This file is indexed.

/usr/lib/python2.7/dist-packages/spf.py is in python-spf 2.0.12t-3.

This file is owned by root:root, with mode 0o644.

The actual contents of the file can be viewed below.

   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
#!/usr/bin/python3
from __future__ import print_function
"""SPF (Sender Policy Framework) implementation.

Copyright (c) 2003 Terence Way <terry@wayforward.net>
Portions Copyright(c) 2004,2005,2006,2007,2008,2011,2012 Stuart Gathman <stuart@bmsi.com>
Portions Copyright(c) 2005,2006,2007,2008,2011,2012,2013,2014 Scott Kitterman <scott@kitterman.com>
Portions Copyright(c) 2013,2014 Stuart Gathman <stuart@gathman.org>

This module is free software, and you may redistribute it and/or modify
it under the same terms as Python itself, so long as this copyright message
and disclaimer are retained in their original form.

IN NO EVENT SHALL THE AUTHOR BE LIABLE TO ANY PARTY FOR DIRECT, INDIRECT,
SPECIAL, INCIDENTAL, OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE USE OF
THIS CODE, EVEN IF THE AUTHOR HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH
DAMAGE.

THE AUTHOR SPECIFICALLY DISCLAIMS ANY WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A
PARTICULAR PURPOSE.  THE CODE PROVIDED HEREUNDER IS ON AN "AS IS" BASIS,
AND THERE IS NO OBLIGATION WHATSOEVER TO PROVIDE MAINTENANCE,
SUPPORT, UPDATES, ENHANCEMENTS, OR MODIFICATIONS.

For more information about SPF, a tool against email forgery, see
    http://www.openspf.net/

For news, bugfixes, etc. visit the home page for this implementation at
    http://cheeseshop.python.org/pypi/pyspf/
    http://sourceforge.net/projects/pymilter/
    http://www.wayforward.net/spf/
"""

# CVS Commits since last release (2.0.11):
# $Log: spf.py,v $
# Revision 1.108.2.150  2015/08/05 13:49:45  customdesigned
# Forgot tabnanny
#
# Revision 1.108.2.149  2015/08/05 13:07:09  customdesigned
# Release 2.0.12
#
# Revision 1.108.2.148  2015/08/05 04:49:48  customdesigned
# Reset void_lookups at top of check()
#
# Revision 1.108.2.147  2015/08/05 03:36:59  customdesigned
# Ignore permerror for best_guess
#
# Revision 1.108.2.146  2015/06/05 15:58:18  customdesigned
# Don't crash on null TXT record.
#
# Revision 1.108.2.145  2015/01/14 20:27:42  customdesigned
# Fix list feature
#
# Revision 1.108.2.144  2015/01/13 04:40:07  customdesigned
# Trailing spaces *are* allowed by 4.5/2
#
# Revision 1.108.2.143  2015/01/12 22:51:56  customdesigned
# Trailing space is PermError, but strip for extended result in lax mode.
#
# Revision 1.108.2.142  2015/01/06 14:13:50  customdesigned
# Make CNAME loop result in unknown host.
#
# Revision 1.108.2.141  2015/01/02 01:08:18  customdesigned
# Test case and fix for mixed case CNAME loop.
#
# Revision 1.108.2.140  2015/01/02 00:26:08  customdesigned
# Make CNAME loop check case insensitive.
#
# Revision 1.108.2.139  2014/12/19 00:16:12  kitterma
# Missed a spot bumping to 2.0.12.
#
# Revision 1.108.2.138  2014/12/19 00:15:12  kitterma
# Bump versions, etc. to start 2.0.12 development.
#
# Revision 1.108.2.137  2014/12/13 15:39:27  customdesigned
# Require ipaddress/ipaddr backport with Bytes for python2.
#
# See pyspf_changelog.txt for earlier CVS commits.

__author__ = "Terence Way, Stuart Gathman, Scott Kitterman"
__email__ = "pyspf@openspf.org"
__version__ = "2.0.12: Aug 5, 2015"
MODULE = 'spf'

USAGE = """To check an incoming mail request:
    % python spf.py [-v] {ip} {sender} {helo}
    % python spf.py 69.55.226.139 tway@optsw.com mx1.wayforward.net

To test an SPF record:
    % python spf.py [-v] "v=spf1..." {ip} {sender} {helo}
    % python spf.py "v=spf1 +mx +ip4:10.0.0.1 -all" 10.0.0.1 tway@foo.com a    

To fetch an SPF record:
    % python spf.py {domain}
    % python spf.py wayforward.net

To test this script (and to output this usage message):
    % python spf.py
"""

import re
import sys
import socket  # for inet_ntoa() and inet_aton()
import struct  # for pack() and unpack()
import time    # for time()
try:
    import urllib.parse as urllibparse # for quote()
except:
    import urllib as urllibparse
import sys     # for version_info()
from functools import reduce
try:
    from email.message import Message
except ImportError:
    from email.Message import Message
try:
    # Python standard libarary as of python3.3
    import ipaddress
    if bytes is str:
      from ipaddress import Bytes
except ImportError:
    try:
        import ipaddr as ipaddress
        from ipaddr import Bytes
    except ImportError:
        print('ipaddr module required: http://code.google.com/p/ipaddr-py/')

import DNS    # http://pydns.sourceforge.net
if not hasattr(DNS.Type, 'SPF'):
    # patch in type99 support
    DNS.Type.SPF = 99
    DNS.Type.typemap[99] = 'SPF'
    DNS.Lib.RRunpacker.getSPFdata = DNS.Lib.RRunpacker.getTXTdata

def DNSLookup(name, qtype, strict=True, timeout=20):
    try:
        req = DNS.DnsRequest(name, qtype=qtype, timeout=timeout)
        resp = req.req()
        #resp.show()
        # key k: ('wayforward.net', 'A'), value v
        # FIXME: pydns returns AAAA RR as 16 byte binary string, but
        # A RR as dotted quad.  For consistency, this driver should
        # return both as binary string.
        #
        if resp.header['tc'] == True:
            if strict > 1:
                raise AmbiguityWarning('DNS: Truncated UDP Reply, SPF records should fit in a UDP packet, retrying TCP')
            try:
                req = DNS.DnsRequest(name, qtype=qtype, protocol='tcp', timeout=(timeout))
                resp = req.req()
            except DNS.DNSError as x:
                raise TempError('DNS: TCP Fallback error: ' + str(x))
            if resp.header['rcode'] != 0 and resp.header['rcode'] != 3:
                raise IOError('Error: ' + resp.header['status'] + '  RCODE: ' + str(resp.header['rcode']))
        return [((a['name'], a['typename']), a['data'])
                for a in resp.answers] \
             + [((a['name'], a['typename']), a['data'])
                for a in resp.additional]
    except AttributeError as x:
        raise TempError('DNS ' + str(x))
    except IOError as x:
        raise TempError('DNS ' + str(x))
    except DNS.DNSError as x:
        raise TempError('DNS ' + str(x))

RE_SPF = re.compile(br'^v=spf1$|^v=spf1 ',re.IGNORECASE)

# Regular expression to look for modifiers
RE_MODIFIER = re.compile(r'^([a-z][a-z0-9_\-\.]*)=', re.IGNORECASE)

# Regular expression to find macro expansions
PAT_CHAR = r'%(%|_|-|(\{[^\}]*\}))'
RE_CHAR = re.compile(PAT_CHAR)
RE_INVALID_MACRO = re.compile(r'(?<!%)%[^{%_-]|%$')

# Regular expression to break up a macro expansion
RE_ARGS = re.compile(r'([0-9]*)(r?)([^0-9a-zA-Z]*)')

RE_DUAL_CIDR = re.compile(r'//(0|[1-9]\d*)$')
RE_CIDR = re.compile(r'/(0|[1-9]\d*)$')

PAT_IP4 = r'\.'.join([r'(?:\d|[1-9]\d|1\d\d|2[0-4]\d|25[0-5])']*4)
RE_IP4 = re.compile(PAT_IP4+'$')

RE_TOPLAB = re.compile(
    r'\.(?:[0-9a-z]*[a-z][0-9a-z]*|[0-9a-z]+-[0-9a-z-]*[0-9a-z])\.?$|%s'
        % PAT_CHAR, re.IGNORECASE)

RE_DOT_ATOM = re.compile(r'%(atext)s+([.]%(atext)s+)*$' % {
    'atext': r"[0-9a-z!#$%&'*+/=?^_`{}|~-]" }, re.IGNORECASE)

# Derived from RFC 3986 appendix A
RE_IP6 = re.compile(                 '(?:%(hex4)s:){6}%(ls32)s$'
                   '|::(?:%(hex4)s:){5}%(ls32)s$'
                  '|(?:%(hex4)s)?::(?:%(hex4)s:){4}%(ls32)s$'
    '|(?:(?:%(hex4)s:){0,1}%(hex4)s)?::(?:%(hex4)s:){3}%(ls32)s$'
    '|(?:(?:%(hex4)s:){0,2}%(hex4)s)?::(?:%(hex4)s:){2}%(ls32)s$'
    '|(?:(?:%(hex4)s:){0,3}%(hex4)s)?::%(hex4)s:%(ls32)s$'
    '|(?:(?:%(hex4)s:){0,4}%(hex4)s)?::%(ls32)s$'
    '|(?:(?:%(hex4)s:){0,5}%(hex4)s)?::%(hex4)s$'
    '|(?:(?:%(hex4)s:){0,6}%(hex4)s)?::$'
  % {
    'ls32': r'(?:[0-9a-f]{1,4}:[0-9a-f]{1,4}|%s)'%PAT_IP4,
    'hex4': r'[0-9a-f]{1,4}'
    }, re.IGNORECASE)

# Local parts and senders have their delimiters replaced with '.' during
# macro expansion
#
JOINERS = {'l': '.', 's': '.'}

RESULTS = {'+': 'pass', '-': 'fail', '?': 'neutral', '~': 'softfail',
           'pass': 'pass', 'fail': 'fail', 'permerror': 'permerror',
       'error': 'temperror', 'neutral': 'neutral', 'softfail': 'softfail',
       'none': 'none', 'local': 'local', 'trusted': 'trusted',
           'ambiguous': 'ambiguous', 'unknown': 'permerror' }

EXPLANATIONS = {'pass': 'sender SPF authorized',
                'fail': 'SPF fail - not authorized',
                'permerror': 'permanent error in processing',
                'temperror': 'temporary DNS error in processing',
        'softfail': 'domain owner discourages use of this host',
        'neutral': 'access neither permitted nor denied',
        'none': '',
                #Note: The following are not formally SPF results
                'local': 'No SPF result due to local policy',
                'trusted': 'No SPF check - trusted-forwarder.org',
                #Ambiguous only used in harsh mode for SPF validation
                'ambiguous': 'No error, but results may vary'
        }

DELEGATE = None

# standard default SPF record for best_guess
DEFAULT_SPF = 'v=spf1 a/24 mx/24 ptr'

#Whitelisted forwarders here.  Additional locally trusted forwarders can be
#added to this record.
TRUSTED_FORWARDERS = 'v=spf1 ?include:spf.trusted-forwarder.org -all'

# maximum DNS lookups allowed
MAX_LOOKUP = 10 #RFC 4408 Para 10.1/RFC 7208 4.6.4
MAX_MX = 10 #RFC 4408 Para 10.1/RFC 7208 4.6.4
MAX_PTR = 10 #RFC 4408 Para 10.1/RFC 7208 4.6.4
MAX_CNAME = 10 # analogous interpretation to MAX_PTR
MAX_RECURSION = 20
MAX_PER_LOOKUP_TIME = 20 # Per RFC 7208 4.6.4
MAX_VOID_LOOKUPS = 2 # RFC 7208 4.6.4

ALL_MECHANISMS = ('a', 'mx', 'ptr', 'exists', 'include', 'ip4', 'ip6', 'all')
COMMON_MISTAKES = {
  'prt': 'ptr', 'ip': 'ip4', 'ipv4': 'ip4', 'ipv6': 'ip6', 'all.': 'all'
}

#If harsh processing, for the validator, is invoked, warn if results
#likely deviate from the publishers intention.
class AmbiguityWarning(Exception):
    "SPF Warning - ambiguous results"
    def __init__(self, msg, mech=None, ext=None):
        Exception.__init__(self, msg, mech)
        self.msg = msg
        self.mech = mech
        self.ext = ext
    def __str__(self):
        if self.mech:
            return '%s: %s' %(self.msg, self.mech)
        return self.msg

class TempError(Exception):
    "Temporary SPF error"
    def __init__(self, msg, mech=None, ext=None):
        Exception.__init__(self, msg, mech)
        self.msg = str(msg)
        self.mech = mech
        self.ext = ext
    def __str__(self):
        if self.mech:
            return '%s: %s'%(self.msg, self.mech)
        return self.msg

class PermError(Exception):
    "Permanent SPF error"
    def __init__(self, msg, mech=None, ext=None):
        Exception.__init__(self, msg, mech)
        self.msg = msg
        self.mech = mech
        self.ext = ext
    def __str__(self):
        if self.mech:
            return '%s: %s'%(self.msg, self.mech)
        return self.msg

def check2(i, s, h, local=None, receiver=None, timeout=MAX_PER_LOOKUP_TIME, verbose=False, querytime=20):
    """Test an incoming MAIL FROM:<s>, from a client with ip address i.
    h is the HELO/EHLO domain name.  This is the RFC4408/7208 compliant
    pySPF2.0 interface.  The interface returns an SPF result and explanation
    only.  SMTP response codes are not returned since neither RFC 4408 nor RFC
    7208 does specify receiver policy.  Applications updated for RFC 4408 and
    RFC 7208 should use this interface.  The maximum time, in seconds, this
    function is allowed to run before a TempError is returned is controlled by
    querytime.  When set to 0 the timeout parameter (default 20 seconds)
    controls the time allowed for each DNS lookup.  When set to a non-zero
    value, it total time for all processing related to the SPF check is
    limited to querytime (default 20 seconds as recommended in RFC 7208,
    paragraph 4.6.4).

    Returns (result, explanation) where result in
    ['pass', 'permerror', 'fail', 'temperror', 'softfail', 'none', 'neutral' ].

    Example:
    #>>> check2(i='61.51.192.42', s='liukebing@bcc.com', h='bmsi.com')

    """
    res,_,exp = query(i=i, s=s, h=h, local=local,
        receiver=receiver,timeout=timeout,verbose=verbose,querytime=querytime).check()
    return res,exp

def check(i, s, h, local=None, receiver=None, verbose=False):
    """Test an incoming MAIL FROM:<s>, from a client with ip address i.
    h is the HELO/EHLO domain name.  This is the pre-RFC SPF Classic interface.
    Applications written for pySPF 1.6/1.7 can use this interface to allow
    pySPF2 to be a drop in replacement for older versions.  With the exception
    of result codes, performance in RFC 4408 compliant.

    Returns (result, code, explanation) where result in
    ['pass', 'unknown', 'fail', 'error', 'softfail', 'none', 'neutral' ].

    Example:
    #>>> check(i='61.51.192.42', s='liukebing@bcc.com', h='bmsi.com')

    """
    res,code,exp = query(i=i, s=s, h=h, local=local, receiver=receiver,
        verbose=verbose).check()
    if res == 'permerror':
        res = 'unknown'
    elif res == 'tempfail':
        res =='error'
    return res, code, exp

class query(object):
    """A query object keeps the relevant information about a single SPF
    query:

    i: ip address of SMTP client in dotted notation
    s: sender declared in MAIL FROM:<>
    l: local part of sender s
    d: current domain, initially domain part of sender s
    h: EHLO/HELO domain
    v: 'in-addr' for IPv4 clients and 'ip6' for IPv6 clients
    t: current timestamp
    p: SMTP client domain name
    o: domain part of sender s
    r: receiver
    c: pretty ip address (different from i for IPv6)

    This is also, by design, the same variables used in SPF macro
    expansion.

    Also keeps cache: DNS cache.  
    """
    def __init__(self, i, s, h, local=None, receiver=None, strict=True,
                timeout=MAX_PER_LOOKUP_TIME,verbose=False,querytime=0):
        self.s, self.h = s, h
        if not s and h:
            self.s = 'postmaster@' + h
            self.ident = 'helo'
        else:
            self.ident = 'mailfrom'
        self.l, self.o = split_email(s, h)
        self.t = str(int(time.time()))
        self.d = self.o
        self.p = None   # lazy evaluation
        if receiver:
            self.r = receiver
        else:
            self.r = 'unknown'
        # Since the cache does not track Time To Live, it is created
        # fresh for each query.  It is important for efficiently using
        # multiple results provided in DNS answers.
        self.cache = {}
        self.defexps = dict(EXPLANATIONS)
        self.exps = dict(EXPLANATIONS)
        self.libspf_local = local    # local policy
        self.lookups = 0
        # New processing limit in RFC 7208, section 4.6.4
        self.void_lookups = 0
        # strict can be False, True, or 2 (numeric) for harsh
        self.strict = strict
        self.timeout = timeout
        self.querytime = querytime # Default to not using a global check
                                   # timelimit since this is an RFC 4408 MAY
        if querytime > 0:
            self.timeout = querytime
        self.timer = 0
        self.ipaddr = None
        if i:
            self.set_ip(i)
        # Document bits of the object model not set up here:
        # self.i = string, expanded dot notation, suitable for PTR lookups
        # self.c = string, human readable form of the connect IP address
        # single letter lowercase variable names (e.g. self.i) are used for SPF macros
        # For IPv4, self.i = self.c, but not in IPv6
        # self.iplist = list of IPv4/6 addresses that would pass, collected
        #               when list or list6 is passed as 'i'
        # self.addr = ipaddr/ipaddress object representing the connect IP 
        self.default_modifier = True
        self.verbose = verbose
        self.authserv = None # Only used in A-R header generation tests

    def log(self,mech,d,spf):
        print('%s: %s "%s"'%(mech,d,spf))

    def set_ip(self, i):
        "Set connect ip, and ip6 or ip4 mode."
        self.iplist = False
        if i.lower() == 'list':
            self.iplist = []
            ip6 = False
        elif i.lower() == 'list6':
            self.iplist = []
            ip6 = True
        else:
            try:
                try:
                    self.ipaddr = ipaddress.ip_address(i)
                except AttributeError:
                    self.ipaddr = ipaddress.IPAddress(i)
            except ValueError as x:
                raise PermError(str(x))
            if self.ipaddr.version == 6:
                if self.ipaddr.ipv4_mapped:
                    self.ipaddr = ipaddress.IPv4Address(self.ipaddr.ipv4_mapped)
                    ip6 = False
                else:
                    ip6 = True
            else:
                ip6 = False
            self.c = str(self.ipaddr)
        # NOTE: self.A is not lowercase, so isn't a macro.  See query.expand()
        if ip6:
            self.A = 'AAAA'
            self.v = 'ip6'
            if self.ipaddr:
              self.i = '.'.join(list(self.ipaddr.exploded.replace(':','').upper()))
            self.cidrmax = 128
        else:
            self.A = 'A'
            self.v = 'in-addr'
            if self.ipaddr:
              self.i = self.ipaddr.exploded
            self.cidrmax = 32

    def set_default_explanation(self, exp):
        exps = self.exps
        defexps = self.defexps
        for i in 'softfail', 'fail', 'permerror':
            exps[i] = exp
            defexps[i] = exp

    def set_explanation(self, exp):
        exps = self.exps
        for i in 'softfail', 'fail', 'permerror':
            exps[i] = exp

