| 00:08.19 | *** join/#devuan Xiku (~Xiku@189.149.122.183) |
| 00:10.02 | *** join/#devuan CenturionDan (~Centurion@office.centurion.net.nz) |
| 00:24.20 | *** join/#devuan Drugo (~andrea@host236-89-dynamic.23-79-r.retail.telecomitalia.it) |
| 00:25.40 | *** join/#devuan Markus_KMi (uid95611@gateway/web/irccloud.com/x-sivdgimziebriwic) |
| 01:11.16 | *** join/#devuan Defiant (erik@x4e361d3e.dyn.telefonica.de) |
| 02:04.24 | *** join/#devuan githogori (~githogori@c-73-70-12-67.hsd1.ca.comcast.net) |
| 02:09.46 | *** join/#devuan Hoshpak (~Hoshpak@p2003005B4B35E700E8654FE916243C63.dip0.t-ipconnect.de) |
| 02:17.57 | *** join/#devuan pecg (~pecg@unaffiliated/pecg) |
| 02:31.48 | *** join/#devuan obarun (~obarun@154.67.14.64) |
| 02:50.36 | *** join/#devuan ffurrywol (~randyg@68-27-253-151.pools.spcsdns.net) |
| 03:14.04 | *** join/#devuan ffurrywol (~randyg@184-194-110-24.pools.spcsdns.net) |
| 03:27.17 | *** join/#devuan AlexLikeRock (~AlexLikeR@unaffiliated/alexlikerock) |
| 03:57.43 | *** join/#devuan pecg (~pecg@unaffiliated/pecg) |
| 04:00.51 | *** join/#devuan DocScrutinizer05 (~saturn@openmoko/engineers/joerg) |
| 04:32.25 | *** join/#devuan towo^work (~towo@unaffiliated/towo/x-4064351) |
| 04:35.06 | *** join/#devuan golinux (~golinux@unaffiliated/golinux) |
| 06:03.18 | *** join/#devuan pav5088 (~markp@124-170-166-25.dyn.iinet.net.au) |
| 06:34.55 | *** join/#devuan wesbl (~wesbl___@95.239.200.116) |
| 07:02.38 | *** join/#devuan Bercik (~Yotsuba@unaffiliated/bercik) |
| 07:04.53 | *** join/#devuan wesbl (~wesbl___@95.239.200.116) |
| 07:12.35 | *** join/#devuan wesbl (~wesbl___@95.239.200.116) |
| 07:37.08 | *** join/#devuan joered (~pingus@93-58-161-247.ip159.fastwebnet.it) |
| 08:03.15 | *** join/#devuan TwistedFate (~TwistedFa@unaffiliated/twistedfate) |
| 08:21.15 | *** join/#devuan gnu_srs (~srs@178-78-231-178.customers.ownit.se) |
| 08:26.18 | *** join/#devuan ruenoak (~chatzilla@122-59-243-80.jetstream.xtra.co.nz) |
| 08:39.49 | *** join/#devuan premoboss (~utente@62.212.13.223) |
| 09:01.10 | *** join/#devuan cocoadaemon (~foo@110.8.19.93.rev.sfr.net) |
| 09:07.01 | *** join/#devuan SylvieLorxu (~TheLastPr@541B7AAC.cm-5-4b.dynamic.ziggo.nl) |
| 09:33.51 | *** join/#devuan wesbl (~wesbl___@95.239.200.116) |
| 09:40.18 | *** join/#devuan sxpert (~sxpert@193.107.127.220) |
| 10:04.43 | *** join/#devuan premoboss (~utente@62.212.13.223) |
| 10:36.05 | *** join/#devuan golinux (~golinux@unaffiliated/golinux) |
| 11:48.13 | *** join/#devuan Humanoid (~anonymous@unaffiliated/humanoid) |
| 11:51.37 | *** join/#devuan wesbl (~wesbl___@95.239.200.116) |
| 11:52.43 | Humanoid | Systemd is a great operating system; it only lacks a decent init system. |
| 11:55.04 | telst4r | how was it ... "cannot start init.service, try journalctl -xn" |
| 12:07.40 | djph | no journals found |
| 12:10.53 | *** join/#devuan Chanku (~Chanku@c-68-52-188-31.hsd1.tn.comcast.net) |
| 12:11.11 | hellekin | Humanoid: nice one :) |
| 12:11.16 | Chanku | ? |
| 12:32.06 | *** join/#devuan kelsoo1 (~kelsoo@86.53.50.178) |
| 12:34.42 | *** join/#devuan amphi (~amphi@59.red-83-54-160.dynamicip.rima-tde.net) |
| 12:34.45 | *** join/#devuan amphi_ (~amphi@59.red-83-54-160.dynamicip.rima-tde.net) |
| 12:35.07 | *** join/#devuan NeonLicht (~rho@shell.bshellz.net) |
| 12:36.52 | *** join/#devuan kelsoo2 (~kelsoo@86.53.50.178) |
| 12:46.25 | telst4r | Here's a dummie: does devuan provide security updates? |
| 12:47.02 | jaromil | yes. masking debian's. working on it currently infrastructure-wise |
| 12:49.39 | telst4r | Great, so in the meantime I could do deb http://security.debian.org and not worry? |
| 12:50.45 | telst4r | The other thing I've been hoping for is to have https apt mirrors in some day :) other than that I'm very pleased with dev1 |
| 12:51.40 | jaromil | best is keep using jessie-security from devuan repos |
| 12:53.18 | telst4r | Oh, there is -security and -updates now =) |
| 12:56.09 | *** join/#devuan Bercik (~Yotsuba@unaffiliated/bercik) |
| 12:59.49 | *** join/#devuan n4dir (~user@pD9EF0224.dip0.t-ipconnect.de) |
| 13:03.48 | jaromil | yep |
| 13:05.45 | *** join/#devuan fsmithred (~fsmithred@68-184-46-18.dhcp.oxfr.ma.charter.com) |
| 13:08.04 | *** join/#devuan clopez (~tau@neutrino.es) |
| 13:08.24 | *** join/#devuan pav5088 (~markp@124.171.233.234) |
| 13:30.36 | *** join/#devuan firegarden (~dionysos@host84-166-dynamic.10-87-r.retail.telecomitalia.it) |
| 13:40.38 | *** join/#devuan pav5088_ (~markp@124-170-190-75.dyn.iinet.net.au) |
| 13:40.49 | *** part/#devuan n4dir (~user@pD9EF0224.dip0.t-ipconnect.de) |
| 13:45.30 | *** join/#devuan kelsoo3 (~kelsoo@86.53.50.178) |
| 13:48.57 | *** join/#devuan kelsoo4 (~kelsoo@86.53.50.178) |
| 13:51.52 | *** join/#devuan TemporalBeing1 (~Ben_Meyer@50.57.17.244) |
| 13:53.15 | *** join/#devuan kelsoo5 (~kelsoo@86.53.50.178) |
| 14:02.06 | *** part/#devuan Humanoid (~anonymous@unaffiliated/humanoid) |
| 14:06.58 | *** join/#devuan kelsoo (~chatzilla@88-144-127-130.host.pobb.as13285.net) |
| 14:14.35 | *** join/#devuan kelsoo5 (~kelsoo@86.53.50.178) |
| 14:18.23 | *** join/#devuan obarun (~obarun@154.67.14.64) |
| 14:20.41 | *** join/#devuan pav5088_ (~markp@124-150-70-212.dyn.iinet.net.au) |
| 14:25.19 | *** join/#devuan furrywolf (~randyg@70-0-101-96.pools.spcsdns.net) |
| 14:29.23 | *** join/#devuan blueness (~blueness@gentoo/developer/blueness) |
| 14:32.27 | *** join/#devuan golinux (~golinux@2605:6000:ebc6:d200:96de:80ff:febc:9ea8) |
| 14:32.27 | *** join/#devuan golinux (~golinux@unaffiliated/golinux) |
| 14:39.35 | DocScrutinizer05 | idly wonders when http://www.pathname.com/fhs/pub/fhs-2.3.html#PURPOSE2 "went out of style", driven by fools who claim /usr/?bin was mandatory for booting a system |
| 14:42.38 | furrywolf | eh? when did that become obsolete? |
| 14:43.18 | DocScrutinizer05 | look at recent distros and how the maintainers .configure --prefix= them. |
| 14:43.31 | DocScrutinizer05 | aka ~poettering |
| 14:44.14 | Lydia_K | That whole converstaion came back up not too long ago, I forget where it ended |
| 14:44.40 | Lydia_K | Personally I agree that /usr should not be required for boot, / should give you a minimal system so that you can recover your trashed /usr |
| 14:45.09 | furrywolf | to be honest, I wouldn't mind a complete re-shaping of the filesystem... create a /something under which each package (or hand-compiled program) has its own directory, in which every single file for the program lives, possibly with bin etc etc dirs. then purging a package is as easy as rm -r /whatever/pkgname... much cleaner than the current state of scattering each package's files all over the system. |
| 14:45.14 | Lydia_K | But I dunno, some young devopts types probably argued docker was necessary for boot and SSDs and boot times or something |
| 14:45.36 | Lydia_K | I've through about that too furrywolf |
| 14:46.05 | Lydia_K | I've considered a structure like that with a crapload (technical term) of symlinks to preserve the old structure for compatibility. |
| 14:46.11 | DocScrutinizer05 | furrywolf: have fun with your PATH |
| 14:46.14 | *** join/#devuan spirou (~usermkaym@78-69-249-162-no180.tbcn.telia.com) |
| 14:46.15 | furrywolf | yep |
| 14:46.26 | Lydia_K | DocScrutinizer05: But symlinks! |
| 14:46.26 | jonadab | I've always used a separate /boot filesystem, admittedly. |
| 14:46.32 | jonadab | That's not the same as requiring /usr though. |
| 14:46.50 | DocScrutinizer05 | Lydia_K: that's done since ages |
| 14:46.50 | telst4r | C:\Program files\my files\ .. :) |
| 14:46.56 | DocScrutinizer05 | the dir called /opt |
| 14:47.15 | Lydia_K | Wait, do we not need a separate /boot anymore? |
| 14:47.20 | jonadab | furrywolf: That creates the kind of PATH management problems Windows command-line users have. |
| 14:47.25 | jonadab | WHich is rather terrible. |
| 14:47.32 | Lydia_K | I mean.. now that I think about it.. you really don't.. |
| 14:47.33 | jonadab | Unless you propose a lot of symlinks. |
| 14:47.36 | Lydia_K | I just always have |
| 14:47.37 | KatolaZ | well Lydia_K, with UEFI you'd probably be better off with a separate /boot anyway.... |
| 14:47.54 | Lydia_K | I very much dislike UEFI |
| 14:48.03 | KatolaZ | me too :D |
| 14:48.04 | jonadab | But yes, if you use a gazillion symlinks, then it could work. |
| 14:48.05 | furrywolf | jonadab: as Lydia_K said, you can make a /usr/bin with a lot of symlinks... or a shell with the ability to parse /foo/*/bin/ in the path. :) |
| 14:48.24 | KatolaZ | technically, it is not needed to have a PATH variable |
| 14:48.25 | DocScrutinizer05 | rotfl |
| 14:48.28 | jonadab | And you just end up with a lot of dead symlinks if you use rm -r to remove packages. Implying, you normally wouldn't do it that way. |
| 14:48.33 | furrywolf | curls up on Lydia_K, still tired after not getting a good night's sleep |
| 14:48.34 | KatolaZ | it's just a common practice |
| 14:48.48 | DocScrutinizer05 | shell patterns in PATH, yeah, what could possibly go wrong? |
| 14:48.50 | jonadab | KatolaZ: Without it, all the shells would have to be taught how to find things. |
| 14:49.07 | Lydia_K | pets furrywolf |
| 14:49.09 | KatolaZ | you could have a shell which looks into all "*/bin/" subfolders of the subfolders of "/opt" |
| 14:49.15 | jonadab | And yes, UEFI is on the list of technologies I want to uninvent. |
| 14:49.29 | Lydia_K | jonadab: a very simple script could just check the filesystem and remove dead symlinks |
| 14:49.35 | jonadab | KatolaZ: You could; but none of the existing shells do that, so it would be a major change. |
| 14:49.41 | jonadab | Lydia_K: True. |
| 14:49.51 | jonadab | Though I personally woudln't want that. |
| 14:49.55 | KatolaZ | jonadab: what I mean is that PATH should not necessariuly be a list of directories... |
| 14:49.58 | Lydia_K | part of the package manager I would imagine |
| 14:50.11 | jonadab | Because sometimes dead symlinks are dead only temporarily (e.g., because a filesystem isn't mounted that normally would be). |
| 14:50.33 | jonadab | But yes, if you still have a real package manager, it could clean up the symlinks when it removes a package. |
| 14:50.36 | jonadab | That would work. |
| 14:50.38 | KatolaZ | you won't need symlinks if your shell was trained to look also for /opt/*/bin... |
| 14:50.39 | KatolaZ | :) |
| 14:50.43 | DocScrutinizer05 | Lydia_K: use a script to clean up the mess caused by not using a decent package manager? sounds like a plan¡ ;-P |
| 14:51.07 | KatolaZ | DocScrutinizer05: you have just described apt, btw :D |
| 14:51.34 | furrywolf | I think systems are a lot more complicated now than they were in 1970 or so... there's a LOT of files in /usr/bin for example on a system with a lot of crap. |
| 14:51.41 | jonadab | apt was a big step forward in its day, though I can't escape the feeling that it could be improved upon. |
| 14:51.42 | Lydia_K | DocScrutinizer05: Still a better solution than systemd. |
| 14:51.46 | furrywolf | one directory made sense when having 50 files in it was normal. :) |
| 14:51.56 | KatolaZ | furrywolf: I agree |
| 14:52.05 | DocScrutinizer05 | furrywolf: huh? |
| 14:52.13 | DocScrutinizer05 | what's yourt problem guys? |
| 14:52.45 | Lydia_K | I was just making a joke. |
| 14:53.01 | spirou | hi |
| 14:53.07 | spirou | what are the chance of a beta2 (with a 32bit dvd) to download within two months? |
| 14:53.23 | Lydia_K | anyway, I don't think there is really an "easy" solution to filesytem structure, hence why people argue about it once in a while, but it rarely changes. |
| 14:53.23 | DocScrutinizer05 | last I heard such rationale was on BS-M where libraries (the 'dirs' of that system) were limuted to namespace fub.aa to fub.zz |
| 14:53.59 | DocScrutinizer05 | and that were 625 possible files |
| 14:54.09 | DocScrutinizer05 | per fub "dir" |
| 14:54.09 | jonadab | Lydia_K: The reason it rarely changes in any major way suddenly is because such changes break existing software that makes assumptions that seemed reasonable at the time it was written. |
| 14:54.18 | spirou | (or actually before 31'th aug while I still can download) |
| 14:55.33 | Lydia_K | jonadab: True in part, backwards compatibility tends to hold back a lot of changes which may or may not be for the better, but I've also seen the various debates around file system structure more than a few times now and I don't think I've ever seen everyone agree on a way to do things. |
| 14:55.38 | DocScrutinizer05 | furrywolf: why do you even _care_ if a dir has 50 or 2000 files in it? |
| 14:55.45 | Lydia_K | spirou: I don't think there are any hard dates, sorry :( |
| 14:56.03 | Lydia_K | It'll be ready when it's ready |
