File: dinfermodel.c

package info (click to toggle)
paraview 5.13.2%2Bdfsg-3
  • links: PTS, VCS
  • area: main
  • in suites: forky, sid, trixie
  • size: 544,220 kB
  • sloc: cpp: 3,374,605; ansic: 1,332,409; python: 150,381; xml: 122,166; sql: 65,887; sh: 7,317; javascript: 5,262; yacc: 4,417; java: 3,977; perl: 2,363; lex: 1,929; f90: 1,397; makefile: 170; objc: 153; tcl: 59; pascal: 50; fortran: 29
file content (1640 lines) | stat: -rw-r--r-- 44,716 bytes parent folder | download
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
/**
 * @file
 *
 * Infer as much as possible from the omode + path.
 * Rewrite the path to a canonical form.
 *
 * Copyright 2018 University Corporation for Atmospheric
 * Research/Unidata. See COPYRIGHT file for more info.
*/

#include "config.h"
#include <stdlib.h>
#ifdef HAVE_UNISTD_H
#include <unistd.h>
#endif
#ifdef HAVE_SYS_TYPES_H
#include <sys/types.h>
#endif

#include "ncdispatch.h"
#include "ncpathmgr.h"
#include "netcdf_mem.h"
#include "fbits.h"
#include "ncbytes.h"
#include "nclist.h"
#include "nclog.h"
#include "ncrc.h"
#ifdef ENABLE_BYTERANGE
#include "nchttp.h"
#ifdef ENABLE_S3_SDK
#include "ncs3sdk.h"
#endif
#endif

#ifndef nulldup
 #define nulldup(x) ((x)?strdup(x):(x))
#endif

#undef DEBUG

/* If Defined, then use only stdio for all magic number io;
   otherwise use stdio or mpio as required.
 */
#undef USE_STDIO

/**
Sort info for open/read/close of
file when searching for magic numbers
*/
struct MagicFile {
    const char* path;
    struct NCURI* uri;
    int omode;
    NCmodel* model;
    long long filelen;
    int use_parallel;
    int iss3;
    void* parameters; /* !NULL if inmemory && !diskless */
    FILE* fp;
#ifdef USE_PARALLEL
    MPI_File fh;
#endif
#ifdef ENABLE_BYTERANGE
    char* curlurl; /* url to use with CURLOPT_SET_URL */
    NC_HTTP_STATE* state;
#ifdef ENABLE_S3_SDK
    NCS3INFO s3;
    void* s3client;
    char* errmsg;
#endif
#endif
};

/** @internal Magic number for HDF5 files. To be consistent with
 * H5Fis_hdf5, use the complete HDF5 magic number */
static char HDF5_SIGNATURE[MAGIC_NUMBER_LEN] = "\211HDF\r\n\032\n";

#define modelcomplete(model) ((model)->impl != 0)

#ifdef DEBUG
static void dbgflush(void)
{
    fflush(stdout);
    fflush(stderr);
}

static void
fail(int err)
{
    return;
}

static int
check(int err)
{
    if(err != NC_NOERR)
	fail(err);
    return err;
}
#else
#define check(err) (err)
#endif

/*
Define a table of "mode=" string values
from which the implementation can be inferred.
Note that only cases that can currently
take URLs are included.
*/
static struct FORMATMODES {
    const char* tag;
    const int impl; /* NC_FORMATX_XXX value */
    const int format; /* NC_FORMAT_XXX value */
} formatmodes[] = {
{"dap2",NC_FORMATX_DAP2,NC_FORMAT_CLASSIC},
{"dap4",NC_FORMATX_DAP4,NC_FORMAT_NETCDF4},
{"netcdf-3",NC_FORMATX_NC3,0}, /* Might be e.g. cdf5 */
{"classic",NC_FORMATX_NC3,0}, /* ditto */
{"netcdf-4",NC_FORMATX_NC4,NC_FORMAT_NETCDF4},
{"enhanced",NC_FORMATX_NC4,NC_FORMAT_NETCDF4},
{"udf0",NC_FORMATX_UDF0,0},
{"udf1",NC_FORMATX_UDF1,0},
{"nczarr",NC_FORMATX_NCZARR,NC_FORMAT_NETCDF4},
{"zarr",NC_FORMATX_NCZARR,NC_FORMAT_NETCDF4},
{"bytes",NC_FORMATX_NC4,NC_FORMAT_NETCDF4}, /* temporary until 3 vs 4 is determined */
{NULL,0},
};

/* Replace top-level name with defkey=defvalue */
static const struct MACRODEF {
    char* name;
    char* defkey;
    char* defvalues[4];
} macrodefs[] = {
{"zarr","mode",{"nczarr","zarr",NULL}},
{"dap2","mode",{"dap2",NULL}},
{"dap4","mode",{"dap4",NULL}},
{"s3","mode",{"s3","nczarr",NULL}},
{"bytes","mode",{"bytes",NULL}},
{"xarray","mode",{"zarr", NULL}},
{"noxarray","mode",{"nczarr", "noxarray", NULL}},
{"zarr","mode",{"nczarr","zarr", NULL}},
{NULL,NULL,{NULL}}
};

/*
Mode inferences: if mode contains key value, then add the inferred value;
Warning: be careful how this list is constructed to avoid infinite inferences.
In order to (mostly) avoid that consequence, any attempt to
infer a value that is already present will be ignored.
This effectively means that the inference graph
must be a DAG and may not have cycles.
You have been warned.
*/
static const struct MODEINFER {
    char* key;
    char* inference;
} modeinferences[] = {
{"zarr","nczarr"},
{"xarray","zarr"},
{"noxarray","nczarr"},
{"noxarray","zarr"},
{NULL,NULL}
};

/* Mode negations: if mode contains key, then remove all occurrences of the inference and repeat */
static const struct MODEINFER modenegations[] = {
{"bytes","nczarr"}, /* bytes negates (nc)zarr */
{"bytes","zarr"},
{"noxarray","xarray"},
{NULL,NULL}
};

/* Map FORMATX to readability to get magic number */
static struct Readable {
    int impl;
    int readable;
} readable[] = {
{NC_FORMATX_NC3,1},
{NC_FORMATX_NC_HDF5,1},
{NC_FORMATX_NC_HDF4,1},
{NC_FORMATX_PNETCDF,1},
{NC_FORMATX_DAP2,0},
{NC_FORMATX_DAP4,0},
{NC_FORMATX_UDF0,1},
{NC_FORMATX_UDF1,1},
{NC_FORMATX_NCZARR,0}, /* eventually make readable */
{0,0},
};

