2016-09-01 06:16:19 <ScrumpyJack> morning
2016-09-01 06:44:09 <ncopa> morning
2016-09-01 06:48:11 <scadu> o/
2016-09-01 07:15:34 <clandmeter> he is back
2016-09-01 07:15:49 <clandmeter> ncopa: should we talk americano now?
2016-09-01 07:26:42 <ncopa> hola!
2016-09-01 07:27:11 <ncopa> volto de america
2016-09-01 07:27:23 <ncopa> si deveriamos hablar americano ahora!
2016-09-01 07:34:29 <clandmeter> :)
2016-09-01 08:24:03 <^7heo> wait wat?
2016-09-01 08:24:09 <^7heo> no nintendo.
2016-09-01 10:01:47 <jbguerraz> Hello guys :)
2016-09-01 10:02:17 <^7heo> hey jbguerraz
2016-09-01 10:02:23 <jbguerraz> I've got a small doubt.. probably you can help me out clarifying
2016-09-01 10:02:37 <jbguerraz> I've made a new subpackage (so to an existing package in main repo)
2016-09-01 10:02:42 <^7heo> Another Frenchie. If you have the same level of skills we're good ;)
2016-09-01 10:02:42 <jbguerraz> https://github.com/alpinelinux/aports/pull/254
2016-09-01 10:03:22 <jbguerraz> but... I wonder if instead of moving openldap to testing I shouldn't have updated the one  in main repo directly
2016-09-01 10:03:45 <jbguerraz> it's my first contrib.. so I'm not yet used to the flow and probably I didn't RFTM enough ))) sorry for that!!!
2016-09-01 10:04:49 <jbguerraz> ^7heo: you're a frenchie too ? ;)
2016-09-01 10:05:05 <^7heo> I know Frenchies use to brag about their level of skills, but I was talking about skarnet
2016-09-01 10:05:55 <^7heo> sorry skarnet, gave the game away...
2016-09-01 10:07:42 <skarnet> I am but an unskilled student of the great ways of Unix.
2016-09-01 10:07:59 <jbguerraz> The Skarnet from S6 ?
2016-09-01 10:08:16 <^7heo> skarnet: yeah, like O'Sensei was an unskilled student of the great ways of Martial arts.
2016-09-01 10:08:23 <jbguerraz> if so, gg btw!
2016-09-01 10:08:28 <^7heo> skarnet: but he pretty much invented Aikido, nevertheless.
2016-09-01 10:08:29 <skarnet> jbguerraz: pour vous servir :)
2016-09-01 10:08:41 <^7heo> jbguerraz: the one and only.
2016-09-01 10:09:01 <jbguerraz> skarnet: et vous nous servez! :) merci!
2016-09-01 10:09:23 <^7heo> It must be cool to have written some software and know it's used and useful.
2016-09-01 10:09:41 <^7heo> (depends what kind of software, tho)
2016-09-01 10:09:44 <jbguerraz> about my PR (https://github.com/alpinelinux/aports/pull/254) should I keep it like this or update directly the main repo openldap APKBUILD ? :)
2016-09-01 10:10:09 <^7heo> (I wonder how the Orcale people sleep at night)
2016-09-01 10:11:14 <^7heo> that's a rather big PR
2016-09-01 10:11:20 <^7heo> I'll trust skarnet if he has time ;)
2016-09-01 10:11:35 <jbguerraz> in fact it's a small one :) added maybe... max 10 lines
2016-09-01 10:11:50 <jbguerraz> but as openldap is a released package, it was not in testing... maybe my bad was to move it back to testing...
2016-09-01 10:11:58 <jbguerraz> that is my main wonder actually
2016-09-01 10:12:27 <^7heo> wait
2016-09-01 10:12:33 <skarnet> there has been a lot of noise around testing lately
2016-09-01 10:12:49 <skarnet> you should check whether it's in community/ or unmaintained/, just in case
2016-09-01 10:13:05 <^7heo> skarnet: s/testing/the & repo/
2016-09-01 10:13:26 <^7heo> skarnet: because we also discussed about implementing testing as an automated practice.
2016-09-01 10:13:27 <skarnet> that's what I meant, yes.
2016-09-01 10:13:36 <^7heo> skarnet: yeah, I know, but making it more explicit.
2016-09-01 10:13:40 <jbguerraz> it's not in community, nor in unmaintained
2016-09-01 10:13:49 <^7heo> so it's in main or testing.
2016-09-01 10:14:22 <jbguerraz> for now it's in main
2016-09-01 10:14:34 <^7heo> also jbguerraz, I didn't look at the PR thoroughly, but if my understanding is correct, you should make atomic changes in git
2016-09-01 10:14:38 <^7heo> esp. when it's for a PR
2016-09-01 10:15:14 <jbguerraz> I copied it to testing, as I was thinking I should start from testing and then maybe the community dudes will move it to main if they think it's cool :)
2016-09-01 10:16:13 <^7heo> ah no, we don't do that. Copy-pasting source is a great evil.
2016-09-01 10:16:25 <^7heo> you should reference the existing, or include, or something.
2016-09-01 10:16:38 <jbguerraz> ok, so I gonna update in main directly and close that PR then
2016-09-01 10:16:42 <^7heo> I dunno exactly, 'cause again, I didn't look much.
2016-09-01 10:16:57 <^7heo> well, either you do that, or if possible, you do a separate package.
2016-09-01 10:17:07 <^7heo> like git and perl-git
2016-09-01 10:17:13 <skarnet> jbguerraz: yes. If it's in main, it means it's actively maintained, so PRs you submit will be looked at, tested, and integrated if they're good
2016-09-01 10:17:17 <^7heo> (for --patch support and all)
2016-09-01 10:17:24 <skarnet> you can't move a package across repos like that :)
2016-09-01 10:17:26 <leo-unglaub> hey my friends :)
2016-09-01 10:17:31 <^7heo> hey leo-unglaub
2016-09-01 10:17:35 <skarnet> leo-cutie!
2016-09-01 10:17:43 <jbguerraz> all clear! thx dudes!
2016-09-01 10:17:51 <^7heo> skarnet: is it a sexist comment?
2016-09-01 10:18:02 <^7heo> skarnet: unglaub means "not thinking" afaik
2016-09-01 10:18:25 <^7heo> skarnet: so cutie == mindless is a bit sexist :P
2016-09-01 10:18:38 <skarnet> You are putting gender references where there's none
2016-09-01 10:18:39 <^7heo> unless my translation is too approximative.
2016-09-01 10:18:54 <^7heo> or maybe it's just a last name.
2016-09-01 10:19:08 <skarnet> also, it wasn't a direct translation of "unglaub", which I always understood as being Leo's last name
2016-09-01 10:19:25 <leo-unglaub> it is my last name
2016-09-01 10:19:30 <^7heo> ok
2016-09-01 10:19:35 <leo-unglaub> but it means "incredible" in english
2016-09-01 10:19:40 <^7heo> ah
2016-09-01 10:19:45 <^7heo> so it's not thinking but believing.
2016-09-01 10:19:53 <^7heo> sorry for the incorrect approximation.
2016-09-01 10:20:26 <^7heo> so "Ich glaub nicht" means "I don't beleive so", not "I don't think"
2016-09-01 10:20:48 <skarnet> Ich glaube dass es ganz off-topic ist
2016-09-01 10:20:53 <^7heo> even if "I don't think" is often used to mean "I don't think so" but that's another point. And a discussion for #alpine-offtopic
2016-09-01 10:21:04 <^7heo> skarnet: ja ja Ich weiß
2016-09-01 10:21:18 <^7heo> skarnet: Ich hab so sagen.
2016-09-01 10:21:45 <^7heo> s/hab/hat/
2016-09-01 10:25:23 <jbguerraz> https://github.com/alpinelinux/aports/pull/254 fixed!
2016-09-01 10:25:42 <jbguerraz> 16 lines, small PR ;)
2016-09-01 10:26:06 <^7heo> good.
2016-09-01 10:31:34 <jbguerraz> what's the next steps before having it available in edge ? (so for us to be able to use it easilly)
2016-09-01 10:32:19 <jbguerraz> surely first of all one should review and accept the PR, but then ? :)
2016-09-01 10:33:26 <^7heo> jbguerraz: if you want to be able to use a package easily, just put your signature in the /etc/apk/keys directory in all machines
2016-09-01 10:33:31 <^7heo> jbguerraz: and scp your apk there.
2016-09-01 10:33:51 <^7heo> jbguerraz: then all you have to do is 'sudo apk add yourapk.apk'
2016-09-01 10:35:30 <jbguerraz> yep, I've tested the package that way. but that's not really the easy way  :) need some more ansible magic to do so :) that's why I'm wondering about the steps (and so the delay) needed to get it in edge repo. that info will tell me if we should go the "upload the signature and apk" way or wait for edge to have it.
2016-09-01 10:36:53 <^7heo> jbguerraz: then you could have your own repo
2016-09-01 10:36:59 <^7heo> jbguerraz: and use that on all machines.
2016-09-01 10:37:23 <^7heo> jbguerraz: that would have the advantage of allowing you to resist downtimes from the main repos.
2016-09-01 10:41:48 <jbguerraz> ^7heo: interesting. thx.
2016-09-01 12:11:58 <cgray269> h
2016-09-01 12:12:01 <cgray269> i
2016-09-01 12:27:59 <jirutka> ncopa: could you please look at https://github.com/alpinelinux/aports/pull/245 ?
2016-09-01 12:28:44 <jirutka> jbguerraz: add own repository, you mean be interested in https://github.com/jirutka/aports#how-to-setup-your-own-repository
2016-09-01 12:30:01 <jbguerraz> thx jirutka . I'll surely check it out!
2016-09-01 12:46:03 <jbguerraz> jirutka: PR updated https://github.com/alpinelinux/aports/pull/254 thx for feedback!
2016-09-01 16:44:48 <leo-unglaub> hmm, when i use a gtk application on alpine linux the 4 mouse button always triggers the "back" button ...
2016-09-01 16:44:56 <leo-unglaub> can someone maybe tell me where this is defined?
2016-09-01 18:09:05 <jirutka> clandmeter: ncopa: two PRs waiting for merge into master https://github.com/alpinelinux/aports/pulls?q=is%3Aopen+is%3Apr+label%3Aready
2016-09-01 18:24:13 <jbguerraz> jirutka: you could add mine :p
2016-09-01 18:24:21 <jbguerraz> https://github.com/alpinelinux/aports/pull/254
2016-09-01 18:24:23 <jbguerraz> :)
2016-09-01 18:24:27 <jbguerraz> oh
2016-09-01 18:24:29 <jbguerraz> sorry
2016-09-01 18:24:31 <jbguerraz> you did it already ;)
2016-09-01 18:24:43 <jbguerraz> you're too damn fast :D
2016-09-01 18:26:08 <jbguerraz> thx! and bye bye )
2016-09-01 19:18:18 <^7heo> jbguerraz: he is.
2016-09-02 09:03:31 <jirutka> build-edge-armhf is back! \o/
2016-09-02 09:03:35 <clandmeter> you must be braindamanged to call you primary build script "do"...
2016-09-02 09:04:15 <clandmeter> jirutka: i have a new kernel which should support arm32 :)
2016-09-02 09:04:41 <clandmeter> im cleaning up some aports, then ill try to boot it and see if i can run a armhf container.
2016-09-02 09:04:48 <jirutka> who calls primary build script “do”?
2016-09-02 09:04:57 <clandmeter> cjdns
2016-09-02 09:05:55 <jirutka> `CJDNS_RELEASE_VERSION="${pkgver}" bash 'do'` WAT? XD
2016-09-02 09:06:38 <clandmeter> :D
2016-09-02 09:07:31 <clandmeter> im cleaning it up, but i cant do much better running it without renaming it.
2016-09-02 09:08:22 <clandmeter> I also wonder if we should really include *.md files in doc packages.
2016-09-02 09:08:28 <jirutka> there’s quite a lot packages that weren’t build because of armhf vacation, is it going to automatically build them now?
2016-09-02 09:08:47 <clandmeter> yes
2016-09-02 09:08:56 <jirutka> these are textual files that you can read in terminal, so why not?
2016-09-02 09:08:58 <clandmeter> or on next push
2016-09-02 09:09:36 <clandmeter> like i said, im not sure.
2016-09-02 09:10:21 <clandmeter> thats why ill leave it.
2016-09-02 09:14:54 <clandmeter> jirutka: this is still a list of packages not set to build on aarch64 for whatever reason http://sprunge.us/OTFH
2016-09-02 09:14:55 <jirutka> hm, arm32-jit is not supported, we must disable it :( http://icedtea.classpath.org/bugzilla/show_bug.cgi?id=2942
2016-09-02 09:20:31 <clandmeter> bug in my script, http://sprunge.us/CfBN should be more correct.
2016-09-02 09:23:54 <clandmeter> multiple aports will have set "x86 x86_64" to disable armhf. we should update them to "all !armhf".
2016-09-02 09:27:10 <jirutka> I’m trying to find if there’s somewhere patch for jit on arm32… if not, then I’ll add --disable-arm32-jit for armhf, then it should work
2016-09-02 09:27:42 <clandmeter> istn that the same on v7?
2016-09-02 09:28:34 <jirutka> yes
2016-09-02 09:57:12 <jirutka> let’s see…
2016-09-02 09:57:41 <jirutka> I still don’t understand what exactly is wrong, because arm32-jit has been introduced one year ago
2016-09-02 09:58:42 <jirutka> the good thing is that this should not disable JIT at all (that would be very bad), arm32-jit is some specifically optimized JIT for arm
2016-09-02 10:01:26 <clandmeter> jirutka: why did you bump pkgrel?
2016-09-02 10:02:00 <jirutka> clandmeter: because I’m not sure if it will affect only armhf
2016-09-02 10:02:18 <clandmeter> then add it conditionally
2016-09-02 10:03:35 <jirutka> clandmeter: hm, that would be maybe better, but too late :/
2016-09-02 10:03:54 <jirutka> afk
2016-09-02 10:06:15 <clandmeter> if it takes too long i may break the build and update the builder.
2016-09-02 10:20:34 <clandmeter> where did ncopa go?
2016-09-02 10:37:10 <jirutka> I dunno
2016-09-02 10:37:40 <jirutka> clandmeter: why is build-edge-armhf still stuck on -r0 and not building -r1?
2016-09-02 10:38:24 <clandmeter> no clue
2016-09-02 10:39:11 <jirutka> we need at least list of abuilds in the queue, this is like a blackbox :(
2016-09-02 10:41:05 <jirutka> I told algitbot to retry master and now armhf is building libteam instead of openjdk7
2016-09-02 10:42:22 <jirutka> ha, finally, now it shows libteam AND openjdk7… but no build logs yet
2016-09-02 11:29:07 <jirutka> clandmeter: ncopa: I have patch for http://git.alpinelinux.org/cgit/user/sl/build-server-status – it adds link to buildlog http://tpaste.us/AagB
2016-09-02 12:21:46 <clandmeter> im going to reboot the aarch64 box.
2016-09-02 12:21:52 <clandmeter> jirutka: ^
2016-09-02 13:04:29 <clandmeter> jirutka: \o/
2016-09-02 13:47:09 <jbguerraz> \quit
2016-09-02 13:47:19 <jbguerraz> opps wrong window )))
2016-09-02 14:39:56 <jirutka> fabled: are you here?
2016-09-02 18:09:06 <BitL0G1c> kernel deprecated warnings http://hastebin.com/biloyulani.vhdl
2016-09-02 19:19:03 <BitL0G1c> perl-mime-construct is now arm only ? it is needed on 64bit too
2016-09-02 19:20:57 <clandmeter> BitL0G1c: its moved to unmaintained
2016-09-02 19:21:07 <clandmeter> but arm builder didnt catch up yet
2016-09-02 19:25:29 <BitL0G1c> clandmeter - ok no problem - I will take it over - it is needed for email from logcheck
2016-09-02 19:25:54 <BitL0G1c> I've got logcheck patched to work with busybox builtin tools
2016-09-02 19:26:00 <clandmeter> BitL0G1c: i think you moved a lot of those pacakges to commnity right?
2016-09-02 19:26:14 <clandmeter> those move all in one time commits?
2016-09-02 19:26:34 <BitL0G1c> I fixed the problems with the big move - but they've not been applied yet
2016-09-02 19:28:19 <BitL0G1c> https://patchwork.alpinelinux.org/patch/2348/ -- 2354
2016-09-02 19:34:00 <jirutka> maybe because the last time you’ve sent patch with move of dozens broken packages to community…
2016-09-02 19:34:51 <clandmeter> well that has been fixed and we moved on :)
2016-09-02 19:35:03 <clandmeter> BitL0G1c: would you mind submitting them to github?
2016-09-02 19:35:32 <clandmeter> we have travis there to check them before we push them.
2016-09-02 19:36:21 <BitL0G1c> ok
2016-09-02 19:37:30 <clandmeter> thanks!
2016-09-02 20:09:21 <coredumb> am I the only finding chromium barely usable ? crashing all the time except for a bunch of sites ?
2016-09-02 20:09:45 <clandmeter> coredumb: it doesnt supprise me.
2016-09-02 20:10:01 <coredumb> what's the issue ?
2016-09-02 20:10:01 <clandmeter> but i dont use it.
2016-09-02 20:10:20 <clandmeter> cause its not made to run properly on musl
2016-09-02 20:10:20 <coredumb> ffx is such a crap I need something that doesn't put my CPU to roof
2016-09-02 20:10:31 <clandmeter> so we hack it
2016-09-02 20:10:40 <coredumb> gotcha
2016-09-02 20:11:04 <clandmeter> from the time i started hacking on it, it seems that the security implementation has issues on musl.
2016-09-02 20:11:27 <coredumb> it's way better than previous version
2016-09-02 20:11:38 <coredumb> but it can't even load my static website :D
2016-09-02 20:11:46 <clandmeter> ah ok
2016-09-02 20:11:54 <clandmeter> improvment is always nice
2016-09-02 20:12:04 <clandmeter> ncopa took over, so maybe he knows more about it.
2016-09-02 20:12:09 <clandmeter> i dont use alpine on desktop
2016-09-02 20:12:14 <coredumb> oki doki
2016-09-02 20:12:32 <clandmeter> im going to bed.
2016-09-02 20:12:35 <clandmeter> gnite!
2016-09-02 20:59:21 <ncopa> coredumb: i suspect its stack size
2016-09-02 20:59:36 <ncopa> i can try increase it
2016-09-02 22:03:32 <coredumb> ncopa: I'd be happy to test :)
2016-09-03 02:35:21 <BitL0G1c> clandmeter / jirutka - if you can push https://github.com/alpinelinux/aports/pull/259 please I can start sending the other patches to github
2016-09-03 10:49:32 <nmeum> argh, so many unreviewed patches on patchwork :S
2016-09-03 19:22:51 <BitL0G1c> all my patches in patchwork can be ignored - they are all successfully built on github now. The CI is great - thank you to whoever set it up
2016-09-03 19:24:47 <Klowner> yea, the travis integration is slick
2016-09-03 19:26:55 <BitL0G1c> wish I'd started using is sooner
2016-09-03 19:28:31 <BitL0G1c> nmeum - some notes for moving patchwork patches to github http://hastebin.com/arodepocaz.avrasm
2016-09-03 22:37:00 <ezintz> hello? :)
2016-09-03 22:49:43 <^7heo> yo?
2016-09-03 23:42:59 <ezintz> hey :) I was asking already in the other channel, I experienced problems with the nodejs package. I am using Alpine in all my containers and I really like it, but somehow NPM hangs - any ideas what it could be?
2016-09-03 23:43:39 <ezintz> Same thing on a Debian container works fine.... and before starting to investigate whats going on I thought I am going to ask if someone else has same issues... :(
2016-09-05 06:13:04 <ncopa> ez
2016-09-05 06:13:17 <ncopa> ezintz: how can i reproduce it?
2016-09-05 06:21:50 <ncopa> coredumb: do you have a webpage that make chromium crash?
2016-09-05 06:39:31 <coredumb> ncopa: yep several
2016-09-05 06:39:45 <coredumb> my blod home page crashes 100% of time
2016-09-05 06:39:49 <coredumb> www.mauras.ch
2016-09-05 06:49:31 <coredumb> ncopa: other sites like medium.com blogs are also crashing all the time
2016-09-05 06:50:29 <ncopa> hm
2016-09-05 06:50:36 <ncopa> www.mauras.ch opens for me
2016-09-05 06:50:49 <coredumb> really ?
2016-09-05 06:50:57 <coredumb> my laptop on edge can't :(
2016-09-05 06:51:22 <coredumb> I note that most of the time the crash happens on https website btw
2016-09-05 06:53:20 <ncopa> http://i.imgur.com/WEnJwqn.png
2016-09-05 06:53:32 <scadu> coredumb: --cipher-suite-blacklist ?
2016-09-05 06:53:58 <scadu> just guessing :f
2016-09-05 06:54:43 <coredumb> scadu: nah it ggets killed with a segfault
2016-09-05 07:01:59 <coredumb> mmmmh wait
2016-09-05 07:03:02 <coredumb> http://pastebin.com/XeFvjz8m < isn't that grsec killing it ?
2016-09-05 07:08:15 <ncopa> not necessarily
2016-09-05 07:24:50 <^7heo> wow such activity
2016-09-05 07:24:58 <^7heo> very work
2016-09-05 07:25:43 <^7heo> (might be usual for you but for me this is the middle of my night)
2016-09-05 07:50:49 <^7heo> ncopa: since you're here
2016-09-05 07:50:59 <^7heo> ncopa: I've been struggling to get a backtrace with nlplug-findfs
2016-09-05 07:51:34 <^7heo> ncopa: I know how to use backtrace() from the glibc in conjunction with a SIGSEGV handler, I know how to do it with gdb without PaX
2016-09-05 07:51:48 <^7heo> ncopa: but I don't know how to actually do it with musl + PaX.
2016-09-05 07:51:53 <^7heo> ncopa: how do *you* do it?
2016-09-05 07:54:53 <fabled> ^7heo, use gdb
2016-09-05 07:55:04 <fabled> gdb work for me just fone
2016-09-05 07:55:05 <fabled> fine*
2016-09-05 07:55:18 <fabled> you need debug info
2016-09-05 07:55:28 <^7heo> ahhh yeah
2016-09-05 07:55:30 <^7heo> now I recall
2016-09-05 07:55:32 <fabled> also, it might be sometimes better to create core dump
2016-09-05 07:55:35 <^7heo> generating a coredump
2016-09-05 07:55:40 <fabled> gdb works reliably on core dump
2016-09-05 07:55:42 <^7heo> loading it in gdb
2016-09-05 07:55:43 <^7heo> sorry
2016-09-05 07:55:48 <^7heo> I know I've been asking
2016-09-05 07:55:54 <^7heo> and you gave me a complete answer
2016-09-05 07:55:57 <^7heo> I omitted to note it
2016-09-05 07:56:12 <^7heo> and now I felt stupid again when I took a few minutes to do it; without remembering how.
2016-09-05 07:56:34 <^7heo> Sorry I omitted to note, and asked again.
2016-09-05 07:56:43 <^7heo> I will note this time; hopefully I will remember where to look ;)
2016-09-05 07:56:50 <ncopa> ulimit -c unlimited
2016-09-05 07:57:00 <ncopa> build with -g
2016-09-05 07:57:04 <^7heo> ok
2016-09-05 07:57:07 <ncopa> generate core dump
2016-09-05 07:57:17 <ncopa> gdb --core ./core /path/to/binary
2016-09-05 07:57:20 <^7heo> I'm gonna cat that to a coredump.readme file in the repo
2016-09-05 07:57:35 <^7heo> so next time I don't have to bother you any more.
2016-09-05 07:58:46 <^7heo> done. Thanks a lot.
2016-09-05 07:58:49 <^7heo> Good morning, also.
2016-09-05 08:03:00 <fabled> for some daemons you might need also sysctl:
2016-09-05 08:03:01 <fabled> fs.suid_dumpable = 1
2016-09-05 08:03:01 <fabled> kernel.core_pattern = /tmp/core.%e.%p
2016-09-05 08:03:24 <^7heo> thanks a lot fabled
2016-09-05 08:03:46 <fabled> to enable core dumps and write them to /tmp
2016-09-05 08:04:19 <^7heo> yeah I got that part ;)
2016-09-05 08:04:29 <^7heo> appended to my file.
2016-09-05 08:04:31 <^7heo> thanks again :)
2016-09-05 08:44:30 <coredumb> ncopa: let me know which kind of tests I can do
2016-09-05 08:46:49 <ncopa> same
2016-09-05 08:46:51 <ncopa> i mean
2016-09-05 08:46:56 <ncopa> i had it crashing here too
2016-09-05 08:47:04 <ncopa> i have built it locally with 8MB stacksize
2016-09-05 08:47:15 <ncopa> and are looking for something that trigger the problem
2016-09-05 08:47:34 <ncopa> in theory, it would probably be better to find the code in chromium that uses too much stack and refactor it
2016-09-05 08:48:03 <ncopa> and i'm upgrading the version too
2016-09-05 09:01:57 <^7heo> hmm
2016-09-05 09:05:22 <coredumb> ncopa: oh you mean it should be fixed/better now with the new version right ?
2016-09-05 09:05:30 <coredumb> I'll test once in repos
2016-09-05 09:18:28 <^7heo> ok, gotta ignore join/part from wugy
2016-09-13 15:22:21 <skarnet> those people who complain about the heat >.>
2016-09-13 15:47:16 <jbguerraz> clandmeter: https://giphy.com/gifs/3oz8xRSzM0An2K3upO
2016-09-13 15:48:30 <clandmeter> im going to experiance it right now
2016-09-13 15:48:41 <clandmeter> gonna turn of the airco
2016-09-13 15:48:48 <clandmeter> and jump on my bike
2016-09-13 15:48:50 <jbguerraz> btw I've updated https://github.com/alpinelinux/aports/pull/290
2016-09-13 15:48:55 <jbguerraz> enjoy the ride ;)
2016-09-13 16:36:49 <ncopa> nidan_: seems like there is a patch to fix the xen issue
2016-09-13 17:44:52 <ncopa> nidan_: can you please test: http://dev.alpinelinux.org/~ncopa/alpine/alpine-xen-160913-x86_64.iso
2016-09-13 17:57:25 <leo-unglaub> https://www.andreas-jung.com/contents/dont-use-docker-in-github-repo-names-or-as-twitter-handle
2016-09-13 17:59:32 <skarnet> we should create a company named 'internet'
2016-09-13 18:00:29 <leo-unglaub> *g*
2016-09-13 18:01:54 <leo-unglaub> i am going to register the twitter acocunt DockerSucksHardcore
2016-09-13 18:02:09 <jirutka> leo-unglaub: I’m with you
2016-09-13 18:08:31 <nidan_> ncopa: Will test tonight.
2016-09-13 18:31:41 <coredumb> o/
2016-09-13 18:32:00 <coredumb> jirutka: Other comments on my saltstack PR ?
2016-09-13 18:32:26 <jirutka> coredumb: sorry, I didn’t have time for it yet
2016-09-13 18:32:38 <coredumb> it's ok
2016-09-13 18:32:39 <jirutka> coredumb: I’ll look at it later today
2016-09-13 18:33:39 <coredumb> I've been a good boy and did all you asked for :)
2016-09-13 18:45:52 <jirutka> coredumb: hey, I’m not a dominatrix XD
2016-09-13 18:46:15 <coredumb> :D
2016-09-13 18:47:41 <boingolov> get a dungeon you two
2016-09-13 18:53:07 <coredumb> fcolista: do you mind if I take maintainership of the salt package?
2016-09-13 19:26:24 <coredumb> ooops :D
2016-09-13 20:44:01 <gena01> hi
2016-09-13 20:51:54 <gena01> had a question about aports/patchwork
2016-09-13 21:17:03 <coredumb> jirutka: Ok fixed the syndic thing
2016-09-13 21:17:10 <coredumb> shame on me not noticing this
2016-09-13 21:33:40 <nidan_> ncopa: Initial testing of the alpine-xen-160913-x86_64.iso indicates that it works. I've tested it on Asus/Intel i5 and Gigabyte/Intel i7. They boot and let me log in on the console.
2016-09-13 21:34:18 <jirutka> coredumb: are you sure that py2-futures is NOT a runtime dependency, i.e. it’s needed only for build?
2016-09-13 21:34:55 <nidan_> ncopa: I'll test the Gigabyte/AMD FX, the other Asus/i7 and also a 48-core AMD machine later tonight/tomorrow, I kind of expect them to work but I'll test anyway. =)
2016-09-13 21:35:08 <nidan_> ncopa: What was the fix? Updated grsec patch?
2016-09-13 21:36:28 <^7heo> jirutka: alex is a pain to update
2016-09-13 21:36:38 <jirutka> ^7heo: ?
2016-09-13 21:36:41 <^7heo> jirutka: I'll continue tomorrow
2016-09-13 21:36:43 <nidan_> Should I add this info to http://bugs.alpinelinux.org/issues/6071 ?
2016-09-13 21:36:49 <^7heo> jirutka: main/alex
2016-09-13 21:36:55 <jirutka> what’s that?
2016-09-13 21:37:10 <^7heo> a web accelerator someting
2016-09-13 21:37:15 <^7heo> +h
2016-09-13 21:37:39 <^7heo> and it uses autohell now
2016-09-13 21:38:18 <^7heo> anyway, nacht
2016-09-13 21:38:20 <^7heo> o/
2016-09-13 21:38:34 <gena01> so my question was i sent in a bunch of package updates and such to the aport/patchwork.. some of which over a month ago, wanted to see if there's anything I need to update/fix so this could potentially get reviewd ( I also have a number of packages I haven't e-mailed patched for yet)
2016-09-13 21:51:03 <nidan_> nm, I hadn't reloaded the 6071-ticket...
2016-09-13 23:29:28 <skarnet> clandmeter: PEACH time
2016-09-13 23:29:50 <clandmeter> bed time
2016-09-13 23:30:21 <clandmeter> gnite my dear
2016-09-13 23:34:04 <skarnet> gn!
2016-09-14 01:04:10 <boingolov> gena01: it seems like things tend to move a little more quickly on the github mirror these days
2016-09-14 02:18:44 <gena01> boingolov: i see packages being updated, but patchwork list getting longer, was a bit concerned.. also not sure if I should stop sending stuff in for now and get what's already there reviewed and possibly merged first
2016-09-14 02:20:14 <gena01> some stuff as simple as updating version and checksum for a package
2016-09-14 02:43:46 <boingolov> yeah, I'm not sure what the right answer is there
2016-09-14 02:44:12 <boingolov> but lately I've been doing PRs and those seem to be getting more attention, and I also notice that the traffic on the mailing list seems to be getting more sparse
2016-09-14 02:45:28 <boingolov> the last time I submitted patches through the mailing list it took over a month iirc
2016-09-14 02:45:36 <boingolov> I think it's just slowly falling out of favor
2016-09-14 03:01:41 <gena01> boingolov: thanks
2016-09-14 03:48:57 <coredumb> jirutka: good point, thought it was pulled in by another one but apparently not ...
2016-09-14 03:49:15 <coredumb> seems to work without it though...
2016-09-14 04:35:00 <coredumb> jirutka: yeah reviewed the official SRPM of salt and py2-futures is only a BuildRequire
2016-09-14 04:45:17 <coredumb> ah no it's not
2016-09-14 04:45:44 <coredumb> also I think cloud functions won't work without py2-libcloud but it's in testing right now :(
2016-09-14 04:52:38 <coredumb> jirutka: see my comments in the PR... I'm off to work
2016-09-14 06:37:20 <fcolista> coredumb, feel free. No problem at all for me.
2016-09-14 06:37:38 <coredumb> fcolista: cool
2016-09-14 07:13:58 <ScrumpyJack> morning climbers
2016-09-14 07:23:10 <coredumb> morning
2016-09-14 07:23:12 <jbguerraz> Morning ;)
2016-09-14 07:25:10 <^7heo> yo
2016-09-14 07:30:37 <jbguerraz> about https://github.com/alpinelinux/aports/pull/290
2016-09-14 07:31:26 <jbguerraz> maybe I could add export TERM=dumb from the build() method
2016-09-14 07:31:30 <jbguerraz> as a workaround
2016-09-14 07:31:36 <jbguerraz> what do you think about it guys ?
2016-09-14 07:31:56 <jbguerraz> btw I've also found out that "soon" that will not be an issue anymore : https://code-review.gradle.org/changelog/Gradle?cs=9d0f1d56210cbb63a7253ba6fe12bc79c1c92be7
2016-09-14 07:32:20 <jbguerraz> so no more need for https://github.com/travis-ci/travis-build/pull/241
2016-09-14 07:33:59 <ncopa> nidan_: yes, this patch fixed it: http://git.alpinelinux.org/cgit/aports/tree/main/linux-grsec/fix-xen-ssp.patch
2016-09-14 07:41:53 <jbguerraz> I gonna try this export TERM=dumb out now
2016-09-14 07:47:37 <jbguerraz> anyway it makes sens, we don't want this "scrolling trick" in a build process
2016-09-14 08:13:33 <jbguerraz> the export TERM=dumb in build() is actually working well!
2016-09-14 08:24:07 <jbguerraz> yeah! https://github.com/alpinelinux/aports/pull/296 built!
2016-09-14 08:29:43 <ncopa> jbguerraz: i'd like the `cp ...` lines to abort on error
2016-09-14 08:29:58 <ncopa> othwise it will silently fail
2016-09-14 08:30:08 <ncopa> successfully build a broken package
2016-09-14 08:31:41 <jbguerraz> ncopa: roger that :) going to update that
2016-09-14 08:32:21 <clandmeter> and wrap your lines please
2016-09-14 08:32:47 <ncopa> would be nice to have the lines wrapped yes
2016-09-14 08:36:04 <jbguerraz> could you please clarify about "wrap your lines" ?
