cxf-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Brendan Collins (JIRA)" <j...@apache.org>
Subject [jira] Commented: (CXF-1831) org.apache.cxf.wsdl.EndpointReferenceUtils.resolveResource() changes cause failure in schema import resolution
Date Tue, 30 Sep 2008 21:43:44 GMT

    [ https://issues.apache.org/jira/browse/CXF-1831?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12635867#action_12635867
] 

Brendan Collins commented on CXF-1831:
--------------------------------------

Changing the second line of EndpointReferenceUtils.resolveResource(), as suggested by Dan,
to read as follows:

if (baseURI != null && systemId != null) {  //add additional systemId null check

fixes the issue for me. My only concern/question is
why are systemId and  publicId null? The code path taken after the
change happens not to access either, but does set them both into the
returned impl.

That said everything seems to get resolved correctly.

> org.apache.cxf.wsdl.EndpointReferenceUtils.resolveResource() changes cause failure in
schema import resolution
> --------------------------------------------------------------------------------------------------------------
>
>                 Key: CXF-1831
>                 URL: https://issues.apache.org/jira/browse/CXF-1831
>             Project: CXF
>          Issue Type: Bug
>    Affects Versions: 2.1.1, 2.1.2
>            Reporter: Brendan Collins
>             Fix For: 2.1.3
>
>
> In moving from cxf 2.1.0 to 2.1.1 (or 2.1.2)  I am encountering the following problem:
> org.apache.cxf.wsdl.EndpointReferenceUtils.resolveResource() fails to resolve schema
imports because systemId is null when trying to resolve the URI.
> Interestingly the comment at the top of EndpointReferenceUtils reads as follows:
>    /**
>     * We want to load the schemas, including references to external schemas, into a SchemaFactory
>     * to validate. There seem to be bugs in resolving inter-schema references in Xerces,
so even when we are
>     * handing the factory all the schemas, interrelated with &lt;import&gt; elements,
we need
>     * to also hand over extra copies (!) as character images when requested.
>     *
>     * To do this, we use the DOM representation kept in the SchemaInfo. This has the
bonus
>     * of benefiting from the use of the catalog resolver in there, which is missing from
>     * the code in here.
>     */
> It looks like the code that this comment refers to was removed in 2.1.1, and that code
does in fact resolve my schema's correctly.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message