RE: XMLForm Schematron question

[Delis, Christopher E.]

on a slightly different note, how does one reference another rule context.
i.e.:

let's say i have two elts, "/password" and "/password2":

        <pattern name="Password Validation Pattern" id="password">
                <rule context="/password">
                        <assert test="string-length(.) &gt; 5">Name
                                should be at least 6 characters.</assert>
                        <assert test="string-length(.) &lt; 60">Password
should be less
                                than 60 characters.</assert>
                </rule>
                <rule context="/password2">
                        <assert test="string-length(.) &gt; 5">Name
                                should be at least 6 characters.</assert>
                </rule>
        </pattern>

...how can i validate that "/password" and "/password2" are equal in the
rule for "/password"?  i could use a predicate, but that won't give the user
an error message (from an assert test).  perhaps, this solution is just as
simple (i hope)? =-)  [i'm pretty sure i tried referencing (../password2)
but it didn't seem to work.]


> -----Original Message-----
> From: Rob Johnston [mailto:[EMAIL PROTECTED]
> Sent: Wednesday, March 26, 2003 10:35 AM
> To: [EMAIL PROTECTED]
> Subject: Re: XMLForm Schematron question
> 
> 
> On Wed, 26 Mar 2003, Joerg Heinicke wrote:
> 
> > Hello Rob,
> > 
> > you simply must add a predicate to the context:
> > 
> > <rule context="/data/number[/data/selected = 'number']">
> >   <assert test="number(.)=.">
> >     You must specify a valid number.
> >   </assert>
> > </rule>
> 
> Sweet.  I hadn't even thought about putting a predicate there.  That
> worked perfectly.  I appreciate it!
> 
> thanks,
> rob
> 
> > Furthermore I think you should change the test pattern. If 
> someone writes 
> > 5.0, it can return false. I have in mind 5.0 and 5, if the 
> transformer 
> > tests on strings and not on numbers. (I had the same 
> problem a few days ago 
> > with XSLT. Is the behaviour simply processor dependent or specified 
> > anywhere?) If you change the test to "number() = number()" 
> you are on the 
> > safe side, because "NaN = NaN" returns always false.
> > 
> > Regards,
> > 
> > Joerg
> > 
> > Rob Johnston wrote:
> > >Hello -- I have a quick Schematron/XMLForm question.
> > >
> > >Let's say I have the following XPath:
> > >
> > >/data/selected    which can be either "text" or "number"
> > >/data/number      a valid number
> > >
> > >I want to write a validation pattern that will assert iff 
> /data/selected
> > >is 'number' and /data/number must be a number.
> > >
> > >For instance:
> > >
> > ><rule context="/data/number">
> > >  <assert test="number(.)=.">
> > >    You must specify a valid number.
> > >  </assert>
> > ></rule>
> > >
> > >works fine, but I want the above to happen ONLY if 
> /data/selected is the
> > >value "number".  So if /data/selected is the value "text", it will
> > >completely ignore the check to see if the /data/number field is a
> > >numeric value.
> > >
> > >Any help would be appreciated.
> > >
> > >thanks,
> > >rob
> > 
> > -- 
> > 
> > System Development
> > VIRBUS AG
> > Fon  +49(0)341-979-7419
> > Fax  +49(0)341-979-7409
> > [EMAIL PROTECTED]
> > www.virbus.de
> > 
> > 
> > 
> ---------------------------------------------------------------------
> > To unsubscribe, e-mail: [EMAIL PROTECTED]
> > For additional commands, e-mail: [EMAIL PROTECTED]
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
> 

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