libera/#devuan-dev/ Wednesday, 2020-01-29

rrqgnu_srs1: done06:49
golinuxPad for this week's meet is here: https://pad.dyne.org/code/#/2/code/edit/hG8t4ng49HNbee4POm5nvbW2/07:20
gnu_srs1rrq: Thanks. Pushed. Maybe somebody can issue a build.10:42
LeePengnu_srs1: Maybe you could fix the Vcs-* urls first?11:02
gnu_srs1LeePen: somebody already issued a build, rrq?, sorry.11:03
LeePenYes, I see.11:03
gnu_srs1OK, I did not look ate the Vcs entries, will fix.11:03
LeePenMaybe fix them anyway and it will be there for the next version.11:04
gnu_srs1LeePen: Do you mean fix and push, not build? Anything else?11:09
LeePenCertainly fix and push.11:09
LeePenI didn't notice anything else.11:10
LeePenIf you want a new build you will have to use a new +devuanN changelog entry.11:10
gnu_srs1LeePen: The I have to increase the devuan version, right?11:11
gnu_srs1Then**11:11
LeePenI see i386 failed too. Just looking why.11:11
LeePenSorry, ppc64el failed:11:12
LeePenErr:2 http://localhost/devuan/merged unstable/main ppc64el Packages11:13
LeePen10:00:59   File has unexpected size (8126940 != 8126728). Mirror sync in progress? [IP: ::1 80]11:13
LeePenIt should work again on a retry.11:13
rrqalso an earlier "E: Error calling setupterm: 0" ??11:14
gnu_srs1Shall I increase the devuan version?11:20
gnu_srs1relocating. BBL11:33
gnu_srs1Sorry, I just saw "use a new +devuanN changelog entry.": Will do :D11:37
LeePenrrq: Yes, that is always there. I have never worked out what it meant!11:42
LeePengnu_srs1: did you see #383?12:20
gnu_srsLeePen: (relocated) Yes I did. But did not get any notice at all about it. Found it by searching bugs.devuan.org.15:52
gnu_srsI've seen this once before when up/downgrading 3.2.2-13, 2.2.7-6 and 3.2.9-3. But that was a transient error.15:57
gnu_srsI'll ask the OP about kernel, release (ascii has 3.2.2.13) etc.15:57
gnu_srsI think we maybe have to add -k <kernel-version> to update-initramfs for eudev.postinst. Any ideas on how to get kernel-version easily?16:35
gnu_srsSee also in #devuan.16:35
masongnu_srs: There's always "all"16:35
masonupdate-initramfs -c -k all16:36
masonThat might even be the most reasonable thing to do.16:36
cosurgignu_srs: uname -a16:47
gnu_srscosurgi: thanks, Well see if we will use -u -k $(uname-a) (correct?) or -c -k all (maybe a little time-consuming?)16:50
masongnu_srs: If they've installed eudev, do we want other kernels to boot without it?16:52
masonMaybe they do, but it's worth asking.16:52
gnu_srsmason: update-initramfs -u is the same as used by Debian, obviously a bug.16:56
gnu_srsRegarding having an initrd.img image or not I dunno :(16:57
masonWell. -u is okay if you want the current one, so that's probably also reasonable, but the question of if there's critical userland stuff outside of the initramfs is still there, and I'd have to look to know for sure.16:58
gnu_srsmason: I think upgrading an existing initrd.img with update-initramfs -u _is_ a bug when you are installing  eudev for a new kernel.17:01
masonOh, if it's for a new kernel, then yes, it would be a bug.17:01
fsmithredgnu_srs, 'update-initramfs -u -k all'17:04
fsmithredoh, now I see you already said that17:16
masonwell, -c -k to start clean17:21
fsmithredyeah, I guess that's needed for a postinst script17:22
fsmithredfor the kernel postinst17:22
fsmithredshouldn't be needed for eudev postinst - any installed kernels will already have an initrd (unless they deliberately don't)17:23
masonAlright, that's reasonable.17:25
gnu_srsLeePen: rrq: I've pushed 1.56+nmu1+devuan2 fixing the Vcs-* entries. Can we rebuild please?17:26
gnu_srsNext would be to take a look at debian-installer. No promises though. LeePen: Can you send your buster to beuwulf script output again.17:29
gnu_srsfsmithred: mason: So update-initramfs -u -k all is the right way to go, agreed?17:31
fsmithredyes, I believe so17:31
fsmithred(and that's what I do in the live installer)17:32
masongnu_srs: Yeah, seems reasonable to me.17:33
fsmithredok, I just did a search in /var/lib/dpkg and all I see is 'update-initramfs -u'17:34
fsmithredoh, I see one with -a17:34
fsmithredwhich is undocumented17:35
LeePengnu_srs: We don't really think we need debian-installer. base-installer has been split from it into a separate source. It only contains Debian documentation AFAIK. That is why we haven't bothered with it.17:45
gnu_srsI see one problem here: eudev.preinst has: check_kernel_features() errs out: Since release 198, udev requires support for the following features in17:49
