I'm writing a Django application within Splunk. Part of this is validating xml that my form generates.
When I try to validate the form using lxml, the etree.parse() function raises an exception that is displayed on the browser:
Error: XMLSyntaxError at /en-gb/<app>/<form_url>/ Document is empty, line 1, column 1
My unit tests work fine from outside of Splunk.
My code validator code:
from lxml import etree class XmlValidator(): def __init__(self, xsd_filename): xmlschemadoc = etree.parse(xsd_filename) # this line fails self.xmlschema = etree.XMLSchema(xmlschemadoc)
My form code:
from django import forms from XmlValidator import XmlValidator class MyForm(forms.Form): def __init__(self, schema_filename): self.xsd_filename = schema_filename <snip> def clean(self): validator = XmlValidator(self.xsd_filename) <snip>
I've added code instrumentation to check the file is valid (absolute path, exists, permissions, etc).
I can open the file within the code, read the content and parse the content, just not the file itself. However, I need to pass in the filename so the schemas can include other schemas - this is only supported through the file interface.
Any suggestions appreciated.
The problem is in the parser.
There must be some compatibility issue, because the supplied ETCompatXMLParser works, where the default XMLParser doesn't. (I am also using ElementTree within my application, though I don't know if this is related to my problem).
xmlschemadoc = etree.parse(xsd_filename) # this line fails
parser = etree.ETCompatXMLParser() xmlschemadoc = etree.parse(xsd_filename, parser) # this line now works