| 14:56.25 | furrywolf | DocScrutinizer51: my sense of proper organization says it's ugly. :) |
| 14:56.32 | KatolaZ | oh, how I like to hear this good old line, Lydia_K :) |
| 14:56.40 | *** join/#devuan fsmithred (~fsmithred@68-184-46-18.dhcp.oxfr.ma.charter.com) |
| 14:56.42 | DocScrutinizer05 | your sense is flawed and not compatible to computers |
| 14:56.48 | KatolaZ | DocScrutinizer05: :D |
| 14:56.53 | jonadab | Lydia_K: Another thing to consider: with the current overall structure, if you have a versioning filesystem, you can easily mark entire trees to be versioned (such as /etc and /home) and not versioned (such as /usr). |
| 14:57.17 | Lydia_K | I'm surprised that hasn't gained more steam |
| 14:57.49 | DocScrutinizer05 | introducing arbitrary cruft entropy into pathnames is clutter and mess and ugly |
| 14:58.41 | furrywolf | is waiting for btrfs to be ready |
| 14:58.46 | furrywolf | has been waiting a while now |
| 14:58.48 | DocScrutinizer05 | while a 3000 files in one dir are pretty much clean, as long as they are all same type |
| 14:59.05 | jonadab | And don't tell me automatic versioning is hard to implement. VMS had it aeons ago. |
| 14:59.18 | KatolaZ | DocScrutinizer05: I might argue that I don't see any problem in having files of several hundred types in a directory with thousands files.... |
| 14:59.34 | KatolaZ | jonadab: versioning is normally inefficient |
| 14:59.41 | KatolaZ | or useless |
| 14:59.49 | KatolaZ | if you think of /usr/bin |
| 14:59.50 | DocScrutinizer05 | KatolaZ: go DOS2 |
| 14:59.51 | Lydia_K | furrywolf: I used btrfs for a while, it's not ready. |
| 14:59.53 | KatolaZ | for instance |
| 15:00.00 | Lydia_K | the experience was *horrible* |
| 15:00.04 | DocScrutinizer05 | one drive, one file hierarchy level |
| 15:00.04 | KatolaZ | how often do you change stuff in there? |
| 15:00.23 | jonadab | KatolaZ: That's why you want /usr, or at least most of /usr, to be marked for "only keep one version of each file". |
| 15:00.46 | jonadab | (Which you can override for a subdirectory or individual file if there's a reason to do so.) |
| 15:00.49 | KatolaZ | DocScrutinizer05: you actually do have several dozens types of files in your /usr/bin anyway |
| 15:01.02 | DocScrutinizer05 | meh, sophism |
| 15:01.31 | DocScrutinizer05 | I can easily make up another 300 types |
| 15:02.01 | DocScrutinizer05 | alls those files in /usr/bin that start with a and have 4 chars in filename |
| 15:02.05 | DocScrutinizer05 | etc pp |
| 15:02.09 | DocScrutinizer05 | that's pointless |
| 15:02.19 | furrywolf | mch prefers saphism |
| 15:02.47 | KatolaZ | DocScrutinizer05: ? |
| 15:03.12 | DocScrutinizer05 | KatolaZ: make your point, so far xyou didn't |
| 15:03.29 | spirou | I was more thinking about if someone got a guesstimate like "probably" - "maybe" - "it could happen" - "not a chance" sort of |
| 15:04.03 | KatolaZ | DocScrutinizer05: you said you have no problems with 3000 files in one directory, as long as they are of the same type |
| 15:04.12 | DocScrutinizer05 | so? |
| 15:04.24 | jonadab | Define "same type" |
| 15:04.28 | DocScrutinizer05 | why? |
| 15:04.29 | KatolaZ | so I don't mind the number of the files *and* the type |
| 15:04.37 | jonadab | What if they're all images, but some are in .png format and some are .jpg? |
| 15:04.47 | KatolaZ | I don't mind jonadab |
| 15:04.48 | DocScrutinizer05 | it's my problem, so why should I define the conditions of it (not) existing? |
| 15:05.02 | KatolaZ | DocScrutinizer05: I am not arguing with you mate :) |
| 15:05.08 | KatolaZ | just making my point |
| 15:05.13 | KatolaZ | that is: I don;t mind |
| 15:05.19 | Leander256 | I'm still waiting for winfs to ship with the next version of windows |
| 15:05.23 | KatolaZ | the organisation of the filesystem is just in our minds |
| 15:05.35 | KatolaZ | since unix provides only flat access |
| 15:05.36 | KatolaZ | fullstop |
| 15:05.45 | DocScrutinizer05 | jonadab: then they're all images, but some are in .png format and some are .jpg |
| 15:05.55 | KatolaZ | so we will never agree about "the best organisation" |
| 15:06.10 | KatolaZ | since there is no such thing, IMHO |
| 15:06.14 | KatolaZ | there are common practices |
| 15:06.26 | KatolaZ | you might find some of them good, and some bad |
| 15:06.44 | Lydia_K | Leander256: LOL! |
| 15:07.19 | DocScrutinizer05 | you might find yourself suggesting weird new cluttered practices to group files in many dirs, just to invent something new that has no advantage whatsoever |
| 15:07.35 | KatolaZ | DocScrutinizer05: never said anythong close to that :) |
| 15:07.40 | KatolaZ | please read again the thread |
| 15:07.43 | Lydia_K | The problem with the abstract concept of organization is that everyone has a different preference as to which organization structure "feels" best, so there will never be one answer for everyone, it's like window managers, everyone has different tastes, styles, and workflows. |
| 15:07.55 | DocScrutinizer05 | well, I started it |
| 15:08.06 | furrywolf | I vote we store all files with "systemd" in the name in their own special location, /dev/null |
| 15:08.09 | KatolaZ | I was saying that in principle it is not difficult to have separate dirs for separate applications |
| 15:08.25 | KatolaZ | and it is not impossible to teach a shell to get binaries from a glob pattern |
| 15:08.26 | Leander256 | seriously though, I find the idea of unorganized files very interesting, as long as they are tagged correctly, and creating all sorts of symlinks to files can be seen as a cheap tagging |
| 15:08.32 | DocScrutinizer05 | I doubt I will reread it to find out how it's relevant that you think the files in /usr/bin have several different properties |
| 15:09.00 | KatolaZ | nevermind DocScrutinizer05 |
| 15:09.01 | KatolaZ | :) |
| 15:10.56 | furrywolf | and when are we getting proper acls anyway? :) |
| 15:10.57 | Lydia_K | This whole conversation makes me want to make a distro in which file locations are randomized throughout the filesytem |
| 15:11.04 | furrywolf | lol |
| 15:11.05 | Lydia_K | So every install is it's own unique mess |
| 15:11.19 | furrywolf | devrandomuan! |
| 15:11.26 | Lydia_K | LOL |
| 15:11.29 | KatolaZ | Lydia_K: :D |
| 15:11.31 | Lydia_K | I should do that for fun :D |
| 15:11.47 | DocScrutinizer05 | it's still standing that http://www.pathname.com/fhs/pub/fhs-2.3.html#PURPOSE2 and http://www.pathname.com/fhs/pub/fhs-2.3.html#PURPOSE16 provides a pretty sane rationale why /usr is supposed to be a separate filesystem, and nobody provided a sound argument why this gets nuked now by Poettering and friend |
| 15:11.48 | Lydia_K | Total joke distro, but it'd be cool to solve all the various problems with scattering files everywhere |
| 15:12.10 | furrywolf | DocScrutinizer51: solution: don't install anything poettered. |
| 15:13.49 | KatolaZ | DocScrutinizer05: that is a common practice, again, not anything written in stone |
| 15:13.58 | KatolaZ | I am not defending Poettering in any way |
| 15:14.26 | KatolaZ | I am just saying that whatever we call "unix-gospel" is the result of practice |
| 15:14.46 | KatolaZ | and as practice change, the gospel could in principle be rewritten |
| 15:15.05 | KatolaZ | I simply argue that all the changes should be done for a precise reason, or discarded |
| 15:15.13 | Lydia_K | charages KatolaZ blasphemy |
| 15:15.17 | KatolaZ | :D |
| 15:15.19 | Lydia_K | The sentance is systemd |
| 15:15.29 | KatolaZ | read the last one, please :) |
| 15:15.39 | Lydia_K | LOL |
| 15:15.54 | KatolaZ | and I should add: any change cannot break working stuff |
| 15:15.55 | Leander256 | they nuked it because systemd can't boot the distro without its thousands of files in /usr |
| 15:16.09 | KatolaZ | and that's not a good reason IMHO, Leander256 |
| 15:16.24 | Leander256 | it's as simple as that, and it's totally retarded IMHO |
| 15:16.25 | KatolaZ | this is just second order complexity, due to first order madness |
| 15:16.41 | KatolaZ | Leander256: agree |
| 15:17.06 | Leander256 | DocScrutinizer05 by the way this happened at least a couple years ago |
| 15:17.12 | KatolaZ | Lydia_K: what's the punishment for unix blasphemy these days? :) |
| 15:17.12 | DocScrutinizer05 | >>should be done for a precise reason<< is not sufficient, this precise reason also needs to be higher value than the reason why it's been done different before. For the latter see above URLs pointing to the rationale given in FHS which not been "practice" but long and careful evaluation |
| 15:17.36 | KatolaZ | DocScrutinizer05: I agree with you |
| 15:17.48 | Leander256 | it was what made me realise how nasty this systemd thing was getting |
| 15:17.52 | DocScrutinizer05 | Leander256: when a file systemd needs is in /usr then the distro maintainer misplaced it |
| 15:18.04 | KatolaZ | but you should admit that "higher value" is a subjective quantity |
| 15:18.14 | KatolaZ | that's why we are here |
| 15:18.18 | DocScrutinizer05 | obviously per definitionem any such file belongs into rootfs |
| 15:18.37 | KatolaZ | if it was scientifically proven that systemd was the only logical way forward, I wouldn't be here |
| 15:19.06 | KatolaZ | in reality, my opinion on the matter is just an opinion, and yours as well |
| 15:19.22 | *** join/#devuan level7 (~quassel@31.44.17.250) |
| 15:19.38 | KatolaZ | I still find it amazing how consensus can emerge out of so many different minds ;) |
| 15:19.50 | DocScrutinizer05 | you should admit that >>systemd can't boot the distro without its thousands of files in /usr<< is a flawed rationale anyway that can't override the rationale that any sich file doesn't belong into /usr to start with |
| 15:19.53 | Lydia_K | If you don't care about having a unix box and just want "an OS" then systemd is perfectly rational |
| 15:20.09 | DocScrutinizer05 | such* |
| 15:20.15 | KatolaZ | DocScrutinizer05: I believe that it is a flawed rationale, but I have no scientific proof for that |
| 15:20.22 | KatolaZ | I have only *opinions* about that |
| 15:20.38 | Lydia_K | I agree DocScrutinizer05, but that's really a question for the poettering crowd as to why it is the way it is |
| 15:20.38 | DocScrutinizer05 | well, when you think logic is a notion |
| 15:21.05 | KatolaZ | no, I think that all this stuff does not obey to a single logic |
| 15:21.23 | KatolaZ | simply because the axioms used by Pettering are different from those I chose to use |
| 15:21.35 | KatolaZ | hence, I am not surprired that I don't agree with him |
| 15:22.00 | KatolaZ | in his *logic*, all these steps are part of a coherent plan |
| 15:22.02 | DocScrutinizer05 | that's pretty much a fork of our topic :-) |
| 15:22.17 | furrywolf | ah, but what is the plan? :P |
| 15:22.21 | KatolaZ | maybe :) |
| 15:22.36 | KatolaZ | furrywolf: free beer for everyone, wasn't it? |
| 15:22.37 | KatolaZ | :D |
| 15:22.40 | furrywolf | has a simple plan... keep being a lapwolfy until forced to go do useful things |
| 15:22.55 | DocScrutinizer05 | when Poettering's plann is "let's play idiot and deny the rationale in FHS" then ok, your point |
| 15:23.35 | Lydia_K | I think his rational is that he wants to control the entire core of all linux distros |
| 15:23.45 | Lydia_K | In which case, bravo. |
| 15:23.47 | KatolaZ | precisely Lydia_K |
| 15:23.48 | DocScrutinizer05 | poettering's plan free beer? well: |
| 15:23.50 | DocScrutinizer05 | ~freee |
| 15:23.50 | infobot | from memory, freee is http://www.youtube.com/watch?feature=player_detailpage&v=ZTdUmlGxVo0&t=3296 |
| 15:24.24 | KatolaZ | and in this plan, systemd, pulseaudio, kdbus, and the rest, fit absolutely well |
| 15:24.38 | KatolaZ | they even seem *logical* |
| 15:25.02 | DocScrutinizer05 | yes, of course. Still he could have moved files from /usr/bin to /bin, instead of moving /usr to / |
| 15:25.11 | furrywolf | yes, can we please please make pulseaudio go away too? |
| 15:25.20 | *** join/#devuan pecg (~pecg@200.107.212.122) |
| 15:25.20 | *** join/#devuan pecg (~pecg@unaffiliated/pecg) |
| 15:25.29 | Lydia_K | From his perspective things are going very well! The vast majority of major distros have giving him personal access to PID 1 on all of their installs. |
| 15:25.39 | KatolaZ | exactly |
| 15:25.54 | KatolaZ | he has tons of reasons to think that he is right, and is doing very well overall |
| 15:25.57 | Lydia_K | hate him or not, you can't argue that he hasn't been successful. |
| 15:26.03 | KatolaZ | yep |
| 15:26.50 | Lydia_K | pets furrywolf |
| 15:26.59 | Lydia_K | And hate him I do.. |
| 15:27.03 | KatolaZ | then, also Steve Jobs was successful, but this didn't stop me from disagreeing with almost anyhting he did since 1979 |
| 15:27.16 | furrywolf | mrrf :) |
| 15:27.20 | Lydia_K | Ugh, I hate the way people think steve jobs was so great |
| 15:27.26 | Lydia_K | and how "he invented" all this crap |
| 15:27.39 | Lydia_K | he just yelled at very talented engineers who invented crap |
| 15:27.57 | *** join/#devuan rafalcpp (~racalcppp@84-10-11-234.static.chello.pl) |
| 15:28.19 | Lydia_K | he didn't do shit but wear those stupid turtlenecks and deny scientific progress in medicine all the way into his grave. |
| 15:28.43 | telst4r | I actually appreciate Woz more than Jobs. |
| 15:28.51 | Lydia_K | Well, that's not just, he was actually pretty damned good at marketing. |
| 15:29.00 | Lydia_K | s/just/true/ |
| 15:29.09 | DusXMT | Marketing, Apple's biggest strength |
| 15:29.13 | Lydia_K | Woz <3 |
| 15:29.24 | DocScrutinizer05 | Woz was the genius |
| 15:29.34 | DocScrutinizer05 | Steve the marketeer |
| 15:30.16 | Lydia_K | He used to get home from high school, take whatever the trade magazine was that had the latest chip designs and stuff in it, and sit there with pen and paper and design *better* versions |
| 15:30.21 | Lydia_K | like, holy crap. |
| 15:32.09 | DusXMT | It's funny how Jobs was even fired from Apple once, because of his poor decisions in the company, like shifting their focus from the next-gen Apple 2 computers to the far-inferior Mac (at the time inferior) |
| 15:32.17 | telst4r | exactly. That's what I find interesting in him. and he's talks. How much enthusiasm! |
| 15:32.47 | Lydia_K | I saw him talk at HOPE one year, the Woz was awesome. |
| 15:33.55 | DocScrutinizer05 | stack pop: could we make sure in Devuan /(s)bin _never_ becomes a symlink to /usr/(s)bin ? |
| 15:34.23 | furrywolf | as long as we make sure all poettered software doesn't exist, sure. |
| 15:36.04 | DusXMT | The main reason to have /(s)bin is to have a usable system even when /usr is not mounted; making it a symlink would make it impossible to do system recovery without a live CD, if one has a separate /usr partition |
| 15:36.24 | DusXMT | So it's illogical |
| 15:36.31 | *** join/#devuan rafalcpp (~racalcppp@84-10-11-234.static.chello.pl) |
| 15:38.12 | telst4r | "oh, everybody has a usb key these days so stop worrying" |
| 15:38.31 | DusXMT | (but funny enough, on System V, if you didn't have a separate /usr partition, it would be a symlink to /) |
| 15:43.20 | DocScrutinizer05 | DusXMT: exactly my point. When Poettering needs stuff from /usr/bin for his malware, the distros adopting that malware rather should move stuff that's needed from /usr/bin to /bin, since that's why we got it. Moving /usr to / just because... it's a very obvious evidence of not understanding the concept of /usr at all |
| 15:43.22 | *** join/#devuan asbesto (~asbesto@88-149-239-187.v4.ngi.it) |
| 15:44.39 | Lydia_K | I think part of the problem is that many of the younger kids don't learn about the history, so they don't understand why things are done they way they are. |
| 15:44.54 | DusXMT | DocScrutinizer05: Just out of curiosity, what is it that they're moving? What kind of files are appearing in /? |
| 15:45.12 | Lydia_K | DusXMT: /usr is required for boot with systemd |
| 15:45.24 | Lydia_K | so /usr is part of / now |
| 15:45.35 | DusXMT | Lydia_K: But we're in a systemd-free, yes? |
| 15:45.44 | DusXMT | *systemd-free zone |
| 15:45.47 | Lydia_K | We are yes |
| 15:46.51 | zdzichu | Lydia_K: no it isn't |
| 15:47.00 | zdzichu | why people won't stop spreading misinformation? |
| 15:47.06 | zdzichu | that won't help devuan's case |
| 15:47.41 | Lydia_K | no it isn't what? |
| 15:50.00 | zdzichu | [174512] <Lydia_K> DusXMT: /usr is required for boot with systemd |
| 15:50.23 | zdzichu | you are constantly trying to establish association "devuan == lies" |
| 15:50.28 | zdzichu | whose agent are you? |
| 15:50.33 | zdzichu | you are not welcome here |
| 15:50.48 | furrywolf | ... |
| 15:51.12 | Lydia_K | Well that's the converstaion we've been having for the last hour. Yes, you can build it and *put* the fiels in / (where they should be) but it seems the arguement has been that /usr should just be required for boot, and most distros have been fine with that. |
| 15:51.17 | Lydia_K | who are you zdzichu? |
| 15:51.20 | Lydia_K | I've been here for months. |
| 15:51.36 | furrywolf | from what I've seen, I'll take Lydia_K over zdzichu any day. |
| 15:52.32 | *** join/#devuan rafalcpp (~racalcppp@84-10-11-234.static.chello.pl) |
| 15:52.34 | Lydia_K | I mean, if I'm wrong if what I've head about this history of this particular debate regarding filesystem structure feel free to correct me, but this is what I've heard |
| 15:53.00 | *** join/#devuan thehornet (~thehornet@192.91.75.30) |
| 15:53.19 | Lydia_K | but just throwing around mean sprited accusations is really unproductive. |
| 15:54.35 | Lydia_K | s/if/in/ |
| 15:54.42 | furrywolf | googling suggests /usr/lib/systemd is required for basic boot tasks, including running getty. |
| 15:54.56 | Lydia_K | I just can't type right lately. |
| 15:55.24 | Lydia_K | And yes, I'm sure distros could configure --prefix, which is what they *SHOULD* be doing to put it all in root, which was said more than once over the last hour here. |
| 15:55.33 | *** join/#devuan drogue (~rocket@cust-44148.ptera.net) |
| 15:58.30 | Lydia_K | Well that was interesting. |
| 16:00.44 | furrywolf | lol |
| 16:01.41 | Lydia_K | pokes zdzichu "You alive?" |
| 16:02.53 | Lydia_K | I killed him :( |
| 16:03.20 | furrywolf | logic is often injurious to trolls |
| 16:04.52 | Lydia_K | managed to execute the entire conversation though |
| 16:05.13 | Lydia_K | whiiiccch.. probably should have been in #debianfork.. |
| 16:05.53 | *** join/#devuan buZz (~buzz@space.nurdspace.nl) |
| 16:06.49 | furrywolf | nah, this conversation was very on-topic. |
| 16:07.02 | furrywolf | and it makes the channel non-dead, which is a good thing. |
| 16:07.33 | DocScrutinizer05 | assuring that Devuan sticks with FHS is on topic I'd say |
| 16:08.11 | jaromil | FHS stays to Devuan like water to Venice, I'd say |
| 16:08.11 | DocScrutinizer05 | when Poettering/systemd is introducing BS to file hierarchy, Devuan shouldn't adopt it |
| 16:08.27 | DocScrutinizer05 | jaromil: :-D |
| 16:08.44 | jaromil | rowing rowing rowing |