gnu_srsthe running kernel: looking at /proc/kallsyms but eudev.postinst is issuing update-initramfs ..., too late.; When is  /proc/kallsyms created?17:49
gnu_srsLeePen: Ok, let's forget debian-installer for now then.17:51
gnu_srsLeePen: How come that I did not get noticed about #383?18:40
masonThe "you have chosen to use sysvinit" and "you have chosen to use openrc" reminder screens in the installer are unfortunate.20:50
golinuxWhy are you objecting to a confirmation of a user's choice?21:04
masongolinux: Reminder text is not standard practise, and the user's already had a dialog where presumably they knew what they were choosing, since they had to choose it explicitly. Nowhere else do we have a dialog that demands interaction to remind a user of what they just chose.21:16
masonThe closest we get is asking for clarifying settings where needed.21:16
golinuxmason: At my age, reminders are welcome.  ;)21:45
masonHonestly, for the default install I wouldn't even ask for an init system. That's the very definition of an expert install.21:46
masonBut if it's there, the only interactions we insist on should be necessary ones.21:46
golinuxSometimes I can't remember what I did 5 seconds earlier.  That's a bit of an exaggeration but it FEELS like that sometime.21:46
masonSomething that could be useful and has some precedent is a summary screen before committing to a batch of options. That's done sometimes.21:47
masonYeah, I'm there too. Heh.21:47
masonThat's one of the reasons why I love IRC. Context is always there.21:47
golinuxWell, talk to us about it on a Thursday after Beowulf is released21:48
masonMm.21:48
masonYeah, it's not a huge thing in the grand scheme of things. Just unusual and perhaps something deserving of polish.21:48
masonI should file a bug for it. I'll see about doing that when I free up.21:50
golinuxMaybe see you tomorrow?21:51
masonHoping so. There are no conflicting meetings on the calendar yet.21:52
masonI hope to have a couple of deliverables ready.21:52
rrqmason: choose-init is already updated to avoid the confirmation interaction.21:55
masonAh, cool.21:55
rrqinstead it has an added slight delay upon installing the chosen init, as way of confirming the choice.21:56
masonrrq: You select it, and it sits with that on-screen for a moment?21:56
masonHow does the delay manifest? That's an interesting HCI choice, anyway.21:57
rrqno the next step, which is a progress bar for installing the choice .. or just "do nothing" for the sysvinit choice which already is installed21:57
masonhrm21:58
rrqwell, the progress bar shows up but jumps to 100% immediately21:58
rrqyou have some trial isos at https://borta.rrq.id.au/21:58
masonI hadn't tried them yet as I was waiting for an amd64, but for testing I'll snag one and give it a run22:00
masonThank you.22:00
masonApropos of little, creative abuse of here-docs: https://bpaste.net/G3MA22:01
rrqyeah I've got delayed by a senior moment with someone gifted me with an "rm -r /*" .. almost recovered now22:02
masonArgh. That's quite a gift.22:02
rrqshows that one shouldn't assume about other peoples Makefiles :)22:04
onefangIt's a small gift, but it keeps on "giving".22:04
masonWas it at least their fat-fingering a variable so it was $FOO/*? Or was it actually malicious?22:04
masonAlso! Never build things as root if you can help it.22:05
rrqyes, it was like $FOO/* .. not as root so only my own stuff; most had been checked in somewhere, so the damage was mostly cosmetic. but it takes a while to swallow the pride.22:07
masonYeah. I'm sorry. That sucks. :/22:08
yetiI was told that the son of someone I know has hacked himself and deleted all his own data... he told his mom it was for research... but I think someone told him "hack me! I'm on 127.0.0.1"22:13
masonThat seems like divine justice.22:15
yetihis real life data was gone... no data on a system for experiments...22:15
yetibut i didnt contradict his mother's belief that it were "for research"22:15
yetino need for estryong her picture of her son22:16
yetiping -c1 2130706433 # octal works too... that can help to disguise localhost a bit22:21
masonI don't see disguising localhost as being useful.22:29
gnu_srsHi again. Any other package you need help with for releasing beowulf? (In addition to resolving the eudev intrd.img kernel issues)22:45
yetimason: hack me @01770000000122:51
yeti:-Þ22:51
gnu_srsLeePen: relocating made by britney link being lost: cannot find it by googling. Thanks!23:04
rrqyou mean /britney on pkgmaster?23:44

Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!