    # Compute p macro only if needed
    def getp(self):
        if not self.p:
            p = self.validated_ptrs()
            if not p:
                self.p = "unknown"
            elif self.d in p:
                self.p = self.d
            else:
                sfx = '.' + self.d
                for d in p:
                    if d.endswith(sfx):
                        self.p = d
                        break
                else:
                    self.p = p[0]
        return self.p

    def best_guess(self, spf=DEFAULT_SPF):
        """Return a best guess based on a default SPF record.
    >>> q = query('1.2.3.4','','SUPERVISION1',receiver='example.com')
    >>> q.best_guess()[0]
    'none'
        """
        if RE_TOPLAB.split(self.d)[-1]:
            return ('none', 250, '')
        pe = self.perm_error
        r,c,e = self.check(spf)
        if r == 'permerror':	# permerror not useful for bestguess
          if self.perm_error and self.perm_error.ext:
            r,c,e = self.perm_error.ext
          else:
            r,c = 'neutral',250
          self.perm_error = pe
        return r,c,e

    def check(self, spf=None):
        """
    Returns (result, mta-status-code, explanation) where result
    in ['fail', 'softfail', 'neutral' 'permerror', 'pass', 'temperror', 'none']

    Examples:
    >>> q = query(s='strong-bad@email.example.com',
    ...           h='mx.example.org', i='192.0.2.3')
    >>> q.check(spf='v=spf1 ?all')
    ('neutral', 250, 'access neither permitted nor denied')

    >>> q.check(spf='v=spf1 redirect=controlledmail.com exp=_exp.controlledmail.com')
    ('fail', 550, 'SPF fail - not authorized')
    
    >>> q.check(spf='v=spf1 ip4:192.0.0.0/8 ?all moo')
    ('permerror', 550, 'SPF Permanent Error: Unknown mechanism found: moo')

    >>> q.check(spf='v=spf1 ip4:192.0.0.n ?all')
    ('permerror', 550, 'SPF Permanent Error: Invalid IP4 address: ip4:192.0.0.n')

    >>> q.check(spf='v=spf1 ip6:2001:db8:ZZZZ:: ?all')
    ('permerror', 550, 'SPF Permanent Error: Invalid IP6 address: ip6:2001:db8:ZZZZ::')

    >>> q.check(spf='v=spf1 =a ?all moo')
    ('permerror', 550, 'SPF Permanent Error: Unknown qualifier, RFC 4408 para 4.6.1, found in: =a')

    >>> q.check(spf='v=spf1 ip4:192.0.0.0/8 ~all')
    ('pass', 250, 'sender SPF authorized')

    >>> q.check(spf='v=spf1 ip4:192.0.0.0/8 -all moo=')
    ('pass', 250, 'sender SPF authorized')

    >>> q.check(spf='v=spf1 ip4:192.0.0.0/8 -all match.sub-domains_9=yes')
    ('pass', 250, 'sender SPF authorized')

    >>> q.strict = False
    >>> q.check(spf='v=spf1 ip4:192.0.0.0/8 -all moo')
    ('permerror', 550, 'SPF Permanent Error: Unknown mechanism found: moo')
    >>> q.perm_error.ext
    ('pass', 250, 'sender SPF authorized')

    >>> q.strict = True
    >>> q.check(spf='v=spf1 ip4:192.1.0.0/16 moo -all')
    ('permerror', 550, 'SPF Permanent Error: Unknown mechanism found: moo')

    >>> q.check(spf='v=spf1 ip4:192.1.0.0/16 ~all')
    ('softfail', 250, 'domain owner discourages use of this host')

    >>> q.check(spf='v=spf1 -ip4:192.1.0.0/6 ~all')
    ('fail', 550, 'SPF fail - not authorized')

    # Assumes DNS available
    >>> q.check()
    ('none', 250, '')

    >>> q.check(spf='v=spf1 ip4:1.2.3.4 -a:example.net -all')
    ('fail', 550, 'SPF fail - not authorized')
    >>> q.libspf_local='ip4:192.0.2.3 a:example.org'
    >>> q.check(spf='v=spf1 ip4:1.2.3.4 -a:example.net -all')
    ('pass', 250, 'sender SPF authorized')

    >>> q.check(spf='v=spf1 ip4:1.2.3.4 -all exp=_exp.controlledmail.com')
    ('fail', 550, 'Controlledmail.com does not send mail from itself.')
    
    >>> q.check(spf='v=spf1 ip4:1.2.3.4 ?all exp=_exp.controlledmail.com')
    ('neutral', 250, 'access neither permitted nor denied')
        """
        self.mech = []        # unknown mechanisms
        # If not strict, certain PermErrors (mispelled
        # mechanisms, strict processing limits exceeded)
        # will continue processing.  However, the exception
        # that strict processing would raise is saved here
        self.perm_error = None
        self.mechanism = None
        self.void_lookups = 0
        self.options = {}

        try:
            self.lookups = 0
            if not spf:
                spf = self.dns_spf(self.d)
                if self.verbose: self.log("top",self.d,spf)
            if self.libspf_local and spf: 
                spf = insert_libspf_local_policy(
                    spf, self.libspf_local)
            rc = self.check1(spf, self.d, 0)
            if self.perm_error:
                # lax processing encountered a permerror, but continued
                self.perm_error.ext = rc
                raise self.perm_error
            return rc
                
        except TempError as x:
            self.prob = x.msg
            if x.mech:
                self.mech.append(x.mech)
            return ('temperror', 451, 'SPF Temporary Error: ' + str(x))
        except PermError as x:
            if not self.perm_error:
                self.perm_error = x
            self.prob = x.msg
            if x.mech:
                self.mech.append(x.mech)
            # Pre-Lentczner draft treats this as an unknown result
            # and equivalent to no SPF record.
            return ('permerror', 550, 'SPF Permanent Error: ' + str(x))

    def check1(self, spf, domain, recursion):
        # spf rfc: 3.7 Processing Limits
        #
        if recursion > MAX_RECURSION:
            # This should never happen in strict mode
            # because of the other limits we check,
            # so if it does, there is something wrong with
            # our code.  It is not a PermError because there is not
            # necessarily anything wrong with the SPF record.
            if self.strict:
                raise AssertionError('Too many levels of recursion')
            # As an extended result, however, it should be
            # a PermError.
            raise PermError('Too many levels of recursion')
        try:
            try:
                tmp, self.d = self.d, domain
                return self.check0(spf, recursion)
            finally:
                self.d = tmp
        except AmbiguityWarning as x:
            self.prob = x.msg
            if x.mech:
                self.mech.append(x.mech)
            return ('ambiguous', 000, 'SPF Ambiguity Warning: %s' % x)

    def note_error(self, *msg):
        if self.strict:
            raise PermError(*msg)
        # if lax mode, note error and continue
        if not self.perm_error:
            try:
                raise PermError(*msg)
            except PermError as x:
                # FIXME: keep a list of errors for even friendlier diagnostics.
                self.perm_error = x
        return self.perm_error

    def expand_domain(self,arg):
        "validate and expand domain-spec"
        # any trailing dot was removed by expand()
        if RE_TOPLAB.split(arg)[-1]:
            raise PermError('Invalid domain found (use FQDN)', arg)
        return self.expand(arg)

    def validate_mechanism(self, mech):
        """Parse and validate a mechanism.
    Returns mech,m,arg,cidrlength,result

    Examples:
    >>> q = query(s='strong-bad@email.example.com.',
    ...           h='mx.example.org', i='192.0.2.3')
    >>> q.validate_mechanism('A')
    ('A', 'a', 'email.example.com', 32, 'pass')

    >>> q = query(s='strong-bad@email.example.com',
    ...           h='mx.example.org', i='192.0.2.3')
    >>> q.validate_mechanism('A//64')
    ('A//64', 'a', 'email.example.com', 32, 'pass')

    >>> q.validate_mechanism('A/24//64')
    ('A/24//64', 'a', 'email.example.com', 24, 'pass')
    
    >>> q.validate_mechanism('?mx:%{d}/27')
    ('?mx:%{d}/27', 'mx', 'email.example.com', 27, 'neutral')

    >>> try: q.validate_mechanism('ip4:1.2.3.4/247')
    ... except PermError as x: print(x)
    Invalid IP4 CIDR length: ip4:1.2.3.4/247
    
    >>> try: q.validate_mechanism('ip4:1.2.3.4/33')
    ... except PermError as x: print(x)
    Invalid IP4 CIDR length: ip4:1.2.3.4/33

    >>> try: q.validate_mechanism('a:example.com:8080')
    ... except PermError as x: print(x)
    Invalid domain found (use FQDN): example.com:8080
    
    >>> try: q.validate_mechanism('ip4:1.2.3.444/24')
    ... except PermError as x: print(x)
    Invalid IP4 address: ip4:1.2.3.444/24
    
    >>> try: q.validate_mechanism('ip4:1.2.03.4/24')
    ... except PermError as x: print(x)
    Invalid IP4 address: ip4:1.2.03.4/24
    
    >>> try: q.validate_mechanism('-all:3030')
    ... except PermError as x: print(x)
    Invalid all mechanism format - only qualifier allowed with all: -all:3030

    >>> q.validate_mechanism('-mx:%%%_/.Clara.de/27')
    ('-mx:%%%_/.Clara.de/27', 'mx', '% /.Clara.de', 27, 'fail')

    >>> q.validate_mechanism('~exists:%{i}.%{s1}.100/86400.rate.%{d}')
    ('~exists:%{i}.%{s1}.100/86400.rate.%{d}', 'exists', '192.0.2.3.com.100/86400.rate.email.example.com', 32, 'softfail')

    >>> q.validate_mechanism('a:mail.example.com.')
    ('a:mail.example.com.', 'a', 'mail.example.com', 32, 'pass')

    >>> try: q.validate_mechanism('a:mail.example.com,')
    ... except PermError as x: print(x)
    Do not separate mechnisms with commas: a:mail.example.com,

    >>> q = query(s='strong-bad@email.example.com',
    ...           h='mx.example.org', i='2001:db8:1234::face:b007')    
    >>> q.validate_mechanism('A//64')
    ('A//64', 'a', 'email.example.com', 64, 'pass')

    >>> q.validate_mechanism('A/16')
    ('A/16', 'a', 'email.example.com', 128, 'pass')

    >>> q.validate_mechanism('A/16//48')
    ('A/16//48', 'a', 'email.example.com', 48, 'pass')

    """
        if mech.endswith( "," ):
            self.note_error('Do not separate mechnisms with commas', mech)
            mech = mech[:-1]
        # a mechanism
        m, arg, cidrlength, cidr6length = parse_mechanism(mech, self.d)
        # map '?' '+' or '-' to 'neutral' 'pass' or 'fail'
        if m:
            result = RESULTS.get(m[0])
            if result:
                # eat '?' '+' or '-'
                m = m[1:]
            else:
                # default pass
                result = 'pass'
        if m in COMMON_MISTAKES:
            self.note_error('Unknown mechanism found', mech)
            m = COMMON_MISTAKES[m]

        if m == 'a' and RE_IP4.match(arg):
            x = self.note_error(
              'Use the ip4 mechanism for ip4 addresses', mech)
            m = 'ip4'


        # validate cidr and dual-cidr
        if m in ('a', 'mx'):
            if cidrlength is None:
                cidrlength = 32;
            elif cidrlength > 32:
                raise PermError('Invalid IP4 CIDR length', mech)
            if cidr6length is None:
                cidr6length = 128
            elif cidr6length > 128:
                raise PermError('Invalid IP6 CIDR length', mech)
            if self.v == 'ip6':
                cidrlength = cidr6length
        elif m == 'ip4' or RE_IP4.match(m):
            if m != 'ip4':
              self.note_error( 'Missing IP4' , mech)
              m,arg = 'ip4',m
            if cidr6length is not None:
                raise PermError('Dual CIDR not allowed', mech)
            if cidrlength is None:
                cidrlength = 32;
            elif cidrlength > 32:
                raise PermError('Invalid IP4 CIDR length', mech)
            if not RE_IP4.match(arg):
                raise PermError('Invalid IP4 address', mech)
        elif m == 'ip6':
            if cidr6length is not None:
                raise PermError('Dual CIDR not allowed', mech)
            if cidrlength is None:
                cidrlength = 128
            elif cidrlength > 128:
                raise PermError('Invalid IP6 CIDR length', mech)
            if not RE_IP6.match(arg):
                raise PermError('Invalid IP6 address', mech)
        else:
            if cidrlength is not None or cidr6length is not None:
              if m in ALL_MECHANISMS:
                raise PermError('CIDR not allowed', mech)
            cidrlength = self.cidrmax

        if m in ('a', 'mx', 'ptr', 'exists', 'include'):
            if m == 'exists' and not arg:
                raise PermError('implicit exists not allowed', mech)
            arg = self.expand_domain(arg)
            if not arg:
                raise PermError('empty domain:',mech)
            if m == 'include':
                if arg == self.d:
                    if mech != 'include':
                        raise PermError('include has trivial recursion', mech)
                    raise PermError('include mechanism missing domain', mech)
            return mech, m, arg, cidrlength, result

        # validate 'all' mechanism per RFC 4408 ABNF
        if m == 'all' and mech.count(':'):
            # print '|'+ arg + '|', mech, self.d,
            self.note_error(
            'Invalid all mechanism format - only qualifier allowed with all'
              , mech)
        if m in ALL_MECHANISMS:
            return mech, m, arg, cidrlength, result
        if m[1:] in ALL_MECHANISMS:
            x = self.note_error(
                'Unknown qualifier, RFC 4408 para 4.6.1, found in', mech)
        else:
            x = self.note_error('Unknown mechanism found', mech)
        return mech, m, arg, cidrlength, x

    def check0(self, spf, recursion):
        """Test this query information against SPF text.

        Returns (result, mta-status-code, explanation) where
        result in ['fail', 'unknown', 'pass', 'none']
        """

        if not spf:
            return ('none', 250, EXPLANATIONS['none'])

        # Split string by space, drop the 'v=spf1'.  Split by all whitespace
        # casuses things like carriage returns being treated as valid space
        # separators, so split() is not sufficient.  
        spf = spf.split(' ')
        # Catch case where SPF record has no spaces.
        # Can never happen with conforming dns_spf(), however
        # in the future we might want to give warnings
        # for common mistakes like IN TXT "v=spf1" "mx" "-all"
        # in relaxed mode.
        if spf[0].lower() != 'v=spf1':
            if self.strict > 1:
                raise AmbiguityWarning('Invalid SPF record in', self.d)
            return ('none', 250, EXPLANATIONS['none'])
        # Just to make it even more fun, the relevant piece of the ABNF for
        # term separations is *( 1*SP ( directive / modifier ) ), so it's one
        # or more spaces, not just one.  So strip empty mechanisms.
        spf = [mech for mech in spf[1:] if mech]

        # copy of explanations to be modified by exp=
        exps = self.exps
        redirect = None

        # no mechanisms at all cause unknown result, unless
        # overridden with 'default=' modifier
        #
        default = 'neutral'
        mechs = []

        modifiers = []
        # Look for modifiers
        #
        for mech in spf:
            m = RE_MODIFIER.split(mech)[1:]
            if len(m) != 2:
                mechs.append(self.validate_mechanism(mech))
                continue

            mod,arg = m
            if mod in modifiers:
                if mod == 'redirect':
                    raise PermError('redirect= MUST appear at most once',mech)
                self.note_error('%s= MUST appear at most once'%mod,mech)
                # just use last one in lax mode
            modifiers.append(mod)
            if mod == 'exp':
                # always fetch explanation to check permerrors
                if not arg:
                    raise PermError('exp has empty domain-spec:',arg)
                arg = self.expand_domain(arg)
                if arg:
                    try:
                        exp = self.get_explanation(arg)
                        if exp and not recursion:
                            # only set explanation in base recursion level
                            self.set_explanation(exp)
                    except: pass
            elif mod == 'redirect':
                self.check_lookups()
                redirect = self.expand_domain(arg)
                if not redirect:
                    raise PermError('redirect has empty domain:',arg)
            elif mod == 'default':
                # default modifier is obsolete
                if self.strict > 1:
                    raise AmbiguityWarning('The default= modifier is obsolete.')
                if not self.strict and self.default_modifier:
                    # might be an old policy, so do it anyway
                    arg = self.expand(arg)
                    # default=- is the same as default=fail
                    default = RESULTS.get(arg, default)
            elif mod == 'op':
                if not recursion:
                    for v in arg.split('.'):
                        if v: self.options[v] = True
            else:
                # spf rfc: 3.6 Unrecognized Mechanisms and Modifiers
                self.expand(m[1])       # syntax error on invalid macro

        # Evaluate mechanisms
        #
        for mech, m, arg, cidrlength, result in mechs:

            if m == 'include':
                self.check_lookups()
                d = self.dns_spf(arg)
                if self.verbose: self.log("include",arg,d)
                res, code, txt = self.check1(d,arg, recursion + 1)
                if res == 'pass':
                    break
                if res == 'none':
                    self.note_error(
                        'No valid SPF record for included domain: %s' %arg,
                      mech)
                res = 'neutral'
                continue
            elif m == 'all':
                break

            elif m == 'exists':
                self.check_lookups()
                try:
                    if len(self.dns_a(arg,'A')) > 0:
                        break
                except AmbiguityWarning:
                    # Exists wants no response sometimes so don't raise
                    # the warning.
                    pass

            elif m == 'a':
                self.check_lookups()
                if self.cidrmatch(self.dns_a(arg,self.A), cidrlength):
                    break

            elif m == 'mx':
                self.check_lookups()
                if self.cidrmatch(self.dns_mx(arg), cidrlength):
                    break

            elif m == 'ip4':
                if self.v == 'in-addr': # match own connection type only
                    try:
                        if self.cidrmatch([arg], cidrlength): break
                    except socket.error:
                        raise PermError('syntax error', mech)

            elif m == 'ip6':
                if self.v == 'ip6': # match own connection type only
                    try:
                        if self.cidrmatch([arg], cidrlength): break
                    except socket.error:
                        raise PermError('syntax error', mech)

            elif m == 'ptr':
                self.check_lookups()
                if domainmatch(self.validated_ptrs(), arg):
                    break

        else:
            # no matches
            if redirect:
                #Catch redirect to a non-existant SPF record.
                redirect_record = self.dns_spf(redirect)
                if not redirect_record:
                    raise PermError('redirect domain has no SPF record',
                        redirect)
                if self.verbose: self.log("redirect",redirect,redirect_record)
                # forget modifiers on redirect
                if not recursion:
                  self.exps = dict(self.defexps)
                  self.options = {}
                return self.check1(redirect_record, redirect, recursion)
            result = default
            mech = None

        if not recursion:       # record matching mechanism at base level
            self.mechanism = mech
        if result == 'fail':
            return (result, 550, exps[result])
        else:
            return (result, 250, exps[result])

    def check_lookups(self):
        self.lookups = self.lookups + 1
        if self.lookups > MAX_LOOKUP*4:
            raise PermError('More than %d DNS lookups'%(MAX_LOOKUP*4))
        if self.lookups > MAX_LOOKUP:
            self.note_error('Too many DNS lookups')

    def get_explanation(self, spec):
        """Expand an explanation."""
        if spec:
            try:
                a = self.dns_txt(spec,ignore_void=True)
                if len(a) == 1:
                    return str(self.expand(to_ascii(a[0]), stripdot=False))
            except PermError:
                # RFC4408 6.2/4 syntax errors cause exp= to be ignored
                if self.strict > 1:
                    raise	# but report in harsh mode for record checking tools
                pass
        elif self.strict > 1:
            raise PermError('Empty domain-spec on exp=')
        # RFC4408 6.2/4 empty domain spec is ignored
        # (unless you give precedence to the grammar).
        return None

    def expand(self, s, stripdot=True): # macros='slodipvh'
        """Do SPF RFC macro expansion.

        Examples:
        >>> q = query(s='strong-bad@email.example.com',
        ...           h='mx.example.org', i='192.0.2.3')
        >>> q.p = 'mx.example.org'
        >>> q.r = 'example.net'

        >>> q.expand('%{d}')
        'email.example.com'

        >>> q.expand('%{d4}')
        'email.example.com'

        >>> q.expand('%{d3}')
        'email.example.com'

        >>> q.expand('%{d2}')
        'example.com'

        >>> q.expand('%{d1}')
        'com'

        >>> q.expand('%{p}')
        'mx.example.org'

        >>> q.expand('%{p2}')
        'example.org'

        >>> q.expand('%{dr}')
        'com.example.email'
    
        >>> q.expand('%{d2r}')
        'example.email'

        >>> q.expand('%{l}')
        'strong-bad'

        >>> q.expand('%{l-}')
        'strong.bad'

        >>> q.expand('%{lr}')
        'strong-bad'

        >>> q.expand('%{lr-}')
        'bad.strong'

        >>> q.expand('%{l1r-}')
        'strong'

        >>> q.expand('%{c}',stripdot=False)
        '192.0.2.3'

        >>> q.expand('%{r}',stripdot=False)
        'example.net'

        >>> q.expand('%{ir}.%{v}._spf.%{d2}')
        '3.2.0.192.in-addr._spf.example.com'

        >>> q.expand('%{lr-}.lp._spf.%{d2}')
        'bad.strong.lp._spf.example.com'

        >>> q.expand('%{lr-}.lp.%{ir}.%{v}._spf.%{d2}')
        'bad.strong.lp.3.2.0.192.in-addr._spf.example.com'

        >>> q.expand('%{ir}.%{v}.%{l1r-}.lp._spf.%{d2}')
        '3.2.0.192.in-addr.strong.lp._spf.example.com'

        >>> try: q.expand('%(ir).%{v}.%{l1r-}.lp._spf.%{d2}')
        ... except PermError as x: print(x)
        invalid-macro-char : %(ir)

        >>> q.expand('%{p2}.trusted-domains.example.net')
        'example.org.trusted-domains.example.net'

        >>> q.expand('%{p2}.trusted-domains.example.net.')
        'example.org.trusted-domains.example.net'

        >>> q = query(s='@email.example.com',
        ...           h='mx.example.org', i='192.0.2.3')
        >>> q.p = 'mx.example.org'
        >>> q.expand('%{l}')
        'postmaster'

        """
        # Check for invalid macro syntax
        if s.find('%') >= 0:
            regex = RE_INVALID_MACRO
            for label in s.split('.'):
                if regex.search(s):
                    raise PermError ('invalid-macro-char ', label)
        # expand macros
        end = 0
        result = ''
        for i in RE_CHAR.finditer(s):
            result += s[end:i.start()]
            macro = s[i.start():i.end()]
            if macro == '%%':
                result += '%'
            elif macro == '%_':
                result += ' '
            elif macro == '%-':
                result += '%20'
            else:
                letter = macro[2].lower()
#                print letter
                if letter == 'p':
                    self.getp()
                elif letter in 'crt' and stripdot:
                    raise PermError(
                        'c,r,t macros allowed in exp= text only', macro)
                expansion = getattr(self, letter, self)
                if expansion:
                    if expansion == self:
                        raise PermError('Unknown Macro Encountered', macro) 
                    e = expand_one(expansion, macro[3:-1], JOINERS.get(letter))
                    if letter != macro[2]:
                        e = urllibparse.quote(e,'~')
                    result += e
            end = i.end()
        result += s[end:]
        if stripdot and result.endswith('.'):
            result =  result[:-1]
        if result.count('.') != 0:
            if len(result) > 253:
                result = result[(result.index('.')+1):]
        return result

    def dns_spf(self, domain):
        """Get the SPF record recorded in DNS for a specific domain
        name.  Returns None if not found, or if more than one record
        is found.
        """
        # Per RFC 4.3/1, check for malformed domain.  This produces
        # no results as a special case.
        for label in domain.split('.'):
          if not label or len(label) > 63:
            return None
        # for performance, check for most common case of TXT first
        a = [t for t in self.dns_txt(domain) if RE_SPF.match(t)]
        if len(a) > 1:
            raise PermError('Two or more type TXT spf records found.')
        if len(a) == 1 and self.strict < 2:
            return to_ascii(a[0])
        # check official SPF type first when it becomes more popular
        if self.strict > 1:
            #Only check for Type SPF in harsh mode until it is more popular.
            try:
                b = [t for t in self.dns_txt(domain,'SPF',ignore_void=True)
			if RE_SPF.match(t)]
            except TempError as x:
                # some braindead DNS servers hang on type 99 query
                if self.strict > 1: raise TempError(x)
                b = []
            if len(b) > 1:
                raise PermError('Two or more type SPF spf records found.')
            if len(b) == 1:
                if self.strict > 1 and len(a) == 1 and a[0] != b[0]:
                #Changed from permerror to warning based on RFC 4408 Auth 48 change
                    raise AmbiguityWarning(
'v=spf1 records of both type TXT and SPF (type 99) present, but not identical')
                return to_ascii(b[0])
        if len(a) == 1:
            return to_ascii(a[0])    # return TXT if SPF wasn't found
        if DELEGATE:    # use local record if neither found
            a = [t
              for t in self.dns_txt(domain+'._spf.'+DELEGATE,ignore_void=True)
            if RE_SPF.match(t)
            ]
            if len(a) == 1: return to_ascii(a[0])
        return None

    ## Get list of TXT records for a domain name.
    # Any DNS library *must* return bytes (same as str in python2) for TXT
    # or SPF since there is no general decoding to unicode.  Py3dns-3.0.2
    # incorrectly attempts to convert to str using idna encoding by default.
    # We work around this by assuming any UnicodeErrors coming from py3dns
    # are from a non-ascii SPF record (incorrect in general).  Packages
    # should require py3dns != 3.0.2.
    # 
    # We cannot check for non-ascii here, because we must ignore non-SPF
    # records - even when they are non-ascii.  So we return bytes.
    # The caller does the ascii check for SPF records and explanations.
    # 
    def dns_txt(self, domainname, rr='TXT',ignore_void=False):
        "Get a list of TXT records for a domain name."
        if domainname:
          try:
              dns_list = self.dns(domainname, rr,ignore_void=ignore_void)
              if dns_list:
                  # a[0][:0] is '' for py3dns-3.0.2, otherwise b''
                  a = [a[0][:0].join(a) for a in dns_list if a]
                  # FIXME: workaround for error in py3dns-3.0.2
                  if isinstance(a[0],bytes):
                      return a
                  return [s.encode('utf-8') for s in a]
          # FIXME: workaround for error in py3dns-3.0.2
          except UnicodeError:
              raise PermError('Non-ascii characters found in %s record for %s'
                 %(rr,domainname))
        return []

    def dns_mx(self, domainname):
        """Get a list of IP addresses for all MX exchanges for a
        domain name.
        """
        # RFC 4408/7208 section 5.4 "mx"
        # To prevent DoS attacks, more than 10 MX names MUST NOT be looked up
        # Changed to permerror if more than 10 exist in 7208
        mxnames = self.dns(domainname, 'MX')
        if self.strict:
            max = MAX_MX
            if len(mxnames) > MAX_MX:
                raise PermError(
                    'More than %d MX records returned'%MAX_MX)
            if self.strict > 1:
                if len(mxnames) == 0:
                    raise AmbiguityWarning(
                        'No MX records found for mx mechanism', domainname)
        else:
            max = MAX_MX * 4
        mxnames.sort()
        return [a for mx in mxnames[:max] for a in self.dns_a(mx[1],self.A)]

    def dns_a(self, domainname, A='A'):
        """Get a list of IP addresses for a domainname.
        """
        if not domainname: return []
        r = self.dns(domainname, A)
        if self.strict > 1 and len(r) == 0:
            raise AmbiguityWarning(
                    'No %s records found for'%A, domainname)
        if A == 'AAAA' and bytes is str:
          # work around pydns inconsistency plus python2 bytes/str ambiguity
          return [Bytes(ip) for ip in r]
        return r

    def validated_ptrs(self):
        """Figure out the validated PTR domain names for the connect IP."""
# To prevent DoS attacks, more than 10 PTR names MUST NOT be looked up
        if self.strict:
            max = MAX_PTR
            if self.strict > 1:
                #Break out the number of PTR records returned for testing
                try:
                    ptrnames = self.dns_ptr(self.i)
                    if len(ptrnames) > max:
                        warning = 'More than %d PTR records returned' % max
                        raise AmbiguityWarning(warning, self.c)
                    else:
                        if len(ptrnames) == 0:
                            raise AmbiguityWarning(
                                'No PTR records found for ptr mechanism', self.c)
                except:
                    raise AmbiguityWarning(
                      'No PTR records found for ptr mechanism', self.c)
        else:
            max = MAX_PTR * 4
        cidrlength = self.cidrmax
        return [p for p in self.dns_ptr(self.i)[:max]
            if self.cidrmatch(self.dns_a(p,self.A),cidrlength)]

    def dns_ptr(self, i):
        """Get a list of domain names for an IP address."""
        return self.dns('%s.%s.arpa'%(reverse_dots(i),self.v), 'PTR')

    # We have to be careful which additional DNS RRs we cache.  For
    # instance, PTR records are controlled by the connecting IP, and they
    # could poison our local cache with bogus A and MX records.  

    SAFE2CACHE = {
      ('MX','A'): None,
      ('MX','MX'): None,
      ('CNAME','A'): None,
      ('A','A'): None,
      ('AAAA','AAAA'): None,
      ('PTR','PTR'): None,
      ('TXT','TXT'): None,
      ('SPF','SPF'): None
    }

    # FIXME: move to dnsplug
    def dns(self, name, qtype, cnames=None, ignore_void=False):
        """DNS query.

        If the result is in cache, return that.  Otherwise pull the
        result from DNS, and cache ALL answers, so additional info
        is available for further queries later.

        CNAMEs are followed.

