2015-01-01 23:05:09 <BitL0G1c> where should 32 bit libraries be installed on 64 bit Alpine ? /usr/lib32 ?
2015-01-02 08:29:38 <ginjachris> Good Morning and a Happy New Year!
2015-01-02 15:35:50 <Laurentius1977> Hello there
2015-01-02 15:38:13 <Laurentius1977> I need little help in freeswitch + bluebox on Alpinelinux
2015-01-02 15:38:44 <Laurentius1977> I followed tutorial from wiki http://wiki.alpinelinux.org/wiki/2600hz
2015-01-02 15:39:28 <Laurentius1977> Everything seems to be fine when installing
2015-01-02 15:39:51 <Laurentius1977> But when I try to acces bluebox from browser
2015-01-02 15:39:59 <Laurentius1977> I get an error
2015-01-02 15:40:25 <Laurentius1977> Warning: require(/usr/share/webapps/bluebox/bluebox/config/config.php): failed to open stream: No such file or directory in /usr/share/webapps/bluebox/system/core/Kohana.php on line 413
2015-01-02 15:40:53 <Laurentius1977> Fatal error: require(): Failed opening required '/usr/share/webapps/bluebox/bluebox/config/config.php' (include_path='.:/usr/share/pear') in /usr/share/webapps/bluebox/system/core/Kohana.php on line 413
2015-01-02 15:41:09 <Laurentius1977> I installed also php-pear
2015-01-02 15:41:34 <Laurentius1977> and set it up
2015-01-02 15:41:34 <Laurentius1977> but the error is the same
2015-01-02 15:41:58 <Laurentius1977> do you have an idea what is the reason of this aerror?
2015-01-02 15:42:05 <Laurentius1977> THANKS
2015-01-02 15:46:39 <fcolista> Laurentius1977, which files do you have in this dir? /usr/share/webapps/bluebox/bluebox/config/
2015-01-02 15:49:12 <Laurentius1977> cobfig.php.dist database.php.dist email.php.dist locale.php.dist session.php.dist telephony.php.dist upload.php.dist
2015-01-02 15:49:21 <Laurentius1977> these are all the files
2015-01-02 15:49:43 <fcolista> copy config.php.dist to config.php and customize it
2015-01-02 15:50:20 <Laurentius1977> ok, I will do that
2015-01-02 15:50:29 <Laurentius1977> thank you a lot
2015-01-02 15:50:45 <fcolista> yw Laurentius1977
2015-01-02 16:04:25 <Laurentius1977> It opened the installer page, but wit another error
2015-01-02 16:04:29 <Laurentius1977> An error was detected which prevented the loading of this page. If this problem persists, please contact the website administrator.
2015-01-02 16:04:46 <Laurentius1977> >bluebox/views/installer/welcome.php [1]:
2015-01-02 16:05:06 <Laurentius1977> Non-static method form::open_section() should not be called statically, assuming $this from incompatible context
2015-01-02 16:11:39 <Laurentius1977> Someone encountered the same error in Arch Linux https://aur.archlinux.org/packages/bluebox-git/
2015-01-02 16:22:04 <Laurentius1977> the .preinstall.sh script ran with no problems...
2015-01-02 19:03:47 <Laurentius1977> I reinstalled all again on a clean environment
2015-01-02 19:03:59 <Laurentius1977> it seems that the error persists
2015-01-02 19:04:04 <Laurentius1977> Non-static method form::open_section() should not be called statically, assuming $this from incompatible context
2015-01-02 19:04:39 <Laurentius1977> I think there is a problem in php-5.6
2015-01-02 19:04:51 <Laurentius1977> I google it
2015-01-02 19:05:11 <Laurentius1977> strange
2015-01-02 19:05:43 <Laurentius1977> Does anyone have a clue about it, please...
2015-01-02 19:06:38 <Laurentius1977> if someone can replicate the error would be great appreciated
2015-01-02 19:07:40 <Laurentius1977> I followed the steps from wiki http://wiki.alpinelinux.org/wiki/2600hz
2015-01-02 19:08:41 <Laurentius1977> I installed also php-pear and made a "pear install DB"
2015-01-02 19:09:09 <Laurentius1977> no succes here
2015-01-02 19:12:46 <Laurentius1977> or do you know a GUI for freeswitch that works? I will try also acf-freeswitch but I think it's not fine grained like bluebox
2015-01-03 02:54:03 <darkfader> firstofall wishing you a nice new year
2015-01-03 02:54:20 <darkfader> second, since it's worrying at the moment, any idea about this:
2015-01-03 02:54:20 <darkfader>   vgxen_raid10-vfile--data0: event registration failed: 18696:3 libdevmapper-event-lvm2raid.so dlopen failed: Error loading shared library libdevmapper-event-lvm2raid.so: No such file or directory
2015-01-03 02:54:33 <darkfader> comes when i do lvconvert -m 1 /dev/vgxen_raid10/vfile-data0
2015-01-03 02:55:01 <darkfader> daveh0003:~# find / -name libdevmapper-event-lvm2raid.so
2015-01-03 02:55:01 <darkfader> /lib/device-mapper/libdevmapper-event-lvm2raid.so
2015-01-03 02:55:01 <darkfader> daveh0003:~# ls -l /lib/device-mapper/libdevmapper-event-lvm2raid.so
2015-01-03 02:55:02 <darkfader> -r-xr-xr-x 1 root root 10032 May 13  2014 /lib/device-mapper/libdevmapper-event-lvm2raid.so
2015-01-03 02:55:16 <darkfader> does exist but shitty devmapper is probably blind
2015-01-03 02:56:02 <darkfader> does the symbol error matter?:
2015-01-03 02:56:03 <darkfader> http://hastebin.com/oxufiqedad.coffee
2015-01-03 22:26:51 <ahills> if I accidentally upgraded a package, how can I downgrade it again?
2015-01-03 22:29:19 <ahills> and, on that note, does anyone have the 3.14.20-1-grsec package lying around ;)
2015-01-03 22:31:04 <ahills> ah, it's on the 3.0.5 ISO
2015-01-03 22:32:48 <openfbtd> ahills, you can look in your local package cache
2015-01-03 22:32:54 <openfbtd> /var/cache/pacman
2015-01-03 22:33:17 <openfbtd> Wait
2015-01-03 22:33:24 <openfbtd> I thought I was on #archlinux for a second
2015-01-03 22:33:26 <openfbtd> Sorry :)
2015-01-03 22:33:27 <ahills> I translated that to apk, but there are no .apks in there
2015-01-03 23:31:41 <ThatTreeOverTher> Why does avahi depend on dbus?
2015-01-04 00:18:06 <dne> ThatTreeOverTher: "for IPC with client applications" according to http://avahi.org/wiki/DownloadAvahi
2015-01-04 00:20:01 <ThatTreeOverTher> dne, what's funny is that says it's optional (and apk agrees); however, `avahi-daemon` refuses to start without D-Bus
2015-01-04 00:29:17 <ThatTreeOverTher> I mean, if that's what it takes, then forget it, I'll just set a static IP and write it down...
2015-01-04 00:31:38 <dne> the package seems to be built w/o "--disable-dbus" which I assume means it will depend on dbus
2015-01-04 00:32:33 <ThatTreeOverTher> ah. I believe I can handle a little recompile. Easier than setting up dbus, anyway
2015-01-04 01:01:24 <ThatTreeOverTher> strange. When compiling avahi, I got an error that gcc couldn't create executables. config.log: http://ix.io/fAk
2015-01-04 01:01:39 <ThatTreeOverTher> couldn't even finish the `./configure` properly.
2015-01-04 14:43:06 <malya> Hi guys
2015-01-04 14:43:40 <malya> Who are using i3 window manager?
2015-01-04 14:45:15 <malya> I have a problem with it
2015-01-04 14:49:09 <malya> anyone?
2015-01-04 14:49:26 <tvvdhSL5hEBn> o/
2015-01-04 14:49:36 <malya> 0/
2015-01-04 14:50:11 <malya> Could you help me?
2015-01-04 15:05:09 <tvvdhSL5hEBn> dunno
2015-01-04 18:54:43 <dalias> is there a way to get a more recent firefox by pinning packages? what's involved?
2015-01-04 19:34:17 <dalias> and is anyone working on packaging chromium? :)
2015-01-04 19:48:59 <plueschopath> packaging chromium was discussed sometime ago ... chromium seems to ship it's own libraries for anything and everything iirc ... https://bugs.alpinelinux.org/issues/285
2015-01-04 19:49:07 <plueschopath> I don't think anyone is working on it
2015-01-04 19:50:17 <plueschopath> regarding firefox, I'm not sure if I understand you correctly. is the current package version outdated?
2015-01-04 19:52:46 <dalias> current package is 31 i think
2015-01-04 19:52:52 <dalias> latest seems to be 34
2015-01-04 19:53:24 <dalias> i don't really follow firefox so i may be mistaken. i'd prefer to be using chromium if it were packaged
2015-01-04 20:20:29 <plueschopath> can you create a bug report @ redmine.alpinelinux.org for the outdated firefox package please?
2015-01-04 20:20:54 <dalias> i think it's known that the package doesn't follow bleeding edge
2015-01-04 20:22:18 <plueschopath> 34 seems to be stable :)
2015-01-04 20:22:51 <plueschopath> aah ... 31 seems to be the ESR release
2015-01-04 20:22:56 <plueschopath> https://www.mozilla.org/en-US/firefox/organizations/faq/
2015-01-04 20:24:20 <dalias> :/
2015-01-05 10:57:06 <vkrishn> interesting to read, https://tools.ietf.org/html/rfc7435.txt
2015-01-05 13:05:27 <kocka> ahoi
2015-01-05 13:39:04 <Phiphler> what are some use cases of Alpine Linux? The wiki has some suggestions, but what have you used it for?
2015-01-05 13:41:08 <vitronic> Phiphler, as server is a beauty
2015-01-05 13:41:42 <vitronic> web,mail,firewall,gateway,database,etc,etc,etc ....
2015-01-05 13:41:44 <Phiphler> how recent are the packages compared to other server oriented distributions?
2015-01-05 13:42:46 <rnalrd> hi
2015-01-05 13:43:17 <rnalrd> we have stable releases and edge (rolling release, like Arch)
2015-01-05 13:43:29 <vitronic> I see these very active boys, so I guess it's roling release, but honestly I'm not sure
2015-01-05 13:43:53 <rnalrd> so "edge" by definition gets the most recent versions
2015-01-05 13:44:23 <rnalrd> stable releases are every 6 months
2015-01-05 13:44:45 <Phiphler> ok that seems nice! Is there a list of packages somewhere or should I just install it in a VM and start querying?
2015-01-05 13:44:55 <rnalrd> http://wiki.alpinelinux.org/wiki/Releases
2015-01-05 13:45:30 <rnalrd> http://forum.alpinelinux.org/packages
2015-01-05 13:45:51 <vitronic> http://forum.alpinelinux.org/packages
2015-01-05 13:46:48 <Phiphler> oh nice, the forum link on the site didnt work for me for some reason
2015-01-05 13:48:17 <plueschopath> there is also this: http://git.alpinelinux.org/cgit/aports/tree/
2015-01-05 13:53:16 <jomat> Phiphler: My Infrastructure is based on currently 17 Alpine Linux Servers
2015-01-05 13:53:39 <Phiphler> jomat: do you provide professional services with those servers?
2015-01-05 13:54:07 <jomat> Phiphler: There are webservers, gateways, postgresql-servers, resolvers, vpn-enpoints, ...
2015-01-05 13:54:17 <jomat> Phiphler: that's just hobby :-)
2015-01-05 13:55:02 <Jean-Scotch> Phiphler: Hi. Similar to Jomat here but with 10 servers only ;) And it's my main professional infrastructure
2015-01-05 13:55:46 <Phiphler> jomat: do you have backup schemes deployed with those servers? Backing up other peoples data is something I've been interesting in doing, since I keep having to fix friends and family members computers
2015-01-05 13:55:46 <jomat> Phiphler: not yet, that's stoll on my todo-list... but i think it will be bacula
2015-01-05 13:55:50 <jomat> *still
2015-01-05 13:57:02 <Phiphler> cool stuff
2015-01-05 13:58:13 <jomat> Phiphler: i work for a hosting company with a few hundret debian/ubuntu servers (company policy) where we run bacula
2015-01-05 13:58:39 <Jean-Scotch> The missing piece is RBD support in the kernel and qemu-KVM to be able to deploy a CEPH cluster with Alpine ;)
2015-01-05 13:59:11 <jomat> Jean-Scotch: qemu-KVM should work
2015-01-05 13:59:15 <Jean-Scotch> qemu-kvm works great but without RBD support
2015-01-05 13:59:46 <jomat> ah, ok :-)
2015-01-05 15:42:15 <Phiphler> if I want to run a small server with Alpine, possibly with some Owncloud, small scale mail server and backups, what amount of RAM should I be aiming at?
2015-01-05 15:47:03 <rnalrd> Phiphler, i'm running owncloud, pgsql, roundcube, postfix, DMVPN hub in 384MB RAM
2015-01-05 15:47:37 <rnalrd> that's a disk install
2015-01-05 15:47:37 <Phiphler> whoa, Im impressed. I tried running Owncloud on a raspberry pi and it choked
2015-01-05 15:47:53 <rnalrd> run-from-ram i'd aim at least 512MB
2015-01-05 15:50:01 <Phiphler> well, if you are using owncloud you are gonna need a disk as well to store your data
2015-01-05 15:50:28 <rnalrd> yes, but you can still run-from-ram
2015-01-05 15:50:38 <rnalrd> and have only /var on disk
2015-01-05 15:50:38 <royger> Phiphler: I'm running tftp, rtorrent, NFS, minidnla and conserver on a 256MB rpi and it works perfectly fine, you just need a big swap parition on a USB disk
2015-01-05 15:51:03 <royger> I usually do streaming from there to my ps3 and it is very stable
2015-01-05 15:52:22 <Phiphler> royger: It was in fact not Alpine but arkOS I was running which caused the problems. arkOS is like "server-for-dummies", but Alpine has pretty decent documentation so I might try it instead
2015-01-05 15:53:11 <Phiphler> I'm not very good at Linux yet, I need some kind of guide to set stuff up properly (Arch was my first distro :P )
2015-01-05 15:54:10 <royger> I'm runnign something called raspbian or similar, by the time I set this up alpine was not available on arm, and I don't what to do it all again from scratch
2015-01-05 15:55:30 <Phiphler> mayby I will just try to make a big swap on my arkOS install then. How big is the swap you are using?
2015-01-05 15:58:06 <royger> 4GB
2015-01-05 15:58:45 <Phiphler> royger: ill scrape up a stick of that size and give it a try! Thanks for the advice
2015-01-05 16:00:01 <jomat> swap on usb sticks is slooooow and eats up your usb stick
2015-01-05 16:00:15 <jomat> you could also try to swap on a nfs share
2015-01-05 16:00:45 <Phiphler> jomat: nfs is sadly not an option in this case, and its a raspberry pi so swap on the SD card would be even slower
2015-01-05 16:01:01 <Phiphler> jomat: how about swap on a USB 2.0 harddrive?
2015-01-05 16:01:11 <jomat> thats faster :-)
2015-01-05 16:01:44 <Phiphler> lets try that first then
2015-01-05 16:02:03 <jomat> if you have some hd on your pi i'd suggest to use a partition/lv/swapfile on that :-)
2015-01-05 16:04:26 <royger> if you use a sd/usb stick as swap it will last from christmas to new year
2015-01-05 16:06:03 <jomat> 3 months for my last 16 G stick :-(
2015-01-05 16:17:36 <plueschopath> wow
2015-01-05 16:18:17 <plueschopath> are usb sticks so unreliable?
2015-01-05 16:21:20 <jomat> flash memory has a very limited amount of write cycles
2015-01-05 16:22:05 <jomat> reading is afaik no problem, but writing destroys them
2015-01-05 17:36:54 <ScrumpyJ1ck> Phiphler: welcome to apline linux :)
2015-01-05 17:37:00 <ScrumpyJ1ck> apline?
2015-01-05 17:37:12 <ScrumpyJack> Apline
2015-01-05 17:37:16 <ScrumpyJack> Alpine
2015-01-05 20:58:58 <cdxn> hello
2015-01-05 20:59:20 <cdxn> is there a tutorial on howto run alpine on the raspiberry pi?
2015-01-05 20:59:38 <ScrumpyJack> i think so
2015-01-05 20:59:44 <ScrumpyJack> it's not too tricky
2015-01-05 20:59:49 <ScrumpyJack> hang on
2015-01-05 21:02:51 <cdxn> i did not find one. i just created two partitions, one vfat(boot) and one ext2 and copy all files from the tgz onto the vfat partion. but it does not boot.
2015-01-05 21:03:00 <ScrumpyJack> arg, can't find it. you'll need a sd card with a fat32 partition (W95 FAT32 in fdisk) with fat filesystem on it
2015-01-05 21:04:06 <cdxn> i have done mkfs.vfat is this ok to?
2015-01-05 21:04:38 <ScrumpyJack> should be. have you set the boot flag on the partition?
2015-01-05 21:04:49 <cdxn> yep
2015-01-05 21:05:07 <ScrumpyJack> can you check the partition type in fdisk?
2015-01-05 21:06:56 <cdxn> ;o) you are good, it was wrong, it said linux swap...
2015-01-05 21:07:09 <cdxn> changed it to b(w95 fat32)
2015-01-05 21:07:17 <ScrumpyJack> cool
2015-01-05 21:07:54 <cdxn> thanks much will give it a try now :o)
2015-01-05 21:11:28 <ScrumpyJack> let us know how you get on
2015-01-05 21:12:58 <cdxn> thx ScrumpyJack that was the problem... it boots know
2015-01-05 21:13:10 <cdxn> -k
2015-01-05 21:14:20 <cdxn> i will digg deeper into configuration etc. tomorrow. its really nice to see a musl based distro with grsec kernel on the raspberry
2015-01-05 21:22:07 <cdxn> bye and thx
2015-01-06 04:37:56 <CcxCZ> jomat: actually that was true with SLC NOR flash memories, NAND and MLC both bring troublesome reads too, even adjeccent reads can flip bits etc.; the block relocation chips will rewrite read areas rather often for that reason...
2015-01-06 04:38:53 <CcxCZ> basically the higher information density, the bigger error rates
2015-01-06 04:40:54 <CcxCZ> also there's bunch of "recycled" od "refurbished" memories in usb sticks, so it's quite a lottery on how will one last
2015-01-06 05:21:56 <bougyman> i'm getting a 404 on alplinelinux.org/packages
2015-01-06 07:55:04 <ncopa> bougyman: should be fixed now
2015-01-06 07:55:04 <ncopa> thanks
2015-01-06 08:24:12 <ScrumpyJack> morning
2015-01-06 08:25:42 <plueschopath> morning :)
2015-01-06 08:51:55 <ncopa> morning
2015-01-06 08:57:08 <plueschopath> does anyone know why there are always two "unreachable default dev lo  table unspec  proto kernel  metric 4294967295  error -101" routes? they are also there when no interface is up and no route/ip address is configured
2015-01-06 08:57:42 <plueschopath> ipv6 works great ... but I don't know why these routes are there
2015-01-06 08:58:07 <plueschopath> they only show up by using "ip -6 route show table all"
2015-01-06 09:53:22 <vkrishn> CcxCZ, I wished there were USB ram sticks, nice for AL use
2015-01-06 10:05:21 <vkrishn> even 1-2gb would be very nice
2015-01-06 10:19:37 <jomat> CcxCZ: thx for the infos, didn't know those facts
2015-01-06 11:24:46 <vkrishn> nice , all low power, http://download.intel.com/newsroom/kits/ces/2015/pdfs/5th_Gen_Intel_Core_Factsheet.pdf
2015-01-06 11:29:55 <melville> hi. is anyone using SSD storage with alpine? i'm interested in changing io scheduler to deadline (it is provided as kernel modules on alpine: kernel/block/deadline-iosched.ko). is there a good way to modify sysfs on boot in alpine - some hidden corner of openrc?
2015-01-06 11:36:14 <melville> or should i go with just simple script as "echo deadline > /sys/block/sda/queue/scheduler" in /etc/local.d/ ?
2015-01-06 11:42:48 <melville> will check that in a minute
2015-01-06 11:47:36 <melville> seems to work, cat /sys/block/sda/queue/scheduler -> noop cfq [deadline]
2015-01-06 11:50:05 <darkfader> melville: there's iirc no hidden way to set sysfs on boot
2015-01-06 11:50:30 <darkfader> was there a distro that allows it? i have some memory of something... not much :)
2015-01-06 11:51:17 <plueschopath> melville: http://dev.nuodb.com/techblog/tuning-linux-io-scheduler-ssds
2015-01-06 11:51:31 <plueschopath> Add "elevator=noop" to the GRUB_CMDLINE_LINUX_DEFAULT line.
2015-01-06 11:51:41 <plueschopath> should also work with deadline I guess
2015-01-06 11:53:45 <melville> plueschopath deadline is a kernel module in alpine - don't think it will work when passed as option to kernel
2015-01-06 11:54:20 <plueschopath> ah...
2015-01-06 11:55:28 <melville> what i'm using right now, seems to work: http://pastebin.com/dut7UeSD
2015-01-06 11:57:15 <plueschopath> ah nice...
2015-01-06 11:57:29 <plueschopath> I don't know if there is a "better way" to do it.
2015-01-06 12:02:04 <vkrishn> I think I found a solution for fog/blizzard related car issues on road, use tiny NFC devices, embedded in middle of road, every 2-3meters, maybe loaded with AL
2015-01-06 12:02:11 <vkrishn> or just wait for smart car's ;-)
2015-01-06 13:04:08 <vkrishn> http://source.2600hz.org/browse/bluebox may be down, http://repo.2600hz.com/bluebox/bluebox-v1.0.3.rpm of year 2011
2015-01-06 13:18:42 <vkrishn> awesome, http://www.hpcwire.com/off-the-wire/mellanox-introduces-new-40-gigabit-ethernet-nics-supporting-ocp-2-0-specification/?utm_source=rss&utm_medium=rss&utm_campaign=mellanox-introduces-new-40-gigabit-ethernet-nics-supporting-ocp-2-0-specification
2015-01-06 13:18:54 <vkrishn> does anyone happen to buy one ?
2015-01-06 13:31:27 <vkrishn> with these kinda flow, definetly need bigger storage, ;)
2015-01-06 13:32:00 <vkrishn> http://www.techradar.com/us/news/computing-components/storage/hoarders-rejoice-a-360tb-1-million-year-storage-option-has-arrived-1165462
2015-01-06 13:32:20 <vkrishn> we might have entered another minor disruptive phase :-))
2015-01-06 14:12:05 <melville> is anyone using tmpfs as /tmp on alpine? wondering what flags are commonly used e.g. nosuid, nodev...
2015-01-06 14:17:25 <darkfader> i think i do
2015-01-06 14:17:41 <darkfader> well, no.
2015-01-06 14:18:06 <darkfader> too sleepy to think :/
2015-01-06 14:18:24 <darkfader> normally mode 1777 and yes, nosuid, nodev, noexec
2015-01-06 14:18:35 <darkfader> noexec normally bites me every year but it's healthy
2015-01-06 14:18:39 <jomat> whohoo, suddenly Alpine 3.1.1 %)
2015-01-06 14:18:59 <k0r10n> melville: you can use smth like rw,nosuid,nodev,relatime,size=787208k,mode=755
2015-01-06 14:19:23 <darkfader> what's also helpful is to separate the mysql or php tempdirs (noexec, too)
2015-01-06 14:19:31 <ncopa> does this look ok? http://wwwtest.alpinelinux.org/posts/Alpine-3.1.1-released.html
2015-01-06 14:22:21 <ScrumpyJack> ncopa: we have *got* to go that mqtt mirror sync thing - soo cool
2015-01-06 14:24:35 <ncopa> ScrumpyJack: we need -w options for busybox flock
2015-01-06 14:24:53 <ncopa> support for -w in busybox flock
2015-01-06 14:25:01 <melville> ncopa, announcement seems to be fine, guess you should wait for someone else to proofread it
2015-01-06 14:25:13 <ncopa> melville: thanks
2015-01-06 14:25:15 <ncopa> yeah
2015-01-06 14:25:30 <ncopa> someone english native that can proofread it?
2015-01-06 14:25:41 <ScrumpyJack> reading now
2015-01-06 14:26:00 <melville> thanks darkfader, i was also thinking about going data mode, not sure how well it would work for desktop
2015-01-06 14:26:46 <ScrumpyJack> alpine or Alpine, but it should consistent
2015-01-06 14:27:35 <ScrumpyJack> in fact i'd lean towards sticking to Alpine Linux as much as possible
2015-01-06 14:27:57 <ScrumpyJack> s/with the alpine vanilla iso image/with the Alpine Linux vanilla iso image
2015-01-06 14:29:50 <ScrumpyJack> the rest is fine
2015-01-06 14:30:05 <ncopa> thanks!
2015-01-06 14:30:07 <ncopa> piushed
2015-01-06 14:30:08 <ncopa> pushed
2015-01-06 14:32:15 <ncopa> could anyone please dl the 3.1.1 vanilla iso and check that it actually boots?
2015-01-06 14:32:29 <ncopa> i have tested it in my dev env
2015-01-06 14:32:30 <ncopa> so i believe it works
2015-01-06 14:32:36 <ScrumpyJack> ok
2015-01-06 14:32:36 <ncopa> but would be nice to doublecheck
2015-01-06 14:33:12 <ScrumpyJack> mirrors aren't synced :)
2015-01-06 14:33:25 <Evil_Bob> hi, question, why was support for sdl disabled for qemu? http://git.alpinelinux.org/cgit/aports/commit/main/qemu/APKBUILD?id=42d989462bd7ee65aa08eed229fb34d0e4dcf587
2015-01-06 14:34:18 <tvvdhSL5hEBn> Evil_Bob: it went out with the gtk support together.
2015-01-06 14:34:23 <ScrumpyJack> ncopa: it's on the site, but noone can download it. perhaps we should wait an hour or so before updating the page, that the repos are synced :)
2015-01-06 14:34:48 <ncopa> ScrumpyJack: try nl.alpinelinux.org/alpine/v3.1/releases/
2015-01-06 14:34:52 <ScrumpyJack> yeah got iy
2015-01-06 14:35:11 <ncopa> Evil_Bob: we removed x11 support alltogether
2015-01-06 14:35:22 <tvvdhSL5hEBn> i'm trying to build a qemu-fb version that still supports fb and a qemu-gtk version that runs with gtk. but i fail to do it correctly using subpakcages, so i wait for some more educated people to point out what i'm doing wrong.
2015-01-06 14:35:38 <ncopa> tvvdhSL5hEBn: i will look at it soonish
2015-01-06 14:35:55 <ncopa> hopefully today
2015-01-06 14:35:55 <ncopa> but i'll try flush the patch queue first
2015-01-06 14:35:55 <tvvdhSL5hEBn> no probs
2015-01-06 14:35:57 <tvvdhSL5hEBn> do you stuff
2015-01-06 14:35:57 <ncopa> and finish the 3.1.1 release
2015-01-06 14:36:15 <tvvdhSL5hEBn> i downgraded qemu to old versions until then.
2015-01-06 14:36:20 <Evil_Bob> tvvdhSL5hEBn: i just recompiled with --enable-sdl works like i want again for me, i just wondered why
2015-01-06 14:36:57 <tvvdhSL5hEBn> smaller dedicated packages. none of us really wants x11 as a dependency ;)
2015-01-06 14:37:22 <Evil_Bob> having to use spice or vnc seems silly, atleast in my humble opinion
2015-01-06 14:37:40 <ncopa> Evil_Bob: we wanted get rid of X11 deps
2015-01-06 14:37:54 <ncopa> you can run it graphically with spice, even if cumbersome
2015-01-06 14:38:00 <ncopa> or with virt-manager
2015-01-06 14:38:07 <ncopa> (even if that is *bloaty*)
2015-01-06 14:38:26 <ncopa> but once we link in x11 libs you cannot run without it
2015-01-06 14:38:26 <ncopa> brb
2015-01-06 14:39:07 <Evil_Bob> spice seems to suck in a short test, but thanks for giving an explanation anyway :)
2015-01-06 14:40:49 <ScrumpyJack> Evil_Bob: what feature of sdl do you like/use in a virtual environment?
2015-01-06 14:42:09 <Evil_Bob> ScrumpyJack: nothing in particular, i just want simplfy want to run it in a window
2015-01-06 14:43:00 <Evil_Bob> ScrumpyJack: using VNC over a unix domain socket might be an option, but the VNC clients in ports are horrible
2015-01-06 14:43:41 <ScrumpyJack> the Alpine Linux ports?
2015-01-06 14:43:41 <ncopa> Evil_Bob: i use virt-manager
2015-01-06 14:43:41 <ncopa> gives pretty nice ui
2015-01-06 14:43:47 <ncopa> but it eats *lots* of ram
2015-01-06 14:43:58 <ncopa> virt-manager + libvirt
2015-01-06 14:44:00 <Evil_Bob> cool, ill give it a try
2015-01-06 14:44:02 <ScrumpyJack> if you like a particular VNC client, we can make a aport if you like
2015-01-06 14:44:27 <ncopa> and i do agree that it would be nice with a quick and dirty qemu for x11
2015-01-06 14:44:33 <ncopa> i used the gtk interface
2015-01-06 14:44:40 <ncopa> which is pretty nice
2015-01-06 14:44:48 <ncopa> so the plan is to have a qemu-gtk subpackage
2015-01-06 14:45:04 <ScrumpyJack> :)
2015-01-06 14:45:26 <ncopa> someone sent me a patch for sdl with framebuffer but i havent had time yet to look at it
2015-01-06 14:45:49 <Evil_Bob> ScrumpyJack: cant think of a particular one i like, but the last one i used was tigervnc (which required some manual patching) :)
2015-01-06 14:46:15 <ncopa> for qemu use, spice is nicer
2015-01-06 14:47:34 <ScrumpyJack> vanilla 64bit boots
2015-01-06 14:47:47 <ncopa> perfect
2015-01-06 14:47:49 <ncopa> thanks!
2015-01-06 14:47:54 <ncopa> i assume 32 bit does too
2015-01-06 14:49:11 <ScrumpyJack> can't see the 32bit version
2015-01-06 14:49:23 <ScrumpyJack> ok found it
2015-01-06 14:49:43 <ncopa> http://nl.alpinelinux.org/alpine/v3.1/releases/x86/alpine-vanilla-3.1.1-x86.iso
2015-01-06 14:49:51 <ncopa> oh
2015-01-06 14:49:59 <ncopa> i now have a manual step...
2015-01-06 14:50:13 <ncopa> the signing...
2015-01-06 14:50:23 <ScrumpyJack> by the way my Alpine Linux mirror has been working nicely for a month or so now
2015-01-06 14:50:23 <vkrishn> ncopa, did you change the markdown parser ? or updated ?
2015-01-06 14:51:10 <ncopa> vkrishn: no?
2015-01-06 14:51:15 <ncopa> why?
2015-01-06 14:55:08 <vkrishn> I see validation error fixed
2015-01-06 14:55:33 <vkrishn> http://wwwtest.alpinelinux.org/posts/Alpine-3.1.1-released.html
2015-01-06 14:59:22 <vkrishn> ok, maybe the markdown txt is different
2015-01-06 15:01:17 <vkrishn> the perl Markdown.pl does not produce the error
2015-01-06 15:48:54 <mjohn> is there a mirror I can download Alpine x86 from? The link on the main page is down.
2015-01-06 15:49:40 <rnalrd> http://nl.alpinelinux.org/alpine/MIRRORS.txt
2015-01-06 15:50:06 <mjohn> rnalrd: thank you
2015-01-06 18:55:29 <Phiphler> flockport seems like an interesting project trying to make using LXC less annoying. The are even interested in Alpine. Are there any considerations I should have over running non security oriented distros inside of Alpine?
2015-01-06 21:59:22 <systmkor> hey all
2015-01-06 21:59:36 <systmkor> for some reason I can't get qemu to launch an SDL graphical display anymore
2015-01-06 21:59:40 <systmkor> it just creates a VNC server
2015-01-06 21:59:57 <systmkor> I just made sure my packages are updated and running 3.1.1
2015-01-06 22:00:13 <systmkor> running the linux-grsec kernel and was failing on my custom kernel prior
2015-01-06 22:01:31 <systmkor> I have installed all the SDL packages apart of main/
2015-01-06 22:02:20 <systmkor> this behavior wasn't happening to me a couple of weeks ago, and I can't think of anything I have changed that would alter qemu's behavior
2015-01-06 22:02:45 <tvvdhSL5hEBn> the packages changed
2015-01-06 22:02:52 <melville> sdl/gtk display targets were removed from qemu recently as to kill depedency on x11, there was discussion about it today.
2015-01-06 22:03:06 <melville> you can use vnc/spice for time being
2015-01-06 22:03:09 <tvvdhSL5hEBn> there's efforts to do seperate fb and x11 packages
2015-01-06 22:04:07 <melville> or downgrade qemu package
2015-01-06 22:04:35 <dalias> :/
2015-01-06 22:05:00 <melville> relevant msg from ml: http://lists.alpinelinux.org/alpine-devel/4431.html
2015-01-06 22:05:06 <dalias> it would be nice if qemu solved this by having a separate program/process for the graphics display
2015-01-06 22:05:31 <dalias> i.e. run the UI in a separate program from the emulator and communicate over shared memory and whatnot
2015-01-06 22:06:31 <melville> i think there was some effort do split some features into shared libraries: http://wiki.qemu.org/Features/Modules
2015-01-06 22:06:53 <melville> "Network block drivers (curl, iscsi, rbd, ssh, glusterfs) can be built as shared library modules with "--enable-modules" configure option. " from changelog for qemu 2.0
2015-01-06 22:07:21 <melville> guess it is still far from providing gtk/sdl as separate library
2015-01-06 22:09:40 <tvvdhSL5hEBn> we'll have seperate packages in short time i hope
2015-01-06 22:10:10 <tvvdhSL5hEBn> for x11 and for fb, and lacking any direct gfx support, only vnc
2015-01-06 22:15:41 <systmkor> well i tried using gtk-vnc
2015-01-06 22:15:51 <systmkor> but i can't seem to find the binaries
2015-01-06 22:16:10 <systmkor> i checked in arch and they are saved as gvnc and i think gvnc-capture or something
2015-01-06 22:16:55 <tvvdhSL5hEBn> if you don't need the graphic interface, you can also say -display curses or similar
2015-01-06 22:20:58 <systmkor> i have to do windows crap for an uncle of mine
2015-01-06 22:21:07 <systmkor> so sadly will need graphical display
2015-01-06 22:25:38 <melville> i used spice some time ago
2015-01-06 22:25:45 <melville> http://www.linux-kvm.org/page/SPICE
2015-01-06 22:26:00 <systmkor> well do you happen to know what gtk-vnc
2015-01-06 22:26:04 <systmkor> installs it's binary as
2015-01-06 22:26:19 <systmkor> i don't know how to get apk to show installed files for a package
2015-01-06 22:26:21 <melville> apk info -L gtk-vnc ?
2015-01-06 22:26:24 <systmkor> like for pacman
2015-01-06 22:26:25 <systmkor> okay
2015-01-06 22:26:42 <melville> without question mark
2015-01-06 22:27:01 <systmkor> yah
2015-01-06 22:27:03 <systmkor> that did it
2015-01-06 22:27:07 <systmkor> no /bin or /usr/bin
2015-01-06 22:27:10 <systmkor> just library files
2015-01-06 22:27:47 <systmkor> found it
2015-01-06 22:27:51 <systmkor> gvncviewer
2015-01-06 22:27:57 <systmkor> seperate package i guess?
2015-01-06 22:28:05 <systmkor> yup
2015-01-06 22:28:44 <melville> "gtk-vnc is a VNC viewer widget for GTK. It is built using coroutines allowing it to be completely asynchronous while remaining single threaded. It provides a core C library, and bindings for Python (PyGTK"
2015-01-06 22:29:06 <melville> to gtk-vnc is just one of the generic libraries from gtk
2015-01-06 22:29:14 <systmkor> got it
2015-01-06 22:29:17 <systmkor> yah it works
2015-01-06 22:31:10 <systmkor> thanks
2015-01-06 22:37:26 <melville> "apk --help" is your friend e.g. you can use something like "apk --help info" to get specific help for that part of apktools
2015-01-06 22:37:50 <melville> a side not
2015-01-06 22:37:52 <melville> note
2015-01-06 22:40:33 <systmkor> thank you :D
2015-01-06 22:40:39 <systmkor> oh and to get qemu to use kvm
2015-01-06 22:40:55 <systmkor> other than having the kvm kernel module loaded what-else do I need to do
2015-01-06 22:42:52 <melville> you need kvm-intel or kvm-amd kernel module afair and then just pass -enable-kvm
2015-01-06 22:43:26 <melville> there is also kvm user group, but don't remember if it was needed
2015-01-06 22:43:48 <systmkor> pass -enable-kvm for qemu?
2015-01-06 22:43:55 <melville> yeah
2015-01-06 22:45:44 <systmkor> strange i get a black screen with the new blue windows logo
2015-01-06 22:45:46 <systmkor> and that's it
2015-01-06 22:46:57 <systmkor> i have tried -vga std & vmware but no difference from default
2015-01-06 22:51:18 <melville> std is default right now i think
2015-01-06 22:51:23 <melville> since qemu 2.2
2015-01-06 22:51:32 <melville> you can try -vga cirrus
2015-01-06 22:56:12 <systmkor> i think that did it
2015-01-06 22:56:21 <systmkor> or just waiting long enough
2015-01-06 22:56:22 <systmkor> hizza
2015-01-06 22:56:39 <melville> you can check in monitor if kvm is on
2015-01-06 22:56:51 <systmkor> how do I do that?
2015-01-06 22:57:47 <systmkor> also I can't wait for steam OS to curb stomp Windows & Linux finally gets a good office suite
2015-01-06 22:58:02 <systmkor> burn Microshaft, burnn..... :P
2015-01-06 22:58:34 <melville> next time you can use "-monitor stdio" to attach monitor to terminal
2015-01-06 22:58:41 <melville> and then "info kvm"
2015-01-06 22:59:12 <melville> should return "kvm support: enabled" or so
2015-01-06 23:00:06 <systmkor> ah m'kay
2015-01-06 23:00:17 <systmkor> well i did make sure kvm module was loaded
2015-01-06 23:00:26 <systmkor> and -enable-kvm was there with no errors spat out
2015-01-06 23:00:34 <systmkor> but yah next-time i'll check that
2015-01-06 23:01:03 <systmkor> oh btw I made 1st run of having libressl port-package for Alpine
2015-01-06 23:01:25 <systmkor> i need to go back and fix it, since I forgot to properly put in dependencies in the APKBUILD
2015-01-06 23:01:28 <systmkor> and such
2015-01-06 23:02:13 <melville> you should notice if kvm acceleration is of or not by speed of emulation alone
2015-01-06 23:04:16 <systmkor> would it best for me to mask libressl as openssl and build packages from source
2015-01-06 23:04:21 <systmkor> to see if they work?
2015-01-06 23:04:25 <systmkor> for a full stack test
2015-01-06 23:04:51 <systmkor> so for example ssh compiles using the libressl-dev rather than openssl-dev
2015-01-06 23:07:12 <systmkor> yup kvm is enabled
2015-01-06 23:08:59 <melville> you can ask around if anyone played with libressl on alpine yet (wouldn't be surprised)
2015-01-06 23:09:22 <melville> there was also some conversation on ml: http://lists.alpinelinux.org/alpine-devel/index.html#msg4285
2015-01-06 23:10:34 <systmkor> melville, that was me :P
2015-01-06 23:11:28 <systmkor> oh and I am working on porting a friends password manager that he built to Alpine, it's wicked
2015-01-06 23:19:17 <melville> i'm contemplating on running alpine from ram (probably data mode) - not sure how it will work for desktop, and then using something like /home/username/persistent mounted on another partition
2015-01-06 23:24:25 <systmkor> melville, document on the wiki?
2015-01-06 23:28:17 <melville> you mean that i should document how it goes?
2015-01-06 23:29:09 <systmkor> yes please
2015-01-07 07:42:29 <ncopa> morning
2015-01-07 07:42:44 <dalias> hi
2015-01-07 07:42:55 <ncopa> qemu + spice should work as a temp work around for no qemu with sdl
2015-01-07 07:43:20 <ncopa> qemu-system-x86_64 -enable-kvm -m 512 -boot d -cdrom ~/Downloads/alpine-mini-3.1.1-x86_64.iso -spice port=5930,disable-ticketing
2015-01-07 07:43:34 <ncopa> spicec -h localhost -p 5930
2015-01-07 07:43:51 <ncopa> dalias: btw, i have an issue with qemu-2.2 and musl libc
2015-01-07 07:43:56 <dalias> oh?
2015-01-07 07:44:27 <ncopa> they use some internal field for pthreads i think
2015-01-07 07:44:34 <ncopa> thats not defined in posix
2015-01-07 07:44:50 <dalias> can you show me?
2015-01-07 07:45:09 <ncopa>      host_sevp->_sigev_un._tid = tswap32(target_sevp->_sigev_un._tid);
2015-01-07 07:45:19 <ncopa> generating a sprunge
2015-01-07 07:45:31 <ncopa> http://sprunge.us/QSMJ
2015-01-07 07:45:42 <dalias> ...
2015-01-07 07:45:52 <ncopa> its for linux user so i think they need pass the internal field
2015-01-07 07:48:10 <ncopa> i think its here: http://git.qemu.org/?p=qemu.git;a=blob;f=linux-user/syscall.c;h=aaac6a25ce21d944455da48d2cf46b11ea7642f1;hb=refs/heads/master#l5016
2015-01-07 07:49:49 <ncopa> i think they need it for SIGEV_THREAD_ID
2015-01-07 07:49:57 <ncopa> but its a bit above my head
2015-01-07 07:53:50 <ncopa> was introduced here: http://git.qemu.org/?p=qemu.git;a=patch;h=c065976f2bca9b87bc699c5fdeb4d3ff1299b8c4
2015-01-07 08:22:11 <dalias> they should just define a struct kernel_sigev or something for it
2015-01-07 08:22:27 <dalias> the userspace struct is different from the kernel struct here
2015-01-07 08:27:44 <dalias> hmm no tex packages for alpine?
2015-01-07 08:28:38 <dalias> btw on main page
2015-01-07 08:28:46 <dalias> http://dl-4.alpinelinux.org/alpine//v3.1/releases/x86_64/alpine-3.1.1-x86_64.iso.asc is 404
2015-01-07 09:01:27 <ncopa> dalias: i just signed the isos
2015-01-07 09:01:41 <ncopa> its a manual step in the release process
2015-01-07 09:01:48 <ncopa> that i forgot to put on my release checklist
2015-01-07 09:02:44 <ncopa> no tex packages unfortunally
2015-01-07 09:02:56 <ncopa> i think i looked at building tex and it was a big monster
2015-01-07 09:03:01 <dalias> :)
2015-01-07 09:03:08 <ncopa> since nobody had asked for it then...
2015-01-07 09:03:13 <ncopa> i just gave up
2015-01-07 09:03:43 <ncopa> i can have another look if you need/want it
2015-01-07 09:03:44 <dalias> i use it in preparing report documents for some work i do
2015-01-07 09:03:50 <ncopa> ok
2015-01-07 09:03:52 <dalias> but i can run it on another machine for now
2015-01-07 09:04:13 <ncopa> i'll give it another try later
2015-01-07 09:09:35 <eppa> hi, I just went through a fresh install of 3.1.1 in diskless mode and don’t seem to be able to keep hostapd and libnl installed between reboots?!  booting from SD card
2015-01-07 09:10:19 <eppa> another machine with setup but running 3.1.0 works fine :/
2015-01-07 09:10:24 <eppa> *same setup
2015-01-07 09:22:27 <ncopa> eppa: you need enabl apk cache
2015-01-07 09:22:52 <ncopa> for example:
2015-01-07 09:22:58 <ncopa> mkdir /media/usb/cache
2015-01-07 09:23:10 <ncopa> ln -s /media/usb/cache /etc/apk/cache
2015-01-07 09:23:16 <ncopa> apk sync
2015-01-07 09:24:18 <eppa> ncopa: the /media/usb/cache directory exists and looks like it contains some packages
2015-01-07 09:24:33 <eppa> and the /etc/apk/cache symlink already exists
2015-01-07 09:24:54 <ncopa> are the hostapd and libnl package there?
