IRC Logs for #crux-devel Thursday, 2012-01-12

*** mike_k has joined #crux-devel04:16
*** jue has joined #crux-devel06:07
*** rmull has joined #crux-devel06:32
jueanother joke from the sievers/poettering society -> http://article.gmane.org/gmane.linux.hotplug.devel/1723606:42
juebtw, kmod is using the same silly version-numbers like udev etc., only by chance?06:45
juebbl06:45
teK_they are double agents from Microsoft!06:53
teK_googling for udev yields no "homepage" for it07:02
teK_...07:02
frinnstlol?07:40
juere08:16
frinnstkmod308:16
juewell, it's more the development model behind than the version numbers. Too many releases were a wild mixture of bug fixes, new features and incompatible changes08:17
frinnsthttp://www.youtube.com/watch?v=YP9RnDp_tms08:17
jueI'd expext from such a important stuff at least a stable branch with bug-fixes only08:18
jueI'm talking about udev, btw ;)08:18
Romsterudev is starting to suck more every version with the current dev.08:18
Romsterwe can't go back to static nodes?08:19
juein theory yes, but IMO that's not a real option08:21
Romsteror we just don't bump udev until those nitwits make a stable branch and stop depending on new ports?08:22
Romsterthe very least they are causing alot of noise about of there incompetence.08:24
Romstersmart crow and also frinnst wrong channel :D08:25
juefor the moment we have to stick with udev 175, because we do not do the step from module-init-tools to kmod now, that's at least my opinion08:26
jaegerWhat a pain08:27
rmullWhat's a good alternative for udev?08:27
frinnsta 2.4 kernel? :)08:27
jaegermdev, perhaps08:28
jaegernot as configurable or scriptable08:28
frinnstyeah, busybox uses it, right?08:28
jaegeryes08:28
rmullI've never touched udev in my life. I don't need anything fancy.08:28
jaegerI've occasionally needed to change its rules a bit, though not often08:28
rmulljaeger: Thanks for the mdev suggestion, I'll check it out08:29
jaegernp08:29
Romsterhow did this guy get into udev portaudio and stuff? and into the kernel.08:30
Romsteror is it just me hating his progress.08:30
juesomething behind kmod -> http://tinyurl.com/6mblc4c08:30
rmullA lot of people are bothered by his "progress"08:30
rmullI'm sure some of it isn't all that bad08:31
rmullBut it always causes me to switch away for some reason...08:31
jaegerI wonder sometimes if it's not just a very elaborate troll, heh08:31
jaegerPerhaps he's trying to encourage someone else to make a better alternative08:31
jueyeah, sometimes it looks like kids playing with their toys to me08:32
Romsteri've came to the same conclusion08:32
Romsteroh and i forgot systemd08:33
Romsterhate to think of whats next08:33
rmullthe new kernel logging system08:33
jaegerHe reminds me of an old boss... always coming up with solutions for problems we don't have08:34
Romsterfirmware, kernel logging or something08:34
rmullHey, is /proc deprecated?08:34
jueyeah, it's already part of systemd 38 -> http://lwn.net/Articles/474968/08:34
Romsterlol /proc that'll be next lets all mvoe to /sys08:34
rmullRomster: I thought that's what the idea was08:34
rmullI was wondering if I can get rid of proc instead of keeping them both08:35
Romsterhmm i think so but i think lots of stuff still depend on /proc i know sysctl was made to depreciate the ol' echo foo > /proc/... method.08:36
juesorry, who says that /proc is deprecated?08:36
jaegernobody, yet08:37
Romsterthat's what i thought.08:37
rmullOkay, just checking08:37
Romsterbecuse there is /sys that everone is moving to and sysctl to adjust settings. i'd imagine one day /proc be depreciated or at least pruned down on some things.08:38
Romsterno proof just what i think may happen.08:39
Romsteranyways i' off topic.08:39
Romsteri'm*08:39
juebtw, did I asked for opinions/results/problems wrt rc-ng today? ;)09:03
jaegerheh09:04
jaegerTo be honest I haven't even looked at it yet, had too many other projects going09:04
juejaeger: would be great if you could find the time to have a look at it09:06
jaegerAny link with info about it? Why we want it, etc.09:07
Romsterno problems running it and i've been rebooting a few times.09:07
juenot much, only the thread in crux-devel09:08
pitilloRomster: with lvm setup, right?09:08
Romsterhaven't had a fsck scheduled yet though.09:08
Romsteryep mdadm lvm09:08
pitilloRomster: nice, those are the good reports IMHO09:08
Romsteronly raid1 in mdadm but it assembles it then activates the VG09:09
juejaeger: it looks very promising to me, and there's a commitment from Michal to support it in the future09:12
Romsterwhen i look at rc.conf though i go man where did my 5 lines go.09:14
Romsterhaven't poked much in the other files.09:14
Romsteri'll look at it after i sleep before i go out.09:15
juemost of the lines in rc.conf are comments, though09:15
Romsteryeah09:15
jaegerjust glanced through the thread again, does the benefit of rc-ng outweigh the large jump in complication?09:16
Romsteri haven't used LUKS yet but i wonder how that would in the mix of mdadm lvm luks09:17
pitillojaeger: I think it should for raid/lvm users, this is why Michael seemed to start with this new ng version (better support for them) May be not for normal users09:17
jaegerI'm not against it, just wondering about the gain09:18
Romsterumount 3 stage for layered devices though i dunno if i like that, should umount in the correct reverse order as it started with.09:19
juejaeger: I'd say yes, TBH we never developed the rc stuff further but added stuff as needed and that's how it look nowadays09:19
Romstertimeout on services that sit for ages for some unknown reason.09:19
Romsterother than that i haven't dug into it more.09:19
jaegerI'll try to install it in a VM today if I have time... working on getting my buildbot up and running while I wait for our 10Gbe support people to show up and work on the NIC problem we have09:20
juewe miss things like prober UTF-8 in console, how to deal with RTC etc.09:20
jaegerWhat does UTF-8 in the console add?09:22
jaeger(I'm really not trying to argue or anything, really don't know the answer)09:22
jueheh, not many :)09:23
jaegerI'm completely out of the loop on things like UTF-8 and what it really does for you09:24
jaegerI've been totally focused on multilib and to a lesser extent gnome 3 in my spare time lately, which is why I've not looked at the rc-ng stuff yet09:25
Romsterdon't forget rejmerge09:25
jaegerThis may sound odd but I've never used rejmerge before and I've been using crux for years :)09:26
Romstersarcasm alert :D09:26
jaegerNo idea how that could be construed as sarcasm but I was actually serious09:27
juethis time it's really important, because you should not reject /etc/inittab and rc.{fix,local,modules}09:29
Romsterwell be sure to run it as it wont boot otherwise.09:29
juethe later are links into rc.hooks now09:29
jaegerYou misunderstand, I merge my rejected files by hand09:31
jaegerI don't ignore them completely09:31
Romstersounds painful or is it i never tired doing it by hand without that command.09:32
jaegerIt might be on your systems since you install way more stuff than I do, but it's not bad for my use09:32
jaegerThe udev/gudev source URLs are broken now, seem Kay removed some things09:44
jaegerseems09:44
juejaeger: thanks, fixed now09:51
jaegerthanks :)09:54
jaegerIs there a problem with the current cups and tcp_wrappers? I keep seeing this in the cups logs:10:19
jaegerW [12/Jan/2012:10:13:21 -0600] Connection from localhost refused by /etc/hosts.allow and /etc/hosts.deny rules.10:19
jaegerEven if I allow ALL from 127. in /etc/hosts.allow10:20
juejaeger: I have 'cupsd: 127.0.0. localhost'10:24
jaegerthat doesn't work for me either10:27
jaegerhrmm10:27
jueIIRC there was a change in that region ...10:27
jaegerand now it does, how bizarre10:27
jaegerI restarted cups, it failed, tried again after about 10 seconds, it worked10:28
teK_jue: have you seen the built-in grep for rc-ng?10:28
jaegernever mind, sorry for the noise10:28
jueteK_: not yet10:30
jueteK_: where is itz hidden?10:30
teK_lib10:35
teK_shegrep()10:35
teK_it sums up my experience ;)10:36
teK_it worked ootb10:36
teK_I had to call ntpdate by hand after rebooting after setting timezone (bogusly) to Europe/berlin10:37
teK_but that may be an issue with ntpdate itself10:37
teK_everything else worked like expected10:37
teK_BUT.10:37
teK_As frinnst (?) mentioned on the ML: it's kinda .. big10:38
teK_while test-driving it I thought to myself: why don't we extract the really really neat features of it an integrate them to our existing rc10:38
teK_oh and by the way.. rc-ng fixed (by handling those virtual terminal locales correctly) the bug of mis-displaying umlauts in Terminal if and only if I logged in with slim (vs. manual tty-login && startx)10:39
jueteK_: would be really nice if you could ask that in a mail to crux-devel, so Michal has a chance to read/reply ;)10:40
teK_another thought was: the functions do utilize quite some advanced (at least syntax-wise) features of (ba)sh programming which may make it harder for a possible maintainer10:40
teK_yeah sure10:40
juewell, _IF_ we add that stuff, Michal is the maintainer of it10:41
teK_what if he leaves the project.10:41
teK_this point leads to my last criticism (so far :p): it's (again..) in bash. zsh won't even parse the functions correctly :\10:42
teK_this == the advanced _ba_sh usage10:42
teK_I know that we discussed this a while ago10:42
jaegerIs that really a problem? Do you remove the bash package after install?10:42
teK_I still think it would be cool to not be forced to used bash for certain system-stuff10:42
juesure, it's a bash script, not sh10:42
teK_jaeger: I can't ;)10:42
teK_I had a good laugh wenn I saw the internal grep function btw :> maybe we should opt for putting grep(1) to /bin as Gentoo and others do *g*10:44
jueteK_: IMO the bash/shell discussion is useless at all10:44
teK_yeah :) contrary to my opinion10:44
juebash is a 650k binary, not much nowayds even if you use it only for system start10:45
juewrt grep in /bin, that's in our TODO28 page10:46
teK_so to sum it up: coolest features were: parallel service startup without changing anything in /etc/rc.d/* and the hook-stuff but we should take the road you (?) proposed on the ML10:46
teK_ok10:46
juesorry, which road?10:47
teK_I don't quite remember the difference between the implementation and the proposition, wait a second10:47
teK_I'm referring to [1] of http://lists.crux.nu/pipermail/crux-devel/2011-October/004000.html10:53
teK_though I don't know what's the real difference to michal's solution as I did not test hooks10:54
juenow I remember ;)10:54
teK_jue: what's the md5sum of your named.root download file?10:56
teK_% md5sum /usr/ports/distfiles/named.root; grep named.root .md5sum10:57
teK_d7a1cb305b7b5b72df4e574777f76723  /usr/ports/distfiles/named.root10:57
teK_d7a1cb305b7b5b72df4e574777f76723  named.root10:57
jueteK_: testing now11:00
jueyeah, it's d7a1cb305b7b5b72df4e574777f7672311:00
teK_that's the checksum of the portsversion, too11:02
jueteK_: the only thing I've done is to assign the ticket to you, because it's your port :)11:03
juemight be that you've corrected the port already, but forgotten to close the ticket11:04
juefrom time to time I'm looking over the tickets to see if we can do something11:06
teK_closing it, thanks11:11
jueback to rc-ng, I got the impression that most of you are more against it, which is perfect fine for me, but I really appreciate to see more comments/opinions in crux-devel11:11
teK_I will try to compile a more elaborate report of my impressions11:12
teK_but not before sunday11:12
teK_brb11:12
jaegerI'm not against it at all, myself, just haven't had a chance to look at it yet11:12
jueteK_: cool, thanks11:13
juejaeger: yeah, except you. I should better have said tek/frinnst/sepen instead of 'you'11:14
jueon the more positive side we have me/Romster/pitillo IIRC11:16
jaegerheh, fair enough11:17
frinnstjue: well, i have yet to actually run it. part of that i guess is that its not really something that i find interesting. i *like* our minimal initscripts11:22
frinnstbut sure, they are sometimes a bit inflexible out of the box11:22
frinnstbut the thing is: its easy to fix yourself11:22
frinnstand if that fix is a more complex solution that you call "rc-ng" - hey, you're free to maintain yourself :)11:23
frinnsti guess that's how i feel :)11:27
juefrinnst: i can fully understand your feeling, I'm in a similar position, it's also nothing I really care about11:33
juebut I really like to have someone who cares about it ;)11:34
juewith 'it' I mean the whole rc/udev/mdadm/lvm stuff11:35
jueah, forgot to mention the new kmod construction site11:38
juebtw, if we are here, there's a kmod port in my private repo, if someone is eager to play with it ;)11:41
jueit can be installed togther with module-init-tools, no symlinks yet11:42
jueanother reading wrt kmod -> http://www.jonmasters.org/blog/2011/12/20/libkmod-replaces-module-init-tools/11:46
jueguess, I should shut up now11:47
jaegerkmod does look sorta interesting but it definitely has a bit of a kitchen sink feel12:35
juelol12:38
jaegerThat perception is NOT helped by the "plumber's wishlist" :D12:39
jaegerinteresting, automake breaks opt/grub now12:47
jaegerhttp://lists.gnu.org/archive/html/grub-devel/2011-12/msg00059.html12:52
teK_m(12:56
teK_dammit the clock is wrong, again12:58
teK_Thu Jan 12 20:58:53 CET 201212:58
teK_reboot.12:59
teK_nope13:02
teK_two reboots with HWCLKMODE=localtime and correct TIMEZONE yield this:13:03
teK_Thu Jan 12 20:03:59 CET 201213:03
teK_ah13:03
teK_correct13:03
teK_oops :P13:03
jaegerheh13:12
jaegerstrange13:12
jaegerjue: does your rc port at jue.li include the kbd fixup?13:16
jueyeah, it's the latest version13:18
jaegerok, going to test it in a sandbox VM now13:18
teK_kernel.org down?13:20
rmullteK_: Apparently...13:21
rmullblah13:21
jaegerI take it SYSNAME= is the new HOSTNAME=13:26
jaegerI guess not, I see both13:27
jaegernoisy boot and it kills my net script but so far it seems fine13:31
teK_jaeger: same thoughts on my side13:32
jaegerfsck -t <lots of stuff> :P13:33
jaegerNo problems with it so far, anything in particular I should poke?13:34
jaegerno issues with time or hwclock13:35
*** j^2 has quit IRC14:11
*** j^2 has joined #crux-devel14:16
*** acrux|G4 has joined #crux-devel14:51
*** acrux|G4 has joined #crux-devel14:51
*** acrux|G4 has quit IRC14:51
*** mike_k has quit IRC15:19
*** prologic has quit IRC19:03
*** prologic has joined #crux-devel19:11
*** prologic is now known as Guest4760919:12
*** mavrick61 has quit IRC21:32
*** mavrick61 has joined #crux-devel21:33

Generated by irclog2html.py 2.11.0 by Marius Gedminas - find it at mg.pov.lt!