Forum Discussion

jramos's avatar
jramos
New Contributor
11 years ago

Error loading WSDL with UTF-8 non-ASCII characters

During the load of a WSDL file, SOAPUI displays the following message:

Error loading [file:\J:\TEMP\xxx\Service1.asmx@WSDL]: org.apache.xmlbeans.XmlException: java.io.CharConversionException: Malformed UTF-8 character: 0xe7 0xe3 0x6f

The sequence 0xe7 0xe3 0x6f is really a malformed UTF-8 character. This sequence represents the string "ção" with LATIN-1 encoded. However, this sequence does not exist in the WSDL file. The corresponding sequence in the WSDL file is c3 a7 c3 a3 6f, where "c3 a7" is ç", "c3 a3" is "ã" and 6f is "o".

  <wsdl:portType name="Service1Soap">
    <wsdl:operation name="HelloWorld">
      <wsdl:documentation xmlns:wsdl="http://schemas.xmlsoap.org/wsdl/">Action in portuguese is ação</wsdl:documentation>
      <wsdl:input message="tns:HelloWorldSoapIn" />
      <wsdl:output message="tns:HelloWorldSoapOut" />
    </wsdl:operation>
  </wsdl:portType>


The error occurs loading the WSDL from a file and also reading the web service URL (http:// ... ?wsdl).

I'm using SOAPUI 4.6.4 and Windows 7 32 bit.

VERY IMPORTANT: This error doesn't occur with SOAPUI 4.6.3 installed side-by-side.

Thanks,
Jorge Ramos
  • SmartBear_Suppo's avatar
    SmartBear_Suppo
    SmartBear Alumni (Retired)
    11 years ago
    4.6.4 is stricter when loading WSDL, which is affecting some users that have WSDLs that are not in UTF-8. If you want to use other encodings, you should add an XML declaration like this:

    <?xml version="1.0" encoding="ISO-8859-1" ?>

    This may be annoying to those affected, and I think we should show a friendly error message to the user (if error messages are ever friendly), but SoapUI is really helping you find a potentially serious problem with your WSDL. You shouldn't assume that every client has the same default encoding as you, and if the encoding is not specified, SOAP clients (actually, XML processors generally) are supposed to assume that the encoding is UTF-8.

    Hope this helps!

    /Manne, SoapUI developer
  • jramos's avatar
    jramos
    New Contributor
    11 years ago
    Thank you for your help, guys. There are no problem with the WSDL file. It's encoded correctly with UTF-8, as I demonstrated in the text, and this can be confirmed in the attached file also. SOAPUI 4.6.3 works well and I don't need the enhancements of 4.6.4. For a while I will use the 4.6.3.

    Many thanks,
    Jorge Ramos
  • jramos's avatar
    jramos
    New Contributor
    11 years ago
    Hey guys, after a little investigation in the source code of SOAPUI 4.6.4, I found something interesting.

    In WsdlLoader.java, Line 113, there is a call to readCleanWsdlFrom. This method calls XmlUtils.createXmlObject( new URL(url) ).xmlText(), which returns the XML text decoded from UTF-8 (or whatever encoding used) to standard UNICODE. After some text work, the method returns the XML string, maybe changed a little, to the caller.

    At line 114, the string content (already in standard UNICODE) is loaded into a byte array and submitted (again) to XmlUtils.createXmlObject, which will try to parse an UTF-8 string from a buffer converted to standard UNICODE. Result: Load error.

    I suspect this is a bug introduced by the above change from 4.6.3 to 4.6.4. In 4.6.3 there is only one call to XmlUtils.createXmlObject, and everything goes well.

    Thanks,
    Jorge Ramos
  • MarcusJ's avatar
    MarcusJ
    Moderator
    11 years ago
    Hi,

    I'm able to import the wsdl you attached by adding -Dfile.encoding=UTF8 in the <soapui install>\bin\*.vmoptions file and restarting SoapUI.
  • MarcusJ's avatar
    MarcusJ
    Moderator
    11 years ago
    Hi,

    This is not a bug. 4.6.4 is stricter in loading wsdl's as Manne pointed out in a previous post. There are some wsdl files that are not UTF-8 encoded, and some do not have an encoding attribute present in the xml header.
  • andersrh's avatar
    andersrh
    New Contributor
    11 years ago
    Thats not what he says, he says its a UTF-8 encoded file that does not work with SoapUI. SoapUI should be able to , but if you add a portugese or in my case a danish letter to a WSDL file, it will not be able be parsed by SoapUI.

    Try with 0xe7 or 0xC3(æ), they will mess with SoapUI even though the wsdl file is correct UTF-8.

    try here for instance: https://service-es.bygogmiljoe.dk/Sagsbehandling.svc?singleWsdl
  • SmartBear_Suppo's avatar
    SmartBear_Suppo
    SmartBear Alumni (Retired)
    11 years ago
    Hi Anders,

    I was able to import that WSDL without problem. Have you tried that particular WSDL URL?

    Regards,
    Manne
  • andersrh's avatar
    andersrh
    New Contributor
    11 years ago
    Yup, I just tried it... It worked fine in 4.6.1, but 4.6.4 gives the following error: