at this rate we'll need to find a few more people so my access can be removed. 😄
at this rate we'll need to find a few more people so my access can be removed. 
B
Jok I kicked off main and then I'll do nvidia and then I promise no new things
Bi think that's actually your plan
JUNTIL THE TRAIN COMES TO THE STATION, then no promises
Joh the train was here, ok one more
CDoes anyone remember where manually added udev rules go? I have been preoccupied and don't know where the file is that I need to remove to test that things are working
J/etc/udev/rules.dJiirc your plan was to put your rules there until we sorted them out
CYep, found them. Good to update and reboot to check?
Bnot yet
Byou are on nvidia right?
CYeah
Bit'll be a bit
CFigured
Bmainnvidia
RI was just looking at the packit.dev service. has anyone looked at it before?
rpm build as a service to aim for getting packages to Fedora and CentOS
rpm build as a service to aim for getting packages to Fedora and CentOS
AYeah, I was just thinking that when Yafti is more mature, it might be useful to have it in main to do some post-install setups, but you're right, it's too opinionated for main.
Bdid you see that Fedora 39 has a proposed first-boot service?
https://fedoraproject.org/wiki/Changes/AutoFirstBootServices
https://fedoraproject.org/wiki/Changes/AutoFirstBootServices
AThis should be fine since it's correct modern way to define input device rules.
And if I've even understood Fedora correctly, they dynamically assign people to the access
So I don't think anyone needs to add themselves to
And if I've even understood Fedora correctly, they dynamically assign people to the access
input group files when they are physically logged into the desktop -- via ACLs.So I don't think anyone needs to add themselves to
input either.JI looked into these firstboot services last cycle
Jit won't work for us but I don't remember why
AYou mean yafti, right?
JJno this thing
AAhh it seems similar to yafti, but silent. Probably more limited.
ARegarding device rules... Our
Source: https://bugzilla.redhat.com/show_bug.cgi?id=815093#c4
70-u2f.rules is a bit wonky too, which is a separate issue to look at later... It currently requires the plugdev group. Here's what a Fedora dev said about that:Sure, we must entirely get rid of that crap. "plugdev" is a very misguided
and fundamentally wrong Ubuntuism. Fedora must not solve problems that way,
we do not put users in system groups to grant access to devices, especially
do not define non-sensical groups like: "the group of users that can plug
devices into the system".
Fedora uses individually managed ACLs for granting ordinary user access
to devices. The ACLs are seat (multiple users working simultaneously on
the same host), and they track session activity. It depends on the type
of devices how to hook up the device in the ACL handling:
http://cgit.freedesktop.org/systemd/systemd/tree/src/login/70-uaccess.rules
Please just patch the rules not to use "plugdev" at all, we can do better and
don't want "plugdev" in Fedora. Thanks!
Source: https://bugzilla.redhat.com/show_bug.cgi?id=815093#c4
Bi looked into the source on pagure...
some of the default functionality it ships is definitely not helpful for us... but it could provide a semi-standard hook... that's assuming it's included in rpm-ostree images and not just on workstation, etc
some of the default functionality it ships is definitely not helpful for us... but it could provide a semi-standard hook... that's assuming it's included in rpm-ostree images and not just on workstation, etc
JI think we'd want to use it to launch yafti perhaps, more to replace the dotfile thing we're doing now
Jwe should look at it with fresh eyes though
Bexactly what i was thinking, a more "standard" way to launch yafti vs rolling our own
Bbut not urgent at all
Jyeah when I was looking at the example was layering a codec package and I was like that's going to be a hard pass for me dawg.
AYeah it might be useful, if it's capable of running yafti on every login. Since yafti needs to re-run whenever
yafti.yml changes.Byep, that would have to be disabled for sure
Jyeah plus also it's there now by default, one less thing for us to break
AYep, let's hope it can do "every login" actions.
Bakmods: urgent
redoing first-boot: *not urgent"
redoing first-boot: *not urgent"
Jright, also this is 39 only
Jso probably post-rebase-to-official images
AI suspect that the Fedora thing will be too limited for us though. They probably only support shell stuff that doesn't require a GUI. Running before the user has logged in. Would be my guess.
Jactually, anyone got a few minutes now? let's give the board some love and milestone some crap
J
Boh sure why not
Bi'm waiting for builds so i can reboot
AI'll have a look and help in any discussions we're having about the board. 
I don't know enough about the main repos yet to say whether something is already done etc.
I don't know enough about the main repos yet to say whether something is already done etc.Jit's ok let's just capture and triage, I'll scrummaster until someone who knows what they're actually doing shows up
POne thing I've noticed about yafti: it creates the file that keeps it from rerunning regardless of how it exits. Whether you've visited all the pages, or closed it manually, or it crashes (this was happening to me because I was experimenting adding my own custom pages). Is that intentional? Seems like that behavior could be a little btter.
AWe need to sort out how to connect vscode to boxes, set up devcontainers, all of that stuff.
That's an interesting one. I've been hearing that VSCode flatpak with
remote-containers extension can run things inside podman containers very easily.JGitHub
https://fedoraproject.org/wiki/Changes/AutoFirstBootServices This is for F39 so let's investigate after the switch to official images.
