cmlenz@1: # -*- coding: utf-8 -*-
cmlenz@1: #
cmlenz@66: # Copyright (C) 2006 Edgewall Software
cmlenz@1: # All rights reserved.
cmlenz@1: #
cmlenz@1: # This software is licensed as described in the file COPYING, which
cmlenz@1: # you should have received as part of this distribution. The terms
cmlenz@230: # are also available at http://genshi.edgewall.org/wiki/License.
cmlenz@1: #
cmlenz@1: # This software consists of voluntary contributions made by many
cmlenz@1: # individuals. For the exact contribution history, see the revision
cmlenz@230: # history and logs, available at http://genshi.edgewall.org/log/.
cmlenz@1: 
cmlenz@1: """This module provides different kinds of serialization methods for XML event
cmlenz@1: streams.
cmlenz@1: """
cmlenz@1: 
cmlenz@123: from itertools import chain
cmlenz@1: try:
cmlenz@1:     frozenset
cmlenz@1: except NameError:
cmlenz@1:     from sets import ImmutableSet as frozenset
cmlenz@123: import re
cmlenz@1: 
cmlenz@230: from genshi.core import escape, Markup, Namespace, QName, StreamEventKind
cmlenz@230: from genshi.core import DOCTYPE, START, END, START_NS, TEXT, START_CDATA, \
cmlenz@145:                         END_CDATA, PI, COMMENT, XML_NAMESPACE
cmlenz@1: 
cmlenz@200: __all__ = ['DocType', 'XMLSerializer', 'XHTMLSerializer', 'HTMLSerializer',
cmlenz@200:            'TextSerializer']
cmlenz@1: 
cmlenz@1: 
cmlenz@85: class DocType(object):
cmlenz@85:     """Defines a number of commonly used DOCTYPE declarations as constants."""
cmlenz@85: 
cmlenz@85:     HTML_STRICT = ('html', '-//W3C//DTD HTML 4.01//EN',
cmlenz@85:                    'http://www.w3.org/TR/html4/strict.dtd')
cmlenz@85:     HTML_TRANSITIONAL = ('html', '-//W3C//DTD HTML 4.01 Transitional//EN',
cmlenz@85:                          'http://www.w3.org/TR/html4/loose.dtd')
cmlenz@85:     HTML = HTML_STRICT
cmlenz@85: 
cmlenz@85:     XHTML_STRICT = ('html', '-//W3C//DTD XHTML 1.0 Strict//EN',
cmlenz@85:                     'http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd')
cmlenz@85:     XHTML_TRANSITIONAL = ('html', '-//W3C//DTD XHTML 1.0 Transitional//EN',
cmlenz@85:                           'http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd')
cmlenz@85:     XHTML = XHTML_STRICT
cmlenz@85: 
cmlenz@85: 
cmlenz@123: class XMLSerializer(object):
cmlenz@1:     """Produces XML text from an event stream.
cmlenz@1:     
cmlenz@230:     >>> from genshi.builder import tag
cmlenz@20:     >>> elem = tag.div(tag.a(href='foo'), tag.br, tag.hr(noshade=True))
cmlenz@123:     >>> print ''.join(XMLSerializer()(elem.generate()))
cmlenz@1:     <div><a href="foo"/><br/><hr noshade="True"/></div>
cmlenz@1:     """
cmlenz@123: 
cmlenz@123:     _PRESERVE_SPACE = frozenset()
cmlenz@123: 
cmlenz@123:     def __init__(self, doctype=None, strip_whitespace=True):
cmlenz@85:         """Initialize the XML serializer.
