libera/#devuan-dev/ Sunday, 2019-03-17

rrqfsmithred: is there an issue with the debus -devuan2 packages?02:25
rrqdebus = dbus02:25
fsmithredrrq, I just installed the new dbus a little while ago, so I haven't noticed any problems02:26
fsmithredapt-get update && apt-get install dbus02:26
rrqok. and again I'm "surprised" that ascii changes rather than using ascii-backports. I guess my idea about "stable release" needs un update.02:27
fsmithredascii-proposed-updates might have been the right place for it02:28
fsmithredor maybe that's just ascii-proposed02:29
rrqmmm the ascii-proposed version seems to 1.10.8-1+devuan1~5.gbped435d ... way older02:29
fsmithrednot surprised02:30
fsmithredlot of old cruft in some of the lesser-used repos02:30
golinuxEarlier discussion with that guy discovered he was using slim with elogind and iirc it should be consolekit02:34
golinuxI pointed him to the release notes but I don't think that he changed it.02:35
rrqas I remember: slim from experimental now handles both, and allows both installed. slim from ascii handles either, but not both installed (or something), or requires CK .. not sure02:39
rrqbut slim from ascii has "later" version02:40
rrqby my thinking, slim from experimental, when tested, will become "latest" slim in ascii-backports, in addition to then be from beowulf02:43
rrqthough, apparently, my thinking about these release procedures is a bit off...02:45
golinuxrrq: Me too.  Just can't wrap my mind around all the *kits and version #s03:21
KatolaZrrq: the dbus package is just an update09:41
KatolaZI should have used ascii-updates09:41
KatolaZsorry09:41
KatolaZthe only change is the re-generation of machine-id at each boot09:41
KatolaZfsmithred: ok, thanks09:43
rrqright. yes, it's just a procedural thingy of keeping "stable" "stable"... soon outsted by beowulf anyhow, but might source confusion until.09:45
rrq(I have a dictionary somewhere... I need one at least)09:47
rrq(not to talk about grammar...)09:48
rrqanyhow; I think I can unhold my dbus packages now :)09:49
KatolaZyou are totally right rrq09:50
KatolaZit was a mistake09:50
KatolaZsorry for that :\09:50
KatolaZI know what you mean09:50
KatolaZit should have gone in ascii-upodates09:51
KatolaZwe are still in time09:51
KatolaZwith a hack09:52
rrqyou mean "a well thought through but limited utility solution" I guess :)09:54
KatolaZsure09:54
KatolaZthat's what I mean :P09:54
KatolaZreally it's doable09:55
* onefang hands KatolaZ a hammer.09:55
* rrq :D !09:56
rrqthough tosave on your time we could probably suffer this quirk09:56
KatolaZmy time is not important09:57
rrqI mean to not do anything about this, but leave it09:57
KatolaZconsistency is more important :)09:58
rrqwell, *I* need my beauty sleep...09:58
KatolaZ:)09:58
KatolaZrrq: done10:03
KatolaZso10:03
KatolaZ+devuan2 is now in ascii-updates10:03
KatolaZ+devuan1 is back in ascii10:03
KatolaZrrq: ^10:04
rrqcheers10:04
KatolaZI won't use ascii-proposed-updates since the point release is happening really soon, and I don't think we have other forked packages that will be updated10:04
KatolaZdoes it make sense?10:04
rrqascii-proposed-updates = ascii-proposed ?10:05
KatolaZit's in ascii-proposed10:22
KatolaZascii-proposed-updates should include all the stuff before an update is released10:23
rrqok10:23
KatolaZrrq: are you coming to ams?10:34
rrqyes10:34
rrqneed to pick up my Devuan T-short :)10:35
rrqor shirt10:35
rrqyou?10:36
rrqKatolaZ: after (finally) apt-get update, the ascii/main repo has 1.10.26-0+deb9u1 ?? or is that just an unlucky mirroring?11:01
rrqthat's the dbus package11:04
rrqI used deb.devuan.org ; now trying with pkgmaster.devuan.org .. but my updates are painfully slow11:04
LeePenKatolaZ: Hi. I know it was going in ascii-security rather than for the point release, but the polkit backport of 0.105-25 I did is still waiting review.11:15
KatolaZrrq: please try again11:19
KatolaZLeePen: sorry I did not have much time for that :\11:20
rrq1.10.26-0+deb9u1 at http://pkgmaster.devuan.org/merged ascii/main amd64 Packages11:20
KatolaZffs11:21
LeePenKatolaZ: No worries, just want to make sure it doesn't get lost. When you have time....11:21
KatolaZrrq: we need to rebuild11:22
KatolaZffs11:22
KatolaZpfff11:23
rrq"hack" ;)11:23
KatolaZrrq: I am putting back +devuan211:23
