2013-04-18 07:25:13 <ncopa> vkrishn: http://dev.alpinelinux.org/irclogs
2013-04-18 07:25:43 <vkrishn> thanks :-)
2013-04-18 07:27:14 <vkrishn> thats interesting, I can use some ajax script to tail it
2013-04-18 07:27:16 <vkrishn> :)))
2013-04-18 07:27:39 <ncopa> clandmeter had some log tail js
2013-04-18 07:27:47 <clandmeter> we already have it for build log
2013-04-18 07:27:57 <vkrishn> yes I remember, its the same script I am refering
2013-04-18 07:28:09 <ncopa> or you could just use a normal irc client :)
2013-04-18 07:28:17 <vkrishn> it was supposed to be implemented somewhere
2013-04-18 07:28:20 <clandmeter> ncopa: do you care about history?
2013-04-18 07:28:28 <ncopa> history?
2013-04-18 07:28:32 <clandmeter> my irssi log file
2013-04-18 07:28:34 <clandmeter> :)
2013-04-18 07:28:37 <ncopa> nah
2013-04-18 07:28:42 <ncopa> well
2013-04-18 07:28:49 <vkrishn> I have unreliable 2G connection
2013-04-18 07:28:54 <clandmeter> youll have few years of history
2013-04-18 07:29:00 <ncopa> if it is similar format
2013-04-18 07:29:17 <vkrishn> yes please, previous logs welcome
2013-04-18 07:29:51 <ncopa> clandmeter: i think you can upload them to git.a.o:/var/www/localhost/htdocs/irclogs/
2013-04-18 07:30:00 <ncopa> i *think* you have write perms there
2013-04-18 07:31:05 <clandmeter> format example: http://sprunge.us/ZHaF
2013-04-18 07:31:37 <ncopa> create a dir named old or so
2013-04-18 07:31:41 <ncopa> and just upload them there
2013-04-18 07:32:10 <ncopa> i think this logger does not show when people join and leaves
2013-04-18 07:32:42 <vkrishn> my xchat logs show
2013-04-18 07:32:51 <ncopa> my xchat too
2013-04-18 07:33:07 <vkrishn> maybe its in the config of each app
2013-04-18 07:33:13 <vkrishn> tell it what to log
2013-04-18 07:34:00 <ncopa> sircbot does not care about it
2013-04-18 07:34:17 <ncopa> if we want/need that info we will have to find a better solution
2013-04-18 07:34:24 <ncopa> like eggdrop or so
2013-04-18 07:34:35 <ncopa> but for now i think this will do
2013-04-18 07:35:07 <vkrishn> for time being, I am happy, sometimes my xchat shows its logged on but does not send/receive message
2013-04-18 07:35:24 <vkrishn> leading to communication gap
2013-04-18 07:36:10 <vkrishn> thanks again
2013-04-18 07:36:44 <ncopa> np
2013-04-18 08:02:13 <vkrishn> now I pinned the two logs using firefox "Pin as App tab" ;)
2013-04-18 08:25:30 <build64-edge> failed to build v130313-314-g4dfe0f1. http://distfiles.alpinelinux.org/distfiles/build64-edge.txt
2013-04-18 08:55:38 <ncopa> that keys change broke the builders
2013-04-18 08:55:56 <ncopa> i wonder if it will break for upgraders
2013-04-18 09:07:03 <ncopa> algitbot: build master
2013-04-18 10:17:10 <vkrishn> any possibility of getting KDEmod ported to alpine?
2013-04-18 10:17:33 <vkrishn> read it improves kde performance
2013-04-18 10:57:46 <ncopa> i dont use kde personally
2013-04-18 10:57:53 <ncopa> so i will not spend time on it
2013-04-18 10:58:00 <ncopa> i am thinking of gnome3 though
2013-04-18 11:05:33 <vkrishn> v3 for upcoming 2.7 :)
2013-04-18 11:06:08 <vkrishn> for 2.6
2013-04-18 11:10:37 <ncopa> i kinda wanted that
2013-04-18 11:10:43 <ncopa> but i dont think i will make it
2013-04-18 11:12:40 <ncopa> algitbot: build master
2013-04-18 11:15:09 <ncopa> algitbot: build master
2013-04-18 11:19:43 <ncopa> algitbot: build master
2013-04-18 11:29:04 <ncopa> algitbot: build master
2013-04-18 11:29:49 <ncopa> algitbot: build master
2013-04-18 11:32:11 <vkrishn> OpenAMQ/1.4 allows amqp over http
2013-04-18 11:32:52 <vkrishn> but still marked as unstable
2013-04-18 11:35:55 <ncopa> i'm pretty happy with zeromq
2013-04-18 11:50:04 <vkrishn> still think a amqp based solution to be available v1.3 is stable
2013-04-18 11:50:18 <vkrishn> rabbitmq
2013-04-18 11:51:59 <vkrishn> stomp(websockets) is available
2013-04-18 12:12:47 <vkrishn> I noticed setting up a chroot sdk env , adduser sets default /bin/bash as shell
2013-04-18 12:12:55 <vkrishn> hope its not a bug
2013-04-18 12:17:57 <vkrishn> will having a ln bash -> /bin/busybox in /bin be a problem ?
2013-04-18 12:44:03 <clandmeter> busybox doesnt have bash. so that applet will probably not work
2013-04-18 13:05:41 <vkrishn> cannot recall at the moment, but have seen that link in some booting env..
2013-04-18 13:09:24 <vkrishn> for  the chroot either default shell be set to /bin/sh or bash need be installed
2013-04-18 13:31:48 <vkrishn> clandmeter, http://git.alpinelinux.org/irclogs/old is still planned ?
2013-04-18 14:04:25 <clandmeter> ill take a look at it
2013-04-18 15:25:56 <clandmeter> ncopa: are logfiles gone?
2013-04-18 15:26:54 <clandmeter> algitbot: build master
2013-04-18 22:51:48 <clandmeter> algitbot: build master
2013-04-19 04:56:24 <barthalion> Oh, I need to order him building my commits?
2013-04-19 04:56:25 <barthalion> Cool
2013-04-19 06:09:18 <ncopa> hi
2013-04-19 06:09:24 <ncopa> no you are not supposed to order him to do it
2013-04-19 06:09:30 <ncopa> it is supposed to be automatic
2013-04-19 06:15:08 <fcolista> hi all
2013-04-19 06:17:21 <ncopa> hi fcolista
2013-04-19 06:17:31 <fcolista> hi ncopa!
2013-04-19 06:17:38 <ncopa> i'm trying to debug why other users cannot forward build server messages on edge
2013-04-19 06:17:44 <ncopa> do you think you could help me test it?
2013-04-19 06:17:50 <ncopa> you can abump cifs-utils-6.0
2013-04-19 06:17:52 <fcolista> yes
2013-04-19 06:17:54 <ncopa> (its your pkg too)
2013-04-19 06:19:13 <fcolista> bumped, now's compiling
2013-04-19 06:20:41 <ncopa> interesting
2013-04-19 06:20:51 <fcolista> looks that it works
2013-04-19 06:20:52 <fcolista> :)
2013-04-19 06:20:53 <ncopa> did you get any error messages when you did git push?
2013-04-19 06:20:55 <ncopa> no
2013-04-19 06:20:58 <fcolista> no
2013-04-19 06:20:58 <ncopa> it does not work
2013-04-19 06:21:01 <ncopa> it does not build
2013-04-19 06:21:04 <ncopa> i think...
2013-04-19 06:21:09 <fcolista> no error in push
2013-04-19 06:21:25 <ncopa> the post-update hook never ran
2013-04-19 06:21:47 <fcolista> you're using autobuilder now right?
2013-04-19 06:21:53 <ncopa> yes
2013-04-19 06:22:00 <fcolista> k
2013-04-19 06:22:39 <ncopa> ok
2013-04-19 06:22:41 <fcolista> hwo can i help you?
2013-04-19 06:22:43 <ncopa> i will undo the push
2013-04-19 06:22:44 <fcolista> *how
2013-04-19 06:22:47 <ncopa> and you can try push again?
2013-04-19 06:22:50 <fcolista> sure
2013-04-19 06:24:10 <fcolista> tell me when i can do it
2013-04-19 06:24:55 <ncopa> can you try git push again?
2013-04-19 06:25:21 <ncopa> maybe git fetch first
2013-04-19 06:26:40 <fcolista> i read later git fetch, sorry. I pull --rebase, rebumped, checksummed, pushed
2013-04-19 06:26:44 <fcolista> it's ok?
2013-04-19 06:26:51 <ncopa> yes
2013-04-19 06:28:07 <fcolista> i've a meeting here @work. I'll be back in 30min
2013-04-19 06:28:13 <fcolista> it's a problem for you?
2013-04-19 06:28:24 <fcolista> I can miss the meeting if i've a good excues
2013-04-19 06:28:27 <fcolista> *excuse :)
2013-04-19 06:28:35 <fcolista> umh, build-edge works
2013-04-19 06:30:41 <ncopa> fcolista: can you please try git pull --rebase and git push again?
2013-04-19 07:15:49 <fcolista> done ncopa
2013-04-19 07:15:59 <fcolista> no log in irc
2013-04-19 07:20:40 <vkrishn> ncopa, noticed irclogger.com ?
2013-04-19 07:22:27 <fcolista> vkrishn, wow!
2013-04-19 07:22:35 <vkrishn> thanks, to Chris (cschneid) for accepting to log
2013-04-19 07:24:32 <ncopa> cool
2013-04-19 07:26:46 <ncopa> fcolista: i found the issue
2013-04-19 07:26:54 <ncopa>  ./post-update refs/heads/master
2013-04-19 07:26:54 <ncopa> -sh: ./post-update: Permission denied
2013-04-19 07:27:00 <fcolista> aah
2013-04-19 07:27:05 <fcolista> no +x ?
2013-04-19 07:27:08 <ncopa> yup
2013-04-19 07:27:10 <ncopa> well
2013-04-19 07:27:11 <fcolista> or other user ?
2013-04-19 07:27:17 <ncopa> correct
2013-04-19 07:27:45 <fcolista> k
2013-04-19 07:28:12 <ncopa> -rwxrwSr-x    1 ncopa    alpinede       241 Apr 19 06:26 post-update
2013-04-19 07:28:26 <fcolista> when you have time, i'm very very interesting to understand your zmq code for autobuilder
2013-04-19 07:28:55 <fcolista> S ?
2013-04-19 07:29:38 <ncopa> yeah thats wrong
2013-04-19 07:29:46 <fcolista> why group suid?
2013-04-19 07:29:47 <fcolista> Ah
2013-04-19 07:29:48 <fcolista> k
2013-04-19 07:30:07 <fcolista> no x permission
2013-04-19 07:30:11 <fcolista> for the group
2013-04-19 07:31:38 <ncopa> fcolista: can you try push again?
