IRC Logs for #crux-devel Wednesday, 2012-10-24

*** mike_k_ has joined #crux-devel00:59
*** horrorStruck has quit IRC01:28
juegood morning02:53
juefrinnst: do you have nfs in your SERVICES array?02:53
jueRomster: looks like we lost the symlinks to sr0 sometimes in the freakish udev history ;)02:56
juejaeger: do you have an additional, local rule for the sr0 device?02:57
Romsterjue seems that way.03:00
frinnstjue: yeah03:03
juefrinnst: hmm, so the 'umount -a -t nfs' in /etc/rc.d/nfs doesn't work somehow?03:06
frinnstI guess. but i've yet to dive into it head first :)03:08
frinnstiirc, it gets stuck on /bin/umount -a -r -t nosysfs,noproc03:08
frinnstI have one nfs3 and one nfs4 mount. i think its the nfs4 mount that fails03:09
frinnstpossibly because the filesystem is "nfs4" so it doesnt get unmounted with that command03:10
juehmm, just tested, works for nfs vers=4 too03:13
frinnstdoes mount list the fs as "nfs4" ?03:14
jueoh, I got it, it works with 2.7 but not with 2.803:16
juewith 2.8 I need a umount -a -t nfs403:16
frinnsti think im seeing two different problems. the "nfs4" issue and a rdiff-backup/"nfs unmount" problem03:16
frinnstwe try to unmount nfs via rc.d/ before user processes are killed - that will fail since the nfs-fs is in use03:17
juewell, the start script of nfs-utils is easy to fix and with that it should work for most cases?03:17
frinnstthen the network goes down and all processes are killed leaving nfs filesystems03:18
frinnstbut im at work and probably talking out of my ass :D03:18
frinnsti'll check once i get home03:18
juesure, if a process is using the nfs filesystem the umount will fail03:19
jueas jaeger said, adding a -l or -f, our even both, to our last umount command in rc.shutdown is something we should consider as well03:22
pitillohey jue, did you see sepen's comments about tzdata?03:22
juehi pitillo, yeah, saw it, adding CFLAGS seems like a good idea03:24
juewill do it later03:25
pitillonice jue, thank you. Don't worry, there is no hurry.03:25
juebtw, new prt-get version is on the way, cptn has found the pipe leak :)03:26
pitillojaeger: did you see the comment about HID kernel support for newer releases (I asked frinnst to do that in X86_64)? it could be great for usb keyboards (may be there are some strange people as me using one of them :)03:27
frinnsthey, I do too! :)03:38
*** horrorStruck has joined #crux-devel03:38
frinnststandard keyboard should work without it. but stuff with logitechs wireless thingy might be problematic03:38
frinnstpitillo: what kind of keyboard are you using?03:39
pitillofrinnst: nothing strange, a cheap logitech kbd+mouse combo03:45
frinnstwireless, no?03:52
frinnstunifying reciever thingy03:52
pitillofrinnst: yes03:52
pitilloone of the problematic devices then...03:53
frinnstI have one of those for my mouse. probably why i never noticed it since the keyboard worked during install :)03:53
*** sepen has joined #crux-devel03:54
pitilloummmm I have tested it on my workstation and the rpi, both with problems. If this is a special case and it should work with standard devices... avoid my comment, sorry :)03:55
frinnstwell the unifying thing is quite common03:56
frinnstand it *used* to work without a special driver on older kernels03:56
frinnstdont remember in what kernel they changed it.. 3.2, 3.3?03:57
pitilloI've tested it on 3.5.4 (x86_64) and 3.1.9 (rpi) I haven't checked it on older kernels03:59
frinnstyeah. requires hidraw and the logitech driver these days04:07
pitilloyes, seems so. I thought it could be interesting, but now I see this is more a personal and non confortable problem xD04:08
*** horrorStruck has quit IRC04:43
*** horrorStruck has joined #crux-devel04:45
*** horrorStruck has quit IRC05:57
*** horrorStruck has joined #crux-devel05:57
*** horrorStruck has quit IRC06:11
*** horrorStruck has joined #crux-devel06:23
frinnstjue: the nfs update helped a bunch06:59
frinnstand yeah, lazy / forced unmount as a last resort might do good as well06:59
frinnstthe system is coming down anyways.. :)06:59
