gman 5.0.2 → 5.0.3
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- checksums.yaml +4 -4
- data/config/domains.txt +0 -28
- data/lib/gman/importer.rb +26 -18
- data/lib/gman/version.rb +1 -1
- metadata +1 -1
checksums.yaml
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
---
|
2
2
|
SHA1:
|
3
|
-
metadata.gz:
|
4
|
-
data.tar.gz:
|
3
|
+
metadata.gz: d5f81c5dc5d03e961dc8830f91d8f17a04cc7eff
|
4
|
+
data.tar.gz: 1c4c914aca60114856034e3cfd1254c555552d49
|
5
5
|
SHA512:
|
6
|
-
metadata.gz:
|
7
|
-
data.tar.gz:
|
6
|
+
metadata.gz: e463cf1a41e87be42a1bae6b82e9b5961a210ea4374750ee9e111406589199adaf0e15be60b31dd65f4cee1e49b510a1e6eacc3f3a7c75c7e4c05fb70313bf8c
|
7
|
+
data.tar.gz: be4371cf714f72cf0b0d0bafdc3de1e26c8b549849f70cce3779bb334bbecad5668f4bc65b5356a5e163d8782b270e8cae3b219ffd0cbc58e8f0a83162809c8a
|
data/config/domains.txt
CHANGED
@@ -206,8 +206,6 @@ cochrane.ca
|
|
206
206
|
coldlake.com
|
207
207
|
colemantownship.ca
|
208
208
|
colwood.ca
|
209
|
-
communityofmillertown.ca
|
210
|
-
communityofoleary.com
|
211
209
|
comox.ca
|
212
210
|
conceptionbaysouth.ca
|
213
211
|
consort.ca
|
@@ -596,7 +594,6 @@ municipalite.saint-armand.qc.ca
|
|
596
594
|
municipalite.saint-isidore.qc.ca
|
597
595
|
municipalite.saint-philippe.qc.ca
|
598
596
|
municipalite.saint-valerien.qc.ca
|
599
|
-
municipalite.sainte-angele-de-monnoir.qc.ca
|
600
597
|
municipalite.sainte-martine.qc.ca
|
601
598
|
municipaliteboileau.ca
|
602
599
|
municipalitedeboltonouest.com
|
@@ -608,7 +605,6 @@ municipalityofmcdougall.com
|
|
608
605
|
municipalityofshelburne.ca
|
609
606
|
muniles.ca
|
610
607
|
munmilan.qc.ca
|
611
|
-
munpontiac.com
|
612
608
|
munstdenis.com
|
613
609
|
munstratford.qc.ca
|
614
610
|
mycallander.ca
|
@@ -883,7 +879,6 @@ snowlake.com
|
|
883
879
|
sooke.ca
|
884
880
|
sourismanitoba.com
|
885
881
|
sourispei.com
|
886
|
-
southbrook.tripod.com
|
887
882
|
southey.ca
|
888
883
|
southhuron.ca
|
889
884
|
southwestmiddlesex.ca
|
@@ -1092,7 +1087,6 @@ townofgrandbank.com
|
|
1092
1087
|
townofhague.com
|
1093
1088
|
townofhampton.ca
|
1094
1089
|
townofhighprairie.com
|
1095
|
-
townofhmcclv.com
|
1096
1090
|
townofhudsonbay.com
|
1097
1091
|
townofindianhead.com
|
1098
1092
|
townofkearney.com
|
@@ -1218,7 +1212,6 @@ villageofalliance.ca
|
|
1218
1212
|
villageofbitternlake.ca
|
1219
1213
|
villageofbriercrest.ca
|
1220
1214
|
villageofcentreville.ca
|
1221
|
-
villageofchauvin.ca
|
1222
1215
|
villageofconquest.ca
|
1223
1216
|
villageofdorintosh.sasktelwebsite.net
|
1224
1217
|
villageofduchess.com
|
@@ -3255,7 +3248,6 @@ vardanalys.se
|
|
3255
3248
|
vinnova.se
|
3256
3249
