MessageImpl changed to set the header part if a name is null

[ibek]

This change is needed for the SWITCHYARD-703.

[jeffyu]

Hi Ivo,

Thanks a lot for your contribution.

It would be great if you rebase the code when you make the pull request next time. :). This time, as your updated branch is not based on the latest code, I couldn't merge it. But I reviewed the code, it is a small change, therefore I used the github built-in function of merging pull request.

Regards
Jeff

[jeffyu]

Hi Ivo,

Sorry for re-opening this issue, however, on applying your patch to the RiftSaw, found that this fix seem to break the unit tests of the RiftSaw. Therefore re-open it. :-(

If you can rebase the riftsaw-ode codebase, and then build it with this update, when you run the build for riftsaw, you would see the failed test cases.

Could you please take a look at it to see if they are any updates that might affect your update, or your update might not work as you thought with latest code?

Thanks
Jeff

[ibek]

Hi Jeff

I will investigate it. It's weird because there is almost nothing to be broken. :)

[jeffyu]

I was thinking of the same, so must be some minor thing there.

Just to let you know, I also revert your update from the repo for now.

Regards
Jeff

[ibek]

So, I pulled from upstream and I successfully built it without any failures or errors in unit tests. Maybe something bad happened with merging. However if you have a moment you can try it here first https://github.com/ibek/riftsaw-ode/tree/SWITCHYARD-703

Should I create new pull request?

[jeffyu]

Hi Ivo,

Just to quick check. I am saying that the fix would break the RiftSaw project, not the RiftSaw-ODE project itself.

Regards
Jeff

[ibek]

Hi Jeff,

thank you, I thought the tests in riftsaw-ode, don't know why ... it was friday :)
finally I reproduced it with current riftsaw master branch... do you think "testSimpleCorrelationVersioned(org.riftsaw.engine.BPELEngineTest): Failed: org.apache.ode.bpel.iapi.ContextException: Deploy failed; Deployment Unit "simple_correlation_HelloGoodbye-1" already deployed!"?

After I change the branch to SWITCHYARD-703 in RiftSaw I don't have this issue. It worked even for master but in the version before 4 months https://github.com/ibek/riftsaw . So, I would try to merge pull request for riftsaw-ode and simultaneously pull request for riftsaw.

Thanks, Ivo

[jeffyu]

Hi Ivo,

Yes, I got the same error, but the root cause for this 'already deployed' was because the previous error. If you look into the failed test case, that would be great.

Regards
Jeff

[ibek]

I slightly edited the BPELEngineTest to undeploy processes properly but I'm not sure if it will help because I had the exception only once and I cannot reproduce it :(
The changes are here ibek/riftsaw@db7392f

[jeffyu]

Thanks Ivo for the update.

Was curious why it was causing the undeployment throwing the error. Your fix seem to catch the error and then ignore it, please let me know if I misunderstood something here. I'll take a look at it tomorrow and then get back to you.

Regards
Jeff

[ibek]

Hi Jeff,

I think that a process should be undeployed even if a test fails because otherwise it causes failures somewhere else. But yes, I probably ignore the error.

BTW, I found another way to put pull request into upstream ... here (the Patch and Apply - https://help.github.com/articles/using-pull-requests) you can find how to use patch instead of merge and the commits don't have to be based on latest code ;)

Regards, Ivo

[jeffyu]

Hi Ivo,

Yeah, your point makes sense to me, I'll just not ignore the error thrown by undeployment procedure though.

Thanks for the Patch and Apply link, didn't went that route yet. :-)

Sorry for the late response, was kept distracted by other things these two days.

Regards
Jeff