From ee6546bc2a08200b805d5e6f57c679a4355c1f3e Mon Sep 17 00:00:00 2001
From: loyalsoldier <loyalsoldier@outlook.com>
Date: Fri, 10 Jan 2020 16:10:06 +0800
Subject: [PATCH] Validate domains

---
 .github/workflows/build.yml | 23 +++++++++++++++++------
 1 file changed, 17 insertions(+), 6 deletions(-)

diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml
index 37e727b1..f8f87c3e 100644
--- a/.github/workflows/build.yml
+++ b/.github/workflows/build.yml
@@ -29,12 +29,12 @@ jobs:
           echo "::set-env name=APPLE_DOMAINS_URL::https://raw.githubusercontent.com/felixonmars/dnsmasq-china-list/master/apple.china.conf"
           echo "::set-env name=GFWLIST_DOMAINS_URL::https://cokebar.github.io/gfwlist2dnsmasq/gfwlist_domain.txt"
           echo "::set-env name=GREATFIRE_DOMAINS_URL::https://raw.githubusercontent.com/wongsyrone/domain-block-list/master/domains.txt"
+          echo "::set-env name=PROFILES_URL::https://raw.githubusercontent.com/ConnersHua/Profiles/master/Shadow/Pro.conf"
           echo "::set-env name=GEQ1AN_RULES_APPLE_URL::https://raw.githubusercontent.com/GeQ1an/Rules/master/QuantumultX/Filter/Apple.list"
           echo "::set-env name=GEQ1AN_RULES_MICROSOFT_URL::https://raw.githubusercontent.com/GeQ1an/Rules/master/QuantumultX/Filter/Microsoft.list"
           echo "::set-env name=GEQ1AN_RULES_GLOBAL_MEDIA_URL::https://raw.githubusercontent.com/GeQ1an/Rules/master/QuantumultX/Filter/GMedia.list"
           echo "::set-env name=GEQ1AN_RULES_OUTSIDE_URL::https://raw.githubusercontent.com/GeQ1an/Rules/master/QuantumultX/Filter/Outside.list"
           echo "::set-env name=GEQ1AN_RULES_REJECT_URL::https://raw.githubusercontent.com/GeQ1an/Rules/master/QuantumultX/Filter/AdBlock.list"
-          echo "::set-env name=PROFILES_REJECT_URL::https://raw.githubusercontent.com/ConnersHua/Profiles/master/Shadow/Pro.conf"
           echo "::set-env name=LHIE1_RULES_REJECT_URL::https://raw.githubusercontent.com/lhie1/Rules/master/Auto/REJECT.conf"
           echo "::set-env name=GOPATH::$(dirname $GITHUB_WORKSPACE)"
           echo "::add-path::$(dirname $GITHUB_WORKSPACE)/bin"
@@ -73,6 +73,7 @@ jobs:
           curl -sSL $GREATFIRE_DOMAINS_URL >> temp-proxy.txt
           curl -sSL $GOOGLE_DOMAINS_URL | awk -F '/' '{print $2}' >> temp-proxy.txt
           curl -sSL $APPLE_DOMAINS_URL | awk -F '/' '{print $2}' >> temp-proxy.txt
+          curl -sSL $PROFILES_URL | awk -F ',' '/^DOMAIN(,|\-SUFFIX,).+PROXY/ {print $2}' >> temp-proxy.txt
           curl -sSL $GEQ1AN_RULES_APPLE_URL | awk -F ',' '/^(HOST|DOMAIN)(,|\-SUFFIX,).+Apple/ {print $2}' >> temp-proxy.txt
           curl -sSL $GEQ1AN_RULES_MICROSOFT_URL | awk -F ',' '/^(HOST|DOMAIN)(,|\-SUFFIX,).+Microsoft/ {print $2}' >> temp-proxy.txt
           curl -sSL $GEQ1AN_RULES_GLOBAL_MEDIA_URL | awk -F ',' '/^(HOST|DOMAIN)(,|\-SUFFIX,).+GMedia/ {print $2}' >> temp-proxy.txt
@@ -80,7 +81,7 @@ jobs:
           
       - name: Get and add reject domains into temp-reject.txt file
         run: |
-          curl -sSL $PROFILES_REJECT_URL | awk -F ',' '/^DOMAIN(,|\-SUFFIX,).+REJECT/ {print $2}' > temp-reject.txt
+          curl -sSL $PROFILES_URL | awk -F ',' '/^DOMAIN(,|\-SUFFIX,).+REJECT/ {print $2}' > temp-reject.txt
           curl -sSL $GEQ1AN_RULES_REJECT_URL | awk -F ',' '/^(HOST|DOMAIN)(,|\-SUFFIX,).+AdBlock/ {print $2}' >> temp-reject.txt
           curl -sSL $LHIE1_RULES_REJECT_URL | awk -F ',' '/^DOMAIN(,|\-SUFFIX,).+REJECT/ {print $2}' >> temp-reject.txt
 
@@ -89,11 +90,20 @@ jobs:
           cat proxy.txt >> temp-proxy.txt
           cat direct.txt >> temp-direct.txt
 
-      - name: Remove repeated domains and write domains to appropriate list
+      - name: Remove repeated domains, sort domains and write domains to appropriate list
         run: |
-          cat temp-proxy.txt | sort --ignore-case -u > $GOPATH/src/$GEOSITE_REPO/data/proxylist
-          cat temp-direct.txt | sort --ignore-case -u > $GOPATH/src/$GEOSITE_REPO/data/directlist
-          cat temp-reject.txt | sort --ignore-case -u > $GOPATH/src/$GEOSITE_REPO/data/rejectlist
+          cat temp-proxy.txt | sort --ignore-case -u > proxy-sort.txt
+          cat temp-direct.txt | sort --ignore-case -u > direct-sort.txt
+          cat temp-reject.txt | sort --ignore-case -u > reject-sort.txt
+
+      - name: Filter valid domains
+        run: |
+          cat proxy-sort.txt | awk '/^([a-zA-Z0-9-]+\.)+[a-zA-Z0-9-]+$/{print $0}' > $GOPATH/src/$GEOSITE_REPO/data/proxylist
+          cat proxy-sort.txt | awk '!/^([a-zA-Z0-9-]+\.)+[a-zA-Z0-9-]+$/{print $0}' > proxy-filter.txt
+          cat direct-sort.txt | awk '/^([a-zA-Z0-9-]+\.)+[a-zA-Z0-9-]+$/{print $0}' > $GOPATH/src/$GEOSITE_REPO/data/directlist
+          cat direct-sort.txt | awk '!/^([a-zA-Z0-9-]+\.)+[a-zA-Z0-9-]+$/{print $0}' > direct-filter.txt
+          cat reject-sort.txt | awk '/^([a-zA-Z0-9-]+\.)+[a-zA-Z0-9-]+$/{print $0}' > $GOPATH/src/$GEOSITE_REPO/data/rejectlist
+          cat reject-sort.txt | awk '!/^([a-zA-Z0-9-]+\.)+[a-zA-Z0-9-]+$/{print $0}' > reject-filter.txt
 
       - name: Add list into appropriate category file
         run: |
@@ -106,6 +116,7 @@ jobs:
         run: |
           domain-list-community
           mv ./dlc.dat ./publish/geosite.dat
+          mv ./*-filter.txt ./publish/
 
       - name: Release dat files
         uses: Ricky-Hao/action-release@master