Is there some configuration that says: “once any of the jobs in a build fail, cancel all of them”?
I have a setup where 3 jobs launch at the same time. 1 job creates a shared payload that the following two jobs build into an ios and android app, respectively. I’d use the Travis stages feature, but the bootstrapping process for the 2 jobs is so slow that I wanted them to start running / bootstrapping immediately, and once they’ve done that, they poll S3 in a loop until the zipped payload has appeared. This works great, but there’s currently no way for them to detect that the first build has failed; it would be very handy if there were some config that could detect that the build was ultimately going to fail, and just stop all jobs right then and there.
Alternatively, I think I could probably get something working where I just uploaded a non-zip file to the expected location, which would cause the following jobs to eagerly fail, but that seems hacky.