From travelwitek at gmail.com  Fri Jan 28 04:39:25 2011
From: travelwitek at gmail.com (Witek Piet)
Date: Fri Jan 28 04:39:33 2011
Subject: [xsde-users] XML schema validation (XSD) and XSD/e
Message-ID: <AANLkTi=ptuuYEdVd4afTJSZF0rChCx6-zVuRpA3kT341@mail.gmail.com>

Dear Support,

I am using XSD/e for serialization and parsing of data. I as I have a
question regarding which type of validation is performed by XSD/e. More
specifically I have added the following to my XSD-file:

<simpleType name="trigger_interval">
  <restriction base="int">
    <minInclusive value="1"/>
    <maxInclusive value="3600"/>
  </restriction>
</simpleType>


I had expected that the parser/serialization code generated by XSD/e would
include validation of the range for elements of type trigger_interval - but
it does not. Could you please elaborate on which type of validation XSD/e
provides?

Thanks in Advance,
Witek
From boris at codesynthesis.com  Fri Jan 28 05:35:25 2011
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Fri Jan 28 05:21:03 2011
Subject: [xsde-users] XML schema validation (XSD) and XSD/e
In-Reply-To: <AANLkTi=ptuuYEdVd4afTJSZF0rChCx6-zVuRpA3kT341@mail.gmail.com>
References: <AANLkTi=ptuuYEdVd4afTJSZF0rChCx6-zVuRpA3kT341@mail.gmail.com>
Message-ID: <boris.20110128122921@codesynthesis.com>

Hi Witek,

Witek Piet <travelwitek@gmail.com> writes:

> <simpleType name="trigger_interval">
>   <restriction base="int">
>     <minInclusive value="1"/>
>     <maxInclusive value="3600"/>
>   </restriction>
> </simpleType>
> 
> I had expected that the parser/serialization code generated by XSD/e would
> include validation of the range for elements of type trigger_interval - but
> it does not.

The numerical facets are not supported in the current release (3.1.0) but
are implemented in the upcoming 3.2.0 (it is in the final testing stage and
should be out in the next few weeks). If you would like, you can try the
beta version, which is available here:

http://www.codesynthesis.com/products/xsde/download.xhtml

Generally, the list of validated XML Schema constructs for parsing is listed
in the C++/Parser Getting Started Guide:

http://www.codesynthesis.com/projects/xsde/documentation/cxx/parser/guide/#A

And for serialization -- in the C++/Serializer Getting Started Guide:

http://www.codesynthesis.com/projects/xsde/documentation/cxx/serializer/guide/#A

The above documentation is for 3.1.0 so if you would like to see the list
for 3.2.0, you will need to download the beta.

Boris

From ivan.lelann at free.fr  Fri Jan 28 09:24:32 2011
From: ivan.lelann at free.fr (Ivan Le Lann)
Date: Fri Jan 28 09:24:41 2011
Subject: [xsde-users] Feature proposal : throw an exception when accessing
 non-present member
In-Reply-To: <116598094.2848321296224641414.JavaMail.root@zimbra36-e6.priv.proxad.net>
Message-ID: <294721245.2848341296224672302.JavaMail.root@zimbra36-e6.priv.proxad.net>

Hi all,

Here is the typical generated code for a variable length member.

    inline
    ::bar_type& foo_type::
    bar ()
    {
      return *this->bar_;
    }

Obviously, calling bar() when bar_present() is false is an access violation.
For a fix length member, you get something between a random value or a default one.
I consider current behavior as both inconsistent and tricky.

Now, how about throwing an exception with very basic information?
I dislike having to do C-style checks to show decent error report.
The concern is mostly for variable length elements but I'd vote for consistency.

Should this proposal meet success, I'd be glad to help with an implementation.
This should probably be an option, conflicting with no-exception, and whose default 
value has to be carefully chosen.

Now one question is left wide open:
How to apply this to mandatory elements/attributes?
One could consider XML validation exceptions sufficient for those,
but that would keep the process of building from scratch hazardous ...

Regards,
Ivan

From boris at codesynthesis.com  Mon Jan 31 10:16:07 2011
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Mon Jan 31 10:01:03 2011
Subject: [xsde-users] Feature proposal : throw an exception when
	accessing non-present member
In-Reply-To: <294721245.2848341296224672302.JavaMail.root@zimbra36-e6.priv.proxad.net>
References: <116598094.2848321296224641414.JavaMail.root@zimbra36-e6.priv.proxad.net>
	<294721245.2848341296224672302.JavaMail.root@zimbra36-e6.priv.proxad.net>
Message-ID: <boris.20110131165057@codesynthesis.com>

Hi Ivan,

Ivan Le Lann <ivan.lelann@free.fr> writes:

> Obviously, calling bar() when bar_present() is false is an access violation.

Or, more precisely, it is a programming error that leads to an access
violation. One should never call an optional member accessor without
first checking for its presence.


> Now, how about throwing an exception with very basic information?

I think that's a bad idea. This is a programming error, not an unexpected
condition. It will also negatively affect performance and object code
size.


> I dislike having to do C-style checks to show decent error report.

Having the _present() accessor return false should no be an error. If it
is, then you should make this member required (in the schema).

Boris