|
vr.se
|
3257
3250
|
vti.se
|
3258
|
-
y.lst.se
|
3259
3251
|
yhmyndigheten.se
|
3260
3252
|
|
3261
3253
|
// Swiss Cantons (Federal States)
|
@@ -4128,7 +4120,6 @@ jeffco.us
|
|
4128
4120
|
kitcarsoncounty.org
|
4129
4121
|
lakewood.org
|
4130
4122
|
littletongov.org
|
4131
|
-
metrodenvercfc.org
|
4132
4123
|
metromayors.org
|
4133
4124
|
minturn.org
|
4134
4125
|
mountain-village.co.us
|
@@ -4268,7 +4259,6 @@ waterfordct.org
|
|
4268
4259
|
watertownct.org
|
4269
4260
|
west-hartford.com
|
4270
4261
|
westbrookct.us
|
4271
|
-
weston-ct.com
|
4272
4262
|
wethersfieldct.com
|
4273
4263
|
willingtonct.org
|
4274
4264
|
wiltonct.org
|
@@ -4450,7 +4440,6 @@ pbgc.org
|
|
4450
4440
|
pcaobus.org
|
4451
4441
|
pepfar.net
|
4452
4442
|
predict.org
|
4453
|
-
project-open-data.github.io
|
4454
4443
|
psctool.us
|
4455
4444
|
r3coastal.com
|
4456
4445
|
radiosawa.com
|
@@ -4926,7 +4915,6 @@ odumgeorgia.com
|
|
4926
4915
|
oxfordgeorgia.org
|
4927
4916
|
peachcounty.net
|
4928
4917
|
peachtree-city.org
|
4929
|
-
pearson-ga.com
|
4930
4918
|
pembrokega.net
|
4931
4919
|
pinelakega.com
|
4932
4920
|
polkcountygeorgia.us
|
@@ -5555,7 +5543,6 @@ newhavenin.org
|
|
5555
5543
|
newtoncountyin.com
|
5556
5544
|
nirpc.org
|
5557
5545
|
nmanchester.org
|
5558
|
-
nobleco.squarespace.com
|
5559
5546
|
perrycountyil.org
|
5560
5547
|
porterco.org
|
5561
5548
|
portertownship.com
|
@@ -5937,7 +5924,6 @@ rppj.com
|
|
5937
5924
|
ruston.org
|
5938
5925
|
sbpg.net
|
5939
5926
|
sjbparish.com
|
5940
|
-
springhillla.com
|
5941
5927
|
state.lib.la.us
|
5942
5928
|
stfrancisville.net
|
5943
5929
|
stjamesla.com
|
@@ -7637,10 +7623,8 @@ shpnc.org
|
|
7637
7623
|
smithfield-nc.com
|
7638
7624
|
sosnc.com
|
7639
7625
|
southernpines.net
|
7640
|
-
southernshores.org
|
7641
7626
|
spindalenc.net
|
7642
7627
|
spring-lake.org
|
7643
|
-
sprucepineonline.com
|
7644
7628
|
startsquad.org
|
7645
7629
|
startwithyourheart.com
|
7646
7630
|
summerfieldgov.com
|
@@ -7742,7 +7726,6 @@ dvlnd.com
|
|
7742
7726
|
edgeley.com
|
7743
7727
|
edinburgnd.com
|
7744
7728
|
ellendalend.com
|
7745
|
-
emmonscounty.tripod.com
|
7746
7729
|
enderlinnd.com
|
7747
7730
|
fessendennd.com
|
7748
7731
|
finleynd.com
|
@@ -7843,7 +7826,6 @@ westfargo.org
|
|
7843
7826
|
williamsnd.com
|
7844
7827
|
willistonnd.com
|
7845
7828
|
wiltonnd.org
|
7846
|
-
wingnd.com
|
7847
7829
|
wishek-nd.com
|
7848
7830
|
wrtc.com
|
7849
7831
|
|
@@ -8688,7 +8670,6 @@ poncatribe-ne.org
|
|
8688
8670
|
portlions.net
|
8689
8671
|
potawatomi.org
|
8690
8672
|
powhatan.org
|
8691
|
-
prairieisland.org
|
8692
8673
|
pueblodecochiti.org
|
8693
8674
|
