Android dependency inferred incorrectly?



Hi All,

Not sure if this IS a bug, but we have not changed our circle.yml file for a while and suddenly, a new dependency is being inferred somehow?

This seems to be an issue for gradle compileTestJava

On the last successful build, the output was:

export TERM="dumb"
if [ -e ./gradlew ]; then ./gradlew dependencies;else gradle dependencies;fi
Root project
No configurations
Total time: 22.759 secs

Currently it does:

FAILURE: Build failed with an exception.

* What went wrong:
Task 'compileTestJava' not found in root project 'customer-android'.


Chinese character seems to throw off our circleci build

We are experiencing the same issue. This started happening yesterday.


Just noticed it an hour ago as well.


Same problem here


Alright well just now it seemed to pass… I’ll post again if it fails again somewhere


Its on and off for me too,
2 build failed one succeeded


Some of us have been able to rebuild successfully, but many are still failing. :disappointed:


Yep, having the same issue. Rebuilt 3 times but haven’t been able to get it to succeed. My Builds were fine yesterday, haven’t made any changes to my circle.yml since then.


same issue here. My circle.yml overrides the test task:

And nothing for dependencies but it is still running compileTestJava .

This happened when I did re-run a successful build but now keeps happening for all.

update: I could workaround the issue by going over settings and adding a test dependency ./gradlew tasks to override the default behavior.


Found the same issues on my project, no changes in the config since yesterday and even older builds are failing.


Hi guys, I’m looking at this now! Sorry for the trouble.


same problem for me. nothing changed in the yml file.


I found out that this is because of our change. I’m now reverting the change and will be deployed to the whole infrastructure in 1 or 2h or so. Sorry about that.


hey @kimh any updates on the deployment. This issue still exists for our Android builds.


I just checked and still have the issue, not fixed yet.


Sorry for the delay. The fix has been deployed to more than half of all machines now. So, if you tried now it will be likely succeed. Can you trie a few times?


I’ve shipped the fix to all machines now. Can someone try if this is fixed?


I just tried now. Issue still exists. Will keep retrying.


@kimh thanks, the change seems to be working for us now.


Thanks @kimh it is working for me now, thank you