Ignore low-confidence CharlockHolmes guesses when parsing link cards (…

…mastodon#9510) * Add failing test for windows-1251 link cards * Ignore low-confidence CharlockHolmes guesses Fixes mastodon#9466 * Fix no method error when charlock holmes cannot detect charset
ProjectMyosotis · Oct 2, 2019 · 0486c21 · 0486c21
1 parent 9cf0120
commit 0486c21
Show file tree

Hide file tree

Showing 3 changed files with 30 additions and 1 deletion.
diff --git a/app/services/fetch_link_card_service.rb b/app/services/fetch_link_card_service.rb
@@ -137,7 +137,8 @@ def attempt_opengraph
     detector.strip_tags = true
 
     guess      = detector.detect(@html, @html_charset)
-    page       = Nokogiri::HTML(@html, nil, guess&.fetch(:encoding, nil))
+    encoding   = guess&.fetch(:confidence, 0).to_i > 60 ? guess&.fetch(:encoding, nil) : nil
+    page       = Nokogiri::HTML(@html, nil, encoding)
     player_url = meta_property(page, 'twitter:player')
 
     if player_url && !bad_url?(Addressable::URI.parse(player_url))

diff --git a/spec/fixtures/requests/windows-1251.txt b/spec/fixtures/requests/windows-1251.txt
@@ -0,0 +1,17 @@
+HTTP/1.1 200 OK
+server: nginx
+date: Wed, 12 Dec 2018 13:14:03 GMT
+content-type: text/html
+content-length: 190
+accept-ranges: bytes
+
+<!DOCTYPE html>
+<html>
+<head>
+  <meta http-equiv="Content-Type" content="text/html; charset=windows-1251" />
+  <title>����� �����</title>
+</head>
+<body>
+  <p>����� �����</p>
+</body>
+</html>
diff --git a/spec/services/fetch_link_card_service_spec.rb b/spec/services/fetch_link_card_service_spec.rb
@@ -17,6 +17,8 @@
     stub_request(:head, 'https://github.com/qbi/WannaCry').to_return(status: 404)
     stub_request(:head, 'http://example.com/test-').to_return(status: 200, headers: { 'Content-Type' => 'text/html' })
     stub_request(:get, 'http://example.com/test-').to_return(request_fixture('idn.txt'))
+    stub_request(:head, 'http://example.com/windows-1251').to_return(status: 200, headers: { 'Content-Type' => 'text/html' })
+    stub_request(:get, 'http://example.com/windows-1251').to_return(request_fixture('windows-1251.txt'))
 
     subject.call(status)
   end
@@ -57,6 +59,15 @@
       end
     end
 
+    context do
+      let(:status) { Fabricate(:status, text: 'Check out http://example.com/windows-1251') }
+
+      it 'works with windows-1251' do
+        expect(a_request(:get, 'http://example.com/windows-1251')).to have_been_made.at_least_once
+        expect(status.preview_cards.first.title).to eq('сэмпл текст')
+      end
+    end
+
     context do
       let(:status) { Fabricate(:status, text: 'テストhttp://example.com/日本語') }