IRC Logs for #crux Wednesday, 2011-05-11

*** Rotwang has quit IRC00:31
*** Rotwang has joined #crux00:32
pitillogood morning01:03
*** mike_k_ has joined #crux01:30
*** lasso|qt has joined #crux01:34
*** happydawn has joined #crux03:45
*** cruxbot has joined #crux03:52
*** happydawn has quit IRC05:07
*** Evil_Bob has joined #crux06:20
*** childintime has joined #crux06:25
*** sepen has joined #crux06:39
cruxbot[contrib.git/2.7]: p5-extutils-pkgconfig: fixed release version to solve problems with p5-cairo build06:39
sepenteK_: thanks for the report :D06:40
jaegerRomster: to what does your make clean comment refer?07:54
juejaeger: FYI, the reason for the libtirpc linking issue is that commit ->
juewrote to the author but no response yet08:09
jaegerhrmm, interesting08:09
*** rauz_ has quit IRC08:15
cruxbot[opt.git/2.7]: apache: update to 2.2.1808:17
cruxbot[core.git/2.7]: file: update to 5.0708:18
*** happydawn has joined #crux08:23
*** ThePub has quit IRC08:37
*** ThePub has joined #crux08:37
*** happydawn has quit IRC08:54
Romster<jaeger> well, shit. apparently I've lost my updates to the net setup iso <jaeger> it might be time to remove that crap from the Makefile <- to that.09:02
jaegerAh... by "that crap" I was referring to the netinst iso stuff09:04
Romsterthe makefile did it nuke your files or something?09:07
Romsteralso rebuilding midori but i suspect that those files come from another dependency i don't have listed in midori.09:08
Romsteri build everything in a chroot for contrib so it's got a clean no missing files footprint.09:08
jaegerno, it didn't do anything wrong, the netinst stuff is just a pain to maintain and hasn't been updated in ages anyway09:08
jaegerI probably lost the changes with an erroneus git reset or something09:09
Romsteryeah gotta watch that, didn't you commit the changes you had done recently.09:10
Romsterwasn't the saying commit often.09:10
Romstereven if you do it to another branch or git stash it before reverting in case you want them changes later.09:11
jaegeryeah, ought to do more of that09:11
Romsteri try to or else i lose stuff too09:12
Romsterwhich port provides those extra icons?09:29
Romsterfootpint on midori is clean in chroot09:29
Romsterbut gst-plugins-base is not clean
Romstersepen, ^09:31
Romsterah rsvg-convert most likely.09:40
sepenhmm you're right I missed to build them in the safe09:43
Romstersepen, also libnotify is missing a dependency dbus-glib09:48
ThePubis there an equivalent to "provides" like Arch has?  maybe I'm just missing that.09:49
sepenThePub: from command line queries or you mean a Pkgfile section?09:50
Evil_Bobbtw does crux have a linux-firmware package if not why not? :) (i composed one earlier because i needed it for my wireless card)09:51
frinnstthere's radeon firmware in opt09:51
frinnstusually if you query the portdb you'll find that someone has packaged the firmware youre looking for09:51
Evil_Bobcouldnt find it in portdb :)09:52
Romstersepen, one of these days i'll get some sort of a build bot thing happening to test all ports in a chroot.09:52
Romsterfor now i just report what i find on my travels of the ports i maintain.09:53
cruxbot[contrib.git/2.7]: midori: add more packages to depends on line09:53
Romsterjaeger, midori footprint fixed up.09:53
RomsterThePub> is there an equivalent to "provides" like Arch has?  maybe I'm just missing that. <- no there is a nice to have or a optional field but that's not official, if oyu want that stuff your best to go with gentoo.09:55
ThePubsepen: arch has a "provides" line where any package can pretend to be anything else, upon installation of one package it then knows to remove any collision.09:55
Romsteror read the README that comes with the port if it has one.09:55
jaegerRomster: I don't use midori, was that meant for someone else?09:55
Romsteroh sorry that was sepen that said that -_-09:56
*** Rotwang has quit IRC09:56
*** Rotwang has joined #crux09:56
jaegerAh, ok, was confused :)09:56
Romsterthen you asked me about the clean comment right after.09:56
Romsterand how ironic in fixing up midori i find two other defects of sepens :D09:57
Romstersepen, at some point the whole gsreamer and plug-ins need the dependencies carefully checked.09:58
Romsterfor now i'll bump the versions in contrib.09:58
Romsteras things get passed from one plug-in package to another.10:00
Romsteralso 1am here now.10:03
frinnstThePub: but isnt that more for a binary distribution? i could imagine a package that is precompiled with different deps. then it would be useful10:05
Romsteroh provides... that's not what i was thinking of..10:05
Romsteris that a kind of thing when you say install gtk+ it provides glib and friends?10:06
