summaryrefslogtreecommitdiff
path: root/Build/source/libs/icu-xetex/common/putil.c
blob: b371284562eb536ce6f7e0c3020e647943675556 (plain)
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
1716
1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
1756
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
1790
1791
1792
1793
1794
1795
1796
1797
1798
1799
1800
1801
1802
1803
1804
1805
1806
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
1824
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835
1836
1837
1838
1839
1840
1841
1842
1843
1844
1845
1846
1847
1848
1849
1850
1851
1852
1853
1854
1855
1856
1857
1858
1859
1860
1861
1862
1863
1864
1865
1866
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
1877
1878
1879
1880
1881
1882
1883
1884
1885
1886
1887
1888
1889
1890
1891
1892
1893
1894
1895
1896
1897
1898
1899
1900
1901
1902
1903
1904
1905
1906
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
1925
1926
1927
1928
1929
1930
1931
1932
1933
1934
1935
1936
1937
1938
1939
1940
1941
1942
1943
1944
1945
1946
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
/*
******************************************************************************
*
*   Copyright (C) 1997-2005, International Business Machines
*   Corporation and others.  All Rights Reserved.
*
******************************************************************************
*
*  FILE NAME : putil.c (previously putil.cpp and ptypes.cpp)
*
*   Date        Name        Description
*   04/14/97    aliu        Creation.
*   04/24/97    aliu        Added getDefaultDataDirectory() and
*                            getDefaultLocaleID().
*   04/28/97    aliu        Rewritten to assume Unix and apply general methods
*                            for assumed case.  Non-UNIX platforms must be
*                            special-cased.  Rewrote numeric methods dealing
*                            with NaN and Infinity to be platform independent
*                             over all IEEE 754 platforms.
*   05/13/97    aliu        Restored sign of timezone
*                            (semantics are hours West of GMT)
*   06/16/98    erm         Added IEEE_754 stuff, cleaned up isInfinite, isNan,
*                             nextDouble..
*   07/22/98    stephen     Added remainder, max, min, trunc
*   08/13/98    stephen     Added isNegativeInfinity, isPositiveInfinity
*   08/24/98    stephen     Added longBitsFromDouble
*   09/08/98    stephen     Minor changes for Mac Port
*   03/02/99    stephen     Removed openFile().  Added AS400 support.
*                            Fixed EBCDIC tables
*   04/15/99    stephen     Converted to C.
*   06/28/99    stephen     Removed mutex locking in u_isBigEndian().
*   08/04/99    jeffrey R.  Added OS/2 changes
*   11/15/99    helena      Integrated S/390 IEEE support.
*   04/26/01    Barry N.    OS/400 support for uprv_getDefaultLocaleID
*   08/15/01    Steven H.   OS/400 support for uprv_getDefaultCodepage
******************************************************************************
*/

#ifndef PTX

/* Define _XOPEN_SOURCE for Solaris and friends. */
/* NetBSD needs it to be >= 4 */
#ifndef _XOPEN_SOURCE
#define _XOPEN_SOURCE 4
#endif

/* Define __USE_POSIX and __USE_XOPEN for Linux and glibc. */
#ifndef __USE_POSIX
#define __USE_POSIX
#endif
#ifndef __USE_XOPEN
#define __USE_XOPEN
#endif

#endif /* PTX */

/* include ICU headers */
#include "unicode/utypes.h"
#include "unicode/putil.h"
#include "unicode/ustring.h"
#include "putilimp.h"
#include "uassert.h"
#include "umutex.h"
#include "cmemory.h"
#include "cstring.h"
#include "locmap.h"
#include "ucln_cmn.h"

/* Include standard headers. */
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <math.h>
#include <locale.h>
#include <float.h>
#include <time.h>

/* include system headers */
#ifdef U_WINDOWS
#   define WIN32_LEAN_AND_MEAN
#   define VC_EXTRALEAN
#   define NOUSER
#   define NOSERVICE
#   define NOIME
#   define NOMCX
#   include <windows.h>
#elif defined(U_CYGWIN) && defined(__STRICT_ANSI__)
/* tzset isn't defined in strict ANSI on Cygwin. */
#   undef __STRICT_ANSI__
#elif defined(OS400)
#   include <float.h>
#   include <qusec.h>       /* error code structure */
#   include <qusrjobi.h>
#   include <qliept.h>      /* EPT_CALL macro  - this include must be after all other "QSYSINCs" */
#elif defined(XP_MAC)
#   include <Files.h>
#   include <IntlResources.h>
#   include <Script.h>
#   include <Folders.h>
#   include <MacTypes.h>
#   include <TextUtils.h>
#   define ICU_NO_USER_DATA_OVERRIDE 1
#elif defined(OS390)
#include "unicode/ucnv.h"   /* Needed for UCNV_SWAP_LFNL_OPTION_STRING */
#elif defined(U_AIX)
#elif defined(U_SOLARIS) || defined(U_LINUX)
#elif defined(U_HPUX)
#elif defined(U_DARWIN)
#include <sys/file.h>
#include <sys/param.h>
#elif defined(U_QNX)
#include <sys/neutrino.h>
#endif

#ifndef U_WINDOWS
#include <sys/time.h> 
#endif

/*
 * Only include langinfo.h if we have a way to get the codeset. If we later
 * depend on more feature, we can test on U_HAVE_NL_LANGINFO.
 *
 */

#if U_HAVE_NL_LANGINFO_CODESET
#include <langinfo.h>
#endif

/* Define the extension for data files, again... */
#define DATA_TYPE "dat"

/* Leave this copyright notice here! */
static const char copyright[] = U_COPYRIGHT_STRING;

/* floating point implementations ------------------------------------------- */

/* We return QNAN rather than SNAN*/
#define SIGN 0x80000000U
#if defined(__GNUC__) || defined(_MSC_VER)
/*
    This is an optimization for when u_topNBytesOfDouble
    and u_bottomNBytesOfDouble can't be properly optimized by the compiler
    or when faster infinity and NaN usage is helpful.
*/
#define USE_64BIT_DOUBLE_OPTIMIZATION 1
#else
#define USE_64BIT_DOUBLE_OPTIMIZATION 0
#endif

#if USE_64BIT_DOUBLE_OPTIMIZATION
/* gcc 3.2 has an optimization bug */
static const int64_t gNan64 = INT64_C(0x7FF8000000000000);
static const int64_t gInf64 = INT64_C(0x7FF0000000000000);
static const double * const fgNan = (const double * const)(&gNan64);
static const double * const fgInf = (const double * const)(&gInf64);
#else

#if IEEE_754
#define NAN_TOP ((int16_t)0x7FF8)
#define INF_TOP ((int16_t)0x7FF0)
#elif defined(OS390)
#define NAN_TOP ((int16_t)0x7F08)
#define INF_TOP ((int16_t)0x3F00)
#endif

/* statics */
static UBool fgNaNInitialized = FALSE;
static UBool fgInfInitialized = FALSE;
static double gNan;
static double gInf;
static double * fgNan = &gNan;
static double * fgInf = &gInf;
#endif

/*---------------------------------------------------------------------------
  Platform utilities
  Our general strategy is to assume we're on a POSIX platform.  Platforms which
  are non-POSIX must declare themselves so.  The default POSIX implementation
  will sometimes work for non-POSIX platforms as well (e.g., the NaN-related
  functions).
  ---------------------------------------------------------------------------*/

#if defined(U_WINDOWS) || defined(XP_MAC) || defined(OS400)
#   undef U_POSIX_LOCALE
#else
#   define U_POSIX_LOCALE    1
#endif

/* Utilities to get the bits from a double */
#if !USE_64BIT_DOUBLE_OPTIMIZATION
static char*
u_topNBytesOfDouble(double* d, int n)
{
#if U_IS_BIG_ENDIAN
    return (char*)d;
#else
    return (char*)(d + 1) - n;
#endif
}
#endif

static char*
u_bottomNBytesOfDouble(double* d, int n)
{
#if U_IS_BIG_ENDIAN
    return (char*)(d + 1) - n;
#else
    return (char*)d;
#endif
}

#if defined(U_WINDOWS)
typedef union {
    int64_t int64;
    FILETIME fileTime;
} FileTimeConversion;   /* This is like a ULARGE_INTEGER */

/* Number of 100 nanoseconds from 1/1/1601 to 1/1/1970 */
#define EPOCH_BIAS  INT64_C(116444736000000000)
#define HECTONANOSECOND_PER_MILLISECOND   10000

#endif

/*---------------------------------------------------------------------------
  Universal Implementations
  These are designed to work on all platforms.  Try these, and if they
  don't work on your platform, then special case your platform with new
  implementations.
---------------------------------------------------------------------------*/

/* Return UTC (GMT) time measured in milliseconds since 0:00 on 1/1/70.*/
U_CAPI UDate U_EXPORT2
uprv_getUTCtime()
{
#ifdef XP_MAC
    time_t t, t1, t2;
    struct tm tmrec;

    uprv_memset( &tmrec, 0, sizeof(tmrec) );
    tmrec.tm_year = 70;
    tmrec.tm_mon = 0;
    tmrec.tm_mday = 1;
    t1 = mktime(&tmrec);    /* seconds of 1/1/1970*/

    time(&t);
    uprv_memcpy( &tmrec, gmtime(&t), sizeof(tmrec) );
    t2 = mktime(&tmrec);    /* seconds of current GMT*/
    return (UDate)(t2 - t1) * U_MILLIS_PER_SECOND;         /* GMT (or UTC) in seconds since 1970*/
#elif defined(U_WINDOWS)