cmlenz@85:         
cmlenz@85:         @param doctype: a `(name, pubid, sysid)` tuple that represents the
cmlenz@85:             DOCTYPE declaration that should be included at the top of the
cmlenz@85:             generated output
cmlenz@123:         @param strip_whitespace: whether extraneous whitespace should be
cmlenz@123:             stripped from the output
cmlenz@85:         """
cmlenz@85:         self.preamble = []
cmlenz@85:         if doctype:
cmlenz@85:             self.preamble.append((DOCTYPE, doctype, (None, -1, -1)))
cmlenz@212:         self.filters = [EmptyTagFilter()]
cmlenz@123:         if strip_whitespace:
cmlenz@123:             self.filters.append(WhitespaceFilter(self._PRESERVE_SPACE))
cmlenz@1: 
cmlenz@123:     def __call__(self, stream):
cmlenz@1:         ns_attrib = []
cmlenz@141:         ns_mapping = {XML_NAMESPACE.uri: 'xml'}
cmlenz@143:         have_doctype = False
cmlenz@143:         in_cdata = False
cmlenz@1: 
cmlenz@123:         stream = chain(self.preamble, stream)
cmlenz@123:         for filter_ in self.filters:
cmlenz@123:             stream = filter_(stream)
cmlenz@1:         for kind, data, pos in stream:
cmlenz@1: 
cmlenz@212:             if kind is START or kind is EMPTY:
cmlenz@1:                 tag, attrib = data
cmlenz@1: 
cmlenz@1:                 tagname = tag.localname
cmlenz@123:                 namespace = tag.namespace
cmlenz@123:                 if namespace:
cmlenz@123:                     if namespace in ns_mapping:
cmlenz@123:                         prefix = ns_mapping[namespace]
cmlenz@1:                         if prefix:
cmlenz@123:                             tagname = '%s:%s' % (prefix, tagname)
cmlenz@123:                     else:
cmlenz@123:                         ns_attrib.append((QName('xmlns'), namespace))
cmlenz@136:                 buf = ['<', tagname]
cmlenz@1: 
cmlenz@397:                 if ns_attrib:
cmlenz@397:                     attrib += tuple(ns_attrib)
cmlenz@397:                 for attr, value in attrib:
cmlenz@1:                     attrname = attr.localname
cmlenz@397:                     attrns = attr.namespace
cmlenz@397:                     if attrns:
cmlenz@397:                         prefix = ns_mapping.get(attrns)
cmlenz@1:                         if prefix:
cmlenz@69:                             attrname = '%s:%s' % (prefix, attrname)
cmlenz@136:                     buf += [' ', attrname, '="', escape(value), '"']
cmlenz@123:                 ns_attrib = []
cmlenz@1: 
cmlenz@397:                 buf.append(kind is EMPTY and '/>' or '>')
cmlenz@1: 
cmlenz@397:                 yield Markup(u''.join(buf))
cmlenz@1: 
cmlenz@69:             elif kind is END:
cmlenz@1:                 tag = data
cmlenz@1:                 tagname = tag.localname
cmlenz@1:                 if tag.namespace:
cmlenz@26:                     prefix = ns_mapping.get(tag.namespace)
cmlenz@26:                     if prefix:
cmlenz@69:                         tagname = '%s:%s' % (prefix, tag.localname)
cmlenz@1:                 yield Markup('</%s>' % tagname)
cmlenz@1: 
cmlenz@69:             elif kind is TEXT:
cmlenz@143:                 if in_cdata:
cmlenz@143:                     yield data
cmlenz@143:                 else:
cmlenz@143:                     yield escape(data, quotes=False)
cmlenz@1: 
cmlenz@89:             elif kind is COMMENT:
cmlenz@89:                 yield Markup('<!--%s-->' % data)
cmlenz@89: 
cmlenz@136:             elif kind is DOCTYPE and not have_doctype:
cmlenz@136:                 name, pubid, sysid = data
cmlenz@136:                 buf = ['<!DOCTYPE %s']
cmlenz@136:                 if pubid:
cmlenz@397:                     buf.append(' PUBLIC "%s"')
cmlenz@136:                 elif sysid:
cmlenz@397:                     buf.append(' SYSTEM')
