IRC Logs for #crux-devel Tuesday, 2014-07-08

*** mavrick61 has quit IRC02:38
*** mavrick61 has joined #crux-devel02:39
*** deus_ex has quit IRC03:45
Romsternrxtx, i was thinking of adding a .multilib file for your use to wine even though pkgmk does not use it.04:27
Romsteractaully nrxtx if port has .*-32 and no .32bit file assume it's got 32bit and 64bit04:28
Romsteryou already parse the depends on line04:29
Romsterand look for the .32bit file04:29
Romsterand stick a multilib tag on them like wine.04:29
Romsterjaeger, any idea if there is anything left of 3.1 release?04:30
Romsterother than cleaning up contrib for perl footprints04:31
*** novak has quit IRC05:10
*** novak has joined #crux-devel05:13
*** deus_ex has joined #crux-devel05:32
*** mike_k has joined #crux-devel05:48
*** mike_k has quit IRC06:08
frinnstjaeger: wtf, you want me to do it alone? :)06:17
frinnsthmm, i should start testing xorg-server 1.16. Some big changes for me since it merged glamor06:35
Romsteroh you sure you wanan do that before crux 3.1 is released?06:37
Romsterwell xorg-server 1.16 isn't out yet currently06:38
frinnstyeah its an rc06:44
Romsterearly days then.07:39
frinnstthats why i should *start* testing it :p08:10
Romsterheh, hmm should we be also testing wayland protocol?08:14
frinnstthat is probably a few years away, but if you feel like it - sure08:16
frinnstIm not holding my breath wrt. wayland. but hey, thats just me08:16
Romsterwell i feel it should be usable now on gtk3 programs.08:16
Romsterall it is is a new protocol that removes the old legacy code out of X08:17
frinnstyeah but x will still be the core of any unix desktop for years and years to come while wayland will be "that pesky extra crap i need to run program xxx"08:19
frinnstand wayland will run nested in X08:19
frinnstbut maybe I am wrong :)08:19
Romsteryeah wayland will work along side the existing protocols08:22
*** jue has quit IRC08:34
*** jue has joined #crux-devel08:35
*** jue has quit IRC08:38
*** jue has joined #crux-devel08:39
jaegerfrinnst: if they didn't do it in the middle of a work day, I'd be there12:29
jaegerfrinnst: get jue to go with :)12:29
frinnstwell im not to keen to do it myself12:30
frinnstheh, yeah i'll ask him again :)12:30
jaegerWhy not? Not like there's anything to prove12:30
jaegerjue: how long before you go on vacation? can we wrap up 3.1 before then? Seems like there have been quite a few updates since rc312:31
jaegerthough not a lot that affects the ISO.... lzo, nss, python would be nice to update12:45
frinnstfirefox 31 or whatever will soon be out too :)12:48
jaegernaturally, heh12:49
frinnstas long as there are no setup-helper ports on the near horizon i think we should try and get it out asap12:50
jaegerI think so, too12:50
juejaeger: would be nice if we can do it before, I'm here until thursday12:51
frinnstxorg-server 1.16 will probably be out in a few weeks time too that would require some user actions if they have glamor installed. if we wait much longer im gonna start pushing us to wait for 1.16 :)12:51
jaegerjue: ok, what do you see as things we must do still?12:52
jaegeradding a link to the udev persistent naming docs to the release notes is still on my list12:52
jaegeroh, someone already put that link in the release notes. :)12:54
frinnstI did some changes to the configure part wrt. networking. Just before you changed dhcp to be the default and i had to redo everything :D12:55
jaegersorry :/12:56
frinnsthehe np12:56
jaegerAnyone mind if I change "details can be found at the freedesktop wiki" to contain the actual URL instead? The text version of the release notes on the ISO won't show it otherwise12:57
jaegerMaking DHCP the default is something we probably should have done long ago, I think13:03
jueno, I don't see anything important besides the ChangeLog13:06
juejaeger: you've build the ISO already?13:07
jaegerThe last build was on the 3rd, can build a new one any time13:09
jaegerShall I build the final one today?13:14
jueyeah, I'm all for it ;)13:20
jaegerok :)13:21
juegreat, I'll tag our git and do the ChangeLog13:21
