Fixing URL shortening for Arabic URLs.

We have to check if the URL itself contains Arabic characters, not only the full text where the URL is in. Fixes: CV2-3814.
meedan · Oct 5, 2023 · 5da8afa · 5da8afa
1 parent dfbb1bc
commit 5da8afa
Show file tree

Hide file tree

Showing 2 changed files with 13 additions and 3 deletions.
diff --git a/lib/url_rewriter.rb b/lib/url_rewriter.rb
@@ -23,7 +23,7 @@ def self.utmize(url, source)
   def self.shorten_and_utmize_urls(input_text, source = nil, owner = nil)
     text = input_text
     # Encode URLs in Arabic which are not detected by the URL extraction methods
-    text = text.gsub(/https?:\/\/[\S]+/) { |url| Addressable::URI.escape(url) } if input_text =~ /\p{Arabic}/
+    text = text.gsub(/https?:\/\/[\S]+/) { |url| url =~ /\p{Arabic}/ ? Addressable::URI.escape(url) : url } if input_text =~ /\p{Arabic}/
     entities = Twitter::TwitterText::Extractor.extract_urls_with_indices(text, extract_url_without_protocol: true)
     # Ruby 2.7 freezes the empty string from nil.to_s, which causes an error within the rewriter
     Twitter::TwitterText::Rewriter.rewrite_entities(text || '', entities) do |entity, _codepoints|

diff --git a/test/lib/url_rewriter_test.rb b/test/lib/url_rewriter_test.rb
@@ -54,9 +54,19 @@ def teardown
   test 'should shorten Arabic URL' do
     shortened = nil
     stub_configs({ 'short_url_host_display' => 'https://chck.media' }) do
-      shortened = UrlRewriter.shorten_and_utmize_urls('Visit https://fatabyyano.net/هذا-المقطع-ليس-لاشتباكات-حديثة-بين-الج/ for more information.', nil)
+      shortened = UrlRewriter.shorten_and_utmize_urls('Visit https://fatabyyano.net/هذا-المقطع-قديم،-ولا-يبين-لحظة-إنقاذ-شا/ for more information.', nil)
     end
-    assert_equal 'https://fatabyyano.net/%D9%87%D8%B0%D8%A7-%D8%A7%D9%84%D9%85%D9%82%D8%B7%D8%B9-%D9%84%D9%8A%D8%B3-%D9%84%D8%A7%D8%B4%D8%AA%D8%A8%D8%A7%D9%83%D8%A7%D8%AA-%D8%AD%D8%AF%D9%8A%D8%AB%D8%A9-%D8%A8%D9%8A%D9%86-%D8%A7%D9%84%D8%AC/', Shortener::ShortenedUrl.last.url
+    assert_equal 'https://fatabyyano.net/%D9%87%D8%B0%D8%A7-%D8%A7%D9%84%D9%85%D9%82%D8%B7%D8%B9-%D9%82%D8%AF%D9%8A%D9%85%D8%8C-%D9%88%D9%84%D8%A7-%D9%8A%D8%A8%D9%8A%D9%86-%D9%84%D8%AD%D8%B8%D8%A9-%D8%A5%D9%86%D9%82%D8%A7%D8%B0-%D8%B4%D8%A7/', Shortener::ShortenedUrl.last.url
     assert_match /^Visit https:\/\/chck\.media\/[a-zA-Z0-9]+ for more information\.$/, shortened
   end
+
+  test 'should not shorten decoded Arabic URL' do
+    url = 'https://fatabyyano.net/%da%af%d8%b1%d8%aa%db%95-%da%a4%db%8c%d8%af%db%8c%db%86%db%8c%db%8c%db%95%da%a9%db%95-%d8%b3%d8%a7%d8%ae%d8%aa%db%95%db%8c%db%95-%d9%88-%d9%84%d8%a7%d9%81%d8%a7%d9%88-%d9%88-%d8%b2%d8%b1%db%8c%d8%a7/'
+    shortened = nil
+    stub_configs({ 'short_url_host_display' => 'https://chck.media' }) do
+      shortened = UrlRewriter.shorten_and_utmize_urls("فتبينوا  | Visit #{url} for more information.", nil)
+    end
+    assert_equal url, Shortener::ShortenedUrl.last.url
+    assert_match /^فتبينوا  \| Visit https:\/\/chck\.media\/[a-zA-Z0-9]+ for more information\.$/, shortened
+  end
 end