    FileTimeConversion winTime;
    GetSystemTimeAsFileTime(&winTime.fileTime);
    return (UDate)((winTime.int64 - EPOCH_BIAS) / HECTONANOSECOND_PER_MILLISECOND);
#else
/*
    struct timeval posixTime;
    gettimeofday(&posixTime, NULL);
    return (UDate)(((int64_t)posixTime.tv_sec * U_MILLIS_PER_SECOND) + (posixTime.tv_usec/1000));
*/
    time_t epochtime;
    time(&epochtime);
    return (UDate)epochtime * U_MILLIS_PER_SECOND;
#endif
}

/*-----------------------------------------------------------------------------
  IEEE 754
  These methods detect and return NaN and infinity values for doubles
  conforming to IEEE 754.  Platforms which support this standard include X86,
  Mac 680x0, Mac PowerPC, AIX RS/6000, and most others.
  If this doesn't work on your platform, you have non-IEEE floating-point, and
  will need to code your own versions.  A naive implementation is to return 0.0
  for getNaN and getInfinity, and false for isNaN and isInfinite.
  ---------------------------------------------------------------------------*/

U_CAPI UBool U_EXPORT2
uprv_isNaN(double number)
{
#if IEEE_754
#if USE_64BIT_DOUBLE_OPTIMIZATION
    /* gcc 3.2 has an optimization bug */
    /* Infinity is 0x7FF0000000000000U. Anything greater than that is a NaN */
    return (UBool)(((*((int64_t *)&number)) & U_INT64_MAX) > gInf64);

#else
    /* This should work in theory, but it doesn't, so we resort to the more*/
    /* complicated method below.*/
    /*  return number != number;*/

    /* You can't return number == getNaN() because, by definition, NaN != x for*/
    /* all x, including NaN (that is, NaN != NaN).  So instead, we compare*/
    /* against the known bit pattern.  We must be careful of endianism here.*/
    /* The pattern we are looking for id:*/

    /*   7FFy yyyy yyyy yyyy  (some y non-zero)*/

    /* There are two different kinds of NaN, but we ignore the distinction*/
    /* here.  Note that the y value must be non-zero; if it is zero, then we*/
    /* have infinity.*/

    uint32_t highBits = *(uint32_t*)u_topNBytesOfDouble(&number,
                              sizeof(uint32_t));
    uint32_t lowBits  = *(uint32_t*)u_bottomNBytesOfDouble(&number,
                             sizeof(uint32_t));

    return (UBool)(((highBits & 0x7FF00000L) == 0x7FF00000L) &&
      (((highBits & 0x000FFFFFL) != 0) || (lowBits != 0)));
#endif

#elif defined(OS390)
    uint32_t highBits = *(uint32_t*)u_topNBytesOfDouble(&number,
                        sizeof(uint32_t));
    uint32_t lowBits  = *(uint32_t*)u_bottomNBytesOfDouble(&number,
                        sizeof(uint32_t));

    return ((highBits & 0x7F080000L) == 0x7F080000L) &&
      (lowBits == 0x00000000L);

#else
    /* If your platform doesn't support IEEE 754 but *does* have an NaN value,*/
    /* you'll need to replace this default implementation with what's correct*/
    /* for your platform.*/
    return number != number;
#endif
}

U_CAPI UBool U_EXPORT2
uprv_isInfinite(double number)
{
#if IEEE_754
#if USE_64BIT_DOUBLE_OPTIMIZATION
    /* gcc 3.2 has an optimization bug */
    return (UBool)(((*((int64_t *)&number)) & U_INT64_MAX) == gInf64);
#else

    /* We know the top bit is the sign bit, so we mask that off in a copy of */
    /* the number and compare against infinity. [LIU]*/
    /* The following approach doesn't work for some reason, so we go ahead and */
    /* scrutinize the pattern itself. */
    /*  double a = number; */
    /*  *(int8_t*)u_topNBytesOfDouble(&a, 1) &= 0x7F;*/
    /*  return a == uprv_getInfinity();*/
    /* Instead, We want to see either:*/

    /*   7FF0 0000 0000 0000*/
    /*   FFF0 0000 0000 0000*/

    uint32_t highBits = *(uint32_t*)u_topNBytesOfDouble(&number,
                        sizeof(uint32_t));
    uint32_t lowBits  = *(uint32_t*)u_bottomNBytesOfDouble(&number,
                        sizeof(uint32_t));

    return (UBool)(((highBits  & ~SIGN) == 0x7FF00000U) &&
      (lowBits == 0x00000000U));
#endif

#elif defined(OS390)
    uint32_t highBits = *(uint32_t*)u_topNBytesOfDouble(&number,
                        sizeof(uint32_t));
    uint32_t lowBits  = *(uint32_t*)u_bottomNBytesOfDouble(&number,
                        sizeof(uint32_t));

    return ((highBits  & ~SIGN) == 0x70FF0000L) && (lowBits == 0x00000000L);

#else
    /* If your platform doesn't support IEEE 754 but *does* have an infinity*/
    /* value, you'll need to replace this default implementation with what's*/
    /* correct for your platform.*/
    return number == (2.0 * number);
#endif
}

U_CAPI UBool U_EXPORT2
uprv_isPositiveInfinity(double number)
{
#if IEEE_754 || defined(OS390)
    return (UBool)(number > 0 && uprv_isInfinite(number));
#else
    return uprv_isInfinite(number);
#endif
}

U_CAPI UBool U_EXPORT2
uprv_isNegativeInfinity(double number)
{
#if IEEE_754 || defined(OS390)
    return (UBool)(number < 0 && uprv_isInfinite(number));

#else
    uint32_t highBits = *(uint32_t*)u_topNBytesOfDouble(&number,
                        sizeof(uint32_t));
    return((highBits & SIGN) && uprv_isInfinite(number));

#endif
}

U_CAPI double U_EXPORT2
uprv_getNaN()
{
#if IEEE_754 || defined(OS390)
#if !USE_64BIT_DOUBLE_OPTIMIZATION
    if (!fgNaNInitialized) {
        /* This variable is always initialized with the same value,
        so a mutex isn't needed. */
        int i;
        int8_t* p = (int8_t*)fgNan;
        for(i = 0; i < sizeof(double); ++i)
            *p++ = 0;
        *(int16_t*)u_topNBytesOfDouble(fgNan, sizeof(NAN_TOP)) = NAN_TOP;
        fgNaNInitialized = TRUE;
    }
#endif
    return *fgNan;
#else
    /* If your platform doesn't support IEEE 754 but *does* have an NaN value,*/
    /* you'll need to replace this default implementation with what's correct*/
    /* for your platform.*/
    return 0.0;
#endif
}

U_CAPI double U_EXPORT2
uprv_getInfinity()
{
#if IEEE_754 || defined(OS390)
#if !USE_64BIT_DOUBLE_OPTIMIZATION
    if (!fgInfInitialized)
    {
        /* This variable is always initialized with the same value,
        so a mutex isn't needed. */
        int i;
        int8_t* p = (int8_t*)fgInf;
        for(i = 0; i < sizeof(double); ++i)
            *p++ = 0;
        *(int16_t*)u_topNBytesOfDouble(fgInf, sizeof(INF_TOP)) = INF_TOP;
        fgInfInitialized = TRUE;
    }
#endif
    return *fgInf;
#else
    /* If your platform doesn't support IEEE 754 but *does* have an infinity*/
    /* value, you'll need to replace this default implementation with what's*/
    /* correct for your platform.*/
    return 0.0;
#endif
}

U_CAPI double U_EXPORT2
uprv_floor(double x)
{
    return floor(x);
}

U_CAPI double U_EXPORT2
uprv_ceil(double x)
{
    return ceil(x);
}

U_CAPI double U_EXPORT2
uprv_round(double x)
{
    return uprv_floor(x + 0.5);
}

U_CAPI double U_EXPORT2
uprv_fabs(double x)
{
    return fabs(x);
}

U_CAPI double U_EXPORT2
uprv_modf(double x, double* y)
{
    return modf(x, y);
}

U_CAPI double U_EXPORT2
uprv_fmod(double x, double y)
{
    return fmod(x, y);
}

U_CAPI double U_EXPORT2
uprv_pow(double x, double y)
{
    /* This is declared as "double pow(double x, double y)" */
    return pow(x, y);
}

U_CAPI double U_EXPORT2
uprv_pow10(int32_t x)
{
    return pow(10.0, (double)x);
}

U_CAPI double U_EXPORT2
uprv_fmax(double x, double y)
{
#if IEEE_754
    int32_t lowBits;

    /* first handle NaN*/
    if(uprv_isNaN(x) || uprv_isNaN(y))
        return uprv_getNaN();

    /* check for -0 and 0*/
    lowBits = *(uint32_t*) u_bottomNBytesOfDouble(&x, sizeof(uint32_t));
    if(x == 0.0 && y == 0.0 && (lowBits & SIGN))
        return y;

#endif

    /* this should work for all flt point w/o NaN and Infpecial cases */
    return (x > y ? x : y);
}

