[ome-devel] XSV crash on wildcard item

[Curtis Rueden]

Thanks Isaak,

This fix will be useful for testing (using validation) OME-XML
produced at LOCI, particularly by our WiscScan acquisition software,
since we use some custom semantic types.

-Curtis

On 12/8/06, Isaak Berg <isaak at cimaging.net> wrote:
> I found out that adding both
> minOccurs="0"
> and
> maxOccurs = "unbounded"
> to the wildcard declaration in CA.xsd resolves the crash and allows the
> processContent="skip" attribute to work as expected, with both xsv
> validation and msXML parser. It seems like maybe explicit declarations of
> values for both minOccurs and maxOccurs are pre-requisites for using the
> processContents directive...
>
> So if you merely update CA.xsd to include both minOccurs and maxOccurs (I
> assume 0 and unbounded are acceptable respective values?) in the xsd:any
> line, validation by XSV and MSXML6.0 parsers can occur correctly, even with
> elements present in the custom attributes for which schema definitions
> cannot be found or aren't available.
>
> Hope that helps someone else too, it will help me a lot if this can be done.
>
> Thanks for all your help!
>
>
> -----Original Message-----
> From: Ilya Goldberg [mailto:igg at nih.gov]
> Sent: Thursday, December 07, 2006 5:21 PM
> To: Isaak Berg
> Cc: ome-devel at lists.openmicroscopy.org.uk
> Subject: Re: [ome-devel] FW: XSV crash on wildcard item
>
> That's rather odd and disquieting.  It should be possible to route around
> the CA schema, because its really meant more for internal use, and doesn't
> really provide anything all that useful anyway.  I'll take a look to see
> what elements we use from CA.  Elimination of the CA dependency should be a
> strong candidate modification for the forthcoming XML update.
> -Ilya
>
>
> On Dec 7, 2006, at 4:02 PM, Isaak Berg wrote:
>
> > For your reference, because XSV validation currently crashes on the
> > wildcard element found in the (Custom Attributes) CA.xsd
> >
> > From: Isaak Berg
> > Sent: Thursday, December 07, 2006 3:47 PM
> > To: 'xmlschema-dev at w3.org'
> > Subject: XSV crash on wildcard item
> >
> > Greetings,
> >
> > I get a crash reported by "XSV 2.9-1 of 2005/04/14 11:49:27" during
> > validation when it encounters a wildcard element. By using 'localized'
> > versions of the seven OME-defined schemas involved, I've isolated the
> > problem to the wildcard element by comparing the results before and
> > after commenting it out
> >
> > (By "localized", I simply mean I downloaded a copy of each, put them
> > all in the same directory, and updated all the
> > schemalocation="<value>" values previously pointing to the
> > www.openmicroscopy.org server to that directory. I also assigned new
> > namespace names although that probably was unnecessary. I know all
> > this was done correctly because I can get my modified version (with
> > the wildcard element in CA.xsd changed to non-wildcard
> > content) to validate the localized version of Samp l e.ome, whereas
> > leaving the wildcard in yields the same error message from my
> > localized setup as I get if I validate Sample.ome directly from its
> > http: link referenced below. Probably pretty much what anyone else
> > would do... unless the XSV command-line can somehow make all that
> > unnecessary?)
> >
> > The subject of the validation when this occurs is the file found at "
> > http://www.openmicroscopy.org/XMLschemas/OME/latest/Sample.ome
> > "
> > and the error occurs while parsing the schema referenced at "
> > http://www.openmicroscopy.org/XMLschemas/CA/RC1/CA.xsd
> > "
> >
> > Any input would be greatly appreciated.
> > Isaak Berg
> >
> > Software Developer
> >
> > isaak at cimaging.net
> >
> > Compix Inc., Imaging Systems
> >
> > 109 Nicholson Road
> >
> > Sewickley, PA 15143
> >
> > 412-741-7920
> >
> > 412-741-7930 Fax
> >
> > www.cimaging.net
> >
> >
> >
> >
> > _______________________________________________
> > ome-devel mailing list
> > ome-devel at lists.openmicroscopy.org.uk
> > http://lists.openmicroscopy.org.uk/mailman/listinfo/ome-devel
>
> _______________________________________________
> ome-devel mailing list
> ome-devel at lists.openmicroscopy.org.uk
> http://lists.openmicroscopy.org.uk/mailman/listinfo/ome-devel
>