-
-
Notifications
You must be signed in to change notification settings - Fork 1k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Google Sign-In issues (fix and workaround available!) #5182
Comments
Also, let me add my personal opinion on this, from an earlier mailinglist mail:
|
Sorry for being possibly a bit off-topic, but out of interest what replacements do you use for GMail and Google Calendar? |
Thanks! I'll check out all of those |
For anyone watching this issue: I just released qutebrowser v1.9.0 with the fix. |
I run radicale and sync with DAVx⁵ as well. radicale works great! I haven't had any issues with it. |
I don't want to turn this into a conversation about Google, but either way: I just read Why I quit using Google – Kyle Piira which is a nice (even if horrifying) post which lists some more alternatives. |
I am a new user of Qutebrowser. I installed Qute on Ubuntu 18.04 using apt-get.
|
|
I have the same problem when I try to test the console project using webbrower control from Visual Studio [Not the default windows browser] |
Seems to work just as well, and is a bit closer to what we actually are. See qutebrowser#5182, qutebrowser#4810
FWIW looks like there's a chance this will break again when Google changes their behavior on January 4th. Right now our workaround still works even with the mentioned header, but with Google, you never know what'll change. Their "browser guidlines" are hypocritical to say the least:
cough ChromeDriver cough |
I encountered this issue today on the google accounts sign-in, despite having
|
@chaorace Oh no... 😟 Can you reproduce this with I have two Google accounts (
Can you elaborate what you mean with "the headers fix still worked when I tried that instead" exactly? |
Regarding the "headers fix", I'm referring to this snippet you included in the issue description: I experienced the issue on a fresh system install (as in a brand new disk, not just qutebrowser). I can't seem to replicate it on my primary disk, now that I'm attempting to do that. That could be related, more likely a red herring imo. In any case, I'll be hopping back onto my fresh install later today and will see if I can get the issue to reproduce once more from there. |
On the system where it breaks with the default setup, can you try if this is broken as well:
If it is, but the Firefox UA you used above helps, that'd be an easy fix. If that's the case, can you try with an updated Firefox UA too please?
(leave the My expectation is that the Edge UA won't help (because that's what the site-specific quirk is doing), but the new Firefox one would still help. |
@chaorace Any update on this? If the workaround needs adjustments, I'd really like to do them before the next release (which might happen soon-ish). |
Sorry for the delayed response. I just can't seem to reproduce the issue anymore. I even tried navigating back to the exact same URL in my history and that worked just fine. (this is all whilst using temp-basedir, of course). I can only guess it's some crazy Google heuristic or (more likely) some kind of flub on my end. All's well that ends well, I suppose? |
That's just how things are with Google... I remember back when this started (late 2019/early 2020) it was quite flaky for a while as well. Let's hope it keeps working now! If that's not the case, please let me know though. |
I was having the same issue, but using the command you recommended worked. |
Might be redundant at this point to mention, but I, too, can reproduce this. With (only) |
I had the same problem just starting today. Using |
I could reproduce with one of my Google accounts (though not another...). I updated the workaround to use an updated Firefox UA and released v2.3.1 with it. |
Hey, I'm glad to see you fighting the good fight here, but this will break, and soon. It only works for Firefox 90 because it's so new. Google gathers fingerprint data from each release of (what they consider to be) a major browser and clusters the results. Once they have enough results to partition the real Firefox 90 from qutebrowser, this will break. That will happen very very soon. Google published a paper explaining exactly how they do this (or rather, how they did this five years ago; what they do now is at least this good and almost certainly better): https://research.google/pubs/pub45581.pdf Unfortunately what you're doing here is only going to work for a few weeks after each release of Edge or Firefox. I came across comments in Mozilla's bugzilla where the Firefox devs noticed this weird adaptive behavior and figured it out (sorry, don't have the bug link anymore). Apparently not even they are looped in on all the details. Sorry to bear bad news. It's pretty evil what Google is doing here, basically blocking anybody from competing with their browser (except for a token nonprofit that is totally dependent on Google funding). |
@a-m-joseph Evidence seems to suggest otherwise, given that:
|
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
Looks like the Edge one doesn't work anymore. Reverts db13e52 and reintroduces the Firefox UA removed in e010afd (but updated). Closes qutebrowser#5182.
Looks like the Edge one doesn't work anymore. Reverts db13e52 and reintroduces the Firefox UA removed in e010afd (but updated). Closes qutebrowser#5182.
I do not know if it helps only saemideluxe's solution worked on me not 90 nor other peoples |
I wonder if we still need the workaround for this? We currently have a quirk which is active by default which sets a user agent of firefox 90. I wonder if we should do one of:
If anyone wants to test (2) you should be able to add "ua-google" to the |
Given that information about this is a bit scattered at the moment, I'm opening (and pinning) this issue in the hope that it's easier to find that way. Here's what's going on:
Explanation
Fix and workaround
content.site_specific_quirks
option was added (see Site-specific quirks overview #4810 and Refactor user agent handling and introduce site-specific quirks #5157 for details). With it enabled (which is the default), qutebrowser sends a Firefox user agent for the Google Accounts page (and other problematic pages like WhatsApp), thus fixing the issue.As a workaround, you can get the equivalent of that fix by running(see below for an updated workaround):set -u https://accounts.google.com/* content.headers.user_agent 'Mozilla/5.0 (X11; Linux x86_64; rv:57.0) Gecko/20100101 Firefox/57.0'
(needs v1.2.0 or newer)Jan 2021 Update
According to some people, this might be coming back - but not all Google Accounts seem to be affected, or at least not to the same degree. If you're seeing this with your account, please answer here!
Aug 2021 Update
This came back for most (all?) accounts now, with the Edge user agent qutebrowser uses as a workaround in v1.14.0 (db13e52). Thankfully, the Firefox one still works.
:set -u https://accounts.google.com/* content.headers.user_agent "Mozilla/5.0 ({os_info}; rv:90.0) Gecko/20100101 Firefox/90.0"
for an equivalent workaround.X11; Linux x86_64
in place of{os_info}
.The text was updated successfully, but these errors were encountered: