04:20 <chadmed_> lmfao was wondering why the built iso is so much bigger than the normal one

04:21 <chadmed_> m1n1 pulls in asahi-sources unconditionally :p

05:11 <chadmed_> hm we may need to slightly uhhh... *abuse* the ebuild versioning schema to make this work

05:12 <chadmed_> dist-kernel_PV_to_KV can't handle _p* versions at all, but ignores ebuild revisions

05:12 <chadmed_> so maybe instead of using _p* to denote the asahi tag we just use -r*

05:13 <chadmed_> at least until i have time to look at fixing dist-kernel_PV_to_KV, if there's even interest in doing so

05:16 <sam_> i dont see why not

05:16 <sam_> although i'd have to maybe check it more closely

05:16 <sam_> but like, im not aware of any immediate reason not to

05:28 <chadmed_> as in, no reason not to switch from _p to -r?

05:30 <sam_> sorry, I mean I don't see why we can't fix dist-kernel_PV_to_KV

05:31 <chadmed_> oh yeah it seems pretty trivial

05:31 <chadmed_> might actually do it tonight

05:31 <chadmed_> how do we go about submitting eclass patches? github or the -dev mailing list?

05:31 <sam_> dev ML but you can do gh first if you want to get some initial feedback

05:31 <chadmed_> ack

05:47 <chadmed_> actually it might just be better to fix catalyst's dracut call to deal with the current eclass behaviour

05:47 <chadmed_> obviously typical livecd builds wont be using unstable _p* kernels so this probably never came up before

05:47 <chadmed_> but the way it passes --kver to dracut is _obviously_ wrong given the eclass's behaviour

07:27 <chadmed> ok yeah the easier thing to do was just fix catalyst

07:28 <sam_> first time anyone has said that

07:32 <chadmed> :p

07:38 <chadmed> i mean it could be that my "fix" is entirely wrong...

07:45 <chadmed> https://github.com/chadmed/catalyst/commit/91da60146a32392cd9bb7c9b37c9f79f33df559e

07:45 <chadmed> if you're curious

07:53 <chadmed> also https://github.com/gentoo/catalyst/pull/15

08:00 <jannau> just for completeness: using -r for asahi tags has the disadvantage that it would make actual revisions of the ebuild harder. I think there is precedent in using _p is other kernel packages but probably nobody has ever tried building live-cd with that

08:13 <chadmed> jannau: yeah, no need now anyway since kmerge.sh is "fixed" :)

08:13 <chadmed> assuming releng are happy with the patch and merge it. if they dont im happy to keep doing builds locally since ive got tooling for it now but still cant really host them long term

09:42 <u154ss> @chadmed Give me a shout when your Gentoo USB .ISO is ready and I will give it a whirl..

09:43 <chadmed> holy fuck having this working just saved my arse (excuse the french)

09:43 <chadmed> accidentally rm -rfed /* and for some reason --no-preserve-root isnt a thing in gentoo?

09:47 <j`ey> chadmed: optimised that argument out, to save disk space

09:49 <chadmed> hehe wouldnt surprise me

09:50 <chadmed> bless btrfs though, my /etc is totally intact and /bin is ~mostly

11:08 <mps> with kernel 6.9.6_p1 and iwd 2.18 I've got this message 'Jun 25 13:03:48 m1pro daemon.info iwd: CMD_SET_CQM failed: Not supported'

11:08 <mps> also, dmesg show '[ 12.417902] ieee80211 phy0: brcmf_dongle_roam: WLC_SET_ROAM_DELTA error (-52)'

11:09 <mps> anyone else have problem with iwd and 6.9.6_p1 kernel?

11:10 <mps> didn't noticed any problem with kernel 6.9.5_p1

11:11 <chadmed> mps: i dont think that error should take out your connection

11:30 <mps> chadmed: hm, maybe I should reboot access point. uptime on it shows 97 days

17:50 <mps> hm, now iwd works fine for about 6 hours. bad spot or EMF noise maybe were issues