Romsterlike the prt-get quickdep foo10:07
Romstermight make more sense for meta packages?10:08
*** nogagplz_ has joined #crux10:33
*** nogagplz has quit IRC10:33
*** nogagplz_ is now known as nogagplz10:33
Romstersepen, libglade needs xorg-libxdamage10:40
sepenoops, too many reports, I'll try to fix them when arrived back to home, thanks10:42
sepenalso I've some logs here if you want to take a look
cruxbot[contrib.git/2.7]: gst-plugins-good: 0.10.26 -> 0.10.2910:43
cruxbot[contrib.git/2.7]: gst-plugins-bad: 0.10.20 -> 0.10.2210:43
cruxbot[contrib.git/2.7]: gst-plugins-ugly: 0.10.16 -> 0.10.1810:43
cruxbot[contrib.git/2.7]: gst-python: 0.10.20 -> 0.10.2110:43
Romsterok i'm done for tonight going to bed, g'night.10:45
Romsterhmm k i'll look over that page later.10:46
teK_sepen: you're welcome10:50
*** ThePub has quit IRC11:06
*** ThePub has joined #crux11:07
*** jse has joined #crux11:29
*** nogagplz_ has joined #crux11:40
*** nogagplz has quit IRC11:43
*** horrorStruck has joined #crux12:47
horrorStruckhi #crux12:47
horrorStruckjue: just ooc, when you move a port from your repo to crux repo (gcc, udev, etc..), will you make a release bump if there's any change or should we check just in case? tia12:54
*** mike_k_ has quit IRC13:02
*** mike_k has joined #crux13:03
*** lasso has joined #crux13:06
*** pitillo has quit IRC13:52
*** pitillo has joined #crux13:54
*** happydawn has joined #crux14:01
*** jse has quit IRC14:19
ThePubfrinnst, Romster: "provides" is a way of "providing" a "virtual" package.  so as an example xterm, xfce-terminal, and gnome-terminal could all provide an "xorg-terminal".  Installing one of the three when another of the three is installed would mean a package collision, hence you would be asked to remove one and then install it or leave it as is.15:33
ThePubfrinnst, Romster: it of course is more useful for libraries which provide the same interfaces but are different libraries themselves.  Like C libraries.15:34
*** childintime has quit IRC15:43
*** lasso has quit IRC15:48
*** happydawn has quit IRC15:53
*** happydawn has joined #crux15:54
*** childintime has joined #crux15:54
*** happydawn has quit IRC16:03
*** uki has joined #crux16:36
ukihi, Im getting a 'filesystem check failed' error during bootup. In mainterance mode I fsck -v /dev/xxx all my filesystems - no errors, what else can I check? (I did an update before reboot - but dont remembere if anything important was being updated)16:39
ukijust before that error I see 'fsck.ext4: No such file or directory while trying to open /dev/sda1' - but the device exists, and manually ran fsck does not complain16:41
jaegerI've been seeing that in a VM recently, no idea what causes it except perhaps udev not populating /dev quickly enough16:48
ukiouch :(16:49
jaegerrebooting a few times will eventually see it work16:49
jaegerat least it did in my case16:50
ukiok, thanks, I shall try it16:51
DaViruza couple of `sleep 60` in the init scripts are never a bad thing16:52
jaegerghetto as hell, though16:57
* uki is editing /etc/rc :D16:58
prologicguys I hadn't actually followed this discussion all that well16:59
Evil_Bobjaeger: i've also seen it on my native machine, i think its a bug in udev16:59
prologicbut if it's what I think it is16:59
prologicwe really need to migrate CRUX to an alternative init16:59
prologicwe are really quite old in that regard16:59
prologicwe're still using a synchronous/serial init process17:00
ukion the plus side, that ancient stuff is so simple I can try to debug it with a few echo calls17:01
Evil_Bobprologic: just curious which alternative init do you want to replace it with ?17:02
prologici don't really think it's up to me alone :)17:02
prologicthe CRUX devs and community should decide this17:02
prologicupstart is nice enough though17:02
prologicand backwards compatible with init scripts17:02
prologicthere are several others (that are nice and backwards compatible)17:02
prologicI just feel that we can't keep using something that's inherently ancient and synchronous - it doesn't lend itself very well to good startup times or good management of services ihmo17:03
prologic(especially on CRUX server boxes)17:03
prologicor maybe desktops too :)17:03
Evil_Bobboots in 3 seconds in my vm :P17:03
prologicthat's because you have a small kernel and no startup services :)17:05
Evil_Bobtrue dat :P17:05
prologicor you've hacked the hell out of /etc/rc :)17:05
* prologic chi ching :)17:05
Evil_Bobonly 2 services, kernel is not supersmall tbh, its like 4mb17:05
prologicanyway I have some time this week and maybe next week17:05
prologicI might look at using upstart in a vm and build a port for it17:05
prologic4MB ?17:06
prologicgeez man :)17:06
prologicdeselect some unnecessary options :)17:06
prologicmine's half that17:06
Evil_Bobwith or without modules ?17:07
prologicwithout modules17:07
prologicI compile all into the kernel that's on my hardware17:07
prologicmy rc.modules is empty17:07
Evil_Bobi have an additional 2.8mb of modules lol, but i share this config with several machines :o17:08
Evil_Bobi used to tweak my kernel, but cba atm  :P17:08
Evil_Bobwhy am i defending this heh17:09
prologicI don't know :)17:09
prologicif you feel the need - go ahead :)17:09
prologicI was talking about an init alternative17:09
prologicI think (therefore I am) (tm) :) -- we should come to a consensus about a new init for CRUX17:10
Evil_Bobim happy with the current one, but if theres an alternative simple one i wouldnt mind i think17:12
*** mike_k has quit IRC17:13
prologicI think we can compromise on simplicity for a little more performance17:13
prologicupstart is simple though ihmo and backwards compatible with init-style scripts17:13
*** nogagplz_ is now known as nogagplz17:14
ukiI like systemd a bit better17:14
Evil_Bobuki: why ?17:15
ukimakes more sense, its super simple17:16
prologicyes why ?17:16
prologicwell this is the sort of stuff we should discuss17:16
prologicpros and cons of other init systems17:16
prologicand we'll pick one that bests fits CRUX17:16
prologicI've seen systemd - it's not too bad17:16
ukion a side note, strategically placed sleep 30 fixed my init script :P17:17
prologicyes but that's horrible17:17
prologicwhich is why I raised this point17:17
prologicsoon as I say your "sleep 60" comment17:18
prologicthat's just soooo wrong :)17:18
Evil_Bobi thought it was a joke... :o17:19
jaegerAnyone have an intel x25-m 120GB SSD?17:20
jaeger - curious if my speeds are reasonable compared to others17:23
ukiok, night project - replace rc scripts on my box with systemd :D cya17:23
*** uki has quit IRC17:23
Evil_Bobnice jaeger :)17:25
*** qid[MOE] has joined #crux17:25
Evil_Bobi want a SSD, but ill wait till they're even cheaper17:25
jaegerI've got an intel x25-m in the windows box and an OCZ Vertex 2 in the CRUX box17:25
jaegerpretty happy with both17:25
prologicso you guys want me to try to implement systemd ?17:28
jaegermaybe you implement upstart and uki does systemd, then compare the two17:29
Evil_Bobi dont like systemd from what i've seen17:30
*** Evil_Bob has quit IRC17:31
prologicheh that's a good idea jaeger  :)17:31
jaegerbest to have options :)17:32
qid[MOE]would sysv init still be an available option? I get the impression it's probably lower-overhead than upstart or systemd, and in some cases that's valuable too17:33
prologicactually from what I've read other inits don't add that much overhead17:35
prologicin fact they vastly improve startup performance overall by asynchronously starting processes / services17:35
prologicbut jaeger is right17:35
prologicwe should try a few different ones out and compare :)17:35
qid[MOE]well, startup performance is a more desktop-oriented optimization17:36
qid[MOE]in some servers, you're lucky if you get to the boot loader inside of two minutes with all the various bioses and such, heh17:38
prologicyes I'm painfully aware of that :)17:38
prologicDell servers for example17:38
prologicwho needs all that crap anyway seriously17:38
qid[MOE]hmm, I can't remember, is there anything special that has to be done when you do a gcc minor version upgrade? I know glibc requires a kernel rebuild and all sorts of fun17:52
Romsterprologic, i've been using runit for the past year or more forget how long exactly. like 20 second boot times with lots of services running.17:57
* prologic doesn't like runit :)17:58
prologicif something takes more than 15mins to work out it's kinda useless :) heh17:58
Romsteryeah the website sucks for docs and stuff.17:59
Romsterbut the init itself is stable.17:59
*** julianj has quit IRC18:02
Romsterwell i could try other init's but i wont touch the one that the save dev that makes pulse audio does.18:02
*** acrux has quit IRC18:02
Romsteranyways later.18:02
*** acrux has joined #crux18:03
*** ente has quit IRC18:06
*** acrux has quit IRC18:52
*** acrux has joined #crux18:54
ThePubjaeger, uki: after starting udev do you run "udevadm settle"?  that stopped most of the device race issues for me.19:11
*** ThePub has quit IRC19:15
*** ThePub has joined #crux19:31
ThePubxsession dies after "fixing" it  ^_^19:33
*** sepen has quit IRC20:16
*** lasso|qt has quit IRC20:18
*** childintime has quit IRC20:47
*** prologic has left #crux21:24
*** laen has quit IRC21:33
*** Dudde has quit IRC21:35
*** Dudde has joined #crux21:37
*** laen has joined #crux21:43
*** DaViruz has quit IRC22:17
*** dinx has quit IRC23:58

Generated by 2.11.0 by Marius Gedminas - find it at!