        If there is no data, [] is returned.

        pre: qtype in ['A', 'AAAA', 'MX', 'PTR', 'TXT', 'SPF']
        post: isinstance(__return__, types.ListType)
        """
        if name.endswith('.'): name = name[:-1]
        if not reduce(lambda x,y:x and 0 < len(y) < 64, name.split('.'),True):
            return []   # invalid DNS name (too long or empty)
        name = name.lower()
        result = self.cache.get( (name, qtype), [])
        if result: return result
        cnamek = (name,'CNAME')
        cname = self.cache.get( cnamek )

        if cname:
            cname = cname[0]
        else:
            safe2cache = query.SAFE2CACHE
            if self.querytime < 0:
                raise TempError('DNS Error: exceeded max query lookup time')
            if self.querytime < self.timeout and self.querytime > 0:
                timeout = self.querytime
            else:
                timeout = self.timeout
            timethen = time.time()
            for k, v in DNSLookup(name, qtype, self.strict, timeout):
                # Force case insensitivity in cache, DNS servers often
                # return random case in domain part of answers.
                k = (k[0].lower(), k[1]) 
                if k == cnamek:
                    cname = v
                if k[1] == 'CNAME' or (qtype,k[1]) in safe2cache:
                    self.cache.setdefault(k, []).append(v)
                    #if ans and qtype == k[1]:
                    #    self.cache.setdefault((name,qtype), []).append(v)
            result = self.cache.get( (name, qtype), [])
            if self.querytime > 0:
                self.querytime = self.querytime - (time.time()-timethen)
        if not result and cname:
            if not cnames:
                cnames = {}
            elif len(cnames) >= MAX_CNAME:
                #return result    # if too many == NX_DOMAIN
                raise PermError('Length of CNAME chain exceeds %d' % MAX_CNAME)
            cnames[name] = cname
            if cname.lower().rstrip('.') in cnames:
                if self.strict > 1: raise AmbiguityWarning('CNAME loop', cname)
            else:
                result = self.dns(cname, qtype, cnames=cnames)
                if result:
                    self.cache[(name,qtype)] = result
        if not result and not ignore_void:
            self.void_lookups += 1
            if self.void_lookups > MAX_VOID_LOOKUPS:
                raise PermError('Void lookup limit of %d exceeded' % MAX_VOID_LOOKUPS)
        return result

    def cidrmatch(self, ipaddrs, n):
        """Match connect IP against a CIDR network of other IP addresses.

        Examples:
        >>> c = query(s='strong-bad@email.example.com',
        ...           h='mx.example.org', i='192.0.2.3')
        >>> c.p = 'mx.example.org'
        >>> c.r = 'example.com'

        >>> c.cidrmatch(['192.0.2.3'],32)
        True
        >>> c.cidrmatch(['192.0.2.2'],32)
        False
        >>> c.cidrmatch(['192.0.2.2'],31)
        True

        >>> six = query(s='strong-bad@email.example.com',
        ...           h='mx.example.org', i='2001:0db8:0:0:0:0:0:0001')
        >>> six.p = 'mx.example.org'
        >>> six.r = 'example.com'

        >>> six.cidrmatch(['2001:0DB8::'],127)
        True
        >>> six.cidrmatch(['2001:0DB8::'],128)
        False
        >>> six.cidrmatch(['2001:0DB8:0:0:0:0:0:0001'],128)
        True
        """
        try:
            try:
                for netwrk in [ipaddress.ip_network(ip) for ip in ipaddrs]:
                    network = netwrk.supernet(new_prefix=n)
                    if isinstance(self.iplist, bool):
                        if network.__contains__(self.ipaddr):
                            return True
                    else:
                        if n < self.cidrmax:
                            self.iplist.append(network)
                        else:
                            self.iplist.append(network.ip)
            except AttributeError:
                for netwrk in [ipaddress.IPNetwork(ip,strict=False) for ip in ipaddrs]:
                    network = netwrk.supernet(new_prefix=n)
                    if isinstance(self.iplist, bool):
                        if network.__contains__(self.ipaddr):
                            return True
                    else:
                        if n < self.cidrmax:
                            self.iplist.append(network)
                        else:
                            self.iplist.append(network.ip)
        except ValueError as x:
            raise PermError(str(x))
        return False

    def parse_header_ar(self, val):
        """Set SPF values from RFC 5451 Authentication Results header.
        
        Useful when SPF has already been run on a trusted gateway machine.

        Expects the entire header as an input.

        Examples:
        >>> q = query('192.0.2.3','strong-bad@email.example.com','mx.example.org')
        >>> q.mechanism = 'unknown'
        >>> p = q.parse_header_ar('''Authentication-Results: bmsi.com; spf=neutral \\n     (abuse@kitterman.com: 192.0.2.3 is neither permitted nor denied by domain of email.example.com) \\n     smtp.mailfrom=email.example.com \\n    (sender=strong-bad@email.example.com; helo=mx.example.org; client-ip=192.0.2.3; receiver=abuse@kitterman.com; mechanism=?all)''')
        >>> q.get_header(q.result, header_type='authres', aid='bmsi.com')
        'Authentication-Results: bmsi.com; spf=neutral (unknown: 192.0.2.3 is neither permitted nor denied by domain of email.example.com) smtp.mailfrom=email.example.com (sender=email.example.com; helo=mx.example.org; client-ip=192.0.2.3; receiver=unknown; mechanism=unknown)'
        >>> p = q.parse_header_ar('''Authentication-Results: bmsi.com; spf=None (mail.bmsi.com: test; client-ip=163.247.46.150) smtp.mailfrom=admin@squiebras.cl (helo=mail.squiebras.cl; receiver=mail.bmsi.com;\\n mechanism=mx/24)''')
        >>> q.get_header(q.result, header_type='authres', aid='bmsi.com')
        'Authentication-Results: bmsi.com; spf=none (unknown: 192.0.2.3 is neither permitted nor denied by domain of email.example.com) smtp.mailfrom=admin@squiebras.cl (sender=admin@squiebras.cl; helo=mx.example.org; client-ip=192.0.2.3; receiver=unknown; mechanism=unknown)'
        """
        import authres
        # Authres expects unwrapped headers according to docs
        val = ' '.join(s.strip() for s in val.split('\n'))
        arobj = authres.AuthenticationResultsHeader.parse(val)
        # TODO extract and parse comments (not supported by authres)
        for resobj in arobj.results:
            if resobj.method == 'spf':
                self.authserv = arobj.authserv_id
                self.result = resobj.result
                if resobj.properties[0].name == 'mailfrom':
                    self.d = resobj.properties[0].value
                    self.s = resobj.properties[0].value
                if resobj.properties[0].name == 'helo':
                    self.h = resobj.properties[0].value
        return

    def parse_header_spf(self, val):
        """Set SPF values from Received-SPF header.
        
        Useful when SPF has already been run on a trusted gateway machine.

        Examples:
        >>> q = query('0.0.0.0','','')
        >>> p = q.parse_header_spf('''Pass (test) client-ip=70.98.79.77;
        ... envelope-from="evelyn@subjectsthum.com"; helo=mail.subjectsthum.com;
        ... receiver=mail.bmsi.com; mechanism=a; identity=mailfrom''')
        >>> q.get_header(q.result)
        'Pass (test) client-ip=70.98.79.77; envelope-from="evelyn@subjectsthum.com"; helo=mail.subjectsthum.com; receiver=mail.bmsi.com; mechanism=a; identity=mailfrom'
        >>> o = q.parse_header_spf('''None (mail.bmsi.com: test)
        ... client-ip=163.247.46.150; envelope-from="admin@squiebras.cl";
        ... helo=mail.squiebras.cl; receiver=mail.bmsi.com; mechanism=mx/24;
        ... x-bestguess=pass; x-helo-spf=neutral; identity=mailfrom''')
        >>> q.get_header(q.result,**o)
        'None (mail.bmsi.com: test) client-ip=163.247.46.150; envelope-from="admin@squiebras.cl"; helo=mail.squiebras.cl; receiver=mail.bmsi.com; mechanism=mx/24; x-bestguess=pass; x-helo-spf=neutral; identity=mailfrom'
        >>> o['bestguess']
        'pass'
        """
        a = val.split(None,1)
        self.result = a[0].lower()
        self.mechanism = None
        if len(a) < 2: return 'none'
        val = a[1]
        if val.startswith('('):
          pos = val.find(')')
          if pos < 0: return self.result
          self.comment = val[1:pos]
          val = val[pos+1:]
        msg = Message()
        msg.add_header('Received-SPF','; '+val)
        p = {}
        for k,v in msg.get_params(header='Received-SPF'):
          if k == 'client-ip':
            self.set_ip(v)
          elif k == 'envelope-from': self.s = v
          elif k == 'helo': self.h = v
          elif k == 'receiver': self.r = v
          elif k == 'problem': self.mech = v
          elif k == 'mechanism': self.mechanism = v
          elif k == 'identity': self.ident = v
          elif k.startswith('x-'): p[k[2:]] = v
        self.l, self.o = split_email(self.s, self.h)
        return p

    def parse_header(self, val):
        """Set SPF values from Received-SPF or RFC 5451 Authentication Results header.
        
        Useful when SPF has already been run on a trusted gateway machine. Auto
        detects the header type and parses it. Use parse_header_spf or parse_header_ar
        for each type if required.

