#meeting:fedoraproject.org log

<@meetbot:fedora.im>

14:56:44

Text Log: https://meetbot.fedoraproject.org/meeting_matrix_fedoraproject-org/2025-05-27/fedora-commops-2025-05-27.2025-05-27-13.58.log.txt

<@nirik:matrix.scrye.com>

15:00:49

!startmeeting riscv-sig

<@meetbot:fedora.im>

15:00:51

Meeting started at 2025-05-27 15:00:49 UTC

<@meetbot:fedora.im>

15:00:51

The Meeting name is 'riscv-sig'

<@kashyapc:fedora.im>

15:01:10

!hi

<@nirik:matrix.scrye.com>

15:01:11

morning everyone. Who's around for a riscv sig meeting?

<@zodbot:fedora.im>

15:01:11

Kashyap Chamarthy (kashyapc)

<@abologna:matrix.org>

15:01:37

!hi

<@zodbot:fedora.im>

15:01:48

None (abologna)

<@kashyapc:fedora.im>

15:01:50

Morning; /me is around

<@nhanlon:beeper.com>

15:01:55

!hi

<@zodbot:fedora.im>

15:01:56

Neil Hanlon (neil) - he / him / his

<@kashyapc:fedora.im>

15:02:18

abologna: Zodbot tells us you don't have a name (see "None")

<@abologna:matrix.org>

15:02:38

yeah, I never quite figured out how to set that up

<@abologna:matrix.org>

15:02:56

or maybe it wanted me to provide additional information that I wasn't comfortable adding

<@kashyapc:fedora.im>

15:03:00

Well, sorry for you, None.

<@nirik:matrix.scrye.com>

15:03:18

it does that when you set 'private' for info in the account system

<@nirik:matrix.scrye.com>

15:03:24

so the bot cannot get your name. ;)

<@abologna:matrix.org>

15:03:39

there you go, I didn't feel like opening the account up

<@nirik:matrix.scrye.com>

15:03:51

(can you tell the bot is written in Python ? ;)

<@kashyapc:fedora.im>

15:04:01

Ah, I see.

<@abologna:matrix.org>

15:04:14

my private information is `None` of its business ;)

<@nirik:matrix.scrye.com>

15:04:25

anyhow, I am not sure what topics we had today... flock and devconf are coming up...

<@kashyapc:fedora.im>

15:04:25

Ba-dum-tss

<@abologna:matrix.org>

15:05:00

I guess whether there's any progress on the Koji server migration

<@abologna:matrix.org>

15:05:24

but it seems that davidlt hasn't showed up

<@kashyapc:fedora.im>

15:05:37

Yeah, I forgot what's the status there. All I know is, we have this: https://riscv-koji.fedoraproject.org

<@irradiated_lemming:matrix.org>

15:05:44

Its more likely that the event bot will get banned

<@nirik:matrix.scrye.com>

15:06:15

yeah, I think folks have been focused on other things lately...

<@nirik:matrix.scrye.com>

15:06:22

I expect this will heat up after devconf...

<@kashyapc:fedora.im>

15:06:47

Sure; I'm not going there, though. I wonder if DavidA shows up at Flock; he said he was looking at flights

<@nirik:matrix.scrye.com>

15:07:49

yeah, I won't be at devconf either. But I will be at flock!

<@abologna:matrix.org>

15:08:18

I'm not going to any conference either. I plan to have switched my focus back to Fedora (the Koji situation in particular) by the time the next SIG meeting comes around

<@abologna:matrix.org>

15:08:53

that might be primarily bothering David ;)

<@davidlt:matrix.org>

15:09:07

!hi

<@zodbot:fedora.im>

15:09:09

David Abdurachmanov (davidlt)

<@abologna:matrix.org>

15:09:20

speak of the David!

<@davidlt:matrix.org>

15:09:55

Sorry, I somehow lost tack of time 😄

<@kashyapc:fedora.im>

15:10:03

That's fine

<@davidlt:matrix.org>

15:10:29

Most focus (from my side) has been updating F42 and fixing some leftovers I dropped before.

<@nirik:matrix.scrye.com>

15:10:31

easy to do. :)

<@davidlt:matrix.org>

15:10:44

In general things are in good shape, actually in better compared to existing repositories.

<@kashyapc:fedora.im>

15:11:11

What do you mean "existing repos"?

<@davidlt:matrix.org>

15:11:20

I think, once there is a new GCC (<2 days left), and the final v6.14.9 kernels show it's kinda time to update repos.

<@davidlt:matrix.org>

15:11:36

Existing dist-repos from which initial disk images were cooked.

<@kashyapc:fedora.im>

15:11:47

Ah, like that.

<@davidlt:matrix.org>

15:12:01

I wanted to get the newer Rust version too, but somehow that's not yet submitted for F42 under Bodhi.

<@davidlt:matrix.org>

15:12:29

Just hoping nothing large and breaking shows up in ~2 days before GCC lands.

<@davidlt:matrix.org>

15:12:53

Otherwise we also have F42 Spacemit tags, for X60/K1/M1 based SBCs. Jason Montleon has cooked the kernel already.

<@nirik:matrix.scrye.com>

15:13:04

rust was untagged

<@nirik:matrix.scrye.com>

15:13:10

it had problems on ppc64le

<@nirik:matrix.scrye.com>

15:13:24

https://pagure.io/releng/issue/12751

<@abologna:matrix.org>

15:13:38

we discussed building updated images for F42 a while ago. is that something that we're still considering? I wonder what would be the advantage of that compared to having people flash the initial F42 image and just running `dnf update`

<@davidlt:matrix.org>

15:13:43

There is a new package from spacemit firmware, and it has sub-package for dracut module. That dracut package will have to be included while building disk images to get firmware land in initramfs.

<@davidlt:matrix.org>

15:14:03

Avoid downloading tons of data 😄

<@abologna:matrix.org>

15:15:10

is riscv.rocks hit by big bills if people download too much? or are you thinking of their convenience?

<@davidlt:matrix.org>

15:15:18

In this case we will get Spacemit images too.

<@abologna:matrix.org>

15:15:43

right, spacemit images showing up would be decent enough justification to just respin the world

<@davidlt:matrix.org>

15:16:04

Convenience mainly, and I like to work "snapshot" based as I cannot guarantee that repos is always in a working condition.

<@davidlt:matrix.org>

15:16:28

Well, updates all the boards at once in my cases does degrade YouTube experience for some minutes 😆

<@abologna:matrix.org>

15:16:50

O:-)

<@davidlt:matrix.org>

15:16:54

(not really a problem)

<@nirik:matrix.scrye.com>

15:17:07

mission critical. ;) But yeah, nice to not have to download more.

<@davidlt:matrix.org>

15:17:16

(or I could finally upgrade to 2.5Gbps line, just kidding 😉)

<@kashyapc:fedora.im>

15:17:39

(I just got called yesterday that they're gonna replace copper with fiber at my place next month)

<@davidlt:matrix.org>

15:18:23

So unless something is critical I don't really expect updates in existing Koji. We will start breaking things apart, I guess.

<@nirik:matrix.scrye.com>

15:18:25

They are supposed to be putting in fiber here this week. ;) Will be wonderfull once it's working.

<@davidlt:matrix.org>

15:19:05

So if something is broken in F42 ping sooner than later as it will be some time before builds start again.

<@nirik:matrix.scrye.com>

15:20:21

Sounds good.

<@nirik:matrix.scrye.com>

15:20:32

Any other topics today? Or shall we call it a short meeting?

<@nirik:matrix.scrye.com>

15:20:48

There was also a lot of announcements around Red Hat summit...

<@rwmj:matrix.org>

15:20:52

Sorry for missing it :-(

<@kashyapc:fedora.im>

15:20:54

davidlt: One quick: the LLVM failures -- did you look any further?

<@davidlt:matrix.org>

15:21:02

Yes, RHEL for RISCV DP.

<@davidlt:matrix.org>

15:21:08

No, I haven't.

<@kashyapc:fedora.im>

15:21:35

Okay, don't worry, I'll sort it out with AlexB; and report back.

<@nirik:matrix.scrye.com>

15:21:59

also https://blog.centos.org/2025/05/initial-centos-support-for-risc-v/

<@davidlt:matrix.org>

15:23:06

Yeah, but nothing yet on CentOS Koji.

<@nirik:matrix.scrye.com>

15:23:18

yeah

<@kashyapc:fedora.im>

15:23:21

Yeah, they have no hardware yet, IIRC

<@davidlt:matrix.org>

15:23:21

We will have to wait to see what lands later on.

<@nhanlon:beeper.com>

15:23:39

question on that.. has anyone found the ISA SIG matrix channel?

<@davidlt:matrix.org>

15:24:00

Ah yeah, I was trying to join their SIG Matrix channel but was not able to do that.

<@davidlt:matrix.org>

15:24:03

I don't think it exist.

<@nhanlon:beeper.com>

15:24:05

cause I also went back and looked at my irc logs and there haven't been (any) meetings in #centos-meeting for--well--months

<@kashyapc:fedora.im>

15:24:24

I don't know; is this CentOS-specific channel?

<@davidlt:matrix.org>

15:24:27

Exactly. The link to their SIG stuff doesn't exist, I think.

<@nhanlon:beeper.com>

15:24:34

so it feels a bit disingenuous to somehow invite to join a SIG that doesn't exist

<@nirik:matrix.scrye.com>

15:24:47

yeah, unclear. There is a centos space under the fedora server... but only a few sigs so far

<@kashyapc:fedora.im>

15:25:13

They say it is this here: https://sigs.centos.org/isa/communication/channels/

<@davidlt:matrix.org>

15:25:17

I think there were meeting mentioned, but again I was not able to find information. Same as with Matrix channel.

<@abologna:matrix.org>

15:25:23

I believe the SIG exists, but it might have been a bit dormant. there might be some more activity coming up due to RISC-V

<@kashyapc:fedora.im>

15:25:37

Nothing disingenuous about it :)

<@kashyapc:fedora.im>

15:25:37

<@kashyapc:fedora.im>

15:25:37

Neil Hanlon: They also say, "For async discussions and announcements we generally use the centos-devel mailing list."

<@davidlt:matrix.org>

15:25:38

`#centos-isa:fedoraproject.org does not exist.`

<@nirik:matrix.scrye.com>

15:25:44

yeah... might be something to bring up on centos-devel mailing list?

<@abologna:matrix.org>

15:26:45

maybe the channel existed back in the IRC days and was not moved properly to Matrix due to the lack of activity? I'm just speculating

<@kashyapc:fedora.im>

15:26:50

Yeah, mailing lists are still a thing

<@davidlt:matrix.org>

15:27:01

Page not found

<@davidlt:matrix.org>

15:27:01

```

<@davidlt:matrix.org>

15:27:01

This page either doesn't exist, or it moved somewhere else.

<@davidlt:matrix.org>

15:27:01

I pressed on `centos-devel` and got:

<@davidlt:matrix.org>

15:27:01

```

<@davidlt:matrix.org>

15:27:01

<@nirik:matrix.scrye.com>

15:27:08

the centos-devel list definitely exists (and gets posts)

<@davidlt:matrix.org>

15:27:16

https://lists.centos.org/mailman/listinfo/centos-devel

<@nirik:matrix.scrye.com>

15:27:31

https://lists.centos.org/hyperkitty/list/devel@lists.centos.org/

<@abologna:matrix.org>

15:27:51

this all smells like outdated information

<@davidlt:matrix.org>

15:27:54

Yeah, links are just broken in all these places, I guess.

<@nirik:matrix.scrye.com>

15:27:58

They moved to mailman3 and I bet some links are still mailman2 set

<@nhanlon:beeper.com>

15:28:30

I will write to centos-devel and see what we can find out, as I'm keenly interested in being involved 🙂

<@kashyapc:fedora.im>

15:28:54

Cool.

<@davidlt:matrix.org>

15:29:23

Neil Hanlon: do you know if Rocky plans to rebase existing riscv64 work on CentOS Stream 10 once riscv64 bits land?

<@abologna:matrix.org>

15:30:22

davidlt: most of the bits have already landed, save a handful of packages

<@nhanlon:beeper.com>

15:30:32

yeah it's mostly all there

<@davidlt:matrix.org>

15:31:58

Well, maybe I picked a wrong package (kernel), and I don't see there config yet.

<@davidlt:matrix.org>

15:32:06

(I didn't check kernel-ark too)

<@abologna:matrix.org>

15:32:16

you did pick one of the handful :) there's a pending MR

<@nhanlon:beeper.com>

15:32:16

https://gitlab.com/redhat/centos-stream/rpms/kernel/-/merge_requests/838 i think is the biggest (which really needs to be against the source repo)

<@davidlt:matrix.org>

15:32:34

GCC changes are definitely in.

<@abologna:matrix.org>

15:33:09

kashyapc: you might want to close that one, it's awfully outdated

<@kashyapc:fedora.im>

15:33:23

abologna: I was _just_ looking at it!

<@nhanlon:beeper.com>

15:33:34

😃 thanks !

<@abologna:matrix.org>

15:33:42

https://gitlab.com/cki-project/kernel-ark/-/merge_requests/3863

<@abologna:matrix.org>

15:33:45

this is the current one

<@kashyapc:fedora.im>

15:34:09

Yes, kernel-ark is the right one; Neil Hanlon

<@davidlt:matrix.org>

15:34:11

Anyways, seems that we should get a full family sooner than later (Fedora, CentOS Stream, Rocky, Alma, RHEL DP).

<@abologna:matrix.org>

15:34:32

we have a few more in flight, e.g. https://gitlab.com/redhat/centos-stream/rpms/python-psutil/-/merge_requests/12 which was merged just earlier today

<@davidlt:matrix.org>

15:35:14

heh, I did similar change recently in Fedora.

<@davidlt:matrix.org>

15:35:25

This should go as-is as MR/PR to Fedora dist-git too.

<@abologna:matrix.org>

15:35:53

https://src.fedoraproject.org/rpms/python-psutil/pull-request/30

<@abologna:matrix.org>

15:36:00

way ahead of you :)

<@kashyapc:fedora.im>

15:36:29

Nice!

<@davidlt:matrix.org>

15:36:36

My kung fu is getting weak 😢

<@abologna:matrix.org>

15:37:22

well you're busy keeping up with the onslaught of changes coming in through Fedora, I don't have to worry about that...

<@abologna:matrix.org>

15:37:33

and it was just coincidence ;)

<@davidlt:matrix.org>

15:37:38

I revived qt5-qtwebengine (thanks Arch Linux!) and qt6-qtwebengine is cooking too (thanks OpenSUSE!).

<@davidlt:matrix.org>

15:38:02

That should unblock quite a good number of KDE apps (and improve repo statistics too).

<@abologna:matrix.org>

15:38:16

lovely!

<@jmontleon:fedora.im>

15:38:53

I'm trying to have a look at scipy and python-pyunicorn. It seems like there is something going on with pytest; something strange going on with some of the tests.

<@jmontleon:fedora.im>

15:39:31

seems like some tests just hang 'forever'. Not really sure what's going on.

<@davidlt:matrix.org>

15:40:38

Is this that thread pool exhaustion thing you posted before?

<@jmontleon:fedora.im>

15:41:42

It looks like that, but I'm not sure that's what is going on exactly. It seems like Failed tests may have a hand in causing problems

<@kashyapc:fedora.im>

15:41:44

Jason Montleon: Might want to check with Trevor Gamblin from Baylibre. He and a few folks from Rivos Inc. are looking at some of these scientific computing and ML Python library packaging. (I learnt it from his talk at RISC-V Summit)

<@abologna:matrix.org>

15:41:56

I'm going to step away from the keyboard to grab some coffee. I'll read the scrollback later. bye!

<@jmontleon:fedora.im>

15:42:29

Thank you kashyapc . Maybe they have an idea what's going on

<@kashyapc:fedora.im>

15:42:40

Yeah, I'll send you his email ID.

<@davidlt:matrix.org>

15:42:58

I have tried running these on different boards to check if this is somehow kernel related or/and hardware.

<@davidlt:matrix.org>

15:43:07

Results are the same on various combinations.

<@davidlt:matrix.org>

15:43:58

There are a couple of packages that segfault GCC. I forgot the names (and that's bad), but hoping that the new GCC could resolve some of that.

<@davidlt:matrix.org>

15:44:14

I checked every GCC change in this update, a good number of them are ICE related.

<@davidlt:matrix.org>

15:44:37

Ah, one of packages is Python 3.14 Beta 1.

<@davidlt:matrix.org>

15:45:46

We need to keep an eye on Python 3.14 as Fedora 43 probably will ship that.

<@davidlt:matrix.org>

15:45:59

It's always fun to bootstrap a new Python version 😄

<@davidlt:matrix.org>

15:46:27

If mistakes are made everything will get broken, including DNF itself 😄

<@davidlt:matrix.org>

15:46:41

https://fedoraproject.org/wiki/Changes/Python3.14

<@davidlt:matrix.org>

15:47:36

I probably have ~200 notes related to Fedora 43 in my inbox to prep before (hopefully) mass rebuild starts.

<@nirik:matrix.scrye.com>

15:49:59

The python folks are good about mass prebuilding, but they don't do riscv yet of course...

<@kashyapc:fedora.im>

15:50:47

At least NumPy (being the main scientific computing package) does test on RISC-V, apparently.

<@kashyapc:fedora.im>

15:51:01

Jason Montleon: Sent an email to Trevor with you in Cc.

<@davidlt:matrix.org>

15:51:32

nirik: churchyard has an account and has helped during one of bootstraps IIRC

<@davidlt:matrix.org>

15:52:06

Once we are in a new Koji we might get more help. We will see.

<@nirik:matrix.scrye.com>

15:52:34

cool.

<@jmontleon:fedora.im>

15:52:44

Thank you kashyapc !

<@nirik:matrix.scrye.com>

15:52:50

ok, if nothing more, lets close out and head back to #riscv:fedoraproject.org to continue?

<@nirik:matrix.scrye.com>

15:52:59

!endmeeting