2013-04-19 07:31:43 <fcolista> sure
2013-04-19 07:31:45 <ncopa> git pull --rebase && git push
2013-04-19 07:32:11 <fcolista> i need to re-commit
2013-04-19 07:32:28 <ncopa> oh , sorry
2013-04-19 07:33:10 <fcolista> yup
2013-04-19 07:33:11 <ncopa> there we go
2013-04-19 07:33:15 <fcolista> \0/
2013-04-19 07:33:24 <ncopa> fcolista: thank you for your help debugging it
2013-04-19 07:33:38 <fcolista> thx for your gr8 job ncopa
2013-04-19 08:01:22 <ncopa> fcolista: re autobuilder:  http://zguide.zeromq.org/page:all
2013-04-19 08:02:49 <ncopa> the buildmsg sets up 2 zmq sockets
2013-04-19 08:03:02 <ncopa> 1 request-reply
2013-04-19 08:03:06 <ncopa> like this: http://zguide.zeromq.org/page:all#Ask-and-Ye-Shall-Receive
2013-04-19 08:03:46 <ncopa> the client (buildmsg-send) sends a command to the buildmsg-server
2013-04-19 08:04:08 <ncopa> command can be either 'build', 'log' or 'irc'
2013-04-19 08:04:39 <ncopa> the buildmsg-server sets up yet another zmq socket, a publish-subscribe
2013-04-19 08:04:46 <ncopa> like this: http://zguide.zeromq.org/page:all#Getting-the-Message-Out
2013-04-19 08:05:07 <fcolista> you implemented then a client7server infrastructure with zmq
2013-04-19 08:05:10 <ncopa> the subscribers (buildservers) can subscribe to messages
2013-04-19 08:05:13 <fcolista> s/7//
2013-04-19 08:05:37 <fcolista> looks very easy
2013-04-19 08:05:42 <ncopa> i implemented both client/server and publisher/subscriber
2013-04-19 08:05:47 <ncopa> it is very easy
2013-04-19 08:06:34 <fcolista> you send a string like 'build' and when server receive this string it do an action
2013-04-19 08:06:35 <ncopa> [client] -> [server/publisher] -> [subscriber]
2013-04-19 08:06:50 <ncopa> client is git hook
2013-04-19 08:07:03 <ncopa> (buildmsg-send)
2013-04-19 08:07:17 <ncopa> builmsg-server is server/publisher
2013-04-19 08:07:26 <ncopa> and the build server are the subscribers
2013-04-19 08:07:43 <ncopa> a subscriber can set a filter on what messages to recieve
2013-04-19 08:07:53 <fcolista> you can have n build server that subscribes to the server
2013-04-19 08:08:00 <ncopa> correct
2013-04-19 08:08:09 <ncopa> but the git hook also sends what git branch
2013-04-19 08:08:25 <ncopa> it sends:  "build <git branch>"
2013-04-19 08:08:38 <fcolista> that's very interesting
2013-04-19 08:08:41 <ncopa> so you can set up a build server to subscribe to a specific git branch
2013-04-19 08:09:10 <ncopa> buildmsg-subscribe 'build 2.5-stable'
2013-04-19 08:09:33 <ncopa> and it will only get the notification when anything is pushed to git 2.5-stable branch
2013-04-19 08:09:35 <fcolista> very very nice
2013-04-19 08:09:40 <ncopa> now
2013-04-19 08:09:55 <ncopa> you can on any box, including your workstation subscribe
2013-04-19 08:10:03 <ncopa> try git clone autobuilder
2013-04-19 08:10:06 <fcolista> :-)
2013-04-19 08:10:14 <ncopa> and run make && sudo make install
2013-04-19 08:10:19 <fcolista> i'm going to do it
2013-04-19 08:10:26 <ncopa> apk add lua-zmq lua-posix
2013-04-19 08:10:36 <ncopa> actually, you dont nee make install to test it
2013-04-19 08:11:04 <fcolista> k
2013-04-19 08:11:14 <ncopa> lua buildmsg-subscribe --publisher=tcp://git.alpinelinux.org:55556 'build'
2013-04-19 08:11:35 <ncopa> and you should get all notifications from all git branches
2013-04-19 08:11:54 <ncopa> or if you only want the notification that comes to irc:
2013-04-19 08:12:04 <ncopa> lua buildmsg-subscribe --publisher=tcp://git.alpinelinux.org:55556 'irc'
2013-04-19 08:12:15 <ncopa> or if you want the log:
2013-04-19 08:12:24 <ncopa> lua buildmsg-subscribe --publisher=tcp://git.alpinelinux.org:55556 'log'
2013-04-19 08:13:11 <vkrishn> ncopa, sub connection does require logging ?
2013-04-19 08:13:29 <ncopa> subscriber?
2013-04-19 08:13:32 <ncopa> no
2013-04-19 08:13:39 <fcolista> i'm going to see log with conky
2013-04-19 08:13:47 <vkrishn> so I can use an zmq client to listen to it?
2013-04-19 08:13:59 <ncopa> vkrishn: ues
2013-04-19 08:14:00 <ncopa> yes
2013-04-19 08:14:26 <vkrishn> conky :)
2013-04-19 08:14:49 <vkrishn> sounds like a easy to setup a dev monitoring station
2013-04-19 08:14:55 <ncopa> correct
2013-04-19 08:15:56 <vkrishn> cool, next step would be controlling using smart mobile app
2013-04-19 08:16:14 <ncopa> you can trigger build command if you want
2013-04-19 08:16:19 <ncopa> like you do from here
2013-04-19 08:16:28 <vkrishn> or may smart remote, like ones by logitech
2013-04-19 08:16:33 <vkrishn> maybe*
2013-04-19 08:16:42 <ncopa> hm
2013-04-19 08:16:50 <ncopa> 'log' does not work
2013-04-19 08:17:47 <vkrishn> smart remote controller*
2013-04-19 08:19:00 <ncopa> log will not show all compile log
2013-04-19 08:19:35 <ncopa> i was thinking adding support for buildmsg in abuild
2013-04-19 08:19:55 <ncopa> so you could get a notify for each package it starts to build
2013-04-19 08:20:27 <ncopa> maybe like
2013-04-19 08:20:45 <ncopa> "<pkg>: fetching sources"
2013-04-19 08:20:52 <ncopa> "<pkg>: building"
2013-04-19 08:21:04 <ncopa> "<pkg>: packaging"
2013-04-19 08:21:44 <vkrishn> is it possible to add a trigger to get verbose
2013-04-19 08:21:51 <vkrishn> like log -vvv
2013-04-19 08:22:16 <ncopa> you mean all the compile output?
2013-04-19 08:22:34 <ncopa> it is possible, but i dont know if we want that....
2013-04-19 08:22:36 <vkrishn> for logs
2013-04-19 08:23:03 <vkrishn> it may be muted like -q
2013-04-19 08:23:58 <vkrishn> so occasionally it needed these can be done
2013-04-19 08:24:03 <vkrishn> if*
2013-04-19 08:24:15 <ncopa> i dont know if we want each compile line passing by the messaging server
2013-04-19 08:24:18 <ncopa> hm
2013-04-19 08:24:48 <vkrishn> it depends on v(nums)
2013-04-19 08:24:52 <vkrishn> vv, v, vvv
2013-04-19 08:24:55 <vkrishn> :)
2013-04-19 08:25:12 <ncopa> the build servers will send it always
2013-04-19 08:25:26 <ncopa> you filter on what you subscribe to
2013-04-19 08:26:51 <vkrishn> filter at client end?
2013-04-19 08:27:17 <ncopa> yes
2013-04-19 08:27:30 <ncopa> well the suscriber tells the buildmsg-server what messages it wants
2013-04-19 08:27:34 <vkrishn> so data still travel, though unintended
2013-04-19 08:27:53 <ncopa> data will travel [buildserver] -> [message server]
2013-04-19 08:28:00 <vkrishn> no I mean when you send a build command
2013-04-19 08:28:11 <vkrishn> pub
2013-04-19 08:28:13 <ncopa> aha
2013-04-19 08:28:20 <ncopa> the git hook does that
2013-04-19 08:28:28 <ncopa> you are not supposed to do it manually
2013-04-19 08:28:31 <ncopa> but sure...
2013-04-19 08:28:56 <vkrishn> sometimes you may be want to know what went wrong
2013-04-19 08:28:57 <ncopa> it is possible
2013-04-19 08:29:21 <ncopa> so you want make it build again but sending output to a zmq listener this time
2013-04-19 08:29:43 <ncopa> currently the build logs will endhere: http://dev.alpinelinux.org/buildlogs/
2013-04-19 08:30:02 <ncopa> and if there is an error, it will make a copy with a timestamp in the filename
2013-04-19 08:30:40 <ncopa> algitbot should give an url with the url to the permanent logfile
2013-04-19 08:34:08 <vkrishn> will filter in subscribe be an overdo?
2013-04-19 08:34:16 <vkrishn> lua buildmsg-subscribe --publisher=tcp://git.alpinelinux.org:55556 'irc -v'
2013-04-19 08:34:53 <ncopa> that will only show messages that are send to server with the prefix 'irc -v'
2013-04-19 08:35:14 <ncopa> messages that starts with 'irc -v ....'
2013-04-19 08:35:42 <ncopa> i would have it send each line from build log to msg server
2013-04-19 08:36:03 <ncopa> but i dont think its worth it
2013-04-19 08:36:08 <ncopa> you have the full build log from http
2013-04-19 08:36:16 <vkrishn> moment,  buildmsg-subscribe is method to get feedback
2013-04-19 08:36:27 <ncopa> yes
2013-04-19 08:36:38 <vkrishn> just a thought
2013-04-19 08:38:25 <vkrishn> I was thinking like send me "irc notification by verbose level 1"
2013-04-19 08:38:34 <ncopa> understand
2013-04-19 08:38:36 <vkrishn> far-fetched
2013-04-19 08:38:42 <vkrishn> idea
2013-04-19 08:38:43 <ncopa> that is not really how zmq works
2013-04-19 08:41:12 <ncopa> ha
2013-04-19 08:41:13 <ncopa> works now
2013-04-19 08:41:18 <ncopa> http://dev.alpinelinux.org/buildlogs/buildmsg.log
2013-04-19 08:42:17 <ncopa> running buildmsg-subscribe locally gave me this:
2013-04-19 08:42:20 <ncopa> http://dpaste.org/ubdfB/
2013-04-19 08:43:36 <vkrishn> :)
2013-04-19 08:44:19 <ncopa> i think i want the apkbuild's 'msg' in there too
2013-04-19 08:45:01 <vkrishn> what does keyword "log" do?