        Examples:
        >>> q = query('0.0.0.0','','')
        >>> p = q.parse_header('''Pass (test) client-ip=70.98.79.77;
        ... envelope-from="evelyn@subjectsthum.com"; helo=mail.subjectsthum.com;
        ... receiver=mail.bmsi.com; mechanism=a; identity=mailfrom''')
        >>> q.get_header(q.result)
        'Pass (test) client-ip=70.98.79.77; envelope-from="evelyn@subjectsthum.com"; helo=mail.subjectsthum.com; receiver=mail.bmsi.com; mechanism=a; identity=mailfrom'
        >>> r = q.parse_header('''None (mail.bmsi.com: test)
        ... client-ip=163.247.46.150; envelope-from="admin@squiebras.cl";
        ... helo=mail.squiebras.cl; receiver=mail.bmsi.com; mechanism=mx/24;
        ... x-bestguess=pass; x-helo-spf=neutral; identity=mailfrom''')
        >>> q.get_header(q.result,**r)
        'None (mail.bmsi.com: test) client-ip=163.247.46.150; envelope-from="admin@squiebras.cl"; helo=mail.squiebras.cl; receiver=mail.bmsi.com; mechanism=mx/24; x-bestguess=pass; x-helo-spf=neutral; identity=mailfrom'
        >>> r['bestguess']
        'pass'
        >>> q = query('192.0.2.3','strong-bad@email.example.com','mx.example.org')
        >>> q.mechanism = 'unknown'
        >>> p = q.parse_header_ar('''Authentication-Results: bmsi.com; spf=neutral \\n     (abuse@kitterman.com: 192.0.2.3 is neither permitted nor denied by domain of email.example.com) \\n     smtp.mailfrom=email.example.com \\n     (sender=strong-bad@email.example.com; helo=mx.example.org; client-ip=192.0.2.3; receiver=abuse@kitterman.com; mechanism=?all)''')
        >>> q.get_header(q.result, header_type='authres', aid='bmsi.com')
        'Authentication-Results: bmsi.com; spf=neutral (unknown: 192.0.2.3 is neither permitted nor denied by domain of email.example.com) smtp.mailfrom=email.example.com (sender=email.example.com; helo=mx.example.org; client-ip=192.0.2.3; receiver=unknown; mechanism=unknown)'
        >>> p = q.parse_header_ar('''Authentication-Results: bmsi.com; spf=None (mail.bmsi.com: test; client-ip=163.247.46.150) smtp.mailfrom=admin@squiebras.cl (helo=mail.squiebras.cl; receiver=mail.bmsi.com; mechanism=mx/24)''')
        >>> q.get_header(q.result, header_type='authres', aid='bmsi.com')
        'Authentication-Results: bmsi.com; spf=none (unknown: 192.0.2.3 is neither permitted nor denied by domain of email.example.com) smtp.mailfrom=admin@squiebras.cl (sender=admin@squiebras.cl; helo=mx.example.org; client-ip=192.0.2.3; receiver=unknown; mechanism=unknown)'
        """

        if val.startswith('Authentication-Results:'):
            return(self.parse_header_ar(val))
        else:
            return(self.parse_header_spf(val))

    def get_header(self, res, receiver=None, header_type='spf', aid=None, **kv):
        """
        Generate Received-SPF or Authentication Results header based on the
         last lookup.

        >>> q = query(s='strong-bad@email.example.com', h='mx.example.org',
        ...           i='192.0.2.3')
        >>> q.r='abuse@kitterman.com'
        >>> q.check(spf='v=spf1 ?all')
        ('neutral', 250, 'access neither permitted nor denied')
        >>> q.get_header('neutral')
        'Neutral (abuse@kitterman.com: 192.0.2.3 is neither permitted nor denied by domain of email.example.com) client-ip=192.0.2.3; envelope-from="strong-bad@email.example.com"; helo=mx.example.org; receiver=abuse@kitterman.com; mechanism=?all; identity=mailfrom'

        >>> q.check(spf='v=spf1 redirect=controlledmail.com exp=_exp.controlledmail.com')
        ('fail', 550, 'SPF fail - not authorized')
        >>> q.get_header('fail')
        'Fail (abuse@kitterman.com: domain of email.example.com does not designate 192.0.2.3 as permitted sender) client-ip=192.0.2.3; envelope-from="strong-bad@email.example.com"; helo=mx.example.org; receiver=abuse@kitterman.com; mechanism=-all; identity=mailfrom'
    
        >>> q.check(spf='v=spf1 ip4:192.0.0.0/8 ?all moo')
        ('permerror', 550, 'SPF Permanent Error: Unknown mechanism found: moo')
        >>> q.get_header('permerror')
        'PermError (abuse@kitterman.com: permanent error in processing domain of email.example.com: Unknown mechanism found) client-ip=192.0.2.3; envelope-from="strong-bad@email.example.com"; helo=mx.example.org; receiver=abuse@kitterman.com; problem=moo; identity=mailfrom'

        >>> q.check(spf='v=spf1 ip4:192.0.0.0/8 ~all')
        ('pass', 250, 'sender SPF authorized')
        >>> q.get_header('pass')
        'Pass (abuse@kitterman.com: domain of email.example.com designates 192.0.2.3 as permitted sender) client-ip=192.0.2.3; envelope-from="strong-bad@email.example.com"; helo=mx.example.org; receiver=abuse@kitterman.com; mechanism="ip4:192.0.0.0/8"; identity=mailfrom'

        >>> q.check(spf='v=spf1 ?all')
        ('neutral', 250, 'access neither permitted nor denied')
        >>> q.get_header('neutral', header_type = 'authres', aid='bmsi.com')
        'Authentication-Results: bmsi.com; spf=neutral (abuse@kitterman.com: 192.0.2.3 is neither permitted nor denied by domain of email.example.com) smtp.mailfrom=email.example.com (sender=strong-bad@email.example.com; helo=mx.example.org; client-ip=192.0.2.3; receiver=abuse@kitterman.com; mechanism=?all)'

        >>> p = query(s='strong-bad@email.example.com', h='mx.example.org',
        ...           i='192.0.2.3')
        >>> p.r='abuse@kitterman.com'
        >>> p.check(spf='v=spf1 redirect=controlledmail.com exp=_exp.controlledmail.com')
        ('fail', 550, 'SPF fail - not authorized')
        >>> p.ident = 'helo'
        >>> p.get_header('fail', header_type = 'authres', aid='bmsi.com')
        'Authentication-Results: bmsi.com; spf=fail (abuse@kitterman.com: domain of email.example.com does not designate 192.0.2.3 as permitted sender) smtp.helo=mx.example.org (sender=strong-bad@email.example.com; client-ip=192.0.2.3; receiver=abuse@kitterman.com; mechanism=-all)'

        >>> q.check(spf='v=spf1 ?all')
        ('neutral', 250, 'access neither permitted nor denied')
        >>> try: q.get_header('neutral', header_type = 'dkim')
        ... except SyntaxError as x: print(x)
        Unknown results header type: dkim
        """
        # If type is Authentication Results header (spf/authres)
        if header_type == 'authres':
            if not aid:
                raise SyntaxError('authserv-id missing for Authentication Results header type, see RFC5451 2.3')
            import authres

        if not receiver:
            receiver = self.r
        client_ip = self.c
        helo = quote_value(self.h)
        resmap = { 'pass': 'Pass', 'neutral': 'Neutral', 'fail': 'Fail',
                'softfail': 'SoftFail', 'none': 'None',
                'temperror': 'TempError', 'permerror': 'PermError' }
        identity = self.ident
        if identity == 'helo':
            envelope_from = None
        else:
            envelope_from = quote_value(self.s)
        tag = resmap[res]
        if res == 'permerror' and self.mech:
            problem = quote_value(' '.join(self.mech))
        else:
            problem = None
        mechanism = quote_value(self.mechanism)
        if hasattr(self,'comment'):
          comment = self.comment
        else:
          comment = '%s: %s' % (receiver,self.get_header_comment(res))
        res = ['%s (%s)' % (tag,comment)]
        if header_type == 'spf':
            for k in ('client_ip','envelope_from','helo','receiver',
                'problem','mechanism'):
                v = locals()[k]
                if v: res.append('%s=%s;'%(k.replace('_','-'),v))
            for k,v in sorted(list(kv.items())):
                if v: res.append('x-%s=%s;'%(k.replace('_','-'),quote_value(v)))
            # do identity last so we can easily drop the trailing ';'
            res.append('%s=%s'%('identity',identity))
            return ' '.join(res)
        elif header_type == 'authres':
            if envelope_from:
                return str(authres.AuthenticationResultsHeader(authserv_id = aid, \
                    results = [authres.SPFAuthenticationResult(result = tag, \
                    result_comment = comment, smtp_mailfrom = self.d, \
                    smtp_mailfrom_comment = \
                    'sender={0}; helo={1}; client-ip={2}; receiver={3}; mechanism={4}'.format(self.s, \
                    self.h, self.c, self.r, mechanism))]))
            else:
                return str(authres.AuthenticationResultsHeader(authserv_id = aid, \
                    results = [authres.SPFAuthenticationResult(result = tag, \
                    result_comment = comment, smtp_helo = self.h, \
                    smtp_helo_comment = \
                    'sender={0}; client-ip={1}; receiver={2}; mechanism={3}'.format(self.s, \
                    self.c, self.r, mechanism))]))
        else:
            raise SyntaxError('Unknown results header type: {0}'.format(header_type))

    def get_header_comment(self, res):
        """Return comment for Received-SPF header.  """
        sender = self.o
        if res == 'pass':
            return \
                "domain of %s designates %s as permitted sender" \
                % (sender, self.c)
        elif res == 'softfail': return \
      "transitioning domain of %s does not designate %s as permitted sender" \
            % (sender, self.c)
        elif res == 'neutral': return \
            "%s is neither permitted nor denied by domain of %s" \
                % (self.c, sender)
        elif res == 'none': return \
            "%s is neither permitted nor denied by domain of %s" \
                  % (self.c, sender)
            #"%s does not designate permitted sender hosts" % sender
        elif res == 'permerror': return \
            "permanent error in processing domain of %s: %s" \
                  % (sender, self.prob)
        elif res == 'temperror': return \
              "temporary error in processing during lookup of %s" % sender
        elif res == 'fail': return \
              "domain of %s does not designate %s as permitted sender" \
              % (sender, self.c)
        raise ValueError("invalid SPF result for header comment: "+res)

def split_email(s, h):
    """Given a sender email s and a HELO domain h, create a valid tuple
    (l, d) local-part and domain-part.

