnopqrstuvwxyz';
    }
    if ($flags & 0x2) {
        $chars .= 'ABCDEFGHIJKLMNOPQRSTUVWXYZ';
    }
    if ($flags & 0x4) {
        $chars .= '0123456789';
    }

    if (($size < 1) || (strlen($chars) < 1)) {
        return '';
    }

    $String = '';
    $j = strlen( $chars ) - 1;
    while (strlen($String) < $size) {
        $String .= $chars{mt_rand(0, $j)};
    }

    return $String;
}

/**
 * Escapes special characters for use in IMAP commands.
 *
 * @param string $str the string to escape
 * @return string the escaped string
 * @since 1.0.3
 */
function quoteimap($str) {
    return str_replace(array('\\', '"'), array('\\\\', '\\"'), $str);
}

/**
 * Create compose link
 *
 * Returns a link to the compose-page, taking in consideration
 * the compose_in_new and javascript settings.
 *
 * @param string $url       The URL to the compose page
 * @param string $text      The link text, default "Compose"
 * @param string $target    URL target, if any (since 1.4.3)
 * @param string $accesskey The access key to be used, if any
 *
 * @return string a link to the compose page
 *
 * @since 1.4.2
 */
function makeComposeLink($url, $text = null, $target='', $accesskey='NONE') {
    global $compose_new_win, $compose_width, 
           $compose_height, $oTemplate;

    if(!$text) {
        $text = _("Compose");
    }

    // if not using "compose in new window", make
    // regular link and be done with it
    if($compose_new_win != '1') {
        return makeInternalLink($url, $text, $target, $accesskey);
    }

    // build the compose in new window link...


    // if javascript is on, use onclick event to handle it
    if(checkForJavascript()) {
        sqgetGlobalVar('base_uri', $base_uri, SQ_SESSION);
        $compuri = SM_BASE_URI.$url;

        return create_hyperlink('javascript:void(0)', $text, '',
                                "comp_in_new('$compuri','$compose_width','$compose_height')",
                                '', '', '',
                                ($accesskey == 'NONE'
                                ? array()
                                : array('accesskey' => $accesskey)));
    }

    // otherwise, just open new window using regular HTML
    return makeInternalLink($url, $text, '_blank', $accesskey);
}

/**
 * version of fwrite which checks for failure
 * @param resource $fp
 * @param string $string
 * @return number of written bytes. false on failure
 * @since 1.4.3
 */
function sq_fwrite($fp, $string) {
        // write to file
        $count = @fwrite($fp,$string);
        // the number of bytes written should be the length of the string
        if($count != strlen($string)) {
                return FALSE;
        }

        return $count;
}

/**
 * sq_get_html_translation_table
 *
 * Returns the translation table used by sq_htmlentities()
 *
 * @param integer $table html translation table. Possible values (without quotes):
 *             <ul>
 *                <li>HTML_ENTITIES - full html entities table defined by charset</li>
 *                <li>HTML_SPECIALCHARS - html special characters table</li>
 *             </ul>
 * @param integer $quote_style quote encoding style. Possible values (without quotes):
 *              <ul>
 *                <li>ENT_COMPAT - (default) encode double quotes</li>
 *                <li>ENT_NOQUOTES -  don't encode double or single quotes</li>
 *                <li>ENT_QUOTES - encode double and single quotes</li>
 *              </ul>
 * @param string $charset charset used for encoding. default to us-ascii, 'auto' uses $default_charset global value.
 * @return array html translation array
 * @since 1.5.1
 */
function sq_get_html_translation_table($table,$quote_style=ENT_COMPAT,$charset='us-ascii') {
  global $default_charset;

  if ($table == HTML_SPECIALCHARS) $charset='us-ascii';

  // Start array with ampersand
  $sq_html_ent_table = array( "&" => '&amp;' );

  // < and >
  $sq_html_ent_table = array_merge($sq_html_ent_table,
                        array("<" => '&lt;',
                              ">" => '&gt;')
                        );
  // double quotes
  if ($quote_style == ENT_COMPAT)
     $sq_html_ent_table = array_merge($sq_html_ent_table,
                            array("\"" => '&quot;')
                            );

  // double and single quotes
  if ($quote_style == ENT_QUOTES)
     $sq_html_ent_table = array_merge($sq_html_ent_table,
                            array("\"" => '&quot;',
                              "'" => '&#39;')
                            );

  if ($charset=='auto') $charset=$default_charset;

