Make RemoteWebProgress event handlers work with out-of-process subframes
Categories
(Core :: DOM: Content Processes, task, P2)
Tracking
()
| Fission Milestone | M4.1 |
People
(Reporter: u608768, Assigned: u608768)
Details
Bug 1510569 (specifically D30252) introduced a mShouldSendWebProgressEventsToParent flag to determine if the BrowserChild is ready to send events to its BrowserParent.
This was required because the RemoteWebProgress rewrite resulted in things being ready earlier than expected, and so we started firing events (specifically onStateChanges) for the initial about:blank load, which broke tests.
The problem now is that we flip this flag in the browser-child.js framescript, which isn't ever loaded for out-of-process iframes.
One possible solution is to just flip this flag at some point after we've created (and hopefully loaded?) the initial about:blank page: https://searchfox.org/mozilla-central/rev/e5327b05c822cdac24e233afa37d72c0552dbbaf/dom/ipc/BrowserChild.cpp#514-515.
|
||
Updated•5 years ago
|
|
||
Comment 1•5 years ago
|
||
Roll some unfixed bugs from Fission Milestone M4 to M5
0ee3c76a-bc79-4eb2-8d12-05dc0b68e732
|
|
||
Comment 2•5 years ago
|
||
Kashav, did you fix this bug as part of your fix for BrowserTestUtils.browserLoaded bug 1578465? This bug blocks bug 1578465.
(In reply to Chris Peterson [:cpeterson] from comment #2)
Kashav, did you fix this bug as part of your fix for BrowserTestUtils.browserLoaded bug 1578465? This bug blocks bug 1578465.
We ended up going a different route for bug 1578465. I may be mistaken, but I think this might still be required for some remaining mochitests.
This isn't required anymore.
Description
•