| 16:10.00 | muep | my impression is that sticking with FHS would not guarantee that /bin is not /usr/bin , but of course the traditional split can be retained otherwise |
| 16:11.34 | Lydia_K | For the moment I think we need to stay focused on the goal of getting a sysvinit debian fork working and stable, then from there we can talk more about whatever changes we may or may not want to make in regards to filesystem layouts and init systems and whatever else. |
| 16:12.02 | DocScrutinizer05 | anyway (which triggered my ranting) maemo is an excellent example of the mess that's ensuing when ignoring FHS and doing the wrong thing regarding exactly /usr |
| 16:12.03 | KatolaZ | Lydia_K for president! |
| 16:12.06 | Lydia_K | Not to suggest we shouldn't talk about it, but just to make it clear that it's nothing that needs to be decided today |
| 16:12.07 | DocScrutinizer05 | ~optification |
| 16:12.07 | infobot | optification is a inventive duct tape workaround to reclaim space in fs root, done due to the fact the systeminit *and* partitioning is FUBAR, http://wiki.maemo.org/Documentation/Maemo_5_Developer_Guide/Packaging,_Deploying_and_Distributing/Installing_under_opt_and_MyDocs, or ""OMG - I wish they looked into FHS and moved /usr to eMMC"", http://www.pathname.com/fhs/pub/fhs-2.3.html#PURPOSE2 bullet1,2 and fhs-2.3.html#PURPOSE16 dot3" |
| 16:12.26 | jaromil | votes Lydia_K |
| 16:12.30 | Lydia_K | :D |
| 16:12.37 | KatolaZ | votes twice Lydia_K |
| 16:16.06 | jaromil | KatolaZ: i'm cleaning up a bit sdk and bringing out the libdevuansdk submodule |
| 16:16.14 | DocScrutinizer05 | http://paste.opensuse.org/26138829 for the shits n giggles (and the agony) |
| 16:16.37 | Lydia_K | DusXMT: Well, to be fair, they had to work around a 100Meg limit |
| 16:16.41 | Lydia_K | I mean DocScrutinizer05 |
| 16:16.51 | DocScrutinizer05 | so what? |
| 16:17.10 | DocScrutinizer05 | ""OMG - I wish they looked into FHS and moved /usr to eMMC"" |
| 16:17.34 | DocScrutinizer05 | exactly our topic we discussed, just maemo instead systemd |
| 16:17.57 | Lydia_K | eMMC? |
| 16:19.20 | KatolaZ | jaromil: sure, I was going to do that at the first opportunity |
| 16:19.32 | KatolaZ | but I haven't had much time lately |
| 16:19.34 | DocScrutinizer05 | when *anything* /usr/* is needed during early boot, before mounting eMMC /opt volume, it's the damn duty of distro maintainers to move that stuff that's needed during boot to rootfs (/bin, /sbin, /lib ...) instead of placing /usr into rootfs and flooding /usr/bin with symlinks |
| 16:19.42 | muep | Lydia_K: I guess it is embedded MMC |
| 16:19.55 | Lydia_K | Ahh, of course. |
| 16:20.03 | DocScrutinizer05 | eMMC = embedded MMC aka flash memory |
| 16:21.23 | muep | the way to address it that seems to be becoming popular is to just add the symlinks from /bin -> usr/bin and /sbin -> usr/sbin and /lib -> usr/lib |
| 16:22.05 | Lydia_K | I think I'm missing something cause the link I'm reading says they moved stuff to /opt, at any rate, I'm a lot more forgiving of breaking conventions with "creative soltuions" when faced with hardware limitations like a 100Meg rootfs. |
| 16:22.12 | muep | of course it has some downsides, but I'd not consider three symlinks an amount that would count as pollution |
| 16:22.51 | *** join/#devuan jsfarinet (d9fabea7@gateway/web/freenode/ip.217.250.190.167) |
| 16:24.30 | furrywolf | am I the only one who thinks 100mb should be plenty for a rootfs? heh |
| 16:24.32 | spirou | a distro shouldn't really put stuff in /op/ should it? isn't that where people would put their own stuff in... softwares not in the repository |
| 16:24.46 | spirou | <PROTECTED> |
| 16:25.23 | Lydia_K | furrywolf: That's a good point, a minimal rootfs should be fine to boot *AND* fit in 100Megs, then the rest can go in /usr /opt whatever. |
| 16:25.38 | Lydia_K | I do agree spirou, they should have put /usr where they put /opt |
| 16:25.56 | telst4r | how about embedded things like routers? |
| 16:26.02 | muep | I think they wanted a bit less minimal OS in the fast small device |
| 16:26.17 | Lydia_K | telst4r: Those should be like busybox style minimal systems anyway, so they should be tiny |
| 16:26.51 | Lydia_K | my impression (not knowing much about Maemo) is that it's trying to be a rather robust, full OS |
| 16:27.12 | Lydia_K | I could be totally wrong about that, but that's just the impression I get without digging into it any deeper. |
| 16:28.51 | furrywolf | has no idea what maemo is |
| 16:29.02 | DocScrutinizer05 | http://paste.opensuse.org/14960612 |
| 16:29.41 | telst4r | I had a Maemo5 phone. Nokia N900 |
| 16:29.46 | Lydia_K | I think I understand what you are saying now DocScrutinizer05 |
| 16:30.17 | DocScrutinizer05 | Lydia_K: you still miss the point? they should have moved whole /usr to /opt, and instead moved the stuff that's _not_ a symlink in /usr/* to /* instead |
| 16:30.19 | Lydia_K | they didn't move /usr to the internal memory where there was 25gigs of space, they put opt there and symlinked the fuck out of everything and left /usr on / |
| 16:30.39 | Lydia_K | Yes, I get your point now :) |
| 16:30.44 | DocScrutinizer05 | /opt is on eMMC and mounted late |
| 16:30.58 | DocScrutinizer05 | /usr is in rootfs |
| 16:31.00 | Lydia_K | Gosh that is dumb, they moved that whole filesystem there via symlinks instead of just moving it there. |
| 16:31.15 | telst4r | haha |
| 16:32.17 | DocScrutinizer05 | http://paste.opensuse.org/36102350 |
| 16:36.47 | DocScrutinizer05 | yes, exactly |
| 16:37.21 | DocScrutinizer05 | maemo been systemd-ready (at least regarding file hierarchy) a 7 years ago already ;-P |
| 16:38.05 | DocScrutinizer05 | so I'm very happy we won't see anything faintly resembiing that, in devuan |
| 16:39.28 | *** join/#devuan Akuli (~Akuli@mobile-access-bceebf-7.dhcp.inet.fi) |
| 16:41.21 | *** join/#devuan kelsoo (~chatzilla@88.144.91.216) |
| 16:41.46 | Lydia_K | Certainly not |
| 16:42.09 | *** join/#devuan Pali (~pali@Maemo/community/contributor/Pali) |
| 16:42.32 | *** join/#devuan Tenhi (~tenhi@static.177.80.201.138.clients.your-server.de) |
| 16:59.47 | *** join/#devuan jsfarinet (d9fabea7@gateway/web/freenode/ip.217.250.190.167) |
| 17:00.10 | DocScrutinizer05 | re >>hardware limitations like a 100Meg rootfs<< Nokia would have done better with a proper /usr volume instead of optification mess. Would have resulted in more space saved in rootfs and way simpler management |