horrorStruckjue: what do you think about adding nodevtmpfs (and maybe notmpfs) to rc.shutdown umount? i always get devtmpfs busy, remounting read-only when shutting down my laptop.07:02
jaegerRomster: line 104 - DEVLINKS=/dev/cdrom /dev/disk/by-id/ata-PIONEER_DVD-RW_DVR-21207:13
jaegerlooks like it should be creating /dev/cdrom07:13
Romsterjaeger, your correct i do have a /dev/cdrom but no /dev/dvd that some programs look for.07:18
horrorStruckjue: this one i'm not too sure but do we really need the two umount?07:22
jaegerRomster: try the sed here:
jaegerpitillo: yes, I saw it08:02
jaegerIt would be trivial to add hid-logitech-dj.ko to the initramfs, it's already built as a module in the ISO kernel and hid/usbhid are already included08:20
Romsterso basically changing PROGRAM="write_cd_rules" too PROGRAM="write_cd_rules by-id"08:21
Romsterchnaged i need to reboot later for newer kernel patch so i'll know after that if that fixes /dev/dvd and friends.08:26
jaegeryou could also try "udevadm control --reload-rules"08:27
Romsterdidn't add anything to /dev/08:39
Romsternot over fussed with it just bugs the hell out of me that udev is so broken.08:40
* frinnst feels we should have stopped updating udev a few versions ago08:44
frinnstbefore the devtmpfs crap and all the breakage08:44
frinnst175 or whatever it was08:45
Romsterprobably should have, everything worked fine then.08:45
Romsterbut doens't everything require a /run directory now?08:45
frinnstonly udev?08:46
Romsteri didn't know what was wrong with /var/run to be honest.08:46
frinnstme neither. only the systemd folks knew08:46
jaeger182 works great for the most part08:47
jaegerthe symlinks are the only issue with it that I can think of08:47
horrorStruckah yes we do need the second umount, i got some option wrong.09:24
*** mike_k has joined #crux-devel09:24
*** mike_k_ has quit IRC09:25
horrorStruckah double wrong makes a right. sh*t i'll stop spamming now but still not sure about the second one09:27
*** mike_k_ has joined #crux-devel10:36
*** mike_k has quit IRC10:37
juehorrorStruck: don't know why have the second mount, it was introduced together with the lvm stuff10:48
jueRomster: wasn't that patch from you?10:49
juehorrorStruck: never heard from / saw your remount-devtmpfs problem, would be nice to have the real reason for the issue10:54
juefrinnst: debian is still at udev 175, with a bunch of patches oc10:56
jueif it makes things better I'd not have objections to use that version too11:01
juehonestly, I really dislike the whole udev stuff and would be really happy if someone else cares about it11:02
juebut well, I think we have more important tasks currently11:03
frinnstheh, yeah. we're just bitching :)11:04
juefor example the release of CRUX 2.8?11:04
frinnstyeah. anything else need doing?11:05
* frinnst would love to start building a x86_64 iso! :)11:06
frinnstmoving grep? anything specific that needs it?11:07
frinnstlooking over the todo i'd say the minor thing that remains can wait for 3.0 or whatever11:08
frinnstwould be real nice to get 2.8 out the door11:09
jueyep, indeed11:09
jaegerI don't have any objections to keeping udev 182 around, we can add that 'by-id' change to the cd generator rules11:14
juejaeger: ok, fine, you are going to add it?11:17
juejaeger: if we update the docs we can relase rc3 as final, right?11:18
jaegersure, I can do that11:18
jaegerdocs and the kernel config, if we want to add the logitech unifying receiver11:19
jaegerer... well, it's already in the kernel config but would need to be added to the initramfs, which is simple11:20
jueI see no reason to wait longer with the release, quite the opposite11:22
jaegerwait longer for what?11:24
juegood question :)11:24
jueRomster: btw, the old udev runtime directory was /dev/.udev/11:28
jueso we agree that we release rc3 with updated docs as final?11:29
jaegerI think so11:29
juegreat :)11:29
jaegerI will of course test the new ISO quickly before posting but it should be fine, same content11:29
juejaeger: could you please look over the release notes and correct my denglish if necessary?11:32