2016-09-14 08:36:52 <clandmeter> cp file..... \ dest || return 1
2016-09-14 08:37:08 <clandmeter> like  you do with the install part
2016-09-14 08:37:18 <clandmeter> always wrap after 80 chars
2016-09-14 08:37:19 <jbguerraz> clear. thx
2016-09-14 08:37:25 <clandmeter> if possible of course
2016-09-14 08:37:29 <jbguerraz> k
2016-09-14 09:15:51 <jbguerraz> clandmeter / ncopa : https://github.com/alpinelinux/aports/pull/296/files
2016-09-14 09:16:08 <jbguerraz> what do you think about it ? :)
2016-09-14 09:21:17 <ncopa> ! -regex ".*.bat$$"
2016-09-14 09:21:27 <ncopa> are you trying to exclude *.bat?
2016-09-14 09:21:41 <ncopa> how about: ! -name '*.bat'
2016-09-14 09:24:24 <jbguerraz> that's lazy copy / paste from rundeck packages (deb/rpm) make file : https://github.com/rundeck/rundeck/blob/9a6e7dd2c5e5acfcad74b833e15f51225b860997/packaging/Makefile#L65
2016-09-14 09:25:23 <clandmeter> why are you not forcing background in initd?
2016-09-14 09:25:45 <clandmeter> seems like gentoo does https://gitweb.gentoo.org/repo/gentoo.git/tree/app-misc/rundeck-bin/files/rundeckd.initd
2016-09-14 09:25:58 <clandmeter> you could also let it log like they do.
2016-09-14 09:26:57 <jbguerraz> clandmeter: If I force background the logging was not working. that's why
2016-09-14 09:27:02 <jbguerraz> will check gento init script now :)
2016-09-14 09:27:25 <jbguerraz> just checked about bat and yes, the only thing it can do is to exclude the .bat
2016-09-14 09:29:57 <jbguerraz> clandmeter: cool. will update it now! indeed it looks much better
2016-09-14 09:30:24 <jbguerraz> also the --make-pid become useless then
2016-09-14 09:30:31 <clandmeter> correct
2016-09-14 09:34:15 <jbguerraz> about "! -regex ".*.bat$$" "
2016-09-14 09:34:21 <jbguerraz> I wonder why they used it then...
2016-09-14 09:34:44 <jbguerraz> will use -name variant
2016-09-14 09:35:09 <clandmeter> what happends when unzip encounters an error?
2016-09-14 09:35:39 <jbguerraz> probably no webapp ;)
2016-09-14 09:35:57 <jbguerraz> || return 1
2016-09-14 09:36:04 <clandmeter> ;)
2016-09-14 09:38:38 <clandmeter> is there a reason you are doing the symlink in 2 steps?
2016-09-14 09:39:10 <jbguerraz> same lazy copy / paste
2016-09-14 09:40:38 <jbguerraz> checking that
2016-09-14 09:41:15 <jbguerraz> hm
2016-09-14 09:41:21 <jbguerraz> really weird as well
2016-09-14 09:41:30 <jbguerraz> I also wonder why they did it..
2016-09-14 09:41:35 <jbguerraz> but yep this sounds weird to me
2016-09-14 09:41:46 <jbguerraz> rewriting it as well
2016-09-14 09:43:26 <skarnet> what symlink operation?
2016-09-14 09:44:02 <skarnet> generally, when you symlink, it's a good think to do it in 2 steps (ln -s then mv) in order for it to be an atomic replacement if the symlink already exists
2016-09-14 09:45:46 <clandmeter> isnt that what -f is for?
2016-09-14 09:45:55 <jbguerraz> https://github.com/alpinelinux/aports/pull/296/files#diff-35db6a066ef47940a3becdb8e1e74656R85
2016-09-14 09:46:19 <jbguerraz> wondering in what case it could already exists at that stage
2016-09-14 09:46:29 <skarnet> -f is for forcing the symlink even when another link or file exists, but doesn't guarantee atomicity
2016-09-14 09:47:09 <skarnet> "ln -s /etc/rundeck/log4j.properties" doesn't make any sense, where's the second argument?
2016-09-14 09:48:32 <clandmeter> it will create a link in the same dir with same name
2016-09-14 09:48:36 <jbguerraz> it creates the symlink in pwd
2016-09-14 09:48:38 <jbguerraz> then it moves it
2016-09-14 09:48:43 <skarnet> clandmeter: every ln implementation (except s6-ln, because I'm cool like that) performs unlink() then symlink() for ln -s. The correct order is symlink() then rename(), and that's what the ln -s... mv sequence achieves.
2016-09-14 09:49:40 <skarnet> jbguerraz: it's not posix. If you want to create a symlink in the cwd, you need "." as a second argument. Else it's a GNU extension.
2016-09-14 09:50:16 <skarnet> And yes, that is a correct sequence of operation, that guarantees that the symlink in the target directory is atomically created.
2016-09-14 09:50:38 <skarnet> Maybe it's not needed, but it's a good habit to have nonetheless. Don't take that out.
2016-09-14 09:52:53 <skarnet> If I were the author of that script, and I saw that some distribution removed my two-step symlink in their packaging, what I would think would be more or less "OMG WTF I HATE NOOB SH*TTY PACKAGERS WHO DON'T KNOW HOW UNIX WORKS, UNINSTALL AND GO PLAY TETRIS F*CKING IDIOT"
2016-09-14 09:53:09 <skarnet> and that would be a reasonable, moderate reaction
2016-09-14 09:54:26 <jbguerraz> ok :)
2016-09-14 09:54:31 <jbguerraz> btw
2016-09-14 09:55:25 <jbguerraz> maybe you have an idea about https://github.com/alpinelinux/aports/pull/296/files#diff-35db6a066ef47940a3becdb8e1e74656R79
2016-09-14 09:55:30 <clandmeter> skarnet can be very persuasive
2016-09-14 09:55:40 <jbguerraz> why they used -regex instead of -name "*.bat" ?
2016-09-14 09:56:45 <jbguerraz> yep, persuasive.. and I really don't feel playing tetris now...
2016-09-14 09:57:09 <jbguerraz> I'd rather prefer to find time to learn deep unix instead :)
2016-09-14 09:57:48 <skarnet> I can't figure out off the top of my head why they're not using -name, but I'm not a find expert
2016-09-14 09:58:09 <skarnet> however, "-and" is not posix, the correct syntax is "-a"
2016-09-14 09:59:04 <jbguerraz> wondering if I risk a reasonable, moderate reaction by rewriting it ;)
2016-09-14 09:59:12 <jbguerraz> maybe if it's for the posix sake..
2016-09-14 09:59:37 <skarnet> well you're not changing the semantics of the operation, and making things more portable
2016-09-14 09:59:57 <skarnet> so unless findutils find, or whatever implementation they use, has specific magic in "-and"
2016-09-14 10:00:09 <skarnet> you probably aren't taking a big risk :)
2016-09-14 10:00:17 <jbguerraz> I'll summon you if at some point the rundeck dude ask me nicely to go and play tetris :p may I ? :)
2016-09-14 10:00:47 <skarnet> nope, you take responsibility. If you don't want to make the change, don't make it. :P
2016-09-14 10:00:49 <jbguerraz> hm I had to add findutils dep to get it working
2016-09-14 10:01:11 <jbguerraz> maybe it's the time to work on that line and remove the findutils dep then
2016-09-14 10:01:31 <jbguerraz> my first approach was "copy / paste from rundeck official packaging makefile"
2016-09-14 10:01:46 <jbguerraz> but since I'm altering it.. I can probably work on that "find" to avoid the findutils dep
2016-09-14 10:02:00 <skarnet> if it works with bb find with -a, then yes, you extracted a dep
2016-09-14 10:02:42 <jbguerraz> thx again dudes! going to work over that find now! other recommended changes are done already (and ln -s stay atomic and I learned a thing today! thx!!!)
2016-09-14 10:04:21 <skarnet> happy to help! (and happy to make a programmer not choke and want to kill you with fire ;))
2016-09-14 10:15:09 <jbguerraz> btw
2016-09-14 10:15:12 <jbguerraz> for the sake of info
2016-09-14 10:15:14 <jbguerraz> http://blog.moertel.com/posts/2005-08-22-how-to-change-symlinks-atomically.html
2016-09-14 10:15:24 <jbguerraz> hm they also miss the "." in second ln arg
2016-09-14 10:15:47 <clandmeter> did they?
2016-09-14 10:16:07 <clandmeter> are you sure they dont use a gnu based system?
2016-09-14 10:16:36 <jbguerraz> I'm rarely sure ;)
2016-09-14 10:16:43 <jbguerraz> in that article case, I just dunon
2016-09-14 10:16:58 <jbguerraz> s/dunon/dunno
2016-09-14 10:17:34 <jbguerraz> same for the rundeck package makefile.. dunno on what system it's intended to be built
2016-09-14 10:17:57 <skarnet> ln -snf new current
2016-09-14 10:18:01 <skarnet> that's 2 arguments :)
2016-09-14 10:18:19 <jbguerraz> skarnet: few lines later
2016-09-14 10:18:32 <jbguerraz> when the mv comes in
2016-09-14 10:18:36 <skarnet> ln -s new current_tmp && mv -Tf current_tmp current
2016-09-14 10:18:47 <skarnet> that's also 2 arguments, in the ln and the mv
2016-09-14 10:18:50 <jbguerraz> damn me
2016-09-14 10:19:13 <jbguerraz> I mis-readed :s
2016-09-14 10:19:29 <jbguerraz> hehe, no more need for the findutils dep! just compiled
2016-09-14 10:19:30 <skarnet> and yes, the -T option isn't posix ^^
2016-09-14 10:19:37 <skarnet> well congrats!
2016-09-14 10:20:34 <clandmeter> so now we decide the keep the atomic safe symlink operation where its actually not needed and creates confusion because nobody (not that i know of) has done it like this before in aports?
2016-09-14 10:20:54 <skarnet> nobody has done it like this before in aports?
2016-09-14 10:21:03 <skarnet> 
ACTION uninstalls Alpine and goes to Void

2016-09-14 10:21:07 <clandmeter> :)
2016-09-14 10:22:11 <skarnet> if it creates confusion, it's the perfect opportunity to educate confused people. :P
2016-09-14 10:22:21 <clandmeter> right
2016-09-14 10:22:33 <clandmeter> we are waiting for your contribution on wiki.a.o :p
2016-09-14 10:22:46 <clandmeter> [[ Atomic safe symlinks ]]
2016-09-14 10:24:08 <skarnet> Who was it yesterday who said Alpine was for power users and wiki.a.o wasn't the place for (shouldbe) common Unix knowledge?
2016-09-14 10:25:09 <skarnet> but yes, if you insist, a mention of the pattern and a link to that good doc from the Ruby guy (who would have thunk?) can't hurt.
2016-09-14 10:25:26 <clandmeter> its good info
2016-09-14 10:27:00 <skarnet> Is there a "general tips" page?
2016-09-14 10:28:10 <clandmeter> maybe create Unix Best Practices
2016-09-14 10:29:39 <clandmeter> would be nice if we could add some posix practices too
2016-09-14 10:29:55 <skarnet> that needs to be a site of its own
2016-09-14 10:30:09 <skarnet> and I'd spend all my free time contributing to it -_-
2016-09-14 10:30:14 <clandmeter> :)
2016-09-14 10:30:35 <jbguerraz> package tested. all ok. updating the PR
2016-09-14 10:32:40 <jbguerraz> https://github.com/alpinelinux/aports/pull/296/files
2016-09-14 10:46:16 <jbguerraz> 
ACTION has gone for a lunch

2016-09-14 11:25:22 <jirutka> skarnet: could you please explain my why exactly it’s a problem when symlink rename is not atomic (i.e. it may not be here for few CPU cycles) *in build script*?
