cljs-dev

ClojureScript compiler & std lib dev, https://clojurescript.org/community/dev
dnolen 2020-05-04T11:13:18.252100Z

ok fixed up the new blocker issues that cropped up over the week

dnolen 2020-05-04T11:15:55.252500Z

cutting 1.10.756 now

mfikes 2020-05-04T12:11:57.253200Z

Cool... https://clojure.atlassian.net/projects/CLJS/issues/CLJS-3245 is a cut-n-dry one we could apply in order to get unit tests running again

dnolen 2020-05-04T13:37:02.253500Z

fixed, building 1.10.758

dnolen 2020-05-04T14:57:37.254300Z

@alexmiller another vote for just hooking in @mfikes CI stuff into the ClojureScript main repo so I can see if a commit is failing test

dnolen 2020-05-04T14:57:56.254800Z

around releases this really like flying blind - manually running everything locally is not practical

dnolen 2020-05-04T14:58:01.255Z

it's too easy to forget something

dnolen 2020-05-04T14:58:22.255400Z

other option would be to do it Jenkins but this seems like a fairly large task for you

alexmiller 2020-05-04T15:05:14.255700Z

would be happy to learn more about what this would mean

alexmiller 2020-05-04T15:05:54.256Z

are you talking about tests in the clojurescript build itself or something external?

alexmiller 2020-05-04T15:19:16.256700Z

I can make a test job to do arbitrary bash whatever

mfikes 2020-05-04T15:21:13.258200Z

ClojureScript has TravisCI and AppVeyor "metadata" files in the root of its tree. For TravisCI, it is pretty much just "go in and turn it on", I think.

alexmiller 2020-05-04T15:21:47.258800Z

it sounds like you want "when there are commits on clojurescript, then run thing and report" and that's not hard structurally, but depends what "thing" requires

alexmiller 2020-05-04T15:21:54.259Z

can you point me more precisely?

alexmiller 2020-05-04T15:24:00.259300Z

oh, .travis.yml

mfikes 2020-05-04T15:24:03.259400Z

mfikes 2020-05-04T15:24:15.259900Z

Yeah, in Travis, minimally it really is just a switch ^

mfikes 2020-05-04T15:24:53.261200Z

Of course, you can mess around with more complicated settings down in there, but fundamentally, it just starts listening to GitHub commits and puts green checkmarks or red Xs in GitHub, and mails people.

alexmiller 2020-05-04T15:25:14.261600Z

the travis one seems like it could pretty trivially be just a bash script, which I could easily embed in a jenkins job

alexmiller 2020-05-04T15:25:52.262300Z

I think it can do that connection back to gh too, but I haven't done that part before

alexmiller 2020-05-04T15:26:47.262700Z

the appveyor one is windows?

alexmiller 2020-05-04T15:29:21.263100Z

looks like there are jenkins plugins that can actually run travis yml directly

alexmiller 2020-05-04T15:30:36.263400Z

hrm, looks out of date

dnolen 2020-05-04T15:38:37.264700Z

yes just all of the tests in ClojureScript itself would be great, nothing external

alexmiller 2020-05-04T15:39:14.265700Z

your last build failed btw, complaining about upload staging rules which seems very strange to me

dnolen 2020-05-04T15:39:39.266Z

"upload staging rules"?

dnolen 2020-05-04T15:39:56.266200Z

oh kicked it again

alexmiller 2020-05-04T15:40:26.266800Z

when you upload a package of artifacts to maven, it applie some staging rules to check that pom has certain attributes, stuff is signed, yada yada

alexmiller 2020-05-04T15:40:42.267200Z

afaik, nothing should have changed though in what you're making or what the rules are

alexmiller 2020-05-04T15:41:21.267500Z

well if it fails again, I'll look deeper

dnolen 2020-05-04T15:54:36.267900Z

@alexmiller hrm, failed again

alexmiller 2020-05-04T16:03:07.268300Z

the rule looks like it's failing because that version already exists

alexmiller 2020-05-04T16:03:28.268500Z

which it does

alexmiller 2020-05-04T16:04:00.268900Z

do you need to bump that or something?

dnolen 2020-05-04T16:06:14.269600Z

oh hrm, not sure what happened ok, then we're ok - the version number is automatically computed

alexmiller 2020-05-04T16:08:02.270600Z

looks like it was built and released at 6:15 am today (not sure if that's my time or eastern), then failed in a rebuild at 8:35 am, and just now

alexmiller 2020-05-04T16:09:58.270900Z

I don't know - I'd double-check that your last changes are in there

alexmiller 2020-05-04T16:10:41.271200Z

it's a little unclear to me what happened there

alexmiller 2020-05-04T16:16:39.271500Z

I'll look at the test stuff this afternoon

mfikes 2020-05-04T17:30:28.272500Z

Canary and (mfikes) CI are all green now (current master) https://github.com/cljs-oss/canary https://github.com/mfikes/clojurescript/commits/master

👍 1
dnolen 2020-05-04T21:48:11.273300Z

@alexmiller huh it appears something did go wrong, I wasn't hallucinating

dnolen 2020-05-04T21:48:12.273400Z

https://github.com/clojure/clojurescript/releases

dnolen 2020-05-04T21:48:30.273800Z

758 never appeared under GitHub releases

dnolen 2020-05-04T21:48:53.274200Z

so it seems weird that Jenkins said it happened but the tag was never created

alexmiller 2020-05-04T21:52:49.275100Z

looking at it again it looks like the part that waits for central to release timed out and so the build job failed and wasn't tagged

alexmiller 2020-05-04T21:53:07.275600Z

but I bet the release did actually happen to maven, just after the timeout

alexmiller 2020-05-04T21:53:40.276Z

and then the next build failed because the release was already theree

alexmiller 2020-05-04T21:53:55.276400Z

so I would touch something so the rev is different and rebuild

alexmiller 2020-05-04T21:54:38.276700Z

I missed the timeout message when I looked before

dnolen 2020-05-04T21:55:52.277Z

@alexmiller or I could just create the tag myself no?

dnolen 2020-05-04T21:56:54.277400Z

(to go avoid going through all the steps when they're really not needed)

alexmiller 2020-05-04T21:58:35.277600Z

that too

dnolen 2020-05-04T22:06:11.277800Z

k

dnolen 2020-05-04T23:46:59.278300Z

@alexmiller website could use a bump whenever you get a chance