  // add entities that depend on charset
  switch($charset){
  case 'iso-8859-1':
    include_once(SM_PATH . 'functions/htmlentities/iso-8859-1.php');
    break;
  case 'utf-8':
    include_once(SM_PATH . 'functions/htmlentities/utf-8.php');
    break;
  case 'us-ascii':
  default:
    break;
  }
  // return table
  return $sq_html_ent_table;
}

/**
 * sq_htmlentities
 *
 * Convert all applicable characters to HTML entities.
 * Minimal php requirement - v.4.0.5.
 *
 * Function is designed for people that want to use full power of htmlentities() in
 * i18n environment.
 *
 * @param string $string string that has to be sanitized
 * @param integer $quote_style quote encoding style. Possible values (without quotes):
 *              <ul>
 *                <li>ENT_COMPAT - (default) encode double quotes</li>
 *                <li>ENT_NOQUOTES - don't encode double or single quotes</li>
 *                <li>ENT_QUOTES - encode double and single quotes</li>
 *              </ul>
 * @param string $charset charset used for encoding. defaults to 'us-ascii', 'auto' uses $default_charset global value.
 * @return string sanitized string
 * @since 1.5.1
 */
function sq_htmlentities($string,$quote_style=ENT_COMPAT,$charset='us-ascii') {
  // get translation table
  $sq_html_ent_table=sq_get_html_translation_table(HTML_ENTITIES,$quote_style,$charset);
  // convert characters
  return str_replace(array_keys($sq_html_ent_table),array_values($sq_html_ent_table),$string);
}

/**
 * Tests if string contains 8bit symbols.
 *
 * If charset is not set, function defaults to default_charset.
 * $default_charset global must be set correctly if $charset is
 * not used.
 * @param string $string tested string
 * @param string $charset charset used in a string
 * @return bool true if 8bit symbols are detected
 * @since 1.5.1 and 1.4.4
 */
function sq_is8bit($string,$charset='') {
    global $default_charset;

    if ($charset=='') $charset=$default_charset;

    /**
     * Don't use \240 in ranges. Sometimes RH 7.2 doesn't like it.
     * Don't use \200-\237 for iso-8859-x charsets. This range
     * stores control symbols in those charsets.
     * Use preg_match instead of ereg in order to avoid problems
     * with mbstring overloading
     */
    if (preg_match("/^iso-8859/i",$charset)) {
        $needle='/\240|[\241-\377]/';
    } else {
        $needle='/[\200-\237]|\240|[\241-\377]/';
    }
    return preg_match("$needle",$string);
}

/**
 * Replacement of mb_list_encodings function
 *
 * This function provides replacement for function that is available only
 * in php 5.x. Function does not test all mbstring encodings. Only the ones
 * that might be used in SM translations.
 *
 * Supported strings are stored in session in order to reduce number of
 * mb_internal_encoding function calls.
 *
 * If you want to test all mbstring encodings - fill $list_of_encodings
 * array.
 * @return array list of encodings supported by php mbstring extension
 * @since 1.5.1 and 1.4.6
 */
function sq_mb_list_encodings() {
    if (! function_exists('mb_internal_encoding'))
        return array();

    // php 5+ function
    if (function_exists('mb_list_encodings')) {
        $ret = mb_list_encodings();
        array_walk($ret,'sq_lowercase_array_vals');
        return $ret;
    }

    // don't try to test encodings, if they are already stored in session
    if (sqgetGlobalVar('mb_supported_encodings',$mb_supported_encodings,SQ_SESSION))
        return $mb_supported_encodings;

    // save original encoding
    $orig_encoding=mb_internal_encoding();

    $list_of_encoding=array(
        'pass',
        'auto',
        'ascii',
        'jis',
        'utf-8',
        'sjis',
        'euc-jp',
        'iso-8859-1',
        'iso-8859-2',
        'iso-8859-7',
        'iso-8859-9',
        'iso-8859-15',
        'koi8-r',
        'koi8-u',
        'big5',
        'gb2312',
        'gb18030',
        'windows-1251',
        'windows-1255',
        'windows-1256',
        'tis-620',
        'iso-2022-jp',
        'euc-cn',
        'euc-kr',
        'euc-tw',
        'uhc',
        'utf7-imap');

    $supported_encodings=array();

    foreach ($list_of_encoding as $encoding) {
        // try setting encodings. suppress warning messages
        if (@mb_internal_encoding($encoding))
            $supported_encodings[]=$encoding;
    }

    // restore original encoding
    mb_internal_encoding($orig_encoding);

    // register list in session
    sqsession_register($supported_encodings,'mb_supported_encodings');

    return $supported_encodings;
}

/**
 * Callback function used to lowercase array values.
 * @param string $val array value
 * @param mixed $key array key
 * @since 1.5.1 and 1.4.6
 */
function sq_lowercase_array_vals(&$val,$key) {
    $val = strtolower($val);
}


/**
 * Function returns number of characters in string.
 *
 * Returned number might be different from number of bytes in string,
 * if $charset is multibyte charset. Detection depends on mbstring
 * functions. If mbstring does not support tested multibyte charset,
 * vanilla string length function is used.
 * @param string $str string
 * @param string $charset charset
 * @since 1.5.1 and 1.4.6
 * @return integer number of characters in string
 */
function sq_strlen($str, $charset=null){
    // default option
    if (is_null($charset)) return strlen($str);

    // lowercase charset name
    $charset=strtolower($charset);

    // use automatic charset detection, if function call asks for it
    if ($charset=='auto') {
        global $default_charset, $squirrelmail_language;
        set_my_charset();
        $charset=$default_charset;
        if ($squirrelmail_language=='ja_JP') $charset='euc-jp';
    }

    // Use mbstring only with listed charsets
    $aList_of_mb_charsets=array('utf-8','big5','gb2312','gb18030','euc-jp','euc-cn','euc-tw','euc-kr');

    // calculate string length according to charset
    if (in_array($charset,$aList_of_mb_charsets) && in_array($charset,sq_mb_list_encodings())) {
        $real_length = mb_strlen($str,$charset);
    } else {
        // own strlen detection code is removed because missing strpos,
        // strtoupper and substr implementations break string wrapping.
        $real_length=strlen($str);
    }
    return $real_length;
}

/**
 * string padding with multibyte support
 *
 * @link http://www.php.net/str_pad
 * @param string $string original string
 * @param integer $width padded string width
 * @param string $pad padding symbols
 * @param integer $padtype padding type
 *  (internal php defines, see str_pad() description)
 * @param string $charset charset used in original string
 * @return string padded string
 */
function sq_str_pad($string, $width, $pad, $padtype, $charset='') {

    $charset = strtolower($charset);
    $padded_string = '';

    switch ($charset) {
    case 'utf-8':
    case 'big5':
    case 'gb2312':
    case 'euc-kr':
        /*
         * all multibyte charsets try to increase width value by
         * adding difference between number of bytes and real length
         */
        $width = $width - sq_strlen($string,$charset) + strlen($string);
    default:
        $padded_string=str_pad($string,$width,$pad,$padtype);
    }
    return $padded_string;
}

/**
 * Wrapper that is used to switch between vanilla and multibyte substr
 * functions.
 * @param string $string
 * @param integer $start
 * @param integer $length
 * @param string $charset
 * @return string
 * @since 1.5.1
 * @link http://www.php.net/substr
 * @link http://www.php.net/mb_substr
 */
function sq_substr($string,$start,$length=NULL,$charset='auto') {

    // if $length is NULL, use the full string length...
    // we have to do this to mimick the use of substr()
    // where $length is not given
    //
    if (is_null($length))
        $length = sq_strlen($length);


    // use automatic charset detection, if function call asks for it
    static $charset_auto, $bUse_mb;

    if ($charset=='auto') {
        if (!isset($charset_auto)) {
            global $default_charset, $squirrelmail_language;
            set_my_charset();
            $charset=$default_charset;
            if ($squirrelmail_language=='ja_JP') $charset='euc-jp';
            $charset_auto = $charset;
        } else {
            $charset = $charset_auto;
        }
    }
    $charset = strtolower($charset);

    // in_array call is expensive => do it once and use a static var for
    // storing the results
    if (!isset($bUse_mb)) {
        if (in_array($charset,sq_mb_list_encodings())) {
            $bUse_mb = true;
        } else {
            $bUse_mb = false;
        }
    }

    if ($bUse_mb) {
        return mb_substr($string,$start,$length,$charset);
    }
    // TODO: add mbstring independent code

