From jean at upfrontsystems.co.za  Thu Oct  2 11:33:18 2003
From: jean at upfrontsystems.co.za (Jean Jordaan)
Date: Sun Aug 10 16:54:54 2008
Subject: [Zope-xml] Expat missing the internal subset
Message-ID: <3F7C453E.808@upfrontsystems.co.za>

Hi all

Has anyone seen behaviour like this? We have python2.2.3 on two
different boxes. Both have pyexpat.EXPAT_VERSION: 'expat_1.95.6'
On both I do this:

"""
 >>> from Products.ParsedXML.ParsedXML import createDOMDocument
 >>> from Products.ParsedXML.ExtraDOM import writeStream
 >>> xml = u"""\
... <!DOCTYPE creme SYSTEM 'creme' [
...   <!ELEMENT Customer (properties|references|Address+|License+)>
...   <!ELEMENT Contact  (properties|references|Address+)>
...   <!ELEMENT Address  (properties|references)>
...   <!ATTLIST Customer id ID #IMPLIED>
...   <!ATTLIST Contact  id ID #IMPLIED>
...   <!ATTLIST Address  id ID #IMPLIED>
...   <!ATTLIST License  id ID #IMPLIED>
... ]>
... <creme/>"""
 >>> dom = createDOMDocument(xml)
 >>> dom
<Document at 0x8337ae8>
 >>> writeStream(dom).getvalue()
"""

On the first box, this returns:

"""
u'<?xml version="1.0" ?>\n<!DOCTYPE creme SYSTEM "creme" [\n  <!ELEMENT Customer 
(properties|references|Address+|License+)>\n  <!ELEMENT Contact 
(properties|references|Address+)>\n  <!ELEMENT Address 
(properties|references)>\n  <!ATTLIST Customer id ID #IMPLIED>\n  <!ATTLIST 
Contact  id ID #IMPLIED>\n  <!ATTLIST Address  id ID #IMPLIED>\n  <!ATTLIST 
License  id ID #IMPLIED>\n]>\n<creme/>\n'
"""

On the second box, I see only:

"""
u'<?xml version="1.0" ?>\n<!DOCTYPE creme SYSTEM "creme">\n<creme/>\n'
"""

I managed to trace up to this point of ParsedXML.DOM.ExpatBuilder:

"""
     def _setup_subset(self, buffer):
         """Load the internal subset if there might be one."""
         if self.document.doctype:
             extractor = InternalSubsetExtractor()
             extractor.parseString(buffer)
             subset = extractor.getSubset()
             if subset is not None:
                 d = self.document.doctype.__dict__
                 d['internalSubset'] = subset
"""

On the first box, when I step into `getSubset`, I see this:

"""
(Pdb) self.subset
[u'[', u'\n  ', u'<!ELEMENT', u' ', u'Customer', u' ', u'(', u'properties', 
u'|', u'references', u'|', u'Address+', u'|', u'License+', u')', u'>', u'\n  ', 
u'<!ELEMENT', u' ', u'Contact', u'  ', u'(', u'properties', u'|', u'references', 
u'|', u'Address+', u')', u'>', u'\n  ', u'<!ELEMENT', u' ', u'Address', u'  ', 
u'(', u'properties', u'|', u'references', u')', u'>', u'\n  ', u'<!ATTLIST', u' 
', u'Customer', u' ', u'id', u' ', u'ID', u' ', u'#IMPLIED', u'>', u'\n  ', 
u'<!ATTLIST', u' ', u'Contact', u'  ', u'id', u' ', u'ID', u' ', u'#IMPLIED', 
u'>', u'\n  ', u'<!ATTLIST', u' ', u'Address', u'  ', u'id', u' ', u'ID', u' ', 
u'#IMPLIED', u'>', u'\n  ', u'<!ATTLIST', u' ', u'License', u'  ', u'id', u' ', 
u'ID', u' ', u'#IMPLIED', u'>', u'\n', u']']
"""

On the second box, when I do that, I don't see the enclosing
square brackets:

"""
(Pdb) self.subset
[u'\n  ', u'<!ELEMENT', u' ', u'Customer', u' ', u'(', u'properties', u'|', 
u'references', u'|', u'Address+', u'|', u'License+', u')', u'>', u'\n  ', 
u'<!ELEMENT', u' ', u'Contact', u'  ', u'(', u'properties', u'|', u'references', 
u'|', u'Address+', u')', u'>', u'\n  ', u'<!ELEMENT', u' ', u'Address', u'  ', 
u'(', u'properties', u'|', u'references', u')', u'>', u'\n  ', u'<!ATTLIST', u' 
', u'Customer', u' ', u'id', u' ', u'ID', u' ', u'#IMPLIED', u'>', u'\n  ', 
u'<!ATTLIST', u' ', u'Contact', u'  ', u'id', u' ', u'ID', u' ', u'#IMPLIED', 
u'>', u'\n  ', u'<!ATTLIST', u' ', u'Address', u'  ', u'id', u' ', u'ID', u' ', 
u'#IMPLIED', u'>', u'\n  ', u'<!ATTLIST', u' ', u'License', u'  ', u'id', u' ', 
u'ID', u' ', u'#IMPLIED', u'>', u'\n']
"""

and so the subset extracted is empty. Any ideas would be most
gratefully considered ..

-- 
Jean Jordaan
http://www.upfrontsystems.co.za