    Examples:
    >>> split_email('', 'wayforward.net')
    ('postmaster', 'wayforward.net')

    >>> split_email('foo.com', 'wayforward.net')
    ('postmaster', 'foo.com')

    >>> split_email('terry@wayforward.net', 'optsw.com')
    ('terry', 'wayforward.net')
    """
    if not s:
        return 'postmaster', h
    else:
        parts = s.split('@', 1)
        if parts[0] == '':
            parts[0] = 'postmaster'
        if len(parts) == 2:
            return tuple(parts)
        else:
            return 'postmaster', s

def quote_value(s):
    """Quote the value for a key-value pair in Received-SPF header field
    if needed.  No quoting needed for a dot-atom value.

    Examples:
    >>> quote_value('foo@bar.com')
    '"foo@bar.com"'
    
    >>> quote_value('mail.example.com')
    'mail.example.com'

    >>> quote_value('A:1.2.3.4')
    '"A:1.2.3.4"'

    >>> quote_value('abc"def')
    '"abc\\\\"def"'

    >>> quote_value(r'abc\def')
    '"abc\\\\\\\\def"'

    >>> quote_value('abc..def')
    '"abc..def"'

    >>> quote_value('')
    '""'

    >>> quote_value(None)
    """
    if s is None or RE_DOT_ATOM.match(s):
      return s
    return '"' + s.replace('\\',r'\\').replace('"',r'\"'
                ).replace('\x00',r'\x00') + '"'

def parse_mechanism(str, d):
    """Breaks A, MX, IP4, and PTR mechanisms into a (name, domain,
    cidr,cidr6) tuple.  The domain portion defaults to d if not present,
    the cidr defaults to 32 if not present.

    Examples:
    >>> parse_mechanism('a', 'foo.com')
    ('a', 'foo.com', None, None)

    >>> parse_mechanism('exists','foo.com')
    ('exists', None, None, None)

    >>> parse_mechanism('a:bar.com', 'foo.com')
    ('a', 'bar.com', None, None)

    >>> parse_mechanism('a/24', 'foo.com')
    ('a', 'foo.com', 24, None)

    >>> parse_mechanism('A:foo:bar.com/16//48', 'foo.com')
    ('a', 'foo:bar.com', 16, 48)

    >>> parse_mechanism('-exists:%{i}.%{s1}.100/86400.rate.%{d}','foo.com')
    ('-exists', '%{i}.%{s1}.100/86400.rate.%{d}', None, None)

    >>> parse_mechanism('mx:%%%_/.Claranet.de/27','foo.com')
    ('mx', '%%%_/.Claranet.de', 27, None)

    >>> parse_mechanism('mx:%{d}//97','foo.com')
    ('mx', '%{d}', None, 97)

    >>> parse_mechanism('iP4:192.0.0.0/8','foo.com')
    ('ip4', '192.0.0.0', 8, None)
    """

    a = RE_DUAL_CIDR.split(str)
    if len(a) == 3:
        str, cidr6 = a[0], int(a[1])
    else:
        cidr6 = None
    a = RE_CIDR.split(str)
    if len(a) == 3:
        str, cidr = a[0], int(a[1])
    else:
        cidr = None

    a = str.split(':', 1)
    if len(a) < 2:
        str = str.lower()
        if str == 'exists': d = None
        return str, d, cidr, cidr6
    return a[0].lower(), a[1], cidr, cidr6

def reverse_dots(name):
    """Reverse dotted IP addresses or domain names.

    Example:
    >>> reverse_dots('192.168.0.145')
    '145.0.168.192'

    >>> reverse_dots('email.example.com')
    'com.example.email'
    """
    a = name.split('.')
    a.reverse()
    return '.'.join(a)

def domainmatch(ptrs, domainsuffix):
    """grep for a given domain suffix against a list of validated PTR
    domain names.

    Examples:
    >>> domainmatch(['FOO.COM'], 'foo.com')
    1

    >>> domainmatch(['moo.foo.com'], 'FOO.COM')
    1

    >>> domainmatch(['moo.bar.com'], 'foo.com')
    0

    """
    domainsuffix = domainsuffix.lower()
    for ptr in ptrs:
        ptr = ptr.lower()
        if ptr == domainsuffix or ptr.endswith('.' + domainsuffix):
            return True

    return False

def expand_one(expansion, str, joiner):
    if not str:
        return expansion
    ln, reverse, delimiters = RE_ARGS.split(str)[1:4]
    if not delimiters:
        delimiters = '.'
    expansion = split(expansion, delimiters, joiner)
    if reverse: expansion.reverse()
    if ln: expansion = expansion[-int(ln)*2+1:]
    return ''.join(expansion)

def split(str, delimiters, joiner=None):
    """Split a string into pieces by a set of delimiter characters.  The
    resulting list is delimited by joiner, or the original delimiter if
    joiner is not specified.

    Examples:
    >>> split('192.168.0.45', '.')
    ['192', '.', '168', '.', '0', '.', '45']

    >>> split('terry@wayforward.net', '@.')
    ['terry', '@', 'wayforward', '.', 'net']

    >>> split('terry@wayforward.net', '@.', '.')
    ['terry', '.', 'wayforward', '.', 'net']
    """
    result, element = [], ''
    for c in str:
        if c in delimiters:
            result.append(element)
            element = ''
            if joiner:
                result.append(joiner)
            else:
                result.append(c)
        else:
            element += c
    result.append(element)
    return result

def insert_libspf_local_policy(spftxt, local=None):
    """Returns spftxt with local inserted just before last non-fail
    mechanism.  This is how the libspf{2} libraries handle "local-policy".
    
    Examples:
    >>> insert_libspf_local_policy('v=spf1 -all')
    'v=spf1 -all'
    >>> insert_libspf_local_policy('v=spf1 -all','mx')
    'v=spf1 -all'
    >>> insert_libspf_local_policy('v=spf1','a mx ptr')
    'v=spf1 a mx ptr'
    >>> insert_libspf_local_policy('v=spf1 mx -all','a ptr')
    'v=spf1 mx a ptr -all'
    >>> insert_libspf_local_policy('v=spf1 mx -include:foo.co +all','a ptr')
    'v=spf1 mx a ptr -include:foo.co +all'

    # FIXME: is this right?  If so, "last non-fail" is a bogus description.
    >>> insert_libspf_local_policy('v=spf1 mx ?include:foo.co +all','a ptr')
    'v=spf1 mx a ptr ?include:foo.co +all'
    >>> spf='v=spf1 ip4:1.2.3.4 -a:example.net -all'
    >>> local='ip4:192.0.2.3 a:example.org'
    >>> insert_libspf_local_policy(spf,local)
    'v=spf1 ip4:1.2.3.4 ip4:192.0.2.3 a:example.org -a:example.net -all'
    """
    # look to find the all (if any) and then put local
    # just after last non-fail mechanism.  This is how
    # libspf2 handles "local policy", and some people
    # apparently find it useful (don't ask me why).
    if not local: return spftxt
    spf = spftxt.split()[1:]
    if spf:
        # local policy is SPF mechanisms/modifiers with no
        # 'v=spf1' at the start
        spf.reverse() #find the last non-fail mechanism
        for mech in spf:
        # map '?' '+' or '-' to 'neutral' 'pass'
        # or 'fail'
            if not RESULTS.get(mech[0]):
                # actually finds last mech with default result
                where = spf.index(mech)
                spf[where:where] = [local]
                spf.reverse()
                local = ' '.join(spf)
                break
        else:
            return spftxt # No local policy adds for v=spf1 -all
    # Processing limits not applied to local policy.  Suggest
    # inserting 'local' mechanism to handle this properly
    #MAX_LOOKUP = 100 
    return 'v=spf1 '+local

if sys.version_info[0] == 2:
  def to_ascii(s):
      "Raise PermError if arg is not 7-bit ascii."
      try:
        return s.encode('ascii')
      except UnicodeError:
        raise PermError('Non-ascii characters found',repr(s))
else:
  def to_ascii(s):
      "Raise PermError if arg is not 7-bit ascii."
      try:
        return s.decode('ascii')
      except UnicodeError:
        raise PermError('Non-ascii characters found',repr(s))

def _test():
    import doctest, spf
    return doctest.testmod(spf)

DNS.DiscoverNameServers() # Fails on Mac OS X? Add domain to /etc/resolv.conf

if __name__ == '__main__':
    import getopt
    try:
       opts,argv = getopt.getopt(sys.argv[1:],"hvs:",
        ["help","verbose","strict"])
    except getopt.GetoptError as err:
       print(str(err))
       print(USAGE)
       sys.exit(2)
    verbose = False
    strict = True
    for o,a in opts:
        if o in ('-v','--verbose'):
           verbose = True
        if o in ('-s','--strict'):
           strict = int(a)
        elif o in ('-h','--help'):
           print(USAGE)
    if len(argv) == 0:
        print(USAGE)
        _test()
    elif len(argv) == 1:
        try:
            q = query(i='127.0.0.1', s='localhost', h='unknown',
                receiver=socket.gethostname())
            print(q.dns_spf(argv[0]))
        except TempError as x:
            print("Temporary DNS error: ", x)
        except PermError as x:
            print("PermError: ", x)
    elif len(argv) == 3:
        i, s, h = argv
        q = query(i=i, s=s, h=h,receiver=socket.gethostname(),verbose=verbose,
                strict=strict)
        r = q.check()
        print('result:',r,q.mechanism)
        if r[0] == 'none':
          print('guessed:',q.best_guess(),q.mechanism)
        if q.perm_error and q.perm_error.ext:
            print('lax:',q.perm_error.ext)
        if q.iplist:
            for ip in q.iplist:
                print(ip)
    elif len(argv) == 4:
        i, s, h = argv[1:]
        q = query(i=i, s=s, h=h, receiver=socket.gethostname(),
            strict=False, verbose=verbose)
        r = q.check(argv[0])
        print('result:',r,q.mechanism)
        if r[0] == 'none':
          print('guessed:',q.best_guess(),q.mechanism)
        if q.perm_error and q.perm_error.ext:
            print('lax:',q.perm_error.ext)
    else:
        print(USAGE)