From e05076f181ce413fa658218551ac367a5372678f Mon Sep 17 00:00:00 2001
From: Dmitry Bochkarev <dimabochkarev@gmail.com>
Date: Fri, 23 Oct 2015 00:29:35 +0500
Subject: [PATCH] =?UTF-8?q?fix(html):=20=D0=BF=D0=BE=D0=B4=D0=B4=D0=B5?=
 =?UTF-8?q?=D1=80=D0=B6=D0=BA=D0=B0=20=D0=BE=D1=82=D0=BD=D0=BE=D1=81=D0=B8?=
 =?UTF-8?q?=D1=82=D0=B5=D0=BB=D1=8C=D0=BD=D1=8B=D1=85=20=D0=BF=D1=83=D1=82?=
 =?UTF-8?q?=D0=B5=D0=B9?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

https://jira.railsc.ru/browse/SERVICES-594
---
 lib/string_tools/html.rb | 16 ++++++++++------
 spec/html_spec.rb        | 24 +++++++++++++++++++++++-
 2 files changed, 33 insertions(+), 7 deletions(-)

diff --git a/lib/string_tools/html.rb b/lib/string_tools/html.rb
index ff4b19d..d3d66b9 100644
--- a/lib/string_tools/html.rb
+++ b/lib/string_tools/html.rb
@@ -1,6 +1,6 @@
 # coding: utf-8
 require 'loofah'
-require 'uri'
+require 'addressable/uri'
 
 module StringTools
   module HTML
@@ -44,16 +44,20 @@ def initialize(options)
 
       def scrub(node)
         return unless node.name == 'a'.freeze
-        uri = URI.parse(node['href'.freeze])
+        href = node['href']
+        return if href.blank?
+        uri = Addressable::URI.parse(href).normalize
+        return unless uri.host
         node.swap(node.children) unless whitelisted? uri.host
-      rescue URI::InvalidURIError => _
+      rescue
+        # в любой непонятной ситуации просто удаляем ссылку
         node.swap(node.children)
       end
 
       def whitelisted?(domain)
-        host_parts = domain.split('.'.freeze).reverse!
-        host = host_parts[0] # com, ru ...
-        1.upto(host_parts.length - 1) do |i|
+        host_parts = domain.split('.'.freeze)
+        host = host_parts[-1] # com, ru ...
+        (host_parts.length - 2).downto(0) do |i|
           subdomain = host_parts[i]
           host = "#{subdomain}.#{host}"
           return true if @whitelist.include? host
diff --git a/spec/html_spec.rb b/spec/html_spec.rb
index 2047ec0..9c94f8a 100644
--- a/spec/html_spec.rb
+++ b/spec/html_spec.rb
@@ -48,11 +48,14 @@
     end
 
     context 'when whitelist passed' do
-      subject { StringTools::HTML.remove_links(html, whitelist: ['yandex.ru']) }
+      subject { StringTools::HTML.remove_links(html, whitelist: ['yandex.ru', 'pulscen.com.ua']) }
 
       context 'domain link match to whitelisted' do
         let(:html) do
           <<-MARKUP
+            <a href="https://firm.pulscen.com.ua">firm.pulscen.com.ua</a>
+            <a href="https://pulscen.com.ua">pulscen.com.ua</a>
+            <a href="https://com.ua">com.ua</a>
             <a href="https://google.com"><span>goo</span><span>gle</span></a>
             <a href="https://yandex.ru"><span>yan</span><span>dex</span></a>
           MARKUP
@@ -60,6 +63,9 @@
 
         it 'should keep only whitelisted links' do
           is_expected.to eq(<<-MARKUP)
+            <a href="https://firm.pulscen.com.ua">firm.pulscen.com.ua</a>
+            <a href="https://pulscen.com.ua">pulscen.com.ua</a>
+            com.ua
             <span>goo</span><span>gle</span>
             <a href="https://yandex.ru"><span>yan</span><span>dex</span></a>
           MARKUP
@@ -99,6 +105,22 @@
           MARKUP
         end
       end
+
+      context 'content with relative links' do
+        let(:html) do
+          <<-MARKUP
+            <a href="https://google.com"><span>goo</span><span>gle</span></a>
+            <a href="yandex.ru"><span>yan</span><span>dex</span></a>
+          MARKUP
+        end
+
+        it 'should keep relative links' do
+          is_expected.to eq(<<-MARKUP)
+            <span>goo</span><span>gle</span>
+            <a href="yandex.ru"><span>yan</span><span>dex</span></a>
+          MARKUP
+        end
+      end
     end
   end
 end