Open
Bug 1465391
Opened 6 years ago
Updated 2 years ago
User cannot log out of Pinterest on a specific platform and OS
Categories
(Core :: DOM: Events, defect, P3)
Tracking
()
NEW
People
(Reporter: vlucaci, Unassigned)
Details
(Keywords: webcompat:site-wait, Whiteboard: [platform-rel-Pinterest][sitewait])
[Affected versions]: - 62.0a1 (2018-05-28) [Affected platforms]: - macOS 10.13 [Steps to reproduce]: 1.Go to Pinterest.com 2.Log into Pinterest with Facebook credentials. 3.Browse the collection for a while. 4. Log out of the account. [Expected result]: -Taping on the "Log Out" button should redirect the user out of his account. [Actual result]: -Tapping on the "Log Out" button does not log the user out of his account. Instead, he can browse the whole page and its subsections freely as if logged in(can even create albums, add pictures or upload) and only after ~5 minutes a message will be displayed "Something went wrong-Authetication Failed" [Regression range]: - Will return with a regression range ASAP. [Additional notes]: -Refreshing the page after tapping the Log Out button has the following two behaviors: 1. After page is refreshed , user appears to be logged out, but will be re-logged in automatically after ~10 second waiting time. 2. After page is refreshed, effective immediately user will be prompted with "Something went wrong-Authetication Failed" and on whatever page the user attempts to reach from Pinterest. 3. This issue was not reproduced on any other FF build, nor on Chrome and not on any other macOS systems. For more information , please consult the attachment provided in the below GDrive url: https://drive.google.com/file/d/1S_RGREdAEzdvyXePmJNwFjA6cY-PD6I3/view?usp=sharing
Updated•6 years ago
|
Keywords: regression
Reporter | ||
Comment 1•6 years ago
|
||
Hello, I have returned with the regression range: Last good revision: dee01445e642e20ea80664ddc28e8948ea8f1af1 First bad revision: f948327f928d4358269227169546f0932a20fb66 Pushlog: https://hg.mozilla.org/integration/autoland/pushloghtml?fromchange=dee01445e642e20ea80664ddc28e8948ea8f1af1&tochange=f948327f928d4358269227169546f0932a20fb66 Hope it helps!
Comment 2•6 years ago
|
||
That pushlog points to a test-only change, which seems rather unlikely.
Flags: needinfo?(vlad.lucaci)
Reporter | ||
Comment 3•6 years ago
|
||
Hello I have returned with a few updates and a new pushlog: The issue no longer occurs as it did when the issue was submitted. Now, when tapping on the log out button , the user is actually logged out, but right before that, a "Something went wrong" prompt is displayed . Keeping this new behavior in mind, I have managed to pull another regression range, and the issue withing it seems to be related to this behavior. Also, attached you will find a video url of the new behavior: https://drive.google.com/file/d/1yZHxtF01sFHsr1z0ZCvM3dEiKiBgKvqy/view?usp=sharing Last good revision: 240114d8acd30222fe7d18b88b1ba25b227d3c9b First bad revision: 497aead7d738c85926e1edf6259dde2eef6d368b Pushlog: https://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=240114d8acd30222fe7d18b88b1ba25b227d3c9b&tochange=497aead7d738c85926e1edf6259dde2eef6d368b
Reporter | ||
Updated•6 years ago
|
Flags: needinfo?(vlad.lucaci)
Comment 4•6 years ago
|
||
(In reply to Vlad Lucaci (:vlucaci) from comment #3) > Last good revision: 240114d8acd30222fe7d18b88b1ba25b227d3c9b > First bad revision: 497aead7d738c85926e1edf6259dde2eef6d368b > Pushlog: > https://hg.mozilla.org/integration/mozilla-inbound/ > pushloghtml?fromchange=240114d8acd30222fe7d18b88b1ba25b227d3c9b&tochange=497a > ead7d738c85926e1edf6259dde2eef6d368b This patch just disables a test -- it doesn't seem related. Is it possible that Pinterest is just having site issues? Vlad, does the bug still reproduce as in Comment #3.
Flags: needinfo?(vlad.lucaci)
Reporter | ||
Comment 5•6 years ago
|
||
(In reply to Mike Taylor [:miketaylr] (62 Regression Engineering Owner) from comment #4) > (In reply to Vlad Lucaci (:vlucaci) from comment #3) > > Last good revision: 240114d8acd30222fe7d18b88b1ba25b227d3c9b > > First bad revision: 497aead7d738c85926e1edf6259dde2eef6d368b > > Pushlog: > > https://hg.mozilla.org/integration/mozilla-inbound/ > > pushloghtml?fromchange=240114d8acd30222fe7d18b88b1ba25b227d3c9b&tochange=497a > > ead7d738c85926e1edf6259dde2eef6d368b > > This patch just disables a test -- it doesn't seem related. > > Is it possible that Pinterest is just having site issues? Vlad, does the bug > still reproduce as in Comment #3. Hey Mike, I would agree with you that Pinterest may have site issues, but after I have tried to reproduce this issue multiple times on the following browsers I would beg to differ: - Chrome 67.0 - Opera 53.0 - Safari 11.1 The above browsers were tested on Windows 10x64 and macOS 10.13 and Ubuntu 16.04(Chromium) and the issue is not present there no matter how many times I tried to reproduce. When trying to reproduce the issue with the latest FF nightly on Windows10x64, macOS 10.13 and Ubuntu 16.04 I get the issues with a repro rate of 80% (8 out 10 attempts). This time, I used the Web Console to see if any errors are triggered when the "Something went wrong" prompt is displayed when logging out and I have managed to record this error contained in the following link: https://drive.google.com/drive/folders/1uV9fGQny9b4kb6SHJlKxzfulBBGQIocW?usp=sharing Hope it helps.