2016-09-14 11:26:18 <jirutka> skarnet: I understand why it may be a problem in the scenario mentioned on http://blog.moertel.com/posts/2005-08-22-how-to-change-symlinks-atomically.html, but I really can’t imagine how it can be problem in APKBUILD
2016-09-14 11:27:19 <skarnet> Nuclear power plants meltdown because people, like you, lack imagination.
2016-09-14 11:27:34 <skarnet> It may or may not be a problem here. I can't be sure.
2016-09-14 11:28:16 <skarnet> It probably isn't a problem. Like you, I can't imagine a situation where it would be a problem - well I can imagine it, but it's relatively far-fetched.
2016-09-14 11:28:26 <skarnet> But I do not want to have to wonder.
2016-09-14 11:28:41 <skarnet> computers ARE wonderfully creative when it's about making far-fetched situations a reality.
2016-09-14 11:29:37 <skarnet> So, if there's a mechanism that will prevent *any* trouble, no matter how far-fetched, from arising, then I don't ask any questions and I use it.
2016-09-14 11:29:43 <jirutka> skarnet: if so, then it’s a bug in ln utility and you’re two step is only mere workaround
2016-09-14 11:29:46 <skarnet> Why take the risk?
2016-09-14 11:30:09 <jirutka> s/you’re/your/
2016-09-14 11:30:13 <skarnet> No, it's not. The ln utility doesn't guarantee atomic link replacement. Everyone knows that.
2016-09-14 11:30:28 <jirutka> if it’s so big problem as you’re suggesting, then it’s a bug
2016-09-14 11:30:30 <jirutka> period
2016-09-14 11:30:46 <jirutka> cargo cults are worst
2016-09-14 11:30:57 <^7heo> < skarnet> computers ARE wonderfully creative when it's about making far-fetched situations a reality.
2016-09-14 11:31:01 <^7heo> This is even named.
2016-09-14 11:31:09 <^7heo> It is called the "Murphy's law"
2016-09-14 11:31:19 <skarnet> jirutka: you can write to The Open Group so they add a requirement for ln -sf to be atomic. I won't stop you.
2016-09-14 11:31:19 <^7heo> it is closely related to computer engineering.
2016-09-14 11:31:35 <skarnet> I doubt you'll have success though.
2016-09-14 11:31:44 <^7heo> the law states that "if there is any chance, no matter how slight, that something can go wrong, it will."
2016-09-14 11:32:31 <^7heo> and the issue lies down in humans not being able to understand (as opposed to "know") how fast exponential progression is.
2016-09-14 11:32:34 <^7heo> and I mean *fast*
2016-09-14 11:34:26 <skarnet> it's not about exponential progression, it's about lack of correct assessments of orders of magnitude. "yeah, the problem may happen once in a billion times, so who cares" -> sure, if you're only performing the operation once, but if that stuff happens a few million times, the problem is much more likely to actually occur.
2016-09-14 11:34:39 <^7heo> (this has many applications, including, but not limited to, the understanding of integration and derivation, etc)
2016-09-14 11:35:28 <^7heo> skarnet: sure, it's about the lack of correct assessments of the order of magnitude. But how do you make correct assessments about the order of magnitude? By *understanding* the progression of something.
2016-09-14 11:35:51 <^7heo> skarnet: now, linear progression is natural, polynomial progression is rather easy to imagine, but exponential is really hard.
2016-09-14 11:36:03 <^7heo> skarnet: especially when people confuse polynomial and exponential.
2016-09-14 11:36:44 <^7heo> people don't even understand the amount of abstraction around.
2016-09-14 11:36:49 <^7heo> it's just "magic"
2016-09-14 11:37:04 <^7heo> call it "mathematics" or "magic" as much as you want, it's just people who have no clue what a transistor actually does.
2016-09-14 11:37:09 <jirutka> skarnet: I’m not arguing that the problem may happen in one of million cases, I’m arguing that it can’t happen at all; just think about what build scripts do
2016-09-14 11:37:11 <skarnet> https://en.wikipedia.org/wiki/Wheat_and_chessboard_problem#Origin_and_story
2016-09-14 11:37:47 <jirutka> skarnet: however, I’ve probably never need to *rename* symlink in abuild, so I haven’t committed any crime in your eyes :)
2016-09-14 11:38:03 <^7heo> skarnet: exactly my point.
2016-09-14 11:38:04 <skarnet> jirutka: you're _probably_ right today. You'll _definitely_ be wrong tomorrow.
2016-09-14 11:38:34 <jirutka> skarnet: no. but I need to do something useful now
2016-09-14 11:38:53 <skarnet> You complain about shitty software (and rightfully so); be aware that shitty software is shitty precisely *because* people lack foresight.
2016-09-14 11:39:12 <skarnet> so they do something that works today for them, but will break tomorrow for someone else.
2016-09-14 11:39:25 <skarnet> Don't fall into the trap you complain about.
2016-09-14 11:39:29 <jirutka> skarnet: but this is not about foresight, but about paranoid cargo cult
2016-09-14 11:40:24 <skarnet> Cargo cult? When I can show you exactly the race condition that can happen?
2016-09-14 11:40:28 <skarnet> Please.
2016-09-14 11:41:06 <skarnet> Also, sticking to IRC instead of moving to Slack is cargo cult. Why would IRC work better? :P
2016-09-14 11:41:15 <jirutka> skarnet: for the fuck’s sake, we’re talking about build script, it runs itself in builddir, not in root of running system, race conditions are related to parallel running, but there’s nothing parallel
2016-09-14 11:41:25 <skarnet> Today.
2016-09-14 11:42:24 <jirutka> so you’re expecting that someday someone will totally rewrite ash to work totally different and our abuilds will work with it or what?
2016-09-14 11:43:28 <skarnet> abuild may be modified in ways you and I don't think about today. It may build several versions of the stuff in parallel at some point, that wouldn't even be stupid.
2016-09-14 11:43:51 <jirutka> then it would run in different directories
2016-09-14 11:44:10 <jirutka> otherwise broken symlink would be the smallest rpoblem you will have
2016-09-14 11:44:33 <jirutka> stop this, I’m already upset and this discussion makes me even more upset
2016-09-14 11:44:47 <skarnet> Just. Don't. Take. The. Risk. You're protesting because it's 2 commands instead of just one; why the heck do you insist on wasting your time and mine complaining about this when you could do the right thing and write the f*cking 2 commands?
2016-09-14 11:45:50 <jirutka> do you understand the problem of code noise and repeated code?
2016-09-14 11:46:18 <jirutka> whatever, I need to do some work now
2016-09-14 11:46:22 <skarnet> *Or*, if you prefer, you can write one single ln -s command, *with* a comment "no race condition here because the build is isolated and sequential"
2016-09-14 11:46:24 <jirutka> and also write to mailing list about anitya
2016-09-14 11:47:06 <skarnet> at least a comment shows that you're aware of the issue and you assess that it's not a risk.
2016-09-14 11:48:01 <skarnet> (and when it breaks, people will look at this and say "ha! that fool was wrong thinking there was no risk" instead of "omg idiot didn't even think about that". Which is arguably better.
2016-09-14 11:48:09 <skarnet> )
2016-09-14 11:50:26 <ncopa> whatsup guys
2016-09-14 11:50:59 <jirutka> ncopa: nothing, just skarnet is fucking paranoid and points to problems where they are not…
2016-09-14 11:51:38 <skarnet> paranoia is the reason why you like my software more than the stuff you use at work.
2016-09-14 11:51:41 <jirutka> ncopa: but nothing really important
2016-09-14 11:51:52 <jirutka> skarnet: I don’t use your software…
2016-09-14 11:52:02 <skarnet> Your loss. :P
2016-09-14 11:52:17 <jirutka> skarnet: so I can’t say if I like it… but I definitely don’t like that you don’t have any (automated of course) tests…
2016-09-14 11:53:15 <skarnet> I don't like it either, want to write some for me?
2016-09-14 11:53:57 <jirutka> skarnet: no, unfortunately I don’t have time for that :(
2016-09-14 11:54:11 <skarnet> nobody has :(
2016-09-14 11:54:25 <jirutka> skarnet: and it’s your software, so your responsibility to write tests
2016-09-14 11:55:08 <jirutka> skarnet: whatever, I don’t want to fight back, let’s end this pointless discussion
2016-09-14 11:56:10 <clandmeter> can i say its very hot again, or did i use all my credits yesterday?
2016-09-14 11:56:48 <skarnet> clandmeter: you're op, you have infinite credits
2016-09-14 11:56:54 <jirutka> yeah, I don’t even want to know how many degrees is in my office, it’s fucking hot here, hard to think in this weather
2016-09-14 11:59:00 <clandmeter> its hot again!
2016-09-14 11:59:11 <jbguerraz> back here )
2016-09-14 11:59:16 <jbguerraz> it's damn hot!
2016-09-14 11:59:18 <skarnet> time to go get peaches
2016-09-14 11:59:42 <jbguerraz> https://giphy.com/gifs/3oz8xRSzM0An2K3upO
2016-09-14 11:59:55 <jbguerraz> I'm not an op but I take the risk of posting it again :p
2016-09-14 12:01:17 <jbguerraz> with all the attention the safe atomic symlink got, maybe one could review https://github.com/alpinelinux/aports/pull/296 ? :D
2016-09-14 12:01:44 <jbguerraz> it's actually there https://github.com/alpinelinux/aports/pull/296/files#diff-35db6a066ef47940a3becdb8e1e74656R85 :p
2016-09-14 12:01:58 <clandmeter> jbguerraz: im sorry we dont review java apps, they are broken by design.
2016-09-14 12:02:06 <jbguerraz> lol )
2016-09-14 12:02:27 <jbguerraz> can't agree more; sadly I know nothing *better" than rundeck for my current need
2016-09-14 12:02:47 <clandmeter> :)
2016-09-14 12:02:54 <jbguerraz> sos-berlin jobscheduler use java anyway; jenkins as well...
2016-09-14 12:03:02 <jbguerraz> damn java everywhere
2016-09-14 12:04:18 <coredumb> seriously are you complaining of the weather ? how can you ??!!
2016-09-14 12:05:04 <coredumb> 14th of sep and stil 27 degrees
2016-09-14 12:05:09 <coredumb> what's to complain about ?
2016-09-14 12:05:52 <jbguerraz> you say that because you're not a polar bear!
2016-09-14 12:06:05 <jbguerraz> ;)
2016-09-14 12:06:22 <jirutka> jbguerraz: I’ll review it later
2016-09-14 12:06:44 <jbguerraz> thx jirutka !
2016-09-14 12:07:02 <coredumb> jbguerraz: I guess I wouldn't spend time on IRC if I was :P
2016-09-14 12:07:17 <skarnet> 
ACTION growls

2016-09-14 12:07:34 <coredumb> you're not a polar bear skarnet
2016-09-14 12:07:40 <skarnet> how can you be sure
2016-09-14 12:07:41 <coredumb> just a bear from what I can see :P
2016-09-14 12:08:13 <skarnet> my fur isn't all white yet
2016-09-14 12:08:48 <jbguerraz> it's all a matter of time...
