It turns out that *print-fn* and *print-err-fn* are documented as accepting strings as arguments. With :checked-arrays we are now dumping an exception to *print-err-fn*.

@ambrosebs thanks, yes an incremental approach would be welcome

@dnolen You commented on old version of the post so some of the edits don't match with the current version

Anyway, I updated the parts that are present in current version. I'll be offline rest of the day, but if you want to publish the post, feel free to edit it.

@juhoteperi k thanks!

@dnolen I made a big list of proposed changes in CLJS-1461 that I'm planning to work through. First small bit of work is ready to be merged https://dev.clojure.org/jira/browse/CLJS-2260

@dnolen I just finished the first iteration on the Canary project, currently there are only two projects participating (planck and cljs-devtools) it generates reports like this: https://github.com/cljs-oss/canary/tree/results/reports/2017/07/18/job-000063-1.9.818-ea70c3f I think we are ready to onboard more projects, tried to document it a bit in the master readme: https://github.com/cljs-oss/canary/tree/master @mfikes @juhoteperi ^^

Very cool!

@mfikes btw. I finally figured out how to run only one job from planck matrix: https://github.com/cljs-oss/canary/blob/master/runner/src/canary/projects/mfikes.clj#L5-L8

Excellent !

also I noticed that your jobs are somehow rate limited or something, it took 40+ minutes for last two jobs to start (when triggering full matrix), don’t know what was going on

running just a single job usually finishes in 10mins or less

@darwin Jobs that involve Mac boxes can take a long time to spin up. I think Travis simply has a long queue relative to very little Mac hardware.

ah ok, that makes sense then