ServiceRobot | hello there tiny peeps | 00:58 |
---|---|---|
ServiceRobot | so I tried out the latest ascii iso and I'm wondering if I can remove sysvinit now that openrc is an option | 01:00 |
ServiceRobot | I selected openrc as the init I wanted to use but still had sysvinit when the installation finished | 01:00 |
gnarface | i don't know the procedure but people have succeeded at this. maybe it's enough to just remove sysvinit once openrc is set up. there might be an extra step. | 01:02 |
ServiceRobot | I tried removing sysvinit and the result was an unbootable system. I assume it's because it's tied to the init binary | 01:03 |
gnarface | had you finished installing openrc? maybe you have to install it AFTER removing sysvinit. i'm just not sure. | 01:03 |
ServiceRobot | if this is the case it should be seperated so more init systems can be added in the future | 01:04 |
ServiceRobot | it does boot with openrc, but I think it's on top of sysvinit | 01:04 |
ServiceRobot | the details in the recent announcement confuse me about this | 01:04 |
gnarface | hmmm. well i don't have any suggestions other than to hang around here until you cross paths with someone else who is using openrc. | 01:04 |
gnarface | you're not the first, i know that for sure. | 01:04 |
golinux | ServiceRobot: I think that openrc uses some sysvinit scripts so you either need to rely on those or replace them with other options to fulfill those functions. | 01:28 |
Juesto | most of devuan is sysvinit | 01:28 |
golinux | You might want to look at this thread https://dev1galaxy.org/viewtopic.php?id=1128 and also here http://troubleshooters.com/linux/init/ | 01:28 |
ServiceRobot | ya, that's annoying. I was hoping devuan would start to allow complete replacements. I'm a bit of a perfectionst when it comes to that | 01:29 |
Juesto | openrc version in devuan lacks important things imo | 01:29 |
golinux | Juesto: openrc is an option that quite a few are using. | 01:29 |
Juesto | openrc has somewhat replaced sysvinit in a few versions later | 01:29 |
Juesto | devuan's 0.23 | 01:29 |
Juesto | 0.24 introduced agetty orcinitscript | 01:30 |
Juesto | and 0.25 replaces sysvinit | 01:30 |
ServiceRobot | ah | 01:30 |
ServiceRobot | wait, so the goal is to completely replace openrc, or make it an option? it seems there should be a main init package from which other init systems can work with | 01:31 |
golinux | There have been discussions about openrc on the mailist, irc and the forum | 01:31 |
ServiceRobot | right, I've been keeping track of the forum a bit | 01:31 |
golinux | Devuan is about choice | 01:31 |
ServiceRobot | that's good to know. "init freedom" is the main slogan on one of your paegs | 01:32 |
ServiceRobot | *pages | 01:32 |
golinux | Indeed. In the beginning we were hoping to also add systemd to the options but then it became apparent that wouldn't be possible. | 01:33 |
golinux | brb | 01:33 |
Juesto | so just fueling the anti systemd stuff after evaluation | 01:34 |
ServiceRobot_ | really? you were gonna make systemd an option? at that point just use debian :L | 02:18 |
djph | that wasn't the point - *init freedom* was | 02:20 |
ServiceRobot_ | yes, and systemd conflicts with that ideology | 02:20 |
djph | and the *freedom* to usy systemd falls solidly into that | 02:20 |
ServiceRobot_ | well either way you guys seem to be making good progress. I just tested everything down stairs and it worked fine. live desktop installer doesn't work for me though | 02:21 |
Juesto | eh | 02:23 |
gnarface | ServiceRobot_: the point is that early on, detractors said it would be hypocritical to make a distro based on init freedom then exclude the most popular new init system. so, man hours were spent evaluating the effort level, at which point it was discovered conclusively that systemd and several things it has tied into (gnome, for example) are already "too far gone" to be redeemed without a costly rewrite that the project | 02:23 |
gnarface | just doesn't have resources to handle internally. | 02:23 |
ServiceRobot_ | that's understandable. everything you guys do is provided with no charge. how long do you think it will be before I can fully removed sysvinit and use a different init though? | 02:24 |
gnarface | i think during this time, the maintainers of systemd also through action made it obvious they would continue to obstruct portability at a core level of their design | 02:24 |
gnarface | i think other than systemd, you can use any init system you want right now | 02:25 |
ServiceRobot_ | sysvinit is default in devuan. minor support for openrc it seems. no support for runit, which I actually like for how simple it is | 02:25 |
gnarface | if what you're asking is how long until it will be an option "out of the box?" ... that i don't know. as they say in #debian; sooner if you help! | 02:25 |
ServiceRobot_ | how do you think I could help speed up development? | 02:26 |
gnarface | i don't actually speak for the project in any official capacity, but there is a path you can follow to help | 02:27 |
gnarface | there is a mailing list | 02:27 |
gnarface | https://devuan.org/os/partners/ways-to-help there is also this, if you're not a coder | 02:27 |
gnarface | there is also i think #devuan-dev | 02:29 |
ServiceRobot_ | the thing is how will this help with allowing more init systems? flexibility like removing an entire init and replacing it? | 02:29 |
djph | thats the idea | 02:29 |
Juesto | it's complicated | 02:31 |
ServiceRobot_ | I can imagine | 02:32 |
ServiceRobot_ | there's not really a standard for multiple inits on 1 distro it seems | 02:32 |
Juesto | Have a kernel capable of multiple bosses | 02:32 |
Juesto | good luck | 02:32 |
Juesto | upstart was on top of sysv | 02:33 |
Juesto | there you go | 02:33 |
ServiceRobot_ | well the distro I'm currently using figured it out. I can run runit or openrc. I wonder if it can be implemented similarly here | 02:33 |
Juesto | well it's similar to openrc to some extent | 02:33 |
Juesto | which distro | 02:33 |
Juesto | openrc is gentoo | 02:33 |
ServiceRobot_ | Artix Linux | 02:33 |
ServiceRobot_ | yes, but gentoo doesn't really seem like the server type of distro. I have to compile everything myself... | 03:05 |
Juesto | yeah... there's some distros that help in that | 03:10 |
Juesto | funtoo | 03:10 |
Juesto | there's many distros that resolve such difficulty | 03:12 |
Juesto | forks* | 03:12 |
ServiceRobot_ | I thought funtoo was just an experimental distro? not really meant for production? | 03:12 |
gnarface | it's really difficult to formulate a helpful response to that without sounding snarky | 03:14 |
ServiceRobot_ | whatever is most informative | 03:15 |
gnarface | if you ask Microsoft, they'll tell you that Linux and Linux distros in their entirety are experimental and not meant for production. of course they have an alternative ready. | 03:15 |
gnarface | the truth is that it's as much about the sysadmin as the distro | 03:16 |
ServiceRobot_ | maybe experimental isn't the right word. I mean not meant for servers | 03:16 |
gnarface | oh, well personally (as a professional) i'd advise against any source-based distro for servers that rely on 3rd party or commercial software | 03:17 |
ServiceRobot_ | what do you mean by source-based? | 03:17 |
