Can someone force rebase and merge web-platform-tests/wpt #27182? It's a change to a README that somehow failed CI

Oh, maybe it's the lack of review that's blocking it... Can someone review?

annevk: blocked now community-tc.services.mozilla.com/tasks/dtv0SRa3TTytNniLHYaBSQ failing

and the UI there tells me:

> Can't be rescheduled past it's deadline of 2021-01-15T07:38:21.275Z.

MikeSmith: let's wait for jgraham to force it

annevk: OK, for now, did “retrigger” instead of “re-run“

but I can actually also force it if you want

if it’s blockin your

MikeSmith: it's fine

k

MikeSmith: it seems like a required check is still failing, admin merge?

hmm weird

now only the sink task is failing

but I thought that was a meta task that only fails when other tasks fail

just now manually retriggered that too

I have yet to bother to look what the actual concrete failure was

annevk: if you know the failure is false positive — or not caused by this PR — then I can admin merge it

OK , sink task still failing

but no error in the log as far as I can see

eExDSnpZMjl3WlhNaU9sc2lZWE56ZFcxbE9tZHBkR2gxWWkxMFpXRnRPbmN6WXk5emRHRm1aaUlzSW1GemMzVnRaVHBuYVhSb2RXSXRkR1ZoYlRwMllXeHBaR0YwYjNJdllXUnRhVzV6SWl3aVlYTnpkVzFsT21kcGRHaDFZaTEwWldGdE9tVnpOUzl2ZDI1bGNuTWlMQ0poYzNOMWJXVTZaMmwwYUhWaUxYUmxZVzA2WldOdFlYTmpjbWx3ZEM5dmQyNWxjbk1pTENKaGMzTjFiV1U2WjJsMGFIVmlMWFJsWVcwNlpYTTJMMjkzYm1WeWN5SXNJbUZ6YzNWdFpUcG5hWFJvZFdJdGRHVmhiVHAzTTJNdmFIUnRiQzEzWnlJc0ltRnpjM1Z0WlRwbmFYUm

9kV0l0ZEdWaGJUcG9kRzFzTld4cFlpOTBaWE4wTFcxcGJtbHZibk1pTENKaGMzTjFiV1U2WjJsMGFIVmlMWFJsWVcwNlpYTXRjM0JsWXk5dmQyNWxjbk1pTENKaGMzTjFiV1U2WjJsMGFIVmlMWFJsWVcwNlpYTTNMMjkzYm1WeWN5SXNJbUZ6YzNWdFpUcG5hWFJvZFdJdGRHVmhiVHBsY3pndmIzZHVaWEp6SWl3aVlYTnpkVzFsT21kcGRHaDFZaTEwWldGdE9uY3pZM1JoWnk5MFlXY3RhWEp5WldkMWJHRnljeUlzSW1GemMzVnRaVHBuYVhSb2RXSXRkR1ZoYlRwM00yTXZkMlZpTFhCc1lYUm1iM0p0TFhSbGMzUnBibWN0WVc1a0xYUnZiMnh6

TFdGa2JXbHVjeUlzSW1GemMzVnRaVHBuYVhSb2RXSXRkR1ZoYlRwM00yTXZZMmhoY25SbGNpMWtjbUZtZEhNdGNIVnphR1Z5Y3lJc0ltRnpjM1Z0WlRwbmFYUm9kV0l0ZEdWaGJUcDNNMk12WVRFeGVYTnNZV05yWlhJdFlXUnRhVzV6SWl3aVlYTnpkVzFsT21kcGRHaDFZaTEwWldGdE9uY3pZeTl6ZEhKaGRHVm5lU0lzSW1GemMzVnRaVHBuYVhSb2RXSXRkR1ZoYlRwM00yTXZZWEpqYUdsMFpXTjBkWEpsSWl3aVlYTnpkVzFsT21kcGRHaDFZaTEwWldGdE9uY3pZeTl6WlhKMmFXTmxMWGR2Y210bGNuTXRkMmNpTENKaGMzTjFiV1U2WjJsMG

FIVmlMWFJsWVcwNmQyVmlMWEJzWVhSbWIzSnRMWFJsYzNSekwzSmxkbWxsZDJWeWN5SXNJbUZ6YzNWdFpUcG5hWFJvZFdJdGRHVmhiVHAzWldJdGNHeGhkR1p2Y20wdGRHVnpkSE12WVdSdGFXNXpJaXdpWVhOemRXMWxPbWRwZEdoMVlpMTBaV0Z0T25aaGJHbGtZWFJ2Y2k5bmNuVnVkQzFvZEcxc0lpd2lZWE56ZFcxbE9tZHBkR2gxWWkxMFpXRnRPblpoYkdsa1lYUnZjaTluZFd4d0xXaDBiV3dpTENKaGMzTjFiV1U2WjJsMGFIVmlMWFJsWVcwNmQyVmlMWEJzWVhSbWIzSnRMWFJsYzNSekwzZHdkQzFqYjNKbExYUmxZVzBpTENKaGMzTjFi

V1U2WjJsMGFIVmlMWFJsWVcwNmR6TmpMMmgwYld3dGQyY3RZMmhoYVhKeklpd2lZWE56ZFcxbE9tZHBkR2gxWWkxMFpXRnRPbmN6WXk5M00yTXRaM0p2ZFhBdE1UQTJMVzFsYldKbGNuTWlMQ0poYzNOMWJXVTZaMmwwYUhWaUxYUmxZVzA2ZHpOakwzY3pZeTFuY205MWNDMHhNREF0YldWdFltVnljeUlzSW1GemMzVnRaVHBuYVhSb2RXSXRkR1ZoYlRwM00yTXZkek5qTFdkeWIzVndMVEV4TlRVeU1DMXRaVzFpWlhKeklpd2lZWE56ZFcxbE9tZHBkR2gxWWkxMFpXRnRPbmN6WXk5M00yTXRaM0p2ZFhBdE1URTFOVEl3TFdOb1lXbHljeUlzSW

1GemMzVnRaVHBuYVhSb2RXSXRkR1ZoYlRwM00yTXZkek5qTFdkeWIzVndMVEV3TVRFNU5pMXRaVzFpWlhKeklpd2lZWE56ZFcxbE9tZHBkR2gxWWkxMFpXRnRPbmN6WXk5M00yTXRaM0p2ZFhBdE1URTBPVEk1TFcxbGJXSmxjbk1pTENKaGMzTjFiV1U2WjJsMGFIVmlMVzl5WnkxaFpHMXBianAzTTJNaUxDSmhjM04xYldVNloybDBhSFZpTFc5eVp5MWhaRzFwYmpwMllXeHBaR0YwYjNJaUxDSmhjM04xYldVNloybDBhSFZpTFc5eVp5MWhaRzFwYmpwbFkyMWhjMk55YVhCMElpd2lZWE56ZFcxbE9tZHBkR2gxWWkxdmNtY3RZV1J0YVc0NlpY

TTFJaXdpWVhOemRXMWxPbWRwZEdoMVlpMXZjbWN0WVdSdGFXNDZaWE0ySWl3aVlYTnpkVzFsT21kcGRHaDFZaTF2Y21jdFlXUnRhVzQ2WlhNdGMzQmxZeUlzSW1GemMzVnRaVHBuYVhSb2RXSXRiM0puTFdGa2JXbHVPbVZ6TnlJc0ltRnpjM1Z0WlRwbmFYUm9kV0l0YjNKbkxXRmtiV2x1T21Wek9DSXNJbUZ6YzNWdFpUcG5hWFJvZFdJdGIzSm5MV0ZrYldsdU9uZGxZaTF3YkdGMFptOXliUzEwWlhOMGN5SXNJbUZ6YzNWdFpUcHNiMmRwYmkxcFpHVnVkR2wwZVRwbmFYUm9kV0l2TVRrME9UZzBmSE5wWkdWemFHOTNZbUZ5YTJWeUlsMHNJbk

4wWVhKMElqb3hOakV3TmprM09URXpPVGM1TENKbGVIQnBjbmtpT2pFMk1UQTJPVGszTVRNNU56a3NJbk5sWldRaU9pSjVYMWczT0hOc1dWSjJaV1U1WlhkUVRVc3hlR0pSYkhvd2VXUXdNbWRVTUhWeFpHZzBTV3BaWm5rMVp5SXNJbk5wWjI1aGRIVnlaU0k2SW1KeVZtRk1ObTh4TUV4MlZYVm1hbHBIZGxCYVFtMVdjMFk0YmxkWVNIUm1WM05FTDJKTFEyUllWVGc5SWl3aWFYTnpkV1Z5SWpvaWMzUmhkR2xqTDNSaGMydGpiSFZ6ZEdWeUwzZGxZaTF6WlhKMlpYSWlmWDA9

oofs

dang

sorry about that

good lord

MikeSmith: I mean, it's a change to a README file

yup

sorry for all the noise

(famous last words, I'm sure)

hahaha

too late now!

button pushed

Mission Accomplished

time for me to go pick up the kids from daycare anyway

so I gotta drop off now

[Patch Landed Meme]

\o/

MikeSmith: The sync task doesn't get rerun if you retrigger a task

And idk how retriggering that one interacts with everything else

jgraham: We might want to figure out how to handle task retriggers at some point (which may in itself be a taskcluster feature request).

Right now I *think* its still the case you have to close/open a PR and thus re-run all jobs to properly re-run the CI

smcgruer_[EST]: So rerun works if the task didn't expire. You may also need to rerun the sink task if it already ran.

We could perhaps figure out how to identify retriggers in the sync task, so you could retrigger a task and retrigger the sync task and update the status

(generically the problem is that it's unclear if a retrigger should also retrigger all dependent tasks, or what exactly should happen)

Oh, I think I broke Edge runs on Azure with the --py3 change. Remember when I said "oh, we can get away with always updating hosts via --py2" ? :D

Heh

Does hoppipolla.co.uk/410/asserts.png look potentially like the start of something useful?

jgraham: I vaguely recall requests for that, but I'm having a hard time coming up with the use case from scratch 🙂

Well the use case is knowing which things executed. The web audio tests have this whole complex framework that they are using to get one test per assert to see what actual comparisons were made which totally breaks everything because they end up putting audio samples values in the test name, and those are not stable.

I think I've heard similar requests from smaug

oh, is this about the thing which makes mochitests easier to follow?

yes

jgraham: looks very useful

Oh hi, didn't know you were in this channel :)

I must have bothered you also here :)

Cool. Feedback on specifics welcome. I might also have a go at adding some kind of debug mode that adds console logs for various things.

mochitest is even missing that line number, I think

so what you have now is already better

I believe Apple folks have also mentioned concerns that such a view might resolve

jgraham: That looks awesome!