U_CAPI int32_t U_EXPORT2
uprv_max(int32_t x, int32_t y)
{
    return (x > y ? x : y);
}

U_CAPI double U_EXPORT2
uprv_fmin(double x, double y)
{
#if IEEE_754
    int32_t lowBits;

    /* first handle NaN*/
    if(uprv_isNaN(x) || uprv_isNaN(y))
        return uprv_getNaN();

    /* check for -0 and 0*/
    lowBits = *(uint32_t*) u_bottomNBytesOfDouble(&y, sizeof(uint32_t));
    if(x == 0.0 && y == 0.0 && (lowBits & SIGN))
        return y;

#endif

    /* this should work for all flt point w/o NaN and Inf special cases */
    return (x > y ? y : x);
}

U_CAPI int32_t U_EXPORT2
uprv_min(int32_t x, int32_t y)
{
    return (x > y ? y : x);
}

/**
 * Truncates the given double.
 * trunc(3.3) = 3.0, trunc (-3.3) = -3.0
 * This is different than calling floor() or ceil():
 * floor(3.3) = 3, floor(-3.3) = -4
 * ceil(3.3) = 4, ceil(-3.3) = -3
 */
U_CAPI double U_EXPORT2
uprv_trunc(double d)
{
#if IEEE_754
    int32_t lowBits;

    /* handle error cases*/
    if(uprv_isNaN(d))
        return uprv_getNaN();
    if(uprv_isInfinite(d))
        return uprv_getInfinity();

    lowBits = *(uint32_t*) u_bottomNBytesOfDouble(&d, sizeof(uint32_t));
    if( (d == 0.0 && (lowBits & SIGN)) || d < 0)
        return ceil(d);
    else
        return floor(d);

#else
    return d >= 0 ? floor(d) : ceil(d);

#endif
}

/**
 * Return the largest positive number that can be represented by an integer
 * type of arbitrary bit length.
 */
U_CAPI double U_EXPORT2
uprv_maxMantissa(void)
{
    return pow(2.0, DBL_MANT_DIG + 1.0) - 1.0;
}

/**
 * Return the floor of the log base 10 of a given double.
 * This method compensates for inaccuracies which arise naturally when
 * computing logs, and always give the correct value.  The parameter
 * must be positive and finite.
 * (Thanks to Alan Liu for supplying this function.)
 */
U_CAPI int16_t U_EXPORT2
uprv_log10(double d)
{
#ifdef OS400
    /* We don't use the normal implementation because you can't underflow */
    /* a double otherwise an underflow exception occurs */
    return log10(d);
#else
    /* The reason this routine is needed is that simply taking the*/
    /* log and dividing by log10 yields a result which may be off*/
    /* by 1 due to rounding errors.  For example, the naive log10*/
    /* of 1.0e300 taken this way is 299, rather than 300.*/
    double alog10 = log(d) / log(10.0);
    int16_t ailog10 = (int16_t) floor(alog10);

    /* Positive logs could be too small, e.g. 0.99 instead of 1.0*/
    if (alog10 > 0 && d >= pow(10.0, (double)(ailog10 + 1)))
        ++ailog10;

    /* Negative logs could be too big, e.g. -0.99 instead of -1.0*/
    else if (alog10 < 0 && d < pow(10.0, (double)(ailog10)))
        --ailog10;

    return ailog10;
#endif
}

U_CAPI double U_EXPORT2
uprv_log(double d)
{
    return log(d);
}

#if 0
/* This isn't used. If it's readded, readd putiltst.c tests */
U_CAPI int32_t U_EXPORT2
uprv_digitsAfterDecimal(double x)
{
    char buffer[20];
    int32_t numDigits, bytesWritten;
    char *p = buffer;
    int32_t ptPos, exponent;

    /* cheat and use the string-format routine to get a string representation*/
    /* (it handles mathematical inaccuracy better than we can), then find out */
    /* many characters are to the right of the decimal point */
    bytesWritten = sprintf(buffer, "%+.9g", x);
    while (isdigit(*(++p))) {
    }

    ptPos = (int32_t)(p - buffer);
    numDigits = (int32_t)(bytesWritten - ptPos - 1);

    /* if the number's string representation is in scientific notation, find */
    /* the exponent and take it into account*/
    exponent = 0;
    p = uprv_strchr(buffer, 'e');
    if (p != 0) {
        int16_t expPos = (int16_t)(p - buffer);
        numDigits -= bytesWritten - expPos;
        exponent = (int32_t)(atol(p + 1));
    }

    /* the string representation may still have spurious decimal digits in it, */
    /* so we cut off at the ninth digit to the right of the decimal, and have */
    /* to search backward from there to the first non-zero digit*/
    if (numDigits > 9) {
        numDigits = 9;
        while (numDigits > 0 && buffer[ptPos + numDigits] == '0')
            --numDigits;
    }
    numDigits -= exponent;
    if (numDigits < 0) {
        return 0;
    }
    return numDigits;
}
#endif

/*---------------------------------------------------------------------------
  Platform-specific Implementations
  Try these, and if they don't work on your platform, then special case your
  platform with new implementations.
  ---------------------------------------------------------------------------*/

/* Win32 time zone detection ------------------------------------------------ */

#ifdef U_WINDOWS

/*
  This code attempts to detect the Windows time zone, as set in the
  Windows Date and Time control panel.  It attempts to work on
  multiple flavors of Windows (9x, Me, NT, 2000, XP) and on localized
  installs.  It works by directly interrogating the registry and
  comparing the data there with the data returned by the
  GetTimeZoneInformation API, along with some other strategies.  The
  registry contains time zone data under one of two keys (depending on
  the flavor of Windows):

    HKLM\SOFTWARE\Microsoft\Windows\CurrentVersion\Time Zones\
    HKLM\SOFTWARE\Microsoft\Windows NT\CurrentVersion\Time Zones\

  Under this key are several subkeys, one for each time zone.  These
  subkeys are named "Pacific" on Win9x/Me and "Pacific Standard Time"
  on WinNT/2k/XP.  There are some other wrinkles; see the code for
  details.  The subkey name is NOT LOCALIZED, allowing us to support
  localized installs.

  Under the subkey are data values.  We care about:

    Std   Standard time display name, localized
    TZI   Binary block of data

  The TZI data is of particular interest.  It contains the offset, two
  more offsets for standard and daylight time, and the start and end
  rules.  This is the same data returned by the GetTimeZoneInformation
  API.  The API may modify the data on the way out, so we have to be
  careful, but essentially we do a binary comparison against the TZI
  blocks of various registry keys.  When we find a match, we know what
  time zone Windows is set to.  Since the registry key is not
  localized, we can then translate the key through a simple table
  lookup into the corresponding ICU time zone.

  This strategy doesn't always work because there are zones which
  share an offset and rules, so more than one TZI block will match.
  For example, both Tokyo and Seoul are at GMT+9 with no DST rules;
  their TZI blocks are identical.  For these cases, we fall back to a
  name lookup.  We attempt to match the display name as stored in the
  registry for the current zone to the display name stored in the
  registry for various Windows zones.  By comparing the registry data
  directly we avoid conversion complications.

  Author: Alan Liu
  Since: ICU 2.6
  Based on original code by Carl Brown <cbrown@xnetinc.com>
*/

/**
 * Layout of the binary registry data under the "TZI" key.
 */
typedef struct {
   LONG       Bias;
   LONG       StandardBias;
   LONG       DaylightBias; /* Tweaked by GetTimeZoneInformation */
   SYSTEMTIME StandardDate;
   SYSTEMTIME DaylightDate;
} TZI;

typedef struct {
    const char* icuid;
    const char* winid;
} WindowsICUMap;

/**
 * Mapping between Windows zone IDs and ICU zone IDs.  This list has
 * been mechanically checked; all zone offsets match (most important)
 * and city names match the display city names (where possible).  The
 * presence or absence of DST differs in some cases, but this is
 * acceptable as long as the zone is semantically the same (which has
 * been manually checked).
 *
 * Windows 9x/Me zone IDs are listed as "Pacific" rather than "Pacific
 * Standard Time", which is seen in NT/2k/XP.  This is fixed-up at
 * runtime as needed.  The one exception is "Mexico Standard Time 2",
 * which is not present on Windows 9x/Me.
 *
 * Zones that are not unique under Offset+Rules should be grouped
 * together for efficiency (see code below).  In addition, rules MUST
 * be grouped so that all zones of a single offset are together.
 *
 * Comments list S(tandard) or D(aylight), as declared by Windows,
 * followed by the display name (data from Windows XP).
 *
 * NOTE: Etc/GMT+12 is CORRECT for offset GMT-12:00.  Consult
 * documentation elsewhere for an explanation.
 */
static const WindowsICUMap ZONE_MAP[] = {
    "Etc/GMT+12",           "Dateline", /* S (GMT-12:00) International Date Line West */

    "Pacific/Apia",         "Samoa", /* S (GMT-11:00) Midway Island, Samoa */

    "Pacific/Honolulu",     "Hawaiian", /* S (GMT-10:00) Hawaii */

    "America/Anchorage",    "Alaskan", /* D (GMT-09:00) Alaska */

    "America/Los_Angeles",  "Pacific", /* D (GMT-08:00) Pacific Time (US & Canada); Tijuana */

    "America/Phoenix",      "US Mountain", /* S (GMT-07:00) Arizona */
    "America/Denver",       "Mountain", /* D (GMT-07:00) Mountain Time (US & Canada) */
    "America/Chihuahua",    "Mexico Standard Time 2", /* D (GMT-07:00) Chihuahua, La Paz, Mazatlan */

    "America/Managua",      "Central America", /* S (GMT-06:00) Central America */
    "America/Regina",       "Canada Central", /* S (GMT-06:00) Saskatchewan */
    "America/Mexico_City",  "Mexico", /* D (GMT-06:00) Guadalajara, Mexico City, Monterrey */
    "America/Chicago",      "Central", /* D (GMT-06:00) Central Time (US & Canada) */

    "America/Indianapolis", "US Eastern", /* S (GMT-05:00) Indiana (East) */
    "America/Bogota",       "SA Pacific", /* S (GMT-05:00) Bogota, Lima, Quito */
    "America/New_York",     "Eastern", /* D (GMT-05:00) Eastern Time (US & Canada) */

    "America/Caracas",      "SA Western", /* S (GMT-04:00) Caracas, La Paz */
    "America/Santiago",     "Pacific SA", /* D (GMT-04:00) Santiago */
    "America/Halifax",      "Atlantic", /* D (GMT-04:00) Atlantic Time (Canada) */

    "America/St_Johns",     "Newfoundland", /* D (GMT-03:30) Newfoundland */

    "America/Buenos_Aires", "SA Eastern", /* S (GMT-03:00) Buenos Aires, Georgetown */
    "America/Godthab",      "Greenland", /* D (GMT-03:00) Greenland */
    "America/Sao_Paulo",    "E. South America", /* D (GMT-03:00) Brasilia */

    "America/Noronha",      "Mid-Atlantic", /* D (GMT-02:00) Mid-Atlantic */

    "Atlantic/Cape_Verde",  "Cape Verde", /* S (GMT-01:00) Cape Verde Is. */
    "Atlantic/Azores",      "Azores", /* D (GMT-01:00) Azores */

    "Africa/Casablanca",    "Greenwich", /* S (GMT) Casablanca, Monrovia */
    "Europe/London",        "GMT", /* D (GMT) Greenwich Mean Time : Dublin, Edinburgh, Lisbon, London */

    "Africa/Lagos",         "W. Central Africa", /* S (GMT+01:00) West Central Africa */
    "Europe/Berlin",        "W. Europe", /* D (GMT+01:00) Amsterdam, Berlin, Bern, Rome, Stockholm, Vienna */
    "Europe/Paris",         "Romance", /* D (GMT+01:00) Brussels, Copenhagen, Madrid, Paris */
    "Europe/Sarajevo",      "Central European", /* D (GMT+01:00) Sarajevo, Skopje, Warsaw, Zagreb */
    "Europe/Belgrade",      "Central Europe", /* D (GMT+01:00) Belgrade, Bratislava, Budapest, Ljubljana, Prague */

    "Africa/Johannesburg",  "South Africa", /* S (GMT+02:00) Harare, Pretoria */
    "Asia/Jerusalem",       "Israel", /* S (GMT+02:00) Jerusalem */
    "Europe/Istanbul",      "GTB", /* D (GMT+02:00) Athens, Istanbul, Minsk */
    "Europe/Helsinki",      "FLE", /* D (GMT+02:00) Helsinki, Kyiv, Riga, Sofia, Tallinn, Vilnius */
    "Africa/Cairo",         "Egypt", /* D (GMT+02:00) Cairo */
    "Europe/Bucharest",     "E. Europe", /* D (GMT+02:00) Bucharest */

    "Africa/Nairobi",       "E. Africa", /* S (GMT+03:00) Nairobi */
    "Asia/Riyadh",          "Arab", /* S (GMT+03:00) Kuwait, Riyadh */
    "Europe/Moscow",        "Russian", /* D (GMT+03:00) Moscow, St. Petersburg, Volgograd */
    "Asia/Baghdad",         "Arabic", /* D (GMT+03:00) Baghdad */

    "Asia/Tehran",          "Iran", /* D (GMT+03:30) Tehran */

    "Asia/Muscat",          "Arabian", /* S (GMT+04:00) Abu Dhabi, Muscat */
    "Asia/Tbilisi",         "Caucasus", /* D (GMT+04:00) Baku, Tbilisi, Yerevan */

    "Asia/Kabul",           "Afghanistan", /* S (GMT+04:30) Kabul */

    "Asia/Karachi",         "West Asia", /* S (GMT+05:00) Islamabad, Karachi, Tashkent */
    "Asia/Yekaterinburg",   "Ekaterinburg", /* D (GMT+05:00) Ekaterinburg */

    "Asia/Calcutta",        "India", /* S (GMT+05:30) Chennai, Kolkata, Mumbai, New Delhi */

    "Asia/Katmandu",        "Nepal", /* S (GMT+05:45) Kathmandu */

    "Asia/Colombo",         "Sri Lanka", /* S (GMT+06:00) Sri Jayawardenepura */
    "Asia/Dhaka",           "Central Asia", /* S (GMT+06:00) Astana, Dhaka */
    "Asia/Novosibirsk",     "N. Central Asia", /* D (GMT+06:00) Almaty, Novosibirsk */

    "Asia/Rangoon",         "Myanmar", /* S (GMT+06:30) Rangoon */

    "Asia/Bangkok",         "SE Asia", /* S (GMT+07:00) Bangkok, Hanoi, Jakarta */
    "Asia/Krasnoyarsk",     "North Asia", /* D (GMT+07:00) Krasnoyarsk */

    "Australia/Perth",      "W. Australia", /* S (GMT+08:00) Perth */
    "Asia/Taipei",          "Taipei", /* S (GMT+08:00) Taipei */
    "Asia/Singapore",       "Singapore", /* S (GMT+08:00) Kuala Lumpur, Singapore */
    "Asia/Hong_Kong",       "China", /* S (GMT+08:00) Beijing, Chongqing, Hong Kong, Urumqi */
    "Asia/Irkutsk",         "North Asia East", /* D (GMT+08:00) Irkutsk, Ulaan Bataar */

    "Asia/Tokyo",           "Tokyo", /* S (GMT+09:00) Osaka, Sapporo, Tokyo */
    "Asia/Seoul",           "Korea", /* S (GMT+09:00) Seoul */
    "Asia/Yakutsk",         "Yakutsk", /* D (GMT+09:00) Yakutsk */

    "Australia/Darwin",     "AUS Central", /* S (GMT+09:30) Darwin */
    "Australia/Adelaide",   "Cen. Australia", /* D (GMT+09:30) Adelaide */

    "Pacific/Guam",         "West Pacific", /* S (GMT+10:00) Guam, Port Moresby */
    "Australia/Brisbane",   "E. Australia", /* S (GMT+10:00) Brisbane */
    "Asia/Vladivostok",     "Vladivostok", /* D (GMT+10:00) Vladivostok */
    "Australia/Hobart",     "Tasmania", /* D (GMT+10:00) Hobart */
    "Australia/Sydney",     "AUS Eastern", /* D (GMT+10:00) Canberra, Melbourne, Sydney */

    "Asia/Magadan",         "Central Pacific", /* S (GMT+11:00) Magadan, Solomon Is., New Caledonia */

    "Pacific/Fiji",         "Fiji", /* S (GMT+12:00) Fiji, Kamchatka, Marshall Is. */
    "Pacific/Auckland",     "New Zealand", /* D (GMT+12:00) Auckland, Wellington */

    "Pacific/Tongatapu",    "Tonga", /* S (GMT+13:00) Nuku'alofa */
    NULL,                   NULL
};

typedef struct {
    const char* winid;
    const char* altwinid;
} WindowsZoneRemap;

/**
 * If a lookup fails, we attempt to remap certain Windows ids to
 * alternate Windows ids.  If the alternate listed here begins with
 * '-', we use it as is (without the '-').  If it begins with '+', we
 * append a " Standard Time" if appropriate.
 */
static const WindowsZoneRemap ZONE_REMAP[] = {
    "Central European",     "-Warsaw",
    "Central Europe",       "-Prague Bratislava",
    "China",                "-Beijing",
                                               
    "Greenwich",            "+GMT",
    "GTB",                  "+GFT",
    "Arab",                 "+Saudi Arabia",
    "SE Asia",              "+Bangkok",
    "AUS Eastern",          "+Sydney",
    NULL,                   NULL,
};

/**
 * Various registry keys and key fragments.
 */
static const char CURRENT_ZONE_REGKEY[] = "SYSTEM\\CurrentControlSet\\Control\\TimeZoneInformation\\";
static const char STANDARD_NAME_REGKEY[] = "StandardName";
static const char STANDARD_TIME_REGKEY[] = " Standard Time";
static const char TZI_REGKEY[] = "TZI";
static const char STD_REGKEY[] = "Std";

/**
 * HKLM subkeys used to probe for the flavor of Windows.  Note that we
 * specifically check for the "GMT" zone subkey; this is present on
 * NT, but on XP has become "GMT Standard Time".  We need to
 * discriminate between these cases.
 */
static const char* const WIN_TYPE_PROBE_REGKEY[] = {
    /* WIN_9X_ME_TYPE */
    "SOFTWARE\\Microsoft\\Windows\\CurrentVersion\\Time Zones",