cmlenz@136:                 if sysid:
cmlenz@397:                     buf.append(' "%s"')
cmlenz@397:                 buf.append('>\n')
cmlenz@397:                 yield Markup(u''.join(buf), *filter(None, data))
cmlenz@136:                 have_doctype = True
cmlenz@109: 
cmlenz@109:             elif kind is START_NS:
cmlenz@109:                 prefix, uri = data
cmlenz@109:                 if uri not in ns_mapping:
cmlenz@109:                     ns_mapping[uri] = prefix
cmlenz@109:                     if not prefix:
cmlenz@109:                         ns_attrib.append((QName('xmlns'), uri))
cmlenz@109:                     else:
cmlenz@109:                         ns_attrib.append((QName('xmlns:%s' % prefix), uri))
cmlenz@109: 
cmlenz@143:             elif kind is START_CDATA:
cmlenz@143:                 yield Markup('<![CDATA[')
cmlenz@143:                 in_cdata = True
cmlenz@143: 
cmlenz@143:             elif kind is END_CDATA:
cmlenz@143:                 yield Markup(']]>')
cmlenz@143:                 in_cdata = False
cmlenz@143: 
cmlenz@105:             elif kind is PI:
cmlenz@105:                 yield Markup('<?%s %s?>' % data)
cmlenz@105: 
cmlenz@1: 
cmlenz@96: class XHTMLSerializer(XMLSerializer):
cmlenz@96:     """Produces XHTML text from an event stream.
cmlenz@1:     
cmlenz@230:     >>> from genshi.builder import tag
cmlenz@20:     >>> elem = tag.div(tag.a(href='foo'), tag.br, tag.hr(noshade=True))
cmlenz@123:     >>> print ''.join(XHTMLSerializer()(elem.generate()))
cmlenz@96:     <div><a href="foo"></a><br /><hr noshade="noshade" /></div>
cmlenz@1:     """
cmlenz@1: 
cmlenz@18:     NAMESPACE = Namespace('http://www.w3.org/1999/xhtml')
cmlenz@1: 
cmlenz@1:     _EMPTY_ELEMS = frozenset(['area', 'base', 'basefont', 'br', 'col', 'frame',
cmlenz@1:                               'hr', 'img', 'input', 'isindex', 'link', 'meta',
cmlenz@1:                               'param'])
cmlenz@1:     _BOOLEAN_ATTRS = frozenset(['selected', 'checked', 'compact', 'declare',
cmlenz@1:                                 'defer', 'disabled', 'ismap', 'multiple',
cmlenz@1:                                 'nohref', 'noresize', 'noshade', 'nowrap'])
cmlenz@346:     _PRESERVE_SPACE = frozenset([
cmlenz@346:         QName('pre'), QName('http://www.w3.org/1999/xhtml}pre'),
cmlenz@346:         QName('textarea'), QName('http://www.w3.org/1999/xhtml}textarea')
cmlenz@346:     ])
cmlenz@1: 
cmlenz@123:     def __call__(self, stream):
cmlenz@136:         namespace = self.NAMESPACE
cmlenz@141:         ns_attrib = []
cmlenz@141:         ns_mapping = {XML_NAMESPACE.uri: 'xml'}
cmlenz@136:         boolean_attrs = self._BOOLEAN_ATTRS
cmlenz@136:         empty_elems = self._EMPTY_ELEMS
cmlenz@85:         have_doctype = False
cmlenz@143:         in_cdata = False
cmlenz@1: 
cmlenz@123:         stream = chain(self.preamble, stream)
cmlenz@123:         for filter_ in self.filters:
cmlenz@123:             stream = filter_(stream)
cmlenz@1:         for kind, data, pos in stream:
cmlenz@1: 
cmlenz@212:             if kind is START or kind is EMPTY:
cmlenz@1:                 tag, attrib = data
cmlenz@96: 
cmlenz@141:                 tagname = tag.localname
cmlenz@177:                 tagns = tag.namespace
cmlenz@177:                 if tagns:
cmlenz@177:                     if tagns in ns_mapping:
cmlenz@177:                         prefix = ns_mapping[tagns]
cmlenz@141:                         if prefix:
cmlenz@141:                             tagname = '%s:%s' % (prefix, tagname)
cmlenz@141:                     else:
cmlenz@177:                         ns_attrib.append((QName('xmlns'), tagns))
cmlenz@141:                 buf = ['<', tagname]
cmlenz@136: 
cmlenz@397:                 if ns_attrib:
cmlenz@397:                     attrib += tuple(ns_attrib)
cmlenz@397:                 for attr, value in attrib:
cmlenz@141:                     attrname = attr.localname
cmlenz@397:                     attrns = attr.namespace
cmlenz@397:                     if attrns:
cmlenz@397:                         prefix = ns_mapping.get(attrns)
cmlenz@141:                         if prefix:
cmlenz@141:                             attrname = '%s:%s' % (prefix, attrname)
cmlenz@141:                     if attrname in boolean_attrs:
cmlenz@141:                         if value:
cmlenz@141:                             buf += [' ', attrname, '="', attrname, '"']
cmlenz@141:                     else:
cmlenz@141:                         buf += [' ', attrname, '="', escape(value), '"']
cmlenz@141:                 ns_attrib = []
cmlenz@141: 
cmlenz@212:                 if kind is EMPTY:
cmlenz@212:                     if (tagns and tagns != namespace.uri) \
cmlenz@397:                             or tagname in empty_elems:
cmlenz@397:                         buf.append(' />')
cmlenz@96:                     else:
cmlenz@397:                         buf.append('></%s>' % tagname)
cmlenz@141:                 else:
cmlenz@397:                     buf.append('>')
cmlenz@96: 
cmlenz@397:                 yield Markup(u''.join(buf))
cmlenz@96: 
cmlenz@96:             elif kind is END:
cmlenz@96:                 tag = data
cmlenz@141:                 tagname = tag.localname
cmlenz@141:                 if tag.namespace:
cmlenz@141:                     prefix = ns_mapping.get(tag.namespace)
cmlenz@141:                     if prefix:
cmlenz@177:                         tagname = '%s:%s' % (prefix, tagname)
cmlenz@141:                 yield Markup('</%s>' % tagname)
cmlenz@96: 
cmlenz@96:             elif kind is TEXT:
cmlenz@143:                 if in_cdata:
cmlenz@143:                     yield data
cmlenz@143:                 else:
cmlenz@143:                     yield escape(data, quotes=False)
cmlenz@96: 
cmlenz@96:             elif kind is COMMENT:
cmlenz@96:                 yield Markup('<!--%s-->' % data)
cmlenz@96: 
cmlenz@136:             elif kind is DOCTYPE and not have_doctype:
cmlenz@136:                 name, pubid, sysid = data
cmlenz@136:                 buf = ['<!DOCTYPE %s']
cmlenz@136:                 if pubid:
cmlenz@397:                     buf.append(' PUBLIC "%s"')
cmlenz@136:                 elif sysid:
cmlenz@397:                     buf.append(' SYSTEM')
cmlenz@136:                 if sysid:
cmlenz@397:                     buf.append(' "%s"')
cmlenz@397:                 buf.append('>\n')
cmlenz@397:                 yield Markup(u''.join(buf), *filter(None, data))
cmlenz@136:                 have_doctype = True
cmlenz@109: 
cmlenz@141:             elif kind is START_NS:
cmlenz@141:                 prefix, uri = data
cmlenz@141:                 if uri not in ns_mapping:
cmlenz@141:                     ns_mapping[uri] = prefix
cmlenz@141:                     if not prefix:
cmlenz@141:                         ns_attrib.append((QName('xmlns'), uri))
cmlenz@141:                     else:
cmlenz@141:                         ns_attrib.append((QName('xmlns:%s' % prefix), uri))
cmlenz@109: 
cmlenz@143:             elif kind is START_CDATA:
cmlenz@143:                 yield Markup('<![CDATA[')
cmlenz@143:                 in_cdata = True
cmlenz@143: 
cmlenz@143:             elif kind is END_CDATA:
cmlenz@143:                 yield Markup(']]>')
cmlenz@143:                 in_cdata = False
cmlenz@143: 
cmlenz@105:             elif kind is PI:
cmlenz@105:                 yield Markup('<?%s %s?>' % data)
cmlenz@105: 
cmlenz@96: 
cmlenz@96: class HTMLSerializer(XHTMLSerializer):
cmlenz@96:     """Produces HTML text from an event stream.