jaegershould I push iso/setup.dependencies before the tag? I think it changed13:22
jaegerI guess that's not too important since it gets rebuilt regularly during ISO development anyway13:24
jaegermaybe you just meant the ports trees, though, rather than iso.git :)13:25
jueyeah, I meant the ports tree :)13:33
frinnstomg, sepen is alive! he pushed an xterm update13:56
jaegerI might as well get that into the bootstrap since it hasn't gone far14:18
jaegeroh, just 3.0 for now, never mind. it'll get updated later14:19
jaegeryeesh. took a while but I finally got a drbd+postgresql+pacemaker+cman HA cluster going at work14:43
teK__he could update jdk, too15:54
teK__and his mail host15:54
teK__jaeger: where were the difficulties?15:55
jaegerMostly with drbd. Ubuntu 12.04 server sometimes installs linux-image-generic-lts-precise and sometimes installs linux-image-generic-lts-saucy (or quantal, or raring) and I have no idea why it selects different ones15:58
jaegerthe precise kernel has drbd 8.3 and the saucy kernel has 8.415:58
jaegerthe userspace tools are 8.415:58
jaegerthe drbd documentation is terrible regarding which commands and switches work in which version of drbd15:58
jaegerso I had to play with that a lot to get it working15:58
teK__checked /etc/apt/sources.list?15:58
jaegeryeah, I did to make sure I wasn't insane15:59
teK__funny that, while using a 'well' and professionally supported distro, it could have been far easier to package && use hand rolled software15:59
teK__(CRUX \o/)15:59
jaegerI set up a cluster at work and one at home, both installed on ubuntu server 12.04 (precise)15:59
jaegerthe ones at home installed the saucy kernel15:59
jaegerthe ones at work installed the precise kernel15:59
jaegerboth were fresh installs15:59
teK__srsly? -_-16:00
teK__the only difference being the HW makes me wonder..16:00
jaegerThey were also both installed in vmware esxi 5.5 hosts16:00
jaegeridentical VM hardware16:00
teK__which would be _really_ odd16:00
teK__apt, you had ONE job16:00
jaegerIt had to be something the installer did16:00
jaegerheh, indeed16:00
teK__probably, diffed /etc/apt?16:01
jaegerthey are the same16:01
jaegerwhich is why I believe the installer is responsible16:01
teK__odd. really odd.16:01
teK__even then it should update the precies kernel to saucy16:02
teK__any messages on held back packages on apt-get update?16:02
teK__+ how is this even possible.. you can jump releases with dist-upgrade I suppose16:02
teK__I see16:03
teK__[ ] legit16:03
jaegerThe only thing I can imagine is this scenario:16:03
jaegerI installed the work cluster using PXE and a network install from us.archive.ubuntu.com16:03
jaegerI installed the home cluster using an ISO in vmware16:03
jaegerboth 12.04.4 LTS16:03
jaegerso perhaps the net install of 12.04.4 isn't exactly up to date with the ISO install regarding the LTS enablement stack16:04
teK__that's supposed to be regulated by sources.list?!16:04
teK__if not, this stuff is seriously broken16:04
jaegerAfter they're installed it's easy to fix simply by installing the 'linux-image-generic-lts-saucy' package instead of 'linux-image-generic-lts-precise'16:04
jaegerwell, it makes sense in an odd way16:05
jaegerboth the -saucy and the -precise packages exist in the precise repo16:05
jueso, tagged our repos for 3.1 and pushed an initial version of our ChangeLog ;)16:05
jaegerthe *installer* selected the newer kernel16:05
jaegerjue: nice :)16:05
teK__yes, in an odd way..16:05
teK__this story makes me like the CRUX way of thins even more. Seriously16:06
teK__I can fix two or three minor glitches by hand without worrying about the stuff you described above16:06
jaegerthough fortunately this isn't an issue with other packages, just the enablement stuff (which is mostly the kernel)16:06
jaegerthe difference factor is that the package names aren't the same16:06
teK__and this broke drdb or did you have different issues there?16:08
jaegerI think the drbd in the 3.2.x kernel was just broken16:09
jaegerit works now using the 3.11.x kernel and drbd 8.416:09
teK__oh boy :)16:10
jaegerthis might help clarify the kernel thing:16:10
jaeger$ dpkg -l | grep -i linux-image | awk '{ print $2 }' | sort16:10
jaegerthe 3rd and 4th are the meta packages. the 1st and 2nd are the actual ones that get updated16:10
teK__at least things are working16:12
teK__sepen forgot to update the md5 sum of libburn16:13
jaegerbootstrap is done19:43
jaegerI'll build the ISO and upload it, would appreciate anyone who has some time tonight or tomorrow morning to do some final tests19:52
nrxtx<- just post the url :)19:53
jaegerwill do19:54
jaegerThe new ISO (rc99.9) is uploading to now.20:11
jaegerIf no problems occur with this ISO then it should be safe to finalize. The only change needed would be finalizing the release date in the ChangeLog20:12
jaegerI will do some tests with it when I get home tonight, too20:12
*** nrxtx has quit IRC20:36
*** nrxtx has joined #crux-devel20:40
nrxtxuhm jaeger cab you resend me the link please?20:45
jaegerit's in the logs, too20:45
nrxtxis the upload completed?21:12
nrxtxah second download got the correct md521:17
frinnstso the linux action show will be the 3rd and the 5th21:34
frinnstwith yours truly21:34
frinnstyou pussies :p21:34
jaegershould be interesting21:34
jaegeryou can complain about teK__ with many listeners21:34
jaegerJust make sure they know you're from Brazil first :D21:37
teK__I will give my own interview, if frinnst mentions me with a single sentence21:58
jaegerteK__: the conversation in #crux does make me think of a suggestion for the portdb cacher - if a repo is unreachable maybe rather than removing it from the cache, simply add a counter of "days unreachable" to the listing or something22:08
teK__yes that would make sense22:08
teK__but it's written in PHP22:08
teK__it should rebuild the db in the background and then atomically replace the old one, too22:09
jaegerprobably so22:10
nrxtxjaeger: worked fine so far with the iso, ethernet is not working completly but i guess that depends on qemu22:12
jaegernetworking can be tough in qemu sometimes, yeah22:12
jaegerappreciate the report22:12
nrxtxhm weird never had that problem22:21
nrxtxdid someone also tested it in qemu?22:21
jaegerI tested one of the previous RCs in qemu and I think teK__ did as well22:22
teK__I did22:23
teK__rc1 I think22:23
jaegerIt worked well once I figured out how to actually configure the networking at the time :)22:26
teK__I had someone ask about a clean setup with qemu and VMs without bridging because he was not able to figure it out22:27
teK__but in reality it's rather easy as soon as you know what you have to do :>22:27
nrxtxmaybe i only got something wrong for the network configuration, other iso (slax/tinycore) are working fine with ethernet22:29
teK__check kernel support with lspci -k (I hope that it's on our ISO)22:30
nrxtxboot the iso -> dhcpd interface -> got ip but no traffic possible22:30
teK__traffic where to?22:31
teK__the gateway?22:31
teK__the dhcp-server?22:31
teK__the internetz?22:31
nrxtxin qemu gateway is which is pingable but nothing else is reachable22:32
teK__by ping?22:32
teK__did you try an application layer protocol?22:32
nrxtxsame configuration of qemu with slax is working fine22:32
nrxtxwill try some http stuff22:33
jaegerfound it in the logs, heh22:34
teK__well  as soon as ping to the gateway works, every other protocol (i.e. the driver itself) should work. What about routes/default gateways?22:34
teK__yeah qemu can be challenging in the networking corner :]22:35
nrxtxapplication layer is working22:35
nrxtxweird stuff :D22:35
teK__no not weird22:36
teK__you started it as as a user with uid !=0?22:36
teK__icmp/ping requires raw sockets and guess what, only root can create these22:36
nrxtxqemu or in the virtual machine?22:37
teK__I start mine as root but with -runas qemu22:37
teK__so that it will switch privs22:37
nrxtxah damn it sry on this machine not :/22:37
nrxtxjaeger: iso & installation fully working :)22:38
*** nrxtx has quit IRC22:46
*** nrxtx has joined #crux-devel22:47
jaegerGlad to hear it :)23:03

Generated by 2.11.0 by Marius Gedminas - find it at!