    /* WIN_NT_TYPE */
    "SOFTWARE\\Microsoft\\Windows NT\\CurrentVersion\\Time Zones\\GMT"

    /* otherwise: WIN_2K_XP_TYPE */
};

/**
 * The time zone root subkeys (under HKLM) for different flavors of
 * Windows.
 */
static const char* const TZ_REGKEY[] = {
    /* WIN_9X_ME_TYPE */
    "SOFTWARE\\Microsoft\\Windows\\CurrentVersion\\Time Zones\\",

    /* WIN_NT_TYPE | WIN_2K_XP_TYPE */
    "SOFTWARE\\Microsoft\\Windows NT\\CurrentVersion\\Time Zones\\"
};

/**
 * Flavor of Windows, from our perspective.  Not a real OS version,
 * but rather the flavor of the layout of the time zone information in
 * the registry.
 */
enum {
    WIN_9X_ME_TYPE = 0,
    WIN_NT_TYPE = 1,
    WIN_2K_XP_TYPE = 2
};

/**
 * Auxiliary Windows time zone function.  Attempts to open the given
 * Windows time zone ID as a registry key.  Returns ERROR_SUCCESS if
 * successful.  Caller must close the registry key.  Handles
 * variations in the resource layout in different flavors of Windows.
 *
 * @param hkey output parameter to receive opened registry key
 * @param winid Windows zone ID, e.g., "Pacific", without the
 * " Standard Time" suffix (if any).  Special case "Mexico Standard Time 2"
 * allowed.
 * @param winType Windows flavor (WIN_9X_ME_TYPE, etc.)
 * @return ERROR_SUCCESS upon success
 */
static LONG openTZRegKey(HKEY *hkey, const char* winid, int winType) {
    LONG result;
    char subKeyName[96];
    char* name;
    int i;

    uprv_strcpy(subKeyName, TZ_REGKEY[(winType == WIN_9X_ME_TYPE) ? 0 : 1]);
    name = &subKeyName[strlen(subKeyName)];
    uprv_strcat(subKeyName, winid);
    if (winType != WIN_9X_ME_TYPE) {
        /* Don't modify "Mexico Standard Time 2", which does not occur
           on WIN_9X_ME_TYPE.  Also, if the type is WIN_NT_TYPE, then
           in practice this means the GMT key is not followed by
           " Standard Time", so don't append in that case. */
        int isMexico2 = (winid[uprv_strlen(winid)- 1] == '2');
        if (!isMexico2 &&
            !(winType == WIN_NT_TYPE && uprv_strcmp(winid, "GMT") == 0)) {
            uprv_strcat(subKeyName, STANDARD_TIME_REGKEY);
        }
    }
    result = RegOpenKeyEx(HKEY_LOCAL_MACHINE,
                          subKeyName,
                          0,
                          KEY_QUERY_VALUE,
                          hkey);

    if (result != ERROR_SUCCESS) {
        /* If the primary lookup fails, try to remap the Windows zone
           ID, according to the remapping table. */
        for (i=0; ZONE_REMAP[i].winid; ++i) {
            if (uprv_strcmp(winid, ZONE_REMAP[i].winid) == 0) {
                uprv_strcpy(name, ZONE_REMAP[i].altwinid + 1);
                if (*(ZONE_REMAP[i].altwinid) == '+' &&
                    winType != WIN_9X_ME_TYPE) {
                    uprv_strcat(subKeyName, STANDARD_TIME_REGKEY);                
                }
                result = RegOpenKeyEx(HKEY_LOCAL_MACHINE,
                                      subKeyName,
                                      0,
                                      KEY_QUERY_VALUE,
                                      hkey);
                break;
            }
        }
    }

    return result;
}

/**
 * Main Windows time zone detection function.  Returns the Windows
 * time zone, translated to an ICU time zone, or NULL upon failure.
 */
static const char* detectWindowsTimeZone() {
    int winType;
    LONG result;
    HKEY hkey;
    TZI tziKey;
    TZI tziReg;
    DWORD cbData = sizeof(TZI);
    TIME_ZONE_INFORMATION apiTZI;
    char stdName[32];
    DWORD stdNameSize;
    char stdRegName[64];
    DWORD stdRegNameSize;
    int firstMatch, lastMatch;
    int j;

    /* Detect the version of windows by trying to open a sequence of
       probe keys.  We don't use the OS version API because what we
       really want to know is how the registry is laid out.
       Specifically, is it 9x/Me or not, and is it "GMT" or "GMT
       Standard Time". */
    for (winType=0; winType<2; ++winType) {
        result = RegOpenKeyEx(HKEY_LOCAL_MACHINE,
                              WIN_TYPE_PROBE_REGKEY[winType],
                              0,
                              KEY_QUERY_VALUE,
                              &hkey);
        RegCloseKey(hkey);
        if (result == ERROR_SUCCESS) {
            break;
        }
    }

    /* Obtain TIME_ZONE_INFORMATION from the API, and then convert it
       to TZI.  We could also interrogate the registry directly; we do
       this below if needed. */
    uprv_memset(&apiTZI, 0, sizeof(apiTZI));
    GetTimeZoneInformation(&apiTZI);
    tziKey.Bias = apiTZI.Bias;
    uprv_memcpy((char *)&tziKey.StandardDate, (char*)&apiTZI.StandardDate,
           sizeof(apiTZI.StandardDate));
    uprv_memcpy((char *)&tziKey.DaylightDate, (char*)&apiTZI.DaylightDate,
           sizeof(apiTZI.DaylightDate));

    /* For each zone that can be identified by Offset+Rules, see if we
       have a match.  Continue scanning after finding a match,
       recording the index of the first and the last match.  We have
       to do this because some zones are not unique under
       Offset+Rules. */
    firstMatch = lastMatch = -1;
    for (j=0; ZONE_MAP[j].icuid; j++) {
        result = openTZRegKey(&hkey, ZONE_MAP[j].winid, winType);
        if (result == ERROR_SUCCESS) {
            result = RegQueryValueEx(hkey,
                                     TZI_REGKEY,
                                     NULL,
                                     NULL,
                                     (LPBYTE)&tziReg,
                                     &cbData);
        }
        RegCloseKey(hkey);
        if (result == ERROR_SUCCESS) {
            /* Assume that offsets are grouped together, and bail out
               when we've scanned everything with a matching
               offset. */
            if (firstMatch >= 0 && tziKey.Bias != tziReg.Bias) {
                break;
            }
            /* Windows alters the DaylightBias in some situations.
               Using the bias and the rules suffices, so overwrite
               these unreliable fields. */
            tziKey.StandardBias = tziReg.StandardBias;
            tziKey.DaylightBias = tziReg.DaylightBias;
            if (uprv_memcmp((char *)&tziKey, (char*)&tziReg,
                       sizeof(tziKey)) == 0) {
                if (firstMatch < 0) {
                    firstMatch = j;
                }
                lastMatch = j;
            }
        }
    }

    /* This should never happen; if it does it means our table doesn't
       match Windows AT ALL, perhaps because this is post-XP? */
    if (firstMatch < 0) {
        return NULL;
    }
    
    if (firstMatch != lastMatch) {
        /* Offset+Rules lookup yielded >= 2 matches.  Try to match the
           localized display name.  Get the name from the registry
           (not the API). This avoids conversion issues.  Use the
           standard name, since Windows modifies the daylight name to
           match the standard name if there is no DST. */
        result = RegOpenKeyEx(HKEY_LOCAL_MACHINE,
                              CURRENT_ZONE_REGKEY,
                              0,
                              KEY_QUERY_VALUE,
                              &hkey);
        if (result == ERROR_SUCCESS) {
            stdNameSize = sizeof(stdName);
            result = RegQueryValueEx(hkey,
                                     (LPTSTR)STANDARD_NAME_REGKEY,
                                     NULL,
                                     NULL,
                                     (LPBYTE)stdName,
                                     &stdNameSize);
            RegCloseKey(hkey);

            /* Scan through the Windows time zone data in the registry
               again (just the range of zones with matching TZIs) and
               look for a standard display name match. */
            for (j=firstMatch; j<=lastMatch; j++) {
                result = openTZRegKey(&hkey, ZONE_MAP[j].winid, winType);
                if (result == ERROR_SUCCESS) {
                    stdRegNameSize = sizeof(stdRegName);
                    result = RegQueryValueEx(hkey,
                                             (LPTSTR)STD_REGKEY,
                                             NULL,
                                             NULL,
                                             (LPBYTE)stdRegName,
                                             &stdRegNameSize);
                }
                RegCloseKey(hkey);
                if (result == ERROR_SUCCESS &&
                    stdRegNameSize == stdNameSize &&
                    uprv_memcmp(stdName, stdRegName, stdNameSize) == 0) {
                    firstMatch = j; /* record the match */
                    break;
                }
            }
        } else {
            RegCloseKey(hkey); /* should never get here */
        }
    }

    return ZONE_MAP[firstMatch].icuid;
}

#endif /*U_WINDOWS*/

/* Generic time zone layer -------------------------------------------------- */

/* Time zone utilities */
U_CAPI void U_EXPORT2
uprv_tzset()
{
#ifdef U_TZSET
    U_TZSET();
#else
    /* no initialization*/
#endif
}