jaegermade only very minor changes, such as "update" to "updated", "are" to "is", etc.11:38
jaegerI wonder if the logitech unifying driver has to be loaded explicitly or if it will load automatically11:38
jaegerwill test that11:38
jaegerGoing to eat some lunch, hopefully will have the ISO ready not long after that12:07
frinnstcould you push the changes? i'd like to keep the x86_64 iso as similar as possible12:08
*** mike_k_ has quit IRC12:21
jaegerfrinnst: of course :)12:47
jaegerDo we need to add a last-minute patch for the ext4 issue or is that not a problem?12:53
juejaeger: don't think so, ... rebooted twice in fairly quick succession ... isn't a very common practice and most people will use a newer kernel anyway13:15
juebut anyway, what's your opinion?13:23
*** sepen has quit IRC13:24
jaegerI'm not too worried about it13:25
jaegerok, did a fresh install and sysup in virtualbox, seems happy. also tested the logitech unifying driver, worked for my K75013:45
jaegerUploaded the ISO to my homedir on if you guys want to test it before making it public14:05
juewell, let's announce it ;)14:08
jaegerYou want to write the announce this time?14:08
jueTBH, would prefer if you do it14:15
* frinnst starts a x86_64 build14:17
frinnsttag core/opt/xorg maybe?14:26
juejaeger: ok for you?14:26
juefrinnst: yeah, will do that now14:26
jue2.8 is now the default branch in out gitweb14:39
jaegerbuilding the source ISO and setting up the torrent before any announcement14:45
frinnsthah, i always forget to edit iso/pkgmk.conf and always panics because gcc is unable to create binaries :D14:46
jaegeroops :P14:47
jueChangeLog updated14:54
jaegersomething besides the release date?14:54
jueyeah, correct kernel version14:55
*** pedja has quit IRC14:56
jaegerI'm confused, already had 3.5.4 in it14:56
jaegerJust wondering if I need to remake the torrent14:57
jaeger7666e1f0684d055261c492747c95465e  ChangeLog <-- this is the one that's in it14:57
juesorry, couldn't find that revision?15:02
jaegerthat's the md5sum15:02
jaegerShould have clarified15:02
jaeger- kernel/linux2.6.39.4 -> 3.5.415:03
jueyeah, that's exactly what I've committed some minutes ago15:05
jaegerYou didn't change anything after adding the release date and updating the kernel version, right? That's the latest change I see in git15:05
jaegerok, good15:05
jueit's over 2 years since 2.7 ...15:08
jaegertorrent is up and running15:08
teK__is there a torrent file?15:09
teK__and: congratulations!15:09
jaegeryes, in the files/ dir on crux.nu15:14
jaegerthe ISO is also uploaded to my mirror now15:21
juejaeger: many thanks, I'm updating the download page15:28
jaegeruploading the source ISO now, too15:28
jaegerjue: sent you a draft, anything you want to change or add to it?15:43
jaegerIt's concise like the 2.7 announcement15:43
juejaeger: no, looks fine for me15:46
jaegerok, shall I go ahead and send it?15:47
jueyeah, please do :)15:47
jaegerOn a side note I wonder if tilman intends to maintain his mirror15:48
teK__dont forget the wiki :}15:49
*** sepen has joined #crux-devel15:51
jueupdated cr15:52
jaegerupdate the "current release is" bit, too :D15:52
jueyep, done :)15:53
rmullCongrats and thanks, people15:54
jaegerWould anyone object to me removing 2.4, 2.5, 2.6, and 2.7 from the torrent tracker? They're still on the mirrors but I see no reason to keep running the torrents for them15:57
pitillokbd is working right jaeger. Thank you very much for all your work15:57
jaeger2.7.1 and 2.8 can stay15:57
jaegerpitillo: excellent. :)15:57
juejaeger: no, the older ones are useless16:01 has been down for what, a month?16:41
frinnstor is it just me?16:41
jueno, got a connection refused too16:43
jaegeryeah, it has been a while. I have a local copy of the alsa-lib tarball but should probably update the ports to use a mirror17:01
*** sepen has quit IRC18:51
*** deus_ex has joined #crux-devel20:34
*** deus_ex is now known as pedja20:35
jaegerhrmm... why does the handbook say linux-3.6.x instead of 3.5.x? heh21:13
*** __mavrick61 has quit IRC21:26
*** __mavrick61 has joined #crux-devel21:27

Generated by 2.11.0 by Marius Gedminas - find it at!