15 Apr 2024 |
adamw | kinda sounds like you're hitting something different | 17:43:47 |
@tablepc:matrix.org | Lenovo intel | 17:44:08 |
adamw | hmm | 17:44:36 |
adamw | just tried here on my bare metal test box and colors in an area recording seem fine | 17:44:47 |
adamw | !topic Accepted blocker review | 17:44:56 |
adamw | !topic (2242759) dnf system-upgrade fails on some RPi4 due to system boot date that pre-dates gpg key | 17:45:04 |
adamw | !link https://bugzilla.redhat.com/show_bug.cgi?id=2242759 | 17:45:06 |
adamw | !link https://pagure.io/fedora-qa/blocker-review/issue/1435 | 17:45:09 |
adamw | !info Accepted Previous Release Blocker, distribution, NEW | 17:45:12 |
adamw | we still seem to be stuck on this one :/ no great ideas, anyone? | 17:45:19 |
Brandon Nielsen | So, the small service to set a clock based on file age is pretty much how I would solve it. But I don't see how you make that actually happen before the update. | 17:46:06 |
Brandon Nielsen | At least automagically | 17:46:25 |
@tablepc:matrix.org | Window worked | 17:46:59 |
Fabio Valentini | isn't that what Before / After settings in unit files are for? | 17:47:11 |
Brandon Nielsen | Right, but how do you get that service installed and running on a system before you do the update without some kind of additional user interaction? | 17:47:40 |
adamw | ah, it looks like wireplumber update fixes the screencast thing | 17:47:54 |
Brandon Nielsen | And running in the update environment | 17:47:58 |
nirik | I think the update env is just a target... just needs to add that service Before it... | 17:48:51 |
Brandon Nielsen | Basically, any "fix" is just as bad of a story for the user as the currently documented fix. | 17:49:20 |
Brandon Nielsen | At least as I see it. | 17:49:30 |
Brandon Nielsen | Can skip the script and just enable timesyncd or whatever the systemd thing is. | 17:49:52 |
| Brandon joined the room. | 17:50:23 |
Fabio Valentini | In reply to @nielsenb:fedora.im Right, but how do you get that service installed and running on a system before you do the update without some kind of additional user interaction? just run it unconditionally? having the clock set to some date way in the past is always bad on devices like the RPi :) | 17:50:52 |
Brandon Nielsen | I guess by fix I mean workaround in commonbugs. | 17:50:56 |
nirik | I guess it would be nice to have the dnf maintainers weigh in on what they would accept? | 17:51:13 |
adamw | !info we still don't have a great idea for a fix for this, we'll have to discuss further at go/no-go | 17:51:17 |
nirik | because it could just do this as part of the normal system-upgrade | 17:51:26 |
Brandon Nielsen | https://discussion.fedoraproject.org/t/f38-to-f39-40-dnf-system-upgrade-can-fail-on-raspberry-pi/92403 | 17:53:05 |
Brandon Nielsen | Basically, touching a file is just as easy of a thing to tell a user to do, as install a package | 17:53:21 |
Brandon Nielsen | Side note, I feel like common issues on Discourse aren't as favored by Google as what we did before :-/ | 17:54:12 |