[svn r4134] r5093@delle: sbehnel | 2009-03-21 15:54:31 +0100

 support for standalone flag both in tree.docinfo and serialiser

--HG--
branch : trunk

5 files changed, 153 insertions, 18 deletions

diff --git a/CHANGES.txt b/CHANGES.txt
index 349adca4..dc73f851 100644
--- a/CHANGES.txt
+++ b/CHANGES.txt
@@ -5,6 +5,13 @@ lxml changelog
 2.2 (?)
 =======
 
+Features added
+--------------
+
+* Support for ``standalone`` flag in XML declaration through
+  ``tree.docinfo.standalone`` and by passing ``standalone=True/False``
+  on serialisation.
+
 Bugs fixed
 ----------
 
diff --git a/src/lxml/lxml.etree.pyx b/src/lxml/lxml.etree.pyx
index 0891f459..6ac64fec 100644
--- a/src/lxml/lxml.etree.pyx
+++ b/src/lxml/lxml.etree.pyx
@@ -317,6 +317,12 @@ cdef public class _Document [ type LxmlDocumentType, object LxmlDocument ]:
             encoding = funicode(c_doc.encoding)
         return (version, encoding)
 
+    cdef isstandalone(self):
+        if self._c_doc.standalone == -1:
+            return None
+        else:
+            return <bint>(self._c_doc.standalone == 1)
+
     cdef buildNewPrefix(self):
         if self._ns_counter < python.PyTuple_GET_SIZE(_PREFIX_CACHE):
             ns = python.PyTuple_GET_ITEM(_PREFIX_CACHE, self._ns_counter)
@@ -441,6 +447,17 @@ cdef class DocInfo:
             xml_version, encoding = self._doc.getxmlinfo()
             return encoding
 
+    property standalone:
+        u"""Returns the standalone flag as declared by the document.  The possible
+        values are True (``standalone='yes'``), False
+        (``standalone='no'`` or flag not provided in the declaration),
+        and None (unknown or no declaration found).  Note that a
+        normal truth test on this value will always tell if the
+        ``standalone`` flag was set to ``'yes'`` or not.
+        """
+        def __get__(self):
+            return self._doc.isstandalone()
+
     property URL:
         u"The source URL of the document (or None if unknown)."
         def __get__(self):
@@ -1622,10 +1639,10 @@ cdef public class _ElementTree [ type LxmlElementTreeType,
 
     def write(self, file, *, encoding=None, method=u"xml",
               pretty_print=False, xml_declaration=None, with_tail=True,
-              compression=0):
+              standalone=None, compression=0):
         u"""write(self, file, encoding=None, method="xml",
                   pretty_print=False, xml_declaration=None, with_tail=True,
-                  compression=0)
+                  standalone=None, compression=0)
 
         Write the tree to a filename, file or file-like object.
 
@@ -1634,9 +1651,14 @@ cdef public class _ElementTree [ type LxmlElementTreeType,
         The keyword argument 'method' selects the output method: 'xml' or
         'html'.
 
+        Passing a boolean value to the ``standalone`` option will
+        output an XML declaration with the corresponding
+        ``standalone`` flag.
+
         The ``compression`` option enables GZip compression level 1-9.
         """
         cdef bint write_declaration
+        cdef int is_standalone
         self._assertHasRoot()
         # suppress decl. in default case (purely for ElementTree compatibility)
         if xml_declaration is not None:
@@ -1650,10 +1672,19 @@ cdef public class _ElementTree [ type LxmlElementTreeType,
             encoding = encoding.upper()
             write_declaration = encoding not in \
                                   (u'US-ASCII', u'ASCII', u'UTF8', u'UTF-8')
+        if standalone is None:
+            is_standalone = -1
+        elif standalone:
+            write_declaration = 1
+            is_standalone = 1
+        else:
+            write_declaration = 1
+            is_standalone = 0
         if compression is None or compression < 0:
             compression = 0
         _tofilelike(file, self._context_node, encoding, method,
-                    write_declaration, 1, pretty_print, with_tail, compression)
+                    write_declaration, 1, pretty_print, with_tail,
+                    is_standalone, compression)
 