KatolaZor we will have problems11:23
rrqfair enough11:23
KatolaZthen I will rebuild +devuan111:23
KatolaZand put it back11:24
* onefang hands KatolaZ a BIGGER hammer.11:24
KatolaZLeePen: backports always have a debian-revision version equal to "-0"11:28
KatolaZthis is to avoid that an upgrade can break things11:29
KatolaZLeePen: and guarantees that a dist-upgrade to a newer releas does the right thing11:32
KatolaZmmmhhh11:35
KatolaZbut this is actually a security update11:36
rrq1.10.22-1+devuan2 back11:37
KatolaZyes11:40
KatolaZnow I am building +devuan1 for ascii-proposed11:40
KatolaZ:\11:40
KatolaZ'cause the package was dominated by +devuan211:40
KatolaZrrq: anyway, when you have a point release, anything in -updates and -security gets into the main suite11:47
KatolaZrrq: for instance, openssh-server in ascii is -7.4p1-10+deb9u511:47
KatolaZwhich means it was updated in stretch-9.511:48
KatolaZso i think it's not totally wrong to leave +devuan2 in ascii11:48
KatolaZsince we are going to have a point release in one week11:48
KatolaZor not?11:48
rrqright. if stretch has a point release, we inherit those updates (?)11:48
KatolaZwhich updates?11:49
KatolaZthat version of openssh-server is in ascii11:49
KatolaZnot in ascii-updates11:49
rrqthe stretch point release updates.11:49
KatolaZwe get them anyway11:49
KatolaZin ascii11:49
KatolaZsince they are in stretch/main11:49
rrqright; that was what confused me before :)11:49
KatolaZhttps://pkginfo.devuan.org/cgi-bin/d1pkgweb-query?search=openssh&release=ascii11:49
KatolaZbut this has always been like that11:50
KatolaZthe main point of having -updates is to get them as soon as they are out11:50
rrqyes, it rather is my foggy appreciation that's lifting11:50
KatolaZand normally they are exclusively coming from -security11:51
KatolaZso11:51
KatolaZto recap11:51
KatolaZany security update goes into -security11:51
KatolaZafter a while it is transferred into -updates11:51
KatolaZwhen there is a point release, anything in -updates gets included in the main suite11:51
KatolaZthis is the normal flow11:51
KatolaZso normally only security updates get into a point release11:52
rrqwhere's -proposed in that picture?11:52
KatolaZit's for something that should go in -updates but does not come from -security11:53
rrqok11:53
KatolaZe.g., a package tweak11:53
KatolaZa minor package bug11:53
KatolaZetc11:53
rrqso -proposed-updates is pre-staging before -proposed11:53
KatolaZnope11:54
KatolaZbefore -updates actually11:54
rrqhm what's then different with that and plain -proposed ?11:55
KatolaZI think -proposed is used mainly during hard freeze11:57
KatolaZhttps://www.debian.org/releases/proposed-updates.html11:58
KatolaZso11:58
KatolaZ-proposed-updates has all the stuff that should go in the next point-release11:58
KatolaZincluding stuff from -security11:59
KatolaZhttps://wiki.debian.org/StableUpdates12:00
KatolaZrrq ^12:01
KatolaZhere they explain why stable-updates is there12:01
KatolaZ```All packages from stable-updates will be included in point releases.```12:01
rrqok. maybe there is a bax-and-arrow jpg/png somewhere as well, but I think my builtin whiteboard has got it now12:02
rrqbox12:02
KatolaZahahahahah12:04
KatolaZwe should make the box-and-arrow thing12:04
KatolaZtbh, our workflow is much simpler12:04
rrqbeauty sleep as wasted on you anyhow, wasn't it?12:06
KatolaZin conclusion, I would probably leave +devuan2 in ascii12:06
rrqyes12:06
KatolaZok12:06
KatolaZwe are not too much wrong with that12:06
KatolaZsince it will be there when the point release is out anyway12:06
KatolaZbbl12:06
KatolaZo/12:06
KatolaZthe one I am building now will go in ascii-proposed12:06
KatolaZbut we can discard it12:07
KatolaZ:)12:07
KatolaZo/12:07
KatolaZfsmithred: shall we have deb.devuan.org in sources.list then?15:11
KatolaZ(I am talking o desktop-live and minimal-live now)15:11
fsmithredKatolaZ, I added some code to blend_finalize to change pkgmaster to deb and it works correctly.15:31
fsmithredI didn't try to build using deb.devuan.org because it didn't work when I tried it before (months ago)15:32
EvilhamCould it have been a momentary glitch a mirror?15:33
fsmithredNo, parazyd confirmed that it didn't work15:34
fsmithredbut I never heard an explanation of why it fails15:34
fsmithredso, speaking of package versions, updates, point releases...  I want to put a new version of refractainstaller-gui in ascii before point release15:35
fsmithredthe feature to re-use an existing home partition (labeled as experimental in the installer) is screwing people up so that they can't log in. I want to remove that feature ASAP.15:36
KatolaZfsmithred: is the package ready?15:38
fsmithrednope15:38
KatolaZok15:38
KatolaZI am building a minimal-live image now to see if everything works15:39
fsmithredquick fix would be to remove the option from the options menu so nobody can see it15:39
KatolaZthat's obviously not final15:39
fsmithredI don't think the option is in the cli installer15:39
KatolaZok15:40
KatolaZwell, the first minimal-live image won't work15:40
KatolaZok about putting deb.devuan.org in sources.list instead of pkgmaster15:40
KatolaZI have done it already15:40
KatolaZin the finalize script15:40
KatolaZfsmithred: also, we should have package_lists15:40
KatolaZI have added it as well15:40
fsmithredyeah, I added that.15:41
KatolaZit was commented for some reason that I don't remember15:41
fsmithredit wasn't working15:41
KatolaZuh?15:41
fsmithredparazyd gave me the fix...15:41
fsmithredescape the $2 and $3 in the awk command15:41
KatolaZoh shit true15:41
fsmithredand then it works15:41
KatolaZnext time :)15:42
fsmithredI just checked - the re-use old home option is not in the cli installer15:43
fsmithredI would like to put the newest version of installer in ascii because it has some other fixes in it15:45
fsmithrednot just tweak the one that's in ascii now15:45
fsmithredbut I need more coffee before I touch anything15:45
KatolaZfsmithred: is it needed for mkiso_efi to start from an empty /boot?15:51
KatolaZI need to install memtest86 as custdebs...15:51
fsmithreddo it that way if you want memtest in the installed system15:52
KatolaZmmmhhhh15:52
KatolaZstill, it looks a bit unclean to remove everything in /boot15:52
fsmithredif you want it in the live-iso it needs to be in binary/live15:52
KatolaZanyway15:52
KatolaZok15:52
fsmithredor whatever that is15:52
KatolaZyeah I know15:52
KatolaZthe problem is that it gets installed among the packages15:53
KatolaZthen removed in postinstall :)15:53
fsmithredI don't understand what you mean about removing stuff from boot15:53
fsmithredwhy does it get removed?15:53
KatolaZfsmithred: there is a `rm -f /boot' in iso_make_efi15:56
KatolaZI don't explicitly call it, but apparently /boot gets wiped anyway15:56
KatolaZ:\15:56
fsmithredthat's in $efi_work/boot15:57
fsmithrednot /boot15:57
KatolaZdunno15:58
fsmithredoh, I think that's there in case you run it twice15:58
fsmithredthat's kinda weird.15:58
KatolaZuh?15:59
KatolaZrun what twice?15:59
fsmithredmight make more sense to just get rid of $efi_work before creating it15:59
fsmithredinstead of getting rid of the contents in two commands15:59
KatolaZI think I am not using it15:59
fsmithredno, you're not15:59
KatolaZbut it gets called anyway...15:59
fsmithred??16:01
KatolaZtrying again16:01
fsmithredyou don't have mkefi="yes" in config, right?16:03
KatolaZnope16:12
golinuxfsmithred: If you change the refractainstaller, I'll need to change the Visual Install Giodes16:15
golinuxGuides really16:15
KatolaZgolinux: please stop spreading FUD on dev1g :D16:22
KatolaZthere is no automagic usrmerge anywhere16:22
fsmithredwe're ssafe from that with debootstrap install?16:23
KatolaZit was discussed in Debian to make it default in buster, but it is not going to happen, apparently16:23
KatolaZthe change has been reverted16:23
fsmithredwow16:23
KatolaZso there is no usrmerge going no16:23
fsmithredgolinux, if you don't change the installer guide, the options window will show an option that doesn't exist16:24
KatolaZthey might want to try it out in buster+116:24
fsmithredthat's the only diff16:24
onefangWhat's usrmerge?16:24
golinuxKatolaZ:  You never get something wrong?16:25
fsmithredno /sbin16:25
onefangAh.16:25
fsmithredactually /sbin is symlink to /usr/sbin16:25
fsmithredand some /lib* are symlinks16:25
KatolaZgolinux: I do, but once I told that there is no usrmerge, there is no reason to keep the doubt there16:25
onefangI understand, I do that on my embedded work.16:25
KatolaZ:)16:25
golinuxI never got the memo that it was reversed in Debian.16:26
KatolaZgolinux: because you don't follwo debian-devel maybe?16:26
KatolaZ:)16:26
golinuxNot any more.  So are you going to take that option out of the installer that you had us test?16:26
KatolaZmaybe16:27
KatolaZit will not be in the default install for sure16:27
KatolaZlet's see what happens in buster16:27
KatolaZad if nothing change, we will remove it from expert install as well16:27
KatolaZwe should get wet when it has rained, not before16:27
KatolaZ:)16:27
golinuxAnd when it has stopped raining the forecaster should broadcast it.16:30
KatolaZo_O16:30
KatolaZI did16:30
KatolaZin dev1galaxy16:30
golinuxShould I edit that post to explain that Debian reversed that decision?16:30
KatolaZand you keep questioning that there is something with usrmerge16:31
golinuxMy post16:31
KatolaZwhat should I do?16:31
KatolaZ-_-16:31
golinuxIn the current post or elsewhere?  I don;t remember seeing it anywhere.16:32
KatolaZgolinux: you pulled it in, without any reason16:32
golinuxI am not a mindreader but can be forgetful.16:32
KatolaZplease re-read the thread16:32
golinuxI didn't know there wasn't a reason to give people a heads up16:34
KatolaZgolinux: have you tries upgrading ascii to beowulf ever?16:34
KatolaZ~tried16:35
golinuxNope.  Just used the mino.iso16:35
KatolaZthen, please, do not make assumptions16:35
KatolaZ:)16:35
KatolaZit's as easy as that16:35
golinuxAnd chose the "do not merge option"\16:35
KatolaZit has never been a question of upgrades from ascii16:35
KatolaZit has been a question of having new installs with meged-usr16:36
golinuxAh, OK.16:36
KatolaZthat's not a big issue, but people have you in good consideration on d1g16:37
golinuxI still had issues with path stuff in that Beowulf install though.16:37
KatolaZplease report which problems16:37
KatolaZI have made dozens of upgrades to beowulf16:37
KatolaZand have never had a single issue16:37
golinuxpkexec stuff and creating Thunar custom actions for root16:38
KatolaZthat's totally unrelated to usrmerge16:39
KatolaZtotally16:39
KatolaZthat was due to policyit16:39
KatolaZand it has been solved16:39
golinuxWell, fsr gave me a little snippet to fix the /sbin and /usr/sbin confusion and it worked16:41
KatolaZok it's useless16:41
KatolaZI give up :)16:41
golinuxI am not often in beowulf or ascii for that matter16:41
fsmithredno, I gave you a snippet to fix root's path. different issue.16:42
KatolaZthen you maybe refrain from saing stuff in public about ascii or beowulf, please16:42
KatolaZfor the good of Devuan16:42
golinux#!/usr/bin/env bash16:42
golinuxPATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin16:42
golinux/usr/sbin/synaptic16:42
golinuxexit 016:42
golinuxShould I remove those posts?16:44
fsmithredright. In beowulf, 'su' doesn't give you the sbins but lets you start graphical apps as root in a user session. 'su -' gives you the sbins in PATH but won't let you run graphical apps.16:44
KatolaZyes fsmithred, and that was an old Debian's idiosyncratic mistake16:44
KatolaZsince su never provided the user env16:45
KatolaZonlky in Debian16:45
KatolaZand debian-derived stuff16:45
golinuxSo that didn't have to do with the /usr merge?16:45
fsmithrednope16:45
KatolaZnot at all16:45
golinuxKatolaZ: Should I remove those posts on d1g?16:46
KatolaZno way16:46
KatolaZleave them16:46
KatolaZbut please, let's all try to be careful when we speal publicly16:46
golinuxI think I may need to retire soon.  Just can't keep up anymore.16:46
KatolaZthere is far too much noise16:47
KatolaZoh golinux  no way16:47
KatolaZ:)16:47
KatolaZyou don't have an option to retire :D16:47
golinuxIf the frustration level gets to high, I may have to.  I sent Dan most of the stuff to upgrade desktop-base for beowulf last night and it was really difficult for me16:49
golinuxI don't know that I will be able to do another desktop16:49
KatolaZyou should just get it easy IMHO :)16:50
golinuxYou will see these changes in another 30 years or so.  ;)16:50
onefang"take it easy" I think you mean.16:50
KatolaZyeah sorry16:51
KatolaZI am on three chats16:51
KatolaZ:)16:51
KatolaZgolinux: sure, but take it easy16:51
golinuxJust preparing for the trip has been exhausting.  So many things to do.16:51
golinuxMy list is getting shorter though.16:52
fsmithredKatolaZ, there's a commented line in iso_write_isolinux() to copy memtest into place. You can use that if you're not including uefi.17:42
fsmithreddon't need live-overlay17:42
KatolaZit turns out I had a typo in custom_debs17:42
KatolaZ-_-17:42
fsmithreddamn17:42
KatolaZI had added a line to copy it there17:42
fsmithredok17:42
KatolaZbut was not taken17:42
fsmithredit's in zlibs/iso17:42
fsmithredit didn't work?17:43
KatolaZrebuilding17:43
KatolaZFFS it didn't work this time either17:43
KatolaZ:\17:43
KatolaZlet's see what's wrong17:43
fsmithredoh, I guess for the line in iso_setup_isolinux, memtest has to be in the package list so that it's in $strapdir to be copied to binary/live/17:44
KatolaZI did it the other way round17:45
KatolaZput it in custom_debs17:45
KatolaZthen copied from finalize17:45
fsmithredok17:45
fsmithredlinux: three ways to do anything (not counting the other seven ways)17:45
KatolaZ:D17:45
fsmithredone of them is sure to work!17:46
KatolaZD17:46
KatolaZbut you don't know which one in advance17:46
onefangSounds like Second Life booleans.  Yes, no, maybe, perhaps, dunno, and what was the question?17:47
fsmithredLOL17:47
pixelherodevHeads up that if you try installing openrc from testing with librc1 from stable it segfaults during the post-install script, the segfault is in librc from the call to `rc-update -u`19:23
pixelherodevThis is only an issue to me because upgrading openrc didn't force an upgrade to librc1 as well, so I had to chroot in and fix the problem19:23
KatolaZpixelherodev: not tried, but why would you do that?19:23
KatolaZmmhhh19:23
KatolaZthat's strange19:23
pixelherodevNot intentionally19:27
pixelherodevI was having problems installing libgl1-mesa-glx19:27
pixelherodevI had some i386 packages using backports, so installing that for amd64 conflicted with some of them19:27
pixelherodevI just added testing and tried again and it added openrc to the upgrade list. If I'd been paying more attention I probably would have been able to avoid the problem in the first place, but I distractedly hit enter19:28
KatolaZpixelherodev: we should try this one on a clean ascii19:31
pixelherodevThis basically was19:31
pixelherodevIt was a fresh install19:31
pixelherodevI literally installed it because I needed a distro with glibc so I could get PCSX2 working19:31
KatolaZbut you mentioned you had packages from backports and clashes due to multiarch...19:31
KatolaZpixelherodev: you cannot use packages directly from testing like that19:34
KatolaZyou normally need to to a dist-upgrade19:34
KatolaZwhich will preserve deps across the upgrade19:34
pixelherodev... oops. Been a while since I've used a Debianesque distro :P19:37
KatolaZ:)19:38
KatolaZbut why did you need any package from testing at all19:38
KatolaZyou shouldn't19:38
KatolaZanything in backports will have all its deps in stable or in backports19:38
* pixelherodev shrugs19:39
pixelherodevCould probably have made do with backports19:39
KatolaZwhich package did you need from testing?19:51
pixelherodevActually, I think i needed some wayland packages from there so that I could build Sway20:05
KatolaZfsmithred: ok memtest works21:08
KatolaZ:)21:08
pixelherodevIs there a list of mirrors? I can't find one, and my current mirror is providing download speeds below 100KB/s21:41
pixelherodevAnd yes, I did test internet speed, it is - as it should be - in the 3-5MB/s range21:41
Evilhamthat' smore of a #devuan question, but where are you located?21:42
pixelherodevUS21:43
pixelherodev...and wasn't aware that channel existed, thanks21:43
Evilhamno worries, in any case, it is a known issue that there is no US mriror atm, are you using deb.devuan.org?21:46
Evilhamon your sources.list21:46
Evilham?21:46
Evilhampixelherodev:  you may want to use devuan.c3sl.ufpr.br, I seem to recall they have a good link to the US, the list of mirros btw is live on http://pkgmaster.devuan.org/mirror_list.txt which is linked from https://devuan.org/os/21:51
Evilhamyou are free to open a Merge Request on the website if you think it can be made clearer (albeit, it is not a given that it will be accepted, but there are good chances if it is a good one),21:52
EvilhamSee: https://git.devuan.org/devuan-editors/devuan-www/21:52
rrqKatolaZ: dbus-repos #112 failed for dbus_1.10.22-1+devuan123:04

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