"Juergen Hermann" <[EMAIL PROTECTED]> writes:

> On Sat, 01 Sep 2001 11:18:02 -0700, Dean Roddey wrote:
> 
> >If you are going to do the James Clark type tests, you must use a parser
> >that does canonical output, are you doing that? 
> 
> No, I do not yet do output conformance testing at all. Those test results are purely 
> based on the parse result (MUST be ok for valid files, MUST fail for non-wf tests, 
> etc.). With output conformance testing, only MORE errors could be generated, I 
> think.
> 
> This is the current checks, I think they correctly do what http://www.oasis-
> open.org/committees/xml-conformance/suite-v1se/xmlconf-20010315.htm outlines as test 
> rules in sections 3.1 to 3.4:
> 
>         msg = ''
>         severity = 'HI'
>         if type == "valid":
>             if result and result.type != "W":
>                 msg = "Error for valid file"
>         elif type == "invalid":
>             if validate:
>                 if not result or result.type == "W":
>                     msg = "No error for invalid document reported"
>                 elif result.type != "E" and self.strict:
>                     severity = 'LOW'
>                     msg = "Error was not reported as recoverable"
>             elif result and result != "W":
>                 severity = 'MED'
>                 msg = "Non-validating parser reported error for invalid document"
>         elif type == "not-wf":
>             if not result or result.type == "W":
>                 msg = "No error for non-wellformed document reported"
>             elif result.type != "F" and self.strict:
>                 severity = 'LOW'
>                 msg = "Error was not reported as fatal"
>         elif type == "error" and strict and not result:
>             severity = 'LOW'
>             msg = 'Optional error was not reported'

Hey Juergen,

This looks correct by my understanding. I looked at the pirxx page and
noticed that you're often recording two errors for the same test:

  --- valid-sa-094 James Clark XMLTEST cases, 18-Nov-1998 (cf. 2.8) ------------
      This refers to an undefined parameter entity reference within
      a markup declaration in the internal DTD subset, violating
      the PEs in Internal Subset WFC.
  
   HI No error for non-wellformed document reported
      xmltest\valid/sa/094.xml
      None
  
  --- valid-sa-094 James Clark XMLTEST cases, 18-Nov-1998 (cf. 2.8) ------------
      This refers to an undefined parameter entity reference within
      a markup declaration in the internal DTD subset, violating
      the PEs in Internal Subset WFC.
  
   HI No error for non-wellformed document reported
      xmltest\valid/sa/094.xml
      None

Is this one error for the non-validating, and one for the validating
parser?

Also, how are you instructing the parser to be
validating/non-validating? You need to use the setValidationScheme()
method and the Val_Always or Val_Never enums, or it won't work
properly. When I tried Val_Auto, I got lots of errors. 

Can you run DOMCount/SAXCount from the Xerces-C distro on the tests
that are giving you errors to double check if any thing is wrong with
your installation (use the -v=always/never flags to set validation)?
If DOMCount gives a different answer, then it's the Python interface.

I have only run the validating parser. I'll run the non-validating
parser now and report my results.

jas.


---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to