2015-01-07 09:25:00 <eppa> yeah
2015-01-07 09:25:19 <ncopa> hum, then i dont know...
2015-01-07 09:25:28 <ncopa> smells like a bug
2015-01-07 09:28:30 <eppa> ncopa: hmm ok thanks
2015-01-07 09:35:14 <ncopa> eppa: would be nice if you could find out whats going on
2015-01-07 09:42:11 <eppa> ncopa: yeah am investigating
2015-01-07 11:21:06 <darkfader> yt y
2015-01-07 13:32:04 <Frosh> pygrub doesn't exist in the new xen :(
2015-01-07 13:33:28 <Frosh> What is the process for adding the /boot/grub/menu.lst or getting it to boot the first time? usually I use pygrub to boot, pvgrub requires the menu.lst to be in place
2015-01-07 13:44:28 <dne> Frosh: there's the "compat-pvgrub" package, but it didn't work well that for me (e.g. it has hardcoded "root (hd0)")
2015-01-07 13:44:47 <dne> and on alpine the vmlinuz & initrd filenames usually don't change when the kernel is upgraded anyway
2015-01-07 13:45:21 <ncopa> dne: they did change from v3.0 -> v3.1
2015-01-07 13:45:28 <ncopa> vmlinuz and initrd names
2015-01-07 13:45:38 <dne> oh?
2015-01-07 13:46:28 <ncopa> vmlinuz-grsec
2015-01-07 13:46:40 <ncopa> but i dont think it affects disk-installs
2015-01-07 13:47:12 <dne> ok
2015-01-07 13:48:07 <dne> didn't have my own proper install until 3.1 anyway :)
2015-01-07 13:49:31 <Frosh> the way I do it for debian is during installation, i create the menu.lst, but its a hassle if I forget, so I was wondering if there is an alternative, incase I forget to create menu.lst during the install
2015-01-07 14:20:14 <Phiphler> how does running ubuntu/debian etc in LXC work in Alpine Linux? Any problems with the hardened kernel?
2015-01-07 14:21:22 <jomat> I have one Debian-Container on my Alpine hosts running Wordpress/MariaDB/ngnix/php-fpm for my gf... no problems
2015-01-07 14:22:19 <Phiphler> jomat: kewl. I like LXC a lot, Docker just has too many problems and full virtualization is a bit heavy
2015-01-07 14:22:35 <jomat> same here :-)
2015-01-07 14:22:49 <ncopa> same here (i havent done any docker though)
2015-01-07 14:23:01 <ncopa> there might some issues with debootstrap though
2015-01-07 14:23:22 <Phiphler> if Im not mistaken Alpine already implements some kind of access restriction for every installed program? Is there a security advantage in running stuff inside of LXC?
2015-01-07 14:23:39 <jomat> yes, the lxc templates shipped with alpine seem to be problematic
2015-01-07 14:23:55 <jomat> i installed that debian thing somehow manually
2015-01-07 14:24:13 <ncopa> Phiphler: we have grsecurity which has some access control support
2015-01-07 14:24:18 <ncopa> i havent really used it
2015-01-07 14:25:01 <ncopa> lxc is not a security feature
2015-01-07 14:25:07 <Phiphler> The guys over at flockport seem to have taken notice of Alpine and started publishing Alpine based containers
2015-01-07 14:25:15 <ncopa> ah thats nice
2015-01-07 14:25:23 <ncopa> i heard of flockport today
2015-01-07 14:25:48 <ncopa> and i'd think that alpine containers for things like flockport would make much sense
2015-01-07 14:26:06 <ncopa> since base tools are significantly smaller
2015-01-07 14:26:24 <ncopa> i think minimal debian is 110MB or so
2015-01-07 14:26:30 <ncopa> while alpine is 10
2015-01-07 14:26:33 <ncopa> 10MB
2015-01-07 14:26:37 <Phiphler> ncopa: they made a blogpost about how much smaller the containers became. Using Alpine as a host makes a lot of sense as well with the security minded stuff
2015-01-07 14:26:47 <Phiphler> it's really tiny
2015-01-07 14:26:51 <ncopa> yup
2015-01-07 14:27:00 <ncopa> do you have a blog post?
2015-01-07 14:27:04 <ncopa> link to blog post
2015-01-07 14:27:53 <jomat> probably http://www.flockport.com/new-micro-containers-based-on-alpine-linux/
2015-01-07 14:28:18 <ncopa> ah
2015-01-07 14:28:24 <ncopa> i think saw some email on that
2015-01-07 14:28:25 <Phiphler> yes thats it
2015-01-07 14:28:47 <jomat> nice :-)
2015-01-07 14:29:37 <ncopa> yes
2015-01-07 14:30:03 <ncopa> i knew alpine would give smaller containers but i never knew exactly how it compared to others
2015-01-07 14:30:07 <ncopa> very nice with some numbers
2015-01-07 14:32:09 <ncopa> re debootstrap
2015-01-07 14:32:15 <ncopa> i saw some bug on it
2015-01-07 14:32:35 <ncopa> #3523
2015-01-07 14:32:42 <ncopa> algitbot: hello?
2015-01-07 14:32:49 <ncopa> http://bugs.alpinelinux.org/issues/3523
2015-01-07 16:25:57 <bougyman> http://alpinelinux.org/packages is still 404ing
2015-01-07 16:28:44 <Phiphler> forum.alpinelinux.org/packages
2015-01-07 16:40:16 <bougyman> the link on the wiki points to the 404'd one, thanks
2015-01-07 17:03:56 <bougyman> anyone successful building ruby with rvm?
2015-01-07 18:36:55 <karasz> hello all
2015-01-07 18:37:29 <karasz> is anyone using enlightenment as desktop with alpine
2015-01-07 18:37:30 <karasz> ?
2015-01-07 18:43:42 <tvvdhSL5hEBn> is enlightenment even packaged in alpine?
2015-01-07 18:43:56 <karasz> e17 afaik
2015-01-07 18:44:20 <karasz> http://nl.alpinelinux.org/alpine/edge/testing/x86_64/enlightenment-0.17.0-r0.apk
2015-01-07 18:53:40 <tvvdhSL5hEBn> ooh.
2015-01-07 18:53:43 <tvvdhSL5hEBn> nice
2015-01-07 18:55:12 <karasz> quite old
2015-01-07 18:56:31 <fabled> karasz, i updated the packages once, and tested it. there was some issues had never time to finish it.
2015-01-07 18:58:20 <karasz> so xfce is the one more used?
2015-01-07 18:58:56 <jomat> I think a lot of people here use i3, I prefer awesome
2015-01-07 18:59:13 <fabled> xfce4 here; i think ncopa runs that too
2015-01-07 20:03:20 <ScrumpyJack> twm here, since you're asking :)
2015-01-07 20:14:10 <jomat> Ok, I can't generalize who's using which wm %)
2015-01-07 20:16:13 <fpletz> jomat: maybe the i3 and awesome users don't speak up because you already mentioned those wms :)
2015-01-08 01:07:34 <darkfader> anyone know if this should work:
2015-01-08 01:07:34 <darkfader> daveh0003:/sys/fs/bcache/fd765fb2-d8c8-40e7-8ae6-0e386b8f8630/stats_total# for lv in /dev/vgxen_raid10/* ; do lvconvert -f --mirrorlog core $lv ; done Conversion operation not yet supported.
2015-01-08 01:07:39 <darkfader> (nvm the path)
2015-01-08 01:07:56 <darkfader> i see it's in the redhat lvm regression tests but on alpine it seems to fail
2015-01-08 07:17:58 <ncopa> morning
2015-01-08 07:18:01 <ncopa> xfce4 here
2015-01-08 07:19:01 <ncopa> darkfader: "Conversion operation not yet supported"
2015-01-08 07:19:20 <ncopa> what are the requirements for "conversion operation"?
2015-01-08 07:20:13 <ncopa> it might be all you need to do is modprobe <somemodule>
2015-01-08 09:01:49 <jomat> dm-mirror probably
2015-01-08 09:25:52 <ScrumpyJack> silly question. I want to add my keys to /etc/apk. I have a package of my key and i'm adding it like this
2015-01-08 09:26:05 <ScrumpyJack> apk add sj-keys --allow-untrusted @st.ilet.to
2015-01-08 09:26:16 <ScrumpyJack> is that the correct syntax?
2015-01-08 09:36:35 <ScrumpyJack> i get this -> http://sprunge.us/bXjV
2015-01-08 09:44:01 <fabled> ScrumpyJack, apk add --allow-untrusted sj-keys@st.ilet.to
2015-01-08 09:47:41 <ScrumpyJack> thanks fabled
2015-01-08 10:45:00 <ScrumpyJack> what's the best way to install alpine with custom partitioning?
2015-01-08 10:46:00 <ScrumpyJack> if i create partitions before setup-alpine, will the script see them and give me the option to chose or will is use the entire disk
2015-01-08 10:46:30 <jomat> ScrumpyJack: use apk -p
2015-01-08 10:51:23 <ScrumpyJack> jomat: the changes the root of package install. I want to use setup-alpine
2015-01-08 10:51:27 <melville> ScrumpyJack "setup-alpine -q" and then you mount your custom partitions under /mnt and then "setup-disk /mnt" - i'll try to find an wiki article in a minute on thus
2015-01-08 10:51:49 <ScrumpyJack> i've seen something on the wiki
2015-01-08 10:52:09 <ScrumpyJack> http://wiki.alpinelinux.org/wiki/Setting_up_disks_manually
2015-01-08 10:52:54 <ScrumpyJack> but it's short for custom paritioning
2015-01-08 10:53:15 <ScrumpyJack> I'll play around and improve that section
2015-01-08 10:53:42 <melville> http://wiki.alpinelinux.org/wiki/Installing_Alpine_on_HDD_dualbooting
2015-01-08 10:53:47 <melville> this is a bit better
2015-01-08 10:56:12 <melville> and here description of "setup-alpine -q" (quick mode) and also setup-disk: http://wiki.alpinelinux.org/wiki/Setup-alpine#setup-alpine
2015-01-08 10:56:47 <melville> hope this helps
2015-01-08 10:58:09 <melville> setup-disk works nicely when you mount your custom partition under /mnt
2015-01-08 10:59:03 <melville> first you start with your root paritions e.g. "mount /dev/sda1 /mnt", then you mount other partitions like /home "mkdir -p /mnt/home && mount /dev/sda2 /mnt/home" and so on
2015-01-08 11:01:54 <ScrumpyJack> thanks, i'll try iy
2015-01-08 11:02:01 <ScrumpyJack> s/iy/it
2015-01-08 11:04:46 <jomat> Ah ok. the setup-* scripts do too much unwanted stuff for me, so my prefered way is apk -p /mnt/target add alpine-base mkinitfs syslinux linux-grsec ...
2015-01-08 11:05:20 <ScrumpyJack> ah, that sounds interesting. do you have notes?
2015-01-08 11:07:22 <ScrumpyJack> fabled: the video-decoder.apkvol.tar.gz has stuff you prolly don't want in there
2015-01-08 11:07:30 <fabled> oh?
2015-01-08 11:07:39 <fabled> i tried to look at it, but i made it in a bit of rush
2015-01-08 11:07:45 <ScrumpyJack> nameserver 10.252.5.129 in resolv.conf
2015-01-08 11:07:57 <fabled> dhcpcd will overwrite it, but yes
2015-01-08 11:08:49 <fabled> could remove the fi keylayout too
2015-01-08 11:09:06 <ScrumpyJack> are you in finland?
2015-01-08 11:09:10 <fabled> yes
2015-01-08 11:13:13 <fabled> need to go for a while
2015-01-08 11:13:15 <fabled> back later
2015-01-08 11:13:44 <melville> jomat - apk can be definitly also used, it is probably even more flexible e.g. chrooted installation
2015-01-08 11:50:37 <melville> that is quite nice hardware, supposed price is $139: http://www.intel.com/content/www/us/en/compute-stick/intel-compute-stick.html
2015-01-08 11:55:52 <ScrumpyJack> so what, it just plug into a hdmi port?
2015-01-08 11:58:40 <ScrumpyJack> for power?
2015-01-08 11:59:34 <melville> http://www.tomsguide.com/us/intel-compute-stick-hands-on,news-20215.html
2015-01-08 11:59:37 <melville> some more specs
2015-01-08 11:59:58 <melville> bluetooth, wifi
2015-01-08 12:01:53 <melville> i think it is powered via usb
2015-01-08 12:02:22 <melville> mhm, maybe i'm wrong
2015-01-08 12:06:44 <melville> "and while a micro USB port supplies the power, Intel aims to fully power the stick through the HDMI connection in the future." via http://www.extremetech.com/computing/196912-intel-compute-stick-transforms-any-hdmi-screen-into-a-desktop-computer
2015-01-08 12:16:37 <ncopa> ScrumpyJack: you can answer 'none' on disk question in setup-alpine. then you manually partition your disks
2015-01-08 12:16:49 <ncopa> you mount up your disk layout the way you want it
2015-01-08 12:17:02 <ncopa> then finally you run: setup-disk /path/to/mountpoint
2015-01-08 12:17:23 <ScrumpyJack> i'll do that
2015-01-08 12:17:25 <ncopa> setup-disk will detect lvm and raid
2015-01-08 12:18:01 <ncopa> it will expect /boot be ext* so it can run extlinux
2015-01-08 12:18:43 <ncopa> you might need create the MBR manually
2015-01-08 12:19:33 <ScrumpyJack> so i can have /boot in partition 1 ext4 and everything else in partition 2 used as a pvdisk?
2015-01-08 12:22:02 <vkrishn> anybody wanting to publish their indoor data ;)
2015-01-08 12:22:04 <vkrishn> http://docs.opengeospatial.org/is/14-005r3/14-005r3.html
2015-01-08 12:23:14 <vkrishn> melville, the linux version would be cheaper, $89, nice
2015-01-08 12:24:15 <ScrumpyJack> hmm, no tsocks in repros
2015-01-08 12:26:14 <vkrishn> ah, but 1gb ram, instead of 2gb
2015-01-08 12:30:21 <jomat> ScrumpyJack: should work..1
2015-01-08 12:49:40 <ncopa> ScrumpyJack: yes, /boot as ext4 and the rest as something should just work
2015-01-08 13:43:09 <ScrumpyJack> fabled: got omx running, thanks for the flavor files
2015-01-08 13:43:55 <ScrumpyJack> the apkvol didn't work for me, but i used the principle
2015-01-08 13:44:23 <fabled> strange.
2015-01-08 13:44:25 <fabled> ok, but nice.
2015-01-08 13:44:59 <ScrumpyJack> not your fault, that's because I'm behind a proxy, with no dhcp.
2015-01-08 13:45:16 <ScrumpyJack> the is no af_packet module on my rpi
2015-01-08 13:45:28 <ScrumpyJack> which was in your apkvol
2015-01-08 13:46:02 <ScrumpyJack> my rpi want to change root password when i ssh to it :(
2015-01-08 13:46:20 <fabled> hmm
2015-01-08 13:46:49 <ScrumpyJack> do you have a "higher def"  stream for new24 :)
2015-01-08 14:11:55 <fabled> ScrumpyJack, no, i just took it from some public web site
2015-01-08 14:12:14 <fabled> i'm using the same overlay against my private rtmp server
2015-01-08 14:13:58 <ScrumpyJack> ssh is asking me to change my root password every connection
2015-01-08 14:14:24 <ScrumpyJack> on the al-pi
2015-01-08 14:18:44 <ScrumpyJack> i think its my clock :)
2015-01-08 14:20:08 <ScrumpyJack> 1970
2015-01-08 14:55:41 <ScrumpyJack> fabled: rc-update add omxplayer default and then rc-service omxplayer start just start omxplayer but nothing shows
2015-01-08 14:55:57 <ScrumpyJack> if i start it on the cmd, it's fine
2015-01-08 15:11:38 <crow> ScrumpyJack maybe service is missing something like DISPLAY=1.1 omxplayer
2015-01-08 15:13:37 <ScrumpyJack> perhaps. i just used fabled's
2015-01-08 15:14:38 <ScrumpyJack> crom I'm not using X
2015-01-08 15:14:44 <ScrumpyJack> crow even
2015-01-08 15:14:57 <ScrumpyJack> 
2015-01-08 15:18:31 <crow> ScrumpyJack ok seems its not even needed: The advantage of the OMXplayer is that you don't need to start an X server to play a video. It fully uses the GPU capabilites of the Raspi, which makes it a fast and reliable player for full HD video. It also plays sound with the video.
2015-01-08 15:44:42 <ScrumpyJack> crow: yup, loving it at the moment. it's player bbc24 on a monitor in my openspace at work
2015-01-08 15:44:59 <ScrumpyJack> i does seem to die every so often
2015-01-08 18:25:58 <systmkor> ScrumpyJack, hey :D
2015-01-08 18:39:03 <ScrumpyJack> systmkor: hey
2015-01-08 19:49:45 <Jean-Scotch> hello
2015-01-08 19:50:49 <Jean-Scotch> I'd like to try linux-pam and nss-pam-ldapd but I cannot make it work and do not find any doc aboutit :(
2015-01-08 19:51:14 <Jean-Scotch> does anyone have some pointer?
2015-01-08 19:57:15 <Mirsc> hey da
2015-01-08 19:58:24 <Mirsc> i'm tryng to get alpine onto my rasberry...
2015-01-08 19:59:23 <Mirsc> and it seemed that there is no way to do a "sys" installion
2015-01-08 20:33:53 <systmkor> ScrumpyJack, did you get a chance to check out gitian?
2015-01-08 23:00:13 <Jean-Scotch> I suppose I have to recompile busybox to enable PAM ?
2015-01-09 07:31:33 <ScrumpyJack> Mirsc: there is no need to install alpine linux for the rpi.
2015-01-09 07:35:47 <ScrumpyJack> but you prolly figured that out by now :)
2015-01-09 10:14:08 <Jean-Scotch> I suppose I have to recompile busybox to enable PAM ?
2015-01-09 10:56:55 <ncopa> Jean-Scotch: i think so yes
2015-01-09 10:58:35 <Jean-Scotch> ncopa: ok. thanks. Will do that. and will build pam_mkhomedir also (see #3681); it's only one file to compile...
2015-01-09 11:00:10 <ncopa> would it be an idea to ship pam_mkhomedir with pam itself?
2015-01-09 11:00:14 <ncopa> its a single file
2015-01-09 11:00:19 <ncopa> a single commit
2015-01-09 11:01:57 <ncopa> seems like pam_mkhomedir is shipped with pam?
2015-01-09 11:02:17 <ncopa> pkg/linux-pam/lib/security/pam_mkhomedir.so
2015-01-09 11:10:11 <Jean-Scotch> ha! ok. great!
2015-01-09 11:37:21 <ScrumpyJack> i have an abuild query
2015-01-09 11:38:11 <ScrumpyJack> i can build a source using ./configure make manually, but not using abuild containing the same command
2015-01-09 11:38:32 <ScrumpyJack> anyone seen this sort of thing?
2015-01-09 12:27:49 <ScrumpyJack> hi ncopa
2015-01-09 12:28:15 <ScrumpyJack> ncopa: I'm having abuild issues
2015-01-09 12:28:20 <ncopa> ok?
2015-01-09 12:29:04 <ScrumpyJack> i can build on manually, but abuild can't
2015-01-09 12:29:21 <ncopa> what error do you get?
2015-01-09 12:29:25 <jomat> ScrumpyJack: how does the APKBUILD look like?
2015-01-09 12:29:35 <ScrumpyJack> simple
2015-01-09 12:29:39 <ncopa> it could be that CC is set in abuild
2015-01-09 12:29:49 <ncopa> it could be that MAKEFLAGS is set from abuild
2015-01-09 12:30:19 <ScrumpyJack> http://sprunge.us/FALN
2015-01-09 12:31:02 <ncopa> what is the error you get?
2015-01-09 12:31:32 <ncopa> btw, you should do: source="$pkgname-$pkgver.tar.gz::https://github.com/..../v1.8.gz"
2015-01-09 12:31:41 <ncopa> its a rc
2015-01-09 12:31:43 <ncopa> so
2015-01-09 12:31:51 <ncopa> pkgver=1.8_rc1
2015-01-09 12:32:00 <ncopa> _ver=${pkgver/_/-}
2015-01-09 12:32:33 <ScrumpyJack> http://sprunge.us/fHTO
2015-01-09 12:32:35 <ncopa> source="$pkgname-$_ver.tar.gz::https://github.com/vgough/encfs/archive/v$_ver.tar.gz"
2015-01-09 12:32:50 <ncopa> ./.libs/libencfs.so: undefined reference to `libintl_gettext'
2015-01-09 12:33:02 <ncopa> it should add -lintl to linker flags
2015-01-09 12:33:15 <ncopa> or gettext-dev should not be installed at build time
2015-01-09 12:33:35 <ScrumpyJack> i get the same error without gettext-dev
2015-01-09 12:33:59 <ScrumpyJack> and it works manually :(
2015-01-09 12:34:04 <ncopa> thats weird
2015-01-09 12:34:21 <ScrumpyJack> if I run build() manually, it compiles fine
2015-01-09 12:34:34 <ncopa> and if you do abuild build?
2015-01-09 12:34:41 <ScrumpyJack> (i was going to fix source later)
2015-01-09 12:34:51 <ScrumpyJack> I'll try
2015-01-09 12:35:18 <ncopa> it seems that it autodetects gettext from configure script
2015-01-09 12:35:27 <ncopa> and use it
2015-01-09 12:35:32 <ncopa> but does not add -lintl
2015-01-09 12:36:04 <ncopa> i think what happens is
2015-01-09 12:36:06 <ScrumpyJack> some error with abuild build
2015-01-09 12:36:15 <ncopa> during configure it will find gettext headers
2015-01-09 12:36:28 <ncopa> and say, oh we have gettext, lets use it
2015-01-09 12:36:31 <ncopa> then it will check
2015-01-09 12:36:40 <ncopa> do i need to add -lintl?
2015-01-09 12:37:13 <ncopa> and the way it check is to look for symbols in libc by linking a test app without -lintl
2015-01-09 12:37:35 <ncopa> in linking works it will think, og libs provides libintl, i dont need append -lintl
2015-01-09 12:37:49 <ncopa> because musl libc has a limited implementaton of libintl
2015-01-09 12:37:59 <ncopa> earlier it had none
2015-01-09 12:38:25 <ncopa> so back then the configure script would figure out that, linking without -lintl failed and then append -lintl
2015-01-09 12:38:27 <ScrumpyJack> i still don't see what it doen't behave like that when built manaully
2015-01-09 12:38:29 <ncopa> ofcourse
2015-01-09 12:38:40 <ncopa> i have no idea if this is the case
2015-01-09 12:38:50 <ncopa> the configure part will tell you that
2015-01-09 12:40:25 <ncopa> check output from configiure
2015-01-09 12:40:34 <ncopa> look for gettext or nls or intl
2015-01-10 16:29:24 <jomat> It should be possible to see in which order openrc will start all the things... but I don't find anything about it. Tips appreciated :-)
2015-01-10 16:30:37 <BitL0G1c> rc-update show
2015-01-10 16:32:12 <jomat> Hm. That doesn't show networking at all, but it is still started: http://pastebin.com/0dQSkfjD
2015-01-10 16:32:46 <jomat> http://pastebin.com/KWEiLyj6 <-  rc-update show
2015-01-10 16:35:43 <jomat> my problem is, that postgres will only listen on lo that way...
2015-01-10 16:36:41 <jomat> and adding "after networking" to depend() in the postgres init script doesn't help either...
2015-01-10 16:37:32 <jomat> ok, rc-update add networking changes the situation.
2015-01-10 16:41:34 <BitL0G1c> yes - networking is not installed by default in alpine (nor alpine lxc containers)
2015-01-10 16:41:48 <BitL0G1c> as a service
2015-01-10 16:48:14 <jomat> yep, that's ok. But why is it still started, and why is it started after a service that depends on networking?
2015-01-10 17:22:09 <Jean-Scotch> I have pb compiling busybox with PAM. the linker does not find libpam even if it is installed and linux-pam-dev required in the APKBUILD.  I found that ohter distro encounter same behavior but cannot find a working solution.
2015-01-10 17:22:45 <Jean-Scotch> i did open a post on the forum but I have little hope from there http://forum.alpinelinux.org/forum/installation/pb-compiling-busybox-pam
2015-01-10 18:18:46 <nsz> how hard it is to do an alpine install with disk encryption?
2015-01-10 18:19:17 <nsz> is there some ready made solution for partitioning, cryptsetup, extlinux boot settings,.. ?
2015-01-10 18:25:45 <jomat> I think there are some howtos in the wiki. And it works straight forward. Just ensure to add sth. like alpine_start=lvm,cryptsetup to the kernel parameters if you use encrypted volumes in a lvm (and not the other way round, lvm on an ancrypted disk)
2015-01-10 18:26:31 <nsz> ok i'll try it
2015-01-10 18:38:11 <BitL0G1c> @nsz - http://wiki.alpinelinux.org/wiki/LVM_on_LUKS
2015-01-10 18:38:50 <nsz> i dont want lvm
2015-01-10 18:39:13 <nsz> but i found the docs on the wiki
2015-01-10 18:42:45 <BitL0G1c> you will probably need to 'mkinitfs -c $MNT/etc/mkinitfs/mkinitfs.conf -b $MNT '     rather than 'apk fix...... ' to generate the initramfs
2015-01-10 20:35:05 <schoppenhauer> hello. I found a description on how to *run* alpine from pxe. but is there a possibility to run only the installer via pxe?
2015-01-10 20:39:12 <jomat> hi
2015-01-10 20:41:31 <jomat> schoppenhauer: should be possible, but I don't know the installer... I always bootstrapped my stuff... you should be able tu run alpine-setup by hand if that's what you need
2015-01-10 20:41:40 <jomat> *to
2015-01-10 20:41:47 <schoppenhauer> jomat: ok. thx.
2015-01-10 20:44:21 <jomat> schoppenhauer: another way is to fetch and extract the apk-tools-static package (.apks are just renamed .tar.gz files) and install base packages to a mounted future root filesystem: apk install --root /mnt/target alpine-base syslinux linux-grsec
2015-01-10 20:45:35 <schoppenhauer> jomat: unfortunately, I am not expierienced with alpine yet.
2015-01-10 20:46:01 <schoppenhauer> but its size would make it a good candidate for an old thinclient which I want to use as print server
2015-01-10 20:46:28 <jomat> schoppenhauer: What do you want to achieve with a bootable pxe installer? Is it a single installation or do you need a pxe setup for many hosts?
2015-01-10 20:47:01 <schoppenhauer> jomat: a single installation. the thinclient has no drives (except the internal flash-drive) and does not boot from usb-sticks.
2015-01-10 20:47:34 <jomat> if it's just the one thin client you want to install: try it the manual way... mount the flash drive with a grml or whatever live system, or mount it with a card reader on your notebook
2015-01-10 20:47:40 <schoppenhauer> jomat: I'm currently installing debian with an usb-stick attached, but I'd like to have a configuration without th eneed of an usb stick
2015-01-10 20:47:50 <schoppenhauer> jomat: ok.
2015-01-10 20:48:00 <schoppenhauer> jomat: thx!
2015-01-10 20:48:10 <jomat> i assume it is one partition, no encryption, no raid, no lvm?
2015-01-10 20:50:00 <jomat> schoppenhauer: and then to make everything bootable, you need to install a bootloader in the mbr... alpines default bootloader is extlinux, but I use grub (that's left over from debian) in my notebook to boot it
2015-01-10 20:50:24 <schoppenhauer> yes.
2015-01-10 20:50:30 <schoppenhauer> and ssh and cups
2015-01-10 20:50:38 <jomat> schoppenhauer: there are also some wiki pages: http://wiki.alpinelinux.org/wiki/Setting_up_disks_manually
2015-01-10 20:50:48 <jomat> http://wiki.alpinelinux.org/wiki/Installing_Alpine_Linux_in_a_chroot
2015-01-10 20:50:52 <jomat> http://wiki.alpinelinux.org/wiki/Replacing_non-Alpine_Linux_with_Alpine_remotely
2015-01-10 20:51:54 <jomat> If you still have questions, just ask - I installed my servers this way *without* the ability to see the vga output :-)
2015-01-10 20:53:48 <schoppenhauer> ++
2015-01-10 20:53:50 <schoppenhauer> thx
2015-01-10 20:55:05 <jomat> schoppenhauer: aaaand finally, here was a similar topic 2 days ago: http://irclogger.com/.alpine-linux/2015-01-08#1420714421 :-)
2015-01-10 21:09:01 <nsz> there is a todo in http://wiki.alpinelinux.org/wiki/Setting_up_disks_manually
2015-01-10 21:09:05 <nsz> Todo: Does adding the --update option to extlinux ... suffice to make /boot/extlinux.conf be regenerated? Or do we need to manually tweak that file, or run update-extlinux, as well?
2015-01-10 21:09:15 <nsz> the answer is no
2015-01-10 21:09:43 <nsz> you have to do the tweaks manually as far as i can see
2015-01-10 21:10:11 <nsz> update-linux does not seem to be configurable for this situation
2015-01-10 21:13:03 <nsz> the crypt device got mapped to /dev/mapper/crypt i assume this was done through the cryptdm=crypt kernel param, it would be nice if that was documented
2015-01-10 21:13:10 <Phiphler> just clearing this up: are packages in alpine linked statically or dynamically?
2015-01-10 21:15:40 <nsz> um i think most binaries are dynamically linked
2015-01-10 21:19:35 <Phiphler> nsz: Yeah I thought that was it, a friend was just suprised when I mentioned that alpine used musl since he thought musl could only be used statically
2015-01-10 21:20:43 <nsz> that was 3 years ago..
2015-01-10 21:21:03 <nsz> (the last time musl didnt support dynamic linking)
2015-01-10 21:26:36 <nsz> (a few things wouldnt work on alpine without dynlinking, eg xorg requires a working dlopen, most script languages need that too)
2015-01-10 23:25:11 <dhiru> Hi! I just installed alpine on my home machine :)
2015-01-10 23:25:27 <dhiru> I was wondering why Tor isn't in the main repositories.
2015-01-10 23:27:09 <dhiru> ok, got it, enabled "testing" and bam! :)
2015-01-11 09:38:36 <Jean-Scotch> hello
2015-01-11 09:38:48 <Jean-Scotch> I'm looking after kpartx
2015-01-11 09:39:01 <Jean-Scotch> how to fin din which package a file gelongs?
2015-01-11 09:39:11 <Jean-Scotch> *belonogs
2015-01-11 09:39:15 <Jean-Scotch> belongs
2015-01-11 09:43:23 <Jean-Scotch> found it : it's in multipath-tools
2015-01-11 09:43:46 <Jean-Scotch> but my question remains valid about finding specific binaries...
2015-01-11 09:52:15 <ScrumpyJack> Jean-Scotch: hello
2015-01-11 09:52:26 <Jean-Scotch> hi
2015-01-11 09:53:12 <ScrumpyJack> apk info who-own /path/to/binary
2015-01-11 09:53:14 <ScrumpyJack> opos
2015-01-11 09:53:24 <ScrumpyJack> apk info who-owns /path/to/binary
2015-01-11 09:53:42 <Jean-Scotch> thks
2015-01-11 09:54:44 <ScrumpyJack> apk info --who-owns /path/to/binary
2015-01-11 09:54:49 <ScrumpyJack> (butterfingers)
2015-01-11 09:55:50 <Jean-Scotch> ;) i's sunday morning after all
2015-01-11 09:55:56 <melville> or using shorter version with "apk info -W /path/to/binary"
2015-01-11 09:56:55 <melville> "-h" flag is your friend when it comes to apk tools e.g. "apk info -h" will give you hints specific to info
2015-01-11 09:59:46 <Jean-Scotch> I'm currently struggling with LVM to emulate an usb key to launch a KVM-alpine... I use regularly alpine in KVM from LVM but with regular disk install. But this time I want to run from tmpfs...
2015-01-11 10:00:52 <Jean-Scotch> to make some comparaison between FS on NBD.
2015-01-11 10:04:18 <ScrumpyJack> blimey, that is specific
2015-01-11 11:41:21 <nsz> i have a boot issue: i have a new disk, i havent put it into my laptop yet, i installed alpine on it through qemu and external usb->sata converter
2015-01-11 11:42:13 <nsz> boot works in qemu, but when i boot the laptop from the external disk it fails: at mount time the specified disk device is not available
2015-01-11 11:43:41 <nsz> it takes some time until the usb storage gets initialized and the /dev/sdb device appear
2015-01-11 11:44:02 <nsz> so is there a way to delay the 'mounting root' operation?
2015-01-11 11:45:19 <nsz> when the mount fails i get into the initramfs shell where i can manually do the cryptsetup, mount /sysroot then continue (because by that time /dev/sdb is available)
2015-01-11 11:45:35 <nsz> so it is only a timing issue
2015-01-11 11:46:20 <nsz> (i have to go now, but will check back in an hour or two)
2015-01-11 11:55:43 <nsz> based on the logs i think the mount happens around 1.5s and the disk appears at 2.8s
2015-01-11 12:00:33 <nsz> btw extlinux finds the boot partition of the external disk just fine, but linux is slow to find it
2015-01-11 12:11:43 <dne> nsz: try passing usbdelay=N (N in seconds) to the kernel command line
2015-01-11 13:16:37 <nsz> ok
2015-01-11 14:40:53 <nsz> dne: thanks usbdelay worked
2015-01-11 14:42:32 <nsz> btw is there a list of kernel params the init system uses?
2015-01-11 15:53:57 <dne> nsz: not that I've seen. I had a look in the source to find that one: http://git.alpinelinux.org/cgit/mkinitfs/tree/initramfs-init.in
2015-01-11 17:27:22 <nsz> btw /sys/class seems to be only accessible by root (i guess it's a grsec feature) how can a user access acpi info such as battery status?
2015-01-11 18:49:49 <BitL0G1c> anyone managed a usb stick install with multiple partitions ?
2015-01-11 18:51:46 <nsz> multiple partitions on the usb stick?
2015-01-11 18:52:35 <jomat> should work, doesn't it?
2015-01-11 18:54:16 <nsz> yes it should (i got it working with a /boot and encrypted / partition but that was a bit fiddly)
2015-01-11 18:55:37 <BitL0G1c> should i be setting syslinux.cfg to /dev/sdb1 (or /dev/sdb)
2015-01-11 18:55:59 <BitL0G1c> (sda is a windows drive)
2015-01-11 18:57:49 <nsz> i think if it's the 'where to boot from' option then sdb1
2015-01-11 18:58:03 <BitL0G1c> ok
2015-01-11 18:58:16 <nsz> if it's the 'where to cat mbr.bin' then /dev/sdb
2015-01-11 18:58:50 <nsz> i think you only need /dev/sdb for mbr, every other option is a partition with the filesystem set up
2015-01-11 18:59:03 <BitL0G1c> right
2015-01-11 19:02:41 <jomat> syslinux.conf? you mean extlinux.conf?
2015-01-11 19:03:15 <jomat> And the APPEND root= should be your partition of the rootfs
2015-01-11 19:03:16 <nsz> well if he boots from fat filesystem then i think it's syslinux
2015-01-11 19:04:07 <jomat> ah, didn't see the FAT :-)
2015-01-11 19:06:02 <BitL0G1c> yes syslinux - trying to make a bootable usb with a fat32 partition 1 & LVM partition 2 (for testing xen)
2015-01-11 19:07:58 <BitL0G1c> as xen doesn't seem to boot in vmware
2015-01-11 19:08:47 <BitL0G1c> or in kvm inside vmware
2015-01-11 19:34:46 <jomat> Does "whois" work for anyone in here?
2015-01-11 20:26:49 <BitL0G1c> whois: can't connect to remote host (204.74.78.75): Connection refused
2015-01-11 20:27:05 <BitL0G1c> maybe try 'jwhois'
2015-01-11 20:27:43 <BitL0G1c> host 204.74.78.75
2015-01-11 20:27:43 <BitL0G1c> 75.78.74.204.in-addr.arpa domain name pointer host-78-75.centergate.com.
2015-01-11 20:37:43 <melville> jomat, it works but you need to specify whois server address, you can find list of such servers here: http://www.nirsoft.net/whois-servers.txt
2015-01-11 20:38:16 <melville> example (which works for me): $ whois -h whois.pir.org alpinelinux.org
2015-01-11 20:39:15 <jomat> ah thx... do you know how the normal whois looks up the correct server from such a list? does it query one from the internet, or is there usually such a list shipped with it?
2015-01-11 20:40:19 <melville> "Most modern versions of whois try to guess the right server to ask for the specified object. If no guess can be made, whois will connect to whois.networksolutions.com for NIC handles or whois.arin.net for IPv4 addresses and network names. "
2015-01-11 20:40:30 <melville> via manpage from http://www.computerhope.com/unix/uwhois.htm
2015-01-11 20:40:41 <jomat> thx
2015-01-12 07:10:59 <opethfan> Evening / morning all. Has anyone encountered this same bug when running X? http://bugs.alpinelinux.org/issues/3691
2015-01-12 09:23:45 <Jean-Scotch> Hi. I try to mkfs.xfs with external journal but it does not let me do it with external journal. with internal one it's fine. Is it a bug or a wanted feature?
2015-01-12 09:28:22 <ncopa> Jean-Scotch: what are the requirements for external journal?
2015-01-12 09:29:38 <Jean-Scotch> ncopa: just to pass some device: mkfs.xfs -l logdev=device1 device2
2015-01-12 09:30:56 <Jean-Scotch> I checked the APKBUILD of xfsprogs and I see nothing there about disableing anything
2015-01-12 09:31:03 <ncopa> no special kernel compile option is needed? ro some compiletime option for xfsprogs?
2015-01-12 09:31:12 <ncopa> or*
2015-01-12 09:31:52 <ncopa> does anybody know who made the https://registry.hub.docker.com/u/alpinelinux/base/ image?
2015-01-12 09:37:41 <Jean-Scotch> I'm looking at it after it at the moment... will report my findings
2015-01-12 14:15:46 <Jean-Scotch> in fact, other -l options (for tuning journal with mkfs.xfs) are giving errors also.
2015-01-12 14:16:22 <Jean-Scotch> for now, I will use the defaults but will continue to look after this topic
2015-01-12 17:49:46 <m_jones> The web forums are far too quiet, so here I inquire:
2015-01-12 17:50:39 <m_jones> Are there any performance or scalability implications to using Alpine with musl on large-RAM servers that aren't virtualisation hosts?
2015-01-12 17:52:14 <m_jones> I'm running some test guests, but haven't concluded whether Alpine has any applicable advantages over Debian for large AMD64 hosts.
2015-01-12 18:09:13 <m_jones> ehlo, traveller.
2015-01-12 18:11:22 <buckley310> 250-hello
2015-01-12 18:13:33 <m_jones> I don't type 8BITMIME either.
2015-01-12 18:13:40 <ahills> m_jones: I don't understand the question... are you asking if musl is 64-bit compatible?
2015-01-12 18:13:51 <m_jones> ahills: no
2015-01-12 18:14:25 <m_jones> I'm working on a project that will run kernels on metal with 256GB and larger ram
2015-01-12 18:14:38 <ahills> If the Linux kernel supports it, ...
2015-01-12 18:15:17 <m_jones> I like what I see from preliminary use of Alpine, but I'm trying to decide if there are any quantifiable and defendable advantages of using it over Debian.
2015-01-12 18:15:37 <ahills> The package manager is much, much better
2015-01-12 18:15:40 <m_jones> Yes, the stock kernel is newer and has pax and grsec, which are notable
2015-01-12 18:15:48 <ahills> for creating your own packages
2015-01-12 18:16:26 <m_jones> I see no performance analysis of musl vs. glibc, and that can be dependent on architecture ISA
2015-01-12 18:16:27 <ahills> Doesn't Debian package a grsec+PaX kernel?
2015-01-12 18:16:48 <m_jones> but before I start looking at the inline asm in musl for applicability to Haswell I thought I'd ask around ;)
2015-01-12 18:16:51 <ahills> http://www.etalabs.net/compare_libcs.html
2015-01-12 18:17:28 <m_jones> I never looked for pax+grsec in Debian. Let me check 7.7...
2015-01-12 18:18:04 <m_jones> kernel-patch-grsecurity2 - transitional package for Debian Lenny linux-patch-grsecurity2 - grsecurity kernel patch
2015-01-12 18:18:15 <m_jones> maybe not in my repos. Won't bing it now.
2015-01-12 18:18:21 <ahills> I'm assuming that people other than you will want to access these machines
2015-01-12 18:18:43 <m_jones> it's a webservices application, so shell shouldn't matter
2015-01-12 18:18:47 <ahills> I use Debian and Alpine at work, both, on many hosts
2015-01-12 18:19:16 <ahills> The thing that surprises people on Alpine boxes is busybox coreutils and ash vs. GNU coreutils and bash
2015-01-12 18:19:37 <ahills> So, on the Alpine boxes, I have an MOTD with a "Notes for operators"
2015-01-12 18:19:42 <m_jones> although now that you mention it, love the ash. Bash was always an expediant decision for /bin/sh from the earliest days, but that wasn't recognised by the broader community and a lot of people got used to it.
2015-01-12 18:19:49 <ahills> Yeah, I like ash a lot
2015-01-12 18:20:04 <m_jones> plus shellshock.
2015-01-12 18:20:06 <ahills> I also love the busybox commands, they have a very consistent interface
2015-01-12 18:20:24 <ahills> Anyways, leaving an MOTD is enough for my coworkers, who are all at least moderately comfortable on the command line
2015-01-12 18:20:53 <m_jones> But bash used to be 10 times slower and 20 times the memory footprint according to one (analysis? comment?) from the last century.
2015-01-12 18:21:16 <ahills> Most of the machines don't need further configuration ever, though, so if you're going to be installing new packages frequently, it may be worth sticking with Debian for the likelihood that several users of the machine will be able to find all the software packages they need
2015-01-12 18:21:25 <m_jones> ahills: that feedback is useful. I use motd for op notes and change-control also.
2015-01-12 18:21:52 <ahills> For me, it really comes down to how many operators the machine will have... if it's me and my team, Alpine, if it's broader, Debian
2015-01-12 18:22:02 <m_jones> Right now I'm planning on making packages either way, apk or apt/deb
2015-01-12 18:22:15 <ahills> apk makes package creation fun!
2015-01-12 18:22:31 <ahills> you can run your own repo trivially
2015-01-12 18:22:42 <m_jones> The only conflict I've had with busybox is missing the operator -ls from find
2015-01-12 18:22:52 <ahills> never used it
2015-01-12 18:22:59 <ahills> but I'm mostly a BSDer
2015-01-12 18:23:36 <jomat> You can still apk add findutils :-)
2015-01-12 18:23:36 <m_jones> I was as well, and might be again. I haven't checked to see if BSD offers more performance in this application.
2015-01-12 18:23:43 <jomat> that supports -ls :-)
2015-01-12 18:24:11 <m_jones> Free, Net, Open, Dragonfly, and BSD/OS on more archs than you can count. I'm starting to sound like a greybeard.
2015-01-12 18:25:21 <m_jones> I'm aware of that etalabs.net comparison but it wasn't clear to me of the practical implications in a much-larger environment.
2015-01-12 18:25:49 <nsz> practical implication of what?
2015-01-12 18:26:02 <m_jones> thanks for the 'add findutils', I was looking for that conjuration
2015-01-12 18:26:23 <nsz> the comparision compares many different aspects of the c runtime
2015-01-12 18:26:25 <m_jones> nsz: performance mainly, and any other implications in large-server-on-metal environments
2015-01-12 18:27:22 <nsz> there are places where musl performance is better and others where it is worse
2015-01-12 18:27:32 <m_jones> I wanted to know if anyone had tests, or even informed speculation
2015-01-12 18:27:39 <nsz> sure
2015-01-12 18:28:03 <nsz> i dont think there are 'large server environment' tests
2015-01-12 18:28:18 <m_jones> I can dig into the asm in musl and glibc but I'm no assembler expert, and I've never done am on x86 (or I should say AMD64, since that arch drops so much x86 compat)
2015-01-12 18:28:30 <m_jones> it'd be faster to test.