gnarface | gentoo, funtoo, arch, slackware | 03:17 |
ServiceRobot_ | right, which is why I looked at this website for options: http://without-systemd.org/wiki/index.php/Linux_distributions_without_systemd | 03:18 |
gnarface | but the thing is, there's also types of things source-based distros do better, and those things aren't all mutually exclusive with "being a server" | 03:18 |
ServiceRobot_ | I'm just not sure what to go with. What I'm looking for is something with openrc that is fixed release and has enough packages for a LEMP/LAMP setup | 03:19 |
gnarface | well i'd recommend Devuan for that, even if you weren't already here | 03:19 |
ServiceRobot_ | ya, there are sides to rolling or fixed release. I've tried both. rolling is great for me as a desktop user | 03:19 |
ServiceRobot_ | I tried Devuan. not a fan of sysvinit, but I suppose if it's the only and best choice, I have to live with it | 03:20 |
gnarface | whatever superficial issues you're having with openrc, i suspect you're putting way too much weight on them as a factor for this decision. openrc is popular enough they'll get fixed. | 03:20 |
gnarface | also, i can help you with most common struggles new users have with sysvinit | 03:20 |
gnarface | it's really not THAT complicated | 03:21 |
gnarface | you can still probably get accurate info about it's behavior from the debian wiki entries for wheezy | 03:21 |
gnarface | (you might want to make copies of the LSB headers notes before they get removed though) | 03:21 |
ServiceRobot_ | it's just that the init systems I've messed with thus far have been openrc, runit, and systemd (which gives more problems than other inits I'll bite) | 03:23 |
ServiceRobot_ | devuan doesn't have complete support for it yet. it seems they're working towards it, and that's good enough for me. I can wait | 03:23 |
gnarface | for a LAMP setup, the debian packages for PHP are far superior to most other distros "one giant package that you have to rebuild in it's entirety for the tiniest change" approach | 03:25 |
ServiceRobot_ | ya, one small issue. there's a nginx module that doesn't seem to be available... libnginx-rtmp :L | 03:26 |
ServiceRobot_ | very small issue though | 03:26 |
gnarface | of course you'll have to resist the urge to start downloading and building source instead of searching the package repo for the right php module package, but once you get over that it gets easy | 03:26 |
ServiceRobot_ | and not the main reason why I'm waiting | 03:26 |
gnarface | hmm, i can't speak for nginx but whenever you are missing a package or need a newer version of something, check backports | 03:27 |
ServiceRobot_ | it's a module | 03:27 |
ServiceRobot_ | sort of an addon if you will. it can be built directly into nginx though I think | 03:28 |
gnarface | packages.debian.org says it's present in stretch-backports. that means it is also present in ascii-backports | 03:29 |
gnarface | in Debian, the package is actually named "libnginx-mod-rtmp" ... mabye that'll help you search for it better | 03:29 |
DocScrutinizer05 | noob question: could you give me a ultraterse howto for installing devuan on a server? Upload CD image to hoster? | 03:29 |
ServiceRobot_ | I searched for libnginx-mod-rtmp. couldn't find it | 03:30 |
ServiceRobot_ | on devuan I mean | 03:30 |
ServiceRobot_ | maybe it hasn't been added yet | 03:30 |
gnarface | ServiceRobot_: that's because it's not in the main repos. it's in backports see here: https://backports.debian.org/ | 03:30 |
ServiceRobot_ | what about for devuan? | 03:30 |
gnarface | ServiceRobot_: all that's different is the url | 03:31 |
ServiceRobot_ | ah, I see. I'm not exactly an expert on how debian works | 03:31 |
DocScrutinizer05 | !amprolla | 03:31 |
infobot | nextime gave an excellent explanation how amprolla works, at https://botbot.me/freenode/devuan/2016-05-07/?msg=65646427&page=4, or https://git.devuan.org/devuan-infrastructure/amprolla | 03:31 |
gnarface | DocScrutinizer05: pick premade debian image, change sources.list to devuan repos and upgrade :) | 03:33 |
DocScrutinizer05 | boils down to: all that's in debian is also in devuan, unless it got removed/replaced | 03:33 |
DocScrutinizer05 | gnarface: ta! | 03:33 |
ServiceRobot_ | ah. see, with artix, we have to rebuild all arch's packages. I guess that's not the case with devuan? | 03:33 |
ServiceRobot_ | so it should be in backports then huh | 03:33 |
DocScrutinizer05 | it should, yes | 03:34 |
ServiceRobot_ | alright, but my main concern is the init... if I can't completely replace it, I'll just wait. | 03:36 |
DocScrutinizer05 | replace which init by what? | 03:36 |
ServiceRobot_ | replace sysvinit with openrc or runit, or whatever will be supported | 03:37 |
djph | sure, have at iy | 03:37 |
djph | *it | 03:37 |
DocScrutinizer05 | >>When installing from ISO, the expert install option offers a choice of SysVinit and OpenRC. << http://tinyurl.com/y9pta5vm | 03:38 |
ServiceRobot_ | yes, I tried the expert install. it still has sysvinit left over afterwards | 03:45 |
Juesto | DocScrutinizer05: ServiceRobot_: because openrc 0.23 does not replace sysvinit but you can use sucklessinit clone on openrc-init by specifying it at boot | 03:49 |
ServiceRobot_ | ah, right, openrc 0.25 replaces it right? when do you think that gets pushed? | 03:49 |
Juesto | ServiceRobot_: some packages were rebuilt because of systemd dependencies primarily, it's mostly the same Debian environment | 03:50 |
Juesto | No idea, openrc is currently at 0.35.5 | 03:50 |
ServiceRobot_ | wait, but .35 is past .25? I must be confused >_> | 03:53 |
Juesto | ServiceRobot_: yes, it's just the version where the change started to take effect | 03:55 |
Juesto | yes 0.35.5 is latest | 03:55 |
ServiceRobot_ | ah, so that means it's in the process of being changed? | 03:55 |
Juesto | that it's done on upstream not on Devuan, ServiceRobot_ | 03:56 |
golinux | ServiceRobot_: Use the backports in the devuan repos not debian's | 03:56 |
Juesto | it's just a version reference | 03:56 |
ServiceRobot_ | right... | 03:56 |
golinux | Never mix repos. Rule #1 | 03:56 |
ServiceRobot_ | that rule I know | 03:56 |
Juesto | which devuan repositories are available? | 03:57 |
golinux | https://devuan.org/os/etc/apt/sources.list | 03:57 |
golinux | Juesto: ^^^ | 03:57 |
ServiceRobot_ | wait, so the backports has the openrc version I want? | 03:58 |
Juesto | do they? | 03:59 |
golinux | If you see it it's there | 03:59 |
Juesto | i suggest visiting the repository with a browser beforehand, ServiceRobot_ | 03:59 |
ServiceRobot_ | I thought there wasn't a way to search devuan packages yet? | 03:59 |
Juesto | yes | 04:00 |
Juesto | but for extra verification :) | 04:00 |
Juesto | and for when search is unavailable | 04:00 |
golinux | It isn't live yet. | 04:00 |
Juesto | that too | 04:00 |
buZz | apt search ? :D | 04:00 |
golinux | There has got to be a command for that. | 04:00 |
Juesto | buZz: not if not added | 04:01 |
buZz | dpkg -l |grep whatyouneed ? | 04:01 |
golinux | I just search in synaptic. | 04:01 |
buZz | oh, ok | 04:01 |
fsmithred | there's no openrc in any backports repo | 04:01 |
ServiceRobot_ | darn | 04:01 |
Juesto | backports are backports | 04:01 |
golinux | I thought you had checked. | 04:01 |
buZz | openrc/testing 0.23-1+b1 amd64 | 04:01 |