cmlenz@96:     
cmlenz@230:     >>> from genshi.builder import tag
cmlenz@96:     >>> elem = tag.div(tag.a(href='foo'), tag.br, tag.hr(noshade=True))
cmlenz@123:     >>> print ''.join(HTMLSerializer()(elem.generate()))
cmlenz@96:     <div><a href="foo"></a><br><hr noshade></div>
cmlenz@96:     """
cmlenz@96: 
cmlenz@280:     _NOESCAPE_ELEMS = frozenset([QName('script'),
cmlenz@280:                                  QName('http://www.w3.org/1999/xhtml}script'),
cmlenz@280:                                  QName('style'),
cmlenz@280:                                  QName('http://www.w3.org/1999/xhtml}style')])
cmlenz@141: 
cmlenz@141:     def __init__(self, doctype=None, strip_whitespace=True):
cmlenz@141:         """Initialize the HTML serializer.
cmlenz@141:         
cmlenz@141:         @param doctype: a `(name, pubid, sysid)` tuple that represents the
cmlenz@141:             DOCTYPE declaration that should be included at the top of the
cmlenz@141:             generated output
cmlenz@141:         @param strip_whitespace: whether extraneous whitespace should be
cmlenz@141:             stripped from the output
cmlenz@141:         """
cmlenz@141:         super(HTMLSerializer, self).__init__(doctype, False)
cmlenz@141:         if strip_whitespace:
cmlenz@141:             self.filters.append(WhitespaceFilter(self._PRESERVE_SPACE,
cmlenz@305:                                                  self._NOESCAPE_ELEMS))
cmlenz@141: 
cmlenz@123:     def __call__(self, stream):
cmlenz@136:         namespace = self.NAMESPACE
cmlenz@136:         ns_mapping = {}
cmlenz@136:         boolean_attrs = self._BOOLEAN_ATTRS
cmlenz@136:         empty_elems = self._EMPTY_ELEMS
cmlenz@141:         noescape_elems = self._NOESCAPE_ELEMS
cmlenz@96:         have_doctype = False
cmlenz@141:         noescape = False
cmlenz@96: 
cmlenz@123:         stream = chain(self.preamble, stream)
cmlenz@123:         for filter_ in self.filters:
cmlenz@123:             stream = filter_(stream)
cmlenz@96:         for kind, data, pos in stream:
cmlenz@96: 
cmlenz@212:             if kind is START or kind is EMPTY:
cmlenz@96:                 tag, attrib = data
cmlenz@136:                 if not tag.namespace or tag in namespace:
cmlenz@136:                     tagname = tag.localname
cmlenz@136:                     buf = ['<', tagname]
cmlenz@96: 
cmlenz@136:                     for attr, value in attrib:
cmlenz@136:                         attrname = attr.localname
cmlenz@141:                         if not attr.namespace or attr in namespace:
cmlenz@136:                             if attrname in boolean_attrs:
cmlenz@136:                                 if value:
cmlenz@136:                                     buf += [' ', attrname]
cmlenz@136:                             else:
cmlenz@136:                                 buf += [' ', attrname, '="', escape(value), '"']
cmlenz@1: 
cmlenz@397:                     buf.append('>')
cmlenz@1: 
cmlenz@212:                     if kind is EMPTY:
cmlenz@212:                         if tagname not in empty_elems:
cmlenz@397:                             buf.append('</%s>' % tagname)
cmlenz@212: 
cmlenz@397:                     yield Markup(u''.join(buf))
cmlenz@1: 
