[
https://issues.apache.org/jira/browse/AXIS2C-843?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12552895
]
Bill Mitchell commented on AXIS2C-843:
--------------------------------------
Your observation is correct. That's what I get for assuming that the .wsdl I
receive matches the responses I see from the server. As the server side is not
wsdl based, it can, and in this case does, act differently than the wsdl says.
A corrected version of the wsdl would read:
<element name="getExemplarResponse">
<complexType>
<complexContent>
<restriction base="anyType">
<sequence>
<element name="exemplar" minOccurs="1"
maxOccurs="1">
<complexType>
<sequence>
<any namespace="##local"
minOccurs="1" maxOccurs="unbounded"/>
</sequence>
<attribute name="handle" type="string"
use="required"/>
</complexType>
</element>
</sequence>
<attribute name="responseCode" type="integer"
use="optional" default="0"/>
<attribute name="responseMessage" type="string"
use="optional"/>
<attribute name="supportedMethods"
type="fw:MethodListType" use="optional"/>
</restriction>
</complexContent>
</complexType>
</element>
After making this change to the .wsdl, I now see a different set of errors, now
compilation errors. I can detail these in a separate issue, but the one that
is immediately obvious is that the generated .h and .c files refer to the type
axis_array_list_t where I expect it means axutil_array_list_t *.
> Generated deserialize method does not handle complex type containing sequence
> of type any
> -----------------------------------------------------------------------------------------
>
> Key: AXIS2C-843
> URL: https://issues.apache.org/jira/browse/AXIS2C-843
> Project: Axis2-C
> Issue Type: Bug
> Components: code generation
> Affects Versions: 1.1.0
> Environment: Windows XP, Visual Studio 2005, guththila parser, libcurl
> Reporter: Bill Mitchell
> Assignee: Dimuthu Gamage
> Attachments: adb_exemplar_type0.new.c, adb_exemplar_type0.orig.c,
> unrolled.wsdl
>
>
> When WSDL2C generates a stub for the following response message description,
> it constructs a stub for entity exemplar. For this entity, it defines a
> single property of type axiom_node_t, the deserialize routine parses only the
> first of the "any" elements, and ignores the remainder. The second issue is
> that it parses the "any" entity only if its name is "extraElement", which is
> not very useful as it has some name, unknown at compilation time, but which
> is not likely to be "extraElement".
> <element name="getExemplarResponse">
> <complexType>
> <complexContent>
> <restriction base="anyType">
> <sequence>
> <element name="exemplar">
> <complexType>
> <sequence>
> <any namespace="##local"
> minOccurs="1" maxOccurs="1"/>
> </sequence>
> <attribute name="handle"
> type="string" use="required"/>
> </complexType>
> </element>
> </sequence>
> <attribute name="responseCode" type="integer"
> use="optional" default="0"/>
> <attribute name="responseMessage" type="string"
> use="optional"/>
> <attribute name="supportedMethods"
> type="fw:MethodListType" use="optional"/>
> </restriction>
> </complexContent>
> </complexType>
> </element>
> My suggestion is that the comparison for the QName "extraElement" should be
> removed, and the property of extraElement should be an array of axiom_node_t
> representing all of the sequence of "any" entities. This impacts the
> deserialize and serialize routines, and the resetter method to avoid the
> potential memory leak. I've seen such code generated for a sequence of
> elements of a simple type, so it should not be too hard to generate the same
> code here in a complexType.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