U_CAPI int32_t U_EXPORT2
uprv_timezone()
{
#ifdef U_TIMEZONE
    return U_TIMEZONE;
#else
    time_t t, t1, t2;
    struct tm tmrec;
    UBool dst_checked;
    int32_t tdiff = 0;

    time(&t);
    uprv_memcpy( &tmrec, localtime(&t), sizeof(tmrec) );
    dst_checked = (tmrec.tm_isdst != 0); /* daylight savings time is checked*/
    t1 = mktime(&tmrec);                 /* local time in seconds*/
    uprv_memcpy( &tmrec, gmtime(&t), sizeof(tmrec) );
    t2 = mktime(&tmrec);                 /* GMT (or UTC) in seconds*/
    tdiff = t2 - t1;
    /* imitate NT behaviour, which returns same timezone offset to GMT for
       winter and summer*/
    if (dst_checked)
        tdiff += 3600;
    return tdiff;
#endif
}

/* Note that U_TZNAME does *not* have to be tzname, but if it is,
   some platforms need to have it declared here. */ 

#if defined(U_TZNAME) && (defined(U_IRIX) || defined(U_DARWIN) || defined(U_CYGWIN))
/* RS6000 and others reject char **tzname.  */
extern U_IMPORT char *U_TZNAME[];
#endif

#if defined(U_DARWIN)   /* For Mac OS X */
#define TZZONELINK      "/etc/localtime"
#define TZZONEINFO      "/usr/share/zoneinfo/"
static char *gTimeZoneBuffer = NULL; /* Heap allocated */
#endif

U_CAPI const char* U_EXPORT2
uprv_tzname(int n)
{
#ifdef U_WINDOWS
    char* id = (char*) detectWindowsTimeZone();
    if (id != NULL) {
        return id;
    }
#endif

#if defined(U_DARWIN)
    int ret;

    char *tzenv;

    tzenv = getenv("TZFILE");
    if (tzenv != NULL) {
        return tzenv;
    }

#if 0
    /* TZ is often set to "PST8PDT" or similar, so we cannot use it. Alan */
    tzenv = getenv("TZ");
    if (tzenv != NULL) {
        return tzenv;
    }
#endif
    
    /* Caller must handle threading issues */
    if (gTimeZoneBuffer == NULL) {
        gTimeZoneBuffer = (char *) uprv_malloc(MAXPATHLEN + 2);

        ret = readlink(TZZONELINK, gTimeZoneBuffer, MAXPATHLEN + 2);
        if (0 < ret) {
            gTimeZoneBuffer[ret] = '\0';
            if (uprv_strncmp(gTimeZoneBuffer, TZZONEINFO, sizeof(TZZONEINFO) - 1) == 0) {
                return (gTimeZoneBuffer += sizeof(TZZONEINFO) - 1);
            }
        }

        uprv_free(gTimeZoneBuffer);
        gTimeZoneBuffer = NULL;
    }
#endif

#ifdef U_TZNAME
    return U_TZNAME[n];
#else
    return "";
#endif
}

/* Get and set the ICU data directory --------------------------------------- */

static char *gDataDirectory = NULL;
#if U_POSIX_LOCALE
 static char *gCorrectedPOSIXLocale = NULL; /* Heap allocated */
#endif

static UBool U_CALLCONV putil_cleanup(void)
{
    if (gDataDirectory && *gDataDirectory) {
        uprv_free(gDataDirectory);
    }
    gDataDirectory = NULL;
#if U_POSIX_LOCALE
    if (gCorrectedPOSIXLocale) {
        uprv_free(gCorrectedPOSIXLocale);
        gCorrectedPOSIXLocale = NULL;
    }
#endif
    return TRUE;
}

/*
 * Set the data directory.
 *    Make a copy of the passed string, and set the global data dir to point to it.
 *    TODO:  see bug #2849, regarding thread safety.
 */
U_CAPI void U_EXPORT2
u_setDataDirectory(const char *directory) {
    char *newDataDir;
    int32_t length;

    if(directory==NULL || *directory==0) {
        /* A small optimization to prevent the malloc and copy when the
        shared library is used, and this is a way to make sure that NULL
        is never returned.
        */
        newDataDir = (char *)"";
    }
    else {
        length=(int32_t)uprv_strlen(directory);
        newDataDir = (char *)uprv_malloc(length + 2);
        uprv_strcpy(newDataDir, directory);

#if (U_FILE_SEP_CHAR != U_FILE_ALT_SEP_CHAR)
        {
            char *p;
            while(p = uprv_strchr(newDataDir, U_FILE_ALT_SEP_CHAR)) {
                *p = U_FILE_SEP_CHAR;
            }
        }
#endif
    }

    umtx_lock(NULL);
    if (gDataDirectory && *gDataDirectory) {
        uprv_free(gDataDirectory);
    }
    gDataDirectory = newDataDir;
    ucln_common_registerCleanup(UCLN_COMMON_PUTIL, putil_cleanup);
    umtx_unlock(NULL);
}

U_CAPI UBool U_EXPORT2
uprv_pathIsAbsolute(const char *path) 
{
  if(!path || !*path) { 
    return FALSE; 
  }

  if(*path == U_FILE_SEP_CHAR) {
    return TRUE;
  }

#if (U_FILE_SEP_CHAR != U_FILE_ALT_SEP_CHAR)
  if(*path == U_FILE_ALT_SEP_CHAR) {
    return TRUE;
  }
#endif

#if defined(U_WINDOWS)
  if( (((path[0] >= 'A') && (path[0] <= 'Z')) ||
       ((path[0] >= 'a') && (path[0] <= 'z'))) &&
      path[1] == ':' ) {
    return TRUE;
  }
#endif

  return FALSE;
}

U_CAPI const char * U_EXPORT2
u_getDataDirectory(void) {
    const char *path = NULL;

    /* if we have the directory, then return it immediately */
    umtx_lock(NULL);
    path = gDataDirectory;
    umtx_unlock(NULL);

    if(path) {
        return path;
    }

    /*
    When ICU_NO_USER_DATA_OVERRIDE is defined, users aren't allowed to
    override ICU's data with the ICU_DATA environment variable. This prevents
    problems where multiple custom copies of ICU's specific version of data
    are installed on a system. Either the application must define the data
    directory with u_setDataDirectory, define ICU_DATA_DIR when compiling
    ICU, set the data with udata_setCommonData or trust that all of the
    required data is contained in ICU's data library that contains
    the entry point defined by U_ICUDATA_ENTRY_POINT.

    There may also be some platforms where environment variables
    are not allowed.
    */
#   if !defined(ICU_NO_USER_DATA_OVERRIDE) && (!defined(UCONFIG_NO_FILE_IO) || !UCONFIG_NO_FILE_IO)
    /* First try to get the environment variable */
    path=getenv("ICU_DATA");
#   endif

    /* ICU_DATA_DIR may be set as a compile option */
#   ifdef ICU_DATA_DIR
    if(path==NULL || *path==0) {
        path=ICU_DATA_DIR;
    }
#   endif

    if(path==NULL) {
        /* It looks really bad, set it to something. */
        path = "";
    }

    u_setDataDirectory(path);
    return gDataDirectory;
}





/* Macintosh-specific locale information ------------------------------------ */
#ifdef XP_MAC

typedef struct {
    int32_t script;
    int32_t region;
    int32_t lang;
    int32_t date_region;
    const char* posixID;
} mac_lc_rec;

/* Todo: This will be updated with a newer version from www.unicode.org web
   page when it's available.*/
#define MAC_LC_MAGIC_NUMBER -5
#define MAC_LC_INIT_NUMBER -9

static const mac_lc_rec mac_lc_recs[] = {
    MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, 0, "en_US",
    /* United States*/
    MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, 1, "fr_FR",
    /* France*/
    MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, 2, "en_GB",
    /* Great Britain*/
    MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, 3, "de_DE",
    /* Germany*/
    MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, 4, "it_IT",
    /* Italy*/
    MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, 5, "nl_NL",
    /* Metherlands*/
    MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, 6, "fr_BE",
    /* French for Belgium or Lxembourg*/
    MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, 7, "sv_SE",
    /* Sweden*/
    MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, 9, "da_DK",
    /* Denmark*/
    MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, 10, "pt_PT",
    /* Portugal*/
    MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, 11, "fr_CA",
    /* French Canada*/
    MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, 13, "is_IS",
    /* Israel*/
    MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, 14, "ja_JP",
    /* Japan*/
    MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, 15, "en_AU",
    /* Australia*/
    MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, 16, "ar_AE",
    /* the Arabic world (?)*/
    MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, 17, "fi_FI",
    /* Finland*/
    MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, 18, "fr_CH",
    /* French for Switzerland*/
    MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, 19, "de_CH",
    /* German for Switzerland*/
    MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, 20, "el_GR",
    /* Greece*/
    MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, 21, "is_IS",
    /* Iceland ===*/
    /*MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, 22, "",*/
    /* Malta ===*/
    /*MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, 23, "",*/
    /* Cyprus ===*/
    MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, 24, "tr_TR",
    /* Turkey ===*/
    MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, 25, "sh_YU",
    /* Croatian system for Yugoslavia*/
    /*MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, 33, "",*/
    /* Hindi system for India*/
    /*MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, 34, "",*/
    /* Pakistan*/
    MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, 41, "lt_LT",
    /* Lithuania*/
    MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, 42, "pl_PL",
    /* Poland*/
    MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, 43, "hu_HU",
    /* Hungary*/
    MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, 44, "et_EE",
    /* Estonia*/
    MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, 45, "lv_LV",
    /* Latvia*/
    /*MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, 46, "",*/
    /* Lapland  [Ask Rich for the data. HS]*/
    /*MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, 47, "",*/
    /* Faeroe Islands*/
    MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, 48, "fa_IR",
    /* Iran*/
    MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, 49, "ru_RU",
    /* Russia*/
    MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, 50, "en_IE",
    /* Ireland*/
    MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, 51, "ko_KR",
    /* Korea*/
    MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, 52, "zh_CN",
    /* People's Republic of China*/
    MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, 53, "zh_TW",
    /* Taiwan*/
    MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, 54, "th_TH",
    /* Thailand*/

