[oXygen-user] Validation of Schema that Imports Multiple XSD Schemas

Wed Aug 15 03:57:27 CDT 2012

Hello Bob,

We managed to identify this is a Saxon problem. If I validate 
"Relation_Image_List13Aug12.xml" with Saxon from command line the 
validation takes about 60 seconds on my computer. We will report this 
issue to Saxon. Until it is resolved, we can provide you a beta version 
of oXygen 14.1 that includes also Xerces as an XML Schema 1.1 engine.

Regards,
Octavian
---------------
Octavian Nadolu
<oXygen/> XML Editor
http://www.oxygenxml.com

On 14.08.2012 21:23, George Cristian Bina wrote:
> Dear Bob,
>
> Since version 14 release we were working on improving the XML Schema 1.1
> support. That includes also supporting Xerces as an XML Schema 1.1
> engine. Validating your Relation_Image_List13Aug12.xml with Xerces
> finishes in less than one second while using Saxon I got 36 seconds.
> Thanks for sending the sample files, we will be able to investigate what
> happens when we use Saxon for validation and determine if the issue is
> in how we use Saxon or in Saxon itself.
>
> Best Regards,
> George
> --
> George Cristian Bina
> <oXygen/> XML Editor, Schema Editor and XSLT Editor/Debugger
> http://www.oxygenxml.com
>
> On 8/14/12 9:08 PM, Robert Leif wrote:
>> Hi Adrian,
>>
>> I checked the paths of the files in the project. They were all in the
>> same directory. oXygen_Global_Options.xml contains the system
>> information. I created and validatedRelation_Image_List13Aug12 with
>> XMLSpy. Validation took 1 to 2 seconds. The same file in oXygen already
>> has the description at over 7,500 items and growing! 50% of the 2 CPUs
>> is being used. I obtained the following message from the bottom panel:
>> Description: Loading catalog:
>> file:/C:/Program%20Files/Oxygen%20XML%20Editor%2014/frameworks/catalog.xml
>>
>>
>> I attempted to validate after about an hour< I received an error message
>> that the system had run out of heap space.
>>
>> System ID: D:\CytometryML\ACS\xsd\Relation_Image_List13Aug12 .xml
>>
>> Main validation file:
>> D:\CytometryML\ACS\xsd\Relation_Image_List13Aug12 .xml
>>
>> Schema: D:\CytometryML\ACS\xsd\relations_image.xsd
>>
>> Engine name: Saxon-EE 9.4.0.3
>>
>> Severity: error
>>
>> Description: Java heap space
>>
>> If the problem is the number of schemas imported, the parser can look in
>> each schema and determine if the prefix is used, if not a comment could
>> be placed next to each import element. The user could then erase it or
>> comment it out. This is equivalent to a compiler parser looking for
>> unused code.
>>
>> Thank you.
>>
>> Yours,
>>
>> Bob Leif
>>
>> -----Original Message-----
>>
>> From: Oxygen XML Editor Support [mailto:support at oxygenxml.com]
>> <mailto:[mailto:support at oxygenxml.com]>
>>
>> Sent: Monday, August 13, 2012 6:21 AM
>>
>> To: rleif at rleif.com <mailto:rleif at rleif.com>
>>
>> Cc: oXygen-user at oxygenxml.com <mailto:oXygen-user at oxygenxml.com>
>>
>> Subject: Re: [oXygen-user] Validation of Schema that Imports Multiple
>> XSD Schemas
>>
>> Hello Robert,
>>
>> When validating XML files, Oxygen provides a caching mechanism for the
>> schema used for validation, so only the first validation will evaluate
>> the entire schema hierarchy. The following XML validations will use the
>> cached schema.
>>
>> When validating XML schema files, the entire schema hierarchy starting
>> from the current schema is validated every time. However, the validation
>> time that you report seems extreme for an XML Schema validation, even if
>> a large schema hierarchy is validated. Do you have a validation scenario
>> with multiple validation units or are you using the default validation
>> as it is offered by Oxygen?
>>
>> Make sure the schemas are all local. In situations where schemas are
>> imported from remote locations, make sure these remote URIs are resolved
>> through an XML catalog to a local copy. My guess is some of the schemas
