[
Lists Home |
Date Index |
Thread Index
]
----- Original Message -----
From: "Henry S. Thompson" <ht@cogsci.ed.ac.uk>
To: "Dare Obasanjo" <kpako@yahoo.com>
Cc: "xml-dev" <xml-dev@lists.xml.org>
Sent: Friday, January 04, 2002 5:05 AM
Subject: Re: [xml-dev] Question on using xs:pattern to restrict xs:dateTime
> > From what I've seen the pattern facet can easily be made to create
> > illegal restrictions which none of the validating parsers I have
> > tried warns about.
>
> Um, now I'm confused -- how does your example's use of pattern create
> 'illegal restrictions'?
I guess "illegal restriction" is the wrong terminology.
> > > > <xsd:simpleType name="myTimeType">
> > > > <xsd:restriction base="xsd:dateTime">
> > > > <xsd:pattern value="\d\d:\d\d(:\d\d)?" />
> > > > <xsd:whiteSpace value="collapse" />
> > > > </xsd:restriction>
> > > > </xsd:simpleType>
>
> It's true there are no valid instances of this type, but that's not
> illegal, just useless. It's open to processors to issue warnings when
> a type has no members, but they're not required to.
>
This is unfortunate because a schema author may create the above type and
assume it is a valid restriction of the xs:dateTime type since validating
parsers never complain only for it to fail whenever used with a [seemingly]
cryptic error message.
I guess it probably would be difficult to validate that the contents of the
value space of the pattern facet are a subset of the contents of the value
space of the original type and thus most (or all?) validating parsers just
don't do it. . *sigh*
--
THINGS TO DO IF I BECOME AN EVIL OVERLORD #16
I will never utter the sentence "But before I kill you, there's just
one thing I want to know."
_________________________________________________________
Do You Yahoo!?
Get your free @yahoo.com address at http://mail.yahoo.com
|
