1722462 - browser_device_change.js (and many other RDM tests) fail with server targets

Status: RESOLVED FIXED

(DevTools :: Responsive Design Mode, defect)

Description

[Alexandre Poirot [:ochameau]]

This fails because we are closing the RDM over here:
https://searchfox.org/mozilla-central/rev/072710086ddfe25aa2962c8399fefb2304e8193b/devtools/client/responsive/test/browser/browser_device_change.js#110-112
And expect it to reload the page. We then assert values which depends on reloading the page.
But we don't reload with server targets, because over there:
https://searchfox.org/mozilla-central/rev/072710086ddfe25aa2962c8399fefb2304e8193b/devtools/client/responsive/ui.js#252-259
this.responsiveFront is destroyed...

I imagine that during a target switch, when doing a reload this front may be destroyed during a short period of time.
We should probably change this check to something else.

Comment 1

[Alexandre Poirot [:ochameau]]

Attachment: Bug 1722462 - [devtools] Fix browser_device_change.js

This test expect RDM to reload the page on destroy,
but because of a race condition -or- because of a behavior change with target switching,
responsiveFront ends up being destroyed and we don't reload.

(And may be also browser_max_touchpoints.js, browser_page_redirection.js, browser_touch_event_should_bubble.js, /browser_touch_pointerevents.js, ...)

Comment 2

[Nicolas Chevobbe [:nchevobbe]]

Attachment: Bug 1722462 - [devtools] Wait for dom-complete in waitForViewportLoad. r=ochameau

Depends on D121059

Comment 3

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/c1b816e832bb
[devtools] Wait for dom-complete in waitForViewportLoad. r=ochameau

Comment 4

[Atila Butkovits]

https://hg.mozilla.org/mozilla-central/rev/c1b816e832bb