1554225 - Push broken on nightly

Status: VERIFIED FIXED

(Core :: DOM: Push Subscriptions, defect)

Description

[Edouard Oger [:eoger]]

I've noticed I was not receiving any push notifications for Send Tab on Desktop Nightly.
STR: On a recent Nightly check for the dom.push.connection.enabled pref, it's false whereas it should be true by default.
I ran mozregression which gave me:

11:12.34 INFO: Last good revision: 55c360c47f31698fe123b0d2838d07d390048e0a
11:12.34 INFO: First bad revision: 1efcba0d56db1d3e12f0fecf3211af6d12a3a646
11:12.34 INFO: Pushlog:
https://hg.mozilla.org/integration/autoland/pushloghtml?fromchange=55c360c47f31698fe123b0d2838d07d390048e0a&tochange=1efcba0d56db1d3e12f0fecf3211af6d12a3a646

I have no idea why testing/ prefs are affecting Nightly though.

Comment 1

[Lina Butler [:lina]]

dom.push.connection.enabled is also false by default for me in Nightly. Marco, this might be what you were seeing, too.

Comment 2

[Henrik Skupin [:whimboo][⌚️UTC+2]]

I wonder why this didn't cause any perma fail. I assume we don't have any kind of test in tree which verifies that this preference should be set to true by default?

The code in Marionette which sets this is preference to false only gets run when Marionette is actually enabled. And by default this is not the case.

It's late here and I won't have the time to check that myself until Monday.

Comment 3

[Lina Butler [:lina]]

(In reply to Henrik Skupin (:whimboo) [⌚️UTC+2] from comment #2)

I wonder why this didn't cause any perma fail. I assume we don't have any kind of test in tree which verifies that this preference should be set to true by default?

We don't, all the push tests set it to true before running. 😞 Not doing that for any prefs that we expect to be set by default is a really good idea, though.

Comment 4

[Lina Butler [:lina]]

(In reply to Henrik Skupin (:whimboo) [⌚️UTC+2] from comment #2)

It's late here and I won't have the time to check that myself until Monday.

Thanks for looking into it!

Comment 5

[Marco Castelluccio [:marco]]

I got the same regression range.

I think marionette.js prefs are actually loaded in all cases when ENABLE_MARIONETTE.
https://searchfox.org/mozilla-central/rev/4606c7974a68cab416c038acaedcae49eed93822/browser/installer/package-manifest.in#200
https://searchfox.org/mozilla-central/rev/4606c7974a68cab416c038acaedcae49eed93822/modules/libpref/Preferences.cpp#4412

Comment 6

[Henrik Skupin [:whimboo][⌚️UTC+2]]

This is clearly a bustage as caused by bug 1542244. The following file should never have been changed:
https://searchfox.org/mozilla-central/source/testing/marionette/prefs/marionette.js

I will request a backout.

Comment 7

[Henrik Skupin [:whimboo][⌚️UTC+2]]

Fixed by the backout on bug 1542244:
https://hg.mozilla.org/mozilla-central/rev/944c410b7e9185a0cb90a4fbc0970299f1ff3e2b

Comment 8

[Emma Humphries ☕️🎸🧞‍♀️✨ (she/they) [:emceeaich] (Pacific Time) use needinfo]

This bug has been identified as part of a pilot on determining root causes of blocking and dot release drivers.

It needs a root-cause set for it. Please see the list at https://docs.google.com/document/d/1FFEGsmoU8T0N8R9kk-MXWptOPtXXXRRIe4vQo3_HgMw/.

Add the root cause as a whiteboard tag in the form [rca - <cause> ] and remove the rca-needed keyword.

If you have questions, please contact :tmaity.