Merge pull request #878 from ruby/pz-invalid-url

Only parse valid URLs
ruby · Apr 13, 2022 · bf519a6 · bf519a6
2 parents 32aca21 + 2bd8fcd
commit bf519a6
Show file tree

Hide file tree

Showing 2 changed files with 4 additions and 1 deletion.
diff --git a/lib/rdoc/markup/to_html.rb b/lib/rdoc/markup/to_html.rb
@@ -61,12 +61,14 @@ def initialize options, markup = nil
   #
   # These methods are used by regexp handling markup added by RDoc::Markup#add_regexp_handling.
 
+  URL_CHARACTERS_REGEXP_STR = /[A-Za-z0-9\-._~:\/\?#\[\]@!$&'\(\)*+,;%=]/.source
+
   ##
   # Adds regexp handlings.
 
   def init_regexp_handlings
     # external links
-    @markup.add_regexp_handling(/(?:link:|https?:|mailto:|ftp:|irc:|www\.)\S+\w/,
+    @markup.add_regexp_handling(/(?:link:|https?:|mailto:|ftp:|irc:|www\.)#{URL_CHARACTERS_REGEXP_STR}+\w/,
                                 :HYPERLINK)
     init_link_notation_regexp_handlings
   end

diff --git a/test/rdoc/test_rdoc_markup_to_html.rb b/test/rdoc/test_rdoc_markup_to_html.rb
@@ -707,6 +707,7 @@ def test_convert_TIDYLINK_irc
   def test_convert_with_exclude_tag
     assert_equal "\n<p><code>aaa</code>[:symbol]</p>\n", @to.convert('+aaa+[:symbol]')
     assert_equal "\n<p><code>aaa[:symbol]</code></p>\n", @to.convert('+aaa[:symbol]+')
+    assert_equal "\n<p><code>https:</code>-foobar</p>\n", @to.convert('<tt>https:</tt>-foobar')
     assert_equal "\n<p><a href=\":symbol\">aaa</a></p>\n", @to.convert('aaa[:symbol]')
   end