     def getpath(self, _Element element not None):
         u"""getpath(self, element)
@@ -2537,9 +2568,11 @@ def dump(_Element elem not None, *, pretty_print=True, with_tail=True):
     _dumpToFile(sys.stdout, elem._c_node, pretty_print, with_tail)
 
 def tostring(element_or_tree, *, encoding=None, method=u"xml",
-             xml_declaration=None, pretty_print=False, with_tail=True):
+             xml_declaration=None, pretty_print=False, with_tail=True,
+             standalone=None):
     u"""tostring(element_or_tree, encoding=None, method="xml",
-                xml_declaration=None, pretty_print=False, with_tail=True)
+                 xml_declaration=None, pretty_print=False, with_tail=True,
+                 standalone=None)
 
     Serialize an element to an encoded string representation of its XML
     tree.
@@ -2557,11 +2590,15 @@ def tostring(element_or_tree, *, encoding=None, method=u"xml",
     The keyword argument 'method' selects the output method: 'xml',
     'html' or plain 'text'.
 
+    Passing a boolean value to the ``standalone`` option will output
+    an XML declaration with the corresponding ``standalone`` flag.
+
     You can prevent the tail text of the element from being serialised
     by passing the boolean ``with_tail`` option.  This has no impact
     on the tail text of children, which will always be serialised.
     """
     cdef bint write_declaration
+    cdef int is_standalone
     if encoding is _unicode:
         if xml_declaration:
             raise ValueError, \
@@ -2575,14 +2612,23 @@ def tostring(element_or_tree, *, encoding=None, method=u"xml",
         write_declaration = xml_declaration
     if encoding is None:
         encoding = u'ASCII'
+    if standalone is None:
+        is_standalone = -1
+    elif standalone:
+        write_declaration = 1
+        is_standalone = 1
+    else:
+        write_declaration = 1
+        is_standalone = 0
 
     if isinstance(element_or_tree, _Element):
         return _tostring(<_Element>element_or_tree, encoding, method,
-                         write_declaration, 0, pretty_print, with_tail)
+                         write_declaration, 0, pretty_print, with_tail,
+                         is_standalone)
     elif isinstance(element_or_tree, _ElementTree):
         return _tostring((<_ElementTree>element_or_tree)._context_node,
                          encoding, method, write_declaration, 1, pretty_print,
-                         with_tail)
+                         with_tail, is_standalone)
     else:
         raise TypeError, u"Type '%s' cannot be serialized." % \
             python._fqtypename(element_or_tree)
@@ -2601,7 +2647,7 @@ def tostringlist(element_or_tree, *args, **kwargs):
 def tounicode(element_or_tree, *, method=u"xml", pretty_print=False,
               with_tail=True):
     u"""tounicode(element_or_tree, method="xml", pretty_print=False,
-                 with_tail=True)
+                  with_tail=True)
 
     Serialize an element to the Python unicode representation of its XML
     tree.
@@ -2623,10 +2669,11 @@ def tounicode(element_or_tree, *, method=u"xml", pretty_print=False,
     """
     if isinstance(element_or_tree, _Element):
         return _tostring(<_Element>element_or_tree, _unicode, method,
-                          0, 0, pretty_print, with_tail)
+                          0, 0, pretty_print, with_tail, -1)
     elif isinstance(element_or_tree, _ElementTree):
         return _tostring((<_ElementTree>element_or_tree)._context_node,
-                         _unicode, method, 0, 1, pretty_print, with_tail)
+                         _unicode, method, 0, 1, pretty_print, with_tail,
+                         -1)
     else:
         raise TypeError, u"Type '%s' cannot be serialized." % \
             type(element_or_tree)
diff --git a/src/lxml/serializer.pxi b/src/lxml/serializer.pxi
index 576a4ce5..84a189d6 100644
--- a/src/lxml/serializer.pxi
+++ b/src/lxml/serializer.pxi
@@ -75,7 +75,7 @@ cdef _textToString(xmlNode* c_node, encoding, bint with_tail):
 