2013-04-19 08:45:36 <ncopa> it will save it in http://dev.alpinelinux.org/buildlogs/buildmsg.log and forward to subscribers
2013-04-19 08:46:38 <vkrishn> so can't there be a preprocessor to this, like a route
2013-04-19 08:47:29 <vkrishn> so "log-v" would do some filtering and save it to buildmsg-L1.log
2013-04-19 08:47:58 <vkrishn> then forward it
2013-04-19 08:53:05 <ncopa> vkrishn: it could
2013-04-19 08:53:30 <ncopa> but someone needs to send the 'log-v ...' message
2013-04-19 08:53:35 <vkrishn> but it may be a slight heavy on server
2013-04-19 08:53:53 <ncopa> it should be able to handle it
2013-04-19 08:54:18 <vkrishn> I was thinking like 'log-error'
2013-04-19 08:54:19 <ncopa> i think zmq is designed for high freq trading
2013-04-19 08:54:23 <vkrishn> show only error lines
2013-04-19 08:54:32 <vkrishn> good for mobile devices
2013-04-19 08:54:46 <ncopa> could also send error messsages prefixed with error
2013-04-19 08:54:49 <vkrishn> like a short flag/line
2013-04-19 08:55:02 <ncopa> so you could subscribe for: 'log error'
2013-04-19 08:55:10 <vkrishn> :)
2013-04-19 08:55:15 <ncopa> currently the messages are prefixed by hostname
2013-04-19 08:58:13 <vkrishn> I think now I would check a bit  php/react's zmq implementation
2013-04-19 08:58:30 <vkrishn> see if a web interface for the logs is easy
2013-04-19 08:59:00 <vkrishn> may be something like this also for acf
2013-04-19 08:59:19 <vkrishn> acf for developers
2013-04-19 08:59:23 <vkrishn> ;)
2013-04-19 09:41:40 <ncopa> fcolista: command suitable for conky: ./buildmsg-subscribe --publisher=tcp://git.alpineli
2013-04-19 09:41:40 <ncopa> nux.org:55556 'log'
2013-04-19 15:02:23 <ncopa> algitbot: build master
2013-04-19 15:02:42 <ncopa> cool
2013-04-19 15:02:46 <ncopa> have a nice weekend
2013-04-19 15:14:25 <kaniini> :D
2013-04-19 17:09:47 <kaniini> hmm
2013-04-19 18:38:29 <barthalion> \m/
2013-04-21 20:36:21 <barthalion> Hm…
2013-04-21 20:36:34 <barthalion> Have you tried to boot Alpine iso from grub2?
2013-04-21 20:37:29 <barthalion> I'm stuck at mounting roots, not sure if it's matter of alpine_dev or modules
2013-04-21 20:45:52 <barthalion> Ok, rather alpine_dev
2013-04-21 20:46:02 <barthalion> But still no idea what I should set
2013-04-21 21:02:00 <barthalion> I tried UUID of partition where iso is placed and filename
2013-04-22 07:11:23 <build64-2-5> files from v2.5.4-47-g53bca70 uploaded
2013-04-22 07:16:03 <ncopa> clandmeter: i think there was some of the xbmc deps that had upgrades that broke abi
2013-04-22 07:16:17 <ncopa> libcec-2.1.1
2013-04-22 07:17:45 <clandmeter> not sure what you mean. did you upgrade libcec already?
2013-04-22 07:17:51 <ncopa> no
2013-04-22 07:17:54 <ncopa> i did not upgrade it
2013-04-22 07:18:04 <ncopa> because an upgrade would require a rebuild of xbmc
2013-04-22 07:18:14 <clandmeter> oh ok
2013-04-22 07:18:17 <clandmeter> let me take a look
2013-04-22 07:18:33 <ncopa> i mean that we should have upgraded that in same shot
2013-04-22 07:18:42 <ncopa> but its always easy to be after-wise :)
2013-04-22 07:18:57 <ncopa> we will have to rebuild xbmc once again as r1
2013-04-22 07:19:07 <ncopa> it does not matter
2013-04-22 07:19:14 <ncopa> just a little extra work
2013-04-22 07:19:15 <clandmeter> i understand. i didnt look at the deps.
2013-04-22 07:19:45 <ncopa> i am upgrading gcc to 4.7.3 now
2013-04-22 07:19:59 <ncopa> cleaning up the subpackage a bit
2013-04-22 07:20:02 <ncopa> better gcc-objc
2013-04-22 07:20:20 <clandmeter> there was talk about it here some days ago
2013-04-22 07:20:24 <ncopa> stripping dbug info from static libs
2013-04-22 07:20:26 <ncopa> ok?
2013-04-22 07:20:49 <ncopa> i think we mentioned that we will wait with gcc-4.8
2013-04-22 07:21:45 <clandmeter> i think kaniini and barthalion were talking about it. but i dont remember the details anymore.
2013-04-22 07:23:58 <ncopa> we have a handful broken deps
2013-04-22 07:24:01 <ncopa> http://dpaste.org/nPUga/
2013-04-22 07:24:05 <ncopa> we need fix those asap
2013-04-22 07:24:29 <clandmeter> should i upgrade libva?
2013-04-22 07:24:43 <ncopa> vill it break abi?
2013-04-22 07:24:47 <vkrishn> clandmeter, xbmc moved from unstable to main?
2013-04-22 07:24:49 <ncopa> i think we should upgrade it anyways
2013-04-22 07:25:06 <ncopa> clandmeter: run checkapk after bumping it
2013-04-22 07:25:30 <clandmeter> vkrishn: yes
2013-04-22 07:25:36 <ncopa> check i SONAME change from libsomething.so.x to libsomething.so.x+1
2013-04-22 07:25:48 <clandmeter> yes ill build it right now
2013-04-22 07:25:52 <ncopa> thanks
2013-04-22 07:26:07 <ncopa> i also saw that there is a gstreamer plugin for libvaapi or something
2013-04-22 07:26:23 <ncopa> http://www.phoronix.com/scan.php?page=news_item&px=MTM1NjA
2013-04-22 07:27:24 <ncopa> those are the circular deps we must fix this week: http://dpaste.org/3PO1p/
2013-04-22 07:27:39 <ncopa> i am working libobjc as we speak
2013-04-22 07:28:10 <ncopa> rnalrd: do you think you have a chance to look at freeradius?
2013-04-22 07:28:16 <ncopa> if not, let me know and i'll do it
2013-04-22 07:28:18 <rnalrd> k
2013-04-22 07:28:44 <ncopa> i think there are 2 circular deps there
2013-04-22 07:28:55 <ncopa> freeradius-libs -> freeradius-libs
2013-04-22 07:29:21 <ncopa> and freeradius-libs -> freeradius-radclient -> freeradius-libs
2013-04-22 07:29:46 <ncopa> it sounds like it is just to remove freeradius-libs from depends and let abuild autodetect it
2013-04-22 07:30:47 <ncopa> yes, try remove freeradius-lib from all "depend="..."
2013-04-22 07:43:14 <clandmeter> libva has no so name diff. I guess i can push it.
2013-04-22 07:49:18 <ncopa> yes
2013-04-22 07:56:44 <hl> kaniini: will alpine ever support selinux and is that a stupid idea?
2013-04-22 07:57:20 <ncopa> hl: most likely never
2013-04-22 07:58:21 <ncopa> why not is doc'ed here: http://grsecurity.net/lsm.php
2013-04-22 07:59:00 <ncopa> ugh
2013-04-22 07:59:07 <ncopa> i need to stop the 2-6 builder
2013-04-22 07:59:44 <clandmeter> so soon we'll have 2 gpu's to accelerate on :)
2013-04-22 08:01:53 <hl> hmm
2013-04-22 08:01:58 <hl> that page says that grsec is compatible with other lsms
2013-04-22 08:11:08 <clandmeter> ncopa: buildboxes have ccache enabled?
2013-04-22 08:15:09 <ncopa> clandmeter: no
2013-04-22 08:15:16 <ncopa> no point with that
2013-04-22 08:15:26 <ncopa> they are supposed to build packages one time only
2013-04-22 08:16:06 <clandmeter> in a perfect world you would be right.
2013-04-22 09:18:13 <vkrishn> ncopa, zmq irc subscription from command line does not show lines with aports:master ?
2013-04-22 09:19:51 <vkrishn> I am testing zmp-php and have current subscribed to 'irc' and  'log'
2013-04-22 09:19:57 <ncopa> it show the lines as they are published
2013-04-22 09:20:44 <ncopa> buildmsg-send "build ${refname##refs/heads/}"
2013-04-22 09:21:27 <ncopa> an thats the message to build boxes
2013-04-22 09:21:48 <vkrishn> I was thinking for another sub like 'aports',
2013-04-22 09:22:05 <vkrishn> That way a web-interface would be possible for live feeds
2013-04-22 09:22:21 <ncopa> what should be published there?
2013-04-22 09:22:24 <ncopa> and from who?
2013-04-22 09:23:31 <vkrishn> lines like. quote from algibot as above " aports:master | Natanael Copa..."
2013-04-22 09:25:06 <vkrishn> I could then build a web-page with a globe/map showing live commits from various geographical region, in LIVE
2013-04-22 09:25:29 <ncopa> ok
2013-04-22 09:25:39 <ncopa> a per commit message in other words
2013-04-22 09:25:40 <ncopa> ok
2013-04-22 09:25:48 <vkrishn> yes
2013-04-22 09:26:34 <vkrishn> since IP is required to get geolocation, currently commiters name can be hard-coded to geolocation
2013-04-22 09:26:58 <ncopa> what happens if i push a commit when i am on vacation :)
2013-04-22 09:27:29 <vkrishn> yes, possible
2013-04-22 09:27:37 <vkrishn> approaches are:
2013-04-22 09:27:59 <vkrishn> 1. update your new geolocation in some web accessible file
2013-04-22 09:28:11 <vkrishn> 2. also add ip to sub
2013-04-22 09:30:00 <ncopa> the algitbot commit messages here does not go via buildmsg
2013-04-22 09:30:04 <ncopa> but i can change that
2013-04-22 09:30:42 <vkrishn> is sub "log error" useful
2013-04-22 09:30:45 <ncopa> so you can catch those with an 'irc' subscriber
2013-04-22 09:31:08 <vkrishn> 'aports'
2013-04-22 09:31:58 <vkrishn> I understand that there would not be and build-servers consuming these subs
2013-04-22 09:32:08 <vkrishn> any* build..
2013-04-22 09:32:25 <vkrishn> same for 'log error'
2013-04-22 09:37:02 <vkrishn> ofcourse, when on vacation you could add location as <hidden-do-not-disturb>
2013-04-22 09:37:04 <vkrishn> ;)
2013-04-22 09:40:28 <vkrishn> got it in sub ;)
2013-04-22 09:41:22 <ncopa> i changed the git commit hook script
2013-04-22 09:42:19 <ncopa> lets see if it works...