/* Define the known URL protocols and their interpretation */
static struct NCPROTOCOLLIST {
    const char* protocol;
    const char* substitute;
    const char* fragments; /* arbitrary fragment arguments */
} ncprotolist[] = {
    {"http",NULL,NULL},
    {"https",NULL,NULL},
    {"file",NULL,NULL},
    {"dods","http","mode=dap2"},
    {"dap4","http","mode=dap4"},
    {"s3","s3","mode=s3"},
    {NULL,NULL,NULL} /* Terminate search */
};

/* Forward */
static int NC_omodeinfer(int useparallel, int omode, NCmodel*);
static int check_file_type(const char *path, int omode, int use_parallel, void *parameters, NCmodel* model, NCURI* uri);
static int processuri(const char* path, NCURI** urip, NClist* fraglist);
static int processmacros(NClist** fraglistp);
static char* envvlist2string(NClist* pairs, const char*);
static void set_default_mode(int* cmodep);
static int parseonchar(const char* s, int ch, NClist* segments);
static int mergelist(NClist** valuesp);

static int openmagic(struct MagicFile* file);
static int readmagic(struct MagicFile* file, long pos, char* magic);
static int closemagic(struct MagicFile* file);
static int NC_interpret_magic_number(char* magic, NCmodel* model);
#ifdef DEBUG
static void printmagic(const char* tag, char* magic,struct MagicFile*);
static void printlist(NClist* list, const char* tag);
#endif
static int isreadable(NCURI*,NCmodel*);
static char* list2string(NClist*);
static int parsepair(const char* pair, char** keyp, char** valuep);
static NClist* parsemode(const char* modeval);
static const char* getmodekey(const NClist* envv);
static int replacemode(NClist* envv, const char* newval);
static void infernext(NClist* current, NClist* next);
static int negateone(const char* mode, NClist* modes);
static void cleanstringlist(NClist* strs, int caseinsensitive);

/*
If the path looks like a URL, then parse it, reformat it.
*/
static int
processuri(const char* path, NCURI** urip, NClist* fraglenv)
{
    int stat = NC_NOERR;
    int found = 0;
    NClist* tmp = NULL;
    struct NCPROTOCOLLIST* protolist;
    NCURI* uri = NULL;
    size_t pathlen = strlen(path);
    char* str = NULL;
    const char** ufrags;
    const char** p;

    if(path == NULL || pathlen == 0) {stat = NC_EURL; goto done;}

    /* Defaults */
    if(urip) *urip = NULL;

    ncuriparse(path,&uri);
    if(uri == NULL) goto done; /* not url */

    /* Look up the protocol */
    for(found=0,protolist=ncprotolist;protolist->protocol;protolist++) {
        if(strcmp(uri->protocol,protolist->protocol) == 0) {
	    found = 1;
	    break;
	}
    }
    if(!found)
	{stat = NC_EINVAL; goto done;} /* unrecognized URL form */

    /* process the corresponding fragments for that protocol */
    if(protolist->fragments != NULL) {
	int i;
	tmp = nclistnew();
	if((stat = parseonchar(protolist->fragments,'&',tmp))) goto done;
	for(i=0;i<nclistlength(tmp);i++) {
	    char* key=NULL;
    	    char* value=NULL;
	    if((stat = parsepair(nclistget(tmp,i),&key,&value))) goto done;
	    if(value == NULL) value = strdup("");
	    nclistpush(fraglenv,key);
    	    nclistpush(fraglenv,value);
	}
	nclistfreeall(tmp); tmp = NULL;
    }

    /* Substitute the protocol in any case */
    if(protolist->substitute) ncurisetprotocol(uri,protolist->substitute);

    /* capture the fragments of the url */
    ufrags = ncurifragmentparams(uri);
    if(ufrags != NULL) {
        for(p=ufrags;*p;p+=2) {
	    const char* key = p[0];
	    const char* value = p[1];
	    nclistpush(fraglenv,nulldup(key));
	    value = (value==NULL?"":value);
	    nclistpush(fraglenv,strdup(value));
	}
    }
    if(urip) {
	*urip = uri;
	uri = NULL;
    }

done:
    nclistfreeall(tmp);
    nullfree(str);
    if(uri != NULL) ncurifree(uri);
    return check(stat);
}

/* Split a key=value pair */
static int
parsepair(const char* pair, char** keyp, char** valuep)
{
    const char* p;
    char* key = NULL;
    char* value = NULL;

    if(pair == NULL)
        return NC_EINVAL; /* empty pair */
    if(pair[0] == '\0' || pair[0] == '=')
        return NC_EINVAL; /* no key */
    p = strchr(pair,'=');
    if(p == NULL) {
	value = NULL;
	key = strdup(pair);
    } else {
	ptrdiff_t len = (p-pair);
	if((key = malloc(len+1))==NULL) return NC_ENOMEM;
	memcpy(key,pair,len);
	key[len] = '\0';
	if(p[1] == '\0')
	    value = NULL;
	else
	    value = strdup(p+1);
    }
    if(keyp) {*keyp = key; key = NULL;};
    if(valuep) {*valuep = value; value = NULL;};
    nullfree(key);
    nullfree(value);
    return NC_NOERR;
}

#if 0
static int
parseurlmode(const char* modestr, NClist* list)
{
    int stat = NC_NOERR;
    const char* p = NULL;
    const char* endp = NULL;

    if(modestr == NULL || *modestr == '\0') goto done;

    /* Split modestr at the commas or EOL */
    p = modestr;
    for(;;) {
	char* s;
	ptrdiff_t slen;
	endp = strchr(p,',');
	if(endp == NULL) endp = p + strlen(p);
	slen = (endp - p);
	if((s = malloc(slen+1)) == NULL) {stat = NC_ENOMEM; goto done;}
	memcpy(s,p,slen);
	s[slen] = '\0';
	nclistpush(list,s);
	if(*endp == '\0') break;
	p = endp+1;
    }

done:
    return check(stat);
}
#endif

/* Split a string at a given char */
static int
parseonchar(const char* s, int ch, NClist* segments)
{
    int stat = NC_NOERR;
    const char* p = NULL;
    const char* endp = NULL;

    if(s == NULL || *s == '\0') goto done;

    p = s;
    for(;;) {
	char* q;
	ptrdiff_t slen;
	endp = strchr(p,ch);
	if(endp == NULL) endp = p + strlen(p);
	slen = (endp - p);
	if((q = malloc(slen+1)) == NULL) {stat = NC_ENOMEM; goto done;}
	memcpy(q,p,slen);
	q[slen] = '\0';
	nclistpush(segments,q);
	if(*endp == '\0') break;
	p = endp+1;
    }

done:
    return check(stat);
}

/* Convert a key,value envv pairlist into a delimited string*/
static char*
envvlist2string(NClist* envv, const char* delim)
{
    int i;
    NCbytes* buf = NULL;
    char* result = NULL;

    if(envv == NULL || nclistlength(envv) == 0) return NULL;
    buf = ncbytesnew();
    for(i=0;i<nclistlength(envv);i+=2) {
	const char* key = nclistget(envv,i);
	const char* val = nclistget(envv,i+1);
	if(key == NULL || strlen(key) == 0) continue;
	assert(val != NULL);
	if(i > 0) ncbytescat(buf,"&");
	ncbytescat(buf,key);
	if(val != NULL && val[0] != '\0') {
	    ncbytescat(buf,"=");
	    ncbytescat(buf,val);
	}
    }
    result = ncbytesextract(buf);
    ncbytesfree(buf);
    return result;
}

