rule_scala is broken on ci.bazel.io about rules_scala

[colinmarc]

Look like the issue is that it's running bazel build //... instead of test_run.sh. There are some build files which we expect to fail in there.

[damienmg]

We generally recommend doing integration test that set-up a workspace and call bazel in it so we can actually runs them from Bazel.

ci.bazel.io does build //... and test //..., we can change those pattern if needed.

Also @johnynek if you accept the invite to the bazelbuild org and make the membership public you can trigger the test on ci.bazel.io with "test this please"

[johnynek]

@damienmg where can I find that invitation? I have looked several times and see no notification on github for this, and I can't find a good search string:

https://www.dropbox.com/s/n0yu9qsnfqdls9e/Screenshot%202016-06-09%2005.40.52.png?dl=0

[damienmg]

I just reinvited you, I hope it work this time.

On Thu, Jun 9, 2016 at 5:40 PM P. Oscar Boykin [email protected]
wrote:

@damienmg https://github.com/damienmg where can I find that invitation?
I have looked several times and see no notification on github for this, and
I can't find a good search string:

https://www.dropbox.com/s/n0yu9qsnfqdls9e/Screenshot%202016-06-09%2005.40.52.png?dl=0

—
You are receiving this because you were mentioned.

Reply to this email directly, view it on GitHub
#73 (comment),
or mute the thread
https://github.com/notifications/unsubscribe/ADjHf2bS1uIvRUFPWcq02r13hEAOcOPHks5qKDQGgaJpZM4Ixtv9
.

[johnynek]

@damienmg got it.

[damienmg]

Ping? What should be done here?

[johnynek]

Well. In the short term you could change the test to manually exclude the
failing directories, but as my birthday was Thursday I have taken a few
days off. I can refactor this week to avoid the test script and call Bazel
recursively (I think is your recommendation) so that we set up failing
examples in sh rules I guess?

On Friday, June 10, 2016, Damien Martin-Guillerez [email protected]
wrote:

Ping? What should be done here?

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
#73 (comment),
or mute the thread
https://github.com/notifications/unsubscribe/AAEJdiaTh7Rb2nje04e3lhCY7mJIxQ2Eks5qKVAigaJpZM4Ixtv9
.

P. Oscar Boykin, Ph.D. | http://twitter.com/posco | http://pobox.com/~boykin

[damienmg]

Plan SGTM. Yes having it has a bazel test is better.

We can just mark the breaking test as "noci" then, I'll send a PR for that.

[damienmg]

Send #75 for this.