Fedora 40 Release Thread: Important for
Fedora 40 Release Thread: Important for all approvers! (Will file issue in github after discussion)
37 Replies
Alright, we have three things in flight that will conflict so we're going to have to go in order.
1) F40 enable goes first, @bsherman will merge to unblock the other two, we are still waiting on confirmation from timothee on the final image names, so even though he is merging, the images WILL PROBABLY BE RENAMED so if you decide to yolo yourself onto F40 you are on your own if you're not paying attention
2) HWE deconflicting - this has served us well and has gotten us an awesome relationship with Framework and other interested parties. Now it's time to trim it down
3) ARM images - RJ got a mac but was kind enough to not give into the shiny light, once we've gotten the first two items done this should go way smoother.
Let's try to get these in order because if we go the other way it will suck worse.
That leaves ISOs to further discuss, but those seem well in hand and what we have now is "good enough" but offline flatpaks would be awesome, plenty of time.
ANYTHING ELSE FOR FEDORA 40?
It's past free practice and qualifying, everything not bolted down should either be sunset or archived, let's go into GA with as little technical debt as possible.
(Anything not on this list won't be brought forward into F40, so old experiments and repos will remain at the version they are and naturally fall out of support)
I created an issue for tracking: https://github.com/ublue-os/nvidia/issues/206
thank you, cramped plane typing isn't fun lol
On the F40 naming, I've added my thoughts to the PR and approved.
In short, I vote for keeping what we have for now, and switching as soon as we find out upstream's naming convention.
Then later we decide if we will maintian the old image names or not (I don't think we should).
Looks like we've got skew in main
@Kyle Gospo got time to take a look?
Out on a walk but I can take a look soon
ack
Isn't it the usual RPMFusion issues that work themselves out over time, without us needing to do anything?
Aye, that looks like it
yeah, I still can't tell from the output
we could turn it off, file the issue and then revisit since it's the only one?
aptx is useful af. 😄
Sorry, not sure what you mean by "turn it off". Failing builds are currently all F40 images, but were building successfully the last run.
If we simply merge now, it will pass when RPMFusion and Fedora provide compatible packages. Until then, it'll fail
If we simply merge now, it will pass when RPMFusion and Fedora provide compatible packages. Until then, it'll fail
no I mean not do the aptx packge
or would we need to special case f40?
I expect F40 to fail a lot between now and launch. That’s the other reason to not stress about preview builds
yep, let it fail
We can merge and let it fail
ok so want me to merge it?
ON IT
old on
hold on
let me say what @bsherman wanted to say
"you wanted F40, here you go."
"but it doesn't work"
"I know."
Thanks @j0rge you get me
Will prep the Bluefin PR, but there's other dependencies that we will be waiting for, like blur-my-shell hasn't got F40 building
"damn bro that's crazy, happy for you or sorry to hear that, whatever"
bluefin is always one behind, I ain't sweating 40
I'm more worried about the 99 other things ahead of it, hahaha
I'm more worried about the 99 other things ahead of itLooking at your list, I don't think there's any show-stoppers except for that F40 build. HWE is a DevExp change that can happen at any point? It should be invisible to end users as I assume the image names will remain the same.
oh I meant the other bluefin parts
anything else for tonight from this hitlist?
ARM and Aarch is way over my head, so I'll leave that to the people who know what they're doing.
Can't think of anything else.
Can't think of anything else.
Who asked for bluefin or bazzite f40? If we are talking about doing all that. We should just hold this whole thing
How so?
I thought the next steps on the F40 builds would be enabling it in the downstream images, right? Why should we hold?
I thought the next steps on the F40 builds would be enabling it in the downstream images, right? Why should we hold?
we were thinking main first for all of this
like, 1,2,3 for main
then down the line
Because it’s massively unstable prerelease
not 1 main, bazitte, blufin, 2 main bazzite, bluefin, 3, bazzite,bluefin
Ack. So you want to get ARM building first and enable everything downstream at once
right
all main with 1,2,3, then down the line
40 for main, then hwe consolidation for main, then arm for main
then bazzifin, and most of hwe will be gone so ideally less work to just archive those
sorry I should have been more clear on that initially
maybe we have one ublue-os/hardware repo for all of these variants?
That is the current plan to my understanding.
We did some project planning with notes/issues/etc in github today.
Here's an outline of the F40 release: https://github.com/ublue-os/main/issues/529
GitHub
F40 Release · Issue #529 · ublue-os/main
A list of tasks to complete for F40 release. Details can be discussed in comments. add F40 builds to akmods disable kmod builds for negative17 sourced kmods (not yet supporting F40) enable kmod bui...
Note that I'm going to finish the HWE consolidation before adding F40 to any HWE repos