Don't leave builds broken!

-10 points to some combination of @raff, @gutkowski, and @adamg for leaving the Reference Application build broken since 11 hours ago.

https://ci.openmrs.org/browse/REFAPP-OMODDISTRO-5019

If you break a build, and you can’t quickly fix it (perhaps by reverting your change) than please send a message to the #dev category so people know what the situation is, and whether/how they should act.

2 Likes

We need a discourse karma plugin…

that would be lovely.

The tests are fixed, but build is still failing due to other issue.

You can expect more test failures as we are trying to improve UI tests and it takes many trials to get them right. We’ll try not to leave the build red for more than 8 hours.

Thanks for the update!

I understand that tests will break while you work on this. As long as you’re either (a) on IRC, and able to answer a “should this be broken”?, or (b) have left a message on talk saying “we know they’re broken overnight”, that should be fine.

I usually check CI a few minutes before scrum because the scrum master @bholagabbar asks me why the builds are failing. So if we can ensure that the scrum master knows by scrum time, that would be perfect. :slight_smile:

@r0bby is any of your team handling this? https://ci.openmrs.org/browse/REFAPP-OMODDISTRO-5027/log

Yes, I will handle it – likely in a couple hrs now that I am awake – it will be fixed before 9PM tonight. I apologize to everyone – I stayed awake far longer than I wanted to ensure this didn’t happen…missed one :frowning:

Fixed. :ship: it.

@r0bby that was fast! Thanks man! :smile:

We manually generate the known_hosts file for the bamboo agents :slight_smile: I thought i got all the servers but missed that one.

Full build running: https://ci.openmrs.org/browse/REFAPP-OMODDISTRO-5028

:green_heart:

@dkayiwa, @preethi_s, the EMR API build is broken after merging the latest PR: https://ci.openmrs.org/browse/EMRAPI-EMRAPI-1778

What’s up?

@darius @dkayiwa I have added MRS2.0 profile in the above commit to run tests against openmrs 2 which essentially uses BaseModuleContextSensitiveTest that is compiled in java 8. Thats why when running tests in jdk 7, it throws java.lang.UnsupportedClassVersionError.

It passes in travis ci because it uses jdk8.

Have we faced this kind of problem before, since there are many modules which run tests against mrs 1.x and 2.x? If so, how was it solved?