 cdef _tostring(_Element element, encoding, method,
                bint write_xml_declaration, bint write_complete_document,
-               bint pretty_print, bint with_tail):
+               bint pretty_print, bint with_tail, int standalone):
     u"""Serialize an element to an encoded string representation of its XML
     tree.
     """
@@ -111,7 +111,7 @@ cdef _tostring(_Element element, encoding, method,
     with nogil:
         _writeNodeToBuffer(c_buffer, element._c_node, c_enc, c_method,
                            write_xml_declaration, write_complete_document,
-                           pretty_print, with_tail)
+                           pretty_print, with_tail, standalone)
         tree.xmlOutputBufferFlush(c_buffer)
         if c_buffer.conv is not NULL:
             c_result_buffer = c_buffer.conv
@@ -155,12 +155,13 @@ cdef void _writeNodeToBuffer(tree.xmlOutputBuffer* c_buffer,
                              xmlNode* c_node, char* encoding, int c_method,
                              bint write_xml_declaration,
                              bint write_complete_document,
-                             bint pretty_print, bint with_tail) nogil:
+                             bint pretty_print, bint with_tail,
+                             int standalone) nogil:
     cdef xmlDoc* c_doc
     cdef xmlNode* c_nsdecl_node
     c_doc = c_node.doc
     if write_xml_declaration and c_method == OUTPUT_METHOD_XML:
-        _writeDeclarationToBuffer(c_buffer, c_doc.version, encoding)
+        _writeDeclarationToBuffer(c_buffer, c_doc.version, encoding, standalone)
 
     # write internal DTD subset, preceding PIs/comments, etc.
     if write_complete_document:
@@ -203,14 +204,20 @@ cdef void _writeNodeToBuffer(tree.xmlOutputBuffer* c_buffer,
         tree.xmlOutputBufferWrite(c_buffer, 1, "\n")
 
 cdef void _writeDeclarationToBuffer(tree.xmlOutputBuffer* c_buffer,
-                                    char* version, char* encoding) nogil:
+                                    char* version, char* encoding,
+                                    int standalone) nogil:
     if version is NULL:
         version = "1.0"
     tree.xmlOutputBufferWrite(c_buffer, 15, "<?xml version='")
     tree.xmlOutputBufferWriteString(c_buffer, version)
     tree.xmlOutputBufferWrite(c_buffer, 12, "' encoding='")
     tree.xmlOutputBufferWriteString(c_buffer, encoding)
-    tree.xmlOutputBufferWrite(c_buffer, 4, "'?>\n")
+    if standalone == 0:
+        tree.xmlOutputBufferWrite(c_buffer, 20, "' standalone='no'?>\n")
+    elif standalone == 1:
+        tree.xmlOutputBufferWrite(c_buffer, 21, "' standalone='yes'?>\n")
+    else:
+        tree.xmlOutputBufferWrite(c_buffer, 4, "'?>\n")
 
 cdef void _writeDtdToBuffer(tree.xmlOutputBuffer* c_buffer,
                             xmlDoc* c_doc, char* c_root_name,
@@ -351,7 +358,8 @@ cdef int _closeFilelikeWriter(void* ctxt):
 
 cdef _tofilelike(f, _Element element, encoding, method,
                  bint write_xml_declaration, bint write_doctype,
-                 bint pretty_print, bint with_tail, int compression):
+                 bint pretty_print, bint with_tail, int standalone,
+                 int compression):
     cdef python.PyThreadState* state = NULL
     cdef _FilelikeWriter writer
     cdef tree.xmlOutputBuffer* c_buffer
@@ -408,7 +416,7 @@ cdef _tofilelike(f, _Element element, encoding, method,
 