| 17:03.11 | *** join/#devuan n4dir (~user@pD9EF0224.dip0.t-ipconnect.de) |
| 17:03.35 | Lydia_K | I totally agree now that I understand better. |
| 17:04.09 | Lydia_K | Previously I was under them impression that they moved stuff from root to opt cause there wasn't enough space, I didn't realize they move /usr to /opt instead of just moving /usr. |
| 17:11.11 | DocScrutinizer05 | http://paste.opensuse.org/51727521 |
| 17:12.10 | Lydia_K | That's just dumb, they could have had one symlink instead of a thousand |
| 17:12.33 | Lydia_K | especially since they are wasting 4k (assume default block size) with every symlink |
| 17:12.40 | Lydia_K | *assuming |
| 17:12.48 | DocScrutinizer05 | no, since then they would move the few files that are needed during boot out of rootfs |
| 17:14.12 | DocScrutinizer05 | I.E. all files in /user/bin (and sbin) that are _not_ symlinks: http://paste.opensuse.org/99944179 |
| 17:14.58 | DocScrutinizer05 | all those files rather should have been move from /usr/* to /* and then move whole remaining /usr dir to /home fs |
| 17:15.36 | DocScrutinizer05 | according to "anything in /usr that's needed during early boot doesn't belong where it is" |
| 17:15.53 | Lydia_K | damn that's a lot of stuff in /usr that's needed for boot |
| 17:16.00 | DocScrutinizer05 | indeed |
| 17:16.25 | DocScrutinizer05 | that's due to the fact maemo is for a phone, where even bootscreen is a movie |
| 17:16.29 | DocScrutinizer05 | with audio |
| 17:16.39 | Lydia_K | Ah, true. |
| 17:16.46 | Lydia_K | I hate that about phones so much. |
| 17:16.52 | DocScrutinizer05 | yep |
| 17:16.57 | DocScrutinizer05 | user experience |
| 17:17.40 | DocScrutinizer05 | much in line with poetterisms like "you need audio bootscreen, for vision impared users" |
| 17:18.14 | DocScrutinizer05 | "you need I18N so bootscreen is in your language" |
| 17:18.16 | *** join/#devuan wildlander (~wild@unaffiliated/wildlander) |
| 17:18.34 | *** join/#devuan TwistedFate (~TwistedFa@unaffiliated/twistedfate) |
| 17:19.00 | DocScrutinizer05 | "you need camera and face recognition during boot, to authenticate to BIOS config screen" ;-P |
| 17:19.18 | telst4r | you don't need to load 30 secs for that bootscreen and play a 45sec video on it, when you can just start up and play a simple "blingbling" when it's ready. |
| 17:20.10 | DocScrutinizer05 | yes, a lot of maemo N900 users nuked that shaking hands video |
| 17:20.51 | telst4r | i wonder if xLennart works on N900... |
| 17:23.01 | DocScrutinizer05 | the video load audio which loads PA which loads all codecs. Somebody moved one codec to /opt so it wasn't available during early boot. PA never recovers and thus audio was broken permanently (until next boot with codec available) |
| 17:23.37 | DocScrutinizer05 | it's aterrible mess |
| 17:32.34 | *** join/#devuan SuicideJunkie (~nick@CPE000f6638c2b8-CM0c473de90d90.cpe.net.cable.rogers.com) |
| 17:37.51 | *** join/#devuan SylvieLorxu (~TheLastPr@541B7AAC.cm-5-4b.dynamic.ziggo.nl) |
| 17:40.19 | *** join/#devuan firegarden (~dionysos@host84-166-dynamic.10-87-r.retail.telecomitalia.it) |
| 17:45.06 | *** join/#devuan n4dir (~user@pD9EF0224.dip0.t-ipconnect.de) |
| 17:45.22 | DocScrutinizer05 | ((a lot of stuff in /usr that's needed for boot)) I have to admit that there's probably a lot of stuff that simply didn't get optified despite it should have been |
| 17:46.26 | Akuli | i wonder what would happen if i'd actually erase my /usr |
| 17:46.26 | *** join/#devuan level7_ (~quassel@31.44.17.250) |
| 17:46.48 | DocScrutinizer05 | on maemo? |
| 17:47.01 | Akuli | on any distro, lets say devuan |
| 17:47.13 | Akuli | i'm pretty sure i'd get a bash prompt... but its not really enough to get things done anymore :) |
| 17:47.18 | DocScrutinizer05 | on devuan the impact should be rather minor |
| 17:48.13 | Akuli | isn't X and DE and all that stuff in /usr? |
| 17:48.24 | DocScrutinizer05 | everything core utils has to be on rootfs |
| 17:48.41 | DocScrutinizer05 | X11 for sure is not rootfs |
| 17:48.42 | Akuli | usually coreutils is in /bin |
| 17:49.11 | DocScrutinizer05 | yes, and /bin is rootfs |
| 17:49.17 | DocScrutinizer05 | see FHS |
| 17:50.15 | DocScrutinizer05 | http://www.pathname.com/fhs/pub/fhs-2.3.html#PURPOSE2 |
| 17:52.55 | DocScrutinizer05 | and http://www.pathname.com/fhs/pub/fhs-2.3.html#PURPOSE16 |
| 17:54.47 | DocScrutinizer05 | note that >>Locally-installed system administration programs should be placed into /usr/local/sbin<< holds a gotcha regarding any such system admin tools that are needed specifically to recover this particular system |
| 17:54.58 | *** join/#devuan pecg (~pecg@200.107.212.122) |
| 17:54.58 | *** join/#devuan pecg (~pecg@unaffiliated/pecg) |
| 17:55.00 | *** join/#devuan pav5088_ (~markp@124-150-93-40.dyn.iinet.net.au) |
| 17:58.00 | DocScrutinizer05 | (proprietary) tools needed to manage hw RAID controller come to mind |
| 17:59.14 | Lydia_K | Stuff like that is really on a case by case basis |
| 17:59.23 | DocScrutinizer05 | yep |
| 17:59.26 | Lydia_K | "Would you need this if /usr/ got nuked?" yes/no |
| 17:59.36 | Lydia_K | fdisk, yes! |
| 17:59.45 | Lydia_K | gconf-editor, f-no |
| 17:59.51 | DocScrutinizer05 | hehe |
| 18:01.09 | Akuli | hmm |
| 18:01.15 | Akuli | gnu parted is in /sbin :D |
| 18:02.20 | Lydia_K | considering last time I checked fdisk didn't support gpt partition tables it might be necessary these days |
| 18:04.27 | telst4r | only if you have 2TB+ harddrives. Which I don't, so no. I wouldn't need parted. Somebody else might, though. |
| 18:07.27 | DocScrutinizer05 | discussing this topic we must not miss that it's actually two purposes folded into one: A) bring up the system with /usr not mounted yet. B) fix a system where /usr is borked. While A) is self evident, B) is arguable and could get achieved by other means (rescue system for example) as well |
| 18:07.59 | *** part/#devuan n4dir (~user@pD9EF0224.dip0.t-ipconnect.de) |
| 18:08.21 | DocScrutinizer05 | originally FHS implied that B > A and thus covers it completely |
| 18:08.55 | Lydia_K | I think having just enough OS to fix a destroyed /usr is kinda the idea |
| 18:09.08 | Lydia_K | otherwise why bother being able to boot without /usr at all? |