2013-04-22 09:42:38 <vkrishn> got it
2013-04-22 10:00:58 <ncopa> algitbot: build master
2013-04-22 13:38:24 <barthalion> clandmeter, ncopa: correct, kaniini said that we will stay on 4.7.2
2013-04-22 13:38:37 <ncopa> i just pushed 4.7.3
2013-04-22 13:38:41 <barthalion> Cool :)
2013-04-22 13:38:46 <ncopa> we wil stay on 4.7.x
2013-04-22 13:38:56 <ncopa> not do 4.8
2013-04-22 13:39:11 <ncopa> at least for v2.6 release
2013-04-22 13:39:28 <barthalion> I thought so, but I had too weak CPU to test it properly
2013-04-22 13:39:42 <barthalion> And btw I'll ask again: someone managed to boot Alpine iso from grub2?
2013-04-22 13:42:33 <ncopa> i never bothered to even try grub2
2013-04-22 13:46:12 <barthalion> I'm missing something in alpine_dev
2013-04-22 13:46:39 <barthalion> I'm not sure where grub2 "mounts" iso
2013-04-22 14:37:40 <kaniini> hl: i generally don't have interest in selinux -- it seems bloated and overdesigned.
2013-04-22 14:51:41 <vkrishn> ncopa, would it be a problem if I leave my php zmq client on(listening) 'log' + 'irc' ?
2013-04-22 14:51:53 <ncopa> i think not
2013-04-22 14:52:08 <ncopa> thanks for asking
2013-04-22 14:52:20 <vkrishn> :)
2013-04-22 14:52:51 <vkrishn> I noticed - though it shows its connected but shows no msgs
2013-04-22 14:53:23 <vkrishn> is it bcoz that your end server restarted
2013-04-22 14:53:37 <ncopa> i dont think so
2013-04-22 14:53:42 <vkrishn> tcp connection shows its alive
2013-04-22 14:53:47 <ncopa> i think zmq should nadle that i restart server
2013-04-22 14:53:53 <ncopa> handle*
2013-04-22 14:54:43 <vkrishn> yes, that's what is, while I was reading the docs
2013-04-22 15:31:35 <vkrishn> will there be any official email/wikipage for gcc to clang switch ?
2013-04-22 16:22:29 <bpiotrowski> Any thoughts on moving to libusbx?
2013-04-22 18:43:20 <build64-2-5> files from v2.5.4-48-gb0da569 uploaded
2013-04-22 18:57:36 <kaniini> hl: however, i will say that AppArmor is interesting
2013-04-22 18:57:48 <kaniini> hl: and we technically already ship the modules
2013-04-22 18:57:54 <hl> kaniini: oh? some stackoverflow post said it was leaky or something
2013-04-22 18:58:12 <hl> kaniini: grsec's rbac thing looks apparmor-ish
2013-04-22 18:58:24 <kaniini> it is, but it is also not stable
2013-04-22 18:58:31 <hl> it isn't? damn.
2013-04-22 18:58:34 <kaniini> every time grsec is updated you have to go update EVERY profile
2013-04-22 18:58:39 <kaniini> that is why i gave up on it
2013-04-22 18:59:05 <hl> also, why exactly is pax not merged into the kernel
2013-04-22 22:52:13 <kaniini> hl: because linus dislikes it
2013-04-23 07:51:21 <ncopa> <hl> also, why exactly is pax not merged into the kernel
2013-04-23 07:51:35 <ncopa> hl: linus dislikes it because it is too big and too intrusive
2013-04-23 07:52:01 <ncopa> you cannot really split it up in small acceptable step by step patches
2013-04-23 10:36:47 <clandmeter> seems overlayfs will get merged into 3.10
2013-04-23 10:38:18 <ncopa> algitbot: build master
2013-04-23 11:39:45 <ncopa> algitbot: build master
2013-04-23 12:22:33 <clandmeter> ncopa: i recently tried to build my own iso with alpine-iso. It seems something has changed? it now wants to sign the apkindex but after boot it fails to verify it.
2013-04-23 12:24:01 <ncopa> hum
2013-04-23 12:24:17 <ncopa> probably the alpine-keys package split
2013-04-23 12:25:32 <clandmeter> so alpine keys are not included?
2013-04-23 12:26:23 <ncopa> i thought they got pulled in automatic
2013-04-23 12:26:28 <ncopa> as a dep
2013-04-23 12:26:47 <ncopa> atleast when i tested to build the iso it worked
2013-04-23 12:27:01 <ncopa> but that is the only thing that has recently changed that i think could affect it
2013-04-23 12:28:08 <clandmeter> i guess thats it, i dont see it in apk dir.
2013-04-23 12:29:07 <ncopa> i thought it shoudl be pulled in as dep for alpine-base
2013-04-23 12:30:59 <clandmeter> hmm there is seomting wrong with the deps i think
2013-04-23 12:32:05 <clandmeter> it just grabbed 800 deps in alpine-mini
2013-04-23 12:32:17 <clandmeter> and then borked out
2013-04-23 12:34:12 <clandmeter> typo in profile name :)
2013-04-23 17:30:42 <hl> So I figured out why postconf -m was segfaulting.
2013-04-23 17:30:59 <hl> It looks like the machine code generated truncates a pointer later passed to a printf-like function.
2013-04-23 17:31:45 <hl> show_maps() at postconf_other.c:78, call to concatenate, which returns a char*, is followed by instruction 'cltq', which sign-extends EAX to RAX. Ego, a 64-bit pointer is truncated to a 32-bit sign-extended pointer.
2013-04-23 17:34:18 <hl> Hmm... I wonder...
2013-04-23 17:38:03 <hl> Fixed.
2013-04-23 17:38:57 <hl> postconf_other.c unlike all other postconf_*.c files did not #include <stringops.h>, which meant that the string-concatenation function concatenate() was left undefined, so gcc assumed a default prototype, incorrectly.
2013-04-23 17:39:26 <kaniini> patch? :D
2013-04-23 18:11:44 <hl> kaniini: sent
2013-04-23 18:17:07 <kaniini> ok
2013-04-24 06:21:17 <ncopa> kaniini: do you mind if i wipe out gconf and gnome-vfs from gstreamer 0.10?
2013-04-24 06:21:29 <ncopa> i think we might need purge gconf
2013-04-24 06:21:41 <ncopa> orbit2 no longer builds
2013-04-24 06:23:21 <ncopa> libbonobo no longer builds
2013-04-24 06:23:33 <ncopa> activation-server-main.c:71:39: error: 'G_STATIC_REC_MUTEX_INIT'
2013-04-24 06:24:29 <ncopa> sigh... looks like we *must* move on to gnome3
2013-04-24 06:32:31 <ncopa> no
2013-04-24 06:32:33 <ncopa> not yet
2013-04-24 06:32:39 <ncopa> i found a patch for orbit2
2013-04-24 09:55:18 <build64-2-5> files from v2.5.4-49-g069962f uploaded
2013-04-24 12:53:22 <vkrishn> ncopa, will this lib be useful to add, http://libzfl.zeromq.org
2013-04-24 12:53:54 <ncopa> probably
2013-04-24 12:54:01 <ncopa> but im busy with the v2.6 release now
2013-04-24 12:54:15 <vkrishn> :)
2013-04-24 12:57:29 <vkrishn> would be nice if anyone could have a look and if useful, make it to 2.6
2013-04-24 13:41:17 <build-2-4> files from v2.4.10-35-gc5ffd88 uploaded
2013-04-24 13:41:30 <build64-2-4> files from v2.4.10-35-gc5ffd88 uploaded
2013-04-24 13:41:42 <build-2-5> files from v2.5.4-50-g43ad5a8 uploaded
2013-04-24 13:46:39 <bpiotrowski> What is the policy on packaging software from git?
2013-04-24 13:58:12 <ncopa> bpiotrowski: we try avoid snapshots if possible
2013-04-24 13:58:21 <ncopa> you can use a git tag
2013-04-24 13:58:46 <ncopa> then we normally make a tarball that we upload on http://dev.alpinelinux.org/archive
2013-04-24 13:58:54 <ncopa> (basically we make the release for upstream)
2013-04-24 13:59:05 <build-2-2> files from v2.2.3-221-g5a33772 uploaded
2013-04-24 13:59:07 <build64-2-2> files from v2.2.3-221-g5a33772 uploaded
2013-04-24 13:59:21 <bpiotrowski> Well, I don't see any useful release of libzfl
2013-04-24 13:59:33 <ncopa> use <latestrelease>_git<snapshotdate>
2013-04-24 13:59:45 <bpiotrowski> aye
2013-04-24 13:59:59 <bpiotrowski> vkrishn: I'll package it to the end of week
2013-04-24 14:00:05 <ncopa> https://github.com/zeromq/libzfl/tags
2013-04-24 14:00:31 <bpiotrowski> ncopa: 3 years ago, and "build but has no functionality"
2013-04-24 14:00:37 <ncopa> lol yes
2013-04-24 14:00:41 <ncopa> so
2013-04-24 14:00:46 <bpiotrowski> That's why I asked
2013-04-24 14:00:56 <ncopa> use 0.1.0_git<date>
2013-04-24 14:00:59 <bpiotrowski> so, 0.1.0_git2013042X
2013-04-24 14:01:05 <ncopa> yes
2013-04-24 14:01:40 <ncopa> we have some apkbuilds that creates a snapshot
2013-04-24 14:01:56 <clandmeter> abuild supports snapshots
2013-04-24 14:01:57 <vkrishn> thanks, note czmq may be needed for it
2013-04-24 14:02:01 <clandmeter> abuild snapshot
2013-04-24 14:02:16 <bpiotrowski> Is it documented on wiki?
2013-04-24 14:02:21 <ncopa> hm no
2013-04-24 14:02:23 <ncopa> i think not
2013-04-24 14:02:25 <ncopa> we should do that
2013-04-24 14:02:28 <clandmeter> no but its inside abuild help
2013-04-24 14:02:37 <bpiotrowski> clandmeter: Thanks, I'll take a look
2013-04-24 14:02:39 <ncopa> we should doc the magic git vars
2013-04-24 14:02:47 <clandmeter> yes i think so
2013-04-24 14:03:16 <clandmeter> we can copy the comment from abuild and put it on wiki
2013-04-24 14:03:20 <ncopa> http://git.alpinelinux.org/cgit/abuild/tree/abuild.in#n1690
2013-04-24 14:03:47 <bpiotrowski> Cool
2013-04-24 14:03:54 <bpiotrowski> Much better idea than in makepkg
2013-04-24 14:04:40 <ncopa> giturl
2013-04-24 14:04:41 <ncopa> verbase
2013-04-24 14:04:48 <ncopa> disturl
2013-04-24 14:05:32 <clandmeter> the comment has a mistake: basever should be verbase
2013-04-24 14:11:30 <vkrishn> looks like v2.6 is all geared for building a cool HTPC
2013-04-24 14:11:49 <vkrishn> maybe we code name v2.6 as htpc ;)
2013-04-24 14:26:03 <ncopa> atm the v2.6 is all geared for building
2013-04-24 14:26:09 <ncopa> :)
2013-04-24 14:26:12 <ncopa> as in compiling
2013-04-24 14:28:55 <vkrishn> :)
2013-04-24 14:41:11 <bpiotrowski> ncopa: Would it be big issue if I push the newest phpmyadmin to 2.6, but after its release?