    // use vanilla string functions as last option
    return substr($string,$start,$length);
}

/**
  * This is a replacement for PHP's substr_replace() that is
  * multibyte-aware.
  *
  * @param string $string      The string to operate upon
  * @param string $replacement The string to be inserted
  * @param int    $start       The offset at which to begin substring replacement
  * @param int    $length      The number of characters after $start to remove
  *                            NOTE that if you need to specify a charset but
  *                            want to achieve normal substr_replace() behavior
  *                            where $length is not specified, use NULL (OPTIONAL;
  *                            default from $start to end of string)
  * @param string $charset     The charset of the given string.  A value of NULL
  *                            here will force the use of PHP's standard substr().
  *                            (OPTIONAL; default is "auto", which indicates that
  *                            the user's current charset should be used).
  *
  * @return string The manipulated string
  *
  * Of course, you can use more advanced (e.g., negative) values
  * for $start and $length as needed - see the PHP manual for more
  * information:  http://www.php.net/manual/function.substr-replace.php
  *
  */
function sq_substr_replace($string, $replacement, $start, $length=NULL,
                           $charset='auto')
{

   // NULL charset?  Just use substr_replace()
   //
   if (is_null($charset))
      return is_null($length) ? substr_replace($string, $replacement, $start)
                              : substr_replace($string, $replacement, $start, $length);


   // use current character set?
   //
   if ($charset == 'auto')
   {
//FIXME: is there any reason why this cannot be a global flag used by all string wrapper functions?
      static $auto_charset;
      if (!isset($auto_charset))
      {
         global $default_charset;
//FIXME - do we need this?
global $squirrelmail_language;
         set_my_charset();
         $auto_charset = $default_charset;
//FIXME - do we need this?
if ($squirrelmail_language == 'ja_JP') $auto_charset = 'euc-jp';
      }
      $charset = $auto_charset;
   }


   // standardize character set name
   //
   $charset = strtolower($charset);


/* ===== FIXME: this list is not used in 1.5.x, but if we need it, unless this differs between all our string function wrappers, we should store this info in the session
   // only use mbstring with the following character sets
   //
   $sq_substr_replace_mb_charsets = array(
      'utf-8',
      'big5',
      'gb2312',
      'gb18030',
      'euc-jp',
      'euc-cn',
      'euc-tw',
      'euc-kr'
   );


   // now we can use our own implementation using
   // mb_substr() and mb_strlen() if needed
   //
   if (in_array($charset, $sq_substr_replace_mb_charsets)
    && in_array($charset, sq_mb_list_encodings()))
===== */
//FIXME: is there any reason why this cannot be a global array used by all string wrapper functions?
   if (in_array($charset, sq_mb_list_encodings()))
   {

      $string_length = mb_strlen($string, $charset);

      if ($start < 0)
         $start = max(0, $string_length + $start);

      else if ($start > $string_length)
         $start = $string_length;

      if ($length < 0)
         $length = max(0, $string_length - $start + $length);

      else if (is_null($length) || $length > $string_length)
         $length = $string_length;

      if ($start + $length > $string_length)
         $length = $string_length - $start;

      return mb_substr($string, 0, $start, $charset)
           . $replacement
           . mb_substr($string,
                       $start + $length,
                       $string_length, // FIXME: I can't see why this is needed:  - $start - $length,
                       $charset);

   }


   // else use normal substr_replace()
   //
   return is_null($length) ? substr_replace($string, $replacement, $start)
                           : substr_replace($string, $replacement, $start, $length);

}

/**
 * Wrapper that is used to switch between vanilla and multibyte strpos
 * functions.
 * @param string $haystack
 * @param mixed $needle
 * @param integer $offset
 * @param string $charset
 * @return string
 * @since 1.5.1
 * @link http://www.php.net/strpos
 * @link http://www.php.net/mb_strpos
 */
function sq_strpos($haystack,$needle,$offset,$charset='auto') {
    // use automatic charset detection, if function call asks for it
    static $charset_auto, $bUse_mb;

    if ($charset=='auto') {
        if (!isset($charset_auto)) {
            global $default_charset, $squirrelmail_language;
            set_my_charset();
            $charset=$default_charset;
            if ($squirrelmail_language=='ja_JP') $charset='euc-jp';
            $charset_auto = $charset;
        } else {
            $charset = $charset_auto;
        }
    }
    $charset = strtolower($charset);

