01:11
tg has joined #fedora-riscv
03:07
troglodi1o is now known as troglodito
03:22
davidlt has joined #fedora-riscv
05:28
allison has joined #fedora-riscv
06:00
dtometzki has quit [*.net *.split]
06:00
defolos has quit [*.net *.split]
06:00
skip77 has quit [*.net *.split]
06:00
skip77 has joined #fedora-riscv
06:01
dtometzki has joined #fedora-riscv
06:02
defolos has joined #fedora-riscv
06:53
jcajka has joined #fedora-riscv
09:27
allison is now known as lis_
09:28
lis_ is now known as lis
10:20
<
davidlt[m] >
Conan Kudo: There is a good chance that I might join Fedora Asahi efforts down the road (once M3 happens). I have been tracking blogs, live streams, etc. from Asahi for a year or more. Got to play with M2 for a longer and it's really tempting.
10:20
<
Eighth_Doctor >
hah
10:22
<
davidlt[m] >
I might as well wait for that rumoured 15" M3. That should be interesting one. That's enough time for actual interesting silicon changes, and 15" would be kinda nice.
10:23
<
davidlt[m] >
M1 is relatively cheap, but it doesn't have MagSafe, which I kinda love (from very old Macbook days). // I used to live in Apple eco-system 10+ years ago.
10:23
<
Eighth_Doctor >
the 2021 MBPs have MagSafe
10:23
<
Eighth_Doctor >
that's what I have
10:24
<
davidlt[m] >
I keep joking that I spent 10 years in Windows camp, 10 years in Mac and 10+ years in Linux in professional capacity. (20+ years if we count from the 1st Linux experience).
10:24
<
davidlt[m] >
Yeah, but MBP IIRC is a lot more expensive.
10:25
<
davidlt[m] >
Thus I will wait for M3 and grab a cheaper one to join Asahi club with somewhat broken experience to begin with :D
10:26
<
davidlt[m] >
Lina (and others of course) are doing an epic work.
10:26
<
davidlt[m] >
I mainly watch Hector and Lina livestreams on YouTube in the background.
10:27
<
davidlt[m] >
I could sunset my PineBook Pro 14" :) // AArch64 laptop too.
10:33
<
davidlt[m] >
The issues with broken *.pc files in Fedora continues.
10:34
<
davidlt[m] >
Found another way it could be broken. Apparently white space before the entry makes pkg-conf ignore the PC file.
10:35
<
davidlt[m] >
Already building patched version in Fedora/RISCV.
10:35
<
davidlt[m] >
We also passed 18'000 builds in our F38 Koji tag.
10:36
<
davidlt[m] >
(where dist == f38)
10:36
<
davidlt[m] >
Going back to building GCC 13 bumps made things slower and our build rate dropped a lot.
10:37
<
davidlt[m] >
I will launch a new GCC 13 build again later today. Hopefully Jakub will move towards his more traditional every 7+ days bump.
10:38
<
davidlt[m] >
I still see packages failing due to GCC 13 headers stuff. Surprised those are not fixed yet.
12:33
zsun has joined #fedora-riscv
14:38
zsun has quit [Quit: Leaving.]
14:54
zsun has joined #fedora-riscv
14:58
zsun has quit [Client Quit]
15:52
nirik-gone has quit [Remote host closed the connection]
15:52
nirik-libera has joined #fedora-riscv
16:12
jcajka has quit [Quit: Leaving]
17:00
<
davidlt[m] >
Why libifp is still tagged in Fedora 38?!
17:00
<
davidlt[m] >
The last time it was built was 2022-01-20 18:33:11
17:01
<
davidlt[m] >
There are more than one package that requires libusb-devel dep which is no longer in Fedora.
17:03
<
nirik >
When we branch, all the latest tagged packages are cloned into the new tag... so yeah, that can happen...
17:30
<
davidlt[m] >
nirik: the script that checks for packages to be deprecated don't check all the conditions.
