hmlendea/gfn-electron

Login Unsuccessful. We are sorry to report we were unable to log you into GeForce NOW.

sethidden opened this issue · 15 comments

  1. Download latest release's zip on Arch Linux (geforcenow_1.1.0_linux.zip) or Windows 10 (npm install from repo)
  2. Run the launcher ./geforcenow (or npm start on Windows 10)
  3. Try to log in - in my case it opens a new window saying 'you'll be logged in automatically in 5 seconds' (I logged in before with the same error I'm reporting) then it says:
Login unsuccesful
We are sorry to report we were unable to log you into GeForce NOW. Please try again
[Close] [Retry]

However login works on:

  1. Lutris launcher for GeForce NOW (Arch Linux)
  2. Google Chrome (both Arch and Windows 10)
  3. Windows 10 desktop client

Arch Linux output:

t430 :: ~/Downloads » ./geforcenow
libva error: vaGetDriverNameByIndex() failed with unknown libva error, driver_name = (null)
GeForce NOW
GeForce NOW
NVIDIA
Sign in with NVIDIA
NVIDIA GeForce NOW

The libva error is probably because I'm on an Intel HD4000 graphics card - but the app launches fine nonetheless and I can proceed until the "Login Unsuccesful" modal. Windows 10 output is the same just that it doesn't have the libva error, so it's probably not that anyways.

I added

  globalShortcut.register("Super+Shift+I", () => {
    BrowserWindow.getAllWindows().forEach((__window) => {
      __window.toggleDevTools();
    })
  });

to the app to see what's happening in the main and login windows.

The login window sends the login data and it receives a normal 302 request with a location header of the following format:

https://play.geforcenow.com/redirect/starfleet-oauth-redirect.html?ui_locales=en_US&nonce=[nonce here]&token_type=Bearer&expires_in=3600&access_token=[access_token here]

This looks like a "everything is ok" request to me. If it was wrong the 302 would redirect to some redirect?error="some_error_id" or whatever.

I'm suspecting the main window's (the one with the games, not the login window) service worker fails to read the new login data somehow. It's just that I can't see the incoming/outgoing messages from a service worker

Hi, having this exact issue with the login, here's what the terminal shows for me.

In case its important, I am on the latest version of Ubuntu studio, and I did install it via snapd, though given the nature of snaps I'm not sure if my OS really has much to do with it in that regard.

fenrave@fenrave-ms7b79:~$ geforcenow 
Gtk-Message: 19:24:43.564: Failed to load module "colorreload-gtk-module"
GeForce NOW
GeForce NOW
NVIDIA
Sign in with NVIDIA
GeForce NOW
NVIDIA
Sign in with NVIDIA
NVIDIA GeForce NOW
GeForce NOW
NVIDIA
Sign in with NVIDIA
Error occurred in handler for 'GUEST_WINDOW_MANAGER_WINDOW_METHOD': Error: Invalid guestId: 3
    at electron/js2c/browser_init.js:193:1580
    at electron/js2c/browser_init.js:201:579
    at Object.<anonymous> (electron/js2c/browser_init.js:161:10005)
    at Object.emit (events.js:315:20)
fenrave@fenrave-ms7b79:~$ 

No way to brute-force it either by selecting automatic login, It'll just tell you the same thing.

Hello, I have this issue too.

libva error: vaGetDriverNameByIndex() failed with unknown libva error, driver_name = (null)
GeForce NOW
GeForce NOW
NVIDIA
Sign in with NVIDIA
NVIDIA GeForce NOW

Same issue here.