    // in_array call is expensive => do it once and use a static var for
    // storing the results
    if (!isset($bUse_mb)) {
        if (in_array($charset,sq_mb_list_encodings())) {
            $bUse_mb = true;
        } else {
            $bUse_mb = false;
        }
    }
    if ($bUse_mb) {
        return mb_strpos($haystack,$needle,$offset,$charset);
    }
    // TODO: add mbstring independent code

    // use vanilla string functions as last option
    return strpos($haystack,$needle,$offset);
}

/**
 * Wrapper that is used to switch between vanilla and multibyte strtoupper
 * functions.
 * @param string $string
 * @param string $charset
 * @return string
 * @since 1.5.1
 * @link http://www.php.net/strtoupper
 * @link http://www.php.net/mb_strtoupper
 */
function sq_strtoupper($string,$charset='auto') {
    // use automatic charset detection, if function call asks for it
    static $charset_auto, $bUse_mb;

    if ($charset=='auto') {
        if (!isset($charset_auto)) {
            global $default_charset, $squirrelmail_language;
            set_my_charset();
            $charset=$default_charset;
            if ($squirrelmail_language=='ja_JP') $charset='euc-jp';
            $charset_auto = $charset;
        } else {
            $charset = $charset_auto;
        }
    }
    $charset = strtolower($charset);

    // in_array call is expensive => do it once and use a static var for
    // storing the results
    if (!isset($bUse_mb)) {
        if (function_exists('mb_strtoupper') &&
            in_array($charset,sq_mb_list_encodings())) {
            $bUse_mb = true;
        } else {
            $bUse_mb = false;
        }
    }

    if ($bUse_mb) {
        return mb_strtoupper($string,$charset);
    }
    // TODO: add mbstring independent code

    // use vanilla string functions as last option
    return strtoupper($string);
}

/**
 * Counts 8bit bytes in string
 * @param string $string tested string
 * @return integer number of 8bit bytes
 */
function sq_count8bit($string) {
    $count=0;
    for ($i=0; $i<strlen($string); $i++) {
        if (ord($string[$i]) > 127) $count++;
    }
    return $count;
}

/**
 * Callback function to trim whitespace from a value, to be used in array_walk
 * @param string $value value to trim
 * @since 1.5.2 and 1.4.7
 */
function sq_trim_value ( &$value ) {
    $value = trim($value);
}

/**
  * Truncates the given string so that it has at
  * most $max_chars characters.  NOTE that a "character"
  * may be a multibyte character, or (optionally), an
  * HTML entity , so this function is different than
  * using substr() or mb_substr().
  *
  * NOTE that if $elipses is given and used, the returned
  *      number of characters will be $max_chars PLUS the
  *      length of $elipses
  *
  * @param string  $string    The string to truncate
  * @param int     $max_chars The maximum allowable characters
  * @param string  $elipses   A string that will be added to
  *                           the end of the truncated string
  *                           (ONLY if it is truncated) (OPTIONAL;
  *                           default not used)
  * @param boolean $html_entities_as_chars Whether or not to keep
  *                                        HTML entities together
  *                                        (OPTIONAL; default ignore
  *                                        HTML entities)
  *
  * @return string The truncated string
  *
  * @since 1.4.20 and 1.5.2 (replaced truncateWithEntities())
  *
  */
function sm_truncate_string($string, $max_chars, $elipses='',
                            $html_entities_as_chars=FALSE)
{

   // if the length of the string is less than
   // the allowable number of characters, just
   // return it as is (even if it contains any
   // HTML entities, that would just make the
   // actual length even smaller)
   //
   $actual_strlen = sq_strlen($string, 'auto');
   if ($max_chars <= 0 || $actual_strlen <= $max_chars)
      return $string;


   // if needed, count the number of HTML entities in
   // the string up to the maximum character limit,
   // pushing that limit up for each entity found
   //
   $adjusted_max_chars = $max_chars;
   if ($html_entities_as_chars)
   {

      // $loop_count is needed to prevent an endless loop
      // which is caused by buggy mbstring versions that
      // return 0 (zero) instead of FALSE in some rare
      // cases.  Thanks, PHP.
      // see: http://bugs.php.net/bug.php?id=52731
      // also: tracker $3053349
      //
      $loop_count = 0;
      $entity_pos = $entity_end_pos = -1;
      while ($entity_end_pos + 1 < $actual_strlen
          && ($entity_pos = sq_strpos($string, '&', $entity_end_pos + 1)) !== FALSE
          && ($entity_end_pos = sq_strpos($string, ';', $entity_pos)) !== FALSE
          && $entity_pos <= $adjusted_max_chars
          && $loop_count++ < $max_chars)
      {
         $adjusted_max_chars += $entity_end_pos - $entity_pos;
      }


      // this isn't necessary because sq_substr() would figure this
      // out anyway, but we can avoid a sq_substr() call and we
      // know that we don't have to add an elipses (this is now
      // an accurate comparison, since $adjusted_max_chars, like
      // $actual_strlen, does not take into account HTML entities)
      //
      if ($actual_strlen <= $adjusted_max_chars)
         return $string;

   }


   // get the truncated string
   //
   $truncated_string = sq_substr($string, 0, $adjusted_max_chars);


   // return with added elipses
   //
   return $truncated_string . $elipses;

}

/**
  * Gathers the list of secuirty tokens currently
  * stored in the user's preferences and optionally
  * purges old ones from the list.
  *
  * @param boolean $purge_old Indicates if old tokens
  *                           should be purged from the
  *                           list ("old" is 2 days or
  *                           older unless the administrator
  *                           overrides that value using
  *                           $max_token_age_days in
  *                           config/config_local.php)
  *                           (OPTIONAL; default is to always
  *                           purge old tokens)
  *
  * @return array The list of tokens
  *
  * @since 1.4.19 and 1.5.2
  *
  */
function sm_get_user_security_tokens($purge_old=TRUE)
{

   global $data_dir, $username, $max_token_age_days;

   $tokens = getPref($data_dir, $username, 'security_tokens', '');
   if (($tokens = unserialize($tokens)) === FALSE || !is_array($tokens))
      $tokens = array();

   // purge old tokens if necessary
   //
   if ($purge_old)
   {
      if (empty($max_token_age_days)) $max_token_age_days = 2;
      $now = time();
      $discard_token_date = $now - ($max_token_age_days * 86400);
      $cleaned_tokens = array();
      foreach ($tokens as $token => $timestamp)
         if ($timestamp >= $discard_token_date)
            $cleaned_tokens[$token] = $timestamp;
      $tokens = $cleaned_tokens;
   }

   return $tokens;

}

/**
  * Generates a security token that is then stored in
  * the user's preferences with a timestamp for later
  * verification/use.
  *
  * NOTE: The administrator can force SquirrelMail to generate
  * a new token every time one is requested (which may increase
  * obscurity through token randomness at the cost of some
  * performance) by adding the following to
  * config/config_local.php:   $do_not_use_single_token = TRUE;
  * Otherwise, only one token will be generated per user which
  * will change only after it expires or is used outside of the
  * validity period specified when calling sm_validate_security_token()
  *
  * WARNING: If the administrator has turned the token system
  *          off by setting $disable_security_tokens to TRUE in
  *          config/config.php or the configuration tool, this
  *          function will not store tokens in the user
  *          preferences (but it will still generate and return
  *          a random string).
  *
  * @param boolean $force_generate_new When TRUE, a new token will
  *                                    always be created even if current
  *                                    configuration dictates otherwise
  *                                    (OPTIONAL; default FALSE)
  *
  * @return string A security token
  *
  * @since 1.4.19 and 1.5.2
  *
  */
function sm_generate_security_token($force_generate_new=FALSE)
{

   global $data_dir, $username, $disable_security_tokens, $do_not_use_single_token;
   $max_generation_tries = 1000;

   $tokens = sm_get_user_security_tokens();

   if (!$force_generate_new && !$do_not_use_single_token && !empty($tokens))
      return key($tokens);

   $new_token = GenerateRandomString(12, '', 7);
   $count = 0;
   while (isset($tokens[$new_token]))
   {
      $new_token = GenerateRandomString(12, '', 7);
      if (++$count > $max_generation_tries)
      {
         logout_error(_("Fatal token generation error; please contact your system administrator or the SquirrelMail Team"));
         exit;
      }
   }

   // is the token system enabled?  CAREFUL!
   //
   if (!$disable_security_tokens)
   {
      $tokens[$new_token] = time();
      setPref($data_dir, $username, 'security_tokens', serialize($tokens));
   }

   return $new_token;

}

/**
  * Validates a given security token and optionally remove it
  * from the user's preferences if it was valid.  If the token
  * is too old but otherwise valid, it will still be rejected.
  *
  * "Too old" is 2 days or older unless the administrator
  * overrides that value using $max_token_age_days in
  * config/config_local.php
  *
  * WARNING: If the administrator has turned the token system
  *          off by setting $disable_security_tokens to TRUE in
  *          config/config.php or the configuration tool, this
  *          function will always return TRUE.
  *
  * @param string  $token           The token to validate
  * @param int     $validity_period The number of seconds tokens are valid
  *                                 for (set to zero to remove valid tokens
  *                                 after only one use; set to -1 to allow
  *                                 indefinite re-use (but still subject to
  *                                 $max_token_age_days - see elsewhere);
  *                                 use 3600 to allow tokens to be reused for
  *                                 an hour) (OPTIONAL; default is to only
  *                                 allow tokens to be used once)
  *                                 NOTE this is unrelated to $max_token_age_days
  *                                 or rather is an additional time constraint on
  *                                 tokens that allows them to be re-used (or not)
  *                                 within a more narrow timeframe
  * @param boolean $show_error      Indicates that if the token is not
  *                                 valid, this function should display
  *                                 a generic error, log the user out
  *                                 and exit - this function will never
  *                                 return in that case.
  *                                 (OPTIONAL; default FALSE)
  *
  * @return boolean TRUE if the token validated; FALSE otherwise
  *
  * @since 1.4.19 and 1.5.2
  *
  */
function sm_validate_security_token($token, $validity_period=0, $show_error=FALSE)
{

   global $data_dir, $username, $max_token_age_days,
          $disable_security_tokens;

   // bypass token validation?  CAREFUL!
   //
   if ($disable_security_tokens) return TRUE;

   // don't purge old tokens here because we already
   // do it when generating tokens
   //
   $tokens = sm_get_user_security_tokens(FALSE);

   // token not found?
   //
   if (empty($tokens[$token]))
   {
      if (!$show_error) return FALSE;
      logout_error(_("This page request could not be verified and appears to have expired."));
      exit;
   }

   $now = time();
   $timestamp = $tokens[$token];

   // whether valid or not, we want to remove it from
   // user prefs if it's old enough (unless requested to
   // bypass this (in which case $validity_period is -1))
   //
   if ($validity_period >= 0
    && $timestamp < $now - $validity_period)
   {
      unset($tokens[$token]);
      setPref($data_dir, $username, 'security_tokens', serialize($tokens));
   }

   // reject tokens that are too old
   //
   if (empty($max_token_age_days)) $max_token_age_days = 2;
   $old_token_date = $now - ($max_token_age_days * 86400);
   if ($timestamp < $old_token_date)
   {
      if (!$show_error) return FALSE;
      logout_error(_("The current page request appears to have originated from an untrusted source."));
      exit;
   }

   // token OK!
   //
   return TRUE;

}

/**
  * Wrapper for PHP's htmlspecialchars() that
  * attempts to add the correct character encoding
  *
  * @param string $string The string to be converted
  * @param int $flags A bitmask that controls the behavior of htmlspecialchars()
  *                   (See http://php.net/manual/function.htmlspecialchars.php )
  *                   (OPTIONAL; default ENT_COMPAT)
  * @param string $encoding The character encoding to use in the conversion
  *                         (OPTIONAL; default automatic detection)
  * @param boolean $double_encode Whether or not to convert entities that are
  *                               already in the string (only supported in
  *                               PHP 5.2.3+) (OPTIONAL; default TRUE)
  *
  * @return string The converted text
  *
  */
function sm_encode_html_special_chars($string, $flags=ENT_COMPAT,
                                      $encoding=NULL, $double_encode=TRUE)
{
   if (!$encoding)
   {
      global $default_charset;
      if ($default_charset == 'iso-2022-jp')
         $default_charset = 'EUC-JP';
      $encoding = $default_charset;
   }

// TODO: Is adding this check an unnecessary performance hit?
   if (check_php_version(5, 2, 3))
      return htmlspecialchars($string, $flags, $encoding, $double_encode);

   return htmlspecialchars($string, $flags, $encoding);
}