I haven't seen any significant differences between test results for android gradle vs non-gradle builds. Do we need to keep running the non-gradle tests? If so, could we reduce the set of tests? If they are valuable, they can certainly continue to run. I'm just challenging the status quo here in hopes of increasing efficiency.

(In reply to Geoff Brown [:gbrown] from comment #0) > I haven't seen any significant differences between test results for android > gradle vs non-gradle builds. Do we need to keep running the non-gradle > tests? If so, could we reduce the set of tests? > > If they are valuable, they can certainly continue to run. I'm just > challenging the status quo here in hopes of increasing efficiency. I'd like to keep something for non-Gradle to smoketest, but just Robocop would achieve that. None of the M, R, X are adding much value. I've been thinking of https://bugzilla.mozilla.org/show_bug.cgi?id=1414415 again... I should do that. But for now, just a smoketest, please!

Sure, let's go with just Robocop for now then. Will put together a patch tomorrow.

Pushed by gbrown@mozilla.com: https://hg.mozilla.org/integration/mozilla-inbound/rev/1bda868b03a3 Stop running most Android non-gradle tests; r=jmaher

Gradle tests are still considered Tier 2, FWIW.

Oh nevermind, the tests we label "Gradle" are really non-Gradle ones. That's not at all misleading.

Right, it's the tier 2 tests that are being reduced here.