    /* fallback is en_US*/
    MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER, MAC_LC_MAGIC_NUMBER,
    MAC_LC_MAGIC_NUMBER, "en_US"
};

#endif

#if U_POSIX_LOCALE
/* Return just the POSIX id, whatever happens to be in it */
static const char *uprv_getPOSIXID(void)
{
    static const char* posixID = NULL;
    if (posixID == 0) {
        posixID = getenv("LC_ALL");
        if (posixID == 0) {
            posixID = getenv("LANG");
            if (posixID == 0) {
                /*
                * On Solaris two different calls to setlocale can result in 
                * different values. Only get this value once.
                */
                posixID = setlocale(LC_ALL, NULL);
            }
        }
    }

    if (posixID==0)
    {
        /* Nothing worked.  Give it a nice value. */
        posixID = "en_US";
    }
    else if ((uprv_strcmp("C", posixID) == 0)
        || (uprv_strchr(posixID, ' ') != NULL)
        || (uprv_strchr(posixID, '/') != NULL))
    {   /* HPUX returns 'C C C C C C C' */
        /* Solaris can return /en_US/C/C/C/C/C on the second try. */
        /* Maybe we got some garbage.  Give it a nice value. */
        posixID = "en_US_POSIX";
    }
    return posixID;
}
#endif

/* NOTE: The caller should handle thread safety */
U_CAPI const char* U_EXPORT2
uprv_getDefaultLocaleID()
{
#if U_POSIX_LOCALE
/*
  Note that:  (a '!' means the ID is improper somehow)
     LC_ALL  ---->     default_loc          codepage
--------------------------------------------------------
     ab.CD             ab                   CD
     ab@CD             ab__CD               -
     ab@CD.EF          ab__CD               EF

     ab_CD.EF@GH       ab_CD_GH             EF

Some 'improper' ways to do the same as above:
  !  ab_CD@GH.EF       ab_CD_GH             EF
  !  ab_CD.EF@GH.IJ    ab_CD_GH             EF
  !  ab_CD@ZZ.EF@GH.IJ ab_CD_GH             EF

     _CD@GH            _CD_GH               -
     _CD.EF@GH         _CD_GH               EF

The variant cannot have dots in it.
The 'rightmost' variant (@xxx) wins.
The leftmost codepage (.xxx) wins.
*/
    char *correctedPOSIXLocale = 0;
    const char* posixID = uprv_getPOSIXID();
    const char *p;
    const char *q;
    int32_t len;

    /* Format: (no spaces)
    ll [ _CC ] [ . MM ] [ @ VV]

      l = lang, C = ctry, M = charmap, V = variant
    */

    if (gCorrectedPOSIXLocale != NULL) {
        return gCorrectedPOSIXLocale; 
    }

    if ((p = uprv_strchr(posixID, '.')) != NULL) {
        /* assume new locale can't be larger than old one? */
        correctedPOSIXLocale = uprv_malloc(uprv_strlen(posixID));
        uprv_strncpy(correctedPOSIXLocale, posixID, p-posixID);
        correctedPOSIXLocale[p-posixID] = 0;

        /* do not copy after the @ */
        if ((p = uprv_strchr(correctedPOSIXLocale, '@')) != NULL) {
            correctedPOSIXLocale[p-correctedPOSIXLocale] = 0;
        }
    }

    /* Note that we scan the *uncorrected* ID. */
    if ((p = uprv_strrchr(posixID, '@')) != NULL) {
        if (correctedPOSIXLocale == NULL) {
            correctedPOSIXLocale = uprv_malloc(uprv_strlen(posixID));
            uprv_strncpy(correctedPOSIXLocale, posixID, p-posixID);
            correctedPOSIXLocale[p-posixID] = 0;
        }
        p++;

        /* Take care of any special cases here.. */
        if (!uprv_strcmp(p, "nynorsk")) {
            p = "NY";

            /*      Should we assume no_NO_NY instead of possible no__NY?
            * if (!uprv_strcmp(correctedPOSIXLocale, "no")) {
            *     uprv_strcpy(correctedPOSIXLocale, "no_NO");
            * }
            */
        }

        if (uprv_strchr(correctedPOSIXLocale,'_') == NULL) {
            uprv_strcat(correctedPOSIXLocale, "__"); /* aa@b -> aa__b */
        }
        else {
            uprv_strcat(correctedPOSIXLocale, "_"); /* aa_CC@b -> aa_CC_b */
        }

        if ((q = uprv_strchr(p, '.')) != NULL) {
            /* How big will the resulting string be? */
            len = (int32_t)(uprv_strlen(correctedPOSIXLocale) + (q-p));
            uprv_strncat(correctedPOSIXLocale, p, q-p);
            correctedPOSIXLocale[len] = 0;
        }
        else {
            /* Anything following the @ sign */
            uprv_strcat(correctedPOSIXLocale, p);
        }

        /* Should there be a map from 'no@nynorsk' -> no_NO_NY here?
         * How about 'russian' -> 'ru'?
         */
    }

    /* Was a correction made? */
    if (correctedPOSIXLocale != NULL) {
        posixID = correctedPOSIXLocale;
    }
    else {
        /* copy it, just in case the original pointer goes away.  See j2395 */
        correctedPOSIXLocale = (char *)uprv_malloc(uprv_strlen(posixID) + 1);
        posixID = uprv_strcpy(correctedPOSIXLocale, posixID);
    }

    if (gCorrectedPOSIXLocale == NULL) {
        gCorrectedPOSIXLocale = correctedPOSIXLocale;
        ucln_common_registerCleanup(UCLN_COMMON_PUTIL, putil_cleanup);
        correctedPOSIXLocale = NULL;
    }

    if (correctedPOSIXLocale != NULL) {  /* Was already set - clean up. */
        uprv_free(correctedPOSIXLocale); 
    }

    return posixID;

#elif defined(U_WINDOWS)
    UErrorCode status = U_ZERO_ERROR;
    LCID id = GetThreadLocale();
    const char* locID = uprv_convertToPosix(id, &status);

    if (U_FAILURE(status)) {
        locID = "en_US";
    }
    return locID;

#elif defined(XP_MAC)
    int32_t script = MAC_LC_INIT_NUMBER;
    /* = IntlScript(); or GetScriptManagerVariable(smSysScript);*/
    int32_t region = MAC_LC_INIT_NUMBER;
    /* = GetScriptManagerVariable(smRegionCode);*/
    int32_t lang = MAC_LC_INIT_NUMBER;
    /* = GetScriptManagerVariable(smScriptLang);*/
    int32_t date_region = MAC_LC_INIT_NUMBER;
    const char* posixID = 0;
    int32_t count = sizeof(mac_lc_recs) / sizeof(mac_lc_rec);
    int32_t i;
    Intl1Hndl ih;

    ih = (Intl1Hndl) GetIntlResource(1);
    if (ih)
        date_region = ((uint16_t)(*ih)->intl1Vers) >> 8;

    for (i = 0; i < count; i++) {
        if (   ((mac_lc_recs[i].script == MAC_LC_MAGIC_NUMBER)
             || (mac_lc_recs[i].script == script))
            && ((mac_lc_recs[i].region == MAC_LC_MAGIC_NUMBER)
             || (mac_lc_recs[i].region == region))
            && ((mac_lc_recs[i].lang == MAC_LC_MAGIC_NUMBER)
             || (mac_lc_recs[i].lang == lang))
            && ((mac_lc_recs[i].date_region == MAC_LC_MAGIC_NUMBER)
             || (mac_lc_recs[i].date_region == date_region))
            )
        {
            posixID = mac_lc_recs[i].posixID;
            break;
        }
    }

    return posixID;

#elif defined(OS400)
    /* locales are process scoped and are by definition thread safe */
    static char correctedLocale[64];
    const  char *localeID = getenv("LC_ALL");
           char *p;

    if (localeID == NULL)
        localeID = getenv("LANG");
    if (localeID == NULL)
        localeID = setlocale(LC_ALL, NULL);
    /* Make sure we have something... */
    if (localeID == NULL)
        return "en_US_POSIX";

    /* Extract the locale name from the path. */
    if((p = uprv_strrchr(localeID, '/')) != NULL)
    {
        /* Increment p to start of locale name. */
        p++;
        localeID = p;
    }

    /* Copy to work location. */
    uprv_strcpy(correctedLocale, localeID);

    /* Strip off the '.locale' extension. */
    if((p = uprv_strchr(correctedLocale, '.')) != NULL) {
        *p = 0;
    }

    /* Upper case the locale name. */
    T_CString_toUpperCase(correctedLocale);

    /* See if we are using the POSIX locale.  Any of the
    * following are equivalent and use the same QLGPGCMA
    * (POSIX) locale.
    * QLGPGCMA2 means UCS2
    * QLGPGCMA_4 means UTF-32
    * QLGPGCMA_8 means UTF-8
    */
    if ((uprv_strcmp("C", correctedLocale) == 0) ||
        (uprv_strcmp("POSIX", correctedLocale) == 0) ||
        (uprv_strncmp("QLGPGCMA", correctedLocale, 8) == 0))
    {
        uprv_strcpy(correctedLocale, "en_US_POSIX");
    }
    else
    {
        int16_t LocaleLen;

        /* Lower case the lang portion. */
        for(p = correctedLocale; *p != 0 && *p != '_'; p++)
        {
            *p = uprv_tolower(*p);
        }

        /* Adjust for Euro.  After '_E' add 'URO'. */
        LocaleLen = uprv_strlen(correctedLocale);
        if (correctedLocale[LocaleLen - 2] == '_' &&
            correctedLocale[LocaleLen - 1] == 'E')
        {
            uprv_strcat(correctedLocale, "URO");
        }

        /* If using Lotus-based locale then convert to
         * equivalent non Lotus.
         */
        else if (correctedLocale[LocaleLen - 2] == '_' &&
            correctedLocale[LocaleLen - 1] == 'L')
        {
            correctedLocale[LocaleLen - 2] = 0;
        }

        /* There are separate simplified and traditional
         * locales called zh_HK_S and zh_HK_T.
         */
        else if (uprv_strncmp(correctedLocale, "zh_HK", 5) == 0)
        {
            uprv_strcpy(correctedLocale, "zh_HK");
        }

        /* A special zh_CN_GBK locale...
        */
        else if (uprv_strcmp(correctedLocale, "zh_CN_GBK") == 0)
        {
            uprv_strcpy(correctedLocale, "zh_CN");
        }

    }

    return correctedLocale;
#endif

}


static const char*  
int_getDefaultCodepage()
{
#if defined(OS400)
    uint32_t ccsid = 37; /* Default to ibm-37 */
    static char codepage[64];
    Qwc_JOBI0400_t jobinfo;
    Qus_EC_t error = { sizeof(Qus_EC_t) }; /* SPI error code */

    EPT_CALL(QUSRJOBI)(&jobinfo, sizeof(jobinfo), "JOBI0400",
        "*                         ", "                ", &error);

    if (error.Bytes_Available == 0) {
        if (jobinfo.Coded_Char_Set_ID != 0xFFFF) {
            ccsid = (uint32_t)jobinfo.Coded_Char_Set_ID;
        }
        else if (jobinfo.Default_Coded_Char_Set_Id != 0xFFFF) {
            ccsid = (uint32_t)jobinfo.Default_Coded_Char_Set_Id;
        }
        /* else use the default */
    }
    sprintf(codepage,"ibm-%d", ccsid);
    return codepage;

#elif defined(OS390)
    static char codepage[64];
    sprintf(codepage,"%s" UCNV_SWAP_LFNL_OPTION_STRING, nl_langinfo(CODESET));
    return codepage;

#elif defined(XP_MAC)
    return "macintosh"; /* TODO: Macintosh Roman. There must be a better way. fixme! */

#elif defined(U_WINDOWS)
    static char codepage[64];
    sprintf(codepage, "windows-%d", GetACP());
    return codepage;

#elif U_POSIX_LOCALE
    static char codesetName[100];
    char *name = NULL;
    char *euro = NULL;
    const char *localeName = NULL;

    uprv_memset(codesetName, 0, sizeof(codesetName));

    /* Check setlocale before the environment variables
       because the application may have set it first */

    /* Use setlocale in a nice way.
       Maybe the application used setlocale already.
       Normally this won't work. */
    localeName = setlocale(LC_CTYPE, NULL);
    if (localeName != NULL && (name = (uprv_strchr(localeName, '.'))) != NULL) {
        /* strip the locale name and look at the suffix only */
        name = uprv_strncpy(codesetName, name+1, sizeof(codesetName));
        codesetName[sizeof(codesetName)-1] = 0;
        if ((euro = (uprv_strchr(name, '@'))) != NULL) {
           *euro = 0;
        }
        /* if we can find the codset name from setlocale, return that. */
        if (*name) {
            return name;
        }
    }
    /* else "C" was probably returned. That's underspecified. */

    /* Use setlocale a little more forcefully.
       The application didn't use setlocale */
    localeName = setlocale(LC_CTYPE, "");
    if (localeName != NULL && (name = (uprv_strchr(localeName, '.'))) != NULL) {
        /* strip the locale name and look at the suffix only */
        name = uprv_strncpy(codesetName, name+1, sizeof(codesetName));
        codesetName[sizeof(codesetName)-1] = 0;
        if ((euro = (uprv_strchr(name, '@'))) != NULL) {
           *euro = 0;
        }
        /* if we can find the codset name from setlocale, return that. */
        if (*name) {
            return name;
        }
    }
    /* else "C" or something like it was returned. That's still underspecified. */

#if U_HAVE_NL_LANGINFO_CODESET
    if (*codesetName) {
        uprv_memset(codesetName, 0, sizeof(codesetName));
    }
    /* When available, check nl_langinfo first because it usually gives more
       useful names. It depends on LC_CTYPE and not LANG or LC_ALL */
    {
        const char *codeset = nl_langinfo(U_NL_LANGINFO_CODESET);
        if (codeset != NULL) {
            uprv_strncpy(codesetName, codeset, sizeof(codesetName));
            codesetName[sizeof(codesetName)-1] = 0;
            return codesetName;
        }
    }
#endif

    /* Try a locale specified by the user.
       This is usually underspecified and usually checked by setlocale already. */
    if (*codesetName) {
        uprv_memset(codesetName, 0, sizeof(codesetName));
    }
    localeName = uprv_getPOSIXID();
    if (localeName != NULL && (name = (uprv_strchr(localeName, '.'))) != NULL) {
        /* strip the locale name and look at the suffix only */
        name = uprv_strncpy(codesetName, name+1, sizeof(codesetName));
        codesetName[sizeof(codesetName)-1] = 0;
        if ((euro = (uprv_strchr(name, '@'))) != NULL) {
           *euro = 0;
        }
        /* if we can find the codset name, return that. */
        if (*name) {
            return name;
        }
    }

    if (*codesetName == 0)
    {
        /* if the table lookup failed, return US ASCII (ISO 646). */
        uprv_strcpy(codesetName, "US-ASCII");
    }
    return codesetName;
#else
    return "US-ASCII";
#endif
}


U_CAPI const char*  U_EXPORT2
uprv_getDefaultCodepage()
{
    static char const  *name = NULL;
    umtx_lock(NULL);
    if (name == NULL) {
        name = int_getDefaultCodepage();
    }
    umtx_unlock(NULL);
    return name;
}


/* end of platform-specific implementation -------------- */

/* version handling --------------------------------------------------------- */

U_CAPI void U_EXPORT2
u_versionFromString(UVersionInfo versionArray, const char *versionString) {
    char *end;
    uint16_t part=0;

    if(versionArray==NULL) {
        return;
    }

    if(versionString!=NULL) {
        for(;;) {
            versionArray[part]=(uint8_t)uprv_strtoul(versionString, &end, 10);
            if(end==versionString || ++part==U_MAX_VERSION_LENGTH || *end!=U_VERSION_DELIMITER) {
                break;
            }
            versionString=end+1;
        }
    }

    while(part<U_MAX_VERSION_LENGTH) {
        versionArray[part++]=0;
    }
}

U_CAPI void U_EXPORT2
u_versionToString(UVersionInfo versionArray, char *versionString) {
    uint16_t count, part;
    uint8_t field;

    if(versionString==NULL) {
        return;
    }

    if(versionArray==NULL) {
        versionString[0]=0;
        return;
    }

    /* count how many fields need to be written */
    for(count=4; count>0 && versionArray[count-1]==0; --count) {
    }

    if(count <= 1) {
        count = 2;
    }

    /* write the first part */
    /* write the decimal field value */
    field=versionArray[0];
    if(field>=100) {
        *versionString++=(char)('0'+field/100);
        field%=100;
    }
    if(field>=10) {
        *versionString++=(char)('0'+field/10);
        field%=10;
    }
    *versionString++=(char)('0'+field);

    /* write the following parts */
    for(part=1; part<count; ++part) {
        /* write a dot first */
        *versionString++=U_VERSION_DELIMITER;

        /* write the decimal field value */
        field=versionArray[part];
        if(field>=100) {
            *versionString++=(char)('0'+field/100);
            field%=100;
        }
        if(field>=10) {
            *versionString++=(char)('0'+field/10);
            field%=10;
        }
        *versionString++=(char)('0'+field);
    }

    /* NUL-terminate */
    *versionString=0;
}

U_CAPI void U_EXPORT2
u_getVersion(UVersionInfo versionArray) {
    u_versionFromString(versionArray, U_ICU_VERSION);
}

/*
 * Hey, Emacs, please set the following:
 *
 * Local Variables:
 * indent-tabs-mode: nil
 * End:
 *
 */