From 73823e3eb9b3695a6777dc02879c286e14e061b4 Mon Sep 17 00:00:00 2001
From: n1474335 <n1474335@gmail.com>
Date: Wed, 6 Sep 2017 16:43:30 +0000
Subject: [PATCH] Improved domain name regex

---
 .travis.yml                        | 1 -
 src/core/config/OperationConfig.js | 2 +-
 src/core/operations/Extract.js     | 7 ++-----
 src/core/operations/StrUtils.js    | 2 +-
 4 files changed, 4 insertions(+), 8 deletions(-)

diff --git a/.travis.yml b/.travis.yml
index c6bd204c..5f039562 100644
--- a/.travis.yml
+++ b/.travis.yml
@@ -4,7 +4,6 @@ node_js:
 install: npm install
 before_script:
   - npm install -g grunt
-  - if [ "$TRAVIS_TAG" ]; then git checkout -b travis-build; fi
 script:
   - grunt lint
   - grunt test
diff --git a/src/core/config/OperationConfig.js b/src/core/config/OperationConfig.js
index 4493bf9c..bc2bf399 100755
--- a/src/core/config/OperationConfig.js
+++ b/src/core/config/OperationConfig.js
@@ -2140,7 +2140,7 @@ const OperationConfig = {
         ]
     },
     "Extract domains": {
-        description: "Extracts domain names with common Top-Level Domains (TLDs).<br>Note that this will not include paths. Use <strong>Extract URLs</strong> to find entire URLs.",
+        description: "Extracts domain names.<br>Note that this will not include paths. Use <strong>Extract URLs</strong> to find entire URLs.",
         run: Extract.runDomains,
         inputType: "string",
         outputType: "string",
diff --git a/src/core/operations/Extract.js b/src/core/operations/Extract.js
index 3f9d2d29..d822830a 100755
--- a/src/core/operations/Extract.js
+++ b/src/core/operations/Extract.js
@@ -187,11 +187,8 @@ const Extract = {
      * @returns {string}
      */
     runDomains: function(input, args) {
-        let displayTotal = args[0],
-            protocol = "https?://",
-            hostname = "[-\\w\\.]+",
-            tld = "\\.(?:com|net|org|biz|info|co|uk|onion|int|mobi|name|edu|gov|mil|eu|ac|ae|af|de|ca|ch|cn|cy|es|gb|hk|il|in|io|tv|me|nl|no|nz|ro|ru|tr|us|az|ir|kz|uz|pk)+",
-            regex = new RegExp("(?:" + protocol + ")?" + hostname + tld, "ig");
+        const displayTotal = args[0],
+            regex = /\b((?=[a-z0-9-]{1,63}\.)(xn--)?[a-z0-9]+(-[a-z0-9]+)*\.)+[a-z]{2,63}\b/ig;
 
         return Extract._search(input, regex, null, displayTotal);
     },
diff --git a/src/core/operations/StrUtils.js b/src/core/operations/StrUtils.js
index b771eb36..71823053 100755
--- a/src/core/operations/StrUtils.js
+++ b/src/core/operations/StrUtils.js
@@ -40,7 +40,7 @@ const StrUtils = {
         },
         {
             name: "Domain",
-            value: "(?:(https?):\\/\\/)?([-\\w.]+)\\.(com|net|org|biz|info|co|uk|onion|int|mobi|name|edu|gov|mil|eu|ac|ae|af|de|ca|ch|cn|cy|es|gb|hk|il|in|io|tv|me|nl|no|nz|ro|ru|tr|us|az|ir|kz|uz|pk)+"
+            value: "\\b((?=[a-z0-9-]{1,63}\\.)(xn--)?[a-z0-9]+(-[a-z0-9]+)*\\.)+[a-z]{2,63}\\b"
         },
         {
             name: "Windows file path",