puyallup-tribe.com
|
8694
8675
|
redding-rancheria.com
|
@@ -8792,7 +8773,6 @@ attica.org
|
|
8792
8773
|
auroranewyork.us
|
8793
8774
|
austerlitzny.com
|
8794
8775
|
avon-ny.org
|
8795
|
-
b-ville.com
|
8796
8776
|
bainbridgeny.org
|
8797
8777
|
baldwinsville.org
|
8798
8778
|
batavianewyork.com
|
@@ -9328,7 +9308,6 @@ townofstonypoint.org
|
|
9328
9308
|
townofsweden.org
|
9329
9309
|
townofthompson.com
|
9330
9310
|
townofthurston.com
|
9331
|
-
townofticonderoga.com
|
9332
9311
|
townoftully.org
|
9333
9312
|
townofulster.org
|
9334
9313
|
townofunion.com
|
@@ -10151,7 +10130,6 @@ lehighcounty.org
|
|
10151
10130
|
lehmantownship.com
|
10152
10131
|
libertytwp-adamscounty.com
|
10153
10132
|
limerickpa.org
|
10154
|
-
lincolnboro.tripod.com
|
10155
10133
|
littlebritain.org
|
10156
10134
|
littlestownboro.org
|
10157
10135
|
liverpoolpa.us
|
@@ -10176,7 +10154,6 @@ manortwp.org
|
|
10176
10154
|
mantwp.com
|
10177
10155
|
marpletwp.com
|
10178
10156
|
marysvilleboro.com
|
10179
|
-
mckeesport.org
|
10180
10157
|
mckeesrocks.com
|
10181
10158
|
mechanicsburgborough.org
|
10182
10159
|
mediaborough.com
|
@@ -10384,7 +10361,6 @@ wesleyvilleborough.com
|
|
10384
10361
|
west-chester.com
|
10385
10362
|
westbradford.org
|
10386
10363
|
westcaln.org
|
10387
|
-
westconshohockenboro.com
|
10388
10364
|
westdeertownship.com
|
10389
10365
|
westeastonborough.org
|
10390
10366
|
westgroveborough.org
|
@@ -10542,7 +10518,6 @@ fortmillsc.org
|
|
10542
10518
|
fountaininn.org
|
10543
10519
|
gastonsc.org
|
10544
10520
|
georgetowncountysc.org
|
10545
|
-
greatfallssc.net
|
10546
10521
|
greenvillecounty.org
|
10547
10522
|
hamptoncountysc.org
|
10548
10523
|
hamptonsc.net
|
@@ -10561,7 +10536,6 @@ libertysc.com
|
|
10561
10536
|
lincolnvillesc.com
|
10562
10537
|
littlemountainsc.info
|
10563
10538
|
llronline.com
|
10564
|
-
mayesvillesc.com
|
10565
10539
|
midlandsareacfc.org
|
10566
10540
|
mtcroghan.com
|
10567
10541
|
mullinssc.us
|
@@ -11361,7 +11335,6 @@ vaao.org
|
|
11361
11335
|
vaboard.org
|
11362
11336
|
vadrs.org
|
11363
11337
|
vadsa.org
|
11364
|
-
vafamilyconnections.com
|
11365
11338
|
vafire.com
|
11366
11339
|
valottery.com
|
11367
11340
|
vaports.com
|
@@ -11513,7 +11486,6 @@ stjvt.com
|
|
11513
11486
|
swcrpc.org
|
11514
11487
|
thetfordvermont.us
|
11515
11488
|
tinmouthvt.org
|
11516
|
-
town-of-orwell.org
|
11517
11489
|
townofbrandon.com
|
11518
11490
|
townofcraftsbury.com
|
11519
11491
|
townofgeorgia.com
|
data/lib/gman/importer.rb
CHANGED
@@ -55,20 +55,23 @@ class Gman
|
|
55
55
|
|
56
56
|
def valid_domain?(domain, options={})
|
57
57
|
return false if domain.empty?
|
58
|
-
return reject(domain, "home. regex")
|
59
|
-
return reject(domain, "user. regex")
|
60
|
-
return reject(domain, "sites. regex")
|
61
|
-
return reject(domain, "weebly")
|
62
|
-
return reject(domain, "wordpress")
|
63
|
-
return reject(domain, "govoffice")
|
64
|
-
return reject(domain, "homestead")
|
65
|
-
return reject(domain, "wix.com")
|
66
|
-
return reject(domain, "blogspot.com")
|
67
|
-
return reject(domain, "
|
68
|
-
return reject(domain, "
|
69
|
-
return reject(domain, "
|
70
|
-
return reject(domain, "
|
71
|
-
return reject(domain, "
|
58
|
+
return reject(domain, "home. regex") if domain =~ /^home\./
|
59
|
+
return reject(domain, "user. regex") if domain =~ /^users?\./
|
60
|
+
return reject(domain, "sites. regex") if domain =~ /^sites?\./
|
61
|
+
return reject(domain, "weebly") if domain =~ /weebly\.com$/
|
62
|
+
return reject(domain, "wordpress") if domain =~ /wordpress\.com$/
|
63
|
+
return reject(domain, "govoffice") if domain =~ /govoffice\d?\.com$/
|
64
|
+
return reject(domain, "homestead") if domain =~ /homestead\.com$/
|
65
|
+
return reject(domain, "wix.com") if domain =~ /wix\.com$/
|
66
|
+
return reject(domain, "blogspot.com") if domain =~ /blogspot\.com$/
|
67
|
+
return reject(domain, "tripod.com") if domain =~ /tripod\.com$/
|
68
|
+
return reject(domain, "squarespace.com") if domain =~ /squarespace\.com$/
|
69
|
+
return reject(domain, "github.io") if domain =~ /github\.io$/
|
70
|
+
return reject(domain, "locality") if domain =~ Gman::LOCALITY_REGEX
|
71
|
+
return reject(domain, "blacklist") if BLACKLIST.include?(domain)
|
72
|
+
return reject(domain, "duplicate") if !options[:skip_dupe] && current.domains.include?(domain)
|
73
|
+
return reject(domain, "invalid") unless PublicSuffix.valid?(".#{domain}")
|
74
|
+
return reject(domain, "academic") if Swot::is_academic?(domain)
|
72
75
|
|
73
76
|
if !options[:skip_dupe] && subdomain = current.domains.find { |c| domain =~ /\.#{Regexp.escape(c)}$/}
|
74
77
|
return reject(domain, "subdomain of #{subdomain}")
|
@@ -122,13 +125,18 @@ class Gman
|
|
122
125
|
@resolver ||= Resolv::DNS.new(:nameserver => ["8.8.8.8","8.8.4.4"])
|
123
126
|
end
|
124
127
|
|
128
|
+
def resolve_without_errors
|
129
|
+
yield
|
130
|
+
rescue Resolv::ResolvError
|
131
|
+
false
|
132
|
+
end
|
133
|
+
|
125
134
|
# Verifies that the given domain has an MX record, and thus is valid
|
126
135
|
def domain_resolves?(domain)
|
127
136
|
domain = Addressable::URI.new(:host => domain).normalize.host
|
128
|
-
resolver.
|
129
|
-
resolver.getresource(domain, Resolv::DNS::Resource::IN::
|
130
|
-
|
131
|
-
false
|
137
|
+
resolve_without_errors { resolver.getaddress(domain) } ||
|
138
|
+
resolve_without_errors { resolver.getresource(domain, Resolv::DNS::Resource::IN::NS) } ||
|
139
|
+
resolve_without_errors { resolver.getresource(domain, Resolv::DNS::Resource::IN::MX) }
|
132
140
|
end
|
133
141
|
end
|
134
142
|
end
|
data/lib/gman/version.rb
CHANGED