     _writeNodeToBuffer(c_buffer, element._c_node, c_enc, c_method,
                        write_xml_declaration, write_doctype,
-                       pretty_print, with_tail)
+                       pretty_print, with_tail, standalone)
     error_result = c_buffer.error
     if error_result == xmlerror.XML_ERR_OK:
         error_result = tree.xmlOutputBufferClose(c_buffer)
diff --git a/src/lxml/tests/test_etree.py b/src/lxml/tests/test_etree.py
index b4e0a811..d35be101 100644
--- a/src/lxml/tests/test_etree.py
+++ b/src/lxml/tests/test_etree.py
@@ -2313,6 +2313,78 @@ class ETreeOnlyTestCase(HelperTestCase):
         result = tostring(a, with_tail=True)
         self.assertEquals(result, _bytes("<a><b/>bTAIL<c/></a>aTAIL"))
 
+    def test_standalone(self):
+        tostring = self.etree.tostring
+        XML = self.etree.XML
+        ElementTree = self.etree.ElementTree
+        Element = self.etree.Element
+
+        tree = Element("root").getroottree()
+        self.assertEquals(None, tree.docinfo.standalone)
+
+        tree = XML(_bytes("<root/>")).getroottree()
+        self.assertEquals(None, tree.docinfo.standalone)
+
+        tree = XML(_bytes(
+            "<?xml version='1.0' encoding='ASCII' standalone='yes'?>\n<root/>"
+            )).getroottree()
+        self.assertEquals(True, tree.docinfo.standalone)
+
+        tree = XML(_bytes(
+            "<?xml version='1.0' encoding='ASCII' standalone='no'?>\n<root/>"
+            )).getroottree()
+        self.assertEquals(False, tree.docinfo.standalone)
+
+    def test_tostring_standalone(self):
+        tostring = self.etree.tostring
+        XML = self.etree.XML
+        ElementTree = self.etree.ElementTree
+
+        root = XML(_bytes("<root/>"))
+
+        tree = ElementTree(root)
+        self.assertEquals(None, tree.docinfo.standalone)
+
+        result = tostring(root, xml_declaration=True, encoding="ASCII")
+        self.assertEquals(result, _bytes(
+            "<?xml version='1.0' encoding='ASCII'?>\n<root/>"))
+
+        result = tostring(root, xml_declaration=True, encoding="ASCII",
+                          standalone=True)
+        self.assertEquals(result, _bytes(
+            "<?xml version='1.0' encoding='ASCII' standalone='yes'?>\n<root/>"))
+
+        tree = ElementTree(XML(result))
+        self.assertEquals(True, tree.docinfo.standalone)
+
+        result = tostring(root, xml_declaration=True, encoding="ASCII",
+                          standalone=False)
+        self.assertEquals(result, _bytes(
+            "<?xml version='1.0' encoding='ASCII' standalone='no'?>\n<root/>"))
+
+        tree = ElementTree(XML(result))
+        self.assertEquals(False, tree.docinfo.standalone)
+
+    def test_tostring_standalone_in_out(self):
+        tostring = self.etree.tostring
+        XML = self.etree.XML
+        ElementTree = self.etree.ElementTree
+
+        root = XML(_bytes(
+            "<?xml version='1.0' encoding='UTF-8' standalone='yes'?>\n<root/>"))
+
+        tree = ElementTree(root)
+        self.assertEquals(True, tree.docinfo.standalone)
+
+        result = tostring(root, xml_declaration=True, encoding="ASCII")
+        self.assertEquals(result, _bytes(
+            "<?xml version='1.0' encoding='ASCII'?>\n<root/>"))
+
+        result = tostring(root, xml_declaration=True, encoding="ASCII",
+                          standalone=True)
+        self.assertEquals(result, _bytes(
+            "<?xml version='1.0' encoding='ASCII' standalone='yes'?>\n<root/>"))
+
     def test_tostring_method_text_encoding(self):
         tostring = self.etree.tostring
         Element = self.etree.Element
diff --git a/src/lxml/tree.pxd b/src/lxml/tree.pxd
index 78ca566f..521fbc5f 100644
--- a/src/lxml/tree.pxd
+++ b/src/lxml/tree.pxd
@@ -131,6 +131,7 @@ cdef extern from "libxml/tree.h":
         xmlDoc* doc
         xmlDict* dict
         xmlHashTable* ids
+        int standalone
         char* version
         char* encoding
         char* URL