Flags: needinfo?(vlad.lucaci)
Updated•6 years ago
|
Flags: needinfo?(miket)
Updated•6 years ago
|
tracking-firefox62:
--- → +
Updated•6 years ago
|
Flags: in-qa-testsuite?
Comment 6•6 years ago
|
||
Hey Mike, Can I assign this to you as the REO for Fx62 to determine if there is really any action here that a dev can take? If there is something for a dev to dig into, can you reach out to Justin for an owner?
Assignee: nobody → miket
Updated•6 years ago
|
Component: General → Untriaged
Comment 7•6 years ago
|
||
Yep, thanks. I can reproduce now. Some (maybe irrelevant) errors in the browser console when I see this: Name: TypeError Message: u._showErrorCallback is not a function ---------- Stack: BVzu/u.showHtmlError@https://s.pinimg.com/webapp/js/entryChunk-www-unauth-362f585707be558e5b54.js:9:57790 BVzu/u.showError@https://s.pinimg.com/webapp/js/entryChunk-www-unauth-362f585707be558e5b54.js:9:57727 e/<@https://s.pinimg.com/webapp/js/entryChunk-www-unauth-362f585707be558e5b54.js:29:28156 a@https://s.pinimg.com/webapp/js/vendor-react-d20f99c48b5d58e4821c.js:4:35163 F/<@https://s.pinimg.com/webapp/js/vendor-react-d20f99c48b5d58e4821c.js:4:35285 s@https://s.pinimg.com/webapp/js/vendor-react-d20f99c48b5d58e4821c.js:3:48070 <unavailable> Name: Message: ---------- Stack: <unavailable> Failed to load ‘https://s.pinimg.com/webapp/js/pjs-locale-en_US-lite-3dcf38fa608036c641ca.js’. A ServiceWorker passed an opaque Response to FetchEvent.respondWith() while handling a ‘cors’ FetchEvent. Opaque Response objects are only valid when the RequestMode is ‘no-cors’. sw.js:144:6 However, this reproduces for me going back to at least v51 nightly. We don't need to track this as a regression, and certainly not for 62 (IMHO).
status-firefox63:
--- → affected
Flags: needinfo?(miket) → needinfo?(lhenry)
Keywords: regression,
regressionwindow-wanted
Comment 8•6 years ago
|
||
Dennis, can you look into the above error (Comment #7) and STR (in Comment #0)?
Flags: needinfo?(dschubert)
Comment 9•6 years ago
|
||
(Note: this still reproduces even when spoofing as Chrome)
Updated•6 years ago
|
Flags: needinfo?(lhenry)
Comment 10•6 years ago
|
||
Sorry for the late reply, I was on PTO, a conference, then got sick and had a way too large backlog to work through. I am a bit lost here. I do neither see the console errors mentioned in Comment 7, and I also can log out without any issues. The only thing I see in the console is a lot of warnings about ignored CSPs, but this is to be expected since they specified strict-dynamic. However, it should be noted that I get force-redirected to pinterest.de all the time, even when using a US IP. As soon as I login via any of the offered methods, I end up on the German domain. Tom, do you have better luck?
Flags: needinfo?(dschubert) → needinfo?(twisniewski)
Reporter | ||
Comment 11•6 years ago
|
||
Hello all, Coming back with an update: This issue is still present with the exact same repro steps and rate, as well as on the same specified platforms. @Dennis ,as I have specified in Comment 3, the problem isn't that the user is not logged out of his account. The problem is that the logout process is faulty only on FF. . A "Something went wrong" prompt is displayed right before he is being completely logged out. For more information, I have created an attachment with the behavior in its current state: https://drive.google.com/file/d/10yLUALoyCUFOuGfafr7nvzvpqsx9yw8y/view?usp=sharing
Reporter | ||
Updated•6 years ago
|
Component: Untriaged → DOM: Events
Product: Firefox → Core
Comment 12•6 years ago
|
||
Adam, can we get in touch with Pinterest about this? Perhaps they would have some insight.
Flags: needinfo?(astevenson)
Whiteboard: [platform-rel-Pinterest]
Comment 13•6 years ago
|
||
Apologies for the delay. Contacted via mailing list.
Flags: needinfo?(astevenson)
Whiteboard: [platform-rel-Pinterest] → [platform-rel-Pinterest][sitewait]
Comment 14•5 years ago
|
||
See bug 1547409. Moving webcompat whiteboard tags to keywords.
Keywords: webcompat:site-wait
Updated•5 years ago
|
Priority: -- → P3
Comment 15•5 years ago
|
||
Apologies for the ridiculous delay here, but I can't reproduce this one with the given STR. Do I need to browse for a longer time than ~3 minutes before it reproduces, or has this perhaps been fixed? I tested on today's nightly build on OSX 10.14.4.
Flags: needinfo?(twisniewski) → needinfo?(vlad.lucaci)
Reporter | ||
Comment 16•5 years ago
|
||
Hello all,
I tried reproducing this issue with FF Nightly 68.0a1 (2019-05-15), FF 67.0 and DevEdition 68.0b1 on macOS 10.14 and macOS 10.13 with the same repro steps and this issue is no longer present.
Flags: needinfo?(vlad.lucaci)
Updated•4 years ago
|
Assignee: miketaylr+bz → nobody
Updated•2 years ago
|
Severity: normal → S3
You need to log in
before you can comment on or make changes to this bug.
Description
•