Re: Integration testing - we need some :-)

2006-10-19 Thread Jeremy Boynes
By making the test classloader a child of the plugin's we are able to run the tests - at least the scenario I had below. That's not ideal as it makes the plugin's classes visible to the tests but at least we can make progress on writing cases. So, with that thought in mind, where should we

Re: Integration testing - we need some :-)

2006-10-19 Thread Jim Marino
On Oct 18, 2006, at 11:12 PM, Jeremy Boynes wrote: By making the test classloader a child of the plugin's we are able to run the tests - at least the scenario I had below. That's not ideal as it makes the plugin's classes visible to the tests but at least we can make progress on writing

Re: Integration testing - we need some :-)

2006-10-18 Thread Bert Lamb
I actually set up a Continuum server to run Tuscany builds here in my office. I just wanted to check it out. It is quite easy to set up. Unfortunately, I do not have a machine I can donate to do this full time or that would be accessible externally. If you'd like I could write up some quick

Re: Integration testing - we need some :-)

2006-10-18 Thread Jeremy Boynes
That would be great - I have a remote machine we can use. I'm working on some itest mojo that starts/stop Tuscany inside Maven. I was thinking that could be used for in-VM testing or for running test clients that could use SCA to connect to a remote test server. -- Jeremy On Oct 18, 2006,

Re: Integration testing - we need some :-)

2006-10-18 Thread ant elder
On 10/17/06, Jeremy Boynes [EMAIL PROTECTED] wrote: In cutting M2 the lack of integration testing has been quite painful. We need to start adding the post-build functional/integration testing we've talked about before but which we've not actually implemented. I'm going to start work on some

Re: Integration testing - we need some :-)

2006-10-18 Thread Raymond Feng
@ws.apache.org Sent: Wednesday, October 18, 2006 6:16 AM Subject: Re: Integration testing - we need some :-) That would be great - I have a remote machine we can use. I'm working on some itest mojo that starts/stop Tuscany inside Maven. I was thinking that could be used for in-VM testing

Re: Integration testing - we need some :-)

2006-10-18 Thread Brent Daniel
I can take a look at doing some of this, perhaps starting with getting a simple sca application automated and going from there. Any direction from the list on specific areas that need integration test coverage would be helpful. Brent On 10/17/06, Jeremy Boynes [EMAIL PROTECTED] wrote: In

Re: Integration testing - we need some :-)

2006-10-18 Thread Bert Lamb
I did a quick write up on how I set my server up here: http://wiki.apache.org/ws/Tuscany/TuscanyJava/ContinuumForTuscany Hope this helps! -Bert On 10/18/06, Jeremy Boynes [EMAIL PROTECTED] wrote: That would be great - I have a remote machine we can use. I'm working on some itest mojo that

Re: Integration testing - we need some :-)

2006-10-18 Thread Jeremy Boynes
On Oct 18, 2006, at 9:20 AM, Bert Lamb wrote: I did a quick write up on how I set my server up here: http://wiki.apache.org/ws/Tuscany/TuscanyJava/ContinuumForTuscany Hope this helps! Thanks - I'll take a look and see if I can get it running on my server later today. -- Jeremy

Re: Integration testing - we need some :-)

2006-10-18 Thread Jeremy Boynes
On Oct 18, 2006, at 8:53 AM, Raymond Feng wrote: +1 from me. I can help as well. I already have a simple mortage approval application in my sandbox and I'll convert it to whatever integration test environment we come out when it's ready. What I'm messing with is a maven plugin that can

Re: Integration testing - we need some :-)

2006-10-18 Thread Jeremy Boynes
I manage to get most of this working but I am having problems with Surefire's classloader isolation. The StartMojo deploys the application and binds the CompositeContext to the thread. However, the SCA class is loaded from two different classloaders so the value returned from getContext()

Integration testing - we need some :-)

2006-10-17 Thread Jeremy Boynes
In cutting M2 the lack of integration testing has been quite painful. We need to start adding the post-build functional/integration testing we've talked about before but which we've not actually implemented. I'm going to start work on some projects in the sandbox to try and automate

Re: Integration testing - we need some :-)

2006-10-17 Thread Jim Marino
On Oct 17, 2006, at 12:53 PM, Jeremy Boynes wrote: In cutting M2 the lack of integration testing has been quite painful. We need to start adding the post-build functional/ integration testing we've talked about before but which we've not actually implemented. I'm going to start work on