Same here

same for me

Hello, I have this issue too.

libva error: vaGetDriverNameByIndex() failed with unknown libva error, driver_name = (null)
GeForce NOW
GeForce NOW
NVIDIA
Sign in with NVIDIA
NVIDIA GeForce NOW

Does instaling libva fix the problem?
e.g. on Arch Linux with Intel I have libva and intel-media-driver

The libva error is probably because I'm on an Intel HD4000 graphics card

That's the same card I'm having. The error occurs because there is no VAAPI driver installed on the system, and the application fails to use hardware acceleration (which will result in input lag)

I will need the following info:

  • Which version did you install? (snap, deb, other...)
  • Is this the first time you run this app on that system?
  • Is this happening only after updating the app, while the account was sucessfully logged in prior to the update?
  • Does deleting ~/.config/GeForce NOW and trying again help?
uncon commented

I am running Arch, have community/intel-media-driver (21.1.3-1) installed, and have tried removing "~/.config/GeForce NOW" to no avail.

  • Which version did you install? (snap, deb, other...)
    • geforcenow-electron in AUR
  • Is this the first time you run this app on that system?
    • Yes
  • Is this happening only after updating the app, while the account was sucessfully logged in prior to the update?
    • No; I have only recently (after the most recent update) tried this app
  • Does deleting ~/.config/GeForce NOW and trying again help?
    • No

I tried both the appimage and deb versions.
Is not the first time I use this app.
It did worked before the update.
Deleting ~/.config/GeForce NOW doesn't work.

What works though is: use a previous version (1.0.9) to login, close it and then immediately start the current version (1.1.0).
I only tested this with the appimage versions, but this way I managed to get connected.

Arch & Wayland Sway + X11 libs

Deleting ~/.config/GeForce NOW doesn't work.

Login unsuccesful
We are sorry to report we were unable to log you into GeForce NOW. Please try again
[Close] [Retry]

Ubuntu 20.04. I'm using the 1.1.0 .deb. Same problem.

Also, oddly enough, the 1.1.0 from popey on the snap store was working for me when upgrading for previous version. I remained logged in. However, when I logged out, it wouldn't log me back in.

This is a regression, as installing the 1.0.8 version logs me in. I then upgrade to 1.1.0 and am still logged in. This is a workaround to get hardware acceleration.

Hello.
I have same problem.

Ubuntu 20.04, 1.1.0 - Hardware acceleration

LAtest release in use:
yana@sergey-R425D-R525D:~/geforcenow-electron$ git checkout M package-lock.json Your branch is up-to-date with «origin/master». yana@sergey-R425D-R525D:~/geforcenow-electron$ git pull Already up-to-date.

`yana@sergey-R425D-R525D:~/geforcenow-electron$ npm start

geforcenow@1.1.0 start /home/yana/geforcenow-electron
electron --enable-accelerated-mjpeg-decode --enable-accelerated-video --ignore-gpu-blacklist --enable-native-gpu-memory-buffers --enable-gpu-rasterization .

GeForce NOW
GeForce NOW
Log in to GFN
NVIDIA GeForce NOW
Error occurred in handler for 'GUEST_WINDOW_MANAGER_WINDOW_METHOD': Error: Invalid guestId: 2
at electron/js2c/browser_init.js:193:1580
at electron/js2c/browser_init.js:201:579
at Object. (electron/js2c/browser_init.js:161:10005)
at Object.emit (events.js:315:20)
`

/var/log/syslog
Apr 13 23:46:43 sergey-R425D-R525D dbus-daemon[845]: [system] Activating via systemd: service name='org.bluez' unit='dbus-org.bluez.service' requested by ':1.456' (uid=1001 pid=18610 comm="/home/yana/geforcenow-electron/node_modules/electr" label="unconfined")

изображение

Version 1.0.8 - Maintenance works correctly

Successfully reproduced the issue and confirmed this as a regression due to the latest electron version, not the hardware acceleration.

I'm still debating what approach should I take, but as a last resort I can temporarily downgrade electron to 11.4.2

EDIT: Proper fix identified. There was a breaking change in electron v12 and I've updated this app accordingly. GFN Login now works using the latest stable electron version 12.0.2. A new release with the fix will arrive shortly.

Version 1.1.1 that fixes this issue just got released.

I will close this thread. If any more login issues occur, please reopen.