2015-01-12 18:28:30 <nsz> but i can tell you that eg musl regex is slower than glibc's but in turn it is more correct
2015-01-12 18:28:47 <m_jones> s/am/asm/g
2015-01-12 18:28:59 <nsz> you dont need to dig into much asm i think
2015-01-12 18:29:45 <m_jones> I was just looking at the avx2 support in the kernel raid6 code for similar reasons. Haven't looked at the libcs.
2015-01-12 18:30:05 <nsz> on server threads may matter a lot, to our knowledge musl pthread implementation is more robust than glibc's (there is a long list of glibc pthread bugs)
2015-01-12 18:30:21 <m_jones> But the implication for the former is that if you have a haswell or newer, or anything that supports AVX2 really, your raid6 resilvers should benefit measurably.
2015-01-12 18:30:27 <nsz> libc has not much to do special cpu support
2015-01-12 18:30:32 <m_jones> the raid6 code also uses sse and sse2.
2015-01-12 18:30:56 <nsz> since libc linked binaries want to be portable across the same abi to different systems
2015-01-12 18:31:29 <m_jones> static binaries, per Rich Felker, right?
2015-01-12 18:31:39 <nsz> no?
2015-01-12 18:32:13 <nsz> if you optimize your binary to avx2 then it will only run a few select archs.. you wont be able to test it in some emulators or other systems
2015-01-12 18:32:36 <nsz> this is independent of static vs dynamic linking
2015-01-12 18:32:37 <m_jones> right, but if the asm is in the libc then it wouldn't matter on dynamically linked libs.
2015-01-12 18:33:03 <m_jones> I'm not contemplating inline asm in the application, but in the libc.
2015-01-12 18:33:32 <dalias> m_jones, there's almost nothing in libc that would benefit from avx2
2015-01-12 18:33:34 <nsz> i see, yes if you dont want to move the libc you can optimize it for a specific arch
2015-01-12 18:33:52 <dalias> m_jones, perhaps memcpy and some string functions could be faster in some cases with it
2015-01-12 18:34:12 <dalias> but that's about all
2015-01-12 18:34:26 <m_jones> dalias: I wouldn't know, but I'm told there's some amount of x86/amd64 asm in glibc?
2015-01-12 18:34:53 <m_jones> not necessarily avx2, which is an extremely new instruction set.
2015-01-12 18:35:04 <nsz> (there is less and less because they figured out most of them actually made things slower)
2015-01-12 18:35:25 <m_jones> well, that's kind-of a musl-specific question, and I was also looking for general feedback. ahill has given me some.
2015-01-12 18:36:05 <dalias> asm is subject to what i like to call asmrot
2015-01-12 18:36:20 <m_jones> nsz: fair enough, I could see that. I won't dig into it unless I find further evidence that it matters.
2015-01-12 18:36:24 <dalias> unless you constantly tweak it, within a couple years it's likely to be much slower than the compiled C code on new cpsu
2015-01-12 18:36:27 <dalias> cpus*
2015-01-12 18:36:31 <nsz> you will see some speed up of standard unix tools just because of the locale system
2015-01-12 18:36:47 <m_jones> yeah, plus clang is keeping gcc more honest these days.
2015-01-12 18:37:02 <nsz> (the way non C locale is handled on glibc makes string processing slower than on musl)
2015-01-12 18:37:14 <nsz> (in exchange you only have utf8 on musl)
2015-01-12 18:37:19 <dalias> btw i don't have any good explanations for _why_ yet, but apparently musl's dynamic linking runs faster for lots of things
2015-01-12 18:37:26 <dalias> despite being completely unoptimized
2015-01-12 18:37:36 <m_jones> nsz: ok, good. didn't realize there was a performance implication.
2015-01-12 18:38:31 <m_jones> I probably need to do some tests. For which I will need to order hardware.
2015-01-12 18:38:54 <m_jones> Because I now have some time and a need, but no longer have a big pile of big iron on which to play.
2015-01-12 18:39:29 <m_jones> I'm kinda intrigued with x32 on <4GB core systems, too.
2015-01-12 18:40:30 <m_jones> totally unrelated question:
2015-01-12 18:40:52 <m_jones> I can't manage to get my root filesystem mounted with option 'discard' at bootup
2015-01-12 18:41:19 <m_jones> I can mount other fileysystems that way from fstab, and I can remount -o discard root, but not from boot.
2015-01-12 18:42:00 <m_jones> I regenerated initrd to no effect. I feel a bit silly, but I don't know if the problem might be specific to grsec or other localisation, or just my unfamiliarity with linux, or what.
2015-01-12 18:47:31 <m_jones> actually, there's something in the logs I missed earlier. The root filesystem remounts at 1.5 seconds into boot, but the log only says that options are data=ordered
2015-01-12 18:50:02 <m_jones> It seems like there's a...capability? in System.map-grsec for blkdev_issue_discard
2015-01-12 18:52:17 <m_jones> Does anyone know from where mkinitfs gets the fstab for the root filesystem?
2015-01-12 18:52:38 <nsz> i think system.map are just kernel internal symbols
2015-01-12 18:53:15 <nsz> mkinitfs is an alpine linux tool, you need to read the source i think
2015-01-12 18:54:40 <nsz> http://git.alpinelinux.org/cgit/mkinitfs/tree/mkinitfs.in
2015-01-12 18:55:05 <m_jones> nsz: yes, looks like kernel symbols and not access-control capabilities, compared to Debian. I haven't had to dig into this before.
2015-01-12 18:56:01 <m_jones> nsz: thanks, tracking it down now.
2015-01-12 18:59:28 <m_jones> ha, busybox 'more' doesn't support searching
2015-01-12 18:59:59 <m_jones> probably everyone but me calls 'less'
2015-01-12 19:01:48 <dalias> imo to notice a major performance difference on server by switching libc, either the programs you're running or one of the libcs would have to be doing stuff that's badly wrong :)
2015-01-12 19:02:27 <dalias> the most plausible causes for difference would be string.h functions
2015-01-12 19:09:25 <m_jones> dalias: mmm. Also, it's looking like one of my big processes is the dreaded c++, and I actually don't know how much libstdc++ uses of libc
2015-01-12 19:09:57 <m_jones> like they say on the big iron: you just have to test your codes and see.
2015-01-12 19:10:10 <dalias> libstdc++ is rather huge, and the underlying libc probably doesn't make a lot of difference to it
2015-01-12 19:11:37 <vkrishn> I wonder if phoronox would take the task of benchmarking barebone alpinelinux vs debian/ubuntu/rh
2015-01-12 19:11:39 <m_jones> really I should test alpine, debian, and freebsd for app performance and for networking performance, as those are my areas of concern. I was hoping to find clear indications without doing the testing, but I certainly haven't!
2015-01-12 19:12:18 <m_jones> vkrishn: now that's a good idea. The P.R. would be great if nothing else.
2015-01-12 19:12:48 <vkrishn> also slackware as that would cover musl/eglibc/glibc
2015-01-12 19:12:52 <m_jones> vkrishn: I know I like <30MB of core used after boot, compared to Debian
2015-01-12 19:13:10 <ahills> yeah, that's the biggest speedup I notice on Alpine at work
2015-01-12 19:13:19 <ahills> oops, forgot to scroll
2015-01-12 19:13:20 <vkrishn> when comes to size for servers, AL beats all
2015-01-12 19:13:43 <ahills> was referring to locale speedup re: nsz
2015-01-12 19:13:50 <dalias> vkrishn, that would be nice to see
2015-01-12 19:13:56 <dalias> anyone who can make it happen?
2015-01-12 19:13:57 <ahills> shell scripts run notably faster on systems that run lots of them
2015-01-12 19:14:21 <m_jones> ahills: are you comparing ash to ash, or ash to dash or bash?
2015-01-12 19:14:21 <dalias> ahills, i suspect shell script performance is due to musl not having insane program-startup overhead
2015-01-12 19:14:36 <ahills> m_jones: both
2015-01-12 19:14:36 <m_jones> Debian uses dash
2015-01-12 19:14:38 <vkrishn> phorox as all setups
2015-01-12 19:14:43 <vkrishn> phoronix
2015-01-12 19:14:43 <dalias> (strace any of the standard utils often used in shell scripts)
2015-01-12 19:14:43 <ahills> m_jones: I should say, all three
2015-01-12 19:15:05 <ahills> individual commands aren't noticably faster on the command line, but my cron jobs finish quicker, etc
2015-01-12 19:15:32 <dalias> you're not going to notice the 500us difference in startup time for a single command
2015-01-12 19:15:32 <ahills> no longer consider writing tools in other languages for performance
2015-01-12 19:15:40 <m_jones> I always disliked bash. I've mellowed in recent years and converted to it, as part of my de-customisation of working setups
2015-01-12 19:15:43 <vkrishn> I think best not to tweak default setups, so ash for AL, bash for debian... likewise
2015-01-12 19:15:44 <dalias> you will notice it when ./configure runs 100k+ commands
2015-01-12 19:16:12 <dalias> debian uses bash for /bin/sh??
2015-01-12 19:16:16 <dalias> i thought they used dash
2015-01-12 19:16:20 <ahills> after using Alpine for a few machines, I tried using busybox on Debian/Ubuntu/RHEL, but it is not the source of the speedups I saw
2015-01-12 19:16:26 <m_jones> bash-dependency is so ingrained in Linux users that it's hard to get people to write portable scripts
2015-01-12 19:16:34 <m_jones> dalias: they use dash for a few years now
2015-01-12 19:16:48 <m_jones> but historically virtually all Linux used bash for /bin/sh since 1991.
2015-01-12 19:16:54 <dalias> if you're writing bash code it's simple: write #!/bin/bash
2015-01-12 19:17:00 <m_jones> I've personally been a hater since 1994.
2015-01-12 19:17:03 <dalias> not #!/bin/sh
2015-01-12 19:17:15 <dalias> bash is full of bugs/vulns and bloat
2015-01-12 19:17:17 <m_jones> dalias: vehemently agreed.
2015-01-12 19:17:18 <vkrishn> advantage of busybox is on once set, shut and forget servers
2015-01-12 19:17:52 <m_jones> it's hard to get people to shebang bash, and it's separately hard to get them to decide to target Bourne instead of bash
2015-01-12 19:17:59 <m_jones> even Google explicitly uses bash not sh.
2015-01-12 19:18:27 <ahills> don't trust large companies to use good tools
2015-01-12 19:18:33 <ahills> I've never seen one that does
2015-01-12 19:18:48 <ahills> Google also writes tons of Java code
2015-01-12 19:18:54 <m_jones> Debian and Ubuntu finally fixed one of the bigger problems I had with Linux for many years. (I used mostly Solaris, *BSD, and Alphas until around 2005).
2015-01-12 19:18:55 <ahills> and chose Java as their platform for Android
2015-01-12 19:19:28 <vkrishn> android,  ah I remember why I logged in here :)
2015-01-12 19:19:34 <m_jones> ahills: that application needs a bytecode platform for obvious portability reasons. Also, Bionic. See also: Dalvik vs. ART.
2015-01-12 19:19:46 <vkrishn> dalias , how does the new AL website look on cm ?
2015-01-12 19:19:52 <ahills> yes, but it doesn't need THAT platform
2015-01-12 19:19:54 <m_jones> android is a separate beast that also uses the Linux kernel
2015-01-12 19:19:55 <vkrishn> we tweaked more
2015-01-12 19:20:02 <dalias> vkrishn, the logo renders wrong on my version :(
2015-01-12 19:20:21 <vkrishn> then I have strong feeling about the brower that comes with cm
2015-01-12 19:20:37 <dalias> others couldn't reproduce the problem
2015-01-12 19:20:42 <dalias> it's probably a bug in the specific version i have
2015-01-12 19:20:46 <vkrishn> try downloading chrome
2015-01-12 19:21:00 <dalias> hmm cert if also rejected now...
2015-01-12 19:21:21 <m_jones> seems fine on my 5.x Chrome stocker.
2015-01-12 19:21:24 <dalias> is StartCom a legitimate CA ?
2015-01-12 19:21:52 <m_jones> yes
2015-01-12 19:22:00 <m_jones> I tested with iceweasel earlier
2015-01-12 19:22:19 <m_jones> I'm working with TLS, so let's take a look at AL.org
2015-01-12 19:22:20 <dalias> weird. my phone doesn't have it
2015-01-12 19:22:43 <dalias> logo issue has changed a bit...
2015-01-12 19:22:51 <m_jones> google is leading edge on tls and has been talking about stripping unused CAs
2015-01-12 19:23:03 <m_jones> but unless I misremember, startcom is pretty significant
2015-01-12 19:24:05 <m_jones> maybe I misremember w.r.t. startcom
2015-01-12 19:25:11 <dalias> imgur.com/WQShdWy
2015-01-12 19:26:09 <m_jones> The StartSSL certificate is included by default in Mozilla Firefox 2.x and higher, in Apple Mac OS X since version 10.5 (Leopard), all Microsoft operating systems since 24 September 2009,[7][8] and Opera since 27 July 2010.[9] Since Google Chrome, Apple Safari and Internet Explorer use the certificate store of the operating system, all major browsers include support for StartSSL certificates.
2015-01-12 19:26:27 <m_jones> startcom does the free ones. I was conflating with rapidSSL.
2015-01-12 19:26:38 <m_jones> however!
2015-01-12 19:27:05 <vkrishn> hmm.. issue with logo only now, rest seems lok
2015-01-12 19:27:14 <m_jones> * Connected to alpinelinux.org (88.159.20.184) port 443 (#0) * successfully set certificate verify locations: *   CAfile: none   CApath: /etc/ssl/certs * SSLv3, TLS handshake, Client hello (1): * SSLv3, TLS handshake, Server hello (2): * SSLv3, TLS handshake, CERT (11): * SSLv3, TLS alert, Server hello (2): * SSL certificate problem: unable to get local issuer certificate * Closing connection #0
2015-01-12 19:27:39 <m_jones> vkrishn: caching an issue at all?
2015-01-12 19:28:08 <jomat> StartCom wants money if you want to revoke your cert i. e. because of Heartbleed
2015-01-12 19:28:43 <dalias> ah startcom is startssl?
2015-01-12 19:29:03 <vkrishn> m_jones, caching ?
2015-01-12 19:29:06 <dalias> jomat, i really don't have much of a problem with that
2015-01-12 19:29:17 <jomat> Yes
2015-01-12 19:29:28 <m_jones> vkrishn: didn't know if you changed the logo.
2015-01-12 19:29:42 <dalias> it's better than making it impossible to have a secure site without paying
2015-01-12 19:29:45 <vkrishn> nope, I cannot, ncopa only
2015-01-12 19:29:57 <m_jones> * SSL connection using ECDHE-RSA-AES256-GCM-SHA384 <-- very nice; just need an ECC cert now ;)
2015-01-12 19:30:14 <m_jones> paying for cert revocation is pretty legit
2015-01-12 19:30:25 <m_jones> revocation via OCSP or CRL is resource-intensive
2015-01-12 19:30:29 <dalias> of course let's encrypt will fix the problem much better
2015-01-12 19:30:32 <vkrishn> the new site was made live couple of weeks back
2015-01-12 19:30:44 <m_jones> Check out Adam Langley/Google's stuff about it
2015-01-12 19:31:54 <m_jones> Let's Encrypt should be excellent when it arrives. I also have even higher hopes for DANE.
2015-01-12 19:32:32 <m_jones> Less need to trust a lot of _pro_forma_ state-sponsored and semi-competent CAs with DANE
2015-01-12 19:33:02 <m_jones> also, Certificate Transparency and related proposals help a lot except for the paying part, which Let's Encrypt should fix
2015-01-12 19:34:12 <dalias> yes, DANE is better but harder to get deployed well
2015-01-12 19:34:21 <kaniini> i do not like DANE because it's a trap for DNSSEC
2015-01-12 19:34:27 <dalias> kaniini, ?
2015-01-12 19:34:34 <kaniini> and DNSSEC is a trap for ensuring USA maintains control of the DNS
2015-01-12 19:34:41 <m_jones> I'm still working on this discard issue for trim/unmap, but on a TLS note: any talk of switching to libressl-portable?
2015-01-12 19:34:47 <kaniini> since ICANN controls the trust root
2015-01-12 19:36:11 <m_jones> kaniini: fair observation, but I don't distrust icann. there's overwhelming transparency on root zone signing. Do you have a scenario that would persuade me otherwise?
2015-01-12 19:37:00 <m_jones> DANE is fairly hard to make functional, yes. I'm working on something related at the moment.
2015-01-12 19:37:10 <melville> mhm about performance difference, it would be interesting to see a flame graphs for same application working under similar workload both on musl and glibc:http://www.brendangregg.com/FlameGraphs/cpuflamegraphs.html
2015-01-12 19:37:41 <melville> since it probably would show a bit of internals of both libs
2015-01-12 19:37:43 <m_jones> at the end of the day we should have competition between traditional CAs and DANE, and competition is rarely bad.
2015-01-12 19:38:26 <m_jones> melville: probably needs one of the new instrumentation interfaces to be practical. Gregg is not a fan of systemtap, and he's used it a lot more than me.
2015-01-12 19:39:03 <m_jones> ktap maybe; I don't know enough to have an opinion. Hopefully we get some concensus before long.
2015-01-12 19:40:43 <melville> seen some dtrace talks, where systemtap was trashed for possibility of crashing system, were you prolly want to profile production system - maybe that is part of what you relate to.
2015-01-12 19:41:53 <Diftraku> kaniini: "Trusting trust"
2015-01-12 19:42:58 <m_jones> I'm using the Qualys site to check the alpinelinux.org cert chaining because I don't seem to have the root CA in my Debian 7.7 /etc/ssh/certs, and the AL.org cert is signed by an intermediate
2015-01-12 19:43:11 <m_jones> it might be related to the fact that the AL.org cert is SHA2 (SHA384)
2015-01-12 19:44:17 <m_jones> melville: yes, we've probably seen some of the same talks by Gregg. I've also never personally had systemtap put a smile on my face, but that could be entirely on me.
2015-01-12 19:46:03 <m_jones> so alpine ships openssl but no certs...
2015-01-12 19:48:26 <dalias> you just have to install the certs package
2015-01-12 19:48:37 <dalias> apk add ca-certificates
2015-01-12 19:48:37 <m_jones> yes, done and done
2015-01-12 19:48:54 <dalias> :)
2015-01-12 19:49:18 <m_jones> had to mount media because that KVM guest is now having a networking problem on this setup
2015-01-12 19:50:00 <m_jones> I'm doing two things that used to be anathema: typing extensively on a laptop, and using 802.11 for semi-serious work
2015-01-12 19:50:20 <melville> yeah his talk also gave some insight into culture diffrences e.g. shipping unstriped binaries with full profiles and (if i recall) debug symbols on systems of solaris origin
2015-01-12 19:51:02 <m_jones> I used to be a big solaris fan but stop using it at its perigee of Solaris 9, right before they turned it around totally with Solaris 10
2015-01-12 19:53:21 <m_jones> 
ACTION plugs a mouse into the thinkpad and looks over at the boxed 4K screen...

2015-01-12 19:53:46 <ahills> I found Solaris 10 to be a huge pain in the ass... the ksh that ships with it is awful
2015-01-12 19:54:32 <m_jones> 
ACTION finds a Sun mousepad from 2001 and uses it ironically with a Microsoft optical mouse.

2015-01-12 19:54:44 <m_jones> ksh people are weird.
2015-01-12 19:54:52 <mitchty> i miss dtrace+zfs the most from solaris 10
2015-01-12 19:55:06 <m_jones> whenever someone professes to be a fan of ksh, I ask them if they spent formative years on AIX (my least-favourite Unix)
2015-01-12 19:55:07 <ahills> hmm, it was a rather long sentence...
2015-01-12 19:55:21 <ahills> I love the pdksh on OpenBSD; it's my favorite shell, hands-down
2015-01-12 19:55:43 <m_jones> I used to use tcsh, but de-customised to bash a number of years ago.
2015-01-12 19:56:05 <m_jones> Ran into a kid recently who switched to zsh after shellshock; I thought it was a bit odd.
2015-01-12 19:56:32 <ahills> It has fancier versions of bash features without a lot of the bash nonsense... what is odd?
2015-01-12 19:57:25 <m_jones> These MSFT mice are great, and I have them on all my Linux workstations. Also, middle scroll wheel I found to be compelling, to my surprise.
2015-01-12 19:58:07 <m_jones> ahills: I think it's heavier than bash, and features aren't always great if they require customisations.
2015-01-12 19:58:22 <m_jones> zsh was the native shell on the Isilons (FreeBSD based) that I ran until recently, though
2015-01-12 19:58:32 <ahills> people keep their .zshrc, .bashrc, etc on github now
2015-01-12 19:58:36 <ahills> so customize once, download forever
2015-01-12 19:58:53 <m_jones> I use so many machines that even keeping customizations in git didn't seem like the thing to do
2015-01-12 19:59:01 <m_jones> I might reconsider though
2015-01-12 19:59:20 <ahills> everyone has their own method... I build custom install images at work that put my preferences in /etc/profile >:D
2015-01-12 19:59:35 <m_jones> sensible defaults are terribly important. For example, I went to try out i3wm recently, only to find out that it's so customisable that there's essentially no defaults
2015-01-12 20:00:30 <ahills> it really depends on the software's scope and the user's purpose
2015-01-12 20:00:53 <m_jones> customisations don't help when you're on someone else's machine, or in restricted environmnts. I'll give it some new consideration, though.
2015-01-12 20:01:03 <ahills> so much software has defaults I don't want that I designed a system for including my default configurations and keeping them up-to-date
2015-01-12 20:01:40 <m_jones> bad customisations are bad. Like, imo, 'll' for ls -la is ubiquitous on Linux
2015-01-12 20:02:19 <m_jones> but i3 has no default for the super/meta key, and then no defaults for anything else either. So there's no chance of coming up to speed in 3 minutes, or of having someone show you something new in 2 minutes.
2015-01-12 20:04:23 <m_jones> https://www.ssllabs.com/ssltest/analyze.html?d=alpinelinux.org
2015-01-12 20:04:33 <m_jones> I was wrong, that isn't a SHA2 cert
2015-01-12 20:05:22 <m_jones> and the cert chain isn't complete, as I suspected
2015-01-12 20:05:32 <m_jones> that's probably the only problem with the cert
2015-01-12 20:07:22 <m_jones> non-browser clients, and I bet some browsers, don't automatically download the intermediates
2015-01-12 20:17:04 <m_jones> damnit, busybox find accepts -iname but treats it like -name, instead of case-insensitive
2015-01-12 20:17:31 <m_jones> A silent fail is full of fail.
2015-01-12 20:18:34 <dalias> that should be fixed in next musl release. it's partly our fault for defining FNM_CASEFOLD but not supporting that nonstandard extension
2015-01-12 20:18:43 <dalias> but it's stupid of busybox to be using it too :)
2015-01-12 20:19:01 <m_jones> dalias: awesome; I was checking busybox in preparation for filing a ticket
2015-01-12 20:19:39 <m_jones> I can't easily install findutils until I figure out my new networking problem
2015-01-12 20:19:59 <m_jones> KVM hands out DHCP, but does it bridge or nat?
2015-01-12 20:20:31 <m_jones> It ust bridge with -user, which means my uplink isn't NATing any subnets beside the one is hands out on 802.11n
2015-01-12 20:20:53 <m_jones> this is why I used to be  wireless hater. Well, that and being a snob.
2015-01-12 20:22:51 <m_jones> actually it would be qemu not kvm I guess.
2015-01-12 21:01:52 <m_jones> I think my KVM guest networking stopped working very close to them time when upgraded from 3.1.0 to 3.1.1
2015-01-12 21:02:10 <m_jones> there's no udev, so the device and MAC shouldn't have changed....hmm
2015-01-12 21:18:53 <m_jones> is there an apk command to list installed packages, like dpkg -l or rpm -qa?
2015-01-12 21:36:09 <buckley310> "apk info"
2015-01-12 21:55:01 <m_jones> buckley310: thanks, I thought that 'info' only worked with an argument to see a specific package in the cache.
2015-01-12 21:55:30 <buckley310> mhm
2015-01-12 21:56:02 <buckley310> you can also do "cat /etc/apk/world" to see all packages without dependants
2015-01-12 21:56:31 <buckley310> not sure if that's the "correct" way
2015-01-12 21:57:10 <clandmeter> correct enough for me.
2015-01-12 21:59:59 <opethfan> dalias, I'm encountering / filed a bug identical to this issue you responded to in 2013 regarding X module dependencies and lazy binding. I'm curious if it's an issue that's widespread or if it only affects my use case (or if it's the X server's issue and out of your hands) http://www.openwall.com/lists/musl/2013/05/17/6
2015-01-12 22:02:48 <dalias> it's not fixed upstream. alpine fixes it with a conf file that force-loads modules in dependency order
2015-01-12 22:03:44 <opethfan> upstream in musl?
2015-01-12 22:03:51 <nsz> xorg
2015-01-12 22:04:25 <opethfan> Ah, it would be those guys :P thanks
2015-01-12 22:04:44 <opethfan> I'll investigate why that conf file isn't being loaded automatically for me, then
2015-01-12 22:17:01 <dalias> there is a possible workaround i could make in musl, but i'd rather not
2015-01-12 22:17:29 <dalias> basically we'd be emulating lazy binding by deferring binding to subsequent dlopen calls instead of to call-time
2015-01-12 22:23:41 <m_jones> I only want to see aded packages though -- does bare 'info' show me world?
2015-01-12 22:23:45 <m_jones> s/aded/added/
2015-01-12 22:28:21 <m_jones> oh, I guess 'world' is locally installed userland then
2015-01-12 22:28:29 <m_jones> I was thinking that meant thr whole repo.
2015-01-12 22:30:17 <opethfan> if it's an Xorg issue, it'd likely be better fixed there rather than making unneeded changes to musl
2015-01-12 22:36:04 <dalias> xorg is relying on RTLD_LAZY actually precluding the checking of undefined symbols at load time, so that they don't cause the load to fail
2015-01-12 22:36:28 <dalias> this is outside the specified behavior. the specified behavior is just that it allows, but does not require, the implementation to resolve them lazily
2015-01-12 22:38:11 <opethfan> But because glibc resolves them lazily, there hasn't been effort upstream by xorg devs to change the behavior, i'd assume
2015-01-12 22:39:41 <dalias> right
2015-01-12 22:39:55 <dalias> basically they want to do their own dependency tracking and manually dlopen deps rather than using DT_NEEDED
2015-01-12 22:40:43 <dalias> but this prevents dlopen from being able to resolve all the relocations when it needs to
2015-01-12 22:43:44 <opethfan> x never ceases to cause me grief
2015-01-12 22:48:42 <dalias> *nod*
2015-01-12 22:50:18 <dalias> it's very intentional that musl doesn't do lazy binding at call time -- it's generally a pessimization, it's a security flaw (incompatible with relro protection), it requires tons of arch-specific fragile code that often breaks with new vector instructions/registers being added, etc.
2015-01-12 22:50:45 <opethfan> Thanks for the insight. I know gui stuff isn't high on the list for most alpine rollouts, so I'll do my best to figure out a workable solution
2015-01-12 22:50:58 <Diftraku> x needs to die, imo
2015-01-12 22:51:31 <dalias> however we could do fake lazy binding at subsequent dlopen time -- the idea is that we would save the list of as-yet-unresolved relocations in modules loaded RTLD_LAZY, and re-process them after each subsequent dlopen using newly-available libs
2015-01-12 22:52:20 <dalias> opethfan, fwiw i have X working fine on alpine, out of the box
2015-01-12 22:52:57 <dalias> diftraku, those who don't understand X are doomed to reimplement it, poorly :)@musllibc
2015-01-12 22:53:04 <dalias> oops paste error at EOL
2015-01-12 22:53:13 <Diftraku> :3
2015-01-12 22:53:18 <Diftraku> wayland!
2015-01-12 22:53:21 <opethfan> that's strange, then. maybe I'm missing an apk?
2015-01-12 22:53:21 <dalias> lol
2015-01-12 22:53:36 <dalias> wayland is exactly what i was talking about
2015-01-12 22:54:00 <opethfan> you don't feel wayland is being implemented correctly?
2015-01-12 22:54:13 <dalias> nope
2015-01-12 22:54:34 <Diftraku> we need daliasd!
2015-01-12 22:54:52 <Diftraku> x that is not x
2015-01-12 22:55:20 <dalias> the right solution to Xorg sucking is a reimplementation of X based on the spec and the useful features/extensions, possibly with the rest omitted or implemented in inefficient-but-trivial ways
2015-01-12 22:55:22 <ScrumpyJack> Xdalias
2015-01-12 22:56:03 <dalias> throwing out the networked and well-factored designed of X is stupid
2015-01-12 22:56:20 <opethfan> I was under the impression that wayland was developed by X devs, but is it the compositors that are troublesome?
2015-01-12 22:57:06 <dalias> and wayland did that
2015-01-12 22:57:15 <dalias> first stupid idea was putting the window manager and display server together
2015-01-12 22:57:42 <dalias> then when they realized people wouldn't like this, they said hey let's split it up with shared libraries
2015-01-12 22:57:44 <opethfan> the idea of duplicating a lot of the work for each wm seems silly to me
2015-01-12 22:58:04 <dalias> which is the idioitic windows-esque way to do something
2015-01-12 22:58:19 <dalias> the sane way is to use separate processes which communicate over a clean, well-defined protocol
2015-01-12 22:58:22 <opethfan> whoops, my auto scroll isn't working, i'm saying things the smart people have already said :P
2015-01-12 22:58:25 <dalias> which would be..... oops, X
2015-01-12 23:01:14 <ScrumpyJack> dbus and however dbus will be assimilated into systemd
2015-01-12 23:37:00 <systmkor> ScrumpyJack, hey
2015-01-13 05:12:11 <mdekkers> hi all, new to alpine, trying to install nginx come up with a weird error: alpine:~# apk add nginx-1.6.2-r1 ERROR: unsatisfiable constraints:   nginx-1.6.2-r1 (missing):     required by: world[nginx-1.6.2-r1]
2015-01-13 05:12:30 <mdekkers> just running apk add nginx just returns nothing
2015-01-13 12:40:25 <vkrishn> anyone tried sending file from mobile -> AL server via bluethooth ?
2015-01-13 13:29:10 <ScrumpyJack> i need to get apache 2.0 working
2015-01-13 13:29:53 <ScrumpyJack> hopefully, Alpine Linux goes back that far :)
2015-01-13 13:30:32 <ScrumpyJack> nope AL2.0 has apache 2.2
2015-01-13 15:10:59 <openfbtd> Hehe. X in an Alpine chroot
2015-01-13 15:39:37 <ScrumpyJack> xen_roger: your a xen user :) what do you use as a "backing store"
2015-01-13 15:45:48 <xen_roger> assumed I use xen-live ?
2015-01-13 15:45:55 <BitL0G1c> using lvm logical volumes as the disks for guests will give the best i/o performance on xen
2015-01-13 15:46:31 <xen_roger> ScrumpyJack: atm I use lvm on luks
2015-01-13 15:50:45 <BitL0G1c> xen_roger: do you use openvswitch for the bridge ?
2015-01-13 15:51:19 <xen_roger> BitL0G1c: yes, sir
2015-01-13 15:52:03 <BitL0G1c> :) ok
2015-01-13 15:52:05 <xen_roger> BitL0G1c: also with multiple physical devices attacht to multiple switches ... works really nice
2015-01-13 15:53:16 <BitL0G1c> I use openvswitch for LXC guests in a KVM host (works great)
2015-01-13 15:53:53 <xen_roger> use it also for lxc ...
2015-01-13 15:55:01 <BitL0G1c> have you ever bridged wlan0 ?
2015-01-13 15:56:47 <xen_roger> no... no business case ;)  so never tried ... but could work ... tell me if you have success with it
2015-01-13 16:06:09 <BitL0G1c> xen_roger: does XEN work inside vmware ? (not worked for me so far)
2015-01-13 16:06:39 <xen_roger> wich vm? kvm?
2015-01-13 16:07:08 <BitL0G1c> vmware workstation 11
2015-01-13 16:07:49 <BitL0G1c> booting the xen kernel seems to hang during hardware initialization
2015-01-13 16:08:38 <xen_roger> nested virt is needed ... maby workstation does not support it... try kvm and have a look at  "modinfo kvm_intel | grep -i nested"
2015-01-13 16:09:21 <xen_roger> if not build as module check "cat /sys/module/kvm_intel/parameters/nested" should be Y to work
2015-01-13 16:09:39 <ScrumpyJack> xen_roger: i meant how do you carve up the fs for each vm?
2015-01-13 16:09:50 <ScrumpyJack> btrfs, lvm etc
2015-01-13 16:09:57 <ScrumpyJack> (if at all :)
2015-01-13 16:11:31 <xen_roger> ScrumpyJack: lvm
2015-01-13 16:20:52 <ScrumpyJack> do you use snapshots?
2015-01-13 16:21:23 <xen_roger> not atm ... there is no need currently
2015-01-13 16:21:52 <darkfader> btw if you're building new xen hosts, be sure to raise the lvm metadata size from the default 255... i didn't know that and hit the limit when adding lvm mirrors
2015-01-13 16:22:55 <ScrumpyJack> just wondering if it's better to :-
2015-01-13 16:23:39 <ScrumpyJack> a) mount the snapshot and boot the vm, do stuff and delete the snapshot when finished or
2015-01-13 16:24:50 <ScrumpyJack> b) mount the original lvm dev and boot the vm, do stuff and merge the (empty) snapshot onto the lv
2015-01-13 16:25:30 <ScrumpyJack> (in both cases to revert to original data on the volume)
2015-01-13 16:26:05 <darkfader> like the first better since it's easier to add more snapshots like that (and retry)
2015-01-13 16:26:27 <darkfader> but maybe i'm too tired
2015-01-13 17:21:49 <xen_roger> Don't know who asks about "XEN in VM" - but I did check it with current alpine-xen on KVM an it works ( 2gb ram - alpine-xen live - 4 cores )
2015-01-13 17:27:55 <BitL0G1c> ok will try again
2015-01-13 17:28:33 <BitL0G1c> I was trying xen in kvm in vmware workstation (I have AMD-V)
2015-01-13 17:41:12 <xen_roger> crazzyyy ;)
2015-01-13 18:51:45 <melville> turtles all the way down
2015-01-13 18:59:54 <z3bra> Hi !
2015-01-13 19:00:05 <z3bra> How goes ?
2015-01-14 11:51:12 <alias_neo> Has anyone got experience running alpine as KVM host? I'm having some issues. It mostly runs fine, but I get an error when launching libvirt, and virt-manager won't manage my network interfaces,
2015-01-14 11:52:21 <alias_neo> the interface error is: this function is not supported by the connection driver: virConnectNumOfInterfaces
2015-01-14 11:53:09 <alias_neo> the other error on start, not sure if a problem or not, but I hate errors non the less: "error : virDriverLoadModule:73 : failed to load module /usr/lib/libvirt/connection-driver/libvirt_driver_interface.so Error relocating /usr/lib/libvirt/connection-driver/libvirt_driver_interface.so: udev_enumerate_unref: symbol not found" Couldn't find anything about it on the web
2015-01-14 12:02:29 <alias_neo> I also get an error starting virt-manager if I try to do it with X forwarding (runs on the alpine box): (virt-manager:4419): WARNING **: Couldn't connect to accessibility bus: Failed to connect to socket /tmp/dbus-IG4S0KQl6N: Connection refuse
2015-01-14 12:02:52 <alias_neo> followed by a dialog with module has no attribute: VIR_DOMAIN_BLOCKED
2015-01-14 12:03:08 <alias_neo> So I can't seem to run virt-manager reliable either locally or remotely
2015-01-14 12:07:05 <ncopa> alias_neo: i use libvirt on my work desktop
2015-01-14 12:08:15 <alias_neo> ncopa: Is that under Alpine?
2015-01-14 12:08:17 <ncopa> yes
2015-01-14 12:08:29 <alias_neo> And everything works?
2015-01-14 12:08:31 <ncopa> libvirt + virt-manager
2015-01-14 12:08:37 <ncopa> as far i can see yes
2015-01-14 12:08:46 <alias_neo> Perhaps I missed some module or package?
2015-01-14 12:08:56 <ncopa> i think you might need dbus
2015-01-14 12:09:10 <alias_neo> I coulodn't find instructions anyway, I managed to get kvm modules loaded, udev, dbus, dnsmasq installed
2015-01-14 12:09:39 <alias_neo> dbus is runnig
2015-01-14 12:09:49 <alias_neo> I haven't changed any libvirt or libvirtd configs from default
2015-01-14 12:09:57 <ncopa>  Failed to connect to socket /tmp/dbus-IG4S0KQl6N: Connection refuse
2015-01-14 12:10:05 <ncopa> what is the permission on /tmp ?
2015-01-14 12:10:18 <alias_neo> drwxrwxrwt    4 root     root          4096 Jan 14 09:34 tmp
2015-01-14 12:10:28 <ncopa> looks correct
2015-01-14 12:10:35 <alias_neo> I just noticed root isn't a member of group libvirt, could that be an issue?
2015-01-14 12:10:43 <ncopa> shouldnt matter
2015-01-14 12:11:08 <ncopa> my ncopa user is member of netdev readproc kvm
2015-01-14 12:11:24 <alias_neo> root isn't a member of kvm on my box
2015-01-14 12:11:32 <ncopa> audio video and cdrom too
2015-01-14 12:11:42 <alias_neo> root bin daemon sys adm disk wheel floppy dialout tape video
2015-01-14 12:11:50 <ncopa> same here
2015-01-14 12:11:52 <ncopa> shouldnt matter
2015-01-14 12:12:20 <ncopa> i think i maybe defined my networks with virsh
2015-01-14 12:12:32 <ncopa> oh
2015-01-14 12:12:49 <ncopa> to be able to connect with virt-manager to locally running libvirt
2015-01-14 12:13:05 <ncopa> i think consolekit and polkit is needed
2015-01-14 12:13:16 <ncopa> + some polkit agent
2015-01-14 12:13:29 <ncopa> like polkit-gnome
2015-01-14 12:14:19 <alias_neo> Installed those, still get the VIR_DOMAIN_BLOCKED attributen ot found issue
2015-01-14 12:14:33 <alias_neo> also says it failed to load canberra gtk module, even though I installed it
2015-01-14 12:17:49 <alias_neo> Would love to get this fixed, got a dedicated box I'd like to run Alpine as KVM host
2015-01-14 12:32:25 <alias_neo> ncopa: THe VIR_DOMAIN_BLOCKED error seems to indicate that the libvirt python module is missing the attribute VIR_DOMAIN_BLOCKED
2015-01-14 12:34:23 <ncopa> when do you see that error?
2015-01-14 12:34:28 <ncopa> when starting virt-manager?
2015-01-14 15:10:14 <alias_neo> ncopa: yes, when I start virt-manager, I see it in a local dialog (I'm forwarding over ssh),
2015-01-14 15:11:20 <ncopa> i have only use it locally
2015-01-14 15:11:41 <ncopa> i think you can connect to a remote libvirt over ssh:// or similar too
2015-01-14 15:12:32 <alias_neo> That's what I normally do, but the network interface page says the connection doens't support managing network interfaces
2015-01-14 15:15:14 <alias_neo> If I can get that fixed, it'll be fully working, and I can write up a guide or something. But I need something fully functional as my KVM host with virt-manager. It all worked flawlessly under Ubuntu Server 14.04 but I'd rather run Alpine if I can make it fully functional
2015-01-14 15:19:22 <alias_neo> My server is headless sadly so I can't test it locally (no gpu)
2015-01-14 15:57:35 <ncopa> i dont know what is wrong
2015-01-14 15:59:09 <alias_neo> ncopa: no prob, thanks. I'll debug it a bit more, otherwise I'll have to go back to Ubuntu crying :(
2015-01-14 15:59:28 <alias_neo> Maybe somebody else will be around that could help
2015-01-14 16:01:29 <ncopa> i would expect it to work that you run libvirt on the headless host, without X11 and virt-manager, then run virt-manager from a desktop
2015-01-14 16:01:44 <ncopa> and connet to ssh+qemu:// or what they call it
2015-01-14 16:11:55 <alias_neo> Well, remote connection works (as in, virt-manager runs on my desktop), and I can manage the machines mostly, but it's missing the capability to manage network devices. But trying to forward the server's birt-manager over ssh with X forwarding just doesn't work
2015-01-14 16:12:37 <alias_neo> from hours of reading last night i believe the network management was an issue of missing tun/tap module, but i've got that installed and loaded and still no dice
2015-01-14 16:14:32 <ScrumpyJack> ncopa: does chroot work for you in 3.1.1?
2015-01-14 16:18:21 <ScrumpyJack> ignore that
2015-01-14 16:25:37 <ScrumpyJack> lxc-debian template doesn't work
2015-01-14 17:40:35 <alias_neo> ncopa: the VIR_DOMAIN_BLOCKED issue seems to be resolved by commenting out the entire section vm_status_icons = { on line 41 of /usr/share/virt-manager/virtManager/domain.py it seems those libvirt.X don't exxist in the version of libvirt-python that I have. THe virt-manager starts with those commented out, but breaks over something else that's missing later on when creating connections
2015-01-14 19:10:41 <mlaine> hi
2015-01-14 19:11:10 <mlaine> is there a preferred way to do bonding/trunking on alpine?
2015-01-14 19:11:26 <mlaine> to have an auto failover interface
2015-01-14 19:15:42 <mlaine> oh, found alpine wiki and bonding :)
2015-01-14 19:31:08 <systmkor> mlaine, what's your project?
2015-01-14 19:31:38 <mlaine> systmkor: testing!
2015-01-14 19:31:52 <systmkor> oh cool
2015-01-14 19:31:59 <systmkor> is it routing?
2015-01-14 19:32:05 <systmkor> or firewall stuff?
2015-01-14 19:32:20 <mlaine> routing
2015-01-14 20:08:19 <systmkor> oh cool
2015-01-14 20:08:29 <systmkor> please document on alpine wiki :D
2015-01-14 20:21:14 <systmkor> mlaine, one of my side projects is creating a package for mininet
2015-01-14 20:21:26 <systmkor> mlaine, http://mininet.org/
2015-01-14 20:21:59 <systmkor> mlaine, and hopefully get a guide or some code to easily map a topology from mininet to lxc/docker containers or VMs
2015-01-14 20:22:34 <mlaine> hmm, looks interesting
2015-01-14 20:26:02 <systmkor> mlaine, it would also be night to get a fair amount of the OpenDaylight distro packages
2015-01-14 20:26:11 <systmkor> ported to Alpine
2015-01-14 20:27:34 <systmkor> mlaine, I also am curious how hard it would be to port https://qubes-os.org/
2015-01-14 20:27:35 <systmkor> with Alpine
2015-01-14 20:30:10 <darkfader> systmkor: what do you mean with alpine?
2015-01-14 20:30:49 <darkfader> alpine in the dom0? or all pieces that are fedora to be alpine?
2015-01-14 20:31:09 <systmkor> well make sure the alpine dom0 xen version works as the hypervisore
2015-01-14 20:31:17 <systmkor> and all the pieces that are fedora to alpine
2015-01-14 20:31:35 <systmkor> so Pax/Grsec/MUSL benefits
2015-01-14 20:31:41 <darkfader> ok
2015-01-14 20:31:43 <systmkor> rather than SELinux & glibc
2015-01-14 20:31:50 <darkfader> i'd say, 6 months of work :)
2015-01-14 20:31:55 <systmkor> yah probably
2015-01-14 20:32:14 <darkfader> i sent out qubes dvds to all my customers for christmas :)
2015-01-14 20:32:29 <systmkor> darkfader, nice :D
2015-01-14 20:32:41 <darkfader> and i *sigh* need to buy a new desktop that will run it since mine won't finish the installer
2015-01-14 20:32:44 <darkfader> fedora...
2015-01-14 20:32:51 <systmkor> yuuuuuup
2015-01-14 20:33:04 <systmkor> fuck dat shit
2015-01-14 20:33:11 <darkfader> i'll try the new intel NUC when it comes out (soon)
2015-01-14 20:33:15 <darkfader> i think it should work
2015-01-14 20:33:51 <darkfader> i see your point, putting in a more robust OS for the heavy lifting would help
2015-01-14 20:34:07 <darkfader> i'd probably say the desktop on fedora is "what users like"
2015-01-14 20:34:16 <systmkor> also Alpine makes it way easier to strip down
2015-01-14 20:35:02 <systmkor> certain parts I miss using Ubuntu/Fedora and such for more comprehensive easy install processes
2015-01-14 20:35:47 <systmkor> darkfader, that said gradm needs to work out of the box with Alpine which it kinda doesn't
2015-01-14 20:36:15 <systmkor> also I think having a per package & extra database of grsec policy
2015-01-14 20:36:20 <systmkor> would help immensily
2015-01-14 20:36:20 <darkfader> i forgot to find out if qubes also uses xen security extensions of some sort
2015-01-14 20:36:29 <darkfader> pretty sure noone had them in use on alpine so far
2015-01-14 20:36:42 <systmkor> darkfader, I don't know, probably can check in their spec 0.3 paper
2015-01-14 20:36:58 <darkfader> i'll dig for it and read it on the flight home from vacation
2015-01-14 20:37:08 <darkfader> as for gradm, yes, you're right
2015-01-14 20:37:25 <systmkor> darkfader, http://files.qubes-os.org/files/doc/arch-spec-0.3.pdf
2015-01-14 20:37:28 <darkfader> i remember someone recently said like "why, it works for me", i didn't poke myself
2015-01-14 20:37:31 <darkfader> heh, ty
2015-01-14 20:38:02 <systmkor> yah I don't know how it would work best but having like
2015-01-14 20:38:06 <systmkor> vsftpd-gpol
2015-01-14 20:38:20 <systmkor> be a standard policy  for vsftpd setups
2015-01-14 20:38:27 <systmkor> would help immensily
2015-01-14 20:38:41 <systmkor> even if it is a lowest common denominator w/ examples to add
2015-01-14 20:38:49 <systmkor> would help me building and modify new systems
2015-01-14 20:39:36 <systmkor> darkfader, also with having their network domain being segmented
2015-01-14 20:39:49 <systmkor> it would be cool to see how hard to drop in other options with that
2015-01-14 20:39:58 <systmkor> e.g. OpenVSwitch & OpenFlow
2015-01-14 20:40:08 <systmkor> or maybe OpenBSD's PF stack
2015-01-14 20:40:20 <systmkor> also if running on top of Xen
2015-01-14 20:40:42 <systmkor> darkfader, you can run Mirage OS with the host system
2015-01-14 20:40:43 <systmkor> http://www.openmirage.org/
2015-01-14 20:41:06 <darkfader> yeah i know
2015-01-14 20:41:13 <systmkor> which is basically a suite of OpenCaml libraries for building unikernels & services
2015-01-14 20:41:19 <systmkor> that would be such a fucking sweet system
2015-01-14 20:41:26 <darkfader> i'm not sure if everything will come down in pieces if you try to combine too many advanced pieces
2015-01-14 20:41:26 <systmkor> so much varied security options
2015-01-14 20:41:43 <systmkor> darkfader, all at once in one big sprint
2015-01-14 20:41:46 <systmkor> ay *boom*
2015-01-14 20:41:48 <darkfader> or rather, i've done stuff like that once. and the whole thing gained like two users
2015-01-14 20:41:51 <systmkor> but adding them in one by one
2015-01-14 20:41:57 <systmkor> and making sure things integrate upstream
2015-01-14 20:42:27 <systmkor> then I think it's duable
2015-01-14 20:42:49 <systmkor> darkfader, gained two users?
2015-01-14 20:43:07 <systmkor> darkfader, oh you mean like a distro only had two users?
2015-01-14 20:43:10 <darkfader> yeah
2015-01-14 20:43:14 <systmkor> ahh
2015-01-14 20:43:25 <systmkor> yah that's why I wouldn't want to fork
2015-01-14 20:43:37 <systmkor> I think it it's within the scope of Alpine
2015-01-14 20:43:41 <darkfader> might help, yes
2015-01-14 20:43:43 <systmkor> and just keep integrating back in
2015-01-14 20:43:54 <systmkor> darkfader, yah maybe helps short term
2015-01-14 20:43:59 <systmkor> but long term I ddon't think so
2015-01-14 20:47:00 <darkfader> what sets qubes apart is how much documentation and end user friendly utils they have
2015-01-14 20:47:16 <darkfader> i.e. name any other distro that has a command to directly mount an iso to a xen vm
2015-01-14 20:47:34 <darkfader> the user friendly bling bling and a lot of blog posts might really help
2015-01-14 20:47:48 <systmkor> yup
2015-01-14 20:47:58 <systmkor> longterm goals
2015-01-14 20:48:27 <darkfader> i also don't mean to discourage you, i'd even try to help. just meaning you need to focus on rallying people, not on any technical aspect :)
2015-01-14 20:48:33 <systmkor> just wish could find more people with the security interest of OpenBSD but focused on Linux and minus the elitism
2015-01-14 20:48:39 <darkfader> hehe
2015-01-14 20:48:47 <systmkor> darkfader, agreed
2015-01-14 20:49:49 <systmkor> darkfader, know any way to do that?
2015-01-14 20:50:09 <systmkor> I am slowly learning on disto construction and trying to contribute more to Alpine
2015-01-14 20:50:30 <darkfader> no, i'm the wrong person to ask about starting communities
2015-01-14 20:51:07 <darkfader> this could be intersting: http://lwn.net/Articles/120584/
2015-01-14 20:51:12 <darkfader> i read it many years ago
2015-01-14 20:51:51 <darkfader> but idk, ask ncopa here for a start!
2015-01-14 20:53:02 <systmkor> darkfader, also started a go port of the Nix package manager
2015-01-14 20:54:13 <systmkor> seems like Nix package/config management connected with a configuration management server
2015-01-14 20:54:19 <systmkor> like Ansible
2015-01-14 20:54:22 <systmkor> much dev so OPs
2015-01-14 20:54:38 <systmkor> :P
2015-01-14 20:57:52 <systmkor> darkfader, also i think Alpine needs to follow suite of Debian and utilize the gitian project style of
2015-01-14 20:58:02 <systmkor> actually verifiable builds
2015-01-14 21:01:32 <darkfader> i just know debian has this "doser" thing for testing builds, but what are verifyable builds?
2015-01-14 21:03:19 <systmkor> darkfader, presentation https://www.youtube.com/watch?v=5pAen7beYNc
2015-01-14 21:03:30 <darkfader> thx
2015-01-14 21:03:30 <systmkor> darkfader, website https://gitian.org/
2015-01-14 21:04:05 <darkfader> oh, i didn't go to that talk and pretty much forgot it
2015-01-14 21:04:13 <darkfader> but i think i fetched it at home hehe
2015-01-14 21:04:50 <darkfader> nice
2015-01-14 21:05:06 <systmkor> darkfader, https://dl.acm.org/citation.cfm?id=358210
2015-01-14 21:09:17 <systmkor> darkfader, to me it seems like distrobutions are commonly to focused on end-user experience
2015-01-14 21:09:28 <systmkor> and miss building a really guide base linux system
2015-01-14 21:09:58 <darkfader> a good system that isn't good to use is a shitty systems
2015-01-14 21:10:04 <darkfader> it needs to be both
2015-01-14 21:10:11 <systmkor> well yah
2015-01-14 21:10:20 <systmkor> but what I mean is basically
2015-01-14 21:10:34 <darkfader> what it doesn't need is shit like dbus that is "fake" user friendly
2015-01-14 21:10:37 <systmkor> i think there are at least to focuses of what we consider a "distro"
2015-01-14 21:10:56 <systmkor> one side focuses on providing packages
2015-01-14 21:11:02 <systmkor> a great  build system
2015-01-14 21:11:11 <systmkor> and all the low level things that Alpine does
2015-01-14 21:11:19 <systmkor> and similar that of gentoo and arch
2015-01-14 21:11:30 <systmkor> then their is the ElementaryOS/Ubuntu
2015-01-14 21:12:10 <systmkor> which focuses more on the end user experience ( granted that's more for the lay person )
2015-01-14 21:12:48 <systmkor> I think there should be good level of seperation between e.g. Alpine as a base system
2015-01-14 21:13:05 <systmkor> and a MATE desktop distro that makes it as easy to use as maybe linux mint for the average person
2015-01-14 21:13:15 <systmkor> but without doing the bullshit hacks that come with Ubuntu and such
2015-01-14 21:13:28 <systmkor> were their focus isn't security or how solid the system is
2015-01-14 21:13:38 <mlaine> is there any devel pkg or something that has the common tools for c compiling etc
2015-01-14 21:13:43 <systmkor> no real proof to this but just kinda rant thought
2015-01-14 21:13:53 <systmkor> for Alpine?
2015-01-14 21:13:56 <mlaine> yes
2015-01-14 21:13:58 <darkfader> mlaine: yes
2015-01-14 21:14:05 <systmkor> should be gcc and binutils
2015-01-14 21:14:07 <systmkor> right?
2015-01-14 21:14:13 <darkfader> systmkor: there's a bigger one, sec
2015-01-14 21:14:39 <darkfader> alpine-sdk
2015-01-14 21:14:44 <darkfader> has all you should need
2015-01-14 21:15:12 <mlaine> that's the one
2015-01-14 21:15:13 <mlaine> cheers
2015-01-14 21:16:37 <systmkor> yah
2015-01-14 21:17:11 <systmkor> mlaine, darkfader idk what do you guys think?
2015-01-14 21:19:38 <mlaine> i guess there's some need for yet another distro
2015-01-14 21:19:58 <mlaine> that provides desktops etc
2015-01-14 21:21:09 <systmkor> mlaine, i wouldn't say another distro
2015-01-14 21:21:24 <systmkor> just a segmentation/abstraction between what a "distro" is
2015-01-14 21:21:37 <mlaine> yeah that's what i meant
2015-01-14 21:22:10 <systmkor> mlaine, so really I guess it would be for example providing tools to an existing distro
2015-01-14 21:22:14 <systmkor> like ElementaryOS
2015-01-14 21:22:22 <systmkor> such as a base of Alpine
2015-01-14 21:22:25 <mlaine> alpine cannot be considered a desktop distro atm
2015-01-14 21:22:28 <mlaine> so yeah
2015-01-14 21:22:36 <systmkor> oh and I don't think it really should
2015-01-14 21:22:55 <systmkor> but I could imagine a group that basically operates on top of Alpine that produces a Desktop Distro
2015-01-14 21:23:06 <mlaine> the seperation between base and something else should be very clear
2015-01-14 21:23:42 <systmkor> mlaine, plus then you would only except upstream custom tools such as an install script or maybe
2015-01-14 21:23:50 <systmkor> an "install language" just as hypothetical
2015-01-14 21:23:56 <systmkor> that doesn't bloat the base system
2015-01-14 21:24:14 <systmkor> but makes it way easier for different higher level distros work faster and cleaner
2015-01-14 21:31:04 <mlaine> yay installed alpine on my laptop
2015-01-14 21:31:21 <systmkor> mlaine, nice
2015-01-14 21:31:30 <systmkor> I've been running it off my laptop for a couple months now
2015-01-14 21:31:37 <mlaine> networking and X with my favourite wm, dwm configured
2015-01-14 21:31:43 <systmkor> it's really nice except for getting that whole media experience from the internet
2015-01-14 21:31:45 <mlaine> zero problems so far :)
2015-01-14 21:31:57 <systmkor> well that's me
2015-01-14 21:33:09 <systmkor> mlaine, :thumbs_up:
2015-01-14 21:33:49 <mlaine> compiling dwb (web browser) might not be so easy
2015-01-14 21:34:49 <mlaine> i've had troubles with it on many distros and on bsd too
2015-01-14 21:35:18 <systmkor> mlaine, also don't even try using TPE with Grsec to run any web-browser
2015-01-14 21:35:24 <systmkor> that shit will just blow up
2015-01-14 21:36:16 <mlaine> :D
2015-01-14 21:43:31 <systmkor> mlaine, it would be nice to push program development to addhear to TPE
2015-01-14 21:43:47 <systmkor> but that would need some serious power to do that
2015-01-14 21:45:28 <systmkor> also the tools to aid developers comply with that aren't there
2015-01-14 21:55:29 <mlaine> hmm when i quit dwm i get only a blank screen
2015-01-14 21:56:52 <mlaine> also cannot change to another tty
2015-01-14 22:00:11 <systmkor> weird the ctrl+alt+F<num>
2015-01-14 22:00:12 <systmkor> works for me
2015-01-14 22:00:53 <mlaine> when i startx (exec dwm) i get X on tty7
2015-01-14 22:01:20 <mlaine> and all other ttys just blank
2015-01-14 22:01:40 <systmkor> try hitting return?
2015-01-14 22:01:56 <mlaine> won't respond to anything
2015-01-14 22:02:02 <systmkor> idk sorry
2015-01-14 22:02:10 <mlaine> i hate debugging these
2015-01-14 22:02:14 <mlaine> :D
2015-01-14 22:12:47 <BitL0G1c> for TPE - just put yourself in the trusted ID group - saves a lot of hassle
2015-01-14 22:16:29 <systmkor> BitL0G1c, i do that
2015-01-14 22:16:35 <systmkor> but it feels like a hack
2015-01-14 22:19:15 <BitL0G1c> otherwise everything executed must be owned by root which is a bit of a pain on a desktop machine
2015-01-14 22:28:43 <systmkor> ahh
2015-01-14 22:28:47 <systmkor> right then the TPE group
2015-01-14 22:29:07 <systmkor> is just it can't have any global execute or write options right?
2015-01-14 22:30:39 <systmkor> BitL0G1c, thanks for clarifying
2015-01-14 23:22:34 <bougyman> what supplies stdio.h ?
2015-01-14 23:28:53 <dalias> musl-dev
2015-01-14 23:29:26 <dalias> maybe there's a metapackage you can install to get all the baseline stuff needed for compiler to work; i dunno
2015-01-15 00:30:54 <BitL0G1c> systmkor: I think the default group for grsec_proc = 1001 & grsec_tpe = 1005
2015-01-15 00:32:47 <BitL0G1c> sorry - CONFIG_GRKERNSEC_PROC_GID=30
2015-01-15 00:37:30 <BitL0G1c> setting 'kernel.grsecurity.tpe=1' in /etc/sysctl.conf enables TPE
2015-01-15 00:40:40 <BitL0G1c> but looks like the kernel is built without a TPE group set
2015-01-15 00:47:33 <BitL0G1c> or the ability to turn it on - 'ls /proc/sys/kernel/grsecurity' to see what can be configured at runtime in sysctl
2015-01-15 00:51:55 <BitL0G1c> http://en.wikibooks.org/wiki/Grsecurity/Appendix/Sysctl_Options
2015-01-15 05:01:57 <systmkor> yah I compile a custom version of the kernel with most security related flags of the all cgroup/pax/chroot/tpe/extra grsec ones
2015-01-15 06:57:30 <opendredd> So alpine in a chroot works nicely. Next step: Alpine as host, Arch in chroot
2015-01-15 06:59:04 <opendredd> I wonder if pacman will build on musl...
2015-01-15 07:38:43 <ahills> probably
2015-01-15 07:38:48 <ahills> I've hacked on it before, it's pretty clean
2015-01-15 07:39:00 <ahills> most of the important stuff is separated out into a lib
2015-01-15 07:58:40 <ncopa> i think we have pacman packaged already
2015-01-15 07:59:18 <ncopa> systmkor: i like your idea of an alpine derivative that does desktop
2015-01-15 08:20:35 <ScrumpyJack> why make it a derivative? :)
2015-01-15 08:21:51 <ncopa> or another iso release
2015-01-15 08:21:57 <ncopa> alpine-desktop
2015-01-15 08:22:00 <ncopa> or so
2015-01-15 08:23:01 <ncopa> reason to make it derivative is that then can the core devs move more independent
2015-01-15 08:23:26 <ncopa> it would be nice for me to be able to focus on the core system and core system functionality
2015-01-15 08:23:45 <ncopa> and that others work on the desktop and polishing parts
2015-01-15 08:29:35 <bougyman> i'm dissapointed with the lack of useradd/userdel
2015-01-15 08:29:46 <bougyman> is there something in apk which provides these which i'm missing/
2015-01-15 08:30:03 <ncopa> adduser from busybox
2015-01-15 08:30:15 <bougyman> in most distros the 'passwd' or 'shadow' package would provide useradd and userdel.
2015-01-15 08:30:27 <bougyman> problem here, is, we're trying to chefify alpine
2015-01-15 08:30:35 <bougyman> for massive deployments (thousands of em)
2015-01-15 08:30:54 <bougyman> to do that, it has to respond to certain protocols, in this case, 'useradd/userdel'
2015-01-15 08:31:51 <bougyman> I don't mind the lack of pan.
2015-01-15 08:31:54 <bougyman> *pam
2015-01-15 08:32:04 <ncopa> did you check the shadow package?
2015-01-15 08:32:11 <bougyman> but no useradd/userdel means we need to monkeypatch chef libraries deeply
2015-01-15 08:32:25 <bougyman> i did an apk search shadow and it came up with nothin'
2015-01-15 08:32:36 <ncopa> hum
2015-01-15 08:34:15 <ncopa> its in testing
2015-01-15 08:35:56 <ncopa> you cannot change the chef config to use adduser instead of useradd?
2015-01-15 08:38:55 <bougyman> it's not a config thing
2015-01-15 08:39:01 <bougyman> it'd be a core lin thing
2015-01-15 08:39:14 <bougyman> I was surprised I got chef working at all.
2015-01-15 08:39:22 <bougyman> I still need to tell ohai about alpine
2015-01-15 08:39:30 <bougyman> lsb-release would be great, for that.
2015-01-15 08:40:15 <bougyman> if we had lsb-release, that is, the ahai integration would be smoother.
2015-01-15 08:40:46 <bougyman> it does a case on $(</etc/lsb-release), and falls back to some uncommon /etc/*release stuff.
2015-01-15 08:40:52 <bougyman> but no alpine-
2015-01-15 09:54:36 <ncopa> heh
2015-01-15 09:54:45 <ncopa> i think we do support the /etc/os-release thingy
2015-01-15 09:56:41 <ncopa> what does chef expect from useradd?
2015-01-15 09:56:57 <ncopa> we could maybe provide some alpine-chef wrapper scripts
2015-01-15 11:04:32 <opendredd> Heh
2015-01-15 11:04:35 <opendredd> No terminus package
2015-01-15 12:40:00 <ncopa> opendredd: terminus as a font?
2015-01-15 13:01:29 <Tsutsukakushi> opendredd: the default font not good enough?
2015-01-15 13:37:51 <Jean-Scotch>  Hi. I did recompile busybox with PAM enabled, added and configured nss-pam-ldapd (nslcorrectly. But still can't login with "Login incorrect". Even if pam_ldap(login:auth): authentication succeeded and pam_ldap(login:account): authorization succeeded...
2015-01-15 13:38:57 <Jean-Scotch> I find nothing on this topic in any webpage, forum, etc :(
2015-01-15 13:41:37 <Jean-Scotch> my /etc/pam.d/login is http://sprunge.us/RMMc
2015-01-15 13:43:32 <Jean-Scotch> anyone using pam-ldap here?
2015-01-15 13:52:50 <ncopa> Jean-Scotch: sounds like a busybox bug
2015-01-15 13:53:17 <Jean-Scotch> looks like...
2015-01-15 14:01:51 <Jean-Scotch> I'm looking after this for a couple of days now. I did relearn all I knew about pam. At least that was usefull. But I don't want to be forced to fall back to another distro as I'm moving all my (debian) boxes to Alpine :-/ I'm not sure to possess the skills needed to debug busybox.
2015-01-15 14:02:36 <ncopa> so what you want to do is let users log in with ldap authentication
2015-01-15 14:03:20 <ncopa> i wonder if it would work if you used login from package shadow in testing
2015-01-15 14:03:20 <Jean-Scotch> yes
2015-01-15 14:03:52 <ncopa> i also wonder if we would like a long-term solution for that
2015-01-15 14:04:36 <ncopa> problem is that i dont want link core apps like busybox with pam libs
2015-01-15 14:05:29 <ncopa> however, i think bsdauth is a bit interesting
2015-01-15 14:05:42 <Jean-Scotch> I define users in ldap. when a new user logs in (with sftp trough ssh), his defined home is created on the fly. and the webservers are serving it directly
2015-01-15 14:06:07 <ncopa> makes sense
2015-01-15 14:06:41 <Jean-Scotch> I understand you don'want pam in base package. It's seems a wise choice to me. I cn compile it for the rare boxes I need it.
2015-01-15 14:07:42 <ncopa> bsdauth is fork/exec of a suid root binary
2015-01-15 14:07:46 <ncopa> reads password from stdin
2015-01-15 14:07:54 <Jean-Scotch> I share this $HOME between the sftp, the lighttpd and the php-fpm LXC with a bind mount
2015-01-15 14:07:55 <ncopa> i think it writes the status back
2015-01-15 14:10:25 <ncopa> hum
2015-01-15 14:10:26 <Jean-Scotch> and I also use ldap to manage de mailboxes and some other services
2015-01-15 14:10:36 <ncopa> you need the mkhomedir thingy too
2015-01-15 14:10:39 <ncopa> so pam makes sense
2015-01-15 14:10:40 <Jean-Scotch> yes
2015-01-15 14:11:06 <ncopa> otherwise you could do authorized_keys
2015-01-15 14:11:13 <ncopa> http://jmorano.moretrix.com/2013/09/openssh-6-2-x-ldap-authentication/
2015-01-15 14:11:21 <Jean-Scotch> my setup works like that on debian boxes for ages
2015-01-15 14:11:36 <ncopa> but
2015-01-15 14:11:51 <ncopa> i think openssh does not call busybox login
2015-01-15 14:11:56 <ncopa> or does it?
2015-01-15 14:12:14 <Jean-Scotch> no, I define a sshd service in pam for it
2015-01-15 14:12:35 <Jean-Scotch> I did recompile openssh with pam support on alpine
2015-01-15 14:12:56 <Jean-Scotch> login is easyer to test my pam rules
2015-01-15 14:13:28 <Jean-Scotch> maybe I my just forget about local login and just use it for sshd
2015-01-15 14:13:42 <Jean-Scotch> s+my+may+
2015-01-15 14:14:18 <ncopa> it does sound like something in busybux is broke yes
2015-01-15 14:15:23 <Jean-Scotch> but I still need pam for mkhomedir with sftp and postfix (or qmail)
2015-01-15 14:16:21 <Jean-Scotch> my users are to dumb to use authorized_keys :(
2015-01-15 14:16:58 <Jean-Scotch> even sftp was a PITA to force upon them
2015-01-15 14:27:18 <ncopa> :)
2015-01-15 14:29:45 <ncopa> i have added the gpg key to frontpage: wwwtest.alpinelinux.org
2015-01-15 14:30:01 <ncopa> and on downloads page: http://wwwtest.alpinelinux.org/downloads/
2015-01-15 14:30:29 <ncopa> lync users cannot hover the gpg key link on frontpage to see the key fingerprint
2015-01-15 14:30:54 <ncopa> so my question is, do we need to show the gpg key finger print?
2015-01-15 14:31:01 <ncopa> or is it enough to display it on downloads page?
2015-01-15 14:31:23 <ncopa> i mean, do we need display the gpg key finger print on front page?
2015-01-15 14:34:43 <tvvdhSL5hEBn> not really. in case of a compromise this could be changed, if the person downloading has no other way to verify your key
2015-01-15 14:34:55 <tvvdhSL5hEBn> the best thing would be to have your key signed by lots of people
2015-01-15 14:44:55 <ginjachris> Greetings!
2015-01-15 14:48:44 <ncopa> hi ginjachris
2015-01-15 14:49:11 <ginjachris> How are things @ncopa?
2015-01-15 14:49:23 <ginjachris> Busy as always I imagine....
2015-01-15 14:49:47 <ginjachris> I must thank you again for Alpine, it's pure genius!
2015-01-15 14:53:59 <ncopa> ginjachris: good, but busy yes :)
2015-01-15 14:54:02 <ncopa> thanks
2015-01-15 14:54:55 <ginjachris> I hope to have a chance to try Alpine on rpi soonish
2015-01-15 14:56:28 <ginjachris> after using Alpine, everything else is pants.  Especially on ARM processor
2015-01-15 14:56:58 <ncopa> i got an email that they have sent my 2 wandboards
2015-01-15 14:57:27 <ncopa> hopefully i will reach to set them up before vacation
2015-01-15 15:07:17 <ginjachris> @ncopa, a well deserved holiday!  Off to anywhere nice?  I too am aiming to finally take a holiday this year.....it's been years since I got off this damn island I find myself stuck on!
2015-01-15 15:12:14 <ncopa> i'm going to brazil 4 weeks in feb
2015-01-15 15:13:39 <ginjachris> @ncopa - awesome!  Enjoy :)  Bring us back a gift? (More Dev's!) :)
2015-01-15 15:41:50 <darkfader> ncopa: try to make it 6 weeks, my 4 weeks are just enough to work the backlog away (while in the sun)
2015-01-15 15:41:57 <darkfader> and i hope you enjoy it alot
2015-01-15 15:42:56 <ncopa> i will :)
2015-01-15 16:08:33 <wgreenhouse> is the LBU system of alpine the equivalent of "persistence" on other live systems, or is it possible to use something like aufs or unionfs to have a copy-on-write update of the live system in real time?
2015-01-15 16:12:35 <wgreenhouse> 
ACTION is uncertain if he should use "sys", "data", or live+LBUs

2015-01-15 16:14:49 <BitL0G1c> @wgreenhouse - for a USB install i just did a normal disk install (-m sys) & set /tmp /run /var/log as tmpfs
2015-01-15 16:16:07 <wgreenhouse> BitL0G1c: thanks for the suggestion.
2015-01-15 16:17:20 <wgreenhouse> BitL0G1c: that probably makes updating easier, too; rather than having to reinstall the livesystem for kernel updates, etc.
2015-01-15 16:18:29 <BitL0G1c> yes exactly
2015-01-15 16:18:32 <BitL0G1c> http://www.hastebin.com/giheweqagu.hs
2015-01-15 16:19:09 <BitL0G1c> am using it to test XEN - I've also worked out how to bridge wifi
2015-01-15 16:19:20 <BitL0G1c> with openvswitch
2015-01-15 16:22:18 <BitL0G1c> @xen_roger: http://www.hastebin.com/ohewoluxoc.hs
2015-01-15 16:22:59 <BitL0G1c> tested with LXC containers so far - but should also work for XEN guests
2015-01-15 16:24:44 <xen_roger> BitL0G1c: you do not commit changes with lbu ? or why create xenbr0 everytime ?
2015-01-15 16:25:53 <xen_roger> BitL0G1c: also this way is routing / nat traffic over wlan0 is not "using wifi on ovs-bridge" ^
2015-01-15 16:29:38 <royger> I don't think you can bridge wlan interfaces
2015-01-15 16:32:31 <BitL0G1c> you can but it needs either ebtables to masquerade the MAC addresses - or use nat as i've used
2015-01-15 16:32:48 <BitL0G1c> the lxc guest has connectivity
2015-01-15 16:33:24 <BitL0G1c> I'm not using lbu
2015-01-15 16:35:44 <ScrumpyJack> anyone familiar with lxc networking?
2015-01-15 16:35:55 <BitL0G1c> yes
2015-01-15 16:36:08 <BitL0G1c> (with openvswitch)
2015-01-15 16:37:13 <BitL0G1c> wgreenhouse: additional sysctl.conf settings for a usb install http://www.hastebin.com/otihoxebiq.hs
2015-01-15 16:38:09 <wgreenhouse> thx BitL0G1c
2015-01-15 16:40:10 <BitL0G1c> ScrumpyJack: - this works very well http://blog.scottlowe.org/2014/01/23/automatically-connecting-lxc-to-open-vswitch/
2015-01-15 16:41:56 <ScrumpyJack> unfortunatly sir, I don't have openvswitch
2015-01-15 16:42:00 <ScrumpyJack> :)
2015-01-15 16:42:59 <Jean-Scotch> I also use LXC (and KVM) ith openvswitch. It works like a charm.
2015-01-15 16:43:33 <jomat> ScrumpyJack: I'm a little bit familiar with it
2015-01-15 16:44:12 <jomat> ScrumpyJack: ... at least the physical interfaces of my servers are in a lxc container which does the routing
2015-01-15 16:44:41 <jomat> not to mention: running alpine ;-)
2015-01-15 16:45:09 <BitL0G1c> http://www.hastebin.com/iqusakaqal.bash settings for lxc / ovs
2015-01-15 16:45:17 <ScrumpyJack> jomat: woah! so you give your physical interface to a lxc container? with lxc.network.type = physical?
2015-01-15 16:45:19 <Jean-Scotch> ncopa: for my ssh + pam + mkhomedir, I just setup a LXC with debian and everything was properly setup and configured from scratch in les than 20 min... the only drawback is 400MB for this LXC versus 77 for the corresponding alpine one
2015-01-15 16:45:25 <jomat> ScrumpyJack: yes
2015-01-15 16:45:59 <jomat> so the physical machine just has a veth interface
2015-01-15 16:46:03 <Jean-Scotch> Il will continue to investigat but at least I can go to production with that alpine box
2015-01-15 16:46:22 <ScrumpyJack> jomat: that's crazy :)
2015-01-15 16:46:28 <ScrumpyJack> anyway
2015-01-15 16:46:34 <jomat> :-)
2015-01-15 16:46:37 <jomat> thx :-)
2015-01-15 16:46:37 <ScrumpyJack> here's my problem
2015-01-15 16:46:51 <ScrumpyJack> i only have one physical interface :(
2015-01-15 16:46:55 <jomat> me too
2015-01-15 16:47:21 <ScrumpyJack> ok. perhaps you can help me. i'll try to describe
2015-01-15 16:47:33 <jomat> i don't even have vga or serial console on this stuff :-)
2015-01-15 16:48:04 <ScrumpyJack> eth0 on host has IP a.b.c.d
2015-01-15 16:48:38 <jomat> Aaah, I love hexadecimal representation of IP adresses
2015-01-15 16:49:52 <ScrumpyJack> i want IP 5.6.7.8 to go straight to a lxc guest
2015-01-15 16:50:21 <jomat> from the outside? so you'll need ARP proxying
2015-01-15 16:50:30 <jomat> resp. ndp proxy for v6
2015-01-15 16:50:38 <jomat> ooooor you use a bridge
2015-01-15 16:50:46 <jomat> but i'm no fan of bridges
2015-01-15 16:51:06 <ScrumpyJack> no no, i don't want to bridge either if i can help it
2015-01-15 16:51:35 <jomat> yeah
2015-01-15 16:52:55 <jomat> (internet)<->[(eth0)   (veth0)<-[->(veth1)]]    <--- is that the picture you want? veth1 with 5.6.7.8 and reachable from the Internet?
2015-01-15 16:53:09 <jomat> And you need to fill the gap`
2015-01-15 16:53:10 <jomat> ?
2015-01-15 16:53:55 <ScrumpyJack> i did try adding a macvlan interface to eth0 on the host, then giving the guest lxc.network.type = phys and lxc.network.link = macvlan0. the macvlan0 interface vanished from the host as expected, and appeared in the guest with it's original mac address, but when i gave the guest a public IP, nothing
2015-01-15 16:54:54 <ScrumpyJack> that would have been cool if it worked
2015-01-15 16:56:28 <ScrumpyJack> on the host: ip link add link eth0 mac0 type macvlan
2015-01-15 16:56:46 <ScrumpyJack> and the guest got the same mac address as mac0 :)
2015-01-15 16:58:09 <BitL0G1c> macvlan's are for guest to guest networking
2015-01-15 16:59:30 <BitL0G1c> i just used an openswitch bridge / nat for an nginx reverse proxy
2015-01-15 17:02:42 <ScrumpyJack> BitL0G1c: yes macvlan in lxc.network.type is for lxc guest to guest networking, but more generally it's part of 8021q and can do way more
2015-01-15 17:04:35 <ScrumpyJack> so here is my macvlan device in host: mac0      Link encap:Ethernet  HWaddr 1E:64:34:90:CA:5F
2015-01-15 17:05:07 <ScrumpyJack> and here it is in my guest: eth1      Link encap:Ethernet  HWaddr 1e:64:34:90:ca:5f
2015-01-15 17:08:00 <jomat> ScrumpyJack: sorry, had a phone call... now I'm back :-)
2015-01-15 17:10:00 <jomat> ScrumpyJack: Do you use some firewall scripts like shorewall?
2015-01-15 17:10:13 <ScrumpyJack> it's looking good. i think it's my firewall rules blocking!
2015-01-15 17:10:26 <ScrumpyJack> i use iptables yes
2015-01-15 17:10:31 <ScrumpyJack> ;)
2015-01-15 17:10:35 <jomat> ;-)
2015-01-15 17:11:06 <jomat> that's why I added "scripts" :-)
2015-01-15 17:11:37 <jomat> Well, but how I do it basically is:
2015-01-15 17:11:38 <ScrumpyJack> i find iptables easier to script than wrappers around iptables, if that makes sense
2015-01-15 17:12:05 <jomat> run the container with a veth-pair (i. e. veth0 on the host and veth1 in the container)
2015-01-15 17:13:01 <jomat> configure i. e. 5.6.7.8/24 to veth1 in the container
2015-01-15 17:13:37 <jomat> ehr, and with the peer address a.b.c.d
2015-01-15 17:14:07 <jomat> can be done in /etc/network/interfaces with "pointopoint a.b.c.d"
2015-01-15 17:14:15 <ScrumpyJack> i think what i'm doing works, tcpdump shows incoming to 5.6.7.8/32 when i give 5.6.7.8/32 to the guest
2015-01-15 17:14:59 <jomat> add a host route: "ip r a a.b.c.d via dev veth1" in the container (not sure right now if it's needed)
2015-01-15 17:15:10 <BitL0G1c> a default logging rule helps debugging iptables
2015-01-15 17:15:24 <BitL0G1c> -A OUTPUT ! -o lo -m limit --limit 2/min -j LOG --log-prefix "DROP OUTPUT: " --log-level 4
2015-01-15 17:15:39 <jomat> and then enable arp proxy somewhere in /proc (i do it with shorewall, that's quite comfortable :-D )
2015-01-15 17:15:46 <jomat> and routing of course
2015-01-15 17:16:10 <jomat> but perhaps it's possible with macvlan, too :-)
2015-01-15 17:16:18 <jomat> but I never tried that :-)
2015-01-15 17:20:14 <ScrumpyJack> on host IP is 1.2.3.4/30
2015-01-15 17:20:30 <ScrumpyJack> on guest IP is 5.6.7.8/32
2015-01-15 17:21:14 <ScrumpyJack> (this is with the macvlan interface of host lxc.network.type=phys attached to the guest)
2015-01-15 17:21:48 <ScrumpyJack> when i access 5.6.7.8 from outside the host, i see this in tcpdump
2015-01-15 17:21:59 <ScrumpyJack> 1.2.3.4 > 5.6.7.8
2015-01-15 17:22:28 <ScrumpyJack> not outside > 5.6.7.8 as i'd expect
2015-01-15 17:22:36 <ScrumpyJack> is that wierd?
2015-01-15 17:26:14 <ScrumpyJack> ok, i need to go home. back later
2015-01-15 17:30:50 <ScrumpyJack> actually, what i see in tcpdump is:
2015-01-15 17:31:25 <ScrumpyJack> outside > guest
2015-01-15 17:31:29 <ScrumpyJack> host > guest
2015-01-15 17:31:33 <ScrumpyJack> houst > guest
2015-01-15 17:31:36 <ScrumpyJack> outside > guest
2015-01-15 17:31:37 <ScrumpyJack> houst > guest
2015-01-15 17:31:39 <ScrumpyJack> houst > guest
2015-01-15 17:31:40 <ScrumpyJack> etc
2015-01-15 18:10:06 <ScrumpyJack> back
2015-01-15 18:12:07 <jomat> ScrumpyJack: where does the tcpdump run, and on which interface(s)?
2015-01-15 18:12:46 <jomat> ScrumpyJack: do you see anything with inside the container? (i. e. tcpdump -ni veth1)
2015-01-15 18:13:31 <jomat> s/with//
2015-01-15 18:13:45 <jomat> ScrumpyJack: i. e. if you ping 5.6.7.8 from the outside
2015-01-15 18:18:00 <ScrumpyJack> good point, but i have no network working inside the container, so i'll need to get the packages from outside the container
2015-01-15 18:18:35 <ScrumpyJack> i'll trying from inside the container with tcpdump on the host
2015-01-15 18:22:42 <jomat> ScrumpyJack: you can lxc-attach -n yourcontainer
2015-01-15 18:23:12 <jomat> ScrumpyJack: and it should be possible to install packages with just a chroot /var/lib/lxc/yourcontainer/rootfs
2015-01-15 18:23:26 <ScrumpyJack> this tcpdump is from the host. tcpdump -i eth0 host 5.6.7.8 (IP of guest)
2015-01-15 18:23:27 <jomat> better stop the container for that, but not sure if neccesary
2015-01-15 18:24:02 <ScrumpyJack> when is ping the host IP 1.2.3.4 from within the guest ...
2015-01-15 18:24:22 <jomat> ScrumpyJack: ah ok. have you tried neutralizing your firewall with. sth. like iptables -P FORWARD ACCEPT; iptables -I FORWARD -j ACCEPT
2015-01-15 18:24:53 <ScrumpyJack> the only thing i see is ARP who-has host tell guest
2015-01-15 18:27:47 <ScrumpyJack> so traffic is going from host to guest, but not guest to host?
2015-01-15 18:28:20 <jomat> what does ip r g 5.6.7.8 on the host say?
2015-01-15 18:34:48 <ScrumpyJack> 5.6.7.8 via host gw dev eth0 src 1.2.3.4
2015-01-15 18:36:09 <ScrumpyJack> do i need to remove the gw for the guest?
2015-01-15 18:36:19 <ScrumpyJack> and make it a host route?
2015-01-15 18:37:36 <ScrumpyJack> right, dinner time
2015-01-15 18:37:41 <ScrumpyJack> biab
2015-01-15 19:51:58 <ScrumpyJack> back
2015-01-15 20:07:27 <systmkor> ScrumpyJack, any luck?
2015-01-15 20:13:48 <ScrumpyJack> not yet :(
2015-01-15 20:14:08 <systmkor> wish I could help but I don't know much about iptables
2015-01-15 20:17:10 <ScrumpyJack> i've cleared them
2015-01-15 20:17:32 <ScrumpyJack> i'm going to add tcpdump to container (it's debian)
2015-01-15 20:17:39 <ScrumpyJack> host is AL
2015-01-15 20:18:47 <ScrumpyJack> actually, i'll prolly just reboot lxc with veth and add some tools. veth works with NAT just fine
2015-01-15 20:31:17 <ScrumpyJack> added tcpdump - testing now
2015-01-15 20:37:45 <ScrumpyJack> tcpdump from inside the guest shows nothing when filtering on IP of guest
2015-01-15 20:37:58 <ScrumpyJack> but there is a whole lot of other chatter on there
2015-01-15 20:47:40 <ScrumpyJack> so the guest is seeing my provider's private IPs 10.3.x.x but nothing from the intertubes
2015-01-15 20:48:26 <ScrumpyJack> and only ARP req/replies
2015-01-15 21:59:26 <systmkor> interesting
2015-01-15 21:59:58 <systmkor> ScrumpyJack, I am trying to compile tor on alpine but it keeps throwing an error that it can't find libevent
2015-01-15 22:00:12 <systmkor> even when I specify it's in /usr/lib
2015-01-15 22:00:25 <systmkor> can't find or can't find a linkable version
2015-01-15 22:00:42 <systmkor> is there something related compiling with MUSL that changes anything for libevent ?
2015-01-16 07:51:49 <ScrumpyJack> .
2015-01-16 08:29:35 <paper_face> so i'm setting up a alpinelinux system as a gateway, and i need pppoe, what should i use? regular Paul's PPP Package or Roaring Penguin's ppp client
2015-01-16 08:29:51 <paper_face> anyone have any opinions either way?
2015-01-16 08:45:29 <paper_face> oops i dropped there did anyone say anything?
2015-01-16 08:56:00 <ncopa> paper_face: nobody said anything
2015-01-16 08:56:14 <ncopa> use regular ppp package
2015-01-16 08:56:39 <paper_face> any specific reason? there wasn't much info on either page
2015-01-16 08:57:04 <paper_face> i did notice you maintain the ppp-pppoe package
2015-01-16 08:57:05 <ncopa> i think it is good integrated with /etc/network/interfaces
2015-01-16 08:57:09 <paper_face> ah
2015-01-16 08:58:09 <ncopa> there are ppp-daemon and ppp-pppoe packages
2015-01-16 08:58:19 <ncopa> i think those sould be enough
2015-01-16 08:59:20 <paper_face> thanks for the info.
2015-01-16 09:00:05 <ncopa> create a /etc/ppp/peers/my-isp file
2015-01-16 09:00:22 <ncopa> containing a line:
2015-01-16 09:00:25 <ncopa> plugin rp-pppoe.so eth0
2015-01-16 09:00:27 <ncopa> or similar
2015-01-16 09:00:41 <ncopa> user <your user>
2015-01-16 09:00:52 <ncopa> password <yourpassword>
2015-01-16 09:00:53 <ncopa> etc
2015-01-16 09:01:08 <ncopa> then in /etc/network/interfaces you have:
2015-01-16 09:01:25 <ncopa> auto ppp0
2015-01-16 09:01:25 <ncopa> iface ppp0 inet ppp
2015-01-16 09:01:44 <ncopa>    provider my-isp
2015-01-16 09:01:49 <paper_face> could you censor yours and put them in a pastebin, am thinking of making a wiki article for this
2015-01-16 09:01:54 <paper_face> as it's a fairly common usecase
2015-01-16 09:02:08 <paper_face> (course i shall test it first)
2015-01-16 09:02:14 <ncopa> might be you also need 'pre-up ip link set dev eth0 up'
2015-01-16 09:02:17 <ncopa> ok
2015-01-16 09:02:59 <paper_face> I'm new to alpinelinux, but so far very impressed
2015-01-16 09:03:22 <paper_face> lbu looks really neat
2015-01-16 09:03:49 <paper_face> one thing i did actually also think about was the caveats of USB stick booting vs CD
2015-01-16 09:04:02 <paper_face> and the possibility of modified firmware on the usb stick being a vector of attack
2015-01-16 09:04:12 <paper_face> was thinking of using one of these http://kanguru.com/storage-accessories/kanguru-flashtrust-secure-firmware.shtml
2015-01-16 09:04:35 <paper_face> i seem to remember an article some time ago about a lot of usb sticks being able to have the firmware modified
2015-01-16 09:05:08 <paper_face> but i guess "all bets are off" if you have physical access to the machine.
2015-01-16 09:05:35 <Tsutsukakushi> bad usb
2015-01-16 09:05:44 <ncopa> this is what i used for /etc/ppp/peers/telenor in norway: http://sprunge.us/MEVb
2015-01-16 09:05:44 <paper_face> yeah that http://arstechnica.com/security/2014/07/this-thumbdrive-hacks-computers-badusb-exploit-makes-devices-turn-evil/
2015-01-16 09:05:50 <Tsutsukakushi> affects one biiig usb manufacteur iirc
2015-01-16 09:06:17 <Tsutsukakushi> also affects the usb controllers in your mobo
2015-01-16 09:06:19 <Tsutsukakushi> so
2015-01-16 09:06:24 <Tsutsukakushi> tough luck
2015-01-16 09:06:45 <paper_face> yeah im thinking of using a fit-PC4 for my new router
2015-01-16 09:06:58 <paper_face> has nice shiny intel ethernet cards
2015-01-16 09:07:40 <paper_face> ncopa: got the /etc/network/interfaces files there to go with that?
2015-01-16 09:08:07 <ncopa> paper_face: the relevant part: http://pastie.org/9834856
2015-01-16 09:08:24 <ncopa> i am not sure that 'pre-up ip link set dev eth0 up' is needed
2015-01-16 09:08:29 <paper_face> yep then the other one i shall make a static address
2015-01-16 09:08:33 <ncopa> telenor is the name of the isp
2015-01-16 09:08:39 <paper_face> figured
2015-01-16 09:09:20 <Tsutsukakushi> is there any busybox shred alternative?
