IRC Logs for #crux-devel Wednesday, 2012-09-12

*** mike_k_ has joined #crux-devel01:14
*** deus_ex is now known as pedja01:53
*** mike_k has joined #crux-devel02:11
*** mike_k_ has quit IRC02:14
*** mike_k has quit IRC03:28
*** mike_k has joined #crux-devel04:24
*** acrux has joined #crux-devel04:32
*** acrux has joined #crux-devel04:32
*** acrux has quit IRC04:39
*** acrux has joined #crux-devel04:40
*** acrux has joined #crux-devel04:40
*** Romster has quit IRC04:42
*** acrux_ has joined #crux-devel05:05
*** acrux has quit IRC05:05
jueRFC: symlink /etc/mtab to /proc/self/mounts05:29
juethe biggest advantage I see if we do that with 2.8 is, that we can simplify our /etc/rc05:30
juethe whole mounting with -n and remounting to write to /ect/mtab etc. is not necessary if we use kernels mtab05:31
jueI've tested it on my laptop, the diff for /etc/rc is here -> http://919dffa9c5b81918.paste.se/05:31
jueadditionally I've removed the entries for /proc and /sys from /etc/fstab05:32
* horrorStruck knows he's not a dev but he'd +1 this anyway05:36
*** Nomius has quit IRC06:59
*** Nomius has joined #crux-devel07:13
horrorStruckjue: OK, did the switch too. good news is, as expected, it solved the fuse issue I had with util-linux-ng07:32
jueyeah07:34
jueyou're following my discussion with kzak in #util-linux?07:34
horrorStrucknope but i will now07:40
frinnstjue yeah I did that too07:49
horrorStruckjue: thanks for helping with my issue on #u-l07:49
frinnstbiggest "drawback" (or, feature) as users will see is probably the extra crap that gets displayed with "mount"07:49
*** sepen has joined #crux-devel07:53
juesepen: your mount problem is fixed :)08:02
jaegerjue: did you find anything that says the symlink is the proper way to go?08:14
jaegerI've got no objections if it's the right way but nobody seems to know for sure08:14
juejaeger: no, but I didn't searched a lot about that issue08:25
jaegerok08:26
jaegerin my experience searching won't yield much useful info in this area08:26
jaegerdoes the u-l guy have any input on that?08:26
juehmm, no idea08:27
horrorStrucki just asked08:38
horrorStruck·20:40·  kzak - horrorStruck: libmount (and mount) supports regular mtab. The current mainstream distros usually use symlink, because it's compatible with systemd and mount namespaces. The kernel is able to  provide per namespace /proc/mounts, regular mtab is shared between all processes and namespaces (it also means that you have to lock mtab before update).08:41
jaegerthat doesn't really answer the question, except to say that either way works depending on whether or not you use systemd08:44
jaegeror am I reading that wrong?08:44
horrorStruck·20:46·  kzak - horrorStruck: and in userspace we don't have all necessary information (for example you can mount read-write nfs, but the node on server is exported read-only then kernel will mount it read-only  -- the result is 'rw' in mtab, 'ro' in /proc/mounts)08:51
horrorStruckjaeger: frankly, this is not an area I'm confortable with at all, just copy/pasting08:52
jaegeralright08:52
jaegerI guess the takeaway is that we should just pick a method that works for crux and go with it because there's no "correct" answer08:54
horrorStruckbut it doesnt seem like a "systemd only" approach08:54
jueI tend to lean against the symlink approach because the 'mount -n foo ; unmout foo ; mount foo' steps in our /etc/rc are not very nice IMO09:00
horrorStruckbut simplifying rc and having the job done by the kernel sound like good things. also, it solved an issue i was having but this is working with regular mtab now anyway (well, next release)09:00
jaegeryou guys have done more research on it than I have, I will defer to your opinions here09:00
juebut as frinnst said:09:01
jue14:49 < frinnst> biggest "drawback" (or, feature) as users will see is probably the extra crap that gets displayed with "mount"09:01
jaegeryeah09:02
sepenjue: sorry I was out09:22
sepenjue: ohhh I see https://github.com/karelzak/util-linux/commits/master09:23
sepen"libmount: segfaults if neither..."09:24
sepenthanks jue ;)09:24
juesepen: np, thanks for your report09:25
sepenI wish I had been the time to do it, but nowadays I'm a bit busy at home to concentrate on something09:33
sepento have my girlfriend near is less productive than the boss ;)09:38
jaegerhrmm... chrome's new pepperflash stuff doesn't agree with twinview09:38
jaegercan only see half of a fullscreen flash window :P09:39
sepenjaeger: no stable updates so far http://googlechromereleases.blogspot.com.es/09:42
jaegerI wonder if I can use adobe flash in it still or if they've removed that ability09:43
jaegerlooks like the answer is yes and it was easy to do09:47
sepenjaeger: do you know about chrome://flags/ ?09:48
jaegerI've looked there, don't know a lot about it09:49
sepenyeah too many options to crash your browser :)09:52
horrorStrucksepen: not using chromium myself but it seems there's a newer release http://googlechromereleases.blogspot.com.es/2012/08/stable-channel-update_30.html09:54
sepenhorrorStruck: thanks I forgot this one09:55
horrorStrucknp09:55
sepenI moved all my dev stuff from the office to home, now I'm still finishing that migration09:55
*** c0x has quit IRC13:08
*** c0x has joined #crux-devel13:18
*** sammi`_ has quit IRC13:35
*** sammi` has joined #crux-devel13:44
*** c0x` has joined #crux-devel13:54
*** c0x has quit IRC13:57
*** c0x` has quit IRC14:02
*** c0x has joined #crux-devel14:03
*** mike_k has quit IRC15:06
*** Romster has joined #crux-devel17:53
*** Romster has joined #crux-devel17:53
*** sepen has quit IRC17:53
*** mavrick61 has quit IRC21:30
*** mavrick61 has joined #crux-devel21:31

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