| 18:09.19 | DocScrutinizer05 | it's one of the ideas, booting the system at all is the other one |
| 18:09.21 | Lydia_K | If you just need a rescue disk anyway then it's sorta pointless. |
| 18:09.53 | Lydia_K | I feel like it needs a reason to be that way other than "because I can", but that's just like, my opinion man. |
| 18:10.12 | DocScrutinizer05 | there are valid other reasons for A) see maemo N900. However B) is not exactly mandatory for maemo |
| 18:10.30 | Lydia_K | Oh totally, there are exceptions to every rule |
| 18:10.40 | Lydia_K | especially when it comes to unusual hardware |
| 18:11.08 | DocScrutinizer05 | "unusual hardware" should be the usual case for linux |
| 18:11.24 | Lydia_K | I tend to think in terms of servers |
| 18:11.35 | DocScrutinizer05 | which is why linux flies, to start with |
| 18:12.14 | DocScrutinizer05 | now poettering tries to transform linux into a OS that only works on industry standard platforms |
| 18:12.46 | Lydia_K | The direction they are going seems more like it's trying to push it into being a consumer grade desktop |
| 18:13.03 | DocScrutinizer05 | exactly |
| 18:13.07 | Lydia_K | Their use cases are built around workflows that non-technical people use |
| 18:13.11 | spirou | don't like being a consumer |
| 18:13.12 | Lydia_K | close eveything when I log out? WTF? |
| 18:13.14 | DocScrutinizer05 | just another way to put it |
| 18:13.32 | Tear | i don't like consuming my own pc |
| 18:13.36 | Tear | it doesn't taste very good |
| 18:13.45 | spirou | try it with ketchup! |
| 18:14.08 | Lydia_K | everything is better with ketchup |
| 18:14.21 | *** join/#devuan wesbl (~wesbl___@95.239.200.116) |
| 18:16.41 | telst4r | libSoup? |
| 18:17.05 | DocScrutinizer05 | anyway the maemo example shows we possibly might want a means to install packages to different --prefix (/sbin vs /usr/sbin, etc), depending on desination hw platform |
| 18:17.17 | DocScrutinizer05 | and system usage scenario |
| 18:17.46 | DocScrutinizer05 | isn't there a thing like dpkg divert, or sth like that? |
| 18:18.43 | DocScrutinizer05 | dpkg-divert |
| 18:19.44 | furrywolf | reluctantly slides off Lydia_K's lap and heads out for work |
| 18:20.03 | Lydia_K | Awwww |
| 18:26.48 | *** join/#devuan dimkr (~dimkr@unaffiliated/dimkr) |
| 18:35.37 | *** join/#devuan kraiskil (~kraiskil@a91-152-241-34.elisa-laajakaista.fi) |
| 18:40.28 | DocScrutinizer05 | hmm, maybe PROVIDES could provide a few meta properties, like "provides recover capabilities for GPT" and some job hooked into apt or whatever package manager you use to install the system would then decide which packages to install to alternative location rootfs since they are needed there, based on comparing those meta properties to a set of system options you can select from |
| 18:40.54 | Lydia_K | That's a pretty interesting idea |
| 18:53.25 | golinux | wonders if Lydia_K has the day off or chatting on company time. |
| 18:53.38 | Lydia_K | Oh I'm working while I'm talking |
| 18:53.40 | Lydia_K | it's a slow day |
| 18:53.58 | golinux | A VERY slow day, I'd say! LOL! |
| 18:54.06 | Lydia_K | I've just been running some performance tests and deploying some new machines |
| 18:54.14 | Lydia_K | LOL |
| 18:54.24 | Lydia_K | This whole week has been super slow |
| 18:54.29 | Lydia_K | so many people took time off |
| 18:57.57 | *** join/#devuan blueness (~blueness@gentoo/developer/blueness) |
| 19:03.43 | *** join/#devuan SylvieLorxu (~TheLastPr@541B7AAC.cm-5-4b.dynamic.ziggo.nl) |
| 19:08.39 | *** join/#devuan cdanderson42 (~nuhrin@50.234.32.186) |
| 19:19.38 | DocScrutinizer05 | yet another maybe even simpler idea: a metapackage "recover_GPT+network" (s/GPT+network/$whatever-you-seem-fit/) that pulls in the needed packages like dunno gparted or whatever, and dpk-divert's them in postinst job |
| 19:21.42 | *** join/#devuan Drugo (~andrea@host236-89-dynamic.23-79-r.retail.telecomitalia.it) |
| 19:25.18 | Lydia_K | That seems like it would fit into the current infrastructure the easiest with the least amount of messing with anything else. |
| 19:25.51 | DocScrutinizer05 | yes, that's why I had no calm mind until it occurred to me :-) |
| 19:26.07 | *** join/#devuan n4dir (~user@pD9EF0224.dip0.t-ipconnect.de) |
| 19:28.11 | DocScrutinizer05 | actually I think it involves ziclh messing with anything else, except maybe moving stuff like parted to /usr/sbin per default |
| 19:29.02 | Lydia_K | It does sound like a metapackge would be a very clean way to impliment that sort of thing. |
| 19:30.02 | DocScrutinizer05 | yep, simply pick mp_server, mp_embedded_without_recovery, mp_embedded_recovery, or whatever fits your usecase, as additional package on top of base install |
| 19:30.25 | *** join/#devuan kelsoo (~chatzilla@88.144.88.129) |
| 19:31.31 | Lydia_K | I like that a lot :) |
| 19:33.17 | DocScrutinizer05 | doing the mps should be pretty easy |
| 19:34.02 | DocScrutinizer05 | and we could keep base install as small as it gets |
| 19:43.29 | DocScrutinizer05 | depending on your hw platform you might want the whole ubi* stuff either in rootfs, or on /usr, or possibly you don't want it at all |
| 19:43.35 | DocScrutinizer05 | >>ubiattach ubicrc32 ubicrc32.pl ubidetach ubiformat ubigen ubimirror ubimkvol ubinfo ubinize ubirename ubirmvol ubiupdatevol<< |
| 19:44.11 | *** join/#devuan blueness (~blueness@gentoo/developer/blueness) |
| 19:44.26 | DocScrutinizer05 | I'm pretty sure there's lots of similar stuff like gparted and ubi* |
| 19:44.38 | Lydia_K | I'm sure there is |
| 19:55.48 | *** join/#devuan oldlaptop (~quassel@50.36.241.195) |
| 20:11.31 | *** join/#devuan drogue (~rocket@cust-44148.ptera.net) |
| 20:54.12 | *** join/#devuan asbesto (~asbesto@88-149-239-187.v4.ngi.it) |
| 21:27.02 | *** join/#devuan Ch0liax (~Thunderbi@93-41-102-120.ip81.fastwebnet.it) |
| 21:27.19 | Ch0liax | Hi! ^_^ |
| 21:29.55 | golinux | Haven't seen you here before |
| 21:30.21 | golinux | But you may have been :) |
| 21:39.27 | Ch0liax | LOL |
| 21:57.44 | *** join/#devuan Infiltrator (~Infiltrat@unaffiliated/infiltrator) |
| 22:07.24 | DocScrutinizer05 | o.O |
| 22:08.01 | gci_admin | O.o |
| 22:52.28 | *** join/#devuan drogue (~rocket@cust-44148.ptera.net) |
| 22:56.13 | *** join/#devuan KiloJuliet (~gplgeek@pdpc/supporter/student/GPLGeek) |
| 23:03.44 | *** join/#devuan kelsoo (~chatzilla@88.144.85.207) |