[cmake-developers] Any idea why continuous tests are failing?

[Stephen Kelly]

David Cole wrote:

> It's only running on machines that are *driven* by a CMake that
> contains policy 19. You need to remove your if(POLICY check in
> CMake/Tests/CMakeLists.txt and then we need to figure out how to get
> the test to run with the built CMake instead of the driving CMake.
> 
> Step 1 should be for you to remove the if(POLICY check.

I'd rather not spend cycles and days trying to debug the dashboard at this 
point. It distracts from my goal of getting the topics into master, and it 
would block merging my topics into master.

I've made the requested change in the wip-target-interface branch of my 
clone though. I can push that patch at some point later if you think it's 
useful. 

I don't think the test is particularly useful anyway, as it doesn't test 
that the warning is not emitted when it should not be.

A better, more focussed test could probably be written for the goal of 
finding out how to get the test to run with the built CMake instead of the 
driving CMake. I don't think that issue needs to block other things.

Additionally, the reasons for the test failures can be debugged far more 
effectively by someone with access to the machines who can do interactive 
debugging (especially the failures from the nightlies today where the 
if(POLICY is used). 

 http://open.cdash.org/testDetails.php?test=167420134&build=2677125
 http://open.cdash.org/testDetails.php?test=167423128&build=2677205

Attempting to debug issues like that from afar without real access is 
something I very much want to avoid. I know from experience that it's no fun 
and takes too long :).


Thanks,

Steve.