17:31
<
davidlt[m] >
E.g. there are also a bunch of packages in Fedora 38 (and 37 and probably further) that never managed to build.
17:31
<
davidlt[m] >
All the builds failed, and yet it still in Fedora 38 and it gets a bump for a mass rebuild to fail again.
17:32
<
davidlt[m] >
Now that I am <5000 packages away from the in the "1st wave of mass rebuild" these will be popping out more and more.
17:33
<
davidlt[m] >
And as I get sometimes tired I keep asking myself "Why this is not yet fixed?" 😺
17:43
<
nirik >
yes, those are all in the boat of 'never was able to build a srpm' which means koji can't tell what package it was.
17:43
<
nirik >
there's a releng ticket to fix it somehow, but it's not an easy problem
17:45
<
davidlt[m] >
koji list-tagged --quiet --latest f38 ?
17:46
<
davidlt[m] >
The result is empty if the package never managed to build. There was nothing to tag.
17:47
<
davidlt[m] >
Dump two lists. All packages in F38 tag. All packages with latest tagged builds in F38. Sort, diff.
17:48
<
davidlt[m] >
The last time I checked there were 23294 packages in F38 tag, but only 23102 tagged packages.
17:48
<
nirik >
it is tagged when we branch
17:48
<
davidlt[m] >
You don't get me. There was nothing to tag.
17:48
<
nirik >
it might be a fc37 or fc36 package
17:48
<
davidlt[m] >
There was never in Koji history a successful build.
17:49
<
nirik >
yeah, I am in the middle of something, so I am not able to focus here... but perhaps you could provide an example package?
17:49
<
davidlt[m] >
There are such packages.
17:49
<
davidlt[m] >
I could check it tomorrow, but I did mention that on IRC or other Fedora Matrix channel.
17:49
<
nirik >
so, look at libifp
17:50
<
nirik >
it hasn't been built since f36
17:50
<
nirik >
is tagged in f37 and f38
17:51
<
davidlt[m] >
autoconf268 is part of F38, F39, but basically never built.
17:52
<
davidlt[m] >
Well, 2015 for EPEL6.
17:52
<
nirik >
thats probibly an epel only package that didn't get retired for fedora...
17:53
<
davidlt[m] >
Another one docker-swarm
17:53
<
davidlt[m] >
This one never had any builds.
17:53
<
davidlt[m] >
But is available since F21.
17:53
<
davidlt[m] >
Still not disabled.
17:54
<
nirik >
fun. Anyhow, if you have a list of these, file a releng ticket and we can clean them up.
17:54
<
davidlt[m] >
Never had any builds.
17:55
<
davidlt[m] >
Nah, I would need to do a better check (way too tired). I noticed that some actually got a build a couple days ago for example.
17:55
<
davidlt[m] >
So I would need to rule out really fresh packages that just got added.
17:55
<
nirik >
so it should just be retired/removed
17:56
<
davidlt[m] >
But it's relatively easy to check other stuff too. Give me a sec.
17:56
* davidlt[m]
spins yet another container
17:58
<
davidlt[m] >
I noticed that in the last ~4 hours I don't exactly know what I am doing, just making a mess :)
17:58
<
davidlt[m] >
I cannot even write CLI command or proper package names 🌝
18:01
<
davidlt[m] >
DecodeIR never managed to build since F36.
18:03
<
davidlt[m] >
There is about ~150 packages that are still from Fedora 36 era (last build) in Fedora 38 (and most likely in Fedora 39).
18:03
<
davidlt[m] >
So let's say there is 150-200 (or more) packages that are basically dead, but don't show up in retirement emails.
18:03
<
davidlt[m] >
I think.
19:29
davidlt has quit [Ping timeout: 265 seconds]
21:03
iooi has quit [Read error: Connection reset by peer]
21:04
iooi has joined #fedora-riscv
21:36
esv has quit [Quit: Leaving]
21:36
esv has joined #fedora-riscv
22:52
esv has quit [Ping timeout: 268 seconds]