Closed Bug 1489257 Opened 6 years ago Closed 6 years ago

Unable to log into any google services with GeckoView

Categories

(GeckoView :: General, defect, P1)

Unspecified
Android
defect

Tracking

(geckoview62blocking fixed, firefox-esr60 unaffected, firefox62 unaffected, firefox63+ fixed, firefox64 fixed)

RESOLVED FIXED
mozilla64
Tracking Status
geckoview62 blocking fixed
firefox-esr60 --- unaffected
firefox62 --- unaffected
firefox63 + fixed
firefox64 --- fixed

People

(Reporter: rbarker, Assigned: droeh)

References

Details

(Whiteboard: [geckoview:fxr:p1])

Attachments

(1 file)

Trying to log into any google service such as gmail or youtube fails with recent builds of GeckoView.
Whiteboard: [geckoview:fxr:p1]
Is this a recent regression in GV 63 or 64?
OS: Unspecified → Android
Priority: -- → P1
This is fallout from bug 1441059; if you comment out the load delegate checking in nsDocShell::InternalLoad[0] (and aborting the InternalLoad call, which is probably the actual issue), you can log in to google services just fine -- I've tested gmail and youtube in GVE. [0]: https://searchfox.org/mozilla-central/source/docshell/base/nsDocShell.cpp#9569
Do we just want to back that out, then? I think we need to consider a different approach.
Dylan, should we just back out bug 1441059? This fallout seems worse than the problem 1441059 fixed. :)
Flags: needinfo?(droeh)
All things considered I'd say it's best to back out bug 1441059 and bug 1478171 from the 62 relbranch; I can probably provide a stopgap patch for 1478171 that doesn't rely on 1441059 if needed.
Flags: needinfo?(droeh)
Ryan is comment 5 something you would handle?
Flags: needinfo?(ryanvm)
Depends on whether Dylan wants to include his stopgap for bug 1478171 in it or not.
Flags: needinfo?(ryanvm) → needinfo?(droeh)
Dylan currently driving here.
Assignee: nobody → droeh
(In reply to Ryan VanderMeulen [:RyanVM] from comment #7) > Depends on whether Dylan wants to include his stopgap for bug 1478171 in it > or not. I don't think we need to include the stopgap, I can land/uplift it later if necessary. Talking this over on Slack, it seems like to meet everyone's release needs the best thing to do at the moment is to back these patches (for bug 1441059 and bug 1478171) out of: 64, 63, and 62 relbranch. If you could handle it, that would be greatly appreciated.
Flags: needinfo?(droeh)
I'll do the relbranch backout, but 63/64 are going to need more manual fixing up than I can help with. https://hg.mozilla.org/releases/mozilla-release/rev/4cdeecd31350b13f8741202dc06d87d8bc3d84db
Flags: needinfo?(droeh)
(In reply to Ryan VanderMeulen [:RyanVM] from comment #10) > I'll do the relbranch backout, but 63/64 are going to need more manual > fixing up than I can help with. > > https://hg.mozilla.org/releases/mozilla-release/rev/ > 4cdeecd31350b13f8741202dc06d87d8bc3d84db Alright, I'll take care of them. Thanks!
Flags: needinfo?(droeh)
Jim, it looks to me like we're failing (specifically, hanging in waitForPageStop or waitForInitialLoad) on tests where the NavigationDelegate is null, but I can't see any reason why backing out the async LoadURIDelegate and redirect patches would cause this. If you've got a moment, could you take a look and see if anything comes to mind? Thanks.
Flags: needinfo?(droeh) → needinfo?(nchen)
You should back out bug 1478171 as well. Otherwise, I bet bug 1480095 is interfering with your backout here. All the test failures here immediately follow tests that were changed by bug 1480095.
Flags: needinfo?(nchen)
That should be bug 1486525 in the first sentence.
(In reply to Jim Chen [:jchen] [:darchons] from comment #16) > That should be bug 1486525 in the first sentence. I was intending to just @Ignore that test for the time being, as we can still support that with sync LoadURIDelegate pretty easily. Thanks for your help!
Status: NEW → RESOLVED
Closed: 6 years ago
Resolution: --- → FIXED
Target Milestone: --- → Firefox 64
Comment on attachment 9008438 [details] [diff] [review] Back out async LoadURIDelegate Approval Request Comment [Feature/Bug causing the regression]: 1441059 [User impact if declined]: Google Services logins will not work in GV [Is this code covered by automated tests?]: No [Has the fix been verified in Nightly?]: No [Needs manual test from QE? If yes, steps to reproduce]: Login to any Google service (gmail, youtube, etc) using GeckoView Example [List of other uplifts needed for the feature/fix]: None [Is the change risky?]: Low risk. [Why is the change risky/not risky?]: We're just backing out changes and restoring previous behavior. [String changes made/needed]: N/A (Note: I'll handle the back out myself, as it's likely to require some manual intervention.)
Attachment #9008438 - Flags: approval-mozilla-beta?
Comment on attachment 9008438 [details] [diff] [review] Back out async LoadURIDelegate Go for it. Approved for 63.0b7.
Attachment #9008438 - Flags: approval-mozilla-beta? → approval-mozilla-beta+
I've just verified that this fixes the regression with the latest nightly of GeckoView 64.0.20180914100149 in Firefox Reality and was able to log into youtube.com
Product: Firefox for Android → GeckoView
Keywords: regression
Target Milestone: Firefox 64 → mozilla64
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: