| 15 Jul 2021 |
 Jan Tojnar | * the best bet would be going to #gdm:gnome.org and asking for some debugging advice there | 17:28:18 |
 hmenke | Jan Tojnar: I have some logs for you. | 18:19:40 |
| hmenke | Download fprint.log | 18:19:50 |
| hmenke | Particularly the bit
Jul 15 20:17:25 laptop .gnome-shell-wr[2989]: Fingerprint service failed almost immediately, considering it unavailable.
Jul 15 20:17:25 laptop gdm-fingerprint][23815]: pam_warn(gdm-fingerprint:auth): function=[pam_sm_authenticate] flags=0 service=[gdm-fingerprint] terminal=[/dev/tty7] user=[henri] ruser=[<unknown>] rhost=[<unknown>]
Jul 15 20:17:25 laptop gdm-fingerprint][23815]: Gdm: GdmSessionWorker: authentication returned 7: Authentication failure
Jul 15 20:17:25 laptop gdm-fingerprint][23815]: Gdm: GdmSessionWorker: uninitializing PAM
Jul 15 20:17:25 laptop .gnome-shell-wr[2989]: Please fix your configuration by running: authselect select --force sssd with-fingerprint with-silent-lastlog
| 18:20:28 |
| hmenke | A search also turned up this https://fedoraproject.org/wiki/Common_F34_bugs#login-authselect-fingerprint-error | 18:21:53 |
| hmenke | This is the linked bug report https://bugzilla.redhat.com/show_bug.cgi?id=1942443 | 18:23:45 |
|  maxine (they/them) joined the room. | 22:58:50 |
| Jan Tojnar | hmenke: I have seen that. But I do not see how that is applicable, we do not have sssd or authselect. | 23:28:28 |
| Jan Tojnar | And it was not clear to me from the bug report what is actually causing the issue, only that Fedora somehow fixed it | 23:28:57 |
| Jan Tojnar |
Maybe the GNOME Shell patch fixed it (but it sounded like a hack and the instructions do not really look like the issue has anything to do with gnome-shell) | 23:29:24 |
| 16 Jul 2021 |
| hmenke | Some more research on this:
- https://src.fedoraproject.org/rpms/gnome-shell/c/db6b665?branch=f34
- https://gitlab.gnome.org/GNOME/gnome-shell/-/merge_requests/1821
| 06:54:55 |
| Jan Tojnar | hmenke: yeah, 0001-gdm-Work-around-failing-fingerprint-auth.patch is the patch I was referring to. The MR !1821 is already in Nixpkgs. | 11:58:39 |
| 17 Jul 2021 |
| maxine (they/them) changed their display name from maxine [they] to maxine (they/them). | 12:04:13 |
| 19 Jul 2021 |
 hpfr | bit of a long shot, but is steam's big picture broken for anyone else on gnome wayland? | 02:57:13 |
| hpfr | nvm, I found it's fixed in a nixpkgs PR | 02:59:55 |
|  mith joined the room. | 17:19:47 |
| 21 Jul 2021 |
|  Josh Ponelat joined the room. | 14:23:36 |
| Josh Ponelat | Hey folks, I recently upgrade (channel) to 21.05, and now gnome is gnome4. Is there anyway I can get it back to gnome3? There is an extension that I need in gnome3 | 14:23:45 |
|  leznom joined the room. | 18:20:09 |
 piegames | No, Gnome 3.38 is end of life | 20:34:19 |
| piegames | You can look at the extension's upstream if they have an experimental 40 support branch. | 20:35:15 |
| 22 Jul 2021 |
| hmenke | Josh Ponelat: I think not so much changed between 3.38 and 40 for extensions, so often it might be sufficient to just add 40.0 to shell-versions in metadata.json. That is basically how Pop_OS! Shell added GNOME 40 support and that is probably the most complex extension out there: https://github.com/pop-os/shell/commit/0143b0b5eb14291cbd9e0b8328eeec93c4871ba9 | 12:53:22 |
| 23 Jul 2021 |
| piegames | I think the most important change was the Gtk3 -> Gtk4 migration? And also the new overview, for all extensions that had their fingers in it | 11:02:51 |
| Josh Ponelat | In reply to @hmenke:matrix.org
Josh Ponelat: I think not so much changed between 3.38 and 40 for extensions, so often it might be sufficient to just add 40.0 to shell-versions in metadata.json. That is basically how Pop_OS! Shell added GNOME 40 support and that is probably the most complex extension out there: https://github.com/pop-os/shell/commit/0143b0b5eb14291cbd9e0b8328eeec93c4871ba9 Thank you kindly! I'll look into that (and yeah, I'm interested in pop shell too 🙂 ) | 11:03:41 |
| hmenke | Josh Ponelat: https://github.com/hmenke/nixos-modules/tree/master/pkgs/pop-os-shell | 12:24:46 |
| maxine (they/them) | How does nix deal with packaging things that have circular dependencies? I am looking at bumping gupnp, but through gssdp it depends on gtk4 now, which ends up eventually depending gupnp again (via gst-plugins-bad). Is the solution providing some sort of "minimal" gtk4 derivation to get things untied? | 23:52:50 |
| 24 Jul 2021 |
 symphorien | That or you could make the gtk unwrapped derivation a distinct attribute set. I suppose that the dependency on gst only comes through wrapGappsHook ? | 02:52:13 |
| 26 Jul 2021 |
|  benedikt_ahle joined the room. | 13:36:50 |
| benedikt_ahle left the room. | 13:38:36 |
| 27 Jul 2021 |
|  @appservice-irc:matrix.org joined the room. | 10:20:27 |
