Changeset 385

Timestamp:

Oct 22, 2006, 4:57:40 PM (17 years ago)

Author:

cmlenz

Message:

• The HTMLParser class and the HTML function now accept an encoding parameter to properly deal with bytestring input (defaults to UTF-8).
• The TemplateLoader class can now also be initialized from a string for the search path, for cases where the search path contains only a single directory.

Location:

trunk/genshi

Files:

• input.py (modified) (4 diffs)
• template.py (modified) (2 diffs)
• tests/input.py (modified) (1 diff)
• tests/template.py (modified) (2 diffs)

trunk/genshi/input.py

@@ -77 +77 @@
 
     def __init__(self, source, filename=None):
-        """Initialize the parser for the given XML text.
+        """Initialize the parser for the given XML input.
         
         @param source: the XML text as a file-like object
@@ -251 +251 @@
                               'param'])
 
-    def __init__(self, source, filename=None):
+    def __init__(self, source, filename=None, encoding='utf-8'):
+        """Initialize the parser for the given HTML input.
+        
+        @param source: the HTML text as a file-like object
+        @param filename: the name of the file, if known
+        @param filename: encoding of the file; ignored if the input is unicode
+        """
         html.HTMLParser.__init__(self)
         self.source = source
         self.filename = filename
+        self.encoding = encoding
         self._queue = []
         self._open_tags = []
@@ -322 +329 @@
 
     def handle_data(self, text):
+        if not isinstance(text, unicode):
+            text = text.decode(self.encoding, 'replace')
         self._enqueue(TEXT, text)
 
@@ -344 +353 @@
 
 
-def HTML(text):
-    return Stream(list(HTMLParser(StringIO(text))))
+def HTML(text, encoding='utf-8'):
+    return Stream(list(HTMLParser(StringIO(text), encoding=encoding)))
 
 def _coalesce(stream):

trunk/genshi/template.py

@@ -1288 +1288 @@
         
         @param search_path: a list of absolute path names that should be
-            searched for template files
+            searched for template files, or a string containing a single
+            absolute path
         @param auto_reload: whether to check the last modification time of
             template files, and reload them if they have changed
@@ -1297 +1298 @@
         if self.search_path is None:
             self.search_path = []
+        elif isinstance(self.search_path, basestring):
+            self.search_path = [self.search_path]
         self.auto_reload = auto_reload
         self._cache = LRUCache(max_cache_size)

trunk/genshi/tests/input.py

@@ -121 +121 @@
             self.assertEqual((None, 1, 6), pos)
 
+    def test_input_encoding(self):
+        text = u'<div>\xf6</div>'.encode('iso-8859-1')
+        events = list(HTMLParser(StringIO(text), encoding='iso-8859-1'))
+        kind, data, pos = events[1]
+        self.assertEqual(Stream.TEXT, kind)
+        self.assertEqual(u'\xf6', data)
+
     def test_unicode_input(self):
         text = u'<div>\u2013</div>'

trunk/genshi/tests/template.py

@@ -1141 +1141 @@
 
 
+    # FIXME
+    #def test_empty_lines(self):
+    #    tmpl = TextTemplate("""Your items:
+    #
+    #    #for item in items
+    #      * ${item}
+    #
+    #    #end""")
+    #    self.assertEqual("""Your items:
+    #      * 0
+    #      * 1
+    #      * 2
+    #    """, tmpl.generate(items=range(3)).render('text'))
+
+
 class TemplateLoaderTestCase(unittest.TestCase):
     """Tests for the template loader."""
@@ -1149 +1164 @@
     def tearDown(self):
         shutil.rmtree(self.dirname)
+
+    def test_search_path_empty(self):
+        loader = TemplateLoader()
+        self.assertEqual([], loader.search_path)
+
+    def test_search_path_as_string(self):
+        loader = TemplateLoader(self.dirname)
+        self.assertEqual([self.dirname], loader.search_path)
 
     def test_relative_include_samedir(self):