>> from your schema hierarchy are requested from a remote location and
>> that's what's causing the huge validation time.
>>
>> To determine if some schemas are retrieved from remote locations, you
>> can enable XML Catalog verbosity in Oxygen from Options > Preferences,
>> XML > XML Catalog. Set "Verbosity" to "All messages". Whenever you
>> validate or perform an Oxygen operation that involves the catalog
>> resolver, the attempts to resolve URIs, systemIDs or namespaces
>> (depending on your Oxygen catalog options) will appear in the Catalogs
>> view from the results panel (at the bottom of the editor window).
>>
>> Note that all attempts at resolving something are listed there, so it's
>> possible you'll see a lot of redundant messages, first some that fail
>> and eventually one that succeeds in resolving the URI.
>>
>> We are not aware of any schema caching mechanisms that Saxon offers in
>> this regard, but we have not investigated this possibility. But, my
>> belief is that the problem you have encountered lies elsewhere. There
>> seems to be a bottleneck somewhere that slows things down significantly
>> and my first guess is that some of the schema files are not immediately
>> accessible.
>>
>> By any chance, have you tried to validate the same schema hierarchy with
>> Saxon from the command line? Note that you'll need a separate Saxon
>> deployment since the one bundled with Oxygen is only licensed to be used
>> from within the application GUI.
>>
>> Regards,
>>
>> Adrian
>>
>> Adrian Buza
>>
>> oXygen XML Editor and Author Support
>>
>> Tel: +1-650-352-1250 ext.202
>>
>> Fax: +40-251-461482
>>
>> support at oxygenxml.com <mailto:support at oxygenxml.com>
>>
>> http://www.oxygenxml.com
>>
>> Robert Leif wrote:
>>
>>  >
>>
>>  > Since my designs follow object-oriented design principles, my XSD
>>
>>  > schemas import multiple XSD schemas, which provide and permit reuse of
>>
>>  > many of their simpleTypes and complexTypes. The time for a successful
>>
>>  > validation of a schema that solely imports 9 other schemas and does
>>
>>  > not have any data-type or element definitions was in excess of 10
>>
>>  > minutes. The subsequent validation took 3 minutes. 2.66  out of 4
>>
>>  > gigabytes of memory and 50% of the CPU were used. I validated with
>>
>>  > Saxon-EE 9.4.0.3 for XSD1.1.
>>
>>  >
>>
>>  > I suspect that the problem is that Saxon validated all of the imported
>>
>>  > schemas. Since these schemas had NOT been changed after their previous
>>
>>  > validation, this was a total waste of time! Does Saxon have a tool to
>>
>>  > check the time of the previous validation against the time of the last
>>
>>  > edit? If not, it needs one. There is a potential problem when two
>>
>>  > schemas import each other. In this case, it probably pays to validate
>>
>>  > both. A quick way to ameliorate this problem is to extend the
>>
>>  > validation-well-formatted pull-down to include Validate only this
>>
>>  > schema. Since a schema tree structure is needed to perform an
>>
>>  > intelligent validation, it would be useful to make this tree available
>>
>>  > to the user and to note if any of the imported schemas were not being
>>
>>  > used.
>>
>>  >
>>
>>  > Has anyone else observed similar long validation times with another
>>
>>  > validator?
>>
>>  >
>>
>>  > Thank you.
>>
>>  >
>>
>>  > Yours truly,
>>
>>  >
>>
>>  > Bob Leif
>>
>>  >
>>
>>  > ----------------------------------------------------------------------
>>
>>  > --
>>
>>  >
>>
>>  > _______________________________________________
>>
>>  > oXygen-user mailing list
>>
>>  > oXygen-user at oxygenxml.com <mailto:oXygen-user at oxygenxml.com>
>>
>>  > http://www.oxygenxml.com/mailman/listinfo/oxygen-user
>>
>>  >
>>
>>
>>
>> _______________________________________________
>> oXygen-user mailing list
>> oXygen-user at oxygenxml.com
>> http://www.oxygenxml.com/mailman/listinfo/oxygen-user
>>
>