Closed
Bug 977173
Opened 10 years ago
Closed 3 years ago
Content interactions navigate user away from viewed page
Categories
(Firefox for Android Graveyard :: General, defect, P5)
Tracking
(firefox27 affected, firefox28 affected, firefox29 affected, firefox30 affected, fennec+)
People
(Reporter: aaronmt, Unassigned)
References
Details
(Keywords: reproducible, Whiteboard: [steps in comment #2])
STR i) Open a private tab, visit http://mail.google.com ii) In the same tab, visit http://facebook.com and log in iii) Tap the menu, and notification buttons See A Cross-Origin Request Blocked warning in console. The user is then redirected back to Gmail or whatever is viewed before Facebook. Don't know if the warning is related to the problem here. Please re-summarize if so.
Reporter | ||
Comment 1•10 years ago
|
||
On tap of Facebook's globe icon e.g, I am booted back to Gmail.
Reporter | ||
Comment 2•10 years ago
|
||
Happens in regular tabs too. I noticed I had to hit the menu, and notifications, and messages icons a few times. Each time Facebook did a 'reload' and eventually I was redirected back to content viewed before accessing Facebook.
Summary: Blocked Cross-Origin Requests in Private Tabs navigate user away from viewed page → Content interactions navigate user away from viewed page
Reporter | ||
Updated•10 years ago
|
status-firefox27:
--- → unaffected
status-firefox28:
--- → unaffected
status-firefox29:
--- → affected
tracking-firefox28:
? → ---
tracking-firefox30:
--- → ?
Reporter | ||
Updated•10 years ago
|
Keywords: regression,
regressionwindow-wanted
Comment 3•10 years ago
|
||
I'm able to reproduce this bug even in release.
tracking-firefox28:
--- → ?
tracking-firefox30:
? → ---
Keywords: regression,
regressionwindow-wanted
Updated•10 years ago
|
tracking-firefox30:
--- → ?
Comment 4•10 years ago
|
||
Bleh, sorry for bugspam. I guess we don't want to track specific firefox releases if this isn't a regression, do we?
Reporter | ||
Comment 5•10 years ago
|
||
Confirming I see this on release too, took a little more effort though.
Reporter | ||
Comment 6•10 years ago
|
||
This seems to happen frequently in interacting with the menu and social toolbar in Facebook in 28.0b6 thus making Facebook (in conjunction with bug 977226) painful to use on beta 28.0
Comment 7•10 years ago
|
||
Apparently this go significantly worse on 28, need to investigate
Assignee: nobody → bnicholson
tracking-fennec: ? → 28+
Comment 8•10 years ago
|
||
With bug 977226 fixed, the behavior on 28/29/30 is now the same as release. It's reproducible, but not as easy to hit. STR are mostly the same as comment 0, but for step iii, repeatedly (and quickly) hit the facebook chat icon. Renoming since we're back to release state.
tracking-fennec: 28+ → ?
Reporter | ||
Comment 10•10 years ago
|
||
Analyzing builds from last April and I am still able to reproduce. Been broken for a long time if not always. With more recent builds, e.g, using Nightly (03/02), today's build, it is far more easier to reproduce. Earlier builds require more effort.
Comment 11•10 years ago
|
||
Mike, this might be a Facebook-specific issue. Could you have a look?
tracking-fennec: ? → +
Flags: needinfo?(miket)
Reporter | ||
Updated•10 years ago
|
Whiteboard: [steps in comment #2]
Comment 12•10 years ago
|
||
The Facebook issue I described in bug #978536 doesn't happen anymore with the last beta from the Google Play store (I'm able to open the Facebook toaster, messages and globe menus again). So it may have been a different bug, but it seems it has been fixed anyway.
Reporter | ||
Comment 13•10 years ago
|
||
The previous candidate had a patch which made it easier to reproduce.
Comment 14•10 years ago
|
||
On PTO until next Thursday, so cc'ing Hallvord to see if he can get to this quicker.
Comment 15•10 years ago
|
||
Hallvord, can you take a look?
Comment 16•10 years ago
|
||
There's some stuff going on here: https://fbstatic-a.akamaihd.net/rsrc.php/v2/yt/r/qtvra8B3kkS.js popSoftState: function (p) { if (!o._enabled) return ; if (o._softState === p) n(function () { if (o._softState === p) history.go( - 1); }, 0); }, <X> pushState: function (p) { if (!o._enabled) return ; if (o._softState) { return o.replaceState(p); } else { o._updateSoftState(p); o._updateXRefererCookie(p); return i.push(new k(p) .normalize() .toString()); } }, I think what happens is that the script is written to "pop" the current state and, when the new "state" loads, "push" it. Now, "popping" the state involves actual history navigation (history.go(-1)) - meaning that if you push a button again before the new feature/state loads and is push()ed, you'll end up calling history.go(-1) twice without an intermediate navigation - the first go(-1) will undo the #hash navigation that's part of Facebook's history mechanism while the second (because no new #hash was added meanwhile) will take you right back to the previous site. (This is still a bit of guesswork - I'll do some logging and see if I can verify it)
Flags: needinfo?(hsteen)
Updated•9 years ago
|
Assignee: bnicholson → nobody
Comment 18•9 years ago
|
||
I assume Facebook has changed since this bug was filed... is this still an issue?
Comment 19•3 years ago
|
||
We have completed our launch of our new Firefox on Android. The development of the new versions use GitHub for issue tracking. If the bug report still reproduces in a current version of [Firefox on Android nightly](https://play.google.com/store/apps/details?id=org.mozilla.fenix) an issue can be reported at the [Fenix GitHub project](https://github.com/mozilla-mobile/fenix/). If you want to discuss your report please use [Mozilla's chat](https://wiki.mozilla.org/Matrix#Connect_to_Matrix) server https://chat.mozilla.org and join the [#fenix](https://chat.mozilla.org/#/room/#fenix:mozilla.org) channel.
Status: NEW → RESOLVED
Closed: 3 years ago
Resolution: --- → INCOMPLETE
Assignee | ||
Updated•3 years ago
|
Product: Firefox for Android → Firefox for Android Graveyard
You need to log in
before you can comment on or make changes to this bug.
Description
•