2016-09-14 12:09:16 <coredumb> ^^
2016-09-14 12:31:48 <clandmeter> wtf
2016-09-14 12:36:18 <clandmeter> vim 8...
2016-09-14 12:46:23 <zwiPZuf9W2Lt> like every language tries to become lips, every editor tries to become emacs...
2016-09-14 12:46:32 <zwiPZuf9W2Lt> s/lips/lisp
2016-09-14 12:47:35 <leo-unglaub> clandmeter: they changed vim8 to have the same stuff as neovim ... some people like it ... but if you dont need it like me you can simply ignore it
2016-09-14 12:47:57 <leo-unglaub> but about the package ... it might be a good idea to remove some of the doc files from that package
2016-09-14 12:48:25 <clandmeter> i need to use shift now to select text, else it will turn on visual mode.
2016-09-14 12:48:36 <clandmeter> select and copy i mean
2016-09-14 12:50:43 <leo-unglaub> hmmm, i am sure there is a config flag for that *g*
2016-09-14 12:57:23 <ncopa> re #5779
2016-09-14 12:57:33 <ncopa> i have fiexd it in git master
2016-09-14 12:57:43 <ncopa> i wonder if we should backport it to v3.4
2016-09-14 12:58:29 <ncopa> it will replace the heimdal dep with a samba bundled version of heimdal
2016-09-14 14:21:32 <clandmeter> leo-unglaub: how fast does your i7 compile 4.4.20?
2016-09-14 14:23:25 <leo-unglaub> depends on the kernel config
2016-09-14 14:23:50 <clandmeter> our current config
2016-09-14 14:23:52 <leo-unglaub> when i store the source on a ram disc an use all kernels i think 4 minutes
2016-09-14 14:23:58 <leo-unglaub> but i am not sure
2016-09-14 14:24:26 <clandmeter> im building vanilla now
2016-09-14 14:24:44 <clandmeter> with -j48
2016-09-14 14:25:02 <clandmeter> which i7 do you have?
2016-09-14 14:25:45 <leo-unglaub> i have an old one
2016-09-14 14:25:46 <leo-unglaub> Intel(R) Core(TM) i7-2600K CPU @ 3.40GHz, 3503.10 MHz
2016-09-14 14:28:53 <leo-unglaub> clandmeter: but i have not done it in a while ..
2016-09-14 14:29:03 <leo-unglaub> i will get back to you on that with a new time
2016-09-14 14:29:08 <clandmeter> leo-unglaub: try it
2016-09-14 14:29:14 <clandmeter> mine build in 9m 26.75s
2016-09-14 14:29:51 <leo-unglaub> clandmeter: i cannot right now .. i an on OpenBSD on my desktop
2016-09-14 14:29:57 <leo-unglaub> i have to check it on my laptop later on
2016-09-14 14:58:07 <clandmeter> https://twitter.com/clandmeter/status/776068182574596098 :D
2016-09-14 14:58:24 <skarnet> grats!
2016-09-14 14:58:45 <skarnet> I'm going to eat a peach in your honor
2016-09-14 14:59:02 <clandmeter> again?
2016-09-14 14:59:22 <skarnet> at least one a day as long as they're good
2016-09-14 14:59:30 <skarnet> it's the end of the season though, so it might not last much longer
2016-09-14 15:00:43 <skarnet> clandmeter: how much does that baby cost?
2016-09-14 15:01:08 <clandmeter> i didnt buy it
2016-09-14 15:01:11 <clandmeter> so i dont know
2016-09-14 15:02:58 <skarnet> I should pretend doing my own distro in order to get gifted hardware
2016-09-14 15:03:39 <skarnet> "but I reaaaaally need to do full builds in parallel for aarch64!"
2016-09-14 15:04:03 <skarnet> jokes aside, that's awesome
2016-09-14 16:16:50 <ScrumpyJack> skarnet: I have a spare aarch64 thread if you like to get you started ;P
2016-09-14 16:17:30 <skarnet> a spare thread? not even a spare process? that won't get us very far.
2016-09-14 16:28:03 <ScrumpyJack> skarnet: i was being facetious, you seemed desperate ;)
2016-09-14 16:29:18 <skarnet> I'm not desperate - but like most of us, I'm still sensitive to hardware porn.
2016-09-14 17:23:00 <jbguerraz> 
ACTION is leaving

2016-09-14 17:23:33 <jbguerraz> I'd be super happy if one could review https://github.com/alpinelinux/aports/pull/296 so I could start using it withiout uploading my key and the apk ;)
2016-09-14 17:23:37 <jbguerraz> thx a lot ! :)
2016-09-14 17:23:41 <jbguerraz> have fun guys
2016-09-14 18:36:29 <jirutka> could someone answer to http://stackoverflow.com/questions/39484723/alpine-linux-makig-a-custom-iso-image ? /cc ncopa fabled clandmeter
2016-09-14 18:57:53 <coredumb> jirutka: hasn't ncopa already answered on the ML ?
2016-09-14 19:13:37 <jirutka> coredumb: I dunno…? could you pls point me to that mail in archive?
2016-09-14 19:16:12 <coredumb> jirutka: https://lists.alpinelinux.org/alpine-user/201609byindex.html
2016-09-14 19:17:38 <coredumb> so should we merge salt or should I put -syndic back as a subpkg ?
2016-09-14 19:17:42 <coredumb> :)
2016-09-14 19:18:20 <nidan_> ncopa: Yes, I sync:ed the aports tree and rebuilt a kernel, seem to work.
2016-09-14 19:18:29 <jirutka> coredumb: I haven’t had time to look at it today yet
2016-09-14 19:19:10 <nidan_> ncopa: The standard kernel also works, tested that one first.
2016-09-14 19:19:54 <nidan_> ncopa: Haven't actually gotten around to booting any VMs yet though. %-)
2016-09-14 19:20:29 <nidan_> How come the alpine_start parameter was removed from the initrd-init btw?
2016-09-14 19:42:47 <coredumb> jirutka: as I was saying in the comment the cloud part won't work before we put py-libcloud in community from testing. py-pygit2 is also a nice to have dep that's in testing as well
2016-09-14 19:44:59 <coredumb> fcolista: oh I see you maintain py-libcloud \o/ do you mind moving it to community ?
2016-09-14 20:29:39 <kaniini> skarnet: the 48 core costs $2750 usd
2016-09-14 20:29:46 <kaniini> skarnet: 96 core costs $4496 us
2016-09-14 20:29:48 <kaniini> d
2016-09-14 20:29:58 <kaniini> i just got a quote
2016-09-14 20:30:03 <skarnet> wow, there's a 96 core model? o.O
2016-09-14 20:30:45 <jirutka> wow, so many cores!
2016-09-14 20:31:19 <skarnet> That's sick. I want one now...
2016-09-14 20:32:14 <skarnet> make -j 96 yeaaaaaaaah
2016-09-14 20:32:29 <^7heo> yeah but the ram's gonna cost a shitload of money
2016-09-14 20:33:39 <kaniini> ^7heo: about $700 from micron for 256gb
2016-09-14 20:33:54 <skarnet> that's almost reasonable
2016-09-14 20:34:14 <^7heo> wow no it's okay.
2016-09-14 20:34:25 <^7heo> how much does it support?
2016-09-14 20:34:32 <kaniini> upto 2 tb
2016-09-14 20:35:14 <skarnet> so basically $10k for the full packed monster
2016-09-14 20:35:16 <kaniini> i am trying to get the vendor i am working with (iXsystems) to dropship directly from gigabyte to the datacentre though
2016-09-14 20:35:33 <kaniini> right now they want to tack on a bunch of shipping nonsense
2016-09-14 20:35:40 <kaniini> but you can get them from iXsystems
2016-09-14 20:36:17 <^7heo> skarnet: actually, 6K€ + 4Ki€
2016-09-14 20:37:51 <^7heo> not sure if it's on purpose, but it's exactly 6K + 4Ki
2016-09-14 20:38:18 <kaniini> 5595 euro for the config i am doing for my own projects
2016-09-14 20:38:25 <kaniini> plus customs i guess
2016-09-14 20:38:35 <kaniini> i would ask clandmeter where to source one inside europe
2016-09-14 20:38:37 <kaniini> :p
2016-09-14 20:38:38 <^7heo> kaniini: are iXsystems delivering in Eu?
2016-09-14 20:38:43 <kaniini> i dont know
2016-09-14 20:40:41 <clandmeter> skarnet: you can also have 384 cores http://www.cavium.com/newsevents-GIGABYTE-announces-384-Core-2U-server-powered-by-Cavium-ThunderX-ARMv8-processors.html
2016-09-14 20:40:59 <^7heo> dafuq
2016-09-14 20:42:10 <clandmeter> ps, the contact info is not uptodate
2016-09-14 20:42:12 <skarnet> at some point you can't mount any more memory
2016-09-14 20:42:14 <clandmeter> kaniini: nice you found one.
2016-09-14 20:42:27 <skarnet> and you get diminishing returns on the # of cores.
2016-09-14 20:43:48 <clandmeter> anyone know how to format a disk image inside a lxc container?
2016-09-14 20:43:49 <skarnet> 2 TB for 96 cores is cool, it's 20 GB per core. Can you get, say, at least 4 TB on the 384 monster?
2016-09-14 20:44:03 <clandmeter> seems losetup doesnt like to read the partitions.
2016-09-14 20:44:19 <skarnet> is the container privileged?
2016-09-14 20:44:27 <clandmeter> no
2016-09-14 20:44:31 <skarnet> well then you can't
2016-09-14 20:44:39 <skarnet>  /dev is protected
2016-09-14 20:44:40 <clandmeter> it is
2016-09-14 20:44:42 <clandmeter> sorry
2016-09-14 20:44:46 <skarnet> oh
2016-09-14 20:44:46 <clandmeter> its not unpriv..
2016-09-14 20:45:03 <clandmeter> i can access my loop devices
2016-09-14 20:45:03 <skarnet> I'm not sure, check your /dev matches the host?
2016-09-14 20:45:30 <clandmeter> but recent version of losetup has --partscan
2016-09-14 20:45:46 <clandmeter> it works outside of container
2016-09-14 20:46:20 <clandmeter> [549298.549118] loop_reread_partitions: partition scan of loop1 () failed (rc=-13)
2016-09-14 20:46:22 <skarnet>  again, /dev is screwy in containers
2016-09-14 20:46:26 <skarnet> check it
2016-09-14 20:46:50 <clandmeter> dev is limited
2016-09-14 20:46:58 <clandmeter> i had to create the loop devices myself
2016-09-14 20:49:20 <skarnet> did you also create the /dev/loop0p1...loop0p8 nodes?
2016-09-14 20:49:52 <skarnet> that's what partscan does
2016-09-14 20:50:05 <skarnet> (it gives you a node per partition)
2016-09-14 20:50:37 <kaniini> skarnet: the 384 is 4 servers in 1 enclosure
2016-09-14 20:50:51 <kaniini> skarnet: like a 2U version of the supermicro fat-twin
2016-09-14 20:51:11 <skarnet> ah, so the memory goes with it? 8TB here I come
2016-09-14 20:52:46 <clandmeter> skarnet: partscan?
2016-09-14 20:52:48 <jirutka> skarnet: but do you know what this will lead to? even less effective bloats, because memory is cheap, you know… ;)
2016-09-14 20:53:04 <clandmeter> you mean partprobe?
2016-09-14 20:53:15 <skarnet> clandmeter: I mean losetup -P (or --partscan)
2016-09-14 20:53:23 <jirutka> clandmeter: why do you need to do it from the LXC container? what about using just a chroot?
2016-09-14 20:53:25 <clandmeter> oh ok
2016-09-14 20:53:53 <jirutka> clandmeter: containers should NOT allow to do what you want to do ;)
2016-09-14 20:54:10 <clandmeter> jirutka: because i want to script the creation of a bootable image inside my container
2016-09-14 20:54:18 <skarnet> jirutka: not if I own that baby. I'll just run a few Dockers and a few Go programs on it!
2016-09-14 20:54:28 <clandmeter> so i need to format the image file (both partitions)
2016-09-14 20:54:37 <jirutka> skarnet: shut up!
2016-09-14 20:54:47 <jirutka> clandmeter: aha, understand now
2016-09-14 20:54:55 <skarnet> if you had to create /dev/loop0 yourself in the container, you also need to create /dev/loop0p1 and /dev/loop0p2
2016-09-14 20:55:19 <jirutka> clandmeter: well, then you basically needs to disable some lxc protections, start with adding capabilities back
2016-09-14 20:55:51 <clandmeter> i can probalby use an offset
2016-09-14 20:56:04 <jirutka> clandmeter: but I don’t know which exactly are needed for this and which other protections will trying to stop you
2016-09-14 20:56:16 <coredumb> jirutka or clandmeter can you please put back WIP tag on the PR 282 ?
2016-09-14 20:56:20 <clandmeter> jirutka: im not sure its any proteciton
2016-09-14 20:56:25 <jirutka> coredumb: yes
2016-09-14 20:56:33 <clandmeter> could be its proc related.
2016-09-14 20:57:43 <jirutka> yes… linux namespaces works like a protection
2016-09-14 20:58:15 <coredumb> jirutka: thx
2016-09-14 20:58:36 <coredumb> I want to have py-libcloud and py-pygit2 in community first
2016-09-14 20:58:49 <coredumb> https://github.com/alpinelinux/aports/pull/298
2016-09-14 20:59:28 <jirutka> coredumb: could you please setup your editor to not mess tabs with spaces?!
2016-09-14 20:59:45 <jirutka> coredumb: and also follow this https://wiki.alpinelinux.org/wiki/APKBUILD_examples:Python
2016-09-14 20:59:59 <coredumb> oh I thought I did fix them :)
2016-09-14 21:00:17 <jirutka> coredumb: there’s also https://wiki.alpinelinux.org/wiki/TODO:py3_packages
2016-09-14 21:03:03 <coredumb> ok I see missing one or two things
2016-09-14 21:03:22 <jirutka> and doing it unnecessary complicated
2016-09-14 21:04:08 <jirutka> also depends must be global in this particular case
2016-09-14 21:04:22 <jirutka> eeh, I mean, depends must be global
2016-09-14 21:05:15 <jirutka> in this particular case, a variable inside a function must be global, because… well… you want to change script’s depends, not to create another variable
2016-09-14 21:05:45 <coredumb> true
2016-09-14 21:07:28 <coredumb> and yeah I shouldn't copy paste from other packages :D
2016-09-14 21:07:44 <jirutka> coredumb: you can, but from correct ones ;) :P
2016-09-14 21:07:58 <coredumb> ^^
2016-09-14 21:10:05 <jirutka> btw there’s also a good reason why I call `_py` as the last statement of the function, and not `replaces=` ;) /cc clandmeter
2016-09-14 21:11:56 <jirutka> (and that reason is not just esthetic, but errors related)
2016-09-14 21:12:40 <clandmeter> jirutka: did you write to the ml already?
2016-09-14 21:12:49 <jirutka> clandmeter: not yet, no time, I’ll now
2016-09-14 21:14:02 <coredumb> jirutka: better now ?
2016-09-14 21:14:47 <jirutka> I’ve lowered temperature in my apartment from 27 °C just to 25.5 after half an hour with open windows… :(
2016-09-14 21:15:05 <coredumb> it's opened all night long here
2016-09-14 21:16:00 <jirutka> coredumb: just move libgit2 to the top-level depends and it will be perfect :)
2016-09-14 21:16:32 <coredumb> really it's better there than in _py() ?
2016-09-14 21:16:46 <jirutka> uh and also replace python-dev with python2-dev
2016-09-14 21:18:08 <jirutka> coredumb: hmm, actually, you’re right… but then you don’t need `$depends` inside the _py’s depends=
2016-09-14 21:18:35 <coredumb> wwwtrue
2016-09-14 21:18:47 <coredumb> push -f is my thing don't worry
2016-09-14 21:19:03 <jirutka> I’ll update the template on wiki
2016-09-14 21:19:51 <coredumb> pushed
2016-09-14 21:19:54 <coredumb> let me compile it
2016-09-14 21:19:59 <jirutka> it doesn’t make much sense to add dependencies for a “metapackage”, so it’s really better to declare only in py2/py3 subpackages
2016-09-14 21:23:37 <coredumb> ok compiles fine
2016-09-14 21:30:04 <coredumb> jirutka: makedepends shouldn't get the same deps than depends ?
2016-09-14 21:30:26 <jirutka> coredumb: yes, this is implicit
2016-09-14 21:30:56 <coredumb> ok so I can basically remove everything except python2-dev from makedepends right ?
2016-09-14 21:35:41 <jirutka> coredumb: it seems that this doesn’t work for depends inside split functions
2016-09-14 21:36:53 <jirutka> coredumb: I should note this, for future improvements
2016-09-14 21:36:57 <coredumb> ok
2016-09-14 21:37:25 <jirutka> coredumb: so in this case you must keep it as it is
2016-09-14 21:37:43 <coredumb> :O
2016-09-14 21:37:54 <coredumb> or
2016-09-14 21:38:12 <coredumb> makedepends="$depends_dev $depends python2-dev"
2016-09-14 21:38:46 <jirutka> coredumb: this will not help
2016-09-14 21:39:06 <coredumb> so I put back everything in there :D
2016-09-14 21:39:25 <jirutka> coredumb: don’t worry about it, I’ve already cloned it and going to merge it
2016-09-14 21:41:24 <coredumb> I pushed anyway
2016-09-14 21:44:18 <jirutka> coredumb: okay, merged; thanks for pointing me to the question about declaring common dependencies inside _py instead of top-level and also this problem with implicit makedepends that doesn’t work with depends in split functions… it’s actually obvious why it doesn’t work, but I haven’t think about it before
2016-09-14 21:47:26 <jirutka> coredumb: I assume that you’re already using Salt, so we can consider py-pygit2 as working and move it to the community…?
2016-09-14 21:47:56 <jirutka> coredumb: eh, pardon, it’s not a dependency of salt
2016-09-14 21:48:17 <jirutka> that’s py2-libcloud
2016-09-14 21:54:25 <jirutka> skarnet: I’m not sure now, was your mention of Go funny provocation related to your today’s discussion? if you meant it seriously, then sorry for my reaction
2016-09-14 21:54:58 <skarnet> it was a joke!
2016-09-14 21:55:31 <skarnet> again: I wouldn't mind *programming* in Go, but I definitely would mind *running* Go programs with the current runtime
2016-09-14 21:56:15 <jirutka> skarnet: okay, so thought it right :) my reaction may look quite offensive without context
2016-09-14 21:56:36 <skarnet> I understood it, no worries, and it was expected :D
2016-09-14 21:57:01 <clandmeter> gnite ppl
2016-09-14 21:57:12 <clandmeter> and dont remember, we have offtopic ;-)
2016-09-14 21:57:25 <skarnet> gn clandmeter. I'm out of peaches, need to go grocery shopping tomorrow.
2016-09-14 21:57:29 <clandmeter> err forget..
2016-09-14 21:58:06 <jirutka> uh and I should finally write that email to ML now
2016-09-14 22:30:36 <vkris> wow, who is getting 96cores !!
2016-09-14 22:32:03 <skarnet> I am, in my dreams
2016-09-14 22:32:24 <vkris> nice ! here is a free joke I just cooked
2016-09-14 22:32:38 <vkris> Sweeheart, I am just starting to compile kernel, wanna .. !!??
2016-09-14 22:32:38 <vkris> (She reply), comming honey.
2016-09-14 22:32:38 <vkris> As she approaches, (he says), nope.. its done.
2016-09-14 22:32:38 <vkris> (She replies), how many cores you said
2016-09-14 22:32:38 <vkris> (reply), 96 dear!
2016-09-14 22:32:40 <vkris> (She replies), oh! I thought you said Soixante-neuf
2016-09-14 22:33:13 <vkris> (royalty free, enjoy)
2016-09-14 22:36:00 <skarnet> the guy who compiles a kernel on a 96-core and thinks he can get a shot in the meantime is the most severe premature optimizator on Earth, and I pity his wife.
2016-09-14 22:38:02 <nidan_> Hahahaha
2016-09-14 22:38:30 <nidan_> Hilarious. =)
2016-09-14 22:48:45 <vkris> there's more way to interpret it ;), here have some peach
2016-09-14 22:49:25 <skarnet> I bet I can eat a peach before my 96-core finishes compiling a kernel
2016-09-14 22:50:09 <vkris> I said chew-n-eat, not gulp it down
2016-09-14 22:52:05 <vkris> gotta go, just finished watching 'citizenfour' and kinda tensed
2016-09-14 22:52:06 <skarnet> You're right. I like to take my time in matters of love, and peaches are love.
2016-09-14 22:52:18 <vkris> ;)
2016-09-15 00:12:12 <BitL0G1c> jirutka - Anitya worked tonight - it triggered an email for "entr" to be updated
2016-09-15 04:12:05 <coredumb> jirutka: I'm using salt but without either pygit2 nor libcloud
2016-09-15 04:12:25 <coredumb> as both were in testing
2016-09-15 04:12:53 <coredumb> I imported the python module for pygit2
2016-09-15 04:18:05 <coredumb> I'm sure we can move them both to community it should be quite quick to validate
2016-09-15 04:20:21 <coredumb> so jirutka #304
2016-09-15 04:24:24 <coredumb> yeah well meant https://github.com/alpinelinux/aports/pull/304
2016-09-15 06:36:12 <fcolista> coredumb, done
2016-09-15 06:36:19 <fcolista> who's Alba Pompeo here ?
2016-09-15 06:45:49 <coredumb> fcolista: thx !
2016-09-15 07:32:14 <ScrumpyJack> morning all
2016-09-15 07:35:08 <^7heo> moin
2016-09-15 08:03:17 <jbguerraz> Hello guys
2016-09-15 08:03:42 <jbguerraz> I've updated my PR for rundeck because of "spaces vs tab" and... https://travis-ci.org/alpinelinux/aports/builds/160096965
2016-09-15 08:03:59 <jbguerraz> "ERROR: http://dl-cdn.alpinelinux.org/alpine/edge/main: temporary error (try again later)"
2016-09-15 08:04:02 <jbguerraz> FYI
2016-09-15 08:06:39 <jbguerraz> btw, how to re-run the build ? have to push again ? or is there any other way ?
2016-09-15 08:23:40 <jbguerraz> it's now ok.. build in progress
2016-09-15 09:11:58 <^7heo> moin jbguerraz
2016-09-15 09:19:14 <jbguerraz> ^7heo: hm ?