2013-04-24 14:41:58 <bpiotrowski> APKBUILD needs major rewrite, I didn't have time to do it yet
2013-04-24 14:42:31 <bpiotrowski> And current version available in supported released is like swiss cheese
2013-04-24 14:45:18 <ncopa> are there any release candidates?
2013-04-24 14:45:46 <bpiotrowski> No, we ship obsolete version
2013-04-24 14:45:55 <ncopa> ah
2013-04-24 14:45:55 <bpiotrowski> very obsolete
2013-04-24 14:45:56 <ncopa> ok
2013-04-24 14:46:00 <ncopa> then we should upgrade it asap
2013-04-24 14:50:45 <ncopa> matt smith is the maintainer
2013-04-24 14:50:49 <ncopa> missing in action
2013-04-24 14:51:13 <ncopa> bpiotrowski, feel free to take maintainership of phpmyadmin
2013-04-24 14:51:18 <ncopa> specially if you use it
2013-04-24 14:51:43 <bpiotrowski> Great
2013-04-24 14:52:09 <bpiotrowski> I'll try to push it on the weekend
2013-04-24 14:55:52 <ncopa> as long as you reach it before May 7 i'm ok
2013-04-24 14:57:14 <bpiotrowski> Long enough :)
2013-04-24 14:57:45 <ncopa> thanks!
2013-04-24 17:57:22 <bpiotrowski> ncopa: #1273 can be closed
2013-04-24 17:57:56 <bpiotrowski> ncopa: #1154 can be moved to 2.7 I guess
2013-04-24 17:58:19 <bpiotrowski> ncopa: #1484 should be closed as well
2013-04-24 18:01:23 <bpiotrowski> ncopa: same with #1752, #1537 and #1484
2013-04-24 18:01:36 <kaniini> ok
2013-04-24 18:01:50 <bpiotrowski> Shouldn't resolved bugs get automatically closed?
2013-04-24 18:03:00 <kaniini> yes
2013-04-24 18:04:22 <kaniini> i created an alpine 2.7.0 milestone
2013-04-24 18:04:27 <kaniini> and assigned 1154 to it
2013-04-24 18:05:23 <bpiotrowski> kaniini: Assign 897 too
2013-04-24 18:11:07 <kaniini> done
2013-04-24 18:11:13 <kaniini> (shouldn't you have this permission?)
2013-04-24 18:12:05 <bpiotrowski> Maybe, I don't anyway :p
2013-04-24 18:23:13 <kaniini> oh.  perhaps it's because you also don't have @ :P
2013-04-24 18:24:29 <bpiotrowski> Just push access, it's fine with me ;)
2013-04-24 18:24:54 <kaniini> yeah.  you don't have any special privilege on redmine
2013-04-24 18:25:02 <kaniini> hmm.
2013-04-24 19:24:58 <vkrishn> bpiotrowski, looks CZMQ lib ma be worth adding
2013-04-24 19:25:22 <vkrishn> libzfl, has not be updated for quite a while
2013-04-24 19:26:09 <vkrishn> http://article.gmane.org/gmane.network.zeromq.devel/13959/match=libzfl
2013-04-24 19:26:30 <vkrishn> CZMQ lib may* be...
2013-04-24 19:26:56 <bpiotrowski> vkrishn: I see
2013-04-24 19:27:23 <bpiotrowski> vkrishn: I'm not directly interested in zeromq packages, so I'll package what you see valuable enough
2013-04-24 19:27:56 <vkrishn> I was toying with filemq app and came across libzfl
2013-04-24 20:36:37 <build-2-2> files from v2.2.3-222-gd1ea0b9 uploaded
2013-04-24 20:36:41 <build-2-3> files from v2.3.6-194-g98f1625 uploaded
2013-04-24 20:37:10 <build64-2-4> files from v2.4.10-36-g3e5610d uploaded
2013-04-24 21:16:19 <build64-2-2> files from v2.2.3-222-gd1ea0b9 uploaded
2013-04-25 06:35:15 <ncopa> bpiotrowski: thanks for looking at the bugs
2013-04-25 06:35:20 <ncopa> kaniini, you too
2013-04-25 06:35:53 <ncopa> does bpiotrowski need more powers on bugs.a.o?
2013-04-25 06:37:23 <ncopa> bpiotrowski: I added you as developer for alpine linux
2013-04-25 07:58:31 <build64-2-5> files from v2.5.4-51-g58d6ca6 uploaded
2013-04-25 07:58:37 <build-2-5> files from v2.5.4-51-g58d6ca6 uploaded
2013-04-25 08:24:31 <ncopa> i know what the problem is
2013-04-25 08:24:52 <ncopa> the problem is tcl
2013-04-25 08:25:01 <ncopa> it silently fails to do make install
2013-04-25 08:42:19 <fcolista> I think that this can be closed: http://bugs.alpinelinux.org/issues/1341
2013-04-25 08:42:36 <fcolista> or we can consider closed only when it is in "main" ?
2013-04-25 08:57:32 <build-2-3> files from v2.3.6-195-g521799b uploaded
2013-04-25 08:59:30 <build64-2-2> files from v2.2.3-223-gf1087a9 uploaded
2013-04-25 08:59:31 <build-2-2> files from v2.2.3-223-gf1087a9 uploaded
2013-04-25 11:03:55 <rnalrd> build64-edge:retry
2013-04-25 11:09:52 <ncopa> rnalrd: the new syntax is: algitbot: build master
2013-04-25 11:10:07 <rnalrd> tnx
2013-04-25 11:10:16 <rnalrd> algitbot: build master
2013-04-25 11:10:41 <rnalrd> not getting output to http://distfiles.alpinelinux.org/log/edge/x86_64/
2013-04-25 11:11:43 <rnalrd> mmm
2013-04-25 11:12:05 <ncopa> log file location moved
2013-04-25 11:12:28 <ncopa> http://dev.alpinelinux.org/buildlogs/
2013-04-25 11:14:23 <ncopa> auth_poly1305_53.c:(.text+0x68): undefined reference to `fegetround'
2013-04-25 11:14:23 <rnalrd> fcolista: ../libsodium/src/libsodium/.libs/libsodium.a(libsodium_la-auth_poly1305_53.o): In function `crypto_onetimeauth_poly1305_53':
2013-04-25 11:14:23 <rnalrd> auth_poly1305_53.c:(.text+0x68): undefined reference to `fegetround'
2013-04-25 11:14:23 <rnalrd> collect2: error: ld returned 1 exit status
2013-04-25 11:14:26 <rnalrd> --
2013-04-25 11:14:41 <ncopa> smells like uclibc problem
2013-04-25 11:14:46 <rnalrd> yup
2013-04-25 11:15:11 <fcolista> umh
2013-04-25 11:15:42 <ncopa> does it link to libm?
2013-04-25 11:16:12 <fcolista> ncaop, i had to apk upgrade with latest alpine packages
2013-04-25 11:16:13 <fcolista> edge
2013-04-25 11:16:23 <ncopa> libtool: link: gcc -std=gnu99 -Wall -Wextra -Wbad-function-cast -Wcast-align -W
2013-04-25 11:16:24 <ncopa> cast-qual -Wchar-subscripts -Wcomment -Wfloat-equal -Wformat=2 -Wimplicit -Wmis
2013-04-25 11:16:24 <ncopa> heses -Wpointer-arith -Wredundant-decls -Wstrict-prototypes -Wswitch-enum -Os -
2013-04-25 11:16:24 <ncopa> sing-declarations -Wmissing-prototypes -Wnormalized=id -Woverride-init -Wparent
2013-04-25 11:16:24 <ncopa> fomit-frame-pointer -fPIC -fPIC -fPIE -fwrapv -fno-strict-aliasing -fno-strict-
2013-04-25 11:16:24 <ncopa> overflow -fstack-protector-all -Winit-self -Wwrite-strings -Wdiv-by-zero -Wl,--
2013-04-25 11:16:26 <ncopa> as-needed -pie -Wl,-z -Wl,relro -Wl,-z -Wl,now -Wl,-z -Wl,noexecstack -o dnscry
2013-04-25 11:16:30 <ncopa> pt-proxy app.o cert.o dnscrypt.o dnscrypt_client.o edns.o logger.o options.o pi
2013-04-25 11:16:34 <ncopa> d_file.o safe_rw.o sandboxes.o stack_trace.o tcp_request.o udp_request.o utils.
2013-04-25 11:16:36 <ncopa> o windows_service.o  -L/usr/local/lib ../libevent-modified/.libs/libevent_extra
2013-04-25 11:16:38 <ncopa> .a ../libevent-modified/.libs/libevent_core.a ../libsodium/src/libsodium/.libs/
2013-04-25 11:16:40 <ncopa> libsodium.a -ldl -lm
2013-04-25 11:16:42 <ncopa> yes it links with libm
2013-04-25 11:16:46 <ncopa> he
2013-04-25 11:16:48 <ncopa> ha
2013-04-25 11:39:59 <ncopa> someone should check that fegetround/fesetround actually works on x86_64
2013-04-25 11:40:03 <ncopa> i think it does
2013-04-25 11:41:21 <ncopa> i really hate the  GConf2 schema stuff
2013-04-25 11:41:43 <ncopa> how is it possible that they implemented something that goes so slow, even on an i7?
2013-04-25 11:58:58 <build64-2-5> files from v2.5.4-52-g44d6bf4 uploaded
2013-04-25 11:58:58 <build-2-5> files from v2.5.4-52-g44d6bf4 uploaded
2013-04-25 12:24:42 <build-2-2> files from v2.2.3-224-g2bcdd42 uploaded
2013-04-25 12:24:49 <build64-2-2> files from v2.2.3-224-g2bcdd42 uploaded
2013-04-25 12:34:37 <rnalrd> ncopa, is spamming my mailbox ;-P
2013-04-25 12:36:15 <ncopa> ah
2013-04-25 12:36:22 <ncopa> changing category...
2013-04-25 12:36:24 <ncopa> sorry
2013-04-25 12:36:42 <jbilyk> ncopa: it's ok, you only generated about 100msgs in < 1min :D
2013-04-25 12:59:55 <ncopa> here comes 80 more...
2013-04-25 13:11:54 <build-2-4> files from v2.4.10-39-g89d02e1 uploaded
2013-04-25 13:11:54 <build64-2-4> files from v2.4.10-39-g89d02e1 uploaded
2013-04-25 13:16:52 <build64-2-2> files from v2.2.3-225-ge0bf41b uploaded
2013-04-25 13:16:52 <build-2-2> files from v2.2.3-225-ge0bf41b uploaded
2013-04-25 13:41:37 <build64-2-4> files from v2.4.10-41-gb588317 uploaded
2013-04-25 13:41:58 <build-2-4> files from v2.4.10-41-gb588317 uploaded
2013-04-25 13:42:15 <build-2-3> files from v2.3.6-199-gdba5b77 uploaded
2013-04-25 13:43:44 <build64-2-5> files from v2.5.4-53-gbe6b131 uploaded
2013-04-25 14:30:58 <ncopa> hmpf
2013-04-25 14:31:05 <ncopa> i thikn i have yet another gcc fix
2013-04-25 16:22:53 <vkrishn> during some tests varnish(kvm setup) hangs
2013-04-25 16:23:02 <vkrishn> I was looking at , http://en.wikibooks.org/wiki/Grsecurity/Application-specific_Settings
2013-04-25 16:23:16 <vkrishn> does varnish need any extra settings
2013-04-25 16:23:19 <vkrishn> ?
2013-04-25 16:24:32 <vkrishn> also is it applicable to other php apps like opcaches', memcache ?
2013-04-25 16:24:51 <vkrishn> hmm... php-apc
2013-04-25 16:26:50 <vkrishn> If I recall correctly, varnish docs also states memory usage to be MUCH more that what is being set
2013-04-25 18:00:18 <Barthalion> ncopa: Thank you
2013-04-25 20:18:29 <Barthalion> Well… if you want pacman for anything involving .install scripts, add bash to deps and set --with-scriptlet-shell=/usr/bin/bash
2013-04-25 20:19:12 <kaniini> i do not
2013-04-25 20:19:23 <kaniini> it is solely for bootstrapping arch chroots
2013-04-25 20:19:40 <Barthalion> So you will stumble upon package using install script
2013-04-25 20:19:49 <Barthalion> And "our" *.install are full of bashisms
2013-04-25 20:19:50 <kaniini> why would we need to add bash to outer deps
2013-04-25 20:20:14 <Barthalion> Because you need to set default interpreter for them
2013-04-25 20:21:02 <kaniini> your scripts will run inside the chroot if you use --root right?
2013-04-25 20:21:18 <kaniini> Barthalion...
2013-04-25 20:21:20 <kaniini> your scripts
2013-04-25 20:21:25 <kaniini> aren't going to be running inside alpine
2013-04-25 20:21:30 <kaniini> they are going to be running in the chroot
2013-04-25 20:21:43 <kaniini> so, the fact that our /bin/sh is not bash should not matteer
2013-04-25 20:22:14 <Barthalion> Just because you set another --root, doesn't mean that you are chrooted into it
2013-04-25 20:28:50 <kaniini> Barthalion: according to pacman docs, scriptlets are always executed in chroot
2013-04-25 20:34:00 <kaniini> wtf
2013-04-25 20:35:27 <Barthalion> Then I'm surprised that finally Allan made it sane way
2013-04-25 20:48:04 <kaniini> ok
2013-04-25 20:48:08 <kaniini> lets test it
2013-04-25 21:17:49 <Barthalion> huh
2013-04-25 21:18:58 <Barthalion> algitbot: retry
2013-04-25 21:19:21 <Barthalion> algitbot: build master
2013-04-26 06:16:12 <ncopa> is tehre anything that uses libunwind?
2013-04-26 06:16:34 <kaniini> not sure
2013-04-26 06:16:37 <kaniini> why ?
2013-04-26 06:16:45 <ncopa> does not build
2013-04-26 06:17:23 <ncopa> and its x86_64 only
2013-04-26 06:17:30 <kaniini> nuke it
2013-04-26 06:21:16 <ncopa> cool! we got pacman!
2013-04-26 06:21:28 <ncopa> algitbot: build master
2013-04-26 06:21:58 <kaniini> ncopa: yeah, for bootstrapping arch chroots
2013-04-26 06:22:09 <ncopa> does archlinux work in lxc?
2013-04-26 06:22:15 <kaniini> it works in xen
2013-04-26 06:22:20 <kaniini> i do not know baout lxc
2013-04-26 06:22:28 <Barthalion> Should work in lxc too
2013-04-26 06:22:43 <ncopa> that would be very nice
2013-04-26 06:22:51 <ncopa> i have issue with debootstrap
2013-04-26 06:23:01 <kaniini> ncopa: what issue?
2013-04-26 06:23:22 <kaniini> you do have to turn off the grsecurity chroot stuff
2013-04-26 06:23:42 <ncopa> i thought i did
2013-04-26 06:23:46 <ncopa> i will test now
2013-04-26 06:24:01 <kaniini> debootstrap is working as expected for me
2013-04-26 06:25:35 <ncopa> maybe it was the grsec chroot that bints me
2013-04-26 06:25:38 <ncopa> bites*
2013-04-26 06:26:21 <ncopa> do i have to disable all of chroot_*?
2013-04-26 06:29:34 <ncopa> yup: http://dpaste.org/kdDmG/
2013-04-26 06:44:48 <ncopa> same happens with all in /proc/sys/kernel/grsecurity/chroot_* set to 0
2013-04-26 06:50:39 <ncopa> i think debootstrap needs to depend on gnu tar
2013-04-26 06:50:49 <kaniini> ah yes
2013-04-26 06:50:51 <kaniini> it does
2013-04-26 08:28:08 <ncopa> hi
2013-04-26 08:28:22 <ncopa> i wonder if we can remove the pkgrel from gcc -v
2013-04-26 08:28:41 <ncopa> --with-pkgversion='Alpine 4.7.3-r5'
2013-04-26 08:29:37 <ncopa> as it is now, we have to rebuild kernel and all 3rd party kernel mods each time we rebuild gcc
2013-04-26 08:30:11 <ncopa> c1: error: fail to initialize plugin /usr/src/linux-headers-3.8.8-0-grsec/tools/gcc/colorize_plugin.so
2013-04-26 08:30:12 <ncopa> cc1: error: incompatible gcc/plugin versions
2013-04-26 08:30:12 <ncopa> cc1: error: fail to initialize plugin /usr/src/linux-headers-3.8.8-0-grsec/tools/gcc/structleak_plugin.so
2013-04-26 08:30:24 <ncopa> and all i have done is moving files around in the gcc package
2013-04-26 08:31:56 <ncopa> fabled: are you ok with me removing the -r$pkgrel from gcc -v?
2013-04-26 08:32:33 <fabled> i think so, yes
2013-04-26 08:33:14 <ncopa> i'm also figthing openjdk6 but i think i have it under control
2013-04-26 08:33:24 <ncopa> some more paxctl was needed
2013-04-26 08:33:28 <ncopa> and earlier
2013-04-26 08:34:45 <ncopa> i think i need paxctl another binary in gcc-java to be able to update openjdk in the stable branches
2013-04-26 08:34:55 <ncopa> but that currently means that i need to rebuild kernel
2013-04-26 08:35:04 <ncopa> so gcc plugin version matches
2013-04-26 09:51:31 <ncopa> i'm going home early today
2013-04-26 10:01:14 <ncopa> i'm going home
2013-04-26 10:01:18 <ncopa> see you monday
2013-04-26 10:34:06 <ncopa> algitbot: build master
2013-04-26 17:09:59 <kaniini> ncopa: i just pushed a critical fix to syslinux
2013-04-26 17:10:34 <kaniini> ncopa: if you could get the edge builders working again, it'd be very appreciated :)
2013-04-26 18:59:58 <Barthalion> kaniini: If you have correct permissions, could you close 1537 and 1540?
2013-04-26 19:03:30 <kaniini> don't have the permissions there, those are for individual software
2013-04-26 19:14:16 <ncopa> algitbot: build master
2013-04-26 19:16:19 <ncopa> algitbot: build master
2013-04-26 19:30:18 <build64-2-5> files from v2.5.4-54-gedb91a6 uploaded
2013-04-27 10:43:19 <Barthalion> hrm
2013-04-27 10:43:22 <Barthalion> algitbot: build master
2013-04-27 10:43:42 <Barthalion> Weird, it's building well locally
2013-04-28 00:24:37 <Kabaka> Hi. I can't get qemu (qemu-system-x86_64) to start as root. It spins around forever under _dl_find_hash according to gdb. :-\
2013-04-28 00:25:07 <kaniini> that is odd, it does not happen for me
2013-04-28 00:26:15 <Kabaka> It's not my args or anything either
2013-04-28 00:26:16 <Kabaka> [kabaka@draco ~] $ sudo qemu-system-x86_64
2013-04-28 00:26:17 <Kabaka> [sudo] password for kabaka:
2013-04-28 00:26:18 <Kabaka> ^C
2013-04-28 00:26:25 <Kabaka> It hangs even with nothing passed to it  :-(
2013-04-28 00:30:11 <kaniini> hmm...
2013-04-28 00:30:17 <kaniini> what libc0.9.32 version
2013-04-28 00:30:23 <kaniini> also, what cpu
2013-04-28 00:33:06 <Kabaka> kaniini: versions of everything are what is currently on edge. CPU is Athlon II X4 645
2013-04-28 00:34:14 <kaniini> does it start as a user?
2013-04-28 00:34:20 <Kabaka> Yes
2013-04-28 00:34:31 <kaniini> why do you n eed to use it as root then? :P
2013-04-28 00:34:36 <Kabaka> USB passthrough
2013-04-28 00:34:45 <Kabaka> Unless you have a solution for non-root access to that
2013-04-28 00:34:46 <Kabaka> :-P
2013-04-28 00:34:52 <kaniini> permissions on /dev/usb/*
2013-04-28 00:36:00 <Kabaka> No such thing. qemu says: husb: opendir /sys/bus/usb/devices: Permission denied
2013-04-28 00:36:08 <Kabaka> I don't think futzing with /sys permissions is a great solution
2013-04-28 00:36:26 <kaniini> hmm.
2013-04-28 00:56:10 <kaniini> Kabaka: try setuid root
2013-04-28 01:00:29 <Kabaka> kaniini: no change
2013-04-28 01:02:12 <kaniini> even running as user ?
2013-04-28 01:02:24 <Kabaka> Wait
2013-04-28 01:02:29 <Kabaka> I ran it about 6 times in a row
2013-04-28 01:02:34 <Kabaka> And it worked on the last time
2013-04-28 01:03:25 <Kabaka> This is similar to the behavior it had a few months ago. It would hang about 80% of the time, but would be fine once it started.
2013-04-28 01:03:51 <Kabaka> I always meant to troubleshoot that. Never remembered when I had time.
2013-04-28 01:18:07 <Kabaka> Weirdly, it is now starting 100% reliably. I guess this is okay, then. :-)
2013-04-28 03:01:47 <kaniini> hmm
2013-04-28 03:01:55 <kaniini> it could be randomization from pax
2013-04-28 03:02:07 <kaniini> fucking with uclibc rtld
2013-04-29 06:37:45 <fabled> ncopa, any idea why my SDL apps seem to hang on opening a window (x86)
2013-04-29 06:38:08 <ncopa> hm
2013-04-29 06:38:12 <fabled> happens with upstream apps like kvm/qemu too
2013-04-29 06:38:23 <kaniini> oh hey!
2013-04-29 06:38:27 <kaniini> someone reported a problem with
2013-04-29 06:38:35 <kaniini> uhm
2013-04-29 06:38:39 <ncopa> morning
2013-04-29 06:38:47 <ncopa> fabled: i have seen something similar
2013-04-29 06:38:52 <kaniini> qemu being stuck in _dl_find_hash()
2013-04-29 06:38:54 <kaniini> on amd64
2013-04-29 06:39:02 <kaniini> Kabaka: ^
2013-04-29 06:39:02 <ncopa> it was related uclibc dlopen i think
2013-04-29 06:39:14 <Kabaka> Yes
2013-04-29 06:39:15 <ncopa> qemu? i use it here daily
2013-04-29 06:39:21 <Kabaka> When starting qemu with sudo
2013-04-29 06:39:46 <kaniini> also, do we want to shoot for linux 3.9?
2013-04-29 06:40:11 <ncopa> are there grsecurity patches for it?
2013-04-29 06:40:25 <kaniini> i think we can roll some with rebasing
2013-04-29 06:40:28 <ncopa> are 3.8 or 3.9 announced to be LTS?
2013-04-29 06:40:38 <kaniini> neither thus far
2013-04-29 06:40:45 <ncopa> ok
2013-04-29 06:41:23 <ncopa> 1577908171b44e3babe248df18af2ddaec85e2cc
2013-04-29 06:41:28 <ncopa> i wonder if that broke sdl
2013-04-29 06:41:34 <kaniini> ``I would like to announce that the 3.8 Linux kernel series is NOT going to be a longterm stable kernel release. I will NOT be maintaining it for long time, and in fact, will stop maintaining it right after the 3.9 kernel is released''
2013-04-29 06:41:46 <kaniini> that answers the question re: 3.8
2013-04-29 06:41:46 <ncopa> ok
2013-04-29 06:42:02 <ncopa> we should probably aim for 3.9 then :-/
2013-04-29 06:42:14 <kaniini> http://www.kroah.com/log/linux/3.8-is_not_longterm_stable.html
2013-04-29 06:42:34 <ncopa> tricky
2013-04-29 06:42:39 <ncopa> 3.9 is not yet out
2013-04-29 06:42:44 <ncopa> so no grsecurity patch
2013-04-29 06:42:54 <ncopa> so we have not had any chance to do any testing
2013-04-29 06:42:59 <ncopa> at the time we can start test it
2013-04-29 06:43:07 <ncopa> the 3.8 is EOL
2013-04-29 06:43:31 <kaniini> 3.9.0 is out
2013-04-29 06:43:39 <ncopa> i think we are pretty much forced to try aim for LTS kernels
2013-04-29 06:43:57 <ncopa> https://www.kernel.org/ says 3.9-rc8
2013-04-29 06:44:34 <ncopa> ah 3.9.0 was tagged 6 hours ago
2013-04-29 06:45:06 <ncopa> yes we should aim for 3.9
2013-04-29 06:45:18 <ncopa> sigh..
2013-04-29 06:46:12 <ncopa> fabled: re sdl
2013-04-29 06:46:17 <ncopa> http://git.alpinelinux.org/cgit/aports/commit/main/sdl/APKBUILD?id=be91866229a76e73f1ecdbdbfec83b8e69cb6e69
2013-04-29 06:46:46 <ncopa> does that look familiar?
2013-04-29 06:47:05 <Kabaka> Hey, that's identical to the bt I got from qemu
2013-04-29 06:47:11 <fabled> right
2013-04-29 06:47:28 <ncopa> clandmeter: you reverted the workaround here: http://git.alpinelinux.org/cgit/aports/commit/main/sdl/APKBUILD?id=1577908171b44e3babe248df18af2ddaec85e2cc
2013-04-29 06:49:11 <ncopa> Kabaka or fabled: do you think you could explain what happens on bugs.a.o?
2013-04-29 06:49:35 <ncopa> then i can revert the clandmeters change with a reference to a bug id
2013-04-29 06:49:49 <kaniini> or perhaps he changed it for a reason?
2013-04-29 06:50:06 <kaniini> i think the real fix is to add recursion guard in _dl_find_hash().
2013-04-29 06:50:08 <fabled> ncopa, actually, i have a hunch
2013-04-29 06:50:19 <ncopa> he should have explained that reason in the commit message
2013-04-29 06:50:35 <ncopa> kaniini: yes you are right, should fix uclibc
2013-04-29 06:50:43 <fabled> it might be my patch that broke it ;)I
2013-04-29 06:51:03 <ncopa> d6df4153c8ef0bdc3b380151d954a94a123dc4dc ?
2013-04-29 06:51:24 <kaniini> also, build64-edge was segfaulting on compiling java... i tried to build it by hand and it segfaulted for me too (dual E5-2620 with 128gb ram server)
2013-04-29 06:51:26 <fabled> no
2013-04-29 06:52:19 <kaniini> hmm... i might actually be able to secure resources at my new company for glibc-based alpine
2013-04-29 06:52:31 <kaniini> we're having to run flash media server
2013-04-29 06:52:34 <fabled> ncopa, uclibc commit f69319d
2013-04-29 06:52:40 <kaniini> for some clients, which is sadness
2013-04-29 06:52:47 <kaniini> because we're having to do this on CentOS
2013-04-29 06:52:57 <kaniini> (i mean, that's just sadness from day 1)
2013-04-29 06:53:07 <fabled> dlsym and dlopen both take lock
2013-04-29 06:53:11 <fabled> which is likely not recursive
2013-04-29 06:53:23 <ncopa> ambiguous argument 'f69319d'
2013-04-29 06:53:49 <kaniini> fabled: that was the fix for openjdk on amd64, right?
2013-04-29 06:53:49 <ncopa> and gdb build broke
2013-04-29 06:54:17 <ncopa> i messed with openjdk last week
2013-04-29 06:54:26 <ncopa> did icedtea6-1.11.11
2013-04-29 06:54:32 <fabled> ncopa, full commit hash f69319d5a7d3a3ccb46b28ee2b0fd9053c6415ac
2013-04-29 06:54:44 <fabled> kaniini, no, that was different
2013-04-29 06:54:50 <kaniini> i vaguely recall us attacking some problem with rtld for openjdk
2013-04-29 06:54:52 <fabled> i think the lock was for vlc which uses dlopen() in multiple threads
2013-04-29 06:54:53 <kaniini> aha
2013-04-29 06:55:13 <fabled> the openjdk commits included $origin support
2013-04-29 06:55:24 <kaniini> maybe we need to make userspace-RCU in rtld a release goal for 2.7
2013-04-29 06:55:58 <fabled> it still uses PTHREAD_RECURSIVE_MUTEX_INITIALIZER_NP, though
2013-04-29 06:56:04 <fabled> not sure if it's honored in uclibc internal mutex
2013-04-29 06:56:31 <kaniini> as i understand it, uclibc internal mutexes are unrelated to pthread, and the pthread stuff wouldn't be available to us in rtld anyway because it's too early
2013-04-29 06:56:53 <ncopa> should i disable-sdl-open meanwhile?
2013-04-29 06:57:16 <kaniini> also, i do believe this is a recursion-related race because Kabaka mentioned he got it to work after a few tries
2013-04-29 06:57:39 <fabled> kaniini, same observation here. it works once in a while.
2013-04-29 06:57:43 <kaniini> so, fabled's analysis seems on the money to me
2013-04-29 06:57:47 <fabled> so definitely  race
2013-04-29 06:57:49 <Kabaka> Yes
2013-04-29 06:58:06 <ncopa> the --disable-sdl-open will force sdl to link directly
2013-04-29 06:58:06 <fabled> the internal macro expands to __pthread_mutex_lock
2013-04-29 06:58:11 <kaniini> odd...
2013-04-29 06:58:24 <kaniini> ncopa: does that induce an ABI break?
2013-04-29 06:58:48 <ncopa> hm
2013-04-29 06:59:11 <ncopa> not sure
2013-04-29 06:59:13 <ncopa> i think not
2013-04-29 06:59:32 <ncopa> i think sdl does tricks to avoid direct linking
2013-04-29 06:59:39 <ncopa> using dlopen
2013-04-29 06:59:46 <kaniini> yes, using a stub it seems.
2013-04-29 07:00:03 <kaniini> (likely to avoid ABI breaks)
2013-04-29 07:01:16 <kaniini> well, if fabled's analysis is right, then reverting the patch on a local build of uclibc should resolve it
2013-04-29 07:01:17 <ncopa> ok i'm gonna look at gdb
2013-04-29 07:01:25 <fabled> would be a good test
2013-04-29 07:01:33 <kaniini> which, i assume fabled is already working on :P
2013-04-29 07:03:16 <fabled> yes
2013-04-29 07:03:19 <fabled> i'll try it out
2013-04-29 07:03:39 <fabled> too sad that then vlc is broken again
2013-04-29 07:03:45 <fabled> need figure out something there
2013-04-29 07:03:50 <ncopa> heh
2013-04-29 07:04:02 <fabled> assuming that is the problem
2013-04-29 07:04:07 <ncopa> i'd rather keep vlc working and --disable-sdl-open in sdl
2013-04-29 07:04:37 <kaniini> i'd like to --disable-sdl-open, keep vlc working, and fix the underlying bug
2013-04-29 07:04:39 <kaniini> :P
2013-04-29 07:04:45 <ncopa> lol
2013-04-29 07:04:47 <fabled> :D
2013-04-29 07:04:59 <ncopa> yes, i agree :)
2013-04-29 07:07:41 <kaniini> https://bitbucket.org/tortoiselabs/appliancekit-ng/src/5481efb874fa28ddcd5d97772f2b765ae63eebc8/appliancekit/specs/arch-base.spec?at=master
2013-04-29 07:07:47 <kaniini> this is the steps needed to bring up arch now
2013-04-29 07:07:59 <kaniini> it is surely nice that they standardized their distribution; too bad it was systemd :P
2013-04-29 07:09:03 <ncopa> gdb-7.6 build appears to be broken
2013-04-29 07:09:10 <kaniini> build log ?
2013-04-29 07:09:18 <ncopa> i wonder if gdb devs even tested it before releasing it
2013-04-29 07:09:26 <ncopa> checking for msgmerge... no
2013-04-29 07:09:26 <ncopa> /home/ncopa/aports/main/gdb/src/gdb-7.6/opcodes/configure: .: line 12678: can't open './configure.host'
2013-04-29 07:09:26 <ncopa> make[1]: *** [configure-opcodes] Error 1
2013-04-29 07:09:36 <ncopa> looks like autoconf issue
2013-04-29 07:09:40 <kaniini> yes
2013-04-29 07:09:45 <kaniini> autoreconf may work
2013-04-29 07:10:04 <kaniini> or, hmm
2013-04-29 07:10:06 <kaniini> no
2013-04-29 07:19:38 <fabled> ncopa, kaniini, it's not locking issue. it's some hashing issue. i have full bt with debug info. will take a look.
2013-04-29 07:21:29 <kaniini> humm
2013-04-29 07:21:35 <kaniini> alright
2013-04-29 07:22:21 <kaniini> ncopa: and of course, debugging autoconf scripts is so great
2013-04-29 07:22:34 <ncopa> yeah.. :-/
2013-04-29 07:22:44 <ncopa> fits perfect with thei cold, rainy monday morning..
2013-04-29 07:22:48 <ncopa> this*
2013-04-29 07:25:01 <kaniini> ugh!
2013-04-29 07:25:14 <kaniini> someone complained to me that their ipv6 was not working on their vps
2013-04-29 07:25:28 <kaniini> i look at openvswitch to find out their vlan does not have ipv6 configuration pushed to it
2013-04-29 07:25:43 <kaniini> "oh, i thought just running 'reboot' would be enough"
2013-04-29 07:25:54 <kaniini> that was like an hour of my time wasted
2013-04-29 07:26:15 <ncopa> sounds like windows users
2013-04-29 07:26:31 <ncopa> :q
2013-04-29 07:26:36 <kaniini> ironically, alpine is easier to learn for those users than gnu/linux
2013-04-29 07:26:55 <kaniini> i guess we win for usability over ubuntu server ;)
2013-04-29 07:27:03 <ncopa> :)
2013-04-29 07:29:43 <ncopa> hm
2013-04-29 07:29:48 <ncopa> gentoo has gdb-7.6
2013-04-29 07:29:55 <ncopa> but no extra patch
2013-04-29 07:30:39 <kaniini> check portage for bug
2013-04-29 07:30:45 <kaniini> erm, gentoo bugzilla
2013-04-29 07:33:08 <kaniini> you know, i don't really like rinse(1) very much
2013-04-29 07:33:33 <kaniini> i should replace it with something in lua
2013-04-29 07:34:02 <ncopa> rinse?
2013-04-29 07:34:23 <kaniini> debootstrap for rpm distributions
2013-04-29 07:34:30 <kaniini> i packaged it alongside rpm
2013-04-29 07:45:25 <ncopa> after working around the autoconf issue in gdb i bump ito the same as gentoo hardened
2013-04-29 07:45:28 <ncopa> https://bugs.gentoo.org/show_bug.cgi?id=467806
2013-04-29 07:46:09 <ncopa> bah its a stupid cpuid asm
2013-04-29 07:46:58 <kaniini> that reminds me, i would like to switch to clang in 2.7 maybe
2013-04-29 07:47:02 <kaniini> who knows
2013-04-29 08:13:06 <ncopa> seems like the zeromq based build message server have been fairly stable
2013-04-29 08:13:21 <ncopa> maybe should port it to v2.5
2013-04-29 08:17:08 <fabled> ncopa, kaniini: ok, i figured out the root cause of the sdl dlopen hang; will try to fix it
2013-04-29 08:17:25 <ncopa> fabled: youre the man!
2013-04-29 08:57:11 <fabled> ncopa, http://sprunge.us/iDEZ
2013-04-29 08:57:23 <fabled> should fix the sdl hang
2013-04-29 08:57:28 <fabled> it was list corruption
2013-04-29 08:57:35 <fabled> i'm off for lunch
2013-04-29 08:58:06 <fabled> (do note, this bug affects other apps that do lot of dlopen(), dlclose() calls with complicated deps)
2013-04-29 08:58:19 <ncopa> ok
2013-04-29 08:58:23 <ncopa> I'll add it
2013-04-29 08:58:24 <ncopa> thanks!
2013-04-29 09:28:23 <ncopa> hm
2013-04-29 09:28:38 <ncopa> that kamailio change broke more things for us
2013-04-29 09:38:27 <fabled> ncopa, argh
2013-04-29 09:38:29 <fabled> the patch is broke
2013-04-29 09:38:52 <ncopa> how bad is it?
2013-04-29 09:39:01 <ncopa> i revert it?
2013-04-29 09:39:05 <fabled> just a sec
2013-04-29 09:39:15 <fabled> it's my mistake while doing the diff
2013-04-29 09:40:01 <fabled> ncopa, proper patch http://sprunge.us/gLcW
2013-04-29 09:42:18 <fabled> sorry about that
2013-04-29 09:42:24 <ncopa> np, happens
2013-04-29 09:42:31 <fabled> that's what you get for doing last minute things
2013-04-29 09:42:36 <ncopa> i hope it does not break the buildboxes
2013-04-29 09:42:37 <ncopa> yes
2013-04-29 09:42:45 <ncopa> thats what you get when you are in a too big hurry
2013-04-29 09:43:52 <fabled> ncopa, do we have 2.5.x release planned?
2013-04-29 09:44:07 <ncopa> not really
2013-04-29 09:44:11 <ncopa> do you need a release?
2013-04-29 09:44:30 <fabled> would be nice at some point. and cherry-pick uclibc and other fixes too.
2013-04-29 09:44:39 <ncopa> we should probably do 2.4.11 too then
2013-04-29 09:44:41 <ncopa> new kernel there
2013-04-29 09:44:46 <fabled> yes, that too
2013-04-29 09:44:53 <ncopa> i should update the kernel for v2.5 too
2013-04-29 09:45:08 <ncopa> but it will be at minimum 1 days work :-/
2013-04-29 09:45:19 <fabled> oh
2013-04-29 09:45:36 <fabled> it's the version where you cherry-pick patches from stable tree?
2013-04-29 09:45:41 <ncopa> yes
2013-04-29 09:45:52 <ncopa> we should avoid that in future if possible
2013-04-29 09:45:55 <ncopa> also
2013-04-29 09:45:58 <fabled> that's inconvenient, yes
2013-04-29 09:46:13 <ncopa> as kaniini said, 3.9 kernel is out and 3.8 is EOL
2013-04-29 09:46:17 <fabled> yes
2013-04-29 09:46:24 <fabled> when was 2.6.0 release date?
2013-04-29 09:46:31 <ncopa> 7 May
2013-04-29 09:46:38 <ncopa> so we have this week
2013-04-29 09:46:38 <Barthalion> not yet, 3.9 is still mainline
2013-04-29 09:46:49 <fabled> that's in one week
2013-04-29 09:47:10 <ncopa> and we need wait for grsecurity patch :-/
2013-04-29 09:47:16 <fabled> ok, let's concentrate on 2.6.0, and do 2.[45].x later if there's an immediate need
2013-04-29 09:47:25 <fabled> i'll be fine to wait for 2.5.0
2013-04-29 09:47:27 <fabled> 2.6.0*
2013-04-29 09:47:36 <ncopa> aw
2013-04-29 09:47:38 <ncopa> <prometheanfire> pipacs: 3.9?
2013-04-29 09:47:39 <ncopa> <pipacs> jeez
2013-04-29 09:47:39 <ncopa> <pipacs> maybe next week ;P
2013-04-29 09:48:08 <ncopa> yes i think so too
2013-04-29 09:48:15 <ncopa> lets focus on v2.6 for now
2013-04-29 09:48:20 <fabled> ncopa, I can verify that all my SDL apps started now working with the new uclibc
2013-04-29 09:48:28 <ncopa> nice!
2013-04-29 09:48:31 <ncopa> very nice!
2013-04-29 09:49:17 <fabled> Kabaka, upgrade to uclibc from edge, and qemu should work again
2013-04-29 09:49:33 <Kabaka> fabled: awesome, will try that tomorrow :-)
2013-04-29 09:50:01 <ncopa> fabled: how about apk-tools?
2013-04-29 09:50:09 <fabled> oh yes
2013-04-29 09:50:14 <fabled> should do release
2013-04-29 09:50:22 <fabled> possibly fix few things as well
2013-04-29 09:50:50 <ncopa> ERROR: Preference overflow
2013-04-29 09:51:07 <ncopa> is what happens when i try apk upgrade my workstation
2013-04-29 11:21:48 <fabled> rebooting, brb
2013-04-29 12:30:52 <ncopa> hmpf
2013-04-29 12:31:06 <ncopa> whats wrong with the openjdk on build-2-6?
2013-04-29 12:31:17 <ncopa> it simply segfaults
2013-04-29 12:31:40 <ncopa> but it built on build64-edge?
2013-04-29 12:31:43 <ncopa> weird
2013-04-29 14:49:00 <vkrishn> ncopa, log subscription does disconnects sometimes, is it because the clients` connection is slow and the pub drops/disconnects the client?
2013-04-29 14:49:40 <ncopa> hm
2013-04-29 14:49:42 <ncopa> i dont know
2013-04-29 14:50:08 <vkrishn> I think I read about something about slow clients, now unable to locate the article
2013-04-29 14:50:54 <vkrishn> I think I would give a try at zeromq list/forum
2013-04-29 15:02:27 <vkrishn> hmm... http://zguide.zeromq.org/page:all#Slow-Subscriber-Detection-Suicidal-Snail-Pattern
2013-04-29 15:22:54 <vkrishn> any chance of getting xapian in main ?
2013-04-29 15:39:15 <vkrishn> setting ZMQ::SOCKOPT_HWM seems the solution, 1000 seems default in v3.x
2013-04-29 18:37:43 <freedomrun> grr can`t figure how to package libreoffice properly
2013-04-30 01:09:20 <build64-2-5> files from v2.5.4-56-gf1b84d1 uploaded
2013-04-30 01:09:28 <build-2-5> files from v2.5.4-56-gf1b84d1 uploaded
2013-04-30 02:19:44 <Mp5shooter> hmmm
2013-04-30 02:19:55 <Mp5shooter> openswan package version is 2.6.38
2013-04-30 02:20:06 <Mp5shooter> but if i do ipsec --version it says it's version 2.5.0
2013-04-30 02:20:07 <Mp5shooter> o.O
2013-04-30 02:23:22 <Mp5shooter> back
2013-04-30 10:01:07 <ncopa> algitbot: build master
2013-04-30 11:02:59 <ncopa> it was me who killed them
2013-04-30 11:03:15 <ncopa> cannot build openjdk on 4 vservers at same time
2013-04-30 12:55:13 <ncopa> algitbot: build master
2013-04-30 14:03:55 <clandmeter> ncopa: regrading --disable-sdl-dlopen commit. you asked me to remove it and see if it breaks anything. I guess you find that out now.
2013-04-30 14:05:13 <ncopa> :)
2013-04-30 14:05:16 <ncopa> i think it was good
2013-04-30 14:05:23 <ncopa> now we got the real issue fixed
2013-04-30 14:05:39 <clandmeter> nice
2013-04-30 14:05:42 <ncopa> algitbot: build master
2013-04-30 14:54:40 <ncopa> i managed to do krb5 and nfs with kerberos
2013-04-30 15:05:26 <ncopa> algitbot: build master
2013-04-30 15:39:27 <ncopa> algitbot: build master
2013-04-30 15:42:42 <ncopa> here comes the packages for v2.6 x86