Juesto | forks | 04:01 |
Juesto | AH | 04:02 |
buZz | is that what you're searching for? | 04:02 |
buZz | oh, you want newer | 04:02 |
buZz | gotcha | 04:02 |
Juesto | yeah | 04:02 |
fsmithred | 0.34-3 buster/beowulf/sid/ceres | 04:02 |
Juesto | is sid mirrored on devuan? | 04:02 |
Juesto | oh it's in sid | 04:03 |
ServiceRobot_ | I don't think sid is even usuable yet? | 04:03 |
Juesto | okay so I broke into a initrd shell | 04:03 |
Juesto | ServiceRobot_: depends on what you call usable | 04:03 |
Juesto | you can probably import the package manually | 04:03 |
fsmithred | we haven't done a lot of work on buster or ceres yet | 04:04 |
Juesto | not a good idea | 04:04 |
djph | sid is always unstable | 04:04 |
Juesto | again, you could just download the specific package without adding a entire repository that would mess up the system | 04:05 |
fsmithred | problem is that among the devuanized packages, there are some newer versions in ascii than in beowulf and ceres | 04:05 |
Juesto | not the case with orc :) | 04:05 |
Juesto | anyway so | 04:05 |
fsmithred | better to backport it | 04:05 |
Juesto | what i could try regarding my initramfs missing executables issue? | 04:05 |
fsmithred | I have no ideas on that. | 04:06 |
ServiceRobot_ | ya, but I don't thin kthe backported version replaces sysvinit yet | 04:06 |
ServiceRobot_ | *think | 04:07 |
Juesto | there's isn't a backported one? | 04:07 |
fsmithred | no openrc in backports | 04:07 |
Juesto | It would replace sysvinit | 04:07 |
Juesto | again, introduced in 0.25 | 04:08 |
ServiceRobot_ | I mean if there was one. sorry, not to be confusing | 04:08 |
ServiceRobot_ | okay, so it was introduced in 0.25, but what is the current version in devuan for ascii? | 04:08 |
fsmithred | you might get away with installing the buster/ceres version in ascii. *might* | 04:08 |
fsmithred | 23 | 04:09 |
Juesto | yeah that's what i suggested | 04:09 |
Juesto | ServiceRobot_: we are repeating the versions stuff.... | 04:09 |
ServiceRobot_ | that would be mixing versions though. doesn't really seem maintainable. maybe I should wait? | 04:09 |
Juesto | uh | 04:09 |
Juesto | ServiceRobot_: you're not mixing anything if you're just trying to install a specific deb downloaded from the repository directly | 04:10 |
Juesto | maybe or not, as he said, you *might* get away with it | 04:11 |
ServiceRobot_ | right, but I'm not sure how well it would work. a lot of important packages rely on sysvinit | 04:11 |
Juesto | again, openrc has compatibility | 04:11 |
Juesto | ServiceRobot_: ask #openrc for such concerns | 04:12 |
fsmithred | is this for a production system? | 04:12 |
fsmithred | or your own box at home? | 04:12 |
ServiceRobot_ | it's a computer I run at home, but I want something stable | 04:12 |
ServiceRobot_ | I'm trying to run it in a production-like way | 04:12 |
Juesto | if you're going prod/daily, you're just being... | 04:13 |
Juesto | uh can't think a good word | 04:13 |
buZz | experimental? :D | 04:13 |
buZz | a guinea pig? | 04:13 |
Juesto | lol | 04:13 |
fsmithred | I just tried a simulated install of openrc=0.34-3 in ascii | 04:13 |
buZz | you wouldnt gain more stability by installing stuff that hasnt finalized testing , ServiceRobot_ | 04:14 |
fsmithred | 0 upgraded, 3 new, 5 to remove, 968 not upgraded | 04:14 |
Juesto | no, he kinda of specifically wants some kind of features | 04:14 |
Juesto | rip | 04:14 |
ServiceRobot_ | which is why I'm considering just waiting | 04:14 |
buZz | i would just wait | 04:14 |
Juesto | :p | 04:14 |
buZz | let other ppl do the experimentation | 04:14 |
buZz | :) | 04:14 |
buZz | have your stable box be stable | 04:14 |
ServiceRobot_ | how long must I wait is the big question >_> | 04:14 |
buZz | <1 year | 04:14 |
buZz | i'm guestimating | 04:14 |
ServiceRobot_ | good enough to me | 04:14 |
Juesto | what your guys version of initramfs-tools and kernel are? | 04:15 |
buZz | 4.9.88 iirc? | 04:15 |
Juesto | and the initramfs filename | 04:15 |
buZz | Linux h81m 4.9.0-6-amd64 #1 SMP Debian 4.9.88-1+deb9u1 (2018-05-07) x86_64 GNU/Linux | 04:15 |
buZz | initrd.img-4.9.0-6-amd64 | 04:15 |
buZz | just plain ascii here | 04:15 |
fsmithred | initramfs-tools 0.130 in ascii, beowulf and ceres | 04:15 |
Juesto | Oh well | 04:16 |
buZz | initramfs-tools/testing,testing,now 0.130 all [installed] | 04:16 |
Juesto | no idea why on earth i have btrfs but not fsck | 04:16 |
Juesto | when the opposite is happening on the installed system | 04:16 |
Juesto | btrfsutils isn't installed | 04:16 |
buZz | Juesto: fsck inside initramfs? | 04:16 |
Juesto | and e2fsck is | 04:16 |
Juesto | buZz: needed for pre-mount sanity checks | 04:17 |
buZz | i guess , yeah | 04:17 |
Juesto | so yea it's a problem | 04:25 |
Juesto | anyone? :/ | 04:25 |
gnarface | Juesto: it's probably in the "btrfs-progs" package. debian/devuan package names aren't all the same as other distros. even ubuntu changes the names of some critical stuff. | 04:44 |
ServiceRobot_ | you know, while I'm waiting on devuan, alpine linux seems like a good alternative? | 04:45 |
gnarface | Juesto: you can save yourself a lot of time by learning how to pass wildcards to the package search | 04:45 |
gnarface | ServiceRobot_: are you aware of the #debianfork channel? | 04:46 |
buZz | ServiceRobot_: 'reinstallation' isnt my idea of stable systems :) | 04:46 |
ServiceRobot_ | what do you mean by 'reinstallation'? | 04:47 |
buZz | moving between distros ;) | 04:47 |
gnarface | ServiceRobot_: he probably means that the best distro to use while waiting on Devuan is Devuan | 04:47 |
ServiceRobot_ | well I'll settle on something eventually | 04:47 |
ServiceRobot_ | heh, I suppose. I have a hard time making up my mind on things | 04:48 |
ServiceRobot_ | because I might end up finding something better | 04:48 |
gnarface | your error in judgement is as clear as day to me, because i've been where you are now. but i don't have a lot of good ways to address it without sounding derogatory | 04:49 |
gnarface | the basic gist of it, is that you're switching distros prematurely | 04:50 |
ServiceRobot_ | no one is above critisism | 04:50 |
ServiceRobot_ | I've actually stuck with the distro I've been using for about a year | 04:50 |
ServiceRobot_ | but I don't think I should use it for servers | 04:50 |
gnarface | it's not that your logical evaluation of the distros is wrong - it's that your own familiarity with any one of them would easily obviate any of the concerns you are currently agonizing over | 04:51 |
buZz | +1 | 04:51 |
ServiceRobot_ | so you're saying that if I become familar with it, I'd be fine with it? | 04:51 |
gnarface | yea. if you spent half as much time today figuring out how to make the openrc install in devuan do what you want instead of kvetching over it, you'd already have a solutino | 04:51 |
gnarface | *a solution | 04:51 |
gnarface | and a decision on distro | 04:52 |
gnarface | it's probably also worth mentioning that i'm not even 100% sure, as someone who doesn't know openrc, whether there's anything even wrong with the install. for all i know you're seeing some sysvinit scripts that were left in openrc as a reverse-compatibility measure and you're overreacting to it. | 04:53 |
ServiceRobot_ | ya, that could be it >_>. | 04:53 |
