Initialize UsesOriginAgentClusters when restoring about:blank with content principal from session history
Categories
(Core :: DOM: Navigation, defect, P3)
Tracking
()
| Tracking | Status | |
|---|---|---|
| firefox-esr115 | --- | unaffected |
| firefox-esr140 | --- | unaffected |
| firefox146 | --- | unaffected |
| firefox147 | --- | unaffected |
| firefox148 | --- | fix-optional |
People
(Reporter: vhilla, Unassigned)
References
(Regression)
Details
(Keywords: regression)
See bug 2004165, esp comment 10.
If an initial about:blank with content principal is restored from session history, EnsureUsesOriginAgentClusterInitialized might fail in the synchronous load path. As preliminary fix, we skip the synchronous path in DoURILoad in that case. We should look into fixing this properly.
| Reporter | ||
Updated•11 months ago
|
| Reporter | ||
Comment 1•11 months ago
|
||
We'll have to see how well the fix works. I'll mark this as S4 for now, hoping defaulting back to the async load happens rarely and causes not much problems.
Comment 2•11 months ago
|
||
Set release status flags based on info from the regressing bug 543435
Comment 3•10 months ago
|
||
Next week is the final week of nightly for Fx148.
:vhilla, what is the current status of the investigation/priority here?
| Reporter | ||
Comment 4•10 months ago
|
||
It's low priority and doesn't need to block 148.
I'm not sure if this bug can impact users at all. If an about:blank is restored with a content principal for which UsesOriginAgentClusters isn't initialized, we'll do an async load. For this sync vs async behavior to have an impact, some page needs to observe the load. But
- We've only seen this with top-level
about:blankand I don't think session restore would restore the window proxy or so - The observing page would need to be same origin, but then that page load should've initialized OAC
Updated•10 months ago
|
| Reporter | ||
Comment 5•3 months ago
|
||
Bug 2055554 might mean this workaround in nsDocShell might not be needed anymore.
We could still keep this bug open because the general issue that OAC might not be initialized when we load initial about:blank still stands.
Description
•