libera/#devuan-dev/ Wednesday, 2018-11-21

LeePenCenturion_Dan: thanks very much.00:23
onefangCenturion_Dan: on this side of the planet, we are always a day early.  B-)02:07
onefangThough I can't turn up to meetings these days, it conflicts with some of my volunteer work.02:10
golinux<Centurion_Dan> golinux: It looks ok, but it would be nice to see it in context with the rest of the theme.07:01
golinuxWell, here's my first stab at it: https://dev1galaxy.org/files/462B25.png07:02
golinuxWill look better when the icons are recolored and there's a proper title bar.07:04
Centurion_Danonefang: pity... it is my volunteer work...07:59
Centurion_DanLeePen: I've been asking if libelogind0 is abi compatible with libsystemd0.  If it is, then debian doesn't need to build a patched polkit at all, and instead replaces libsystemd0 with libelogind0, and for Devuan we do the same.  So can you patch your latest elogind package to create symlinks from libelogind0 to libsystemd0 paths and setup replaces, conflicts etc?09:22
golinuxCenturion_Dan: The stock gnome icons look quite nice on Clearlooks-Phenix-Cinnabar until I can recolor them.09:32
Centurion_Dancool09:32
KatolaZI see you've been busy :)09:53
KatolaZgood stuff :)09:53
KatolaZCenturion_Dan: I guess that's not gonna happen in Debian09:54
_moep_o/09:54
KatolaZwhy should systemd packagers take the burden of packaging elogind?09:54
Centurion_Danthey don't. but libelogind0 is supposedly abi compatible, so we can use it as a replacement for libsystemd0.10:02
KatolaZas we10:03
Centurion_Danfor Debian libelogind can replace libsystemd where they are not using systemd as init.10:03
KatolaZCenturion_Dan: in the end libelogind IS libsystemd...10:04
Centurion_Danit's not, it uses the ABI, but the api is different.10:05
Centurion_Danit handles some things differently.10:05
Centurion_Danbut because it's abi compatable we can simply setup symlinks to the locations of libsystemd0 and put a provides, conflcts and breaks libsytemd0, as an alternative (or replacement) for libsystemd0 (which we then could ban too).10:08
Centurion_Danelogind does the work...10:09
KatolaZlet's build it for experimental then10:11
Centurion_DanI'll see if I can prod LeePen to do it, seen he has just today uploaded and built a new elogind for experimental ;-)10:17
KatolaZyeah I saw that :)10:39
KatolaZ09:56 <+KatolaZ> I see you've been busy :)10:39
KatolaZ^^10:39
KatolaZthat's great10:39
KatolaZCenturion_Dan: I would migrate the new packages we built for unstable straight into beowulf10:41
KatolaZthis might create some breakage, but at least we have a slightly larger user base to test them10:42
Centurion_Danbewulf is so broken, I'm not recommending it anyway.10:47
KatolaZuh?10:50
KatolaZwhy do you say that?10:50
KatolaZI have been using it for about 8 months now10:50
KatolaZnot a single problem10:50
KatolaZif "so broken" is the missing buttons to reboot/shutdown, then that's another story10:51
KatolaZbut it's not at all broken10:51
KatolaZand we should start moving packages there10:51
KatolaZotherwise it will never been released10:51
KatolaZ...10:51
KatolaZsimple equation10:51
KatolaZ:D10:51
KatolaZunstable is frozen at some point in 201610:52
Centurion_DanI agree we should start moving packages.10:53
KatolaZok10:55
KatolaZplease have a look at my proposal in beowulf_todos10:55
Centurion_DanIf you get a chance can you move consolekit2 from experimental to unstable (I need it to build slim) and also move the current elogind to unstable and we can test the version with the patch for replacing libsystemd0 in experimental.10:55
KatolaZCenturion_Dan: if it's in experimental you don't need it in unstable to build for experimental10:56
KatolaZI can move consolekit210:56
KatolaZCenturion_Dan: there is no binary for consolekit2 in experimental10:57
Centurion_Danthe binaries are consolekit.10:58
KatolaZok10:58
KatolaZand we don't have any consolekit in unstable, right?10:59
KatolaZor do we?10:59
Centurion_Danprobably should fix it to either be consolekit (it's a direct continuation but with a new upstream)10:59
KatolaZI guess that's fine11:00
Centurion_Danor to use consolekit2 in the binary and lib names and set Provides to the equivalent consolekit package.11:00
KatolaZlet's keep it like that for now11:00
Centurion_Danok... I'll fix it in the next round...11:01
Centurion_Danwe may need to patch and rebuild policykit to use consolekit2 anyway11:02
KatolaZconsolekit2 is in unstable11:03
KatolaZwill be available in a couple of minutes from pkgmaster11:03
Centurion_DanKatolaZ: I don't think our experimental builds have access to experimental11:03
KatolaZyes they do11:03
KatolaZ:)11:03
KatolaZlook at the binary job11:03
KatolaZif it's built for experimental, the experimental repo is added by default11:04
Centurion_Dan(I have a set of pending issues for jdg to fix this using vendor profiles were we can set all these mappings properly.11:04
Centurion_Dan)11:04
KatolaZanyway, I can move also elogind to unstable11:04
KatolaZthis will probably break things, but it's unstable11:04
KatolaZshall I?11:05
Centurion_DanYou've pushed elogind from unstable into beowulf, or doesn't it have any yet?11:05
KatolaZnope11:07
KatolaZI haven't touched elogind at all11:07
KatolaZthere is only a version in ascii11:07
KatolaZone in ascii-proposed11:07
Centurion_DanI'm more keen to push consolekit2 - that way we can get slim11:07
KatolaZand one in experimental11:07
KatolaZCenturion_Dan: we will have both in the end11:07
Centurion_Danpush the experimental one to unstable then11:08
KatolaZok11:08
Centurion_Danand if it doesn't result in howls of anguish in the next couple of days, we can promote to beowulf too.11:08
KatolaZmoving now11:09
KatolaZdone11:09
KatolaZwill hit the repo soon11:09
KatolaZin theory we should rebuild those unstable packages every week11:09
KatolaZthis would be part of the unstable/testing/stable pipeline11:10
KatolaZI guess we'll get there as soon as unstable is updated11:10
LeePenMorning!11:14
LeePenAFAIK, there are some extra bits in libsystemd over libelogind. That is why11:15
LeePendebian didn't want their pam_elogind to provide pam_systemd.11:16
Centurion_Danyes, but only in the functionality related specifically to systemd and not the abi.  That has been left untouched and provides empty stuns (I think).11:20
Centurion_Dan*stubs11:20
Centurion_Dandebian shouldn't care provided it's abi compatible and is only installed when systemd is not used.11:22
Centurion_DanLeePen: And the benefit is Debian won't have to fork and rebuild packages to support elogind.11:23
LeePenYes I agree. maybe they shouldn't but they did, which is why we have ended up with the -compat package in Debian.11:23
Centurion_DanAnyway, for debian elogind should provide all the functionality we need and we can easier be modified to our needs.11:24
Centurion_Dans/Devuan/debian/11:24
Centurion_Dandooh!11:25
LeePen:D11:25
Centurion_Danwhat I meant to say was:11:25
Centurion_DanAnyway, for Devuan elogind should provide all the functionality we need and we can easier be modified to our needs.11:25
LeePenYes. Ultimately, once elogind is in debian (it is still waiting in incoming)11:26
LeePenthe way I see this is11:26
LeePenDevuan uses elogind directly from debian and11:26
Centurion_Danso a simple patch to add the symlinks and control file changes.11:26
LeePenwe have a compatibility package that we control to manage the symlinks and dependencies11:27
Centurion_DanWhy.  Debian will need it's own special case, lets not bother waiting for their butt hurt maintainers who see systemd utopia slipping away to finally let go...11:27
Centurion_Dan:-P11:28
KatolaZanyway, I guess we need a working polkit series of packages in unstable11:29
Centurion_Danforget compat packages in Devuan.  When Debian get there shit together they can either fork and patch our packages.11:29
LeePenOK. Have you tried a libsystemd/elogind symlink? Does it work or break anything?11:29
Centurion_Dan... or do it on their own.11:29
Centurion_DanNot yet.11:30
LeePenI have tried a pam_systemd/pam_elogind symlink and I think it was OK. That is also in the -compat package we have got on Debian.11:30
Centurion_DanKatolaZ has moved your elogind package to unstable so we can build for experimental and test from there.11:30
LeePenGreat. Thanks to you both.11:31
Centurion_DanThanks to you for taking on a package11:31
* Centurion_Dan high fives LeePen11:32
LeePenIt is good to help11:32
Centurion_DanAnyway I need an earlier night... too many runs past 2am already this week..11:33
LeePenOK. Sleep well. Morning here.11:34
Centurion_DanLeePen: btw, if you create an issue with the subject "build" and assign it to autobuild and set a label: "experimental" then that will kick off a build ;-)11:46
LeePenThanks. I will try it next time.11:47
Centurion_Danfor packages where you are maintainer in git.d.o11:47
LeePenDoes it work if you are only Uploader?11:48
KatolaZdon't think so, unless it has been fixed12:10
LeePenNo worries.12:16
KatolaZtry that anyway12:19
LeePenI will next time.12:30
KatolaZoki12:30
plasma41ping parazyd19:05
plasma41https://lists.dyne.org/lurker/message/20181121.164811.f1e54e85.en.html19:06
KatolaZplasma41: jessie cannot start on pi3B release 201719:07
KatolaZsince jessie was released before that model was put on the market...19:07
KatolaZ:D19:07
plasma41 I tested the Raspberry PI 2 image in question and I can concur that the OS doesn't start, no video other than a black screen.19:08
plasma41KatolaZ: I'm corroborating the report about the RP2 ASCII image.19:10
KatolaZok great19:11
KatolaZthat needs to be fixed19:11
plasma41I'm frankly astounded this wasn't caught sooner.19:13
KatolaZit looks strange to me as well, TBH19:14
golinuxCould someone who understands the bug please respond thanking him for letting us know?19:18
plasma41I suppose I can do that.19:24
golinuxplasma41: Thanks!19:39
plasma41parazyd KatolaZ: I just tested the RP2 ASCII Release Candidate image and it suffers the same issue.19:55
KatolaZplasma41: there are two versions of rpi2B19:57
KatolaZone is 32bit19:57
KatolaZthe other one is 64bits19:57
plasma41I thought the first pi to support 64 bit was the 319:58
KatolaZnope19:58
KatolaZrpi2b ver 1.2 has the same processor as the rpi319:58
plasma41My Pi 2s are both V1.119:59
KatolaZeh19:59
KatolaZI guess parazyd built the image for the ver 1.219:59
KatolaZwhich is the newer one19:59
plasma41:-/19:59
KatolaZ(just guessing)19:59
KatolaZanyway, you could use the rpi1 image there19:59
parazydThe pi2 image always booted on my Pis20:00
plasma41rpi1 image doesn't support hard float20:00
KatolaZyeah parazyd20:00
KatolaZit must be a ver 1.220:00
parazydTry getting a serial console and read what's up.20:00
KatolaZprobably20:00
plasma41hello parazyd20:00
parazydo/20:01
KatolaZo/20:01
KatolaZ \o/20:01
KatolaZ \o\20:01
KatolaZ /o/20:01
KatolaZ |o|20:01
KatolaZ /o\20:02
KatolaZ /.\20:02
plasma41parazyd: I don't have a serial cable for the pi20:02
plasma41The ASCII Beta rpi2 image boots on my pi20:04
parazydAnd the current one does not?20:07
plasma41Correct20:07
plasma41<KatolaZ> rpi2b ver 1.2 has the same processor as the rpi320:09
plasma41<KatolaZ> I guess parazyd built the image for the ver 1.220:09
parazydWhat revision of the board do you have?20:09
plasma41parazyd: is that right ^20:09
plasma411.120:09
parazydI don't build the stuff for any specific rev.20:10
parazydplasma41: Do you know the date of the beta image?20:11
plasma412018-02-2820:13
plasma41Is when I downloaded it20:13
parazydThanks20:13
plasma41The beta released on Valentines Day if memory serves20:13
parazydLiterally nothing changed except the kernel version...20:15
parazydHmm20:15
parazydI can build a new one for you tomorrow if you like.20:16
parazydThen we can debug it. I have some Pi2 somewhere as well.20:16
plasma41You didn't accidentally start compiling the image as 64bit perhaps? That would work fine on V1.2 but not on V1.120:18
parazydNope20:18
plasma41parazyd: What version is your rpi2?20:19
parazydhttps://git.devuan.org/sdk/arm-sdk/blob/master/boards/raspberry-pi2.sh#L2820:19
parazydplasma41: I'll have to check20:19
parazydIt's not at my house.20:19
KatolaZparazyd: we also need to add "--no-merged-usr" to the debootstrap options in devuan-sdk20:21
KatolaZthe latest debootstrap in unstable has --merged-usr enabled by default...20:22
parazydACK20:28

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