C# webservice - wrong minOccurs spec on WSDL ?

[rt.64]

Consider the following C# webservice code and the corresponding WSDL
generated by .NET:

C# Webservice (Fragment)
================
...
public struct SSS
{
[XmlElement("AAA")] public decimal aaa ;
[XmlElement("BBB")] public string bbb ;
}
....
public int (SSS args) {

WSDL (Fragment)
==============
...
<s:complexType name="SSS">
<s:sequence>
<s:element minOccurs="1" maxOccurs="1" name="AAA"
type="s:decimal" />
<s:element minOccurs="0" maxOccurs="1" name="BBB"
type="s:string" />
</s:sequence>
</s:complexType>
...

My problem is: I don't understand why the element AAA is marked as
mandatory while BBB is left as optional on the WSDL description. And
worst,
I don't understand why, despite this, a SOAP request without any AAA
element is
considered Ok by the webservice - I thought the absence of an AAA
element on
the SOAP request would trigger an error.

[Mrozik]

On Tue, 31 Jul 2007 04:59:05 +0200, <rt***@hotmail.comwrote:

Consider the following C# webservice code and the corresponding WSDL
generated by .NET:

C# Webservice (Fragment)
================
...
public struct SSS
{
[XmlElement("AAA")] public decimal aaa ;
[XmlElement("BBB")] public string bbb ;
}
....
public int (SSS args) {

WSDL (Fragment)
==============
...
<s:complexType name="SSS">
<s:sequence>
<s:element minOccurs="1" maxOccurs="1" name="AAA"
type="s:decimal" />
<s:element minOccurs="0" maxOccurs="1" name="BBB"
type="s:string" />
</s:sequence>
</s:complexType>
...

My problem is: I don't understand why the element AAA is marked as
mandatory while BBB is left as optional on the WSDL description. And
worst,
I don't understand why, despite this, a SOAP request without any AAA
element is
considered Ok by the webservice - I thought the absence of an AAA
element on
the SOAP request would trigger an error.

string is a nullable type where decimal not....

Mrozik

[rt.64]

>

string is a nullable type where decimal not....

Ok. But what I also said is that, despite this, a SOAP request WITHOUT
the AAA element (ie., the thought mandatory decimal element) is
considered valid (and the value 0 (zero) is assigned to the variable
'aaa'). If so, why not mark it as optional on the WSDL? And more ..
If zero is the implicit default value for decimal types, why this is
not reflected on the generated WSDL. I think that the 'correct' WSDL
for this element would have 'minOccurs=0' and 'default=0'. By the way,
is there any way of modifying the WSDL serialization proccess so to
have this mentioned syntax as the result?

[Mrozik]

On Tue, 31 Jul 2007 14:57:22 +0200, <rt***@hotmail.comwrote:

>>
string is a nullable type where decimal not....

Ok. But what I also said is that, despite this, a SOAP request WITHOUT
the AAA element (ie., the thought mandatory decimal element) is
considered valid (and the value 0 (zero) is assigned to the variable
'aaa'). If so, why not mark it as optional on the WSDL? And more ..
If zero is the implicit default value for decimal types, why this is
not reflected on the generated WSDL. I think that the 'correct' WSDL
for this element would have 'minOccurs=0' and 'default=0'. By the way,
is there any way of modifying the WSDL serialization proccess so to
have this mentioned syntax as the result?

because default value of decimal variable is 0 - try use decimal? datatype
if you want to send null value to WS and check if paremeter was set.
Mrozik

[rt.64]

I understand that 0 is the default value for decimal datatypes. But
that is not the problem. The problem is that the .NET serialized WSDL
is WRONG. And It's causing me problems with some Java clients
(developed by another company) that are trying to access my
webservice. Basically the client software is reading my WSDL and
understanding that AAA (the decimal element) is mandatory, refusing to
make the request when no value is available for AAA (while it would be
perfectly OK to send the request without AAA, as it's proved that it
would cause no error). So the solution would be to provide a 'correct'
WSDL, ie., one that had AAA marked as optional (with minOccurs=0
instead of minOccurs=1). But I don't know how to do that.

[John Saunders [MVP]]

<rt***@hotmail.comwrote in message
news:11*********************@k79g2000hse.googlegro ups.com...

>I understand that 0 is the default value for decimal datatypes. But
that is not the problem. The problem is that the .NET serialized WSDL
is WRONG. And It's causing me problems with some Java clients
(developed by another company) that are trying to access my
webservice. Basically the client software is reading my WSDL and
understanding that AAA (the decimal element) is mandatory, refusing to
make the request when no value is available for AAA (while it would be
perfectly OK to send the request without AAA, as it's proved that it
would cause no error). So the solution would be to provide a 'correct'
WSDL, ie., one that had AAA marked as optional (with minOccurs=0
instead of minOccurs=1). But I don't know how to do that.

But it _would_ cause an error if AAA was not sent. That would violate the
WSDL and probably cause an error during deserialization.

Perhaps you'd like the WSDL to indicate that AAA is nillable. Have you tried
to declare AAA as "decimal?"
--
John Saunders [MVP]

[rt.64]

But it _would_ cause an error if AAA was not sent. That would violate the
WSDL and probably cause an error during deserialization.

That's exactly the point - there's NO error when AAA is absent
(because it has an implicit default value).

I think we should analyse who cames first, the program or the WSDL. In
this case, the program cames first, and for this particular program
AAA is an OPTIONAL element - it can be supressed from the request with
no consequence but the use of AAA default value. So, in my opinion,
the .NET automatically generated WSDL is wrong, as it does not have
AAA marked as an optional element. If the oposite was true (WSDL
first, with AAA marked as mandatory, as a requirement for the
program), the program should trigger an exception on the absence of
AAA (but it doesn't).

Perhaps you'd like the WSDL to indicate that AAA is nillable. Have you tried
to declare AAA as "decimal?"

Yes, I think this can be a solution. The problem is that the client
software is not able to understand the 'nillable="true"' declaration.
It only looks for "minOccurs=0" or "minOccurs=1". I will declare AAA
as "decimal?" and try to convince then to refine their software.

Thanks all for helping me on this subject

[rt.64]

I tryed the use of the "decimal?" datatype (in this case the
'isNillable="true"' declaration is serialized) but a SOAP request with
'<AAA xsi:nil="true" />' triggers the following error: "Input string
was not in a correct format."

[John Saunders [MVP]]

<rt***@hotmail.comwrote in message
news:11*********************@d55g2000hsg.googlegro ups.com...

>I tryed the use of the "decimal?" datatype (in this case the
'isNillable="true"' declaration is serialized) but a SOAP request with
'<AAA xsi:nil="true" />' triggers the following error: "Input string
was not in a correct format."

Can you include the full stack trace, please? I'm interested in seeing which
piece of code it is that complains about the format.
--
John Saunders [MVP]