[svn r3989] r4826@delle: sbehnel | 2008-10-25 20:19:53 +0200

 fixed doctype serialisation for internal subsets without public/system ID

--HG--
branch : trunk

CHANGES.txt

@@ -17,6 +17,10 @@ Features added
 Bugs fixed
 ----------
 
+* Internal DTD subsets that did not specify a system or public ID were
+  not serialised and did not appear in the docinfo property of
+  ElementTrees.
+
 * Fix a pre-Py3k warning when parsing from a gzip file in Py2.6.
 
 * Test suite fixes for libxml2 2.7.

src/lxml/lxml.etree.pyx

@@ -328,6 +328,9 @@ cdef public class _Document [ type LxmlDocumentType, object LxmlDocument ]:
             return None
         return _elementFactory(self, c_node)
 
+    cdef bint hasdoctype(self):
+        return self._c_doc.intSubset is not NULL
+
     cdef getdoctype(self):
         cdef tree.xmlDtd* c_dtd
         cdef xmlNode* c_root_node
@@ -520,6 +523,8 @@ cdef class DocInfo:
             elif system_url:
                 return u'<!DOCTYPE %s SYSTEM "%s">' % (
                     root_name, system_url)
+            elif self._doc.hasdoctype():
+                return u'<!DOCTYPE %s>' % root_name
             else:
                 return u""
 

src/lxml/serializer.pxi

@@ -189,9 +189,7 @@ cdef void _writeDtdToBuffer(tree.xmlOutputBuffer* c_buffer,
     cdef tree.xmlDtd* c_dtd
     cdef xmlNode* c_node
     c_dtd = c_doc.intSubset
-    if c_dtd == NULL or c_dtd.name == NULL:
-        return
-    if c_dtd.ExternalID == NULL and c_dtd.SystemID == NULL:
+    if c_dtd is NULL or c_dtd.name is NULL:
         return
     if cstd.strcmp(c_root_name, c_dtd.name) != 0:
         return

src/lxml/tests/test_etree.py

@@ -2097,6 +2097,24 @@ def test_docinfo_empty(self):
         self.assertEquals(docinfo.root_name,   'html')
         self.assertEquals(docinfo.doctype, '')
 
+    def test_docinfo_name_only(self):
+        etree = self.etree
+        xml = _bytes('<!DOCTYPE root><root></root>')
+        tree = etree.parse(BytesIO(xml))
+        docinfo = tree.docinfo
+        self.assertEquals(docinfo.encoding,    "UTF-8")
+        self.assertEquals(docinfo.xml_version, "1.0")
+        self.assertEquals(docinfo.public_id,   None)
+        self.assertEquals(docinfo.system_url,  None)
+        self.assertEquals(docinfo.root_name,   'root')
+        self.assertEquals(docinfo.doctype, '<!DOCTYPE root>')
+
+    def test_doctype_name_only_roundtrip(self):
+        etree = self.etree
+        xml = _bytes('<!DOCTYPE root>\n<root/>')
+        tree = etree.parse(BytesIO(xml))
+        self.assertEquals(xml, etree.tostring(tree))
+
     def test_xml_base(self):
         etree = self.etree
         root = etree.XML(_bytes("<root/>"), base_url="http://no/such/url")