2015-01-16 09:09:27 <Tsutsukakushi> or gnu shred
2015-01-16 09:09:30 <Tsutsukakushi> for busybox
2015-01-16 09:09:39 <ncopa> for busybox i dont know
2015-01-16 09:10:14 <ncopa> paper_face: could you please write a http://wiki.alpinelinux.org/wiki/PPP article?
2015-01-16 09:10:24 <paper_face> yes that is the intention
2015-01-16 09:10:44 <paper_face> im going to add a bit that allows for a DHCP server too
2015-01-16 09:10:50 <paper_face> any specific recommendation on dhcp server
2015-01-16 09:11:17 <ncopa> i have always used ISC dhcp
2015-01-16 09:11:25 <ncopa> there is also dnsmasq
2015-01-16 09:11:27 <paper_face> yeah i was thinking of using that
2015-01-16 09:11:52 <ncopa> dnsmasq is nice for small installs
2015-01-16 09:12:31 <ncopa> but i normally want a full dns resolver, like unboud
2015-01-16 09:12:37 <ncopa> unbound
2015-01-16 09:12:39 <paper_face> yeah
2015-01-16 09:13:08 <ncopa> which kind of makes the dns part or dnsmasq useless
2015-01-16 09:21:37 <paper_face> i should be able to do this with two usb sticks
2015-01-16 09:21:48 <paper_face> one that boots alpine linux (read only) and one for the lbu to be stored
2015-01-16 09:22:13 <paper_face> the second one i will obviously want to mount via uuid though
2015-01-16 09:22:18 <paper_face> ie in fstab
2015-01-16 09:41:51 <paper_face> this may seem like a stupid question, but if i am running alpine linux from a usb stick
2015-01-16 09:41:56 <paper_face> and i have an additional usb stick
2015-01-16 09:42:43 <paper_face> i am guessing i want to run setup-disk and select "data"
2015-01-16 09:42:48 <paper_face> and can i store my lbu on there?
2015-01-16 09:45:52 <paper_face> ie /var/lbu
2015-01-16 09:46:32 <ncopa> why not put it on the first, boot stick?
2015-01-16 09:46:49 <paper_face> oh i guess i could do that.
2015-01-16 09:46:50 <ncopa> i'm not sure i'd recommend putting /var on a secondary usb stick
2015-01-16 09:46:58 <paper_face> i suppose
2015-01-16 09:47:10 <paper_face> i was just doing it as a test
2015-01-16 09:47:19 <paper_face> but i guess when i get the fitpc4 it will have a single sata disk in it
2015-01-16 09:47:26 <paper_face> so using one usb stick is probably more realistic
2015-01-16 09:47:39 <paper_face> when i come to doing it on the fitpc it will be a sata disk instead
2015-01-16 09:47:48 <ncopa> i dont think it will find the apkovl.tar.gz on /var/lbu/*
2015-01-16 09:48:03 <paper_face> where does it look for it by default?
2015-01-16 09:48:05 <ncopa> might be it will be found if you put it directly on /var
2015-01-16 09:48:25 <paper_face> (assuming i used one stick)
2015-01-16 09:48:39 <paper_face> i looked at http://wiki.alpinelinux.org/wiki/Alpine_local_backup it didn't really mention
2015-01-16 09:49:02 <paper_face> or is /etc/lbu wheere it is supposed to be
2015-01-16 09:49:51 <paper_face> oh i guess /media/usb
2015-01-16 09:49:55 <paper_face> is probably where it looks for it
2015-01-16 09:50:39 <paper_face> mm that would be read only though
2015-01-16 09:51:59 <ncopa> i think with alpine v3.1 it will try mount everything blkid finds
2015-01-16 09:52:06 <paper_face> my bad, i didn't uncommot the LBU_BACKUPDIR in lbu.conf
2015-01-16 09:52:26 <paper_face> which by default looks like it puts in /root/config-backups
2015-01-16 09:52:41 <ncopa> and it will look for $mnt/*.apkovl.tar.gz
2015-01-16 09:53:02 <ncopa> putting things in /root/confnig-backups sounds not like a good solution to  me:)
2015-01-16 09:53:07 <ncopa> i think default is empty
2015-01-16 09:53:29 <paper_face> oh
2015-01-16 09:54:34 <paper_face> maybe i just wanted to uncomment that LBU_MEDIA="usb"
2015-01-16 10:50:39 <paper_face> hmm
2015-01-16 10:50:44 <paper_face> i followed this http://wiki.alpinelinux.org/wiki/Configure_Networking#Example:_Dual-Stack_Configuration
2015-01-16 10:50:55 <paper_face> i wonder why my interfaces aren't being brought up with a static ip set
2015-01-16 10:51:43 <paper_face> oh there we go :) forgot to start networking
2015-01-16 11:43:46 <paper_face> is it possible to use aes 256 on rsa key generated by ssh-keygen?
2015-01-16 11:54:03 <xen_roger> paper_face: some thing like "ssh-keygen -t rsa -E aes256-cbc " ?
2015-01-16 11:54:43 <paper_face> i t ried that but there's no such option
2015-01-16 11:59:27 <paper_face> someone proposed it here xen_roger http://www.gossamer-threads.com/lists/openssh/bugs/45093
2015-01-16 11:59:44 <paper_face> did it ever happen?
2015-01-16 12:00:57 <paper_face> do you have to encrypt them manually with openssl?
2015-01-16 12:08:23 <paper_face> hmm, how do i prevent sshd on alpinelinux from generating keys ie -A on boot
2015-01-16 12:08:32 <paper_face> commenting out HostKeys didn't seemt o do the trick
2015-01-16 12:09:24 <ncopa> add -A to SSHD_OPTS in /etc/conf.d/sshd
2015-01-16 12:09:31 <paper_face> ah
2015-01-16 12:09:41 <paper_face> because i only want it to use my ssh key :)
2015-01-16 12:10:05 <ncopa> hum
2015-01-16 12:10:42 <ncopa> it looks like it will run gen_keys anyways
2015-01-16 12:11:18 <paper_face> oh.
2015-01-16 12:11:25 <paper_face> but why do i need all those keys to be generated
2015-01-16 12:11:36 <paper_face> i just want to put my rsa key in /etc/ssh
2015-01-16 12:11:57 <paper_face> [22:39:23] <ncopa> add -A to SSHD_OPTS in /etc/conf.d/sshd
2015-01-16 12:12:02 <paper_face> and wouldn't that do the reverse
2015-01-16 12:12:07 <paper_face> ie run sshd with -A
2015-01-16 12:12:13 <ncopa> yes it would
2015-01-16 12:12:18 <ncopa> hum
2015-01-16 12:12:18 <ncopa> ok
2015-01-16 12:12:24 <paper_face> i want to do the opposite :P
2015-01-16 12:12:43 <ncopa> i think we need to add a config option for that
2015-01-16 12:13:01 <ncopa> something like DISABLE_KEYGEN=1 in /etc/conf.d/sshd or similar
2015-01-16 12:13:07 <paper_face> yes.
2015-01-16 12:13:48 <ncopa> paper_face: could you please file a bug on bugs.alpinelinux.org and i'll have a look at it
2015-01-16 12:14:05 <ncopa> set target version to 3.1.2
2015-01-16 12:14:14 <ncopa> or i will forget it
2015-01-16 12:14:17 <paper_face> okay will do
2015-01-16 12:14:50 <ncopa> hum
2015-01-16 12:15:04 <ncopa> other option would be to configure what keys we want it to autogenerate
2015-01-16 12:15:23 <ncopa> gen_keys() {
2015-01-16 12:15:23 <ncopa>         if egrep -q '^[[:space:]]*Protocol[[:space:]]+.*1' "${SSHD_CONFDIR}"/sshd_config ; then
2015-01-16 12:15:23 <ncopa>                 gen_key rsa1 "" || return 1
2015-01-16 12:15:23 <ncopa>         fi
2015-01-16 12:15:23 <ncopa>         gen_key dsa && gen_key rsa && gen_key ecdsa && gen_key ed25519
2015-01-16 12:15:23 <ncopa>         return $?
2015-01-16 12:15:25 <ncopa> }
2015-01-16 12:15:38 <paper_face> well that too
2015-01-16 12:20:12 <ncopa> i think the current init.d script is just a ripoff from gentoo
2015-01-16 12:21:43 <paper_face> also you shouldn't need your private keys in /etc/ssh
2015-01-16 12:21:51 <paper_face> that does not seem good at all
2015-01-16 12:22:00 <paper_face> i put my public key in there and it overwrote it
2015-01-16 12:22:06 <paper_face> because you can't stop it from generating them lol
2015-01-16 12:23:45 <paper_face> that's a good point though
2015-01-16 12:23:55 <paper_face> if someone did get root access they could just copy all the keys from /etc/ssh
2015-01-16 12:24:14 <paper_face> which could potentially give access to elsewhere
2015-01-16 12:24:25 <paper_face> also i dont usually make crappy dsa keys
2015-01-16 12:24:37 <ncopa> huh? how could it give access other places?
2015-01-16 12:24:46 <paper_face> well if u used the ssh key elsewhere
2015-01-16 12:25:01 <paper_face> and they aren't encrypted either :P
2015-01-16 12:25:16 <paper_face> i always use a passphrase
2015-01-16 12:25:23 <ncopa> i think thats what the gen key thingy is about
2015-01-16 12:25:29 <paper_face> in any case i dont see any reason a private key should ever be on a remote machine
2015-01-16 12:25:31 <ncopa> prevent reuse of same host key
2015-01-16 12:27:23 <paper_face> in any case when that bug (the generating keys -A) is fixed one can just dump their public key in /etc/ssh and wont have to have a private one there
2015-01-16 12:27:35 <paper_face> seems if the private one is missing sshd just generates a new pair for that algorithm
2015-01-16 12:28:22 <ncopa> ssh keys should be unique
2015-01-16 12:28:35 <ncopa> ssh host keys i mean
2015-01-16 12:29:45 <ncopa> the host key is only serving as an identification, to make sure that the server you try to connect to, really is that server
2015-01-16 12:30:09 <ncopa> the pub key of that keypair is stored in the clients known_hosts
2015-01-16 12:32:30 <paper_face> oh yeah my bad i confused them for authorized_keys
2015-01-16 14:38:55 <vkrishn> news !! , got my landline broadband reconnected
2015-01-16 14:39:16 <vkrishn> it got disconnected due to road contruction
2015-01-16 14:39:45 <vkrishn> but still at 512kbps, but nice
2015-01-16 14:39:49 <vkrishn> :-))
2015-01-16 14:40:50 <vkrishn> and unlimited !!
2015-01-16 15:11:39 <ncopa> vkrishn: congrats :)
2015-01-16 15:12:51 <vkrishn> :-) thx
2015-01-16 15:15:33 <vkrishn> should be setting couple of AL servers this week
2015-01-16 15:16:23 <vkrishn> ncopa, any solution to booting and auto wifi
2015-01-16 15:16:35 <vkrishn> or I just add a cron script
2015-01-16 15:30:40 <ncopa> wpa_supplicant?
2015-01-16 15:32:10 <vkrishn> ok would check the docs
2015-01-16 15:34:10 <vkrishn> I wifi'ed my simple ADSL router by connecting to TP-LINK-3020
2015-01-16 20:35:47 <darkfader> http://hastebin.com/fizidojiwe.avrasm
2015-01-16 20:35:58 <darkfader> honestly i think the lvm2 error messages need to be supprssed
2015-01-16 20:36:19 <darkfader> this is an excerpt from 100+ lines while i was changing mirrors
2015-01-16 20:36:33 <darkfader> this diag shit easily hides important messages
2015-01-16 20:39:18 <darkfader> 6 lines per mirror add/removal
2015-01-16 20:40:14 <systmkor> lvm2 ?
2015-01-16 20:40:18 <systmkor> as in volume manager?
2015-01-16 20:42:37 <darkfader> ye
2015-01-16 20:43:04 <darkfader> it tries to lock memory for critical operations but that throws an error "per default"
2015-01-16 20:43:47 <darkfader> so far i'd been ignoring it but it's risky if you have something that is safe to ignore only in light usage
2015-01-16 20:44:24 <darkfader> noone really knows why the error is coming, the suggestiong was it would be low memory conditions, but that's not true
2015-01-16 20:44:30 <darkfader> you even get it with 64GB of ram
2015-01-16 20:44:48 <darkfader> it's how they do the lock that is problem
2015-01-16 20:55:31 <systmkor> oh.. that sucks
2015-01-16 20:57:25 <KozakMAN> hi
2015-01-16 20:57:33 <systmkor> KozakMAN, hello
2015-01-16 20:58:37 <KozakMAN> clean install alpine 3.1.1 + asterisk
2015-01-16 20:58:56 <KozakMAN> con't load chan_pjsip.so
2015-01-16 20:59:01 <KozakMAN> symbol not found
2015-01-16 20:59:55 <KozakMAN> error relocating
2015-01-16 21:01:06 <KozakMAN> what can i do to load module ?
2015-01-16 21:25:19 <systmkor> i don't know sorry
2015-01-16 21:30:03 <systmkor> I am running into an issue with trying to compile
2015-01-16 21:30:21 <systmkor> tor but ./configure can't find libevent
2015-01-16 21:30:39 <systmkor> even though it's in /usr/lib/
2015-01-16 21:35:07 <ScrumpyJack> systmkor: i'm interested in compiling tor
2015-01-16 21:35:31 <systmkor> ScrumpyJack, well help would be wonderful :)
2015-01-16 21:35:45 <ScrumpyJack> are you building a package?
2015-01-16 21:35:51 <systmkor> yah
2015-01-16 21:35:55 <ScrumpyJack> or compiling manually
2015-01-16 21:36:03 <systmkor> compiling manually then package
2015-01-16 21:36:15 <ScrumpyJack> can i have your APKBUILD file?
2015-01-16 21:36:35 <systmkor> I didn't make one yet because I wanted to make sure just running ./configure then ./make worked
2015-01-16 21:36:42 <systmkor> it's stuck on the ./configure portion
2015-01-16 21:36:49 <ScrumpyJack> ok
2015-01-16 21:36:56 <systmkor> because it can't find /usr/lib/libevent
2015-01-16 21:37:01 <systmkor> even though its there
2015-01-16 21:37:45 <ScrumpyJack> where's your source for tor?
2015-01-16 21:38:11 <systmkor> https://dist.torproject.org/
2015-01-16 21:39:37 <systmkor> i am trying to compile
2015-01-16 21:39:50 <ScrumpyJack> 2.6.2?
2015-01-16 21:40:23 <systmkor> tor-0.2.5.10.tar.gz
2015-01-16 21:40:33 <systmkor> i believe that's most recent stable
2015-01-16 21:40:37 <systmkor> 2.5.10
2015-01-16 21:48:21 <BitL0G1c> you probably want libevent-dev
2015-01-16 21:48:34 <systmkor> i sware i searched for that and didn't see it ugghh
2015-01-16 21:48:48 <systmkor> fuck me...
2015-01-16 21:48:50 <systmkor> lol my bad
2015-01-16 21:48:57 <BitL0G1c> :)
2015-01-16 21:49:13 <BitL0G1c> there is a libeventlog-dev too
2015-01-16 21:50:15 <systmkor> BitL0G1c, ScrumpyJack did either of you get a chance to check out gitian
2015-01-16 21:50:40 <systmkor> i sent the email twice the mailing list
2015-01-16 21:56:09 <ScrumpyJack> not yet for gitian. it's not the sort of thing i get involved with in Apline Linux. It's ncopa you need to convince. (i'm just a user)  Shame noone took you up on the discussion on the mailing list.
2015-01-16 21:56:40 <ScrumpyJack> tor compiling with openssl-dev and libevent-dev as deps
2015-01-16 21:56:41 <systmkor> ScrumpyJack, just curious on peoples opinions
2015-01-16 22:02:38 <ScrumpyJack> make succeeded for tor with libevent-dev and openssl-dev
2015-01-16 22:02:42 <ScrumpyJack> can you try?
2015-01-16 22:04:39 <systmkor> yah one sec
2015-01-16 22:04:44 <systmkor> sorry talking to co-worker
2015-01-16 22:05:53 <ScrumpyJack> what's your timezone?
2015-01-16 22:08:24 <systmkor> UTC -8:00
2015-01-16 22:09:02 <systmkor> West coast continental U.S.A
2015-01-16 22:10:28 <ScrumpyJack> im UTC, so time for bed :)
2015-01-16 22:10:29 <systmkor> ScrumpyJack, it appears to have compiled for me
2015-01-16 22:10:51 <ScrumpyJack> cool
2015-01-16 22:10:58 <systmkor> ScrumpyJack, Alpine needs more users not in the European area
2015-01-16 22:11:23 <systmkor> ScrumpyJack, I'm also planning on making proxychains a package as well
2015-01-16 22:11:31 <systmkor> i have some web scraping to do :D
2015-01-16 22:11:54 <ScrumpyJack> web scraping? what's that?
2015-01-16 22:13:26 <systmkor> :P
2015-01-16 22:13:35 <systmkor> grabing data from websites in an automated fashion
2015-01-16 22:13:52 <systmkor> ScrumpyJack, http://scrapy.org/
2015-01-16 22:15:05 <ScrumpyJack> ah. that's what curl is for :)
2015-01-16 22:15:52 <systmkor> ScrumpyJack, ehhhhh
2015-01-16 22:16:00 <systmkor> when you need to do more complicated things then no
2015-01-16 22:16:06 <systmkor> but curl is a good simple start
2015-01-16 22:21:39 <ScrumpyJack> systmkor: can you submit your tor and proxychains packages?
2015-01-16 22:24:32 <systmkor> ScrumpyJack, will be doing that within the hour :D
2015-01-16 22:24:53 <systmkor> ScrumpyJack, does anyone have a script to get an lxc container up and running
2015-01-16 22:24:58 <systmkor> with alpine-sdk
2015-01-16 22:25:04 <systmkor> and nothing else
2015-01-16 22:25:14 <systmkor> and compiles the package to make sure no missing dependencies?
2015-01-16 22:26:57 <ScrumpyJack> systmkor: you need to figure out the make dependancies yourself - usually in the INSTALL or README in the source :). abuild will figure out the runtime dependancies for you
2015-01-16 22:29:38 <ScrumpyJack> i don't have a script for an alpine linux build container. there's not much to it apart from getting edge, apk add alpine-sdk and setting up git
2015-01-16 22:30:02 <ScrumpyJack> http://wiki.alpinelinux.org/wiki/Setting_up_the_build_environment_on_HDD
2015-01-16 22:31:23 <ScrumpyJack> i guess you could use docker for that - but i find it quicker in a shell
2015-01-16 22:33:18 <systmkor> ScrumpyJack, yah I can figure it out
2015-01-16 22:33:23 <systmkor> but I meant to validate
2015-01-16 22:35:46 <ScrumpyJack> what do you mean by validate?
2015-01-16 22:36:02 <systmkor> because for example
2015-01-16 22:36:23 <systmkor> I may have installed a dependency for a package I am working on months ago
2015-01-16 22:36:29 <BitL0G1c> I use lxc containers as my build environment - works very well - one for 32 bit one for 64 bit
2015-01-16 22:36:35 <systmkor> and during the build I specify all the dependencies I think it needs
2015-01-16 22:36:43 <systmkor> but forgot one which was already installed, months ago
2015-01-16 22:36:47 <systmkor> it builds correctly
2015-01-16 22:36:54 <systmkor> but when I push the APKBUILD file
2015-01-16 22:37:05 <systmkor> someone else on a new system could be missing the dependency
2015-01-16 22:37:25 <systmkor> BitL0G1c, I just started messing with that yesterday
2015-01-16 22:37:34 <BitL0G1c> abuild detects the dependencies from the binary
2015-01-16 22:37:57 <systmkor> and auto inserts them into the APKBUILD?
2015-01-16 22:38:03 <ScrumpyJack> check /etc/apk/world. it has the packages you added manually (without deps). remove anything suspicious
2015-01-16 22:38:17 <BitL0G1c> (that is built) - inserts them into the control file in the package
2015-01-16 22:38:19 <systmkor> BitL0G1c, also abuild does fail sometimes
2015-01-16 22:38:34 <systmkor> for example scrapy was missing dependencies
2015-01-16 22:38:40 <BitL0G1c> no
2015-01-16 22:38:41 <systmkor> for running the cli portion of the program
2015-01-16 22:38:51 <systmkor> scrapy even though it can be imported in python
2015-01-16 22:39:00 <BitL0G1c> the only dependencies to be put in the APKBUILD are for the compile
2015-01-16 22:39:02 <systmkor> so I updated that in the APKBUILD
2015-01-16 22:39:25 <BitL0G1c> you only need the xxx-dev packages in APKBUILD
2015-01-16 22:39:44 <systmkor> for most C programs yah
2015-01-16 22:39:51 <BitL0G1c> abuild works out the run time dependencies
2015-01-16 22:39:58 <systmkor> but there are python modules that use
2015-01-16 22:40:08 <systmkor> things outside of that
2015-01-16 22:40:14 <systmkor> outside of a dev package
2015-01-16 22:40:26 <BitL0G1c> yes  python is a little different
2015-01-16 22:40:39 <systmkor> BitL0G1c, graunted I am new to abuild so I am not trying to claim better knowledge on it
2015-01-16 22:40:43 <systmkor> just what I experienced
2015-01-16 22:41:20 <BitL0G1c> I had to add a few extra things to get virtualbricks to build & run
2015-01-16 22:44:21 <BitL0G1c> (virtualbricks is a python app)
2015-01-16 22:44:35 <systmkor> ah
2015-01-16 22:46:21 <ScrumpyJack> what BitL0G1c said. not all runtime deps get picked up by abuild. i've not really looked at abuild nor do i want to. it works fine if you know what your package needs to run :)
2015-01-16 22:46:38 <systmkor> ScrumpyJack, got it
2015-01-16 22:46:52 <systmkor> idk how i like a tool automajically detecting dependencies in the background
2015-01-16 22:47:09 <systmkor> without notifying me as someone who is trying to make the package
2015-01-16 22:47:32 <systmkor> or not just automatically puting it in the APKBUILD
2015-01-16 22:47:40 <systmkor> to review
2015-01-16 22:49:38 <BitL0G1c> you should also test the pkg runs after it's built
2015-01-16 22:49:49 <systmkor> BitL0G1c, is this on the wiki?
2015-01-16 22:50:13 <BitL0G1c> not sure - it just from experience
2015-01-16 22:50:15 <systmkor> BitL0G1c, I do test the package after it is built
2015-01-16 22:50:41 <systmkor> i wish though that projects provided automated tests for a package
2015-01-16 22:50:49 <systmkor> and unit-tests
2015-01-16 22:51:20 <systmkor> in a non-unique way
2015-01-16 22:53:15 <systmkor> okay going to go write proxychains & tor packages
2015-01-16 22:53:18 <systmkor> be back shortly
2015-01-16 22:59:41 <ScrumpyJack> automatically test packages? that's interesting. can you elaborate? we pay tester vasts amounts of money here. I'm listening :)
2015-01-16 23:00:22 <ScrumpyJack> ooo. that's testerS, with an S
2015-01-16 23:01:02 <systmkor> ScrumpyJack, well so for example are you familiar with unit-tests?
2015-01-16 23:02:35 <systmkor> ScrumpyJack, Alpine has a budget?
2015-01-16 23:05:16 <systmkor> what do you know tor is already in testing, man I am on the idiot train today
2015-01-16 23:06:39 <ScrumpyJack> systmkor: no i didnt mean testing in Apline, i meant testing at wok
2015-01-16 23:07:49 <ScrumpyJack> and yes, tor is in testing :) doesnt hurt to learn though :)
2015-01-16 23:08:12 <ScrumpyJack> s/wok/work
2015-01-16 23:08:13 <systmkor> yup yup :P
2015-01-16 23:08:22 <systmkor> though proxychains isn't
2015-01-16 23:08:24 <systmkor> so will add that
2015-01-16 23:08:35 <systmkor> ScrumpyJack, does Alpine have any "hackathons"
2015-01-16 23:08:43 <systmkor> like a weekend of just developing for Alpine
2015-01-16 23:09:39 <ScrumpyJack> not that i know of, but "what i don't know, you could fit into the Hollywood bowl"
2015-01-16 23:10:43 <systmkor> :P
2015-01-16 23:15:21 <ScrumpyJack> Im just a happy camper. if you want to know more, you need ncopa: barthalion: clandmeter: fabian_a: etc
2015-01-16 23:22:39 <systmkor> okay cool thanks
2015-01-17 01:28:24 <Mp5shooter> anyone using cacti on musl?
2015-01-17 01:31:18 <systmkor> cacti the monitoring tool?
2015-01-17 01:31:51 <systmkor> Mp5shooter, it is a mainline package
2015-01-17 01:31:58 <Mp5shooter> yes
2015-01-17 01:32:04 <Mp5shooter> says it can't connect to mysql for me
2015-01-17 01:32:21 <Mp5shooter> it's very weird
2015-01-17 01:32:31 <systmkor> i've never used or installed it before so I won't be too much help
2015-01-17 01:32:43 <systmkor> and I assume your MySQL server is up
2015-01-17 01:32:52 <systmkor> and conenectable with the mysql cli client?
2015-01-17 01:32:59 <Mp5shooter> yep
2015-01-17 01:33:05 <Mp5shooter> all settings are correct
2015-01-17 01:33:30 <systmkor> is there any C compilation ?
2015-01-17 01:33:41 <systmkor> meaning for Cacti to be built?
2015-01-17 01:34:43 <Mp5shooter> um
2015-01-17 01:35:13 <systmkor>  just looked at port there isn't
2015-01-17 01:35:41 <Mp5shooter> no it's all php files
2015-01-17 01:36:30 <systmkor> can you connect to mysql from a php-myadmin
2015-01-17 01:36:32 <systmkor> style thing
2015-01-17 01:36:40 <systmkor> so a different php mysql web app?
2015-01-17 01:36:56 <Mp5shooter> yes i can connect from adminer
2015-01-17 01:37:10 <systmkor> hrmmm
2015-01-17 01:37:19 <systmkor> i don't know
2015-01-17 01:37:21 <systmkor> sorry
2015-01-17 01:38:11 <Mp5shooter> yeah it's weird :/
2015-01-17 01:38:28 <systmkor> have you tried a manually installed version of cacti
2015-01-17 01:38:41 <systmkor> meaning like a seperate root htmldocs directory
2015-01-17 01:38:49 <Mp5shooter> no
2015-01-17 01:40:21 <Mp5shooter> usually when something like this happens to me it's a missing package or something
2015-01-17 01:40:30 <Mp5shooter> but nothing is missing this time around
2015-01-17 01:40:31 <Mp5shooter> :P
2015-01-17 02:26:35 <systmkor> thaaat sucks
2015-01-17 02:26:42 <Mp5shooter> yeah
2015-01-17 03:01:18 <systmkor> gully_foyle, lil-bitch, and rapeclown seem to be throwing a masterbatorial lemon party :P
2015-01-17 03:02:51 <gully_foyle> sorry
2015-01-17 03:03:00 <gully_foyle> I kept getting banned on other channels
2015-01-17 03:03:25 <gully_foyle> I'm banned for good on the ubuntu channel now
2015-01-17 03:03:38 <gully_foyle> oh well time for slackware
2015-01-17 03:06:55 <Mp5shooter> bah
2015-01-17 03:06:57 <Mp5shooter> systmkor
2015-01-17 03:06:59 <Mp5shooter> I found the issue
2015-01-17 03:07:11 <Mp5shooter> I changed database type from mysql to mysqli on a whim
2015-01-17 03:07:16 <Mp5shooter> and it made cacti work!
2015-01-17 03:07:21 <Mp5shooter> idk why normal mysql isn't working
2015-01-17 03:09:02 <systmkor> gully_foyle, was just toying with your nick's just funny that's all
2015-01-17 03:09:11 <gully_foyle> oh
2015-01-17 03:09:16 <gully_foyle> alright
2015-01-17 03:09:25 <systmkor> Mp5shooter, awesome
2015-01-17 03:09:41 <Mp5shooter> yeah
2015-01-17 03:09:41 <Mp5shooter> :D
2015-01-17 03:10:47 <Mp5shooter> yeah
2015-01-17 03:11:00 <Mp5shooter> after i installed php-mysqli package adminer worked before
2015-01-17 03:11:07 <Mp5shooter> but i didn't think of that with cacti
2015-01-17 03:11:11 <Mp5shooter> maybe something's wrong with php-mysql
2015-01-17 03:13:18 <systmkor> good question
2015-01-17 03:13:26 <systmkor> document document document :P
2015-01-17 03:14:24 <Mp5shooter> :P
2015-01-17 03:14:34 <Mp5shooter> net-snmp-tools should be another dependency for cacti package too
2015-01-17 03:14:37 <Mp5shooter> since it includes snmpwalk
2015-01-17 03:14:43 <Mp5shooter> and cacti needs that
2015-01-17 03:14:43 <Mp5shooter> :P
2015-01-17 03:47:17 <systmkor> Mp5shooter, email the mailing list
2015-01-17 03:47:35 <systmkor> and cc the maintainer of cacti package
2015-01-17 04:23:13 <Guest766> hello all, a bit confused in getting UTF-8 to work properly, specifically for drawing characters on screen
2015-01-17 04:25:24 <Guest766> well, for reference, looks like export NCURSES_NO_UTF8_ACS=1 did the trick
2015-01-17 05:14:37 <dalias> what terminal/app were affected?
2015-01-17 06:17:18 <paper_face> when setting up dhcpcd on alpine linux, how do you specify the listen interface
2015-01-17 06:17:48 <paper_face> in ubuntu you do this in  /etc/default/isc-dhcp-server
2015-01-17 06:18:09 <paper_face> eg INTERFACES="eth0"
2015-01-17 06:18:53 <paper_face> this is distribution specific isn't it?
2015-01-17 06:19:37 <paper_face> as it looks like systemd based distributions have their own way of doing it ie https://wiki.archlinux.org/index.php/Dhcpd#Listening_on_only_one_interface
2015-01-17 06:24:02 <paper_face> ah probably what i want is In /etc/conf.d/dhcpd set IFACE to "eth0".
2015-01-17 06:24:22 <paper_face> keep forgetting that it uses openrc like gentoo
2015-01-17 10:07:45 <paper_face> ncopa: would i see ppp0 in ip addr?
2015-01-17 10:07:53 <paper_face> or only when a ppp connection is established
2015-01-17 10:08:11 <paper_face> i dont even think the interface exists.
2015-01-17 10:08:29 <paper_face> despite having http://pastie.org/9834856
2015-01-17 11:08:59 <Guest766> midnight commander wasn't showing properly
2015-01-17 12:20:44 <Phiphler> Any gotchas before I install Alpine on my Raspberry Pi? I want to run Owncloud primarily
2015-01-17 12:35:46 <Phiphler> SSH does not seem to work OOB?
2015-01-17 12:41:48 <tvvdhSL5hEBn> that's a bold claim
2015-01-17 12:41:54 <Jean-Scotch> ssh is enabled with setup-alpine. not before that
2015-01-17 12:57:20 <tvvdhSL5hEBn> :)
2015-01-17 13:36:00 <paper_face> ncopa: u there
2015-01-17 13:47:57 <sinetek> hello
2015-01-17 13:48:05 <sinetek> I have a feature request, something simple I believe
2015-01-17 13:48:34 <sinetek> would it be possible to add the Colemak keyboard layout to setup-alpine?? :D
2015-01-17 15:20:08 <paper_face_> can you edit your apkovl.tar.gz file on another distribution
2015-01-17 15:23:27 <sinetek> me?
2015-01-17 15:26:05 <sinetek> i need a newer kernel, how do i go about tis
2015-01-17 15:31:53 <sinetek> where is loadkeys?
2015-01-17 15:31:54 <sinetek> :/
2015-01-17 15:33:35 <Mp5shooter> anyone here using cacti on alpine? have you had an issue with the rrd files not being generated when the poller is run by cron?
2015-01-17 15:54:56 <HarryHaaren> hey all, Alpine linux noobie here, I'm a long time Arch user, and linux-audio developer.
2015-01-17 15:55:25 <HarryHaaren> I'm looking into a small run-from-RAM distro, and hope to compile a -rt (realtime) kernel for low-latency audio processing.
2015-01-17 15:55:51 <HarryHaaren> Is anybody here or on the forums already doing this? If so, please ping me
2015-01-17 16:07:00 <Phiphler> so Im running setup-alpine and I want to put /var on a separate USB harddrive, but the installer warns me that it will wipe the hdd. Any way around wiping it? Cant I just put /var on its own partition?
2015-01-17 16:08:12 <xen_roger> Phiphler: move data after install and add an fstab entry to mount stick to var on boot
2015-01-17 16:08:48 <Phiphler> xen_roger: That will do it, thank you for your answer
2015-01-17 16:10:08 <xen_roger> keep in mind to preserve owner and permission while moving data and delete content of var after moving...
2015-01-17 16:44:04 <paper_face> ncopa: is this the same for alpine-linux
2015-01-17 16:44:06 <paper_face> http://wiki.gentoo.org/wiki/Home_Router#ADSL_and_PPPoE
2015-01-17 16:44:21 <paper_face> maybe thats why pppd just hangs when it tries to make a ppp connection because ppp0 doesn't exist
2015-01-17 16:44:45 <paper_face> although i thought alpinelinux used /etc/network/interface instead of /etc/conf.d/net
2015-01-17 16:45:31 <paper_face> although it could be because of some other reason
2015-01-17 16:45:33 <paper_face> When the DSL interface comes up, it will create a "ppp0" entry to the output of ifconfig command. Although the NIC is called eth1, the IP address is actually bound to ppp0. From now on substitute 'eth1' with 'ppp0'.
2015-01-17 16:45:36 <paper_face> that that is not happening
2015-01-17 16:49:08 <xen_roger> ppp0 is the "tunnel" device on eth1 so it is correct that there are both devices ...
2015-01-17 17:37:34 <Jean-Scotch> the tor init scripts seems broken to me as running the program from command line works fine but the init script dies poorly
2015-01-17 17:55:20 <Jean-Scotch> I think I found the culprit ;)
2015-01-17 18:14:13 <Jean-Scotch> script in not broken. Iwas using a privileged port and the script change user to an unprivileged one...
2015-01-17 19:56:47 <Roosevelt> ls
2015-01-17 19:57:58 <sinetek> $
2015-01-17 20:01:28 <c355E3B> Anyone know why the vmware package for x cant find the `fbCreateDefColormap` symbol?
2015-01-17 20:03:23 <Roosevelt> hey all is nvidia blob possible on here?
2015-01-17 20:04:24 <dalias> ask on #musl; i believe there's been success using it with some hackery
2015-01-17 20:04:55 <dalias> but is there a reason you don't just use the free driver? afaik it's better
2015-01-17 20:05:28 <Roosevelt> you talking about the nouveau
2015-01-17 20:06:07 <Roosevelt> No there is no reason other than using vdpau
2015-01-17 20:07:23 <sinetek> gentoo has a nice musl overlay
2015-01-17 20:12:25 <dalias> how is that related?
2015-01-17 20:21:16 <pcarrier> OMG the lemo is great
2015-01-17 20:21:44 <pcarrier> oops wrong chan :D
2015-01-17 20:24:47 <Roosevelt> i'm installing alpine now is it mandatory to recompile kernel?
2015-01-17 20:25:35 <dalias> no
2015-01-17 20:25:53 <dalias> but you might want to install the vanilla kernel package instead of the grsec one
2015-01-17 20:26:12 <dalias> the grsec one has some extreme hardening that breaks lots of programs
2015-01-17 20:28:18 <sinetek> is there a kernel that's not so old?
2015-01-17 20:29:21 <Roosevelt> how does alpine load modules like the gentoo way?
2015-01-17 20:29:39 <Roosevelt> very old..
2015-01-17 20:30:29 <dalias> sinetek, how old is it?
2015-01-17 20:30:40 <Roosevelt> 3.1 is very old
2015-01-17 20:30:56 <sinetek> 3.14 afaik
2015-01-17 20:31:02 <sinetek> my laptop needs 3.16+
2015-01-17 20:31:03 <sinetek> :p
2015-01-17 20:31:25 <dalias> 3.14 yes
2015-01-17 20:32:19 <Roosevelt> so the grsec is not recommended for desktop usage which is the way I will be using it
2015-01-17 20:32:37 <dalias> yeah i wouldn't recommend grsec for desktop
2015-01-17 20:32:41 <dalias> iirc firefox breaks
2015-01-17 20:32:57 <dalias> and any kind of process/resource monitors using info from /proc break
2015-01-17 20:32:58 <Roosevelt> I need firefox
2015-01-17 20:34:00 <Tsutsukakushi> 3.14 isn't that old
2015-01-17 20:34:07 <Tsutsukakushi> i wouldn't really recommend alpine for desktop anyways
2015-01-17 20:34:24 <Roosevelt> why is that?
2015-01-17 20:34:30 <Tsutsukakushi> desktop would need a lot bigger repos
2015-01-17 20:34:42 <dalias> well
2015-01-17 20:34:55 <Roosevelt> well it's easy to make packages on here
2015-01-17 20:35:05 <dalias> if you use alpine for a desktop, you'll probably find you need to build stuff that's not already packaged
2015-01-17 20:35:13 <Tsutsukakushi> yeah
2015-01-17 20:35:47 <dalias> otoh if you use some desktop-oriented dist like ubuntu, you'll find 50GB of crap installed, constant swapping, random wiping of your hdd from upgrades that change how your partititons are identified, ...
2015-01-17 20:38:36 <Roosevelt> where are modules loaded?
2015-01-17 20:39:24 <Tsutsukakushi> modprobe?
2015-01-17 20:41:38 <Roosevelt> are the modules loaded in /etc/conf.d/modules...like gentoo
2015-01-17 21:09:50 <systmkor> Roosevelt, I believe so
2015-01-17 21:11:02 <systmkor> Roosevelt, OpenRC is the default for Alpine
2015-01-17 21:37:18 <Roosevelt> okay well see installing now
2015-01-17 21:56:18 <Roosevelt> what happen to mkfs.ext4?
2015-01-17 21:57:46 <dalias> ?
2015-01-17 21:58:47 <Roosevelt> never mind apk add e2fsprogs didn't know one could install software on the cd
2015-01-17 22:13:12 <systmkor> Roosevelt, anything confusing document :D
2015-01-17 22:19:15 <Roosevelt> my only problem right now is after setting up my partitions with fstab I get a message you might need fix the MBR to be able to boot I used fdisk and marked my /boot(/dev/sda1) active
2015-01-17 22:19:41 <Roosevelt> after setup-alpine /mnt
2015-01-17 22:45:27 <systmkor> Roosevelt, i sorta remember that
2015-01-17 22:45:32 <systmkor> but not too well
2015-01-18 00:58:15 <systmkor> does anyone have experience on how to properly write an APKBUILD
2015-01-18 00:58:25 <systmkor> that should have $pkgname-doc
2015-01-18 00:58:29 <systmkor> as a sub-package
2015-01-18 00:58:52 <systmkor> looking in aports for existing packages that do that but so far it doesn't help
2015-01-18 01:05:30 <ovf> systmkor: the "simple" example produces -doc: http://wiki.alpinelinux.org/wiki/APKBUILD_examples:Simple
2015-01-18 01:06:05 <ovf> if you need more control over the generation of -doc subpackage, provide your own doc() function: http://wiki.alpinelinux.org/wiki/APKBUILD_examples:Special_Subpackages_Doc
2015-01-18 01:06:45 <systmkor> ahhh okay
2015-01-18 01:06:54 <systmkor> ovf, i have a feeling i'll have to provide my own doc function
2015-01-18 01:07:02 <systmkor> ovf, thank you
2015-01-18 01:16:11 <systmkor> ovf, saweet I think I got it to work
2015-01-18 01:44:00 <c355E3B> Anyone know if there is a workaround for https://bugs.alpinelinux.org/issues/3691 ?
2015-01-18 01:49:43 <Roosevelt> hey is it okay for me to add some documentation to the wiki for the manual installation?
2015-01-18 02:05:47 <systmkor> Roosevelt, go for it :D
2015-01-18 02:06:07 <systmkor> also add commit summary message
2015-01-18 02:28:18 <systmkor> Roosevelt, let me know when you finish your edits
2015-01-18 02:28:21 <systmkor> i'll look over them
2015-01-18 02:31:57 <Roosevelt> okay no problem
2015-01-18 03:26:02 <Roosevelt> @systmkor I don't see the wiki edits?'
2015-01-18 03:27:06 <systmkor> Roosevelt, are you searching for this http://wiki.alpinelinux.org/w/index.php?title=Installation&action=history?
2015-01-18 03:28:10 <systmkor> also you have to make an account before editing if I remember correctly
2015-01-18 03:29:23 <Roosevelt> yeah that's it..
2015-01-18 03:29:49 <Roosevelt> I want to add to the installation method it need tweaking for a manual install
2015-01-18 03:31:23 <systmkor> ah okay one sec
2015-01-18 03:31:30 <systmkor> which page was that specifically?
2015-01-18 03:32:38 <Roosevelt> http://wiki.alpinelinux.org/wiki/Setting_up_disks_manually
2015-01-18 03:33:21 <systmkor> so log in
2015-01-18 03:34:01 <systmkor> then at the top of the article next to the search box
2015-01-18 03:34:25 <systmkor> it should have several tabs: read, edit, view history, (down arrow)
2015-01-18 03:35:19 <systmkor> Roosevelt, any luck?
2015-01-18 03:36:12 <Roosevelt> I had problems signing up let me try again
2015-01-18 03:37:54 <Roosevelt> asking for the confirmation which I keep puting answering no to: are you a spammer?
2015-01-18 05:59:52 <paper_face> ncopa: one thing i did notice is that if ppp0 cannot be brought up the system doesn't continue to boot
2015-01-18 06:01:38 <paper_face> eg:
2015-01-18 06:01:39 <paper_face> auto ppp0
2015-01-18 06:01:39 <paper_face> iface ppp0 inet ppp
2015-01-18 06:01:39 <paper_face> 	pre-up ip link set dev eth0 up
2015-01-18 06:01:39 <paper_face> 	provider myISP
2015-01-18 06:01:39 <paper_face> 	post-down ip link set dev eth0 down
2015-01-18 06:02:04 <paper_face> btw modprobing pppoe fixed the messages i was seeing from pppd about the "kernel not supporting pppoe"
2015-01-18 06:02:23 <paper_face> and i made sure to add it to /etc/modules
2015-01-18 06:30:26 <systmkor> Roosevelt, any luck?
2015-01-18 07:22:23 <systmkor> Roosevelt, i take that as a no
2015-01-18 07:22:33 <systmkor> also checked to see any history difference and didn't see you have a post
2015-01-18 12:35:18 <Jean-Scotch> Hi. I'm trying to install redmine following the wiki but it breaks at the secret_token generation.
2015-01-18 12:35:32 <Jean-Scotch> I know nothing about ruby :(
2015-01-18 12:36:29 <c355E3B> can you gist the full error?
2015-01-18 12:36:55 <Jean-Scotch> rake aborted! Could not find 'rack' (>= 0) among 34 total gem(s)
2015-01-18 12:37:38 <c355E3B> did you run `bundle install` before trying rake?
2015-01-18 12:38:07 <Jean-Scotch> not. just a fresh LXC. then following the wiki
2015-01-18 12:38:18 <c355E3B> and did you setup a `config/database.yml` ?
2015-01-18 12:38:32 <Jean-Scotch> http://wiki.alpinelinux.org/wiki/Redmine#Installing_Redmine_Since_2.1.2
2015-01-18 12:38:35 <Jean-Scotch> yes I did
2015-01-18 12:38:58 <Jean-Scotch> I just added sudo
2015-01-18 12:45:47 <Jean-Scotch> they are 9 files called rack around
2015-01-18 12:47:19 <c355E3B> seems about right
2015-01-18 12:47:41 <c355E3B> can you run rake -T ?
2015-01-18 12:53:25 <Jean-Scotch> ok
2015-01-18 12:53:43 <Jean-Scotch> same error
2015-01-18 12:57:51 <Jean-Scotch> same error for rake without any argument
2015-01-18 12:58:15 <c355E3B> hmm
2015-01-18 15:21:04 <Roosevelt> @systmkor I'm having problems creating the acount
2015-01-18 15:42:33 <ahills> Has anyone else experienced Xen domU hangs?
2015-01-18 15:42:45 <ahills> I'm getting them very frequently, sometimes several times per day on some VMs
2015-01-18 15:42:52 <ahills> Doesn't seem to matter which distro
2015-01-18 15:46:47 <kaniini> what hardware ?
2015-01-18 15:52:16 <xen_roger> ahills: any logs ?
2015-01-18 15:52:50 <ahills> Hardware is old, dual opterons on an ASUS motherboard
2015-01-18 15:54:20 <ahills> xen_roger: the qemu-dm-[domU].log files are all empty, the xl-[domU].log files all just have two lines in them from when I run "xl destroy":
2015-01-18 15:54:50 <ahills> waiting for domain to die, then domain has been destroyed
2015-01-18 15:55:09 <ahills> dmesg isn't showing anything either
2015-01-18 15:56:26 <ahills> another place I could look?
2015-01-18 15:58:34 <xen_roger> ahills: did you limit dom0 memory? are there any failures in hardware in host? maybe mem-corruption ?
2015-01-18 15:59:06 <xen_roger> ahills: don´t really have an idea... just guessing
2015-01-18 15:59:39 <ahills> dom0 memory is limited to 256MB, I have ECC memory but I guess that doesn't mean it's not bad
2015-01-18 15:59:54 <ahills> I ran memtest86+ a few weeks ago and didn't find a problem
2015-01-18 16:00:08 <xen_roger> 256MB ?
2015-01-18 16:00:11 <ahills> no hardware failures that I can detect
2015-01-18 16:00:29 <xen_roger> quite a bit to less... try at least 512MB
2015-01-18 16:00:36 <ahills> hmm, alright, will do
2015-01-18 16:00:51 <ahills> only using 94MB right now according to free -m
2015-01-18 16:01:10 <xen_roger> had some problems in early days with xen and low mem on host ...
2015-01-18 16:01:18 <ahills> alright, thanks
2015-01-18 16:01:35 <xen_roger> report back if or not problem is solved by that
2015-01-18 16:01:44 <ahills> possible qemu processes running out of memory you think?
2015-01-18 16:04:04 <xen_roger> I darkly remember a rule for that but not 100% ... maybe I could find a note for thet
2015-01-18 16:04:13 <xen_roger> /e/a/
2015-01-18 16:36:35 <ahills> xen_roger: upped the memory, now to wait for the next hang
2015-01-18 16:36:38 <ahills> hopefully for my whole life
2015-01-18 16:36:51 <xen_roger> I am with you ;)
2015-01-18 17:04:19 <Roosevelt> Is there any reason why I can't create account on wiki?
2015-01-18 19:41:33 <xen_roger1> ahills: domu(s) still up and running ?
2015-01-18 20:13:14 <ahills> xen_roger1: so far, so good!
2015-01-18 20:41:20 <K350> When trying to connect to a smtp server I use: smtp.server.org:587/tls/user=user The login info is correct and works well in other clients. However in debug mode Alpine says that teh info is sent as PLAIN and that the format is wrong. The connection is then closed and I can't acces. What to do here?
2015-01-18 20:44:13 <K350> I forog tot mention - I use Alpine 2.10
2015-01-18 22:18:05 <ScrumpyJack> Roosevelt: I think there is a bug report for that. Can you add a comment to the bug report? that might get it looked at :)
2015-01-18 22:46:30 <xen_roger> ahills: Yeaaahhhh ;)
2015-01-19 00:52:51 <Roosevelt> ScrumpyJack thanks
2015-01-19 18:48:02 <tze> In case somebody missed it https://blog.xenproject.org/2015/01/15/less-is-more-in-the-new-xen-project-4-5-release/
2015-01-19 21:03:07 <systmkor> dalias, ahh ports anthing you want me to check out
2015-01-19 21:03:27 <systmkor> dalias, probably should of sent that message here instead of in the #musl chatroom
2015-01-19 21:03:28 <systmkor> my b
2015-01-20 01:47:55 <systmkor> dalias, all the suid/sgid files i could find using 'find'
2015-01-20 01:47:57 <systmkor> http://pastebin.com/bQx9S9PM
2015-01-21 20:18:35 <NeoGeo64> hello
2015-01-21 20:48:54 <ScrumpyJack> evening
2015-01-21 21:43:33 <rhapsodhy> are tor packages available in alpine?
2015-01-21 21:44:01 <rhapsodhy> i was looking at ticket https://bugs.alpinelinux.org/issues/1068 but couldn't find the actual package
2015-01-21 21:45:11 <rhapsodhy> ok, found it
2015-01-21 22:24:53 <systmkor> rhapsodhy, proxychains should be committed to /testing within 24 hours
2015-01-21 22:24:57 <systmkor> if you need that as well
2015-01-21 23:22:47 <tze> Hey guys, is there a way to downgrade a system when using sysmode?
2015-01-21 23:23:38 <tze> There seems to be a bug with 3.1.1 when booting with xen, i get a kernel panic on 2 machines, already opened a bug in the tracker
2015-01-22 00:03:30 <systmkor> idk off hand, however most of the users/devs are in Europe so I would recommend on sending an email to mailing list
2015-01-22 00:04:06 <systmkor> probably alpine-devel
2015-01-22 00:05:09 <systmkor> tze, hopefully that helps
2015-01-22 00:05:13 <systmkor> tze, good luck
2015-01-22 00:05:41 <tze> For now i just took an older version of the linux-grsec package from an older iso release. As far as i can see there is no mechanism built into apk, when old package is not in the cache.
2015-01-22 00:10:24 <systmkor> tze, well definitely send a mail to the devel
2015-01-22 00:10:36 <systmkor> maybe it's something that should eventually be offered
2015-01-22 00:10:37 <systmkor> idk
2015-01-22 08:17:54 <ScrumpyJack> ncopa: you about? was wondering if you had the time and enery to add LXC console support to tmux :)
2015-01-22 08:30:48 <systmkor> ScrumpyJack, what do you mean?
2015-01-22 08:41:27 <ScrumpyJack> systmkor: the ability to use tmux to attach to an LXC console
2015-01-22 08:41:41 <systmkor> aaah directly
2015-01-22 08:41:49 <systmkor> rather than just running a lxc-console inside tmux
2015-01-22 08:41:57 <systmkor> ScrumpyJack, that would be nice
2015-01-22 08:43:35 <systmkor> ScrumpyJack, hey so i'm looking into trying to remove all SUIDs & SGIDs from Alpine's base installed system
2015-01-22 08:43:49 <systmkor> ScrumpyJack, do you know any reference material?
2015-01-22 08:54:14 <ScrumpyJack> I'm not sure I know what you mean by reference material
2015-01-22 08:55:48 <ScrumpyJack> do you mean how to look for them?
2015-01-22 08:56:09 <ScrumpyJack> find / -user root -perm -4000  :0
2015-01-22 08:58:57 <Diftraku> and while typing, you slip the mode to 0000
2015-01-22 09:05:12 <crow> ScrumpyJack you was also the one who needed NFS with 3.* ?
2015-01-22 09:08:31 <ScrumpyJack> crow: nope, not me
2015-01-22 09:24:11 <ncopa> ScrumpyJack: any ideas how to implement that?
2015-01-22 09:25:52 <ncopa> i think i recently changed the alpine template to include a /dev/console tty
2015-01-22 09:26:13 <ncopa> whcih means that you now can do: lxc-start -n $container
2015-01-22 09:26:25 <ncopa> and the login happens in current console
2015-01-22 09:26:53 <ncopa> so you could start a tmux server and start the container in a new tmux window
2015-01-22 09:26:55 <ncopa> should work
2015-01-22 09:27:06 <ncopa> i did something similar with xen
2015-01-22 09:27:12 <ncopa> but it does not work 100% perfect
2015-01-22 09:27:29 <ncopa> if i reboot the xen domu then will tmux lose the console
2015-01-22 09:27:50 <ncopa> i dont know if that happened in gnu screen
2015-01-22 09:58:44 <vkrishn> noticing a behaviour will `less`, when I "less somefile.iso", on my knoppix less(444) I see the iso content
2015-01-22 09:58:59 <vkrishn> but not on AL's pkg less
2015-01-22 10:00:50 <vkrishn> does it require some default param ?
2015-01-22 10:07:52 <vkrishn> ok may be this in bashrc
2015-01-22 10:07:53 <vkrishn> # make less more friendly for non-text input files, see lesspipe(1)
2015-01-22 10:07:53 <vkrishn> [ -x /usr/bin/lesspipe ] && eval "$(SHELL=/bin/sh lesspipe)"
2015-01-22 10:12:48 <vkrishn> hmm, did not figured it out yet
2015-01-22 10:49:34 <ScrumpyJack> ncopa: i use lxc-start -d option to background boots.
2015-01-22 10:50:46 <ScrumpyJack> ncopa: do you think we can have #3702 in the kernelconfig at some point?
2015-01-22 10:51:34 <ncopa> it will give performance penalty to fork(2)
2015-01-22 10:52:18 <tvvdhSL5hEBn> fork is already a performance penalty. no?
2015-01-22 10:52:27 <tvvdhSL5hEBn> i try to avoid forking if possible
2015-01-22 10:54:15 <ncopa> yes
2015-01-22 11:12:35 <ScrumpyJack> ncopa: you can pass cgroup_disable=memory to turn it off :)
2015-01-22 11:13:36 <ScrumpyJack> parhaps pass cgroup_disable=memory by default and let people remove it if the need it
2015-01-22 11:14:48 <ScrumpyJack> http://cateee.net/lkddb/web-lkddb/CGROUP_MEM_RES_CTLR.html
2015-01-22 11:15:57 <ScrumpyJack> do you mean MM_OWNER ?
2015-01-22 12:44:56 <ncopa> ScrumpyJack: I'd prefer to have the feature compiled in kernel, but disabled by default, so to enable you would need do cgroup_enable=memory
2015-01-22 12:45:11 <ncopa> iirc that only solves the memory usage penalty
2015-01-22 12:45:26 <ncopa> i dont know if it solves the fork penalty
2015-01-22 12:46:19 <ncopa> that said, i am not totally against enabling cgroup memory knob
2015-01-22 12:46:50 <ncopa> i just wonder what the price is
2015-01-22 12:47:00 <ncopa> and how to do it properly
2015-01-22 13:49:42 <ScrumpyJack> "i am not totally against enabling cgroup memory knob" without an (in)definite article in that sentance, you are calling me a knob!
2015-01-22 13:52:03 <ncopa> sorry :)
2015-01-22 13:52:31 <ncopa> knob: i meant *the* cgroup memory knob
2015-01-22 16:40:47 <Jean-Scotch> Hi, whats the right method to mount a nfs share in alpine LXC on alpine host at startup of the LXC ?
2015-01-22 16:45:15 <Jean-Scotch> I've got a rpcbind[5836]: segfault in ld-musl-x86_64.so.1[ and grsec is banning suid/sgid execs for 15 minutes
2015-01-22 17:38:46 <lexnokon^> hello, if i want to install custom software with alpine, do i need to follow all the steps here http://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package ? is there a tool like dpkg-buildpackage for alpine?
2015-01-22 17:55:31 <lexnokon^> hello
2015-01-22 18:00:53 <lexnokon^> should i post this to the forum?
2015-01-23 08:37:27 <ncopa> <Jean-Scotch
2015-01-23 08:37:30 <ncopa> ẗhats a bug
2015-01-23 08:37:33 <ncopa> segfault
2015-01-23 08:37:45 <ncopa> i have seen rpcbind segfault but have not been able to reproduce it
2015-01-23 08:37:50 <Jean-Scotch> I was thinking so
2015-01-23 08:38:12 <ncopa> i'd like to fix that properly
2015-01-23 08:38:20 <Jean-Scotch> it do it always when trying to mount nfsmount from LXC
2015-01-23 08:38:29 <Jean-Scotch> 'AL container on AL host)
2015-01-23 08:38:39 <ncopa> from within the container?
2015-01-23 08:38:42 <Jean-Scotch> yes
2015-01-23 08:39:13 <ncopa> could you try set up an ubuntu vm and set up lxc there
2015-01-23 08:39:17 <ncopa> and see if it happens there too?
2015-01-23 08:39:32 <ncopa> also
2015-01-23 08:39:39 <ncopa> i am interested if it happens on vanilla kernel too
2015-01-23 08:39:39 <Jean-Scotch> currently the nfs server is debian and also, mountting the share from the AL host has strage behaviour
2015-01-23 08:40:00 <ncopa> i know for sure that idmap does not work
2015-01-23 08:40:11 <ncopa> but basic mount should work
2015-01-23 08:40:18 <Jean-Scotch> the uid/gid are 8+ digits and it is impossible to use the share
2015-01-23 08:40:56 <Jean-Scotch> I will work on that topic this afternoon and prepare a nice bug report
2015-01-23 08:41:00 <ncopa> do you think you could file a bug on bugs.a.o?
2015-01-23 08:41:00 <ncopa> thanks
2015-01-23 08:41:13 <ncopa> try include how to reproduce it
2015-01-23 08:41:25 <Jean-Scotch> for sure
2015-01-23 08:41:46 <Jean-Scotch> otherwise it's kind of useless
2015-01-23 08:41:51 <Jean-Scotch> ;)
2015-01-23 08:42:35 <ncopa> heh, its not that uncommon to get "somethign is broken" tickets :)
2015-01-23 08:43:04 <ncopa> which of course gets closed with "something was fixed"
2015-01-23 08:43:09 <Jean-Scotch> currently I'm scripting the deployment of a new SAN from plugging an AL usb key to working situation ;)
2015-01-23 12:26:11 <Jean-Scotch> ncopa: I plan http://wiki.alpinelinux.org/wiki/User_talk:Jch#NFS_bug_study Do you need more outputs?
2015-01-23 12:38:37 <ncopa> Jean-Scotch: no
2015-01-23 12:38:40 <ncopa> its more than enough
2015-01-23 12:38:47 <ncopa> i think you can drop the baremetal stuff too
2015-01-23 13:44:58 <vkrishn> can I do a usb install and boot similar to cdrom, i.e "modloop-grsec" mount method ?
2015-01-23 15:13:33 <Jean-Scotch> ncopa: http://wiki.alpinelinux.org/wiki/User_talk:Jch#NFS-server_in_KVM-Alpine updated
2015-01-23 17:00:40 <Jean-Scotch> I mean http://wiki.alpinelinux.org/wiki/User_talk:Jch#NFS_bug_study  updated
2015-01-24 05:14:55 <vkrishn> ah, ok figured out the boot issue, found here http://wiki.alpinelinux.org/wiki/Create_a_Bootable_USB
2015-01-24 06:02:29 <ahills> anyone tried to build rustlang on alpine yet?
2015-01-24 06:51:30 <vkrishn> looks like fat32/vfat does not support uuid=<?> boot param
2015-01-24 13:16:36 <pcarrier> is buildlab readily available in stable or edge?
2015-01-26 12:30:40 <vkrishn> ah, noticed "boot to wifi" internet connection happens on boot, but dies after couple of minutes
2015-01-26 13:04:13 <clandmeter> vkrishn: no vacation, just busier with other things.
2015-01-26 13:04:41 <clandmeter> ncopa: any luck on mpd?
2015-01-26 13:14:37 <ncopa> not yet
2015-01-26 13:14:44 <ncopa> but dont stop asking ::)
2015-01-26 14:03:27 <karasz_> anyone has some pointers on how to compile toybox with musl?
2015-01-26 14:04:00 <ncopa> karasz_: never tried
2015-01-26 15:28:37 <Jean-Scotch> bjrk
2015-01-26 16:11:48 <arca_vorago> Hello all
2015-01-26 16:12:19 <arca_vorago> Really interested by Alpine linux, hate selinux and like the idea of grsec
2015-01-26 16:12:48 <arca_vorago> Wondering if anyone has gotten ansible running and functional. Wiki shows it's borked. http://wiki.alpinelinux.org/wiki/Ansible
2015-01-26 17:05:01 <jomat> arca_vorago: the last change of this wiki page was 7 August 2013 ... so if you give it a try (it's also on my todo list), and it works, someone could update the wiki :-)
2015-01-26 18:49:51 <bradfa> Apparently I'm dull, but I'm trying to create an account on the wiki and I keep being told that I've put in an "incorrect or missing confirmation code" which I assume is the "Are you a spammer?" question response.  Any pointers?  I've tried answering various forms of yes and no and "I am not a spammer" to no avail.  Thanks!
2015-01-27 08:35:57 <pcarrier> can't seem to boot alpine-mini-3.1.1-x86_64.iso on UEFI
2015-01-27 08:36:06 <pcarrier> ncopa: is it supposed to work?
2015-01-27 08:37:45 <ncopa> i have never tested that
2015-01-27 08:37:58 <ncopa> so i'm not surprised it does not work
2015-01-27 08:38:08 <pcarrier> all righty
2015-01-27 08:38:15 <ncopa> what does uefi need to boot iso?
2015-01-27 08:38:28 <ncopa> should probably have a look at it for v3.2
2015-01-27 08:38:51 <pcarrier> https://wiki.archlinux.org/index.php/USB_flash_installation_media#BIOS_and_UEFI_Bootable_USB might be useful
2015-01-27 08:38:52 <ncopa> i were able to set up my laptop with uefi and dualboot win8
2015-01-27 08:39:09 <pcarrier> nope, useless
2015-01-27 08:40:30 <ncopa> the reason we dont use isohybrid (so you can dd if=alpine.iso ...) is that we use fat
2015-01-27 08:40:31 <pcarrier> looks like it takes a whole other partition :)
2015-01-27 08:40:38 <ncopa> and fat does not support resize
2015-01-27 08:40:49 <pcarrier> oh wow I just assumed it used isohybrid
2015-01-27 08:41:10 <ncopa> we need be able to to store the config on boot media
2015-01-27 08:41:13 <pcarrier> well this might explain that
2015-01-27 08:41:18 <ncopa> :)
2015-01-27 08:41:33 <pcarrier> well that's... frustrating
2015-01-27 08:42:13 <pcarrier> so there's a good chance that a UEFI bootable media could use FAT for any partition
2015-01-27 08:42:20 <pcarrier> it actually relies on FAT for the system partition
2015-01-27 08:42:24 <ncopa> yes
2015-01-27 08:42:34 <pcarrier> but that partition needs some magic marker
2015-01-27 08:42:40 <ncopa> we shoudl fix that
2015-01-27 08:42:48 <ncopa> will uefi boot a cdrom?
2015-01-27 08:42:57 <ncopa> i suppose not
2015-01-27 08:43:09 <pcarrier> no idea
2015-01-27 08:46:59 <ncopa> i think not
2015-01-27 08:47:04 <ncopa> since uefi needs boot of fat
2015-01-27 08:47:09 <ncopa> and cdrom is iso9660
2015-01-27 09:15:24 <pcarrier> well maybe with El Torino?
2015-01-27 09:15:44 <pcarrier> so UEFI + El Torino?
2015-01-27 09:16:07 <pcarrier> not saying it's a good idea :)
2015-01-27 10:34:23 <vkrishn> got my rapoo X1800 keyboard, at $14 for wireless keyboard seems nice
2015-01-27 10:35:05 <vkrishn> only issues are getting used to the slightly new layout but overall nice, key strokes good response
2015-01-27 10:35:59 <vkrishn> only caution is to insert the connect wifi dongle on may computer and not to a hub that has other usbflash attached, as it disconnects the hub on insert
2015-01-27 10:36:26 <vkrishn> dongle on computer*
2015-01-27 10:39:03 <vkrishn> re: for isohybrid, ^^
2015-01-27 10:39:39 <vkrishn> would it be easy to have param for save_disk=/dev/sdb1
2015-01-27 10:40:26 <vkrishn> that way if someone wants to do dd=<> to the iso file the would have to add the extra param at boot
2015-01-27 10:48:11 <ncopa> or we could at setup-alpine create another partition of the free unpartitioned space
2015-01-27 10:50:21 <vkrishn> hmm, but I thought that was not possible on fat
2015-01-27 10:51:23 <vkrishn> I mean since dd uses the whole disk
2015-01-27 11:01:44 <vkrishn> even for partition will it not need the boot param or it would do auto search, or just increment the /dev/sda<num> ?
2015-01-27 11:03:07 <vkrishn> I had a auto search added to my knoppix setup a year back, code is not clean but works
2015-01-27 11:04:12 <vkrishn> couple of days back I move the readonly sfs files to SD card and locked it ;)
2015-01-27 11:04:53 <vkrishn> worse part is its very hard to get a good SD card reader that honors the lock
2015-01-27 11:06:19 <vkrishn> can't complain for a 50c device
2015-01-27 11:22:25 <vkrishn> I hope its known that any dvd/cd that is written with utf8 support like once done by K3B (by default) will not be readable on AL !!
2015-01-27 11:23:20 <vkrishn> or maybe the version of K3b I am using is not ok :)
2015-01-27 11:45:28 <pcarrier> gah coinbase has a stupid delay to buy bitcoins
2015-01-27 11:45:38 <pcarrier> technology.
2015-01-27 13:08:40 <vkrishn> what would be best experience, use alpine-xen as base for other guests or plain install with lxc/qemu guests?
2015-01-27 14:08:20 <pcarrier> ncopa: stupid question maybe, but why is af_packet in /etc/modules?
2015-01-27 14:14:11 <ncopa> pcarrier: i don't remember
2015-01-27 14:23:21 <pcarrier> foo:~# apk add man => libgcc libstdc++ groff gawk man
2015-01-27 14:23:24 <pcarrier> seriously?!?! :D
2015-01-27 14:24:13 <pcarrier> I'm gonna give http://mdocml.bsd.lv/ a try :)
2015-01-27 15:32:00 <ncopa> mdocml-man needs sqlite which is slightly unfortunate
2015-01-27 15:32:43 <ncopa> apk trigger is horribly slow when you have many -doc packages installed
2015-01-27 15:33:19 <ncopa> because mdocml does not have good support for syncing the db with the dir tree
2015-01-27 15:33:35 <ncopa> i think we might want use the openbsd man implementation like void linux does
2015-01-27 15:35:04 <pcarrier> isn't that based on mdocml?
2015-01-27 15:35:23 <pcarrier> I thought sqlite3 was optional
2015-01-27 15:36:45 <ncopa> i think not with latest mdocml release
2015-01-27 15:37:15 <pcarrier> hu. that's bad. I thought it was meant to be the new thing(c) for openbsd
2015-01-27 15:42:10 <pcarrier> hu, I'm having weird terminfo issues with alpine
2015-01-27 15:42:31 <ncopa> ok?
2015-01-27 15:42:44 <pcarrier> eg without a git config, git grep shows a bunch of garbage
2015-01-27 15:42:59 <ncopa> does it go via busybox less?
2015-01-27 15:43:08 <ncopa> try apk add less
2015-01-27 15:43:09 <pcarrier> I'm gonna guess so
2015-01-27 15:43:12 <ncopa> and see if that helps
2015-01-27 15:43:20 <pcarrier> well I have a git config :)
2015-01-27 15:43:34 <ncopa> there was a patch to fix busybox less
2015-01-27 15:43:55 <pcarrier> he he no systemtap
2015-01-27 15:44:01 <ncopa> but the submitter did the mistake to mention systemd - so the patch never made it in to busybox :)
2015-01-27 15:44:05 <pcarrier> so what's the policy for main vs testing?
2015-01-27 15:44:28 <pcarrier> lol ok
2015-01-27 15:44:59 <ncopa> http://lists.busybox.net/pipermail/busybox/2014-January/080353.html
2015-01-27 15:45:28 <ncopa> we use testing to verify that the builders build packages successfully
2015-01-27 15:45:46 <pcarrier> there are a lot of packages there
2015-01-27 15:46:15 <ncopa> so we can test that the packages works when build from build servers
2015-01-27 15:46:16 <ncopa> yes
2015-01-27 15:46:24 <ncopa> becuase many are package requests
2015-01-27 15:46:30 <ncopa> the developer doesnt even use it
2015-01-27 15:46:33 <ncopa> but packages it
2015-01-27 15:46:39 <ncopa> then never hears anything about it
2015-01-27 15:46:48 <pcarrier> so they just stay in testing?
2015-01-27 15:46:54 <ncopa> the idea was that they should be moved to main once tested
2015-01-27 15:46:56 <ncopa> yes
2015-01-27 15:47:07 <ncopa> we should autopurge stuff there i suppose
2015-01-27 15:47:10 <pcarrier> just sent a patch for dfc :)
2015-01-27 15:47:27 <pcarrier> should I subscribe to the ML, or are reply headers well maintained?
2015-01-27 15:50:57 <jomat> re
2015-01-27 15:53:16 <pcarrier> /etc/init.d/hwdrivers is... interesting.
2015-01-27 15:54:00 <pcarrier> "# we run it twice so we detect all devices" <--- why would twice be reliably enough?
2015-01-27 15:56:17 <ncopa> i think that is no longer needed
2015-01-27 15:56:33 <ncopa> due to mdev.conf has support for MODALIAS
2015-01-27 15:57:00 <ncopa> earlier there was modalias for host controllers
2015-01-27 15:57:29 <ncopa> before the host controller driver module was loaded no devices attached to that controller was visible
2015-01-27 15:57:35 <ncopa> so we ran it twice
2015-01-27 15:58:00 <pcarrier> /etc/local.d/README and /etc/sysctl.d/README feel like they belong in documentation
2015-01-27 15:58:15 <ncopa> example, if you had an usb disk, but the usb host driver was not loaded
2015-01-27 15:58:43 <ncopa> then wouldnt the modalias for the usb disk be visible til after the usb controller driver was loaded
2015-01-27 15:58:53 <pcarrier> yeah, sure
2015-01-27 15:59:05 <pcarrier> but then, there could be more than one layer
2015-01-27 16:00:28 <ncopa> right, and i think that was solved when mdev got MODALIAS support
2015-01-27 16:01:02 <ncopa> i haven't really put too much effort in cleaning that up because i have some ideas how to solve it with a netlink listener
2015-01-27 16:01:06 <ncopa> a'la udev
2015-01-27 16:01:39 <ncopa> something like this: http://git.r-36.net/nldev/
2015-01-27 16:02:38 <ncopa> what i really want is bb mdev read modalias from stdin
2015-01-27 16:02:49 <ncopa> + a timeout
2015-01-27 16:03:23 <ncopa> so you could have mdev to read the modalias from stdin, line by line, and if nothing happens within 1 sec or so, then just exit
2015-01-27 16:04:04 <ncopa> ok gotta go
2015-01-27 16:04:06 <ncopa> see u
2015-01-27 16:04:17 <pcarrier> tty,
2015-01-27 16:48:20 <plueschopaht_mob> Hi. Is Alpine Linux affected by https://news.ycombinator.com/item?id=8953545 CVE-2015-0235 ?
2015-01-27 16:48:39 <plueschopaht_mob> Seems to be glibc specific, so no problem with musl I guess?
2015-01-27 16:50:11 <darkfader> yeah unless someone copied their NSS code
2015-01-27 16:50:16 <darkfader> and i'd hope not
2015-01-27 16:50:41 <plueschopaht_mob> Can we check?
2015-01-27 17:06:48 <darkfader> i don't got time, i'm getting ready to patch all that other linux shit
2015-01-27 17:07:02 <darkfader> but it seems it's end up with just the enterprise distros being affected?
2015-01-27 17:07:32 <darkfader> blame still to be put on the glibc author, if he didn't love his bugs there would be no need to run an old version
2015-01-27 17:07:39 <darkfader> the bugfix was said to be from 2013
2015-01-27 17:16:03 <plueschopaht_mob> musl all the way :3
2015-01-27 20:33:42 <rowan|b613> where can i find the grsec options for the alpine kernel?
2015-01-27 20:38:45 <xen_roger1> is there nfnl_osf in alpine ?
2015-01-27 21:00:34 <jomat> rowan|b613: Do you mean the kernel config? http://git.alpinelinux.org/cgit/aports/tree/main/linux-grsec/kernelconfig.x86#n5751
2015-01-27 21:06:46 <rowan|b613> jomat: yes that's exactly what i was looking for. thank you!
2015-01-27 21:08:03 <jomat> np :-)
2015-01-27 21:08:20 <crow> Hi i am tryint to create correct coredump, but seems i am meesing something, as coredumpctl does not give me right informations, and i installed -debug packages for all involed applications http://sprunge.us/HiEU
2015-01-27 21:08:21 <rowan|b613> i was trying to use checksec.sh but it doesn't return any values for the kernel.
2015-01-28 06:56:31 <kate> hello! i'm picky and i hate everything. i just wanted to say i'm enjoying my first impressions of alpine linux. good work :)
2015-01-28 07:03:52 <ncopa> morning
2015-01-28 07:04:12 <ncopa> glad you like it
2015-01-28 07:19:25 <kate> nice to see a package manager which doesn't have a million options, too. it feels nice to use. i bet it's written in C, isn't it?
2015-01-28 07:19:41 <kate> ah yes, it doesn't link against a bazillion things
2015-01-28 07:20:01 <fabled> kate, yes, http://git.alpinelinux.org/cgit/apk-tools
2015-01-28 07:20:24 <kate> things feel so much nicer when they're not a pile of scripts
2015-01-28 07:21:25 <kate> what, no dependency on apr or glib? K&R indentation? a reasonable comment-to-code ratio? goodness me
2015-01-28 07:22:26 <kate> C99, but i can understand that :)
2015-01-28 07:22:31 <fabled> kate, it's not perfect. but gets the job in relatively decent manner :)
2015-01-28 07:23:24 <kate> IDB for signed types as bitfields, i think: http://git.alpinelinux.org/cgit/apk-tools/tree/src/dot.c#n21
2015-01-28 07:23:30 <kate> you should probably have those unsigned
2015-01-28 07:26:28 <kate> gccism here; arithmetic on void *: http://git.alpinelinux.org/cgit/apk-tools/tree/src/hash.c#n79
2015-01-28 07:26:37 <kate> serves you right for hiding it behind a typedef :)
2015-01-28 07:28:01 <kate> unchecked malloc (and possibly strdup?): http://git.alpinelinux.org/cgit/apk-tools/tree/src/blob.c#n24
2015-01-28 07:28:44 <systmkor> kate, yah its pretty much written in pure c
2015-01-28 07:29:24 <kate> no need to cast void *: http://git.alpinelinux.org/cgit/apk-tools/tree/src/del.c#n127
2015-01-28 07:30:47 <systmkor> kate, wanna do a code audit of it
2015-01-28 07:31:02 <systmkor> and create 100% unit test coverage?
2015-01-28 07:31:10 <systmkor> i would be down to join in
2015-01-28 07:31:59 <kate> no thank you :)
2015-01-28 07:32:04 <systmkor> kate, fuck...
2015-01-28 07:32:06 <kate> it's not bad
2015-01-28 07:32:07 <systmkor> owell
2015-01-28 07:32:27 <systmkor> it isn't but until you actually prove that, it's just at a glance looks goood
2015-01-28 07:32:45 <kate> you can't prove something is okay :) you can only disprove a hypothesis
2015-01-28 07:32:52 <kate> a good start would be to compile it with something other than gcc
2015-01-28 07:33:04 <systmkor> that would be a good start
2015-01-28 07:33:15 <systmkor> also having 100% unit test coverage is another good thing to do
2015-01-28 07:33:30 <systmkor> having a formal spec is also a critical part of validating code
2015-01-28 07:33:36 <kate> perhaps. i think unit tests can be a false economy
2015-01-28 07:33:54 <systmkor> ....
2015-01-28 07:34:10 <systmkor> just like debugging code, yup for sure :|
2015-01-28 07:34:11 <systmkor> :P
2015-01-28 07:34:21 <systmkor> kate, how so?
2015-01-28 07:35:39 <kate> systmkor, i don't have a concise explanation, sorry. maybe i'll try to write about it sometime. i've found a few different kinds of negative effect, and they seem to outweigh the positives, imo
2015-01-28 07:36:06 <systmkor> well in a general sense
2015-01-28 07:36:31 <systmkor> because code has been developed without unit testing for o what the past 60 years producing some of the crappiest of things to ever be produced by keyboards
2015-01-28 07:36:49 <systmkor> and unit-tests have finally started to have some semblance of verified code
2015-01-28 07:36:57 <systmkor> I'm by no means saying it is sufficient
2015-01-28 07:37:52 <kate> i've seen some spectacularly awful unit tested code, too. i don't think the presence of unit tests is the deciding factor; i think it's more to do with the mentality of the people developing it
2015-01-28 07:42:58 <kate> unit tests won't find you things like the items i just pointed out, for example
2015-01-28 07:44:45 <systmkor> again yah it isn't sufficient
2015-01-28 07:45:15 <systmkor> but unit testing starts to push coding community to make less excuses for producing good code
2015-01-28 07:45:31 <kate> i don't think so. i think it gives the illusion of that
2015-01-28 07:45:32 <systmkor> and I think becomes even more salient if their unit tests are terrible
2015-01-28 07:45:50 <systmkor> inately giving the illusion I would disagree
2015-01-28 07:45:55 <kate> i also think people become complacent, when they're under the impression that such tests will catch things
2015-01-28 07:46:01 <systmkor> can it create an illusion from a sub-group of the population
2015-01-28 07:46:02 <systmkor> yah
2015-01-28 07:46:19 <systmkor> so no different than before
2015-01-28 07:46:29 <systmkor> so i don't think it creates any new illusions
2015-01-28 07:46:42 <systmkor> but if you glance at a unit test and it just say
2015-01-28 07:46:45 <systmkor> return True
2015-01-28 07:46:55 <systmkor> *just says
2015-01-28 07:47:20 <systmkor> then it is very indicative that the people/person producing the code
2015-01-28 07:47:40 <systmkor> has little to no care about code quality or good design
2015-01-28 07:48:29 <kate> i can see that well enough from the code itself :)
2015-01-28 07:48:36 <systmkor> if you are going to spend the time to verify your code, and in a reproducable way not just putting in print flags and doing casual manual checks
2015-01-28 07:48:51 <systmkor> then you are effectively already doing unit testing
2015-01-28 07:49:00 <systmkor> kate, you can sorta see it from the code
2015-01-28 07:49:08 <systmkor> but until you have either formally proven that
2015-01-28 07:49:11 <systmkor> or actually tested it
2015-01-28 07:49:26 <systmkor> it just looks good from a higher view
2015-01-28 07:49:44 <systmkor> because even if the code is correct
2015-01-28 07:49:54 <systmkor> that assumes everything underneath it is correct as well
2015-01-28 07:50:40 <systmkor> not to mention not having any communal test bed for a projects developers and anyone who wants to package is
2015-01-28 07:50:53 <systmkor> *package it
2015-01-28 07:50:56 <kate> i didn't mean regarding correctness - i meant quite a few aspects, none of which unit tests can cover: decent design, simplicity, and good use of the language it's written in, for example
2015-01-28 07:51:16 <systmkor> agreed
2015-01-28 07:51:28 <systmkor> well sorta
2015-01-28 07:51:43 <systmkor> you can do unit tests of lexical/syntactic analysis of your code
2015-01-28 07:51:54 <systmkor> for general trends of good code quality
2015-01-28 07:52:23 <pcarrier> I am so very confused
2015-01-28 07:52:26 <pcarrier> foo:~$ sudo sysctl kernel.grsecurity.chroot_deny_chmod=0
2015-01-28 07:52:27 <pcarrier> sysctl: error: 'kernel.grsecurity.chroot_deny_chmod=0' is an unknown key
2015-01-28 07:52:29 <pcarrier> foo:~$ sudo sysctl -a|grep chmod
2015-01-28 07:52:31 <pcarrier> kernel.grsecurity.chroot_deny_chmod = 1
2015-01-28 07:52:55 <kate> that sounds like sysctl is parsing "kernel.grsecurity.chroot_deny_chmod=0" as the key name
2015-01-28 07:52:59 <systmkor> yah
2015-01-28 07:53:06 <systmkor> i've had that problem before
2015-01-28 07:53:14 <pcarrier> yes...
2015-01-28 07:53:47 <pcarrier> well echo 0|sudo tee /proc/sys/kernel/grsecurity/chroot_deny_chmod works so moving on for now
2015-01-28 07:54:04 <systmkor> if it works :P
2015-01-28 07:55:00 <systmkor> pcarrier, https://imgur.com/TwLiMoh
2015-01-28 07:55:01 <pcarrier> gnnn though.
2015-01-28 07:55:41 <systmkor> kate, not to mention unit tests (general notion of automated testing) at least provide some semblance of a central point
2015-01-28 07:55:41 <pcarrier> I mean I have 2 options: get sh*t done or yakshave.
2015-01-28 07:55:57 <systmkor> people can check and validate and contribute to verify code
2015-01-28 07:55:58 <pcarrier> given this is edge, I'll just get sh*t done
2015-01-28 07:56:04 <systmkor> nice
2015-01-28 07:56:33 <systmkor> kate, and not for end users/developers to produce their own tests which just is repeated work without really moving
2015-01-28 07:56:38 <systmkor> a code quality forward
2015-01-28 07:57:38 <kate> systmkor, i think regression tests would be more useful for that
2015-01-28 07:57:58 <kate> systmkor, particularly if you'd like a formal spec
2015-01-28 07:58:10 <systmkor> kate, those as far as I understand it tests more if the code is performing
2015-01-28 07:58:15 <systmkor> any worse or better than before
2015-01-28 07:58:18 <systmkor> which is important
2015-01-28 07:58:33 <kate> no, performance isn't usually what you'd test for
2015-01-28 07:58:41 <pcarrier> is there a guide to go from a blank image to a bootable system?
2015-01-28 07:58:59 <systmkor> pcarrier, what do you mean by a blank image
2015-01-28 07:59:04 <kate> pcarrier, i just ran the install script and pressed enter a lot
2015-01-28 07:59:20 <systmkor> pcarrier, http://wiki.alpinelinux.org/wiki/Installation
2015-01-28 07:59:32 <systmkor> there is a starting point
2015-01-28 07:59:42 <systmkor> kate, then what would you be testing for
2015-01-28 07:59:47 <pcarrier> let's say I have a live alpine sstem
2015-01-28 07:59:51 <systmkor> yup
2015-01-28 07:59:55 <pcarrier> and I want to make a harddrive bootable :)
2015-01-28 08:00:10 <pcarrier> or in my case a loop device
2015-01-28 08:00:21 <systmkor> you mean the HDD is in the system you are running?
2015-01-28 08:00:35 <systmkor> pcarrier, i believe it's in the installation structions
2015-01-28 08:00:35 <pcarrier> well it's a 1GB file here
2015-01-28 08:01:00 <pcarrier> and I'm partitioning myself and co
2015-01-28 08:01:07 <systmkor> this isn't for all cases but you need the boot partition to have the boot flag, and you need to have installed syslinux
2015-01-28 08:01:54 <systmkor> pcarrier, would this be closer to similar approach of what you are doing
2015-01-28 08:01:55 <systmkor> http://wiki.alpinelinux.org/wiki/Installing_Alpine_on_HDD_dualbooting
2015-01-28 08:02:58 <pcarrier> LOL
2015-01-28 08:03:00 <pcarrier> foo:/# rc-update add urandom boot
2015-01-28 08:03:01 <pcarrier> Segmentation fault
2015-01-28 08:03:20 <systmkor> pcarrier, document and send email plz :D
2015-01-28 08:03:48 <pcarrier> will do
2015-01-28 08:03:54 <pcarrier> send email where?
2015-01-28 08:04:00 <pcarrier> open a bug or?
2015-01-28 08:04:27 <systmkor> kate, also if you are writing unit tests to your spec as you should and modify code and it fails that's a form of regression testing
2015-01-28 08:04:31 <systmkor> pcarrier, http://alpinelinux.org/community/
2015-01-28 08:04:34 <systmkor> devel mailing list
2015-01-28 08:04:40 <systmkor> pcarrier, I would also open a bug
2015-01-28 08:04:54 <systmkor> pcarrier, http://bugs.alpinelinux.org/projects/alpine/issues
2015-01-28 08:05:09 <kate> systmkor, then i would not call those unit tests
2015-01-28 08:05:16 <systmkor> but they would be
2015-01-28 08:05:21 <systmkor> you are testing a unit piece of a code
2015-01-28 08:05:37 <systmkor> and which goes from small mechanism in code
2015-01-28 08:05:46 <pcarrier> bug it is
2015-01-28 08:05:55 <systmkor> to larger method calls or chunks with their own tests
2015-01-28 08:06:09 <systmkor> that's unit tests, or everywhere I have seen it
2015-01-28 08:06:52 <kate> systmkor, i'm sorry if this seems a little short, but i've had this conversation lots and lots of times. it always gets increasingly academic, and further detatched from what i think actually works in practice. if you want to have that theoretical discussion, go ahead - but that's not a topic for me :)
2015-01-28 08:07:16 <systmkor> we can keep it non theoretical
2015-01-28 08:07:19 <systmkor> i'm just saying
2015-01-28 08:07:26 <ncopa> :)
2015-01-28 08:08:18 <systmkor> if you consider regression tests test to be testing an older code base or last known working version and then the modified version of code
2015-01-28 08:08:29 <systmkor> and see what breaks or works regression testing
2015-01-28 08:08:36 <systmkor> you get that by virtue of running unit tests
2015-01-28 08:08:50 <systmkor> every time you make chunk of code change
2015-01-28 08:08:59 <systmkor> assuming your unit tests are existing there
2015-01-28 08:09:32 <systmkor> ncopa, sorry about the not propper email thread for one of the ports, my bad will fix within the next day or so
2015-01-28 08:10:01 <ncopa> apk has a "test" mode
2015-01-28 08:10:10 <systmkor> kate, whatever you want to call it unit tests, regression tests, my little pwny tests
2015-01-28 08:10:15 <ncopa> http://git.alpinelinux.org/cgit/apk-tools/tree/src/apk.c#n42
2015-01-28 08:10:27 <ncopa> which runs some kind of regression testing
2015-01-28 08:10:35 <systmkor> i'm just trying to say automated testing is necessary for showing code works
2015-01-28 08:10:48 <systmkor> ncopa, cool
2015-01-28 08:11:22 <ncopa> i wouldnt mind if someone with unit/regression testing experience improved it ;)
2015-01-28 08:11:36 <systmkor> ncopa, did you get a chance to read the ARM  build email in the devel mailing list?
2015-01-28 08:12:13 <systmkor> ncopa, I may start trying to do that. I'm very slowly wanting to make my port of Nix and starting with something less complex would help me get a better understanding of package management
2015-01-28 08:13:24 <ncopa> the email about wandbord boot image?
2015-01-28 08:14:36 <systmkor> yah
2015-01-28 08:15:27 <systmkor> ncopa, I was hoping there could be disussion about it because I'm using Alpine for the start-up I co-founded and will be using Alpine on the arm boards we are using
2015-01-28 08:15:45 <ncopa> cool
2015-01-28 08:15:48 <ncopa> well
2015-01-28 08:15:54 <systmkor> and have been spending time researching ways of doing cross compiling
2015-01-28 08:16:00 <systmkor> automated building of distros
2015-01-28 08:16:08 <ncopa> i have very little experience of arm in general
2015-01-28 08:16:20 <ncopa> i got 2 wandbord boxes the last week
2015-01-28 08:16:26 <systmkor> nice
2015-01-28 08:16:28 <ncopa> and i have a rpi
2015-01-28 08:16:51 <ncopa> so i am supposed to have a look at that soonish
2015-01-28 08:17:01 <fabled> systmkor, we tried cross-build and deemed it too much work
2015-01-28 08:17:10 <fabled> so ended up using wandboards doing native builds
2015-01-28 08:17:26 <kate> ncopa, did you see those few comments i made from briefly looking at the apk code a moment ago? are you the right person to see them?
2015-01-28 08:17:41 <systmkor> fabled, well you have to cross compile at some point
2015-01-28 08:17:42 <ncopa> fabled is the right person for apk-tools
2015-01-28 08:17:51 <systmkor> without a fuck ton of work
2015-01-28 08:17:53 <kate> ah
2015-01-28 08:17:53 <fabled> kate, i wrote most of the apk c code
2015-01-28 08:18:02 <kate> fabled, see above :)
2015-01-28 08:18:35 <ncopa> i suspect the best way to start a discussion or suggest things is with a git patch ;)
2015-01-28 08:18:42 <fabled> systmkor, the problem is that most people don't know how to do proper configure, and build system for cross-compile
2015-01-28 08:18:46 <fabled> so you end up fixing many of those
2015-01-28 08:18:56 <systmkor> fabled, yah i have been searching around
2015-01-28 08:19:09 <systmkor> and it just likes like going from one garbage heap to another
2015-01-28 08:19:16 <ncopa> :)
2015-01-28 08:19:25 <systmkor> but I'm trying to think how do we have a full stack automated build
2015-01-28 08:19:31 <systmkor> native or cross
2015-01-28 08:19:36 <ncopa> cross compile is non-trivial
2015-01-28 08:19:49 <systmkor> very much so
2015-01-28 08:20:15 <systmkor> I mean cross more just to at least get a base system working
2015-01-28 08:20:48 <fabled> yes, porting to new platforms include cross-compiling the core
2015-01-28 08:20:58 <fabled> and i did that for our armhf build
2015-01-28 08:21:12 <systmkor> well okay so pausing on different platforms at the moment
2015-01-28 08:21:26 <systmkor> is there any existing code, designs, or projects
2015-01-28 08:21:45 <systmkor> to automate doing a full build of Alpine with
2015-01-28 08:21:58 <fabled> what i used earlier is at http://dev.alpinelinux.org/~tteras/bootstrap/
2015-01-28 08:22:01 <systmkor> easily swapping out configs, what programs I want on that
2015-01-28 08:22:05 <systmkor> image
2015-01-28 08:22:17 <systmkor> and producing isos, and img, or u-boot img etc.
2015-01-28 08:22:47 <fabled> making u-boot / img is tricky as it is in most cases board specific
2015-01-28 08:22:56 <fabled> most arm boards have different boot sequence
2015-01-28 08:23:15 <systmkor> well yes for it to work specifically that board it needs customization
2015-01-28 08:23:24 <systmkor> but it should require an entirely new build process
2015-01-28 08:23:29 <systmkor> from top to bottom
2015-01-28 08:23:56 <systmkor> in a general algorithmic sense
2015-01-28 08:27:31 <systmkor> fabled, this is really new to me so I don't really even have a full picture of building a distro
2015-01-28 08:31:11 <systmkor> fabled, ncopa anything you recommend I should read, code or papers etc.?
2015-01-28 08:33:37 <kate> i recommend Johnathan Livingstone, Seagull. by Richard Bach. it's a pretty good book
2015-01-28 08:34:23 <ncopa> :)
2015-01-28 08:35:23 <ncopa> i have been waiting for: cross compiling - the movie
2015-01-28 08:36:02 <systmkor> lol
2015-01-28 08:36:08 <kate> *movie voice* ... in a world... where host and target... do not match...
2015-01-28 08:36:26 <systmkor> kate, looks like a good book, will put on to do list
2015-01-28 08:36:41 <systmkor> kate, but hoping for more related to automating the build of a distro
2015-01-28 08:36:53 <kate> it's a book about how now to have to-do lists
2015-01-28 08:36:58 <kate> how not*
2015-01-28 08:37:00 <kate> :)
2015-01-28 08:37:02 <fabled> "LLVM - when single compiler can build to all targets"
2015-01-28 08:37:21 <fabled> waiting for that to happen
2015-01-28 08:37:47 <systmkor> kate, :P to-do list is just my catch all term for communicating with other people
2015-01-28 08:38:11 <systmkor> on things I want to do
2015-01-28 08:41:05 <pcarrier> is there a simple way to restore permissions for all packages with apk?
2015-01-28 08:41:17 <fabled> pcarrier, "apk fix <pkg>"
2015-01-28 08:41:27 <fabled> oh, all
2015-01-28 08:41:44 <fabled> full reinstall
2015-01-28 08:41:56 <fabled> "apk fix -rd $(cat /etc/apk/world)"
2015-01-28 08:42:41 <pcarrier> thanks
2015-01-28 08:42:59 <pcarrier> and a way to show the post-install scripts for a package?
2015-01-28 08:44:02 <fabled> mm... not sure if there's clear command for it. for installed packages you could do "tar tvf /lib/apk/db/scripts.tar" if poking internals is ok
2015-01-28 08:44:13 <pcarrier> it's installed
2015-01-28 08:45:51 <systmkor> fabled, so anything I should checkout or look at?
2015-01-28 08:46:35 <fabled> i don't have any recommendations. i've sucked my knowledge throughout the years from various sources
2015-01-28 08:46:52 <fabled> if you want to go with cross-compile, then maybe buildroot is good starting point
2015-01-28 08:47:11 <systmkor> m'kay I've started looking at that and the yocto project
2015-01-28 08:47:28 <systmkor> will keep at that
2015-01-28 08:47:36 <fabled> yocto might be good too
2015-01-28 08:47:46 <fabled> i haven't looked into it in detail
2015-01-28 08:47:58 <systmkor> m'kay I was looking at it and there may be a decent amount of stuff we need to gut out
2015-01-28 08:48:06 <systmkor> but they do have a meta-musl
2015-01-28 08:48:10 <fabled> but yes, if you need fast build times
2015-01-28 08:48:15 <systmkor> layer or whatever they call it
2015-01-28 08:48:18 <fabled> cross compile is pretty much the only way to do it atm
2015-01-28 08:48:26 <fabled> and it might need lot of effort for some packages
2015-01-28 08:48:41 <systmkor> well if I'm natively building
2015-01-28 08:48:55 <systmkor> is there a good reference for automating that process
2015-01-28 08:54:51 <ncopa> abuild? :)
2015-01-28 08:57:55 <systmkor> ncopa, well maybe full discussion later date
2015-01-28 08:58:07 <systmkor> but abuild can build the entire distro?
2015-01-28 08:58:14 <systmkor> from top to bottom
2015-01-28 08:58:49 <systmkor> ncopa, and then either produce an image or hooks in nicely for something to do that?
2015-01-28 09:00:38 <pcarrier> wait wut
2015-01-28 09:01:04 <systmkor> ncopa, what I'm thinking is basically starting from a set of configurations that would through various steps
2015-01-28 09:01:27 <systmkor> compile, build, etc. to a final point
2015-01-28 09:01:30 <systmkor> or image
2015-01-28 09:02:55 <systmkor> basically a full automation of building Alpine from a set of configs to an ISO
2015-01-28 09:02:55 <systmkor> but thinking more modular and generically
2015-01-28 09:04:53 <systmkor> ncopa, I'm hoping I'm making at least some sense
2015-01-28 09:04:53 <kate> systmkor, sounds like you're after an equivalent of these, for Alpine: https://code.google.com/p/freebsd-wifi-build/wiki/GettingStarted
2015-01-28 09:06:03 <systmkor> kate, that would be near what I'm thinking
2015-01-28 09:06:15 <systmkor> really close but I don't know if exactly what I'm thinking
2015-01-28 09:06:19 <systmkor> kate, thanks :D
2015-01-28 09:06:24 <systmkor> well i need to head to bed ciao
2015-01-28 09:06:39 <kate> it's pretty simple
2015-01-28 09:07:24 <kate> it's just a shell script which is fed a few variables, stating which toolchain to use, where to build, and what to build
2015-01-28 16:20:07 <clandmeter> tvvdhSL5hEBn: have you been using zfs on alpine?
2015-01-28 16:23:23 <tvvdhSL5hEBn> not recently no
2015-01-28 16:24:36 <tvvdhSL5hEBn> the licensing issue with libtirpc was kinda discouraging
2015-01-28 18:57:48 <systmkor> tvvdhSL5hEBn, clandmeter isn't there an open port of ZFS?
2015-01-28 18:58:17 <systmkor> also I don't know but is BTRFS being pushed to replace ZFS?
2015-01-28 19:14:09 <tvvdhSL5hEBn> the open port is zfsonlinux, and it is adapted to glibc, which comes with an internal rpc lib, while musl does not. the one musl brings cannot link per license to zfs which is cddl
2015-01-28 19:17:46 <systmkor> ah
2015-01-28 19:17:58 <systmkor> viva la btrfs ? :P
2015-01-28 19:20:46 <kate> can apk tell you which package a file came from?
2015-01-28 19:27:08 <systmkor> i don't know
2015-01-28 19:27:33 <systmkor> you may already know this but it can list the files installed by package
2015-01-28 19:27:56 <systmkor> so you could just iterate over all installed packages with that and take not which ones match
2015-01-28 19:28:19 <kate> that'll do
2015-01-28 19:34:32 <fabled> kate, "apk info -W"
2015-01-28 19:36:45 <kate> oh great. thanks :)
2015-01-28 19:49:03 <systmkor> fabled, do you have any design docs for apk-tools?
2015-01-28 19:49:56 <fabled> systmkor, i have some random notes.
2015-01-28 19:50:17 <fabled> it sort of evolved from the original script set ncopa had for early versions of alpine
2015-01-28 19:50:36 <systmkor> k
2015-01-28 19:50:52 <systmkor> it's more of just high-level design
2015-01-28 19:51:01 <systmkor> is my interest at moment
2015-01-28 19:51:33 <fabled> no
2015-01-28 19:51:40 <fabled> that is, no such document
2015-01-28 19:51:55 <fabled> we did discuss several aspects of it here
2015-01-28 19:52:10 <fabled> but it was never collected as a doc
2015-01-28 19:52:13 <systmkor> back in the irc logs?
2015-01-28 19:52:20 <fabled> i think it predates the logs
2015-01-28 19:52:24 <systmkor> ah
2015-01-28 19:52:35 <systmkor> well i'll start reading through the code and make notes and post questions
2015-01-28 19:52:41 <fabled> sure
2015-01-28 19:52:48 <systmkor> and i guess reverse engineer the design :P
2015-01-28 19:52:51 <fabled> :)
2015-01-28 19:53:01 <fabled> well
2015-01-28 19:53:02 <systmkor> fabled, thanks :D
2015-01-28 19:53:09 <fabled> the major parts in simplicity
2015-01-28 19:53:25 <fabled> 1) database is set of text files, and tar archive for scripts
2015-01-28 19:53:37 <fabled> we wanted stuff that is editable by vi if needed
2015-01-28 19:53:43 <fabled> or other standard tools
2015-01-28 19:54:09 <fabled> and it needed to be one or just few files (vs. multiple files per pkg as in pretty much all other pkg managers)
2015-01-28 19:54:20 <fabled> due to tmpfs doing allocs on page basis
2015-01-28 19:54:37 <fabled> 2) index, is basically simplified version of database
2015-01-28 19:55:15 <fabled> 3) package is concatenation of three tar.gz files
2015-01-28 19:55:31 <fabled>  in the order of 'signatures', 'meta data' and 'data files'
2015-01-28 19:56:01 <fabled>  (again we wanted stuff that is extractable with standard tools if needed, but still contains everything in a way that apk read/parse fast)
2015-01-28 19:56:15 <systmkor> okay cool, heading out to lunch be back shortly
2015-01-28 19:56:21 <fabled> 4) the solver - is basically a constraint solver
2015-01-28 19:56:39 <fabled> and then the cli to tie it all in
2015-01-28 19:56:47 <fabled> feel free to post questions
2015-01-28 19:56:50 <fabled> i'll try to answer up
2015-01-28 19:56:53 <fabled> now i'm of to sleep
2015-01-28 22:27:07 <kate> what was the last version of alpine which used uclibc?
2015-01-28 22:32:30 <systmkor> kate, 2.6?
2015-01-28 22:32:52 <systmkor> kate, are you wanting to compile and compare against MUSL?
2015-01-28 22:37:57 <kate> effectively, yes
2015-01-28 22:46:26 <kate> actually i'm porting something to as many different (hosted) libc families as i can realistically get my hands on. so other suggestions welcome, too :)
2015-01-28 23:29:37 <BitL0G1c> kate: simplest way to compare alpine uclibc / alpine musl will be in 2 LXC containers
2015-01-28 23:31:38 <kate> oh, i'm not bothered about how to run them :) i do all my development in VMs, anyway
2015-01-28 23:39:37 <kaniini> interesting
2015-01-28 23:39:49 <kaniini> kate: 2.7
2015-01-28 23:39:58 <kaniini> kate: but alpine's branch of uclibc was extensively modified
2015-01-28 23:42:06 <kate> ah, thank you
2015-01-28 23:57:24 <systmkor> kate, what libcs are you compiling with
2015-01-28 23:57:38 <systmkor> dietlibc, uClibc, bionic, glibc, musl, ?
2015-01-28 23:58:50 <systmkor> kate, or plan9'ss libc
2015-01-28 23:59:55 <kate> for linux, currently just eglibc and glibc
2015-01-29 00:01:05 <systmkor> cool
2015-01-29 00:04:36 <systmkor> kate, have you tried using tcc in conjunction with eglibc?
2015-01-29 00:05:30 <kate> do you mean tinycc?
2015-01-29 00:05:44 <kate> the software i'm porting is itself a compiler
2015-01-29 00:08:04 <systmkor> tcc aka tinycc
2015-01-29 00:08:06 <kate> it doesn't matter which compiler is used to build it. the reason the libc matters is quite involved; briefly, this compiler needs to map its own representation of standard APIs onto the libc headers (which is something other compilers do not need to do). part of that means simply being able to parse the libc headers - which is a problem for many headers which have all sorts of gccisms
2015-01-29 00:08:20 <systmkor> i didn't know you were porting a compiler
2015-01-29 00:08:26 <systmkor> kate, what compier custom?
2015-01-29 00:08:40 <kate> the compiler used to build this isn't the compiler which is exposed to those headers - so it doesn't make much difference. but yes, it builds with tinycc
2015-01-29 00:08:54 <kate> it's called tendra
2015-01-29 00:08:54 <systmkor> yah
2015-01-29 00:09:46 <systmkor> cool never heard of that one before
2015-01-29 00:09:56 <systmkor> kate, what are the benefits of tendra over others
2015-01-29 00:10:52 <kate> strictness with regards to API conformance, mainly
2015-01-29 00:11:05 <systmkor> interesting
2015-01-29 00:11:08 <kate> if you're interested, this is a decent overview: http://dioptre.org/tmp/tendra-doc0/tendra-doc/doc/papers/porting/single.xhtml
2015-01-29 00:11:42 <kate> or: http://dioptre.org/tmp/errors.html
2015-01-29 00:12:24 <systmkor> cool
2015-01-29 00:12:51 <kate> not many people think it is cool
2015-01-29 00:14:57 <systmkor> why?
2015-01-29 00:17:57 <kate> well, not many people care about writing correct or portable software - as long as it appears to work on their particular machine
2015-01-29 00:21:31 <kate> i suppose there is little incentive for diversity
2015-01-29 00:25:43 <systmkor> well there is also a difference between correct to an API and secure code
2015-01-29 00:25:58 <systmkor> not implying that glibc is secure
2015-01-29 00:26:52 <kate> i meant correct with regards to API conformance - which glibc certianly is not
2015-01-29 00:27:26 <systmkor> yah i do think API conformance code is important but more of a a means of validating other libcs
2015-01-29 00:27:40 <systmkor> not necessarily implementation code
2015-01-29 00:27:56 <kate> well, read that article, if you want to see what that's about
2015-01-29 00:28:15 <systmkor> kate, it would be cool to see discrepancies between Tendra and MUSL
2015-01-29 00:28:32 <kate> sure. i can give you a list when i'm done. it's not many
2015-01-29 00:28:39 <systmkor> oh cool
2015-01-29 00:28:57 <systmkor> I would be down to make that a wiki page on Alpine
2015-01-29 00:29:02 <kate> i'm only catering for the C89 and POSIX APIs currently, though
2015-01-29 00:29:09 <systmkor> that's a start
2015-01-29 00:29:15 <systmkor> nothing wrong inherently about that
2015-01-29 00:29:15 <kate> i don't think there's much point in making a wiki page about that
2015-01-29 00:29:38 <systmkor> kate, well the part I would say is that to show more new people to Alpine
2015-01-29 00:29:56 <systmkor> that Libcs can very
2015-01-29 00:30:07 <kate> well tendra's entire userbase is currently talking to you
2015-01-29 00:30:09 <systmkor> and also how that could affect security
2015-01-29 00:30:22 <systmkor> lol hey but I care about security and code correctness
2015-01-29 00:30:45 <kate> if you care about those things, then you might also be interested in API conformance
2015-01-29 00:30:48 <systmkor> and *fingers crossed* I may have the money and influence one day that things like this carry more coding community weight
2015-01-29 00:31:18 <systmkor> kate, well I was talking with dalias from MUSL about how it would be cool write a formal "libc" oracle
2015-01-29 00:31:30 <systmkor> that you use only as a in-depth fuzzing and validation tool
2015-01-29 00:32:12 <kate> hm
2015-01-29 00:32:18 <kate> read that article
2015-01-29 00:32:29 <systmkor> will but can't at moment
2015-01-29 00:32:34 <dalias> kate, have you looked at libc-test repo?
2015-01-29 00:32:45 <dalias> we'd like to get more conformance tests added to it
2015-01-29 00:32:56 <kate> dalias, i haven't
2015-01-29 00:35:20 <kate> dalias, very nice
2015-01-29 00:35:31 <kate> pity it mixes C89 and C99
2015-01-29 00:35:33 <dalias> find it ok?
2015-01-29 00:36:12 <dalias> the only part that should be specific to c89 vs c99 is header requirements
2015-01-29 00:36:13 <systmkor> dalias kate but what about C11 :P
2015-01-29 00:36:25 <dalias> interface behavior is generally unchanged
2015-01-29 00:37:15 <kate> dalias, i wasn't talking about testing earlier, incidentally. but i am also interested in testing the implementation of any give libc. but that has nothing to do with the kind of conformance i mentioned a moment ago
2015-01-29 00:37:31 <dalias> kate, what in particular are you interested in?
2015-01-29 00:37:41 <dalias> i can tell you if i know any existing resources for it
2015-01-29 00:39:01 <kate> dalias, for testing? there are two aspects; both validation. 1. for language conformance (as per the plum hall suite); 2. for library implementation conformance (as per your libc-test suite)
2015-01-29 03:14:49 <thiagowfx> hello people, I'm trying out 'simple' distros right now
2015-01-29 03:14:56 <thiagowfx> alpine will be my next one
2015-01-29 03:15:17 <thiagowfx> question: does it support xorg and graphical apps in general?
2015-01-29 03:15:28 <dalias> yes but not a lot of desktop apps are packaged
2015-01-29 03:15:33 <dalias> x works fine tho
2015-01-29 03:16:13 <dalias> xfce and firefox are packaged
2015-01-29 03:16:37 <dalias> and gimp, inkscape, abiword, ....
2015-01-29 03:17:03 <thiagowfx> nice. I'll probably test it as a server, but I'll want a decent web browser and basic applications, nothing fancy
2015-01-29 03:17:06 <dalias> i don't have a full list. there's definitely enough to get you through the basics but your favorite app in a particular category might not be packaged yet
2015-01-29 03:17:13 <systmkor> I think awesome and MATE
2015-01-29 03:17:15 <systmkor> are also packaged
2015-01-29 03:17:31 <dalias> e.g. i'd rather have chromium :)
2015-01-29 03:17:45 <systmkor> dalias, is that working yet?
2015-01-29 03:17:54 <thiagowfx> is i3wm packaged?
2015-01-29 03:18:04 <systmkor> thiagowfx, not yet
2015-01-29 03:18:14 <systmkor> I have tried but there are like one or two issues with building it
2015-01-29 03:18:32 <systmkor> that are dependent on glibc if I remember correctly
2015-01-29 03:18:33 <dalias> supposedly someone's built chromium successfully but iirc a few things, possibly some of the sandboxing, had to be turned off
2015-01-29 03:18:57 <systmkor> dalias, well then I would say not entirely successful yet then
2015-01-29 03:18:58 <systmkor> but awesome
2015-01-29 03:19:01 <dalias> systmkor, if you have a list of failures and want comments on how to fix them i can take a brief look
2015-01-29 03:19:22 <thiagowfx> hum, I think I would be interested in trying to package it, just for the fun
2015-01-29 03:19:25 <dalias> systmkor, iirc there was a sandbox issue but it might have been caused by grsec rather than musl/alpine
2015-01-29 03:19:35 <thiagowfx> I come from Arch Linux, I read alpine`s package manager is heavily inspired on it
2015-01-29 03:19:39 <thiagowfx> APKBUILDs and such
2015-01-29 03:21:52 <systmkor> dalias, you mean about i3(wm)?
2015-01-29 03:22:49 <dalias> systmkor, yeah, or any other software there's demand for
2015-01-29 03:23:07 <systmkor> m'kay cool cool
2015-01-29 03:24:13 <thiagowfx> this muslglibc thing (I probably typed it wrong), is it much different from glibc? I mean, the end user can perceive any differences?
2015-01-29 03:25:03 <thiagowfx> musl libc*
2015-01-29 03:25:25 <systmkor> thiagowfx, will they notice not really
2015-01-29 03:25:33 <systmkor> but similar to like a safer car
2015-01-29 03:25:52 <systmkor> you don't really notice between a unsafe car and a safe car while everything is going well
2015-01-29 03:25:57 <systmkor> but as soon as you hit a brick wall
2015-01-29 03:26:00 <thiagowfx> I found this page -> http://www.etalabs.net/compare_libcs.html
2015-01-29 03:26:09 <systmkor> yup
2015-01-29 03:26:30 <thiagowfx> let me see if I understand
2015-01-29 03:26:32 <systmkor> I don't know if there has been any another studies done. It would be nice to have a automated means of continually comparing
2015-01-29 03:26:51 <thiagowfx> those recent vulnerabilities? heartbleed, gcc
2015-01-29 03:26:58 <thiagowfx> was alpine affected by them?
2015-01-29 03:27:19 <systmkor> heartblead was as far as I know more pertenant to OpenSSLs written code
2015-01-29 03:27:45 <systmkor> for gcc I don't know it depends if it is strictly apart of the compiler (i.e. gcc) or if was from the compiler using a libc (i.e. MUSL)
2015-01-29 03:27:56 <systmkor> thiagowfx, good questions though
2015-01-29 03:30:47 <systmkor> thiagowfx, I wish I had a better paper or resources to describe in depth why there is a difference
2015-01-29 03:40:21 <thiagowfx> final question before I go to install it
2015-01-29 03:40:46 <thiagowfx> can I do a dd if=...alpineiso of=/my/usbflashdrive to boot from alpine?
2015-01-29 03:41:06 <thiagowfx> this isnt documented in the wiki in this way, it recommends using setup scripts
2015-01-29 03:42:01 <thiagowfx> setup-bootable to be more precise
2015-01-29 03:44:27 <dalias> thiagowfx, which gcc issue?
2015-01-29 03:44:37 <dalias> if you're thinking of GHOST that's glibc not gcc
2015-01-29 03:44:44 <dalias> and musl is not affected
2015-01-29 03:45:23 <thiagowfx> yep, ghost
2015-01-29 03:45:26 <thiagowfx> glad to hear that
2015-01-29 03:45:35 <dalias> to be fair though we've also had some dns-related vulns in the past, but they depended on malicious nameservers or spoofing malicious replies
2015-01-29 03:46:32 <dalias> as opposed to ghost which is triggered by attempting to lookup a bad name
2015-01-29 03:50:19 <thiagowfx> thanks people, im going to install alpine now, bye
2015-01-29 09:27:52 <ginjachris> Morning all!
2015-01-29 09:28:23 <ginjachris> It seems the whole world is panicking about this glibc GHOST vulnerability (CVE-2015-0235).  Except, of course, the sensible users of Alpine linux :)
2015-01-29 09:42:03 <fabled> :)
2015-01-29 09:42:24 <fabled> musl is of course unaffected. and surprisingly uclibc is unaffected too
2015-01-29 09:48:11 <ginjachris> awesome isn't it :-)
2015-01-29 13:20:15 <ScrumpyJack> i get errors when adding package
2015-01-29 13:20:16 <ScrumpyJack> (16/31) Installing musl-dev (1.1.5-r1)
2015-01-29 13:20:16 <ScrumpyJack> ERROR: musl-dev-1.1.5-r1: No error information
2015-01-29 13:20:29 <ScrumpyJack> where can i find more info about that?
2015-01-29 13:21:57 <ScrumpyJack> apk update fix that :)
2015-01-29 16:16:16 <Jean-Scotch> j'amais ce sera normalemen
2015-01-29 16:21:29 <Jean-Scotch> oups wrong window ;)
2015-01-29 16:21:46 <Jean-Scotch> and hi!
2015-01-29 16:22:02 <Jean-Scotch> is there something like badblocks in AL ?
2015-01-29 16:22:34 <Jean-Scotch> or any other way to verify a disk at low level?
2015-01-29 16:23:24 <royger> Jean-Scotch: e2fsprogs or similar shjould contain badblocks
2015-01-29 16:23:39 <Jean-Scotch> ok. thanks
2015-01-29 19:04:08 <MaddinXx> Hi everyone. I have problems using awall (getting lua assertion errors). I think the problem is that I am within an OpenVZ container...but all iptable modules are loaded and Debian/CentOS etc. have no problems (and iptables alone (without awall)).
2015-01-29 19:04:32 <MaddinXx> the error is here: http://pastebin.com/2whEjuJB
2015-01-29 19:05:31 <MaddinXx> commenting the assertion out results in another error:
2015-01-29 19:05:38 <MaddinXx> http://pastebin.com/MkadPGQD
2015-01-29 19:06:05 <MaddinXx> any ideas how I can further check why it is failing?
2015-01-29 19:07:10 <MaddinXx> Alpine is 3.1.1 and all packages are up-to-date
2015-01-29 20:46:31 <kaniini> MaddinXx: openvz does not allow custom iptables rules iirc.
2015-01-29 22:22:07 <MaddinXx> @kaniini: yes it does. it has some limitations, but newer releases allow full iptables interaction
2015-01-29 22:22:42 <MaddinXx> @kaniini: e.g. shorewall show capabilities has everything as "available" (and it also works, not only shown like that)
2015-01-29 22:23:23 <MaddinXx> @kaniini: http://openvz.org/Man/vzctl.8#Netfilter_.28iptables.29_control_parameters
2015-01-29 22:35:08 <ahills> are you using twitter as an IRC client?
2015-01-29 22:38:49 <Diftraku> openvz is poop
2015-01-29 22:39:09 <MaddinXx> ahills: is that a funny way to ask me why I am sending three messages instead of one? ;)
2015-01-29 22:44:18 <MaddinXx> Diftraku: why do you think so? What are better alternatives? Xen/KVM are quite different...LXC/Docker not mature enough (and heavily based on stuff merged upstream by the OpenVZ if I am right) ... have been using it for around 3 years now and never had problems expect for Netfilter stuff..but with the new netfilter config option that is painless as well. but would love to hear
2015-01-29 22:44:18 <MaddinXx> your opinion :) (and hey, Alpine runs perfectly fine on it (yeah, no grsecurity of course))
2015-01-29 22:45:01 <Diftraku> my main problem woth openvz is lack of separstion between host and guest
2015-01-29 22:45:12 <Diftraku> they sahre the same kernel image
2015-01-29 22:45:17 <Diftraku> *share
2015-01-29 22:47:27 <MaddinXx> Diftraku: okay. many love it for exactly that reason (because it _should_ be a bit faster) but I can agree with that. it is totally a question of requirement here
2015-01-29 22:47:39 <ahills> MaddinXx: no, you're prefixing kaniini's name with @
2015-01-29 22:48:28 <MaddinXx> ahills: ahhh hehe. it was shown like that. but I see...that's the prefix for moderators. I am not so familiar with IRC :)
2015-01-29 22:52:17 <MaddinXx> Diftraku: for example an Alpine container is around 15M in size and uses 4MB or RAM. not sure if that can be reached with full emulators/hypervisors :) ... yes, it is the question if that is the goal. but just to mention it =)
2015-01-29 22:52:56 <ahills> I'm using alpine as a xen dom0 and it works fantastic
2015-01-29 22:53:02 <ahills> if xen is what you're looking for
2015-01-29 22:53:20 <ahills> I've heard xen performance complaints but I get better disk and net throughput than I do on ESX
2015-01-29 22:54:37 <Diftraku> I don't have a problem with openvz on guests and host you control
2015-01-29 22:55:02 <Diftraku> I do have a problem when you only control one guest
2015-01-29 22:55:23 <MaddinXx> ahills: Xen is still on my list to get familiar with some day...that might indeed be a good idea to use Alpine for that. but right now I am not looking for that. I joined to ask about an error I encounter with awall within an OpenVZ container...and still hoping to get it working (because as said, iptables in general has no problems..only with the awall "interface")
2015-01-29 22:55:40 <Diftraku> some configurations expose a ton ttys to thr guest for no apparent reason
2015-01-29 22:55:49 <ahills> is openvz an important detail?
2015-01-29 22:56:02 <ahills> e.g. does it affect the operation of iptables at all?
2015-01-29 22:56:12 <ahills> awall is just a json-to-iptables converter
2015-01-29 22:57:11 <systmkor> ahills, I thought awall was json to shorewall
2015-01-29 22:57:19 <systmkor> which is a wrapper on top of iptables
2015-01-29 22:57:27 <systmkor> (note: I could totally be wrong)
2015-01-29 22:57:30 <Diftraku> afaik iptables is a kmod and thus is provided in the kernel that in turn is the host's kernel
2015-01-29 22:57:38 <ahills> I don't think it goes through shorewall, as it's not a dependency
2015-01-29 22:57:45 <MaddinXx> ahills: it should not with my configuration. plain iptables does everything I want, same for shorewall. the only thing is that modprobe etc to load iptables modules doesn't work, but that should not be nessecarily anyway. so no, I think the problem is not related to OpenVZ. I only mentioned that to play with open cards :) (and because it probably has some effect I don't know
2015-01-29 22:57:45 <MaddinXx> about)
2015-01-29 22:57:47 <ahills> I don't see shorewall on my system, unless it's installed in a weird place
2015-01-29 22:57:54 <jomat> ahills: "Debian/CentOS etc. have noproblems (and iptables alone (without awall))."
2015-01-29 22:58:22 <systmkor> well then I'm probably wrong
2015-01-29 22:58:25 <ahills> debian/centos and openvz are different categories of thing, are they not?
2015-01-29 22:58:46 <MaddinXx> ahills: well yes, I meant Debian/CentOS OpenVZ containers
2015-01-29 22:58:58 <MaddinXx> like my Alpine is a container on an OpenVZ host
2015-01-29 23:00:11 <systmkor> is there a particular reason for using OpenVZ?
2015-01-29 23:00:24 <ahills> and iptables works fine on alpine?
2015-01-29 23:00:25 <systmkor> (i.e. what is the main difference or better features/benefits)
2015-01-29 23:00:27 <ahills> or just on debian?
2015-01-29 23:00:59 <MaddinXx> ahills: yes, that works fine on Alpine as well. even issueing the rules from /etc/iptables/awall-rules myself does work
2015-01-29 23:01:16 <MaddinXx> just not with awall activate
2015-01-29 23:03:30 <jomat> MaddinXx: I don't use awall myself... are the ip6tables-rules also stored in /etc/iptables/awall-rules ?
2015-01-29 23:03:34 <MaddinXx> systmkor: no, it is just the one I am using since the very beginning....was easy to get started with back then and always played well for me. only hosting some web/db/mail/dns containers on it
2015-01-29 23:04:32 <ScrumpyJack> MaddinXx: when you speak iptables from your guest without using awall and configure your chain/tables/rules does it work?
2015-01-29 23:05:10 <MaddinXx> ScrumpyJack: jup
2015-01-29 23:05:48 <jomat> MaddinXx: in ip6tables is what fails in your pastebin... have you tested that w/o awall, too?
2015-01-29 23:05:59 <MaddinXx> jomat: one strange point is that the service definitions are also placed inside awall-rules (not sure if this is correct?) beside that the content is as:
2015-01-29 23:06:30 <MaddinXx> http://pastebin.com/tuNZtd0s
2015-01-29 23:07:21 <jomat> ah, i see
2015-01-29 23:08:44 <MaddinXx> and the policy is a simple drop all: http://pastebin.com/xBsxTMak
2015-01-29 23:09:59 <MaddinXx> I am happy to give SSH access if anyone if willing to have a look. It's only a playground server
2015-01-29 23:10:17 <jomat> MaddinXx: http://pastebin.com/2whEjuJB .. is that the whole output of `awall activate`?
2015-01-29 23:11:11 <MaddinXx> jomat: yes. the next pastebin was the one after commenting the assertion out in iptables.lua
2015-01-29 23:11:40 <MaddinXx> right now the output therefor it like:
2015-01-29 23:11:48 <MaddinXx> http://pastebin.com/wq8xBigH
2015-01-29 23:12:19 <MaddinXx> (but as said, that is where I got after commenting out the lua assertion...)
2015-01-29 23:13:57 <jomat> Did you press return?
2015-01-29 23:14:24 <MaddinXx> yes
2015-01-29 23:14:48 <MaddinXx> otherwise it blocks at that message
2015-01-29 23:15:05 <MaddinXx> let me quickly record everything with showterm
2015-01-29 23:16:41 <jomat> ok... just asked because the last error in http://pastebin.com/wq8xBigH also scrolls over your prompt
2015-01-29 23:22:49 <MaddinXx> http://showterm.io/ac864570caec7ad38fe68
2015-01-29 23:31:16 <MaddinXx> jomat: if you didn't watch the record. I did not press return ... it fails before
2015-01-29 23:31:30 <jomat> I've seen it
2015-01-29 23:31:50 <jomat> showterm.io looks nice btw
2015-01-29 23:32:03 <jomat> didn't know that yet :-)
2015-01-29 23:32:33 <MaddinXx> yeah, can be of great help. good service
2015-01-29 23:33:33 <jomat> MaddinXx: Do you have the possibility to clone that system to a non-openvz-machine?
2015-01-29 23:33:52 <MaddinXx> any ideas what could be the reason? iptables-restore < rules-save and for ip6tables works.
2015-01-29 23:34:06 <jomat> MaddinXx: because I'm not convinced that its the fault of openvz
2015-01-29 23:34:26 <MaddinXx> hmm....not sure how I could do that. I could however install a fresh copy, yes
2015-01-29 23:35:16 <jomat> is that openvz-container hosted on a server you have access to?
2015-01-29 23:35:23 <jomat> or is it a rented vps?
2015-01-29 23:35:41 <MaddinXx> yes, its on a dedicated box I have rented
2015-01-29 23:36:11 <MaddinXx> (I have access to the HW node)
2015-01-29 23:37:00 <jomat> one approach could be to transfer the rootfs i. e. to your notebook where you start it in i. e. kvm with direct kernel boot (as it doesn't have an own kernel and bootloader)
2015-01-29 23:39:22 <jomat> and if the problem persists, it's more likely a problem of this installation
2015-01-29 23:40:02 <MaddinXx> jomat: good idea. will try to do that. one other hint: iptables-restore on rules-save etc. works as well and rules are applied. so it definitely has to do with what awall activate does behind the scenes
2015-01-29 23:41:33 <MaddinXx> with --force I only get this error: iptables-restore v1.4.21: Can't set policy `ACCEPT' on `INPUT' line 15: Bad built-in chain name
2015-01-29 23:42:07 <jomat> thats odd... and iptables -P INPUT ACCEPT works ofc...
2015-01-29 23:42:24 <MaddinXx> that works, yes
2015-01-29 23:44:29 <MaddinXx> line 15 from which file should that be? because in rules-save line 15 is -A OUTPUT -o lo -j ACCEPT and no ACCEPT INPUT
2015-01-29 23:44:41 <jomat> do you have a clue .... that's what I just wanted to ask :-)
2015-01-29 23:47:33 <MaddinXx> hehe. mind reader
2015-01-29 23:48:08 <MaddinXx> commenting out the revert() call in awall helps that the rules are applied. the error still appears, but at least the iptables rules are set
2015-01-29 23:50:35 <MaddinXx> well, leaving for today. will try to clone to another system tomorrow and see what happens and give feedback here. thanks anyway for the help everybody. very helpful here :)
2015-01-29 23:50:40 <MaddinXx> have a good night & bye
2015-01-29 23:51:06 <jomat> MaddinXx: if that doesn't help, there's a redmine bugtracker
2015-01-29 23:51:21 <jomat> http://bugs.alpinelinux.org/
2015-01-29 23:52:57 <jomat> MaddinXx: good luck and good night :-)
2015-01-29 23:53:05 <MaddinXx> jomat: will keep that in mind. :) googeling the error gives quite a few results (Arch Linux, Ubuntu etc. etc.) so I really hope to find the answer somewhere.
2015-01-29 23:53:06 <MaddinXx> bye
2015-01-30 01:36:25 <jomat> Error relocating /usr/sbin/pdns_server: _ZSt24__throw_out_of_range_fmtPKcz: symbol not found
2015-01-30 01:36:47 <jomat> hm...
2015-01-30 01:37:13 <dalias> perhaps it depends on a newer version of one of its libs and the package is lacking a dependency on it
2015-01-30 01:37:21 <dalias> i've encountered that before on alpine
2015-01-30 01:37:27 <dalias> iirc last time it was with inkscape
2015-01-30 01:39:17 <jomat> shouldn't libs be traced automatically?
2015-01-30 01:39:26 <jomat> while packaging?
2015-01-30 01:42:31 <dalias> i don't know why it fails. dependence on the lib packages is there, but the fact that the binary depends on a particular version gets missed :/
2015-01-30 01:42:41 <dalias> at least that was my experience
2015-01-30 01:42:55 <dalias> if you can find what lib that symbol should be in, try upgrading it manually
2015-01-30 01:43:36 <jomat> dalias: good tip... somehow i still have libstdc++ 4.8.3-r0 installed instead of 4.9.2-r3
2015-01-30 01:43:51 <jomat> after upgrading it works again
2015-01-30 01:43:54 <dalias> that sounds likely to be the cause
2015-01-30 03:41:09 <systmkor> anyone have anythoughts about providing or maybe making pam_tcb as default instead of default shadow
2015-01-30 03:41:11 <systmkor> files
2015-01-30 03:41:30 <systmkor> http://www.openwall.com/tcb/
2015-01-30 03:59:20 <systmkor> dalias, hey I'm having compilation issues with tcp_pam
2015-01-30 03:59:26 <systmkor> i have it's a package issue
2015-01-30 04:00:00 <systmkor> because /usr/include/nss/seccomon.h - includes - #include "prtypes.h"
2015-01-30 04:00:07 <systmkor> which that file doesn't exist in that directory
2015-01-30 04:00:40 <systmkor> and the "prtypes.h" file that does exist, exists in /usr/include/nspr/prtypes.h
2015-01-30 04:04:35 <systmkor> brb
2015-01-30 04:22:33 <systmkor> ncopa, I think the nss libraries are broken for compilation
2015-01-30 04:23:12 <systmkor> ncopa, they all refer to all nspr local c header files such as "nspr.h"
2015-01-30 04:23:29 <systmkor> ncopa, while they live in /usr/include/nspr/*
2015-01-30 13:51:57 <ET> Is anyone else getting a 404 - Not Found error when downloading alpine linux?
2015-01-30 13:57:03 <ncopa> it is probably due to new release
2015-01-30 13:57:08 <ncopa> needs to syn on the mirrors first
2015-01-30 14:01:06 <ET> aha works now
2015-01-30 14:03:27 <ET> eek, download via browser works but wget doesnt
2015-01-30 14:05:31 <ET> righto downloading now
2015-01-30 14:39:24 <ncopa> try this: http://nl.alpinelinux.org/alpine/v3.1/releases/x86_64/
2015-01-30 18:12:15 <steenuil> hey, I've seen that Wayland was installed as a dependency on my Alpine vm for some reason, but is there any way to actually run some wayland desktop without having to compile extra stuff from outside the default repos? Weston is not in there.
2015-01-30 23:39:08 <systmkor> steenuil, to piggy back off of you, is there any way to run wayland at all (minus X) with Alpine?
2015-01-30 23:39:13 <systmkor> if so, any guides how to?
2015-01-30 23:39:31 <steenuil> eeeh not that I know of
2015-01-30 23:40:00 <steenuil> I was trying to do stuff earlier but the compositors I tried didn't really work
2015-01-30 23:40:43 <systmkor> ah dang.. I can't wait till the no longer need of X :D
2015-01-30 23:40:58 <steenuil> I mean I just tried like two
2015-01-30 23:41:14 <systmkor> steenuil, in the mean time I was hoping we could basically do what OpenBSD did for X
2015-01-30 23:41:30 <steenuil> what did they do?
2015-01-30 23:41:30 <systmkor> and remove all the SUIDs/SGIDs et .
2015-01-30 23:41:39 <steenuil> eeh
2015-01-30 23:41:47 <systmkor> it's been done
2015-01-30 23:41:57 <systmkor> and that's been standard for years now
2015-01-30 23:43:05 <systmkor> steenuil, I believe the also added in other security features or just made it the base more secure
2015-01-30 23:43:08 <systmkor> steenuil, https://www.youtube.com/watch?v=2l7ixRE3OCw
2015-01-30 23:43:21 <steenuil> well it's OpenBSD, of course they did.
2015-01-30 23:43:33 <systmkor> yup
2015-01-30 23:43:40 <systmkor> and so should we (i.m.o)
2015-01-30 23:43:41 <steenuil> I was just trying to do that because something didn't quite work out with my X installation
2015-01-30 23:43:59 <steenuil> but I don't really know what it is
2015-01-30 23:48:18 <systmkor> ahhh
2015-01-30 23:49:03 <steenuil> but yeah I mean, using Wayland would be great.
2015-01-30 23:49:43 <systmkor> I also would love to see more regression tests for all of Alpine packages
2015-01-30 23:49:45 <systmkor> :D
2015-01-30 23:50:06 <systmkor> but... not nearly enough people power for that, atm
2015-01-30 23:50:43 <steenuil> yeah, I figured.
2015-01-30 23:50:50 <systmkor> so please contributed :D
2015-01-30 23:51:06 <steenuil> I-I'll try to do something about it
2015-01-30 23:51:10 <steenuil> when I figure out what it is
2015-01-30 23:51:22 <systmkor> and document document document :P
2015-01-31 04:54:09 <systmkor> is there a location we could post desired projects to work on
2015-01-31 04:54:16 <systmkor> or what we would like to see happen
2015-01-31 04:54:31 <systmkor> ncopa, would that be appropriate on the wiki?
2015-01-31 04:56:13 <dalias> that would be nice :)
2015-01-31 04:56:27 <dalias> emacs & chromium :)
2015-01-31 04:56:36 <dalias> those are my top 2 requests
2015-01-31 04:57:48 <dalias> btw are alternate user-db backends a topic of interest for alpine?
2015-01-31 04:57:54 <dalias> ldap/nis/whatever?
2015-01-31 04:58:09 <dalias> because we're working on them in musl now :)
2015-01-31 04:58:33 <systmkor> huh how do those relate to MUSL?
2015-01-31 04:58:51 <systmkor> dalias, well I'll just create a page under Developers for project ideas
2015-01-31 04:58:56 <dalias> getpwnam() etc
2015-01-31 04:58:58 <systmkor> if it needs to be moved, no problem
2015-01-31 04:59:03 <systmkor> what's that?
2015-01-31 04:59:11 <dalias> right now they only support /etc/passwd and /etc/group files
2015-01-31 04:59:36 <dalias> getpwnam() and friends are the standard libc functions to lookup user/group name/id mappings
2015-01-31 04:59:48 <dalias> and supplemental stuff like shell, homedir, realname, ...
2015-01-31 05:00:30 <systmkor> dalias, or do you think under Contribute page?
2015-01-31 05:00:50 <systmkor> getpwnam() is a posix/std libc call?
2015-01-31 05:00:54 <dalias> yes
2015-01-31 05:00:56 <dalias> posix
2015-01-31 05:01:08 <systmkor> so wait is that POSIX or pure Libc?
2015-01-31 05:01:16 <systmkor> or are those synonymous now?
2015-01-31 05:02:23 <systmkor> dalias, oh I have two projects I think you should check out (you may already know of there existence)
2015-01-31 05:02:33 <dalias> in the contexts we're working in, "libc" includes posix
2015-01-31 05:02:38 <systmkor> dalias, https://github.com/NixOS/patchelf
2015-01-31 05:02:57 <systmkor> dalias, why not make that a separate?
2015-01-31 05:03:00 <systmkor> out of curiousity
2015-01-31 05:03:19 <systmkor> dalias, http://www.openwall.com/tcb/
2015-01-31 05:03:58 <dalias> making that separate really isn't feasible
2015-01-31 05:04:07 <dalias> but musl unifies more than historically done into one file
2015-01-31 05:04:11 <dalias> see http://wiki.musl-libc.org/wiki/Design_Concepts
2015-01-31 05:04:37 <dalias> what about tcb?
2015-01-31 05:05:03 <dalias> musl is the first adopter of tcb shadow outside of Owl, so yes i know about it
2015-01-31 05:05:17 <dalias> actually tcb in musl is how solar/openwall first got interested in musl
2015-01-31 05:05:47 <systmkor> cool
2015-01-31 05:05:56 <systmkor> well I'm trying to compile TCB for Alpine
2015-01-31 05:05:59 <systmkor> but running into NSS issues
2015-01-31 05:07:35 <dalias> well you don't need any libc (/nss) side stuff for tcb on musl
2015-01-31 05:07:40 <dalias> it just works
2015-01-31 05:07:55 <dalias> you do need the modified passwd utility, etc. tho to update tcb shadow files
2015-01-31 05:08:08 <systmkor> oh well when I'm compiling
2015-01-31 05:08:15 <systmkor> straight from owl's source
2015-01-31 05:08:18 <systmkor> that's not what I get
2015-01-31 05:08:33 <dalias> the tcb package from owl has stuff that you won't need on musl because equivalent code is part of musl
2015-01-31 05:08:41 <systmkor> ah
2015-01-31 05:09:01 <dalias> but unless you want to edit /etc/tcb/*/shadow by hand, you need modified passwd etc. utils to edit it for you
2015-01-31 05:09:06 <systmkor> well do they have a source of tcb that doesn't compile against glibc
2015-01-31 05:09:25 <systmkor> dalias, well I'm going to make a wish list one sec
2015-01-31 05:10:24 <dalias> for example if you have a working alpine system and you just split /etc/shadow up into /etc/tcb/[username]/shadow (one line file for each user)
2015-01-31 05:10:32 <dalias> then rename/remove shadow
2015-01-31 05:10:38 <dalias> you should still be able to login just fine
2015-01-31 05:10:53 <dalias> but the passwd utility won't be able to change your password then
2015-01-31 05:10:54 <systmkor> oh cool
2015-01-31 05:11:12 <dalias> btw the reason this works....
2015-01-31 05:11:39 <dalias> AGES ago (back in 1998 or 99 or so) i had a hack in libc5 and later glibc that allowed the user's password hash to be in ~/.passwd
2015-01-31 05:11:57 <dalias> and i wanted the same functionality when switching to musl, but didn't want it to be hackish like that
2015-01-31 05:12:18 <dalias> and i ran across owl's tcb that was an existing published system that matched what i was doing, but did it better
2015-01-31 05:12:36 <dalias> so rather than inventing something new and incompatible, i made musl support the owl tcb system
2015-01-31 05:13:09 <systmkor> very awesome
2015-01-31 05:13:28 <systmkor> well I think getting the suid/sgid out of passwd would be wicked
2015-01-31 05:13:31 <systmkor> one sec
2015-01-31 05:14:01 <dalias> that can be done depending on the permissions for /etc/tcb/* dirs
2015-01-31 05:14:09 <dalias> if you allow users to write their own dirs, suid is not needed
2015-01-31 05:14:20 <systmkor> yuuup
2015-01-31 05:14:20 <dalias> but then you can't force users to pick strong passwords
2015-01-31 05:14:24 <dalias> they can store any hash they want
2015-01-31 05:14:36 <dalias> imo this isn't a practical problem
2015-01-31 05:14:46 <dalias> if the utility enforces password strength, that's enough imo
2015-01-31 05:15:02 <dalias> somebody who goes out of their way to manually edit the file to work around this is going to find another way to make their account insecure anyway
2015-01-31 05:15:09 <dalias> like emailing their password in cleartext :-p
2015-01-31 05:18:43 <systmkor> idk I like the goal of at least removing all SUIDs/SGIDs from base alpine
2015-01-31 05:18:50 <systmkor> and tcb seems like a good solution
2015-01-31 05:19:12 <systmkor> Idk if blowfish is really the best hash to go with, especially since now it should be at least
2015-01-31 05:19:14 <dalias> yes
2015-01-31 05:19:28 <systmkor> a hash with a salt and at least 1k iterations
2015-01-31 05:19:30 <systmkor> or so
2015-01-31 05:19:52 <dalias> afaik bcrypt is still considered best practice short of scrypt or similar
2015-01-31 05:19:57 <systmkor> also creating an assymetric login would be nice
2015-01-31 05:20:21 <dalias> well i prefer disabling password login for sshd
2015-01-31 05:20:22 <systmkor> *assymetric crypto login
2015-01-31 05:20:27 <dalias> and using public key only
2015-01-31 05:20:28 <systmkor> oh I do to
2015-01-31 05:20:50 <systmkor> okay so wanna help me populate an initial wishlist page
2015-01-31 05:21:29 <dalias> well for apps, my big wishlist items are emacs and chromium
2015-01-31 05:22:32 <dalias> for security, minimizing/eliminating suids, making use of musl's tcb-shadow support (i.e. making the passwd utility etc. support it and perhaps use it by default), ...
2015-01-31 05:23:25 <dalias> i'm interested in getting things to the point where openwall might be able to recommend alpine as an alternative to owl for users who want a less-conservative, more modern system
2015-01-31 05:23:27 <systmkor> well so I'm making an outline first
2015-01-31 05:23:31 <systmkor> so sections first
2015-01-31 05:23:42 <dalias> security enhancements/hardening
2015-01-31 05:23:46 <dalias> application packagingf
2015-01-31 05:24:06 <dalias> perhaps testing/benchmarking type stuff?
2015-01-31 05:24:30 <systmkor> m'kay
2015-01-31 05:25:43 <dalias> possibly other soft goals in areas like publicity/outreach/etc. ?
2015-01-31 05:26:10 <dalias> imo that should have input from the ppl behind alpine tho
2015-01-31 05:26:19 <dalias> rather than users just making it with no input from them
2015-01-31 05:27:32 <systmkor> well yah. thanks I will put a notice at top of the page that it's unofficial
2015-01-31 05:33:51 <systmkor> dalias, well it's up
2015-01-31 05:33:57 <dalias> url?
2015-01-31 05:34:03 <systmkor> http://wiki.alpinelinux.org/wiki/Wishlist
2015-01-31 05:34:20 <systmkor> May need to be changed to ProjectWishlist in the future but for later discussion
2015-01-31 05:35:43 <systmkor> dalias, go put some stuff in there :D
2015-01-31 05:36:49 <thiagowfx> cat 'TL;DR' >> /wishlist
2015-01-31 05:37:08 <thiagowfx> I'm with so many wiki pages of alpine opened here, I'll go crazy =p
2015-01-31 05:37:14 <dalias> imo it's too broken-down for the amount of content there now :-p
2015-01-31 05:37:24 <thiagowfx> I am liking what I'm reading, though
2015-01-31 05:37:33 <systmkor> dalias, of course but that is all the things I already have ideas of putting stuff into
2015-01-31 05:37:45 <systmkor> I just wanted for you to be able to put some stuff in there first
2015-01-31 05:38:00 <systmkor> thiagowfx, TL;DR are you talking about the tool
2015-01-31 05:38:26 <systmkor> or meaning like there should be a sentence at the top of the page or in the contribute section as TL;DR
2015-01-31 05:38:42 <systmkor> or meaning each page should have a TL;DR sentence :P
2015-01-31 05:38:49 <systmkor> *each wiki page
2015-01-31 05:39:19 <thiagowfx> I mean there should be a (single) wiki page called tl;dr or similar, for newcomers like me
2015-01-31 05:39:30 <thiagowfx> which I am willing to create myself if I decide I liked alpine
2015-01-31 05:39:31 <systmkor> thiagowfx, very good Idea :D
2015-01-31 05:39:44 <systmkor> thiagowfx, I think you would be best suited for that at the moment
2015-01-31 05:39:56 <systmkor> will for sure be down to do editing & review
2015-01-31 05:40:51 <thiagowfx> it is nice to see an active community
2015-01-31 05:40:59 <thiagowfx> I though alpine were a much very small project
2015-01-31 05:41:03 <thiagowfx> was*
2015-01-31 05:41:56 <systmkor> huh? hard time parsing thought. I think I get the idea though.
2015-01-31 05:43:35 <systmkor> did you mean 'I thought alpine was a very small project.' ?
2015-01-31 05:54:17 <systmkor> dalias, okay SGID/SUIDs section added
2015-01-31 05:54:34 <systmkor> definitely needs more, especially siting sources such as articles/papers/& talks of why this is a good idea
2015-01-31 05:54:49 <systmkor> why should anyone give a shit about these projects without good reasons :D
2015-01-31 06:09:27 <thiagowfx> systmkor: yes, that was what I meant
2015-01-31 06:09:42 <thiagowfx> small typo :/
2015-01-31 06:10:19 <systmkor> no problem just trying to make sure I understood you correctly
2015-01-31 06:11:20 <thiagowfx> and I actually missed your question because my internet connection dropped
2015-01-31 06:11:26 <thiagowfx> I just answered to it now because I found http://dev.alpinelinux.org/irclogs/
2015-01-31 06:11:37 <thiagowfx> you guys seem to be well organized :D
2015-01-31 06:12:27 <systmkor> lol, idk about that. I just have a lot of ideas and am really eager to help.
2015-01-31 06:16:48 <systmkor> thiagowfx, thanks though
2015-01-31 06:17:01 <systmkor> main props to rest of Alpine community and devs
2015-01-31 06:17:20 <thiagowfx> sure
2015-01-31 06:20:27 <thiagowfx> logging off, bye
2015-01-31 06:20:52 <systmkor> ciao
2015-01-31 11:35:55 <paper_face> is there a way of making the system boot if a network interface can't be brought up?
2015-01-31 11:36:11 <paper_face> as in a timeout
2015-01-31 11:36:19 <paper_face> before the system will boot anyway and give you a shell
2015-01-31 11:36:55 <paper_face> because when i added auto ppp0 to my networking configuration, i realised if the ppp0 interface can't be brought up for some reason i can't get into the system
2015-01-31 11:37:09 <Zlaxhe> Your latest edition is on Linuxtracker.org, please consider using torrent files to get your distro shared :-)
2015-01-31 11:39:47 <paper_face> am i looking in /etc/conf.d/netmount at this "rc_need="net""
2015-01-31 11:45:22 <paper_face> ncopa: did you find a solution to this? for your ppp0 interface?
2015-01-31 12:06:02 <paper_face> did I miss anything?
2015-01-31 14:01:48 <paper_face_> So I've got my network setup like: http://slexy.org/view/s21QLkRCgV (diagram, ip r, ip addr)
2015-01-31 14:01:49 <paper_face_> I'm trying to add a route so that hosts on 192.168.1.0/24 (LAN) can access 192.168.0.1/30 (modem)
2015-01-31 14:01:49 <paper_face_> I've tried this: # ip r add 192.168.1.0/24 via 192.168.0.2 dev eth0 but I get RTNETLINK answers: File exists anyone messed with routing that might know what is up?
2015-01-31 14:40:39 <acer> I'm having differculties creating a account on http://wiki.alpinelinux.org/, whats the answer to the question "Are you a spammer?"?
2015-01-31 15:09:59 <jzono1> paper_face_: you'll want something like this:
2015-01-31 15:10:01 <jzono1> iptables --table nat --append POSTROUTING --out-interface eth0 -j MASQUERADE
2015-01-31 15:10:04 <jzono1> iptables --append FORWARD --in-interface eth1 -j ACCEPT
2015-01-31 15:10:06 <jzono1> echo 1 > /proc/sys/net/ipv4/ip_forward
2015-01-31 20:07:05 <thiagowfx> when I do apk add, the package manager just extracts tarballs or it also downloads them?
2015-01-31 20:07:39 <thiagowfx> because the output is always 'Installing <package xxx>', I have the impression everything is already downloaded
2015-01-31 20:15:26 <systmkor> I don't know for certain
2015-01-31 20:15:38 <systmkor> but not all packages are already downloaded
2015-01-31 20:15:41 <systmkor> that I'm certain of
2015-01-31 20:21:14 <steenuil> I think that's just to keep a clean output
2015-01-31 20:21:38 <steenuil> since all it does is download and extract packages.
2015-01-31 20:28:50 <thiagowfx> it looks like it downloads them too,I could notice that after adding/installing firefox
2015-01-31 20:29:27 <steenuil> yep. if it cached them it would take ages to apk update and you'd need lots of space too
2015-01-31 20:33:20 <systmkor> thiagowfx, I added more to the wishlist, go add your own stuff :D
2015-01-31 20:34:08 <thiagowfx> systmkor: great, I must get a working xorg server on my alpine box first =p
2015-01-31 20:35:39 <systmkor> m'kay document process
2015-01-31 20:35:56 <thiagowfx> is there something similar to pkgfile in alpine?
2015-01-31 20:36:11 <thiagowfx> to find which package provide a given program
2015-01-31 21:15:32 <thiagowfx> yay
2015-01-31 21:15:39 <thiagowfx> finally managed to compile i3wm under alpine
2015-01-31 21:15:52 <thiagowfx> would anyone be interested in packaging it?
2015-01-31 21:16:27 <steenuil> make an apkbuild
2015-01-31 21:17:40 <thiagowfx> I'll try
2015-01-31 21:17:52 <thiagowfx> what is the equivalent of the AUR in alpine?
2015-01-31 21:18:06 <thiagowfx> a place where I can find sample APKBUILDs
2015-01-31 21:18:28 <steenuil> http://wiki.alpinelinux.org/wiki/Abuild_and_Helpers
2015-01-31 21:27:39 <systmkor> thiagowfx, yes, I would really appreciate that packaged
2015-01-31 21:27:44 <systmkor> thiagowfx, how did you get it to compile
2015-01-31 21:28:05 <tvvdhSL5hEBn> it is packaged
2015-01-31 21:28:13 <systmkor> thiagowfx, I ran into problems were it required some blob code from glibc so I wasn't sure of an easy fix
2015-01-31 21:28:30 <tvvdhSL5hEBn> i3wm is packaged in testing for quite some time
2015-01-31 21:29:37 <tvvdhSL5hEBn> https://github.com/alpinelinux/aports/tree/master/testing/i3wm
2015-01-31 21:31:49 <systmkor> tvvdhSL5hEBn, ughh I swear I looked
2015-01-31 21:32:08 <systmkor> tvvdhSL5hEBn, thanks :D
2015-01-31 21:33:59 <thiagowfx> tvvdhSL5hEBn: oh, nice. And I was reinventing the wheel here
2015-01-31 21:34:55 <thiagowfx> I compiled it differently, without any patch, just by modifying CFLAGS
2015-01-31 21:35:11 <thiagowfx> I'll write it here
2015-01-31 21:36:34 <thiagowfx> make CFLAGS="-DGLOB_TILDE=0 -I/usr/include/pango-1.0 -I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include"
2015-01-31 21:37:16 <thiagowfx> based on this: https://bug.tasktools.org/secure/attachment/10271/0001-Fix-build-with-musl-libc.patch
2015-01-31 21:37:32 <thiagowfx> suggested by Copa himself
2015-01-31 21:37:54 <systmkor> i haven't seen that before
2015-01-31 21:37:58 <systmkor> interesting
2015-01-31 21:38:18 <thiagowfx> I just wondered why I have to include all those folder (with -I)
2015-01-31 21:38:25 <thiagowfx> they should be included automatically AFAIK
2015-01-31 21:38:37 <tvvdhSL5hEBn> for that you also need these defines in the code: +#ifdef GLOB_TILDE
2015-01-31 21:38:44 <tvvdhSL5hEBn> as you can see in the patch: https://github.com/alpinelinux/aports/blob/master/testing/i3wm/musl.patch
2015-01-31 21:38:59 <tvvdhSL5hEBn> otherwisse the -DGLOB_TILDE=0 makes no sense
2015-01-31 21:39:15 <thiagowfx> tvvdhSL5hEBn: it makes even without the patch
2015-01-31 21:39:24 <thiagowfx> in the beginning, I was adding those defines manually
2015-01-31 21:39:35 <tvvdhSL5hEBn> also i thing -DGLOB_TILDE=0 actually defines it, so the #if defined actually kicks in
2015-01-31 21:39:38 <thiagowfx> however, after three files, I decided to stop and to use -DGLOB_TILDE=0
2015-01-31 21:39:49 <thiagowfx> it worked (even without defining manually in the files I missed)
2015-01-31 21:40:01 <tvvdhSL5hEBn> maybe this patch is already upstream?
2015-01-31 21:40:15 <thiagowfx> yeah, the purpose of -D is to define it, so there is no reason to use #ifndefs
2015-01-31 21:40:27 <systmkor> so is that a generic patch to make it easier to compile with MUSL?
2015-01-31 21:40:29 <thiagowfx> I'm not sure, probably not
2015-01-31 21:40:38 <tvvdhSL5hEBn> why guess? why don't you check?
2015-01-31 21:40:39 <thiagowfx> unless someone that used musl reported it
2015-01-31 21:40:54 <tvvdhSL5hEBn> there's git histories and possibly mailinglists related
2015-01-31 21:40:57 <tvvdhSL5hEBn> where you can track this.
2015-01-31 21:41:28 <thiagowfx> don't mind my laziness, I'm kind of multitasking right now
2015-01-31 21:41:30 <thiagowfx> but I'll check it
2015-01-31 21:41:58 <tvvdhSL5hEBn> checking things before reduces redundant work :)
2015-01-31 21:42:34 <thiagowfx> sure
2015-01-31 21:42:53 <thiagowfx> I was creating an APKBUILD for i3 because yesterday somewhere said that it wasn't packaged
2015-01-31 21:43:10 <tvvdhSL5hEBn> someone you still trust tomorrow?
2015-01-31 21:43:11 <thiagowfx> and indeed I didn't found it with apk search today, but now I know about this testing repo
2015-01-31 21:43:17 <systmkor> thiagowfx, that may have been me, idk
2015-01-31 21:43:46 <systmkor> tvvdhSL5hEBn, yah sorta missing documentation for new people looking to see if something is already packaged
2015-01-31 21:43:49 <tvvdhSL5hEBn> apk search also returns results from testing if you have it in /etc/apk/repos
2015-01-31 21:44:03 <tvvdhSL5hEBn> i think that documentation is in the wiki
2015-01-31 21:44:21 <thiagowfx> tvvdhSL5hEBn: I'm a newcomer to alpine, so I'm likely to accept as truth what people tells me here about it
2015-01-31 21:44:25 <thiagowfx> at least for now
2015-01-31 21:46:47 <thiagowfx> okwYotjAkvr0:  in fact, /etc/apk/repsitories has the testing repo commented out
2015-01-31 21:47:07 <okwYotjAkvr0> yeah, testing is not default enabled, it's called testing :)
2015-01-31 21:48:07 <thiagowfx> it could be enabled for searching but not for installation :D
2015-01-31 21:48:15 <thiagowfx> anyway, problem solved
2015-01-31 21:48:25 <okwYotjAkvr0> read more wikis! ;)
2015-01-31 21:49:46 <systmkor> okwYotjAkvr0, that said the wikis are rather out of date and meh for new people to read
2015-01-31 21:51:11 <okwYotjAkvr0> i don't think that the information that is relevant to the discussion the last 20 min is either meh, nor outdated.
2015-01-31 21:53:51 <thiagowfx> Wikis should indeed be read, but I ended up with more than 15 tabs opened. I think there is a compromise between reading and applying what you read
2015-01-31 21:54:57 <thiagowfx> tl;dr read, use, read, use, and gain familiarity with the system bit by bit
2015-01-31 21:55:16 <systmkor> thiagowfx, yup yup :D
2015-01-31 21:55:29 <okwYotjAkvr0> sounds quite american. first shoot, then ask
2015-01-31 21:55:29 <systmkor> thiagowfx, don't forget to document & edit as well
2015-01-31 21:55:33 <thiagowfx> systmkor: that said, I have to keep reading now :)
2015-01-31 21:55:34 <okwYotjAkvr0> !!!
2015-01-31 21:55:36 <thiagowfx> by the way
2015-01-31 21:55:42 <thiagowfx> glad you said that
2015-01-31 21:55:47 <thiagowfx> I cannot register a new user
2015-01-31 21:55:54 <thiagowfx> there is some problem with the captcha
2015-01-31 21:55:58 <systmkor> weird I was able to a month back
2015-01-31 21:56:07 <systmkor> ncopa, wiki captcha problems?
2015-01-31 21:56:34 <thiagowfx> there is this question: "Are you a spammer?"
2015-01-31 21:56:44 <thiagowfx> and nothing else, so I fill "no"
2015-01-31 21:56:54 <thiagowfx> but I think there should be some image
2015-01-31 21:56:56 <thiagowfx> but there is none
2015-01-31 21:57:50 <thiagowfx> okwYotjAkvr0: actually it is more like ask first, but don't spend all day asking
2015-01-31 22:03:32 <systmkor> thiagowfx, okwYotjAkvr0 Idk as long nothing will get hurt
2015-01-31 22:03:36 <systmkor> do first works pretty well
2015-01-31 22:03:59 <systmkor> so shooting definitely falls out of that guide
2015-01-31 22:04:07 <thiagowfx> systmkor: yep
2015-01-31 22:08:39 <okwYotjAkvr0> the asking should come after the reading, since if you ask without any effort, you are waste the time of those, who did actually document their work. so that they can focus on more awesome work.
2015-01-31 22:09:51 <systmkor> okwYotjAkvr0, that said just because it's documented doesn't mean the documentation is any good
2015-01-31 22:11:03 <okwYotjAkvr0> we seem to disagree on this part i think.
2015-01-31 22:11:18 <systmkor> I'm not saying that Alpine's current docs are bad
2015-01-31 22:11:19 <thiagowfx> I don't defend asking before reading; however, there is a thing called 'opportunity cost' that comes to mind when there is too much documentation
2015-01-31 22:11:31 <systmkor> I'm just saying the fact that documentation exists
2015-01-31 22:11:50 <okwYotjAkvr0> yes, and you externalize the cost on those who are most productive with the product.
2015-01-31 22:11:55 <systmkor> doesn't mean its good
2015-01-31 22:11:57 <thiagowfx> I believe that (for example) wasting an entire day to read something just to conclude that you won't need what you just read is a waste of resources
2015-01-31 22:12:12 <thiagowfx> so there is a compromise between reading and asking
2015-01-31 22:12:22 <okwYotjAkvr0> so your time is more valuable than the time of devs
2015-01-31 22:12:35 <okwYotjAkvr0> certainly a reasonable assumption.
2015-01-31 22:12:43 <okwYotjAkvr0> everyone would agree ;)
2015-01-31 22:12:50 <thiagowfx> ¬¬
2015-01-31 22:13:19 <systmkor> okwYotjAkvr0, also arguably if you keep getting the same questions for what already exists for documentation
2015-01-31 22:13:28 <systmkor> that may mean it isn't easy for new people or that demographic
2015-01-31 22:13:34 <systmkor> to find what they are looking for
2015-01-31 22:13:58 <systmkor> okwYotjAkvr0, it's easy to go look up reference material when you wrote it
2015-01-31 22:14:04 <systmkor> but when you are reading it for the first time
2015-01-31 22:14:20 <systmkor> it can be like trying to figure out the illogical nuances of a text adventure game
2015-01-31 22:17:45 <okwYotjAkvr0> all is good and fine. i just wanted to point out that the redundancy can be avoided and used more efficiently. but you do however you like.
2015-01-31 22:17:50 <thiagowfx> I'm not sure if you are really trying to express your point or just to disagree. But I'll end my argument with this: there is a difference between who doesn't read absolutely anything (thus asking first, sometimes the same questions) aaaand those who demonstrate some effort before asking. For example, I could spend here all day asking you how to install x, y, and z, but there is no need for me do to that because it is (well) documented in the
2015-01-31 22:17:51 <thiagowfx> Alpine Wiki. However, if I find a particular program that I couldn't find (i3 in this case), I don't really see any problems in asking *here* if it isn't really present in the repos. Besides, I was creating a package for it before discovering there was one already, so I don't think 'my time is more important than the devs' either.
2015-01-31 22:18:34 <okwYotjAkvr0> you could've used the time to do something non-redundant.
2015-01-31 22:18:53 <okwYotjAkvr0> but yes. sorry if you got the wrong answer here
2015-01-31 22:22:56 <thiagowfx> okwYotjAkvr0: all right
2015-01-31 22:23:29 <okwYotjAkvr0> but i think you also got the right answer, just not at the time of your convenience ;)
2015-01-31 22:26:16 <systmkor> okwYotjAkvr0, thiagowfx I wouldn't argue a user, newb, or developer's time is innately more valuable than any other. It depends on desired outcomes.
2015-01-31 22:26:34 <systmkor> How much reasonable time any party has attempted.
2015-01-31 22:26:40 <systmkor> and other things
2015-01-31 22:27:09 <systmkor> okwYotjAkvr0, because if a newbs time is considered worthless your community will end up like OpenBSD
2015-01-31 22:27:19 <systmkor> forever diminishing
2015-01-31 22:29:50 <systmkor> this said I don't have solid research backing up my claims
2015-01-31 22:33:50 <steenuil> do you have to make any major changes to "port" PKGBUILDs to abuild?
2015-01-31 22:34:13 <okwYotjAkvr0> "newbs time is considered worthless" is a gross misrepresentation of what i said.
2015-01-31 22:34:51 <okwYotjAkvr0> the message was to spend your own time instead of those who already contribute
2015-01-31 22:34:58 <systmkor> okwYotjAkvr0, I didn't say that's what you said
2015-01-31 22:35:08 <systmkor> I'm just saying that position (not yours)
2015-01-31 22:35:14 <systmkor> can be rather poisonous
2015-01-31 22:35:22 <okwYotjAkvr0> agreed
2015-01-31 22:35:43 <systmkor> okwYotjAkvr0, for me that's why I really want to encourage documentation as you learn
2015-01-31 22:35:57 <systmkor> so for those who are new to a project have material that was made
2015-01-31 22:36:01 <systmkor> from their perspective
2015-01-31 22:36:09 <systmkor> not of a veteran user or dev
2015-01-31 22:36:10 <thiagowfx> steenuil: from what I'm reading, it looks like no, just minor ones
2015-01-31 22:36:24 <steenuil> so I just run it and see what doesn't work
2015-01-31 22:36:26 <steenuil> good.
2015-01-31 22:36:37 <thiagowfx> changing names of dependencies
2015-01-31 22:36:41 <systmkor> steenuil, uhhh I don't know if you can just drop in replace but worth a shot
2015-01-31 22:36:55 <steenuil> oh right
2015-01-31 22:37:16 <okwYotjAkvr0> just to not repeat the thing from a few minutes earlier, you sure, the pkg is not in testing?
2015-01-31 22:37:24 <thiagowfx> a few minutes ago I was porting i3 PKGBUILD to a APKBUILD
2015-01-31 22:37:28 <thiagowfx> it wasn't much different, overall
2015-01-31 22:37:33 <thiagowfx> the process
2015-01-31 22:38:04 <okwYotjAkvr0> also you might want to check the "unmaintained" repo
2015-01-31 22:39:02 <steenuil> nope, it's not.
2015-01-31 22:39:24 <steenuil> I wanted to run Wayland so I'm trying to make wlc and a window manager work.
2015-01-31 22:39:53 <okwYotjAkvr0> hmmm, doesnt wayland depend on systemd, and systemd on glibc?
2015-01-31 22:40:05 <steenuil> maybe?
2015-01-31 22:40:21 <okwYotjAkvr0> well the latter i'm sure.
2015-01-31 22:40:27 <okwYotjAkvr0> the former i guess you'll find out ;)
2015-01-31 22:41:06 <steenuil> nop http://forum.alpinelinux.org/apk/main/x86_64/wayland
2015-01-31 22:41:26 <steenuil> everything "requires" systemd these days
2015-01-31 22:41:49 <steenuil> except nothing really does
2015-01-31 22:45:20 <okwYotjAkvr0> wow cool!
2015-01-31 22:46:42 <okwYotjAkvr0> yeah on debian even erlang depends on systemd.
2015-01-31 22:47:11 <dalias> haha
2015-01-31 22:47:44 <dalias> systemd dependency is usually a bug in how packages are built
2015-01-31 22:47:48 <dalias> not a core dependency of the program
2015-01-31 22:48:18 <thiagowfx> okwYotjAkvr0: lol
2015-01-31 22:48:50 <okwYotjAkvr0> i guess the debianguys want to make sure i never come back. ;)
2015-01-31 22:52:13 <okwYotjAkvr0> dalias: do you think it would make sense to submit a bug against erlang?
2015-01-31 22:53:30 <dalias> in debian?
2015-01-31 22:53:53 <dalias> debian is so polarized and i'm so far outside their political circles that i'm not sure whether it would help or hurt
2015-01-31 22:54:27 <dalias> i've pretty much lost interest in debian. systemd wasn't the main factor but it was kinda the last straw
2015-01-31 22:54:46 <dalias> basically debian is just insanely slow and bloated
2015-01-31 22:54:49 <dalias> :-p
2015-01-31 22:55:00 <thiagowfx> dalias: whoa, bloated?
2015-01-31 22:55:09 <thiagowfx> that was the last think I would expect to hear about debian
2015-01-31 22:55:29 <dalias> compare as base alpine installation to a base debian installation
2015-01-31 22:55:34 <thiagowfx> I mean, if you consider the distros that use .deb packages
2015-01-31 22:55:39 <thiagowfx> I think debian is the simplest
2015-01-31 22:55:48 <dalias> that's not saying much
2015-01-31 22:55:56 <okwYotjAkvr0> aboslutely agreed with dalias
2015-01-31 22:56:01 <thiagowfx> yeah, the sense of bloated is relative
2015-01-31 22:56:07 <okwYotjAkvr0> my alpine boots to loginprompt in 2 secs.
2015-01-31 22:56:12 <dalias> debian's gotten impossible to run on a system without ridiculous amounts of ram
2015-01-31 22:56:16 <okwYotjAkvr0> while debian is still looking for grub in that time.
2015-01-31 22:56:45 <dalias> and super-fast hdd or ssd to deal with all the io load
2015-01-31 22:57:18 <dalias> none of these are an option on an energy-efficient, long-batt-life laptop or netbook
2015-01-31 22:57:47 <dalias> where doubling the ram reduces the batt life in suspend mode by 40% or so
2015-01-31 22:58:17 <okwYotjAkvr0> suspend mode. never use that. at least if you practice opsec, and have full disk encryption.
2015-01-31 22:58:38 <dalias> debian has something like 400-800 megs of useless crap running before you even get to any user-facing apps
2015-01-31 22:59:32 <dalias> that's not called opsec
2015-01-31 22:59:35 <dalias> it's called tinfoil hat
2015-01-31 22:59:49 <okwYotjAkvr0> depends on your threatmodell.
2015-01-31 22:59:58 <dalias> nobody is going to perform a live attack on a running system rather than hitting you over the head with a wrench
2015-01-31 23:00:07 <dalias> that's just stupid
2015-01-31 23:00:09 <okwYotjAkvr0> really?
2015-01-31 23:00:13 <okwYotjAkvr0> police controls?
2015-01-31 23:00:19 <okwYotjAkvr0> never run into those?
2015-01-31 23:00:22 <okwYotjAkvr0> i do
2015-01-31 23:00:26 <okwYotjAkvr0> :/
2015-01-31 23:00:28 <dalias> where?
2015-01-31 23:00:38 <okwYotjAkvr0> east europe
2015-01-31 23:00:42 <dalias> and how would they attack the system?
2015-01-31 23:01:04 <okwYotjAkvr0> well, once the .hu government got me drunk and separated me from my laptop for about 12h
2015-01-31 23:01:06 <dalias> probing ram on a live powered system is not impossible but it's also fragile and time-consuming
2015-01-31 23:01:41 <okwYotjAkvr0> also some of my clients are kinda sensible, and have been attacked by similar orgs
2015-01-31 23:01:51 <dalias> :-p
2015-01-31 23:01:52 <thiagowfx> okwYotjAkvr0: were your partitions encrypted?
2015-01-31 23:02:32 <dalias> and you could achieve similar protection with a safety kill
2015-01-31 23:02:40 <dalias> without losing the usability factor of suspend
2015-01-31 23:02:49 <okwYotjAkvr0> well i can boot in 2sec.
2015-01-31 23:02:51 <dalias> personally i find hibernate unusably slow
2015-01-31 23:02:58 <okwYotjAkvr0> that is equvivalent speed, and much improved sec
2015-01-31 23:03:08 <dalias> the saving side is slow. not the resume
2015-01-31 23:03:14 <okwYotjAkvr0> agreed.
2015-01-31 23:03:51 <dalias> and linux doesn't handle it well, because it insists on using swap rather than a dedicated hibernate backing
2015-01-31 23:03:59 <dalias> i generally don't enable swap
2015-01-31 23:04:14 <okwYotjAkvr0> but i stoped also hibernating with linux-3.13. as they do not have /sys/power/disk anymore, and i had no time to figure out what it had been replaced with.
2015-01-31 23:04:54 <okwYotjAkvr0> and poweroff is as convenient as is hibernate. so why bother.
2015-01-31 23:05:46 <dalias> hardly...
2015-01-31 23:05:57 <dalias> manual work to get back to where you left off is horrible ux
2015-01-31 23:06:18 <dalias> lots of browsers can save tabs for you..
2015-01-31 23:06:28 <okwYotjAkvr0> they do
2015-01-31 23:06:33 <dalias> but then they'll bog down the network and cpu reloading 20 things all at once
2015-01-31 23:06:45 <okwYotjAkvr0> lately firefox does lazy loading :/
2015-01-31 23:06:53 <okwYotjAkvr0> so only the active tab is reloaded.
2015-01-31 23:06:54 <dalias> yeah that's ugly in its own way tho
2015-01-31 23:07:02 <okwYotjAkvr0> much less ram, but also much less joy
2015-01-31 23:07:17 <okwYotjAkvr0> so many fuckups with the lazyloading
2015-01-31 23:07:19 <dalias> a good suspend could be secure
2015-01-31 23:07:23 <okwYotjAkvr0> i lost hundreds of tabs
2015-01-31 23:07:33 <okwYotjAkvr0> due to f....ing mozilla
2015-01-31 23:08:05 <dalias> it could encrypt ram and just leave a tiny resume program to accept as passphrase and decrypt on resume
2015-01-31 23:08:14 <okwYotjAkvr0> a good suspend could be secure if you disable your external ports, and run something non-supported by finfisher
2015-01-31 23:08:30 <okwYotjAkvr0> or, what you write ;)
2015-01-31 23:09:07 <okwYotjAkvr0> but if you do the encrypt ram thing, i think you also have to encrypt the kernel ram.
2015-01-31 23:09:11 <okwYotjAkvr0> with the fde keys.
2015-01-31 23:09:20 <dalias> yes
2015-01-31 23:09:34 <okwYotjAkvr0> doesnn't that mean you need take over control of the kernel?
2015-01-31 23:09:49 <dalias> well the kernel would be responsible for doing this
2015-01-31 23:09:52 <dalias> or a hypervisor
2015-01-31 23:10:46 <dalias> anyway threat model is very geographically specific, imo
2015-01-31 23:10:50 <okwYotjAkvr0> is it?
2015-01-31 23:11:01 <okwYotjAkvr0> my friends are also harrassed in the us, and the uk
2015-01-31 23:11:12 <okwYotjAkvr0> e.g.
2015-01-31 23:11:22 <okwYotjAkvr0> also i travel a lot to the east. similar
2015-01-31 23:11:39 <okwYotjAkvr0> but the threat is much less technically advanced the more east i go
2015-01-31 23:11:44 <dalias> :-p
2015-01-31 23:12:28 <okwYotjAkvr0> but then i guess if you get far enough in the east technology can get more advanced again. but ex soviet union is quite ok digitally, but physically, it's quite more difficult.
2015-01-31 23:12:36 <dalias> my experience is that in the us and similar the threat model is to make you fear legal hell if you don't 'voluntarily' comply rather than any technological attacks
2015-01-31 23:12:56 <okwYotjAkvr0> some of my clients have been actually attacked in azerbaidjan, and in kirgiztan. with malware
2015-01-31 23:13:11 <dalias> :/
2015-01-31 23:13:14 <okwYotjAkvr0> governemnt stuff
2015-01-31 23:13:19 <okwYotjAkvr0> has been documented in the press
2015-01-31 23:13:19 <dalias> fun
2015-01-31 23:13:45 <okwYotjAkvr0> it is not boring. i can tell you :)
2015-01-31 23:14:03 <okwYotjAkvr0> the shitting your pants part is the most exciting.
2015-01-31 23:17:14 <dalias> wow new inkscape released
2015-01-31 23:17:18 <dalias> wonder when it'll get packaged
2015-01-31 23:17:50 <systmkor> anyone else have better luck with Debian?
2015-01-31 23:18:32 <dalias> :-p
2015-01-31 23:20:02 <okwYotjAkvr0> /o\
2015-01-31 23:29:16 <systmkor> i guess I'm taking that as a no
2015-01-31 23:29:32 <okwYotjAkvr0> well the question was very generic.
2015-01-31 23:29:37 <okwYotjAkvr0> dalias: http://www.euractiv.com/infosociety/kroes-staff-computers-hacked-bak-news-515996
2015-01-31 23:29:44 <systmkor> true just was curious about general thoughts
2015-01-31 23:30:04 <okwYotjAkvr0> i was a very happy debian user until a few years ago.
2015-01-31 23:31:28 <okwYotjAkvr0> i think it started when the debian packaging purist decided to remove npm or so, because they said, there's dpkg in debian, we don't need no redundant packaging system. i know the guy who was responsible personally, we had an epic irc flame.
2015-01-31 23:32:26 <okwYotjAkvr0> i'm no happy nodejs user, i'm one who is forced to use it. and then the debian people come and make my life more miserable.
2015-01-31 23:33:31 <okwYotjAkvr0> then it got worse when they introduced patches that forced me to disable grsec protections on many binaries. because they need executable stacks, and whatnot
2015-01-31 23:33:48 <dalias> :)
2015-01-31 23:34:01 <okwYotjAkvr0> i submitted abug, but was rejected, that the problem is not existing on a pure debian install, only on ones with non-distro kernels.
2015-01-31 23:34:24 <dalias> you should have submitted it as a security weakness
2015-01-31 23:34:28 <okwYotjAkvr0> i did
2015-01-31 23:34:30 <okwYotjAkvr0> :)
2015-01-31 23:34:31 <dalias> rather than incompatibility with grsec
2015-01-31 23:35:08 <okwYotjAkvr0> somewhat later they introduced more of these patches.
2015-01-31 23:35:18 <okwYotjAkvr0> and linked almost everything to it.
2015-01-31 23:35:32 <okwYotjAkvr0> libffi and libglapi
2015-01-31 23:37:15 <okwYotjAkvr0> i'm actually very hopeful about jaromils devuan. but since i tried alpine, i'm not very interested in anything else. (i need to spend time to get emacs and ksh working with musl though)
2015-01-31 23:37:33 <okwYotjAkvr0> jaromil and dyne rocks!
2015-01-31 23:40:54 <systmkor> okwYotjAkvr0, there is a ksh port already
2015-01-31 23:41:08 <systmkor> oksh - its a Linux port of OpenBSDs ksh
2015-01-31 23:41:14 <okwYotjAkvr0> the real at&t ksh, or some of the bad knockoffs?
2015-01-31 23:41:18 <okwYotjAkvr0> meh. no good
2015-01-31 23:41:55 <okwYotjAkvr0> hmmm. do you know if it does ksh93?
2015-01-31 23:42:03 <systmkor> huh?
2015-01-31 23:42:16 <systmkor> real at&t ksh, or some bad knockoffs?
2015-01-31 23:42:16 <okwYotjAkvr0> (before i dismiss it as "meh, no good", i should be sure it's no good)
2015-01-31 23:42:36 <systmkor> which is no good
2015-01-31 23:42:54 <okwYotjAkvr0> well ksh93 is the standard, the other kshs do not implement it mostly. so are useless
2015-01-31 23:43:06 <systmkor> the standard by whom?
2015-01-31 23:43:23 <okwYotjAkvr0> david korn?
2015-01-31 23:43:28 <okwYotjAkvr0> and at&t
2015-01-31 23:43:30 <okwYotjAkvr0> labs
2015-01-31 23:43:42 <okwYotjAkvr0> the others are just reimplementations, bad ones.
2015-01-31 23:43:55 <okwYotjAkvr0> could as well use bash or ash
2015-01-31 23:44:14 <systmkor> well sure if you want to run 1983 code :P
2015-01-31 23:44:18 <okwYotjAkvr0> (well again, i dunno about oksh, but pdksh is bad)
2015-01-31 23:44:27 <okwYotjAkvr0> ksh is well maintained
2015-01-31 23:44:33 <okwYotjAkvr0> korn is still working on it.
2015-01-31 23:44:38 <systmkor> well oksh is in testing
2015-01-31 23:44:53 <systmkor> i was the one who ported it, I haven't done extensive testing
2015-01-31 23:44:58 <systmkor> I would appreciate testing :D
2015-01-31 23:46:04 <okwYotjAkvr0> sorry! good job! thanks!
2015-01-31 23:46:21 <okwYotjAkvr0> but i'm more into the genuine ksh.
2015-01-31 23:46:52 <systmkor> okwYotjAkvr0, http://www.connochaetos.org/oksh/
2015-01-31 23:46:58 <okwYotjAkvr0> the genuine ksh has last been updated 2012 not 1983
2015-01-31 23:47:18 <systmkor> i figured it's just was just to poke fun
2015-01-31 23:47:24 <okwYotjAkvr0> "The OpenBSD people have cleaned up and enhanced the original PD ksh"
2015-01-31 23:47:39 <okwYotjAkvr0> with great features. E. g.
2015-01-31 23:47:42 <okwYotjAkvr0> PS1=`\u@\h'
2015-01-31 23:47:43 <systmkor> well ksh is OpenBSDs default shell
2015-01-31 23:47:44 <okwYotjAkvr0> works like under bash (try this with a standard PD ksh).
2015-01-31 23:47:52 <okwYotjAkvr0> huh? that is surely a joke.
2015-01-31 23:47:59 <systmkor> no
2015-01-31 23:48:03 <okwYotjAkvr0> (i mean the "great features)
2015-01-31 23:48:06 <okwYotjAkvr0> not the defaultshell
2015-01-31 23:48:09 <systmkor> oh lol
2015-01-31 23:49:01 <systmkor> okwYotjAkvr0, well were is the source for the o.g. ksh?
2015-01-31 23:49:38 <okwYotjAkvr0> http://www2.research.att.com/~astopen/download/
2015-01-31 23:49:57 <okwYotjAkvr0> hmm, perhaps also here: http://kornshell.com/
2015-01-31 23:50:12 <okwYotjAkvr0> no, it's the same two pages.
2015-01-31 23:50:18 <okwYotjAkvr0> or very much connected
2015-01-31 23:50:18 <systmkor> i thought that was the public domain one, memory is wrong then
2015-01-31 23:50:32 <okwYotjAkvr0> no. the license actually is fucked up
2015-01-31 23:50:43 <systmkor> what's the license?
2015-01-31 23:50:49 <okwYotjAkvr0> read the first link, and be horrified.
2015-01-31 23:51:14 <okwYotjAkvr0> "Most AT&T ast software is licensed by the Eclipse Public License 1.0 (EPL-1.0). Before you download any packages you must signify your agreement to the terms of EPL-1.0 by entering a license agreement user name and password."
2015-01-31 23:51:37 <okwYotjAkvr0> fucked up at&t corporatism
2015-01-31 23:54:28 <dalias> :)
2015-01-31 23:57:32 <systmkor> okwYotjAkvr0, well do they validate user-name & password to anything?
2015-01-31 23:57:43 <systmkor> otherwise it's a meaningless string you give them
2015-01-31 23:58:13 <systmkor> sounds like the equivlant requirement of asking for blood donations as a user agreement
2015-01-31 23:58:13 <okwYotjAkvr0> i never licensed ksh properly myself. usually the distros i use take care of that. i guess this might change with alpine?
2015-01-31 23:58:17 <systmkor> seems b.s.
2015-01-31 23:58:33 <systmkor> no idea
2015-01-31 23:58:36 <okwYotjAkvr0> in the end, you just have to bundle the license in the binary pkg
2015-01-31 23:58:43 <systmkor> ahhh
2015-01-31 23:58:59 <systmkor> yah that's handled  by existing alpine packages
2015-01-31 23:59:18 <systmkor> "/usr/share/licenses/" i think