Re: [xml] Schema validity failure for valid document

[Kasimier Buchcik <kbuchcik 4commerce de>]

Hi,

Daniel Veillard wrote:
> On Wed, Jan 12, 2005 at 03:13:00PM +0100, Kasimier Buchcik wrote:
>
> > >  nb = 3
> > >  nbneg = 1
> > >  values[0] = "b|http://FOO";
> > >  values[1] = "*|*"
> > >  values[2] = "c|http://FOO";
> > >  values[3] = "*|http://FOO";
> > >  
> > > I think such an API would allow complete analysis and report without
> > > needing to provide extra implementation details or new strings in the 
> > > regexp format.
> >
> > Cool, adding a nbneg sounds simpler than anything else.
>
>
>   okay, it's in CVS already :-)
>
>
> > But there is still a problem: the negated namespace wildcard is build
> > using two transitions: "*|http://FOO"; leading to the sink state, and
> > "*|*" passing all other elements through. The "*|*" would not be
> > distinguishable from a <any namespace="##any"/> wildcard. Maby the
> > build of the automaton can be changed somehow to avoid this. Hmm,
> > is it possible to extend the sink state detection to handle this,
> > i.e. to surpress the report of such a "only-way-out-of-sink" transition?
>
>
>   I'm not 100% sure what you would like to be reported.
>
>       ------ *|* -------> S1
>      /
>   S0
>      \
>       -- *|http://FOO --> S2

Nice, I like such pictures.

>   S2 is the sink state. When the error is raised on S0 because 
>   {http://FOO}a  is pushed you currently get
>   nb = 1, nbneg = 1
>   values[0] = "*|*"
>   values[1] = "*|http://FOO";
>
> What do you want instead ? 
>   nb = 0
>   nbneg = 1
>   values[0] = "*|http://FOO";

Yes, please :-)

> I'm not sure the low level xmlregexp code is better at interpreting
> why that construct was used, since it is specific to XMLSchemas.

I'm not sure it's specific to XML Schemata; I'm not deep into automata,
and less into it's jargon, so I would naively call the picture above a
"filter" pattern, and filters should not be used with XML Schemata only.

> Seems to me that this reduction is better done at the level which 
> can make an interpretation of the transition values.

Given that every negated namespace comes with an additional "*|*",
remove, for every negated namespace, one "*|*".
All "*|*" left can then be intepreted as <any namespace="##any"/>.
Yes this could be done on the schema side.

Additionally I noticed that a combination of <any namespace="##any"/>
and a negated namespace should never be a result of the error info,
since it would be ambigious; a quick test revealed that the regex
reports a non-determinist content model, if using such a combination.

So indeed, it seems to be complete. Cool :-)


Regards,

Kasimier