Fix the domain regex beeing to greedy. It parsed stuff like "http://example.com/test/bla.net_foo_123.jpg" as the domain "example.com/test/bla.net". Now it's fixed ^.^"

BonsaiDen · BonsaiDen · commit 718c82179a7d · 2010-05-03T15:31:44.000+02:00
diff --git a/tests.py b/tests.py
@@ -57,6 +57,16 @@ def test_url_parentheses(self):
         self.assertEqual(result.html, u'text (<a href="http://example.com">http://example.com</a>)')
         self.assertEqual(result.urls, [u'http://example.com'])
     
+    def test_url_underscore(self):
+        result = self.parser.parse(u'text http://example.com/test/foo_123.jpg')
+        self.assertEqual(result.html, u'text <a href="http://example.com/test/foo_123.jpg">http://example.com/test/foo...</a>')
+        self.assertEqual(result.urls, [u'http://example.com/test/foo_123.jpg'])
+    
+    def test_url_underscore_dot(self):
+        result = self.parser.parse(u'text http://example.com/test/bla.net_foo_123.jpg')
+        self.assertEqual(result.html, u'text <a href="http://example.com/test/bla.net_foo_123.jpg">http://example.com/test/bla...</a>')
+        self.assertEqual(result.urls, [u'http://example.com/test/bla.net_foo_123.jpg'])
+    
     def test_url_amp_lang_equals(self):
         result = self.parser.parse(u'Check out http://search.twitter.com/search?q=avro&lang=en')
         self.assertEqual(result.html, u'Check out <a href="http://search.twitter.com/search?q=avro&amp;lang=en">http://search.twitter.com/s...</a>')
diff --git a/ttp.py b/ttp.py
@@ -45,7 +45,7 @@
 
 # URLs
 PRE_CHARS = ur'(?:[^/"\':!=]|^|\:)'
-DOMAIN_CHARS = ur'([\.-]|[^\s_\!\.])+\.[a-z]{2,}(?::[0-9]+)?'
+DOMAIN_CHARS = ur'([\.-]|[^\s_\!\.\/])+\.[a-z]{2,}(?::[0-9]+)?'
 PATH_CHARS = ur'(?:[\.,]?[%s!\*\'\(\);:=\+\$/%s#\[\]\-_,~@])' % (UTF_CHARS, '%')
 QUERY_CHARS = ur'[a-z0-9!\*\'\(\);:&=\+\$/%#\[\]\-_\.,~]'
 
