It's a maven thing that causes this type of duplication.    Maven only puts 
the non-test code of one project on the classpath of the other.   Thus, the 
file in the first project is not usable in the second.    

There are two ways around it:

1) Create a new "testutils" project that creates a jar of test utilites.   One 
project can depend on that with scope of tests. 

2) Force creation of a "test-jar" in the first project.   Basically, add 
jar:test-jar to the package lifecycle.    Then depend on that in the test 
scope with a test classifier.   The problem is, this REQUIRES an "install" in 
the first project to get the jar in place.   A "mvn test" won't work as maven 
still won't put the stuff in target/classes-test on the classpath.

I know Jeremy, Jim, and I discussed something similar when I started the 
Celtix stuff.   There are a bunch of "mock" objects that are duplicated in 
several of the SCA projects for the same reason.  We have considered pulling 
them out into a testutils project to remove some of that duplication.

Dan


On Thursday June 22 2006 9:47 am, Frank Budinsky wrote:
> Simon, I think that a patch with your changes sounds good. We were already
> thinking that we will want an ignoreComments option anyway.
>
> I'm not sure how there got to be two versions. Jeremy added the first one
> (in the impl project), I believe. Not sure when the tools one crept in,
> but since the tools project depends on the impl project, I would think
> that it should be able to reuse the one from there - unless there's
> something special about it. Does anybody know why there are two?
>
> Frank.
>
> "Simon Laws" <[EMAIL PROTECTED]> wrote on 06/22/2006 08:34:08 AM:
> > I want to add a little automated testing to the interop suite to compare
> > input XML files with output XML files. I note that there is XML
>
> comparison
>
> > function in
> >
> > org.apache.tuscany.sdo.test.TestUtil
> >
> > However this test compares everything and I am hitting the problem which
>
> is
>
> > touched on in http://issues.apache.org/jira/browse/TUSCANY-427, i.e. my
> > input xml files have comments and my output XML files don't so the
>
> current
>
> > equalXML... function always returns false. For my local testing what I
>
> have
>
> > done is edited the class to:
> >
> > 1/ Change all the static operations to non static operations
> > 2/ Add a member variable to describe what should be ignored when testing
>
> for
>
> > equality
> > 3/ Add outputs in each failure case to record what causes XML files to
>
> be
>
> > marked as not equal
> > 4/ Added a string output stream member variable to capture all of this
> > output
> > 5/ Added a getReport method to return the contents of this output stream
>
> if
>
> > you are interested in why the XMLs differ
> >
> > Now this is all very well and good but of course this breaks anything
>
> that
>
> > relies on this class having static methods (it's used in about 50
>
> places).
>
> > So I can make a slight change to maintain the static interface. The
>
> question
>
> > is would you rather I:
> >
> > A/ Make a special version for me and not patch the original (not ideal)
> > B/ Create a patch for the original
> >
> > As an aside there already seems to be two copies of this file, in:
> >
> > tuscany\java\sdo\impl\src\test\java\org\apache\tuscany\sdo\test
> > tuscany\java\sdo\tools\src\test\java\org\apache\tuscany\sdo\test
> >
> > Is there a good reason for this? If not, which one stays?
> >
> > Simon
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]

-- 
J. Daniel Kulp
Principal Engineer
IONA
P: 781-902-8727    C: 508-380-7194   F:781-902-8001
[EMAIL PROTECTED]

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to