1369700 - Delay initialization of marionette.js until delayed startup and startup tests are finished

Status: RESOLVED FIXED

(Remote Protocol :: Marionette, enhancement, P1)

Description

[Florian Quèze [:florian] (on leave until October 9)]

Currently marionette.js starts its initialization right after receiving the "sessionstore-windows-restored" notification.

This is unfortunate for the browser_startup.js browser chrome mochitest I added in bug 1358798, because it causes lots of marionette file to appear to be loaded before the browser is ready to handle user events.

Comment 1

[Florian Quèze [:florian] (on leave until October 9)]

Attachment: Patch

Tree seems green-ish: https://treeherder.mozilla.org/#/jobs?repo=try&revision=b0fa95d744f7ecc52f7d105b4d37ecc2a235010c

Comment 2

[Andreas Tolfsen ❲:ato❳]

Comment on attachment 8873807 [details] [diff] [review]
Patch

Review of attachment 8873807 [details] [diff] [review]:
-----------------------------------------------------------------

r+wc: Please expand commit message with information about the motivation and reason behind this change.

Comment 3

[Andreas Tolfsen ❲:ato❳]

Comment on attachment 8873807 [details] [diff] [review]
Patch

Review of attachment 8873807 [details] [diff] [review]:
-----------------------------------------------------------------

r+wc: Please expand commit message with information about the motivation and reason behind this change.

Comment 4

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/mozilla-inbound/rev/a3f93870b233
Delay initialization of marionette.js until delayed startup and startup tests are finished, r=ato.

Comment 5

[Dão Gottwald [:dao]]

Any particular reason why you used idleDispatch rather than explicitly observing the delayed startup finished notification?

Comment 6

[Sebastian Hengst [:aryx] (needinfo me if it's about an intermittent or backout)]

https://hg.mozilla.org/mozilla-central/rev/a3f93870b233

Comment 7

[Florian Quèze [:florian] (on leave until October 9)]

(In reply to Dão Gottwald [::dao] from comment #5)
> Any particular reason why you used idleDispatch rather than explicitly
> observing the delayed startup finished notification?

There's a lot of stuff that observes the delayed-startup-notification, and cause jank at that point. We are really done with startup once the stuff started from that notification is finished, which means when the event queue is empty. That's what idleDispatch gives us.

Comment 8

[Florian Quèze [:florian] (on leave until October 9)]

(In reply to Florian Quèze [:florian] [:flo] from comment #7)
> (In reply to Dão Gottwald [::dao] from comment #5)
> > Any particular reason why you used idleDispatch rather than explicitly
> > observing the delayed startup finished notification?
> 
> There's a lot of stuff that observes the delayed-startup-notification,

I meant sessionstore-windows-restored.

Also, the code modified by this patch is already running off that notification.