ServiceRobot_ | okay, but I'm also not a fan of graphical installers when it comes to being exact. there's a debootstrap tutorial for devuan, right? not sure if it was updated recently | 04:55 |
gnarface | but we (people who hang out in linux irc support channels in general) see this pattern a lot; as a new linux user, coming from a commercial operating system environment, you're tuned to basically just blow everything out the airlock at the first sign of trouble. around here, we don't reinstall we just alter the install in place. | 04:55 |
gnarface | that's the way Debian was designed to be used primarily; upgraded, not reinstalled | 04:55 |
gnarface | to the extent that the installer isn't even always as well tested as the upgrade process | 04:56 |
ServiceRobot_ | don't get the wrong idea. I blew the airlock off because I wanted to try something different. I've tried ubuntu, linux mint, then arch, then artix (which is working good for me). I like trying different things | 04:56 |
ServiceRobot_ | if there's trouble, I'll do as much research as possible to understand it | 04:56 |
gnarface | so it's not that your evaluation of "when to reinstall" is fundamentally insane, it's just based on ideology foreign to the open source development paradigm. | 04:57 |
ServiceRobot_ | it's more of the question "should I give this a try. is it worth it?" | 04:57 |
gnarface | i would say it's worth it, but even that question is loaded with invalid assumptions | 04:58 |
gnarface | you shouldn't have to purge the old install to try a new distro | 04:58 |
Juesto | blame disk space | 04:59 |
ServiceRobot_ | I've used virtual machines for testing | 04:59 |
gnarface | you could install to a blank partition on the existing drive, or to an external drive or even USB key (it would be slow but work) | 05:00 |
gnarface | you could use a VM but they dont' typically give you a good picture of performance or hardware compatibility | 05:00 |
buZz | if with 'stable machine' you ment 'distro testbed' , then by all mean | 05:00 |
buZz | s | 05:00 |
buZz | just dont expect stability when you delete your knowledge on each reinstall :P | 05:00 |
ServiceRobot_ | nah, any problems I run into I need to deal with. if it's something upstream I report it | 05:01 |
gnarface | ServiceRobot_: debootstrap shouldn't behave any different from debian. just use the devuan repo url instead. | 05:01 |
buZz | ;) my point was most, dont confuse 'stability' with 'hey, lets try this new thing' | 05:01 |
buZz | it would only get you hurt | 05:01 |
gnarface | yea, that's a good point too | 05:02 |
DocScrutinizer05 | tbh I don't even care much at all which distro I run. Until recently the differences were marginal and really nothing you couldn't fix into oblivion after installing whatever flavor and then tweaking the packages and configs | 05:02 |
buZz | sure, and thats all good for a testbed | 05:02 |
buZz | not for stability though | 05:02 |
Juesto | what's the difference between merged and devuan? | 05:03 |
Juesto | packages.devuan.org/devuan | 05:03 |
gnarface | i think it's supposed to be /merged | 05:03 |
Juesto | hm | 05:04 |
Juesto | must be a leftover | 05:04 |
gnarface | probably something outdated, or maybe just a mistake | 05:04 |
gnarface | i don't know | 05:04 |
gnarface | it was supposed to be merged only, last i heard | 05:04 |
Juesto | Ah it's old | 05:04 |
Juesto | hmmmmm | 05:05 |
Juesto | :s | 05:05 |
DocScrutinizer05 | afaik related to !amprolla | 05:05 |
Juesto | !amprolla | 05:06 |
infobot | nextime gave an excellent explanation how amprolla works, at https://botbot.me/freenode/devuan/2016-05-07/?msg=65646427&page=4, or https://git.devuan.org/devuan-infrastructure/amprolla | 05:06 |
Juesto | ~say | 05:07 |
Juesto | hmmmmmmm | 05:07 |
Juesto | !say | 05:07 |
infobot | say what? | 05:07 |
Juesto | lol | 05:07 |
Juesto | anyway | 05:07 |
Juesto | for some reason initramfs doesn't appear to work properly | 05:07 |
Juesto | or I'm doing something wrong | 05:08 |
gnarface | any error messages? | 05:08 |
Juesto | anyway, which being in the initrd shell is not ideal | 05:08 |
Juesto | eh | 05:08 |
Juesto | there's issues with the tools included, gnarface | 05:09 |
Juesto | mount doesn't do -t auto | 05:09 |
Juesto | fsck is missing there | 05:10 |
gnarface | hmmm | 05:10 |
gnarface | did update-initramfs not get run last time you rebuilt your kernel? | 05:11 |
Juesto | gnarface: it's producing the same bad initrd | 05:11 |
gnarface | this was for btrfs you said? did you make sure btrfs-progs was installed before you ran it? | 05:11 |
Juesto | gnarface: I don't have a raid | 05:12 |
Juesto | but guess what | 05:12 |
Juesto | it's generating a initramfs with brtfs but without fsck | 05:13 |
gnarface | that does seem strange | 05:13 |
Juesto | and i have the opposite in the installed system | 05:13 |
Juesto | i don't have brtfs installed | 05:13 |
Juesto | oh nvm | 05:13 |
gnarface | ls -l /sbin/fsck* | 05:13 |
gnarface | some of these filesystems, they've been replacing the normal fsck.* symlink with a broken wrapper script. | 05:14 |
gnarface | as best as i can tell it's a purposeful attempt to sabotage everything but ext4 | 05:14 |
gnarface | (thinly disguised as incompetence, but try to file a bug report about it and you'll realize what's going on) | 05:14 |
gnarface | anyway, find the right binary, delete the broken wrapper script, write an angry letter if it helps you feel better, then put the symlink back and re-run update-initramfs -u -k all | 05:15 |
Juesto | fsck.ext* = symblink to e2fsck | 05:15 |
gnarface | yea, they didn't break the ext* ones | 05:15 |
Juesto | gnarface: this is a initrd problem | 05:16 |
Juesto | i don't know if it's exactly broken or not | 05:16 |
Juesto | because | 05:16 |
Juesto | i just found /usr/share/initramfs-tools | 05:17 |
Juesto | that should be the place to look at since is where the scripts used come from | 05:17 |
Juesto | my fstab consists in "rootfs / rootfs rw 0 0" | 05:18 |
gnarface | heh | 05:18 |
gnarface | doesn't look valid to me | 05:18 |
gnarface | but what do i know | 05:18 |
gnarface | also i'm pretty sure if that fsck.btrfs isn't in /sbin, it won't work | 05:19 |
gnarface | but i don't claim to have tried btrfs | 05:19 |
Juesto | gnarface: it's spec file vfstype mntops freq passno | 05:19 |
gnarface | is it? | 05:20 |
Juesto | i just told you the order of the fstab | 05:21 |
Juesto | what's each thing | 05:21 |
gnarface | what actually happens wrong? | 05:22 |
Juesto | gnarface: the issue is with initrd | 05:23 |
gnarface | i gathered that part but you're gonna have to be more specific if i can help | 05:23 |
gnarface | does it fail to boot? do you get an error? what is the error? | 05:24 |
gnarface | i've never seen "rootfs" used in that context in a valid fstab file. | 05:24 |
gnarface | that doesn't guarantee it's invalid, but right now we have a situation where you've told me nebulously there's something wrong with your initrd.img and shown me an invalid fstab file. | 05:25 |
gnarface | i've got enough information to tell you what's wrong with the fstab file, but not the initrd.img | 05:25 |
Juesto | gnarface: apparently update-initramfs depends on the running fstab... | 05:25 |
gnarface | yea, it might. all kinds of things rely on a valid fstab. (not systemd though) | 05:26 |
Juesto | ... | 05:26 |
gnarface | it might not surprise you to know that "mount -a" also relies on fstab | 05:26 |
Juesto | i already knew that :) | 05:26 |
gnarface | so, you said it can't find the mount points and then show me an invalid fstab. where is the mystery? | 05:27 |
gnarface | you want me to help you make a valid fstab? that's easy, i can help. | 05:27 |
Juesto | gnarface: no, the problem is: i don't get fsck in the initrd | 05:27 |
Juesto | the binary isn't getting added at all | 05:28 |
gnarface | well it would need to be able to find your filesystem and mount it to succeed at that... | 05:29 |
gnarface | which it would need a valid fstab for | 05:29 |
gnarface | i'm not saying that's the only problem. there could be others, but this is definitely one. | 05:29 |
gnarface | Juesto: sorry, i must have missed something important. how did you even get into this situation? are you trying to migrate an existing install to btrfs? | 05:31 |
Juesto | gnarface: no, it's a laptop | 05:32 |
Juesto | I'm trying to recover fsck | 05:32 |
Juesto | get fsck back there | 05:32 |
gnarface | do you have a live cd or usb image you can boot from? | 05:33 |
gnarface | or another machine you can plug the harddrive into? | 05:34 |
gnarface | even if you have an older kernel version still installed it might be bootable | 05:38 |
gnarface | the initrd.img files are paired with individual kernel versions | 05:38 |
gnarface | (assuming you're using stock kernels; if you're not, all bets are off here) | 05:38 |
Juesto | gnarface: i am, i think it's the fstab | 05:54 |
gnarface | well the first "rootfs" needs to be a valid UUID or /dev/ device node file path and the second one needs to be the name of the filesystem driver (ext4 or btrfs or whatever) | 05:56 |
gnarface | "rw" for the fourth parameter will probably work but you should put "defaults" there until you know better | 05:56 |
gnarface | the last two fields should be 0 1 | 05:57 |
gnarface | and that is just the root filesystem. if you have other partitions, they'll be ignored unless you add them here too. | 05:58 |
gnarface | you might very well only have that one data partition but usually there would at least be a second partition for swap | 05:58 |
gnarface | you can find out the existing UUIDs from /dev/disk/by-uuid/ if you can boot it, but be warned that they'll change if you resize or reformat any of those partitions | 05:59 |
Juesto | gnarface: i just said the meaning of the fstab entries | 06:24 |
Juesto | i resized a partition and the uuid didn't change afaik | 06:25 |
aitor | good morning | 06:59 |
aitor | i've isolated the code of the netstatus plugin for LXPanel, ported by Hong Jen Yee (PCManFM) from the original code comming from the GNOME2 netstatus panel applet | 07:02 |
aitor | i want to use it in the backend of simple-netaid | 07:02 |
aitor | it uses an unix socket depending on libiw-dev | 07:03 |
aitor | i isolated it in a main.c and it's building succesfully | 07:03 |
aitor | the applet seems to be a work in progress, because it doesn't appear in my LXPanel | 07:05 |
aitor | surelly, there is a LxDE mailing list; i'll write there... | 07:06 |
aitor | no more news for now :) | 07:06 |
aitor | have a nice day! | 07:06 |
_abc_ | Hello. I have multiple volumes which get mounted or not, and updatedb runs when all are mounted, then locate behaves as if I'd issue locate -e, not finding things on not (yet) mounted volumes. How do I make locate behave normally, showing db contents instead of -e ? | 09:05 |
_abc_ | This is "new" in ascii and may have also been in jessie | 09:05 |
_abc_ | https://bbs.archlinux.org/viewtopic.php?id=167104 again systemd hitting under the beltline. Even locate/updatedb is affected | 09:21 |
earthnative | "again" makes it sound like something new. that's nearly a 5 year old thread | 09:27 |
_abc_ | Yes and it still hurts, sort of. | 10:36 |
_abc_ | How does one do the locate non -e thing please? | 10:36 |
_abc_ | I would like to not read source and recompile mlocate from source to make it work | 10:36 |
DocScrutinizer05 | a friend of mine once built a wget that completely ignored robots.txt. I guess when locate/updatedb got lobotopoetterized then a fork is in order | 10:40 |
DocScrutinizer05 | also >> Note that this may slow down the program a lot, if there are many matches in the database.<< will most certainly apply to this case as well. Ohmy | 11:00 |
_abc_ | DocScrutinizer05: hm? Not following? Im Klartext, bitte? | 11:12 |
_abc_ | I am going to run locate under strace to see what it really does. | 11:12 |
_abc_ | So: 1) used devuan does not have /var/lib/mlocate/mlocate.db group access, requires adding devuan to mlocate group or changing permissions on the db file to permit universal read. | 11:25 |
_abc_ | 2) the problem persists, is not solved by changing permissions. | 11:25 |
_abc_ | additionally, /usr/lib/locale/locale-archive exists which contains alternate / cached locale data? | 11:26 |
_abc_ | Ok, so I strace'd locate (mlocate): when run without option -e, the strace shows each found match in the db is tested for access() before being reported, and is not reported without access() success, even if no -e flag is given. This is wrong. | 11:31 |
_abc_ | Need to get and recompile mlocate from source now. I need this feature. what the hell, don't people notice these things? Index unmounted fs's and search in them while offline using locate?! What the HELL. | 11:31 |
DocScrutinizer05 | _abc_: what's unclear? I quoted a similar fsckup in wghet that a friend fixed, then concluded that for such type of fsckup a for and new build is probably the way to go. Then I mused about performance impact ofg the braindamaged poetterization on locate | 11:53 |
DocScrutinizer05 | and locale != locate | 11:55 |
DocScrutinizer05 | locale-archive afaik is a compiled blob of all .po or whatever localizations | 11:55 |
DocScrutinizer05 | re -e I bet that's a 'feature', no bug. "Needed to ensure security" (as Poettering / freedesktop understands it) | 11:58 |
DocScrutinizer05 | you prolly can find a commit on this | 11:58 |
DocScrutinizer05 | note that -e aiui also hides files in non-x dirs from regular users, even when updatedb is run as root | 12:01 |
DocScrutinizer05 | which prolly is the reason to "FIX that BUUUUG!!!!" - idiots | 12:01 |
DocScrutinizer05 | googlw for "poettering: -e must become default and only allowed behavior in locate" | 12:03 |
DocScrutinizer05 | _abc_: https://github.com/crossroads1112/tlocate | 12:09 |
_abc_ | DocScrutinizer05: http://188.26.183.144:8888/w/mlocate-0.26-patch/ | 12:23 |
_abc_ | DocScrutinizer05: the bug was introduced by Red Hat (per author who works there >;), but not by our Pott Pott friend. | 12:24 |
_abc_ | DocScrutinizer05: the -0.26 mlocate is from 2006 or s, per author date put in there. | 12:24 |
_abc_ | Enjoy. Moving on. | 12:24 |
_abc_ | (/me quips something unprintable about high pressure work environments which leave coders no real time to check their private (??) projects) | 12:25 |
_abc_ | DocScrutinizer05: please copy that out and wherever, it is a temp page in a mini box, it might come down soon, like tonight. | 12:26 |
_abc_ | I will put this on my webpages at peter5.50webs.com soon, have other things to organize before that. | 12:26 |
_abc_ | DocScrutinizer05: got it? | 12:29 |
_abc_ | DocScrutinizer05: u:none p:none on that directory | 12:30 |
_abc_ | DocScrutinizer05: emailed original package author and package debian maintainer | 12:49 |
DocScrutinizer05 | who's that? | 12:50 |
DocScrutinizer05 | OA | 12:50 |
DocScrutinizer05 | yeah, got it | 12:50 |
DocScrutinizer05 | _abc_: - locate $known_file_on_umounted -> must not appear ++ --exists | 12:54 |
_abc_ | DocScrutinizer05: no | 12:54 |
_abc_ | yes ok | 12:54 |
DocScrutinizer05 | that small code snippet in patch already makes me cringe | 12:57 |
DocScrutinizer05 | if (*visible == -1) foo; if (*visible != 1) bar; // Who The Fsck writes such code? | 12:57 |
DocScrutinizer05 | it stinks like bugs | 12:58 |
DocScrutinizer05 | note it checks for -1 and 1 | 12:59 |
DocScrutinizer05 | while "visible" to me sounds like a bool | 13:00 |
DocScrutinizer05 | ooh I see what they do | 13:02 |
DocScrutinizer05 | still stinks like bugs | 13:03 |
* DocScrutinizer05 shoots the coder with >>(conf_check_existence != false)<< and "Boolean logic for the braindamaged" - nobody writes "!= false" | 13:05 | |
DocScrutinizer05 | "if (conf_check_existence != false)" :=: "if (conf_check_existence)" | 13:06 |
_abc_ | Package build question: the makefile in the mlocate package makes a locate with wrong locatedb location with default ./configure, even after patching. | 13:08 |
_abc_ | Where is the real debian makefile for this package located? I got: apt-get source mlocate which retrieved mlocate_0.26 | 13:08 |
_abc_ | This is a general devuan/debian package reqbuild question, | 13:09 |
gnarface | use dpkg-buildpackage | 13:11 |
gnarface | you're not supposed to be running ./configure directly in a source package | 13:11 |
gnarface | there are some scripts in the ./debian directory | 13:11 |
gnarface | you can run some of them directly if you want, but dpkg-buildpackage does that for you | 13:12 |
gnarface | it's probably worth mentioning though that even if it weren't a source package, the default path prefix for configure is always wrong, on purpose (/usr/local) | 13:12 |
DocScrutinizer05 | hah | 13:14 |
* DocScrutinizer05 idly wonders where the locatedb lives | 13:15 | |
DocScrutinizer05 | etc/? /var/? | 13:15 |
DocScrutinizer05 | usr/lib/ ? :-o | 13:16 |
djph | according to the manpage /var/lib/mlocate/mlocate.db | 13:16 |
DocScrutinizer05 | sounds sane | 13:17 |
DocScrutinizer05 | http://paste.ubuntu.com/p/f8jdd4QY5r http://paste.ubuntu.com/p/tjPxrKpf2k UMMM http://paste.ubuntu.com/p/pwM2xGRx96 | 13:25 |
DocScrutinizer05 | http://paste.ubuntu.com/p/RzY78Vr6C6 | 13:30 |
_abc_ | How does one recompile a debian package once the source is gotten with apt-get source package ? | 13:33 |
DocScrutinizer05 | http://paste.ubuntu.com/p/yBqzcmWVXx | 13:33 |
fsmithred | _abc_, 'apt-get build-dep' to get the build dependencies | 13:34 |
fsmithred | dpkg-buildpackage -us -uc -b | 13:34 |
_abc_ | fsmithred: the deps are in, it balks on signing and such | 13:34 |
fsmithred | (there are other ways) | 13:34 |
_abc_ | debuild -b -uc -us seems to do it yes | 13:34 |
_abc_ | fsmithred: tell more. Did you see my link and patch? | 13:34 |
fsmithred | no | 13:34 |
fsmithred | I did read some of the discussion | 13:35 |
fsmithred | and I'm confused, because I was using locate yesterday, and it was showing me files that I know were deleted | 13:35 |
fsmithred | does it handle those differently from the ones that just aren't currently mounted and available? | 13:36 |
_abc_ | fsmithred: debuilder is not in the ascii packages | 13:37 |
djph | the cronjob for updatedb only runs on occasion. | 13:37 |
_abc_ | fsmithred: I don't know. What version do you have? My patch certainly fixes the problem, I straced locate to make sense of it | 13:37 |
_abc_ | fsmithred: yes, differently, it checks the path dir by dir then the file | 13:38 |
DocScrutinizer05 | my debian locate is FUBAR | 13:38 |
_abc_ | how? | 13:38 |
DocScrutinizer05 | doesn't find files in my own ~ | 13:38 |
_abc_ | DocScrutinizer05: you need to add used devuan to group mlocate | 13:38 |
_abc_ | *user | 13:38 |
fsmithred | 0.26-2 (ascii) | 13:38 |
_abc_ | this is also a bug | 13:38 |
_abc_ | That's the one I have too fsmithred | 13:38 |
_abc_ | See my comment above, it is different, if the dir exists, it 'succeeds' | 13:39 |
_abc_ | try: mkdir /tmp/footest; touch /tmp/footest/foo | 13:39 |
DocScrutinizer05 | neither does root find files in user's $HOME | 13:39 |
_abc_ | then updtedb (may not look in tmp!), then locate foo; rm -rf /tmp/footest | 13:39 |
fsmithred | I'm not in mlocate group and I see files in my home | 13:39 |
_abc_ | then locate foo; then locate -e | 13:40 |
_abc_ | fsmithred: it uses a cache when that is the case | 13:40 |
_abc_ | wait a second | 13:40 |
gnarface | fsmithred: you don't need it you know. if it annoys you, just get rid of it. | 13:40 |
gnarface | disks seek so fast anyway these days it seems silly | 13:41 |
fsmithred | it's not annoying me. I was just wondering why I didn't see the behavior that was described | 13:41 |
_abc_ | open("/var/lib/mlocate/mlocate.db", O_RDONLY|O_LARGEFILE) = 3 ;; one type of access, if present and readable | 13:41 |
_abc_ | fsmithred: did you try to delete or umount the directory | 13:41 |
fsmithred | oh, locate is way faster than 'find / -name... | 13:41 |
DocScrutinizer05 | http://paste.ubuntu.com/p/8QsGYVfBxn | 13:42 |
fsmithred | I didn't test with unmounted stuff | 13:42 |
DocScrutinizer05 | PRETTY_NAME="Debian GNU/Linux 9 (stretch)" | 13:42 |
_abc_ | fsmithred: of course, and mlocate uses db splay tree insertion on new files, does not waste time on reindexing | 13:42 |
_abc_ | fsmithred: the goal is to index files on backup/usb stick etc and be able to reference the with locate when not mounted | 13:43 |
_abc_ | fsmithred: this is normal use in many cases | 13:43 |
fsmithred | yeah, that's what I would have expected | 13:43 |
_abc_ | It's enough to delete/rename the dir where the file is | 13:43 |
_abc_ | mv ~/tmp/test ~/tmp/test1 | 13:43 |
DocScrutinizer05 | prolly configured to death on Stretch, or killed by systemd | 13:44 |
_abc_ | Anyway, I can't located debuild anywhere fsmithred | 13:44 |
fsmithred | hang on for debuild | 13:44 |
_abc_ | DocScrutinizer05: this one time, no, but the oem author DOES work for Red Hat... | 13:44 |
fsmithred | maybe devscripts | 13:44 |
_abc_ | google can't find it even by file in package beyond etch? | 13:44 |
_abc_ | *wheezy | 13:45 |
DocScrutinizer05 | _abc_: hmm? | 13:45 |
fsmithred | yeah, debuild is in the devscripts package | 13:45 |
_abc_ | DocScrutinizer05: the maintainer is a Norway guy by name, the author worked at Red Hat in 2006 | 13:45 |
DocScrutinizer05 | I don't care author, my locate doesn't find files in ~user/ | 13:45 |
_abc_ | DocScrutinizer05: sudo chmod 644 /var/lib/mlocate/mlocate.db | 13:46 |
DocScrutinizer05 | wait, maybe it got fucked up on dist upgrade | 13:46 |
_abc_ | DocScrutinizer05: then sudo updatedb, and check your ~user is not mounted on something nixed by /etc/updatedb.conf | 13:46 |
DocScrutinizer05 | root@lagrange:~# ls -l /var/lib/mlocate/mlocate.db | 13:46 |
DocScrutinizer05 | ls: cannot access '/var/lib/mlocate/mlocate.db': No such file or directory | 13:46 |
_abc_ | DocScrutinizer05: I think that that is an oem Red Hat features only, bugs not acknowledged, untested, Red Hat thing | 13:47 |
_abc_ | It's a company mentality apparently. | 13:47 |
DocScrutinizer05 | this is STETCH | 13:47 |
_abc_ | DocScrutinizer05: sudo apt-get install mlocate? | 13:47 |
DocScrutinizer05 | o.O | 13:47 |
_abc_ | Oh I don't know about stretch | 13:47 |
_abc_ | DocScrutinizer05: what does your /usr/bin/locate point at | 13:48 |
fsmithred | it's the same mlocate in debian and devuan | 13:48 |
_abc_ | I'm on ascii he's on stretch? | 13:48 |
DocScrutinizer05 | root@lagrange:~# locate --version | 13:48 |
DocScrutinizer05 | locate (GNU findutils) 4.7.0-git | 13:48 |
_abc_ | ^ | 13:48 |
_abc_ | that's the wrong locate, there are 3 | 13:48 |
_abc_ | mlocate slocate and the one you have | 13:48 |
DocScrutinizer05 | I bet it simply doesn'T run updatedb anymore. THANK YOU SYSTEM D | 13:48 |
fsmithred | locate --version | 13:48 |
fsmithred | mlocate 0.26 | 13:48 |
_abc_ | DocScrutinizer05: yeah, it detects the file by remote accessing a tank of genetically modified ni dolphins in California | 13:49 |
DocScrutinizer05 | http://paste.ubuntu.com/p/MJyPbyfRVt | 13:49 |
_abc_ | DocScrutinizer05: so I see it finds it in your ~? | 13:50 |
DocScrutinizer05 | ^^^ | 13:50 |
DocScrutinizer05 | I bet it simply doesn'T run updatedb anymore. THANK YOU SYSTEM D | 13:50 |
_abc_ | DocScrutinizer05: iirc slocate can do that too, not sure | 13:51 |
_abc_ | fsmithred: mlocate performance is quite good enough to use it as general system file finder . | 13:51 |
DocScrutinizer05 | cronjobs nuked | 13:51 |
_abc_ | fsmithred: I mean add 3 buttons and a search text input for gui and all. I do that sometimes. | 13:52 |
DocScrutinizer05 | during upgrade ro stretch | 13:52 |
DocScrutinizer05 | to* | 13:52 |
fsmithred | I use locate a lot. | 13:52 |
_abc_ | locate -ir is a godsend | 13:52 |
_abc_ | So, anyway, I emailed the maintainer and oem author, let's see if this patch gets in the normal releases by "itself" | 13:53 |
DocScrutinizer05 | http://paste.ubuntu.com/p/JGHgPqfzPC | 13:54 |
DocScrutinizer05 | honestly, I don't think 0.26 is the version to use | 13:57 |
DocScrutinizer05 | unless GNU findutils are systemd infested | 13:58 |
_abc_ | http://188.26.183.144:8888/w/mlocate-0.26-patch/ updated the readme, no other changes | 13:59 |
_abc_ | The patch works fine | 13:59 |
DocScrutinizer05 | then, even maemo (2011) used GNU findutils locate 4.4.2 | 14:00 |
_abc_ | I don't think that that is relevant for the current discussion. There is such a thing as backports, and also security patches | 14:01 |
_abc_ | https://packages.debian.org/stretch/mlocate fyi | 14:05 |
_abc_ | 0.26_2 is current | 14:05 |
_abc_ | Do not confuse findutils version with mlocate version | 14:05 |
_abc_ | fsmithred: it appears persistence volumes used on the live system ascii desktop and maybe others are not picked up as persistence unless the volume name is persistence | 14:07 |
_abc_ | fsmithred: is this normal? | 14:07 |
_abc_ | DocScrutinizer05: rlocate may be better for you, depending on what you do | 14:07 |
_abc_ | http://rlocate.sourceforge.net/ you can read the blurb | 14:08 |
_abc_ | Most people do NOT want slocate excepting for users on limited rights accounts. | 14:08 |
fsmithred | _abc_, if the persistent volume name is not persistence, you need to add a boot option to give it the right name | 14:10 |
fsmithred | to use the right name, I mean | 14:10 |
fsmithred | persistence-label=<your-label> | 14:13 |
_abc_ | oh | 14:14 |
_abc_ | okay I missed that | 14:14 |
DocScrutinizer05 | _abc_: why? my locate works | 14:15 |
_abc_ | You said your locate does not work, I think? Before? | 14:16 |
_abc_ | Brb need to eat something | 14:16 |
DocScrutinizer05 | and it's findutils-gnu, not mlocate | 14:16 |
DocScrutinizer05 | on Devian Stretch | 14:16 |
DocScrutinizer05 | on Suse it's 0.25 but works as expected | 14:17 |
DocScrutinizer05 | 0.26 even | 14:17 |
DocScrutinizer05 | what doesn't work on Stretch is the updatedb cronjob | 14:18 |
DocScrutinizer05 | prolly doesn't even exist anymore, gracefully nuked by dist-upgrade | 14:19 |
buZz | man, armbian is such annoying distro | 14:29 |
buZz | the main benefit is the kernels and general 'flashable images' they make | 14:30 |
buZz | but man | 14:30 |
buZz | premade image for (ok, brandnew) board i'm using, has kernel X | 14:30 |
buZz | but they didnt like to offer kernel headers to match that kernel | 14:30 |
DocScrutinizer05 | tz | 14:41 |
DocScrutinizer05 | bootloader blob too? | 14:41 |
msiism | so, devuan ascii offers gnome, right? | 14:42 |
buZz | offers a lot of window managers, gnome is one, yeah | 14:43 |
buZz | DocScrutinizer05: well, allwinner stuff | 14:43 |
DocScrutinizer05 | gnome sans system D? | 14:43 |
buZz | their argument for 'not having to give source' is 'we built it from the patches in this github dir' | 14:43 |
buZz | no branches, etc | 14:43 |
buZz | oh, no? | 14:44 |
buZz | i dno, i never tried gnome since 2 | 14:44 |
DocScrutinizer05 | hmm >>Desktop Environments including XFCE, KDE, MATE, Cinnamon, LXQT (with others available post-install).<< | 14:44 |
buZz | ah yeah Mate, thats the gnome fork | 14:44 |
DocScrutinizer05 | https://devuan.org/os/debian-fork/ascii-rc-announce-050918 | 14:44 |
DocScrutinizer05 | I don't know if it was already gnome2 when I went scream&run | 14:45 |
msiism | ok, right i could have had a look at that myself... | 14:45 |
msiism | btw, this is not about me liking or wanting to use gnome | 14:45 |
msiism | buZz: gnome is not a window manager | 14:46 |
DocScrutinizer05 | eventually I will learn what's the difference between DE and WM | 14:47 |
msiism | it's easy: a window manager is a program that does window management. a desktop environement is a program suite that facilitates all the usual things you'd do on a desktop system. so, a DE will always have to include a wm. | 14:48 |
DocScrutinizer05 | yeah like kwin | 14:48 |
msiism | right | 14:48 |
DocScrutinizer05 | but a WM on its own is pretty pointless, no? | 14:49 |
fsmithred | no | 14:49 |
djph | nah - see i3wm, etc. | 14:49 |
fsmithred | it's a step up from plain console - at minimum, you get a terminal and can run graphical apps | 14:50 |
DocScrutinizer05 | hmm, prolly I could start a xterm in a WM on X11 plain vanilla | 14:50 |
msiism | DocScrutinizer: well, i would rather say yes, it's kind of pointless. | 14:50 |
msiism | but i guess no one is just running window manager. | 14:51 |
msiism | most people "just running a window manager" will have some sort of self-assembled DE around it. | 14:51 |
fsmithred | and the wm itself probably has a menu and sometimes a panel | 14:51 |
msiism | right. that's a bit confusing, though a menu is nice. | 14:52 |
DocScrutinizer05 | I seem to recall I started konqueror as primary process on X11 | 14:52 |
DocScrutinizer05 | worked | 14:52 |
DocScrutinizer05 | remote, even | 14:53 |
msiism | DocScrutinizer05: i always disliked konqueror for being a web browser and file manager in one. feels a bit unsafe. but i never really investigated that. | 14:54 |
gnarface | firefox has plenty of local filesystem access to cause trouble | 14:56 |
DocScrutinizer05 | yeah, as web browser it sucks donkey balls by now | 14:56 |
