Ok.. I got it... even though I define a method for parsing the value, it
should still confirm to the date format of CCYY-MM-DD. I mis-understood it
as a way of overriding the default validation i.e if I have a method
customizing the conversion of xs:date to jxb:java.util.Calendar, I
understood it as the method that would get executed whenever a xs:date field
is encountered during unmarshalling and the default validation will not be
executed ....
So if a element is defined as xs:date it will always need to be in the
format CCYY-MM-DD ? Is there a way to change this format ?
Looks like the following happens... when a element defined as date and has
the value 20030410
1. An even gets thrown, because it does not fall in the default format of
CCYY-MM-DD
2. My customized method gets executed, which converts it into a
java.util.Caledar object.
My process aborts when there are ValidationEvents.. So even though the
customized convertor, converted the value... because of the ValidationEvent,
the process aborts...
-----Original Message-----
From: Kohsuke Kawaguchi [mailto:Kohsuke.Kawaguchi_at_Sun.COM]
Sent: Thursday, April 10, 2003 3:42 PM
To: JAXB-INTEREST_at_JAVA.SUN.COM
Subject: Re: Customaziation and validation
> Woudn't it be valid according to the Schema specs ? The schema defines
an
> element as type SimpleType and as the one which has different conversion
> rules ?
The conversion rules are defined by JAXB. So it's not something the XML
Schema spec recognizes. And YYYYMMDD is NOT a valid xs:date value.
It's this simple; your schema says the AccidentDate element will have
the xs:date type. YYYYMMDD is not a valid xs:date value. Therefore
<AccidentDate>20030410</AccidentDate> is not valid wrt your schema.
Just try any XML schema processor out there. They should agree with this.
> So does this mean that if I have validation turned on, then Conversion
using
> other methods cannot be done ?
You can. It's just that the meaning of the validation is not affected by
any of the JAXB customization. XML Schema validation is defined purely
on the schema and the instance.
> Just curious, one solution I can think in my situation is to catch that
> event and discard it as if it did not happen...
Would you elaborate what you think a problem is? Because I don't see any
problem.
regards,
--
Kohsuke KAWAGUCHI 408-276-7063 (x17063)
Sun Microsystems kohsuke.kawaguchi_at_sun.com