Vescovi, Matteo wrote:
Hi Lars,
that's odd.
I am using the same setup, running on Linux with a UTF-8 encoding.
Yes, the encoding is relevant as it affects the very first line <?xml
version="1.0" encoding="UTF-8"?>, if nothing else.
Latest revision is 509955, not sure why you are on 509720.
This number seems to be changing with time. If you go to
https://svn.apache.org/repos/asf/incubator/yoko/trunk/tools/, you'll see
that it's now 510161. Probably it's a counter for all commits in the svn
repository (i.e. all apache projects).
I believe the tools generation tests should be improved: instead of
comparing the content of wsdl files as text, they could do a
comparison of the xml content. This would ignore whitespace
differences (including lines breaks and different closing style, i.e.
<empty></empty> vs <empty/> ).
Yes, that's a good idea, and I could do a simple change goes in that
direction and fixes the problem. Just trim the strings before they are
asserted equal.
However, I'd like to understand what's happening here. Somehow I suspect
that one of the development tools is fooling me... Are you using the
Eclipse svn plugin?
Can you check you are using wsdl4j 1.6.1?
Yes, ~/.m2/repository/wsdl4j/wsdl4j/1.6.1/wsdl4j-1.6.1.jar
Cheers,
Lars
Cheers,
- Matteo
Lars Kühne wrote:
Vescovi, Matteo wrote:
Hi Lars,
the tools module builds and tests fine for me.
Do you still get this failure?
Yes, I see this test failure in the Maven build and inside Eclipse.
When I debug IDLToWSDLGenerationTest.testStringGeneration(), I see
that the generated output is
<wsdl:definitions targetNamespace=...>
<corba:typeMapping
targetNamespace="http://schemas.apache.org/yoko/idl/String/typemap">
<corba:anonstring ...
...
<corba:struct ...
...
</corba:struct>
</corba:typeMapping>
<wsdl:types>
The indentation of the closing tag for typeMapping is incorrect (4,
should be 2), and the test catches that.
Not sure why this test fails for me and not for you. I tried to debug
this, but for now I got stuck in
WSDLCorbaWriterImpl.printExtensibilityElements() where the super
class is called, because that is inside wsdl4j.jar which doesn't seem
to have debug information included.
I'm working on Linux, tools revision 509720 (i.e. latest and
greatest). My platform encoding is UTF-8 if that matters (I saw that
the test uses the platform encoding for char<->byte conversion).
Cheers,
Lars