DocScrutinizer05 | then otoh it's so old and bit rotten, it's prolly immune to any contenporary exloit ;-P | 14:57 |
msiism | i like w3m, it's even being maintained again. | 14:57 |
buZz | so fluxbox is a DE? | 15:04 |
buZz | what are the 'usual things i would do on a desktop system' and why does a DE do that and not a WM? | 15:05 |
djph | DEs tend to use coordinate-based windowing (so they can overlap) | 15:06 |
buZz | eh? so overlapping windows makes it a DE? | 15:06 |
fsmithred | no | 15:06 |
djph | WMs can too, but I think they tend to just tile | 15:06 |
buZz | i think DE = WM + xterm | 15:06 |
buZz | or whatever tools you wanna click on | 15:07 |
muep_ | I'd not generalize it like that. tiling wms certainly are popular but there are dozens of stacking WMs | 15:07 |
djph | Further, AIUI, DEs present more "click-based" utilities. WMs are just "oh, I needed a graphical application that one time" | 15:07 |
fsmithred | DE often has added utilities for admin, search, widgets to add to panel, mounting external drives... | 15:09 |
muep_ | a DE almost certainly has some kind of a panel in the first place. many WMs designed for standalone use have one too but providing the panel does go out of the basic tasks of being a WM | 15:10 |
muep_ | e.g. openbox is a WM that AFAIK is not associated with a wider "DE" project but it still does not have a panel. quite often in screenshots that feature openbox there is then some panel program added by the user | 15:11 |
fsmithred | ability for user to reboot and shutdown is also missing from pure WM | 15:14 |
msiism | those terms... well, generally, even if you just run plain X with an xterm in it, that is a desktop enviironment, but a pretty bare one. | 15:14 |
fsmithred | yeah, there's not a real clear boundary between the two | 15:15 |
msiism | and then, "desktop environment" is being used as a general term for what i'd call a "desktop suite". but i'm not gonna imposa that. :) | 15:15 |
buZz | DE is just a WM with too many dependancies | 15:16 |
msiism | buZz: no | 15:16 |
buZz | :P | 15:16 |
msiism | :D | 15:16 |
msiism | as fsmithred has already sort of mentioned, a DE also has some stuff under the hood that are simply beyond window management | 15:16 |
msiism | for example session managment | 15:17 |
msiism | things to facilitate fast user switching and so on | 15:17 |
muep_ | file manager | 15:17 |
msiism | muep_: well, that's not exactly under the hood, but yes, a good example | 15:18 |
muep_ | yes it's not under the hood but something that almost anything called a DE will supply and which is not typically included in a WM | 15:18 |
msiism | and file managers are also often sort of bound to a De, at least with the major DE's file managers | 15:18 |
muep_ | especially if the DE is one that is designed around the "desktop metaphor" | 15:19 |
msiism | muep_: right, if my wm would include a file manager, i'd flip out. | 15:19 |
muep_ | because then there would often be an instance of the file manager running to provide icons and files on the desktop | 15:19 |
msiism | muep_: that do you mean by "desktop metaphor". the opposite of tiling/tabbing? | 15:20 |
msiism | s/that/what | 15:20 |
muep_ | the opposite to tiling/tabbing imo is just a stacking WM | 15:20 |
msiism | ok, right. but what's desktop metaphor then? | 15:20 |
muep_ | following the desktop metaphor will typically involve using a stacking WM but then there's also the concept of representing the filesystem as folders and icons | 15:21 |
muep_ | and manipulation of those in a way that remotely mimics manipulation of stuff on a real desktop | 15:22 |
msiism | ok, i see. that makes sense | 15:22 |
msiism | (i like icons) | 15:22 |
muep_ | you could use emacs as your window manager and then you could have in your WM a file manager, code editor, some games and a couple of IRC clients... | 15:23 |
msiism | :D | 15:23 |
msiism | did you try? | 15:23 |
muep_ | I like keeping the window manager separate, but there is a package for emacs to do this https://github.com/ch11ng/exwm/wiki/Screenshots | 15:24 |
msiism | that's crazy | 15:24 |
muep_ | I think it's quite awesome but I don't think it's for me | 15:25 |
msiism | ok | 15:25 |
muep_ | I don't live that completely in emacs | 15:26 |
muep_ | this is nowadays often taken as a given, but there was a time when some people had to invent the whole concept https://en.wikipedia.org/wiki/Desktop_metaphor | 15:27 |
msiism | nice. | 15:29 |
msiism | i was recently trying to imagine what an operating system really is by thinking of it in terms of how you'd "run" a house. didn't get too far with that, though. | 15:31 |
DocScrutinizer05 | don't know if somebody said it since >><buZz> i think DE = WM + xterm<< but I think a DE has a task bar and a window bar, start menu, multiple screens etc pp | 15:39 |
DocScrutinizer05 | alt+Tab and whatnot | 15:42 |
msiism | that's window management | 15:42 |
DocScrutinizer05 | sure? | 15:43 |
msiism | what does alt+tab do on your system? | 15:43 |
DocScrutinizer05 | opens a window with a list of window captions and favicons and you can navigate that with alt+Tab down and shift+alt+Tab up. when you releast alt, it opens that window | 15:44 |
DocScrutinizer05 | that window with focus | 15:44 |
DocScrutinizer05 | in list | 15:44 |
msiism | well, i'd consider this window management. and so does my window manager. :) | 15:45 |
DocScrutinizer05 | I think particularly that list window is beyond WM | 15:45 |
DocScrutinizer05 | might be wrong | 15:45 |
msiism | well, in openbox you have fast window switching with alt+tab (or whatever key you bind that to) and also a function that shows a list of all windows on alldesktops. | 15:46 |
Lydia_K | openbox <3 | 15:47 |
msiism | it's very nice indeed | 15:47 |
msiism | btw, to bring this conversation back to more devuan-related issues: if anyone is interested in re-creating the openbox-themes package that Debian had up until wheezy, ping me, i've done some work on this already (but no packaging work...). | 15:49 |
msiism | gotta install a blu-ray recorder on this machine. bbl. | 15:56 |
_abc_ | Hello. Re: updatedb patch and mlocate: I had a discussion with the oem package creator, no longer maintainer "Miloslav Trmac" and he said that the package is okay as is since there's a security issue when applying my patch, users being able to see other people's files | 20:19 |
_abc_ | I tend to disagree, since the check is in there anyway, checking the check check now. | 20:20 |
_abc_ | So it appears he is right (hey he wrote the package), and using just 'updatedb --requre-visibility no' compiles a flag into the db header which does what my patch does. Duh. | 20:27 |
_abc_ | The init came in roundabout ways so I missed it when I dissected the code. | 20:27 |
_abc_ | Okay, this ONE time it wasn't systemd's fault ;) ;) | 20:27 |
_abc_ | As I discussed with the oem author, to get the permissions check working with indexed items on umounted media, one has to second guess the kernel's security resolution mechanism, including any selinx acl's etc. And that is not feasible. | 20:28 |
_abc_ | The other way would be to store uid.gid+perm in the db. This is not done now. | 20:29 |
Criggie | proofpoint is a pretty shitty mailwasher | 22:43 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!