/* Given a mode= argument, fill in the impl */
static int
processmodearg(const char* arg, NCmodel* model)
{
    int stat = NC_NOERR;
    struct FORMATMODES* format = formatmodes;
    for(;format->tag;format++) {
	if(strcmp(format->tag,arg)==0) {
            model->impl = format->impl;
	    if(format->format != 0) model->format = format->format;
	}
    }
    return check(stat);
}

/* Given an envv fragment list, do macro replacement */
static int
processmacros(NClist** fraglenvp)
{
    int stat = NC_NOERR;
    const struct MACRODEF* macros = NULL;
    NClist*  fraglenv = NULL;
    NClist* expanded = NULL;

    if(fraglenvp == NULL || nclistlength(*fraglenvp) == 0) goto done;
    fraglenv = *fraglenvp;
    expanded = nclistnew();
    while(nclistlength(fraglenv) > 0) {
	int found = 0;
	char* key = NULL;
	char* value = NULL;
	key = nclistremove(fraglenv,0); /* remove from changing front */
	value = nclistremove(fraglenv,0); /* remove from changing front */
	if(strlen(value) == 0) { /* must be a singleton  */
            for(macros=macrodefs;macros->name;macros++) {
                if(strcmp(macros->name,key)==0) {
		    char* const * p;
		    nclistpush(expanded,strdup(macros->defkey));
		    for(p=macros->defvalues;*p;p++) 
			nclistpush(expanded,strdup(*p));
		    found = 1;		    
		    break;
	        }
	    }
	}
	if(!found) {/* pass thru */
	    nclistpush(expanded,strdup(key));
    	    nclistpush(expanded,strdup(value));
	}
	nullfree(key);
	nullfree(value);
    }
    *fraglenvp = expanded; expanded = NULL;

done:
    nclistfreeall(expanded);
    nclistfreeall(fraglenv);
    return check(stat);
}

/* Process mode flag inferences */
static int
processinferences(NClist* fraglenv)
{
    int stat = NC_NOERR;
    const char* modeval = NULL;
    NClist* newmodes = nclistnew();
    NClist* currentmodes = NULL;
    NClist* nextmodes = nclistnew();
    int i;
    char* newmodeval = NULL;

    if(fraglenv == NULL || nclistlength(fraglenv) == 0) goto done;

    /* Get "mode" entry */
    if((modeval = getmodekey(fraglenv))==NULL) goto done;

    /* Get the mode as list */
    currentmodes = parsemode(modeval);

#ifdef DEBUG
    printlist(currentmodes,"processinferences: initial mode list");
#endif

    /* Do what amounts to breadth first inferencing down the inference DAG. */

    for(;;) {
        NClist* tmp = NULL;
        /* Compute the next set of inferred modes */
#ifdef DEBUG
printlist(currentmodes,"processinferences: current mode list");
#endif
        infernext(currentmodes,nextmodes);
#ifdef DEBUG
printlist(nextmodes,"processinferences: next mode list");
#endif
        /* move current modes into list of newmodes */
        for(i=0;i<nclistlength(currentmodes);i++) {
	    nclistpush(newmodes,nclistget(currentmodes,i));
	}
        nclistsetlength(currentmodes,0); /* clear current mode list */
        if(nclistlength(nextmodes) == 0) break; /* nothing more to do */
#ifdef DEBUG
printlist(newmodes,"processinferences: new mode list");
#endif
	/* Swap current and next */
        tmp = currentmodes;
	currentmodes = nextmodes;
	nextmodes = tmp;
        tmp = NULL;
    }
    /* cleanup any unused elements in currenmodes */
    nclistclearall(currentmodes);

    /* Ensure no duplicates */
    cleanstringlist(newmodes,1);

#ifdef DEBUG
    printlist(newmodes,"processinferences: final inferred mode list");
#endif

   /* Remove negative inferences */
   for(i=0;i<nclistlength(newmodes);i++) {
	const char* mode = nclistget(newmodes,i);
	negateone(mode,newmodes);
    }

    /* Store new mode value */
    if((newmodeval = list2string(newmodes))== NULL)
	{stat = NC_ENOMEM; goto done;}        
    if((stat=replacemode(fraglenv,newmodeval))) goto done;
    modeval = NULL;

done:
    nullfree(newmodeval);
    nclistfreeall(newmodes);
    nclistfreeall(currentmodes);
    nclistfreeall(nextmodes);
    return check(stat);
}


static int
negateone(const char* mode, NClist* newmodes)
{
    const struct MODEINFER* tests = modenegations;
    int changed = 0;
    for(;tests->key;tests++) {
	int i;
	if(strcasecmp(tests->key,mode)==0) {
	    /* Find and remove all instances of the inference value */
	    for(i=nclistlength(newmodes)-1;i>=0;i--) {
		char* candidate = nclistget(newmodes,i);
		if(strcasecmp(candidate,tests->inference)==0) {
		    nclistremove(newmodes,i);
		    nullfree(candidate);
	            changed = 1;
		}
	    }
        }
    }
    return changed;
}

static void
infernext(NClist* current, NClist* next)
{
    int i;
    for(i=0;i<nclistlength(current);i++) {
        const struct MODEINFER* tests = NULL;
	const char* cur = nclistget(current,i);
        for(tests=modeinferences;tests->key;tests++) {
	    if(strcasecmp(tests->key,cur)==0) {
	        /* Append the inferred mode unless dup */
		if(!nclistmatch(next,tests->inference,1))
	            nclistpush(next,strdup(tests->inference));
	    }
        }
    }
}

/*
Given a list of strings, remove nulls and duplicates
*/
static int
mergelist(NClist** valuesp)
{
    int i,j;
    int stat = NC_NOERR;
    NClist* values = *valuesp;
    NClist* allvalues = nclistnew();
    NClist* newvalues = nclistnew();
    char* value = NULL;

    for(i=0;i<nclistlength(values);i++) {
	char* val1 = nclistget(values,i);
	/* split on commas and put pieces into allvalues */
	if((stat=parseonchar(val1,',',allvalues))) goto done;
    }
    /* Remove duplicates and "" */
    while(nclistlength(allvalues) > 0) {
	value = nclistremove(allvalues,0);
	if(strlen(value) == 0) {
	    nullfree(value); value = NULL;
	} else {
	    for(j=0;j<nclistlength(newvalues);j++) {
	        char* candidate = nclistget(newvalues,j);
	        if(strcasecmp(candidate,value)==0)
	            {nullfree(value); value = NULL; break;}
	     }
	}
	if(value != NULL) {nclistpush(newvalues,value); value = NULL;}
    }
    /* Make sure to have at least 1 value */
    if(nclistlength(newvalues)==0) nclistpush(newvalues,strdup(""));
    *valuesp = values; values = NULL;

done:
    nclistfree(allvalues);
    nclistfreeall(values);
    nclistfreeall(newvalues);
    return check(stat);
}

static int
lcontains(NClist* l, const char* key0)
{
    int i;
    for(i=0;i<nclistlength(l);i++) {
        const char* key1 = nclistget(l,i);
	if(strcasecmp(key0,key1)==0) return 1;
    }
    return 0;
}

/* Warning values should not use nclistfreeall */
static void
collectvaluesbykey(NClist* fraglenv, const char* key, NClist* values)
{
    int i;
    /* collect all the values with the same key (including this one) */
    for(i=0;i<nclistlength(fraglenv);i+=2) {
        const char* key2 = nclistget(fraglenv,i);
        if(strcasecmp(key,key2)==0) {
	    const char* value2 = nclistget(fraglenv,i+1);
	    nclistpush(values,value2); value2 = NULL;
	}
    }
}

/* Warning allkeys should not use nclistfreeall */
static void
collectallkeys(NClist* fraglenv, NClist* allkeys)
{
    int i;
    /* collect all the distinct keys */
    for(i=0;i<nclistlength(fraglenv);i+=2) {
	char* key = nclistget(fraglenv,i);
	if(!lcontains(allkeys,key)) {
	    nclistpush(allkeys,key);
	}
    }
}

/* Given a fragment envv list, coalesce duplicate keys and remove duplicate values*/
static int
cleanfragments(NClist** fraglenvp)
{
    int i,stat = NC_NOERR;
    NClist*  fraglenv = NULL;
    NClist* tmp = NULL;
    NClist* allkeys = NULL;
    NClist* newlist = NULL;
    NCbytes* buf = NULL;
    char* key = NULL;
    char* value = NULL;

    if(fraglenvp == NULL || nclistlength(*fraglenvp) == 0) return NC_NOERR;
    fraglenv = *fraglenvp; /* take control of this list */
    *fraglenvp = NULL;
    newlist = nclistnew();
    buf = ncbytesnew();
    allkeys = nclistnew();
    tmp = nclistnew();

    /* collect all unique keys */
    collectallkeys(fraglenv,allkeys);
    /* Collect all values for same key across all fragment pairs */
    for(i=0;i<nclistlength(allkeys);i++) {
	key = nclistget(allkeys,i);
	collectvaluesbykey(fraglenv,key,tmp);
	/* merge the key values, remove duplicate */
	if((stat=mergelist(&tmp))) goto done;
        /* Construct key,value pair and insert into newlist */
	key = strdup(key);
	nclistpush(newlist,key);
	value = list2string(tmp);
	nclistpush(newlist,value);
	nclistclear(tmp);
    }
    *fraglenvp = newlist; newlist = NULL;
done:
    nclistfree(allkeys);
    nclistfree(tmp);
    ncbytesfree(buf);
    nclistfreeall(fraglenv);
    nclistfreeall(newlist);
    return check(stat);
}

/* process non-mode fragment keys in case they hold significance; currently not */
static int
processfragmentkeys(const char* key, const char* value, NCmodel* model)
{
    return NC_NOERR;
}

/*
Infer from the mode + useparallel
only call if iscreate or file is not easily readable.
*/
static int
NC_omodeinfer(int useparallel, int cmode, NCmodel* model)
{
    int stat = NC_NOERR;

    /* If no format flags are set, then use default */
    if(!fIsSet(cmode,NC_FORMAT_ALL))
	set_default_mode(&cmode);

    /* Process the cmode; may override some already set flags. The
     * user-defined formats must be checked first. They may choose to
     * use some of the other flags, like NC_NETCDF4, so we must first
     * check NC_UDF0 and NC_UDF1 before checking for any other
     * flag. */
    if(fIsSet(cmode, NC_UDF0)  || fIsSet(cmode, NC_UDF1))
    {
        if(fIsSet(cmode, NC_UDF0))
        {
	    model->impl = NC_FORMATX_UDF0;
	} else {
	    model->impl = NC_FORMATX_UDF1;
	}
        if(fIsSet(cmode,NC_64BIT_OFFSET)) 
        {
            model->format = NC_FORMAT_64BIT_OFFSET;
        }
        else if(fIsSet(cmode,NC_64BIT_DATA))
        {
            model->format = NC_FORMAT_64BIT_DATA;
        }
        else if(fIsSet(cmode,NC_NETCDF4))
        {
            if(fIsSet(cmode,NC_CLASSIC_MODEL))
                model->format = NC_FORMAT_NETCDF4_CLASSIC;
            else
                model->format = NC_FORMAT_NETCDF4;
        }
        if(! model->format)
            model->format = NC_FORMAT_CLASSIC;
	goto done;
    }

    if(fIsSet(cmode,NC_64BIT_OFFSET)) {
	model->impl = NC_FORMATX_NC3;
	model->format = NC_FORMAT_64BIT_OFFSET;
        goto done;
    }

    if(fIsSet(cmode,NC_64BIT_DATA)) {
	model->impl = NC_FORMATX_NC3;
	model->format = NC_FORMAT_64BIT_DATA;
        goto done;
    }

    if(fIsSet(cmode,NC_NETCDF4)) {
	model->impl = NC_FORMATX_NC4;
        if(fIsSet(cmode,NC_CLASSIC_MODEL))
	    model->format = NC_FORMAT_NETCDF4_CLASSIC;
	else
	    model->format = NC_FORMAT_NETCDF4;
        goto done;
    }

    /* Default to classic model */
    model->format = NC_FORMAT_CLASSIC;
    model->impl = NC_FORMATX_NC3;

done:
    /* Apply parallel flag */
    if(useparallel) {
        if(model->impl == NC_FORMATX_NC3)
	    model->impl = NC_FORMATX_PNETCDF;
    }
    return check(stat);
}

/*
If the mode flags do not necessarily specify the
format, then default it by adding in appropriate flags.
*/

static void
set_default_mode(int* modep)
{
    int mode = *modep;
    int dfaltformat;

    dfaltformat = nc_get_default_format();
    switch (dfaltformat) {
    case NC_FORMAT_64BIT_OFFSET: mode |= NC_64BIT_OFFSET; break;
    case NC_FORMAT_64BIT_DATA: mode |= NC_64BIT_DATA; break;
    case NC_FORMAT_NETCDF4: mode |= NC_NETCDF4; break;
    case NC_FORMAT_NETCDF4_CLASSIC: mode |= (NC_NETCDF4|NC_CLASSIC_MODEL); break;
    case NC_FORMAT_CLASSIC: /* fall thru */
    default: break; /* default to classic */
    }
    *modep = mode; /* final result */
}

/**************************************************/
/*
   Infer model for this dataset using some
   combination of cmode, path, and reading the dataset.
   See the documentation in docs/internal.dox.