cmlenz@141:                     if tagname in noescape_elems:
cmlenz@141:                         noescape = True
cmlenz@141: 
cmlenz@69:             elif kind is END:
cmlenz@1:                 tag = data
cmlenz@136:                 if not tag.namespace or tag in namespace:
cmlenz@136:                     yield Markup('</%s>' % tag.localname)
cmlenz@1: 
cmlenz@141:                 noescape = False
cmlenz@141: 
cmlenz@69:             elif kind is TEXT:
cmlenz@141:                 if noescape:
cmlenz@141:                     yield data
cmlenz@141:                 else:
cmlenz@141:                     yield escape(data, quotes=False)
cmlenz@1: 
cmlenz@89:             elif kind is COMMENT:
cmlenz@89:                 yield Markup('<!--%s-->' % data)
cmlenz@89: 
cmlenz@136:             elif kind is DOCTYPE and not have_doctype:
cmlenz@136:                 name, pubid, sysid = data
cmlenz@136:                 buf = ['<!DOCTYPE %s']
cmlenz@136:                 if pubid:
cmlenz@397:                     buf.append(' PUBLIC "%s"')
cmlenz@136:                 elif sysid:
cmlenz@397:                     buf.append(' SYSTEM')
cmlenz@136:                 if sysid:
cmlenz@397:                     buf.append(' "%s"')
cmlenz@397:                 buf.append('>\n')
cmlenz@397:                 yield Markup(u''.join(buf), *filter(None, data))
cmlenz@136:                 have_doctype = True
cmlenz@109: 
cmlenz@136:             elif kind is START_NS and data[1] not in ns_mapping:
cmlenz@136:                 ns_mapping[data[1]] = data[0]
cmlenz@109: 
cmlenz@105:             elif kind is PI:
cmlenz@105:                 yield Markup('<?%s %s?>' % data)
cmlenz@105: 
cmlenz@1: 
cmlenz@200: class TextSerializer(object):
cmlenz@200:     """Produces plain text from an event stream.
cmlenz@200:     
cmlenz@200:     Only text events are included in the output. Unlike the other serializer,
cmlenz@200:     special XML characters are not escaped:
cmlenz@200:     
cmlenz@230:     >>> from genshi.builder import tag
cmlenz@200:     >>> elem = tag.div(tag.a('<Hello!>', href='foo'), tag.br)
cmlenz@200:     >>> print elem
cmlenz@200:     <div><a href="foo">&lt;Hello!&gt;</a><br/></div>
cmlenz@200:     >>> print ''.join(TextSerializer()(elem.generate()))
cmlenz@200:     <Hello!>
cmlenz@200: 
cmlenz@200:     If text events contain literal markup (instances of the `Markup` class),
cmlenz@200:     tags or entities are stripped from the output:
cmlenz@200:     
cmlenz@200:     >>> elem = tag.div(Markup('<a href="foo">Hello!</a><br/>'))
cmlenz@200:     >>> print elem
cmlenz@200:     <div><a href="foo">Hello!</a><br/></div>
cmlenz@200:     >>> print ''.join(TextSerializer()(elem.generate()))
cmlenz@200:     Hello!
cmlenz@200:     """
cmlenz@200: 
cmlenz@200:     def __call__(self, stream):
cmlenz@200:         for kind, data, pos in stream:
cmlenz@200:             if kind is TEXT:
cmlenz@200:                 if type(data) is Markup:
cmlenz@200:                     data = data.striptags().stripentities()
cmlenz@201:                 yield unicode(data)
cmlenz@200: 
cmlenz@200: 
cmlenz@212: class EmptyTagFilter(object):
cmlenz@212:     """Combines `START` and `STOP` events into `EMPTY` events for elements that
cmlenz@212:     have no contents.
