Make _ElementTree.deepcopy() properly handle comments/PI's before doctype.

3 files changed, 24 insertions, 11 deletions

diff --git a/src/lxml/apihelpers.pxi b/src/lxml/apihelpers.pxi
index 0492b0be..c529384a 100644
--- a/src/lxml/apihelpers.pxi
+++ b/src/lxml/apihelpers.pxi
@@ -1069,13 +1069,21 @@ cdef int _copyNonElementSiblings(xmlNode* c_node, xmlNode* c_target) except -1:
     cdef xmlNode* c_copy
     cdef xmlNode* c_sibling = c_node
     while c_sibling.prev != NULL and \
-            (c_sibling.prev.type == tree.XML_PI_NODE or \
-                 c_sibling.prev.type == tree.XML_COMMENT_NODE):
+            (c_sibling.prev.type == tree.XML_PI_NODE or
+             c_sibling.prev.type == tree.XML_COMMENT_NODE or
+             c_sibling.prev.type == tree.XML_DTD_NODE):
         c_sibling = c_sibling.prev
     while c_sibling != c_node:
-        c_copy = tree.xmlDocCopyNode(c_sibling, c_target.doc, 1)
-        if c_copy is NULL:
-            raise MemoryError()
+        if c_sibling.type == tree.XML_DTD_NODE:
+            c_copy = <xmlNode*>_copyDtd(<tree.xmlDtd*>c_sibling)
+            if c_sibling == <xmlNode*>c_node.doc.intSubset:
+                c_target.doc.intSubset = <tree.xmlDtd*>c_copy
+            else: # c_sibling == c_node.doc.extSubset
+                c_target.doc.extSubset = <tree.xmlDtd*>c_copy
+        else:
+            c_copy = tree.xmlDocCopyNode(c_sibling, c_target.doc, 1)
+            if c_copy is NULL:
+                raise MemoryError()
         tree.xmlAddPrevSibling(c_target, c_copy)
         c_sibling = c_sibling.next
     while c_sibling.next != NULL and \
diff --git a/src/lxml/lxml.etree.pyx b/src/lxml/lxml.etree.pyx
index d5f43f67..74e7e93a 100644
--- a/src/lxml/lxml.etree.pyx
+++ b/src/lxml/lxml.etree.pyx
@@ -1903,12 +1903,6 @@ cdef public class _ElementTree [ type LxmlElementTreeType,
             assert root is not None
             _assertValidNode(root)
             _copyNonElementSiblings(self._context_node._c_node, root._c_node)
-            doc = root._doc
-            c_doc = self._context_node._doc._c_doc
-            if c_doc.intSubset is not NULL and doc._c_doc.intSubset is NULL:
-                doc._c_doc.intSubset = _copyDtd(c_doc.intSubset)
-            if c_doc.extSubset is not NULL and not doc._c_doc.extSubset is NULL:
-                doc._c_doc.extSubset = _copyDtd(c_doc.extSubset)
             return _elementTreeFactory(None, root)
         elif self._doc is not None:
             _assertValidDoc(self._doc)
diff --git a/src/lxml/tests/test_etree.py b/src/lxml/tests/test_etree.py
index 25ad998b..8131881e 100644
--- a/src/lxml/tests/test_etree.py
+++ b/src/lxml/tests/test_etree.py
@@ -576,6 +576,17 @@ class ETreeOnlyTestCase(HelperTestCase):
         self.assertEqual(_bytes("<test/>"),
                           tostring(root2))
 
+    def test_deepcopy_pi_dtd(self):
+        XML = self.etree.XML
+        tostring = self.etree.tostring
+        xml = _bytes('<!-- comment --><!DOCTYPE test [\n<!ENTITY entity "tasty">\n]>\n<test/>')
+        root = XML(xml)
+        tree1 = self.etree.ElementTree(root)
+        self.assertEqual(xml, tostring(tree1))
+
+        tree2 = copy.deepcopy(tree1)
+        self.assertEqual(xml, tostring(tree2))
+
     def test_parse_remove_comments(self):
         fromstring = self.etree.fromstring
         tostring = self.etree.tostring