@param path
@param omode
@param iscreate
@param useparallel
@param params
@param model
@param newpathp
*/

int
NC_infermodel(const char* path, int* omodep, int iscreate, int useparallel, void* params, NCmodel* model, char** newpathp)
{
    int i,stat = NC_NOERR;
    NCURI* uri = NULL;
    int omode = *omodep;
    NClist* fraglenv = nclistnew();
    NClist* modeargs = nclistnew();
    char* sfrag = NULL;
    const char* modeval = NULL;
    char* abspath = NULL;

    /* Phase 1:
       1. convert special protocols to http|https
       2. begin collecting fragments
    */
    if((stat = processuri(path, &uri, fraglenv))) goto done;

    if(uri != NULL) {
#ifdef DEBUG
	printlist(fraglenv,"processuri");
#endif

        /* Phase 2: Expand macros and add to fraglenv */
        if((stat = processmacros(&fraglenv))) goto done;
#ifdef DEBUG
	printlist(fraglenv,"processmacros");
#endif

	/* Cleanup the fragment list */
	if((stat = cleanfragments(&fraglenv))) goto done;

        /* Phase 2a: Expand mode inferences and add to fraglenv */
        if((stat = processinferences(fraglenv))) goto done;
#ifdef DEBUG
	printlist(fraglenv,"processinferences");
#endif

        /* Phase 3: coalesce duplicate fragment keys and remove duplicate values */
        if((stat = cleanfragments(&fraglenv))) goto done;
#ifdef DEBUG
	printlist(fraglenv,"cleanfragments");
#endif

        /* Phase 4: Rebuild the url fragment and rebuilt the url */
        sfrag = envvlist2string(fraglenv,"&");
        nclistfreeall(fraglenv); fraglenv = NULL;
#ifdef DEBUG
	fprintf(stderr,"frag final: %s\n",sfrag);
#endif
        ncurisetfragments(uri,sfrag);
        nullfree(sfrag); sfrag = NULL;

	/* If s3, then rebuild the url */
	if(NC_iss3(uri)) {
	    NCURI* newuri = NULL;
	    if((stat = NC_s3urlrebuild(uri,&newuri,NULL,NULL))) goto done;
	    ncurifree(uri);
	    uri = newuri;
	} else if(strcmp(uri->protocol,"file")==0) {
            /* convert path to absolute */
	    char* canon = NULL;
	    abspath = NCpathabsolute(uri->path);
	    if((stat = NCpathcanonical(abspath,&canon))) goto done;
	    nullfree(abspath);
	    abspath = canon; canon = NULL;
	    if((stat = ncurisetpath(uri,abspath))) goto done;
	}
	
	/* rebuild the path */
        if(newpathp) {
            *newpathp = ncuribuild(uri,NULL,NULL,NCURIALL);
#ifdef DEBUG
	    fprintf(stderr,"newpath=|%s|\n",*newpathp); fflush(stderr);
#endif    
	}

        /* Phase 5: Process the mode key to see if we can tell the formatx */
        modeval = ncurifragmentlookup(uri,"mode");
        if(modeval != NULL) {
	    if((stat = parseonchar(modeval,',',modeargs))) goto done;
            for(i=0;i<nclistlength(modeargs);i++) {
        	const char* arg = nclistget(modeargs,i);
        	if((stat=processmodearg(arg,model))) goto done;
            }
	}

        /* Phase 6: Process the non-mode keys to see if we can tell the formatx */
	if(!modelcomplete(model)) {
	    const char** p = ncurifragmentparams(uri); /* envv format */
	    if(p != NULL) {
	        for(;*p;p+=2) {
		    const char* key = p[0];
		    const char* value = p[1];;
        	    if((stat=processfragmentkeys(key,value,model))) goto done;
	        }
	    }
	}

        /* Phase 7: Special cases: if this is a URL and model.impl is still not defined */
        /* Phase7a: Default is DAP2 */
        if(!modelcomplete(model)) {
	    model->impl = NC_FORMATX_DAP2;
	    model->format = NC_FORMAT_NC3;
        }

    } else {/* Not URL */
	if(newpathp) *newpathp = NULL;
    }

    /* Phase 8: mode inference from mode flags */
    /* The modeargs did not give us a model (probably not a URL).
       So look at the combination of mode flags and the useparallel flag */
    if(!modelcomplete(model)) {
        if((stat = NC_omodeinfer(useparallel,omode,model))) goto done;
    }

    /* Phase 9: Infer from file content, if possible;
       this has highest precedence, so it may override
       previous decisions. Note that we do this last
       because we need previously determined model info
       to guess if this file is readable.
    */
    if(!iscreate && isreadable(uri,model)) {
	/* Ok, we need to try to read the file */
	if((stat = check_file_type(path, omode, useparallel, params, model, uri))) goto done;
    }

    /* Need a decision */
    if(!modelcomplete(model))
	{stat = NC_ENOTNC; goto done;}

    /* Force flag consistency */
    switch (model->impl) {
    case NC_FORMATX_NC4:
    case NC_FORMATX_NC_HDF4:
    case NC_FORMATX_DAP4:
    case NC_FORMATX_NCZARR:
	omode |= NC_NETCDF4;
	if(model->format == NC_FORMAT_NETCDF4_CLASSIC)
	    omode |= NC_CLASSIC_MODEL;
	break;
    case NC_FORMATX_NC3:
	omode &= ~NC_NETCDF4; /* must be netcdf-3 (CDF-1, CDF-2, CDF-5) */
	if(model->format == NC_FORMAT_64BIT_OFFSET) omode |= NC_64BIT_OFFSET;
	else if(model->format == NC_FORMAT_64BIT_DATA) omode |= NC_64BIT_DATA;
	break;
    case NC_FORMATX_PNETCDF:
	omode &= ~NC_NETCDF4; /* must be netcdf-3 (CDF-1, CDF-2, CDF-5) */
	if(model->format == NC_FORMAT_64BIT_OFFSET) omode |= NC_64BIT_OFFSET;
	else if(model->format == NC_FORMAT_64BIT_DATA) omode |= NC_64BIT_DATA;
	break;
    case NC_FORMATX_DAP2:
	omode &= ~(NC_NETCDF4|NC_64BIT_OFFSET|NC_64BIT_DATA|NC_CLASSIC_MODEL);
	break;
    case NC_FORMATX_UDF0:
    case NC_FORMATX_UDF1:
        if(model->format == NC_FORMAT_64BIT_OFFSET) 
            omode |= NC_64BIT_OFFSET;
        else if(model->format == NC_FORMAT_64BIT_DATA)
            omode |= NC_64BIT_DATA;
        else if(model->format == NC_FORMAT_NETCDF4)  
            omode |= NC_NETCDF4;
        else if(model->format == NC_FORMAT_NETCDF4_CLASSIC)  
            omode |= NC_NETCDF4|NC_CLASSIC_MODEL;
        break;
    default:
	{stat = NC_ENOTNC; goto done;}
    }