cmlenz@212:     """
cmlenz@212: 
cmlenz@212:     EMPTY = StreamEventKind('EMPTY')
cmlenz@212: 
cmlenz@212:     def __call__(self, stream):
cmlenz@212:         prev = (None, None, None)
cmlenz@212:         for kind, data, pos in stream:
cmlenz@212:             if prev[0] is START:
cmlenz@212:                 if kind is END:
cmlenz@212:                     prev = EMPTY, prev[1], prev[2]
cmlenz@212:                     yield prev
cmlenz@212:                     continue
cmlenz@212:                 else:
cmlenz@212:                     yield prev
cmlenz@212:             if kind is not START:
cmlenz@212:                 yield kind, data, pos
cmlenz@212:             prev = kind, data, pos
cmlenz@212: 
cmlenz@212: 
cmlenz@212: EMPTY = EmptyTagFilter.EMPTY
cmlenz@212: 
cmlenz@212: 
cmlenz@123: class WhitespaceFilter(object):
cmlenz@123:     """A filter that removes extraneous ignorable white space from the
cmlenz@123:     stream."""
cmlenz@123: 
cmlenz@305:     def __init__(self, preserve=None, noescape=None):
cmlenz@123:         """Initialize the filter.
cmlenz@123:         
cmlenz@141:         @param preserve: a set or sequence of tag names for which white-space
cmlenz@397:             should be preserved
cmlenz@141:         @param noescape: a set or sequence of tag names for which text content
cmlenz@141:             should not be escaped
cmlenz@141:         
cmlenz@346:         The `noescape` set is expected to refer to elements that cannot contain
cmlenz@346:         further child elements (such as <style> or <script> in HTML documents).
cmlenz@123:         """
cmlenz@123:         if preserve is None:
cmlenz@123:             preserve = []
cmlenz@123:         self.preserve = frozenset(preserve)
cmlenz@141:         if noescape is None:
cmlenz@141:             noescape = []
cmlenz@141:         self.noescape = frozenset(noescape)
cmlenz@123: 
cmlenz@219:     def __call__(self, stream, ctxt=None, space=XML_NAMESPACE['space'],
cmlenz@219:                  trim_trailing_space=re.compile('[ \t]+(?=\n)').sub,
cmlenz@219:                  collapse_lines=re.compile('\n{2,}').sub):
cmlenz@123:         mjoin = Markup('').join
cmlenz@141:         preserve_elems = self.preserve
cmlenz@346:         preserve = 0
cmlenz@141:         noescape_elems = self.noescape
cmlenz@141:         noescape = False
cmlenz@123: 
cmlenz@123:         textbuf = []
cmlenz@141:         push_text = textbuf.append
cmlenz@136:         pop_text = textbuf.pop
cmlenz@123:         for kind, data, pos in chain(stream, [(None, None, None)]):
cmlenz@123:             if kind is TEXT:
cmlenz@141:                 if noescape:
cmlenz@141:                     data = Markup(data)
cmlenz@141:                 push_text(data)
cmlenz@123:             else:
cmlenz@123:                 if textbuf:
cmlenz@123:                     if len(textbuf) > 1:
cmlenz@123:                         text = mjoin(textbuf, escape_quotes=False)
cmlenz@123:                         del textbuf[:]
cmlenz@123:                     else:
cmlenz@136:                         text = escape(pop_text(), quotes=False)
cmlenz@141:                     if not preserve:
cmlenz@123:                         text = collapse_lines('\n', trim_trailing_space('', text))
cmlenz@123:                     yield TEXT, Markup(text), pos
cmlenz@141: 
cmlenz@141:                 if kind is START:
cmlenz@346:                     tag, attrs = data
cmlenz@346:                     if preserve or (tag in preserve_elems or
cmlenz@346:                                     attrs.get(space) == 'preserve'):
cmlenz@346:                         preserve += 1
cmlenz@219:                     if not noescape and tag in noescape_elems:
cmlenz@141:                         noescape = True
cmlenz@141: 
cmlenz@141:                 elif kind is END:
cmlenz@346:                     noescape = False
cmlenz@346:                     if preserve:
cmlenz@346:                         preserve -= 1
cmlenz@141: 
cmlenz@305:                 elif kind is START_CDATA:
cmlenz@143:                     noescape = True
cmlenz@143: 
cmlenz@305:                 elif kind is END_CDATA:
cmlenz@143:                     noescape = False
cmlenz@143: 
cmlenz@136:                 if kind:
cmlenz@123:                     yield kind, data, pos