done:
    nullfree(sfrag);
    nullfree(abspath);
    ncurifree(uri);
    nclistfreeall(modeargs);
    nclistfreeall(fraglenv);
    *omodep = omode; /* in/out */
    return check(stat);
}

static int
isreadable(NCURI* uri, NCmodel* model)
{
    int canread = 0;
    struct Readable* r;
    /* Step 1: Look up the implementation */
    for(r=readable;r->impl;r++) {
	if(model->impl == r->impl) {canread = r->readable; break;}
    }
    /* Step 2: check for bytes mode */
    if(!canread && NC_testmode(uri,"bytes") && (model->impl == NC_FORMATX_NC4 || model->impl == NC_FORMATX_NC_HDF5))
        canread = 1;
    return canread;
}

#if 0
static char*
emptyify(char* s)
{
    if(s == NULL) s = strdup("");
    return strdup(s);
}

static const char*
nullify(const char* s)
{
    if(s != NULL && strlen(s) == 0)
        return NULL;
    return s;
}
#endif

/**************************************************/
/**************************************************/
/**
 * Provide a hidden interface to allow utilities
 * to check if a given path name is really a url.
 * If not, put null in basenamep, else put basename of the url path
 * minus any extension into basenamep; caller frees.
 * Return 1 if it looks like a url, 0 otherwise.
 */

int
nc__testurl(const char* path0, char** basenamep)
{
    NCURI* uri = NULL;
    int ok = 0;
    char* path = NULL;

    if(!ncuriparse(path0,&uri)) {
	char* p;
	char* q;
	path = strdup(uri->path);
	if(path == NULL||strlen(path)==0) goto done;
        p = strrchr(path, '/');
	if(p == NULL) p = path; else p++;
	q = strrchr(p,'.');
        if(q != NULL) *q = '\0';
	if(strlen(p) == 0) goto done;
	if(basenamep)
            *basenamep = strdup(p);
	ok = 1;
    }
done:
    ncurifree(uri);
    nullfree(path);
    return ok;
}

/**************************************************/
/* Envv list utilities */

static const char*
getmodekey(const NClist* envv)
{
    int i;
    /* Get "mode" entry */
    for(i=0;i<nclistlength(envv);i+=2) {
	char* key = NULL;
	key = nclistget(envv,i);
	if(strcasecmp(key,"mode")==0)
	    return nclistget(envv,i+1);
    }
    return NULL;
}

static int
replacemode(NClist* envv, const char* newval)
{
    int i;
    /* Get "mode" entry */
    for(i=0;i<nclistlength(envv);i+=2) {
	char* key = NULL;
	char* val = NULL;
	key = nclistget(envv,i);
	if(strcasecmp(key,"mode")==0) {
	    val = nclistget(envv,i+1);	    
	    nclistset(envv,i+1,strdup(newval));
	    nullfree(val);
	    return NC_NOERR;
	}
    }
    return NC_EINVAL;
}

static NClist*
parsemode(const char* modeval)
{
    NClist* modes = nclistnew();
    if(modeval)
        (void)parseonchar(modeval,',',modes);/* split on commas */
    return modes;    
}

/* Convert a list into a comma'd string */
static char*
list2string(NClist* list)
{
    int i;
    NCbytes* buf = NULL;
    char* result = NULL;

    if(list == NULL || nclistlength(list)==0) return strdup("");
    buf = ncbytesnew();
    for(i=0;i<nclistlength(list);i++) {
	const char* m = nclistget(list,i);
	if(m == NULL || strlen(m) == 0) continue;
	if(i > 0) ncbytescat(buf,",");
	ncbytescat(buf,m);
    }
    result = ncbytesextract(buf);
    ncbytesfree(buf);
    if(result == NULL) result = strdup("");
    return result;
}

#if 0
/* Given a comma separated string, remove duplicates; mostly used to cleanup mode list */
static char* 
cleancommalist(const char* commalist, int caseinsensitive)
{
    NClist* tmp = nclistnew();
    char* newlist = NULL;
    if(commalist == NULL || strlen(commalist)==0) return nulldup(commalist);
    (void)parseonchar(commalist,',',tmp);/* split on commas */
    cleanstringlist(tmp,caseinsensitive);
    newlist = list2string(tmp);
    nclistfreeall(tmp);
    return newlist;
}
#endif

/* Given a list of strings, remove nulls and duplicated */
static void
cleanstringlist(NClist* strs, int caseinsensitive)
{
    int i,j;
    if(nclistlength(strs) == 0) return;
    /* Remove nulls */
    for(i=nclistlength(strs)-1;i>=0;i--) {
        if(nclistget(strs,i)==NULL) nclistremove(strs,i);
    }
    /* Remove duplicates*/
    for(i=0;i<nclistlength(strs);i++) {
        const char* value = nclistget(strs,i);
	/* look ahead for duplicates */
        for(j=nclistlength(strs)-1;j>i;j--) {
	    int match;
            const char* candidate = nclistget(strs,j);
            if(caseinsensitive)
	        match = (strcasecmp(value,candidate) == 0);
	    else
		match = (strcmp(value,candidate) == 0);
	    if(match) {char* dup = nclistremove(strs,j); nullfree(dup);}
	}
    }
}


/**************************************************/
/**
 * @internal Given an existing file, figure out its format and return
 * that format value (NC_FORMATX_XXX) in model arg. Assume any path
 * conversion was already performed at a higher level.
 *
 * @param path File name.
 * @param flags
 * @param use_parallel
 * @param parameters
 * @param model Pointer that gets the model to use for the dispatch table.
 * @param version Pointer that gets version of the file.
 *
 * @return ::NC_NOERR No error.
 * @author Dennis Heimbigner
*/
static int
check_file_type(const char *path, int omode, int use_parallel,
		   void *parameters, NCmodel* model, NCURI* uri)
{
    char magic[NC_MAX_MAGIC_NUMBER_LEN];
    int status = NC_NOERR;
    struct MagicFile magicinfo;
#ifdef _WIN32
    NC* nc = NULL;
#endif

    memset((void*)&magicinfo,0,sizeof(magicinfo));

#ifdef _WIN32 /* including MINGW */
    /* Windows does not handle well multiple handles to the same file.
       So if file is already open/created, then find it and just get the
       model from that. */
    if((nc = find_in_NCList_by_name(path)) != NULL) {
	int format = 0;
	/* Get the model from this NC */
	if((status = nc_inq_format_extended(nc->ext_ncid,&format,NULL))) goto done;
	model->impl = format;
	if((status = nc_inq_format(nc->ext_ncid,&format))) goto done;
	model->format = format;
	goto done;
    }
#endif

    magicinfo.path = path; /* do not free */
    magicinfo.uri = uri; /* do not free */
    magicinfo.omode = omode;
    magicinfo.model = model; /* do not free */
    magicinfo.parameters = parameters; /* do not free */
#ifdef USE_STDIO
    magicinfo.use_parallel = 0;
#else
    magicinfo.use_parallel = use_parallel;
#endif

    if((status = openmagic(&magicinfo))) goto done;

    /* Verify we have a large enough file */
    if(magicinfo.filelen < (unsigned long long)MAGIC_NUMBER_LEN)
	{status = NC_ENOTNC; goto done;}
    if((status = readmagic(&magicinfo,0L,magic)) != NC_NOERR) {
	status = NC_ENOTNC;
	goto done;
    }

    /* Look at the magic number */
    if(NC_interpret_magic_number(magic,model) == NC_NOERR
	&& model->format != 0) {
        if (use_parallel && (model->format == NC_FORMAT_NC3 || model->impl == NC_FORMATX_NC3))
            /* this is called from nc_open_par() and file is classic */
            model->impl = NC_FORMATX_PNETCDF;
        goto done; /* found something */
    }

    /* Remaining case when implementation is an HDF5 file;
       search forward at starting at 512
       and doubling to see if we have HDF5 magic number */
    {
	long pos = 512L;
        for(;;) {
	    if((pos+MAGIC_NUMBER_LEN) > magicinfo.filelen)
		{status = NC_ENOTNC; goto done;}
            if((status = readmagic(&magicinfo,pos,magic)) != NC_NOERR)
	        {status = NC_ENOTNC; goto done; }
            NC_interpret_magic_number(magic,model);
            if(model->impl == NC_FORMATX_NC4) break;
	    /* double and try again */
	    pos = 2*pos;
        }
    }
done:
    closemagic(&magicinfo);
    return check(status);
}

/**
\internal
\ingroup datasets
Provide open, read and close for use when searching for magic numbers
*/
static int
openmagic(struct MagicFile* file)
{
    int status = NC_NOERR;

    if(fIsSet(file->omode,NC_INMEMORY)) {
	/* Get its length */
	NC_memio* meminfo = (NC_memio*)file->parameters;
        assert(meminfo != NULL);
	file->filelen = (long long)meminfo->size;
#ifdef ENABLE_BYTERANGE
    } else if(file->uri != NULL) {
#ifdef ENABLE_S3_SDK
	/* If this is an S3 URL, then handle specially */
	if(NC_iss3(file->uri)) {
	    if((status = NC_s3urlprocess(file->uri,&file->s3))) goto done;
	    if((file->s3client = NC_s3sdkcreateclient(&file->s3))==NULL) {status = NC_EURL; goto done;}
	    if((status = NC_s3sdkinfo(file->s3client,file->s3.bucket,file->s3.rootkey,&file->filelen,&file->errmsg)))
	        goto done;
	    file->iss3 = 1;
	} else
#endif
	{
	    /* Construct a URL minus any fragment */
            file->curlurl = ncuribuild(file->uri,NULL,NULL,NCURISVC);
	    /* Open the curl handle */
	    if((status=nc_http_init(&file->state))) goto done;
	    if((status=nc_http_size(file->state,file->curlurl,&file->filelen))) goto done;
	}
#endif /*BYTERANGE*/
    } else {
#ifdef USE_PARALLEL
        if (file->use_parallel) {
	    int retval;
	    MPI_Offset size;
            assert(file->parameters != NULL);
	    if((retval = MPI_File_open(((NC_MPI_INFO*)file->parameters)->comm,
                                   (char*)file->path,MPI_MODE_RDONLY,
                                   ((NC_MPI_INFO*)file->parameters)->info,
                                   &file->fh)) != MPI_SUCCESS) {
#ifdef MPI_ERR_NO_SUCH_FILE
		int errorclass;
		MPI_Error_class(retval, &errorclass);
		if (errorclass == MPI_ERR_NO_SUCH_FILE)
#ifdef NC_ENOENT
		    status = NC_ENOENT;
#else
		    status = errno;
#endif
		else
#endif
		    status = NC_EPARINIT;
		file->fh = MPI_FILE_NULL;
		goto done;
	    }
	    /* Get its length */
	    if((retval=MPI_File_get_size(file->fh, &size)) != MPI_SUCCESS)
	        {status = NC_EPARINIT; goto done;}
	    file->filelen = (long long)size;
	} else
#endif /* USE_PARALLEL */
	{
            if (file->path == NULL || strlen(file->path) == 0)
                {status = NC_EINVAL; goto done;}
            file->fp = NCfopen(file->path, "r");
   	    if(file->fp == NULL)
	        {status = errno; goto done;}
  	    /* Get its length */
	    {
	        int fd = fileno(file->fp);
#ifdef _WIN32
		__int64 len64 = _filelengthi64(fd);
		if(len64 < 0)
		    {status = errno; goto done;}
		file->filelen = (long long)len64;
#else
		off_t size;
		size = lseek(fd, 0, SEEK_END);
		if(size == -1)
		    {status = errno; goto done;}
		file->filelen = (long long)size;
#endif
	    }
	    rewind(file->fp);
	  }
    }
done:
    return check(status);
}

static int
readmagic(struct MagicFile* file, long pos, char* magic)
{
    int status = NC_NOERR;
    NCbytes* buf = ncbytesnew();

    memset(magic,0,MAGIC_NUMBER_LEN);
    if(fIsSet(file->omode,NC_INMEMORY)) {
	char* mempos;
	NC_memio* meminfo = (NC_memio*)file->parameters;
	if((pos + MAGIC_NUMBER_LEN) > meminfo->size)
	    {status = NC_EINMEMORY; goto done;}
	mempos = ((char*)meminfo->memory) + pos;
	memcpy((void*)magic,mempos,MAGIC_NUMBER_LEN);
#ifdef DEBUG
	printmagic("XXX: readmagic",magic,file);
#endif
#ifdef ENABLE_BYTERANGE
    } else if(file->uri != NULL) {
	fileoffset_t start = (size_t)pos;
	fileoffset_t count = MAGIC_NUMBER_LEN;
#ifdef ENABLE_S3_SDK
	if(file->iss3) {
	    if((status = NC_s3sdkread(file->s3client,file->s3.bucket,file->s3.rootkey,start,count,(void*)magic,&file->errmsg)))
	        {goto done;}
    }
    else
#endif
    {
        status = nc_http_read(file->state, file->curlurl, start, count, buf);
        if (status == NC_NOERR) {
            if (ncbyteslength(buf) != count)
                status = NC_EINVAL;
            else
                memcpy(magic, ncbytescontents(buf), count);
        }
    }
#endif
    } else {
#ifdef USE_PARALLEL
        if (file->use_parallel) {
	    MPI_Status mstatus;
	    int retval;
	    if((retval = MPI_File_read_at_all(file->fh, pos, magic,
			    MAGIC_NUMBER_LEN, MPI_CHAR, &mstatus)) != MPI_SUCCESS)
	        {status = NC_EPARINIT; goto done;}
        }
        else
#endif /* USE_PARALLEL */
        { /* Ordinary read */
            long i;
            i = fseek(file->fp, pos, SEEK_SET);
            if (i < 0) { status = errno; goto done; }
            ncbytessetlength(buf, 0);
            if ((status = NC_readfileF(file->fp, buf, MAGIC_NUMBER_LEN))) goto done;
            memcpy(magic, ncbytescontents(buf), MAGIC_NUMBER_LEN);
        }
    }

done:
    ncbytesfree(buf);
    if(file && file->fp) clearerr(file->fp);
    return check(status);
}

/**
 * Close the file opened to check for magic number.
 *
 * @param file pointer to the MagicFile struct for this open file.
 * @returns NC_NOERR for success
 * @returns NC_EPARINIT if there was a problem closing file with MPI
 * (parallel builds only).
 * @author Dennis Heimbigner
 */
static int
closemagic(struct MagicFile* file)
{
    int status = NC_NOERR;

    if(fIsSet(file->omode,NC_INMEMORY)) {
	/* noop */
#ifdef ENABLE_BYTERANGE
    } else if(file->uri != NULL) {
#ifdef ENABLE_S3_SDK
	if(file->iss3) {
	    NC_s3sdkclose(file->s3client, &file->s3, 0, &file->errmsg);
	    NC_s3clear(&file->s3);
	    nullfree(file->errmsg);
	} else
#endif
	{
	    status = nc_http_close(file->state);
	    nullfree(file->curlurl);
	}
#endif
    } else {
#ifdef USE_PARALLEL
        if (file->use_parallel) {
	    int retval;
	    if(file->fh != MPI_FILE_NULL
	       && (retval = MPI_File_close(&file->fh)) != MPI_SUCCESS)
		    {status = NC_EPARINIT; return status;}
        } else
#endif
        {
	    if(file->fp) fclose(file->fp);
        }
    }
    return status;
}

/*!
  Interpret the magic number found in the header of a netCDF file.
  This function interprets the magic number/string contained in the header of a netCDF file and sets the appropriate NC_FORMATX flags.

  @param[in] magic Pointer to a character array with the magic number block.
  @param[out] model Pointer to an integer to hold the corresponding netCDF type.
  @param[out] version Pointer to an integer to hold the corresponding netCDF version.
  @returns NC_NOERR if a legitimate file type found
  @returns NC_ENOTNC otherwise

\internal
\ingroup datasets

*/
static int
NC_interpret_magic_number(char* magic, NCmodel* model)
{
    int status = NC_NOERR;
    int tmpimpl = 0;
    /* Look at the magic number */
    if(model->impl == NC_FORMATX_UDF0 || model->impl == NC_FORMATX_UDF1)
        tmpimpl = model->impl;

    /* Use the complete magic number string for HDF5 */
    if(memcmp(magic,HDF5_SIGNATURE,sizeof(HDF5_SIGNATURE))==0) {
	model->impl = NC_FORMATX_NC4;
	model->format = NC_FORMAT_NETCDF4;
	goto done;
    }
    if(magic[0] == '\016' && magic[1] == '\003'
              && magic[2] == '\023' && magic[3] == '\001') {
	model->impl = NC_FORMATX_NC_HDF4;
	model->format = NC_FORMAT_NETCDF4;
	goto done;
    }
    if(magic[0] == 'C' && magic[1] == 'D' && magic[2] == 'F') {
        if(magic[3] == '\001') {
	    model->impl = NC_FORMATX_NC3;
	    model->format = NC_FORMAT_CLASSIC;
	    goto done;
	}
        if(magic[3] == '\002') {
	    model->impl = NC_FORMATX_NC3;
	    model->format = NC_FORMAT_64BIT_OFFSET;
	    goto done;
        }
        if(magic[3] == '\005') {
	  model->impl = NC_FORMATX_NC3;
	  model->format = NC_FORMAT_64BIT_DATA;
	  goto done;
	}
     }
     /* No match  */
     if (!tmpimpl) 
         status = NC_ENOTNC;         

     goto done;

done:
     /* if model->impl was UDF0 or UDF1 on entry, make it so on exit */
     if(tmpimpl)
         model->impl = tmpimpl;
     /* if this is a UDF magic_number update the model->impl */
     if (strlen(UDF0_magic_number) && !strncmp(UDF0_magic_number, magic,
                                               strlen(UDF0_magic_number)))
     {
         model->impl = NC_FORMATX_UDF0;
         status = NC_NOERR;
     }
     if (strlen(UDF1_magic_number) && !strncmp(UDF1_magic_number, magic,
                                               strlen(UDF1_magic_number)))
     {
         model->impl = NC_FORMATX_UDF1;
         status = NC_NOERR;
     }    

     return check(status);
}

#ifdef DEBUG
static void
printmagic(const char* tag, char* magic, struct MagicFile* f)
{
    int i;
    fprintf(stderr,"%s: ispar=%d magic=",tag,f->use_parallel);
    for(i=0;i<MAGIC_NUMBER_LEN;i++) {
        unsigned int c = (unsigned int)magic[i];
	c = c & 0x000000FF;
	if(c == '\n')
	    fprintf(stderr," 0x%0x/'\\n'",c);
	else if(c == '\r')
	    fprintf(stderr," 0x%0x/'\\r'",c);
	else if(c < ' ')
	    fprintf(stderr," 0x%0x/'?'",c);
	else
	    fprintf(stderr," 0x%0x/'%c'",c,c);
    }
    fprintf(stderr,"\n");
    fflush(stderr);
}

static void
printlist(NClist* list, const char* tag)
{
    int i;
    fprintf(stderr,"%s:",tag);
    for(i=0;i<nclistlength(list);i++) {
        fprintf(stderr," %s",(char*)nclistget(list,i));
	fprintf(stderr,"[%p]",(char*)nclistget(list,i));
    }
    fprintf(stderr,"\n");
    dbgflush();
}


#endif