2016-09-15 09:19:16 <jbguerraz> oh
2016-09-15 09:19:24 <jbguerraz> morning ^7heo :)
2016-09-15 10:12:12 <jirutka> fcolista: Alba Pompeo is phantom, he watches IRC, but never writes here, instead he sends my reactions via email :)
2016-09-15 10:12:47 <fcolista> interesting
2016-09-15 10:13:07 <fcolista> Alba Pompeo, or whatever nick you have, can you please contact me here over irc ?
2016-09-15 10:47:05 <coredumb> lol
2016-09-15 10:54:59 <^7heo> fcolista: good luck ;)
2016-09-15 11:29:42 <jirutka> hi
2016-09-15 11:30:50 <jirutka> my co-worker reported me that Go somehow doesn’t work on armhf… does anyone know about some problem?
2016-09-15 11:32:31 <jirutka> I’ve just wanted to rebuild it on builder-edge-armhf and it’s gone again :(
2016-09-15 11:35:36 <coredumb> I don't have arm platform to test
2016-09-15 12:05:50 <nidan_> What do you folk think about a small patch for mkinitfs; I'd like to add support for an /etc/mkinitfs/conf.d/ directory. Makes it easier to drop in configs (when creating packages) instead of having the package's post-install-script do something like echo 'features="\$features a b c"' > /etc/mkinitfs/mkinitfs.conf.
2016-09-15 12:10:13 <coredumb> jirutka: thx I can test the features in salt now
2016-09-15 12:24:02 <fcolista> bah..i've started to receive anitya's email about flagged packages which are already upgraded.
2016-09-15 12:24:07 <fcolista> jirutka, ^
2016-09-15 12:24:26 <nidan_> Another remark on the topic of mkinitfs; should mkinitfs really get it's features from a feature directory prefixed with $basedir?
2016-09-15 12:24:45 <jirutka> fcolista: which ones?
2016-09-15 12:25:00 <fcolista> cifs-utils
2016-09-15 12:25:24 <nidan_> Imho that seems strange, I'd like to be able to configure in /etc/mkinitfs/ and tell it to use something other than / for basedir when making the initramfs, but not read mkinitfs' config from within that basedir.
2016-09-15 12:25:37 <fcolista> perl-text-diff
2016-09-15 12:25:45 <jirutka> fcolista: really? I don’t see cifs-utils as flagged on http://pkgs.alpinelinux.org/packages?name=cifs-utils&branch=&repo=&arch=&maintainer=
2016-09-15 12:25:55 <fcolista> openvas + grenbone*
2016-09-15 12:26:04 <fcolista> but i got the email
2016-09-15 12:26:17 <fcolista> now are more than 30 email i got with flagged packages
2016-09-15 12:26:52 <fcolista> perl-net-appliance-session
2016-09-15 12:27:22 <jirutka> oh shit, I think that I know why…
2016-09-15 12:27:30 <fcolista> perl-xml-twig
2016-09-15 12:27:35 <jirutka> shit shit shit
2016-09-15 12:28:41 <nidan_> mkinitfs basedir vs. rootdir.
2016-09-15 12:29:12 <jirutka> I should have rather write it from scratch than extend aports-turbo :(
2016-09-15 12:29:52 <fcolista> so far, all the packages claiming to be "upgrade", are already updated to the latest version
2016-09-15 12:30:45 <jirutka> hmm… but old packages are cleaned in existing import script, so there should not be old packages
2016-09-15 12:33:42 <jirutka> fcolista: perl-xml-twig is reallly outdated, for armhf https://pkgs.alpinelinux.org/packages?name=perl-xml-twig&branch=&repo=&arch=&maintainer=
2016-09-15 12:34:17 <jirutka> this may be caused by the fact that builder-edge-armhf is broken too often so many packages have not been built :/
2016-09-15 12:34:29 <fcolista> i think so.
2016-09-15 12:34:34 <jirutka> perl-net-appliance-session is the same case
2016-09-15 12:34:41 <fcolista> APKBUILD has the latest version
2016-09-15 12:35:02 <jirutka> yes, but there’s no package in the repository for armhf
2016-09-15 12:35:13 <fcolista> it's "noarch"
2016-09-15 12:35:27 <jirutka> it works with aports-turbo database, no abuilds repository
2016-09-15 12:35:48 <jirutka> yes, but as you can see on pkgs.a.o, the package for armhf is in older version
2016-09-15 12:37:50 <fcolista> how can i trigger the rebuild for armhf?
2016-09-15 12:37:59 <fcolista> for that package i mean
2016-09-15 12:38:06 <fcolista> jsu bump the pkgrel ?
2016-09-15 12:38:10 <fcolista> *just
2016-09-15 12:38:34 <jirutka> I think so
2016-09-15 12:43:00 <jirutka> fcolista: okay, I’ve checked the packages that you’ve reported me and all of them are outdated only on armhf; so it’s an infra problem, not your as a maintainer
2016-09-15 12:43:38 <clandmeter> he could have fixed the arm builder.
2016-09-15 12:43:46 <jirutka> fcolista: sorry for flood of emails
2016-09-15 12:43:51 <clandmeter> so its also a little his issue ;-)
2016-09-15 12:43:56 <jirutka> clandmeter: he’s the one who has access to the builder?
2016-09-15 12:44:05 <clandmeter> no
2016-09-15 12:44:24 <clandmeter> oh i scared him away
2016-09-15 12:44:28 <fcolista> lol
2016-09-15 12:44:34 <fcolista> ctrl+w in the wrong window
2016-09-15 12:44:55 <clandmeter> jirutka: only fabled and ncopa have access.
2016-09-15 12:45:01 <jirutka> I’ll analyze how many packages for armhf are not built, at evening
2016-09-15 12:45:08 <clandmeter> ncopa: can we move the arm builder to new arm hardware?
2016-09-15 13:00:25 <clandmeter> ok time to go home and have some holidays
2016-09-15 13:00:53 <jirutka> clandmeter: have a nice time, see ya!
2016-09-15 13:01:05 <skarnet> clandmeter: happy holidays!
2016-09-15 13:01:10 <clandmeter> its raining in porto
2016-09-15 13:01:18 <clandmeter> and its 32 here
2016-09-15 13:01:19 <ncopa> have a nice vacation!
2016-09-15 13:01:25 <clandmeter> ist upside down
2016-09-15 13:01:29 <clandmeter> :)
2016-09-15 13:01:30 <fcolista> enojoy clandmeter !
2016-09-15 13:01:43 <clandmeter> fcolista: sorry, no italy this time :)
2016-09-15 13:01:52 <fcolista> sorry for you :)
2016-09-15 13:01:52 <skarnet> are you going somewhere?
2016-09-15 13:02:05 <clandmeter> porto
2016-09-15 13:02:06 <clandmeter> area
2016-09-15 13:02:17 <skarnet> well, have some nice storms then :D
2016-09-15 13:02:30 <clandmeter> :)
2016-09-15 13:02:35 <clandmeter> tomorrow is good again
2016-09-15 13:02:39 <clandmeter> i hope
2016-09-15 13:02:41 <skarnet> "I'm leaving Germany, it's too hot, too much sunshine"
2016-09-15 13:11:08 <barthalion> jirutka: hi, your anitya integration decided to inform me about newer menu-cache release despite the fact I'm not the maintainer, does it sound like a bug?
2016-09-15 13:12:18 <jirutka> barthalion: no, the problem is that you’ve disowned the packages, but didn’t triggered rebuild, so on pkgs.a.o is still the old information (https://pkgs.alpinelinux.org/packages?name=menu-cache&repo=all&=&arch=x86_64&=&maintainer=all&=)
2016-09-15 13:13:19 <jirutka> barthalion: I’ll trigger rebuild of your former packages to update the info in pkgs.a.o
2016-09-15 13:13:28 <barthalion> jirutka: bumping package for updating maintainer info doesn't seem very clever to me
2016-09-15 13:13:52 <jirutka> barthalion: it’s not, but that’s how pkgs.a.o currently works :/
2016-09-15 13:14:04 <jirutka> barthalion: or we can update it manually in the db
2016-09-15 13:14:10 <jirutka> barthalion: that would be probably better
2016-09-15 13:14:20 <barthalion> jirutka: and to be honest, I don't see why maintainership info belongs to package at all, but I digress; thanks for taking care of that
2016-09-15 13:14:35 <jirutka> barthalion: I’ll prepare update query for clandmeter, but he’s now on vacation, so he will run it next week
2016-09-15 13:14:57 <barthalion> that's ok, I can live with it
2016-09-15 13:15:15 <barthalion> that definitely means that your integration works and that I forgot to get rid of my al.o alias
2016-09-15 13:15:22 <jirutka> barthalion: okay, btw how many flag emails did you get?
2016-09-15 13:15:31 <barthalion> just this one
2016-09-15 13:15:47 <jirutka> okay
2016-09-15 13:15:57 <barthalion> I didn't rebuild any of my past packages when I disowned them so it might happen again
2016-09-15 13:17:09 <jirutka> yes, you have still 24 packages according to pkgs.a.o
2016-09-15 13:22:08 <barthalion> I didn't expect pkgs.a.o to work this way :)
2016-09-15 13:22:34 <barthalion> going back to work, bb
2016-09-15 13:22:45 <^7heo> tbh I was also surprised
2016-09-15 13:22:51 <^7heo> that p.a.o did that.
2016-09-15 13:29:29 <jirutka> ^7heo: that’s because pkgs.a.o works only with packages repository, not with abuilds repository
2016-09-15 13:30:03 <jirutka> ^7heo: the information about maintainer is in APKINDEX
2016-09-15 13:30:22 <^7heo> ok
2016-09-15 14:02:17 <jirutka> ^7heo: when you have some time, could you please update community/caddy? it’s some wannabe web server written in goshit (unfortunately my co-worker uses it and complained that it’s outdated)
2016-09-15 14:08:18 <^7heo> ok
2016-09-15 14:08:19 <^7heo> I will.
2016-09-15 14:08:26 <^7heo> jirutka: are you fluent with perl?
2016-09-15 14:12:10 <jbguerraz> jirutka: any chance you would review https://github.com/alpinelinux/aports/pull/296 ? :) pleaaaase :)
2016-09-15 14:13:47 <^7heo> jirutka: it's quite big.
2016-09-15 14:14:02 <^7heo> oops, wrong nick, sorry, irssi.
2016-09-15 14:14:06 <^7heo> jbguerraz: it's quite big.
2016-09-15 14:14:19 <^7heo> jbguerraz: would you mind removing the newlines between the statments?
2016-09-15 14:14:32 <^7heo> it's just making it extra long for no reason.
2016-09-15 14:16:52 <jbguerraz> ^7heo: I don't mind . Doing it now!
2016-09-15 14:17:59 <^7heo> jbguerraz: thanks :)
2016-09-15 14:18:12 <^7heo> damn, this go stuff is annoying.
2016-09-15 14:21:00 <jbguerraz> ^7heo: https://github.com/alpinelinux/aports/pull/296/files done!
2016-09-15 14:21:13 <jbguerraz> my pleasure :p
2016-09-15 14:21:19 <^7heo> :)
2016-09-15 14:21:28 <^7heo> I'm sorry, glide_init from clandmeter isn't working
2016-09-15 14:21:33 <^7heo> I have to check what I need to fix. ;)
2016-09-15 14:42:37 <^7heo> jirutka: caddy done.
2016-09-15 14:42:42 <^7heo> jbguerraz: I'll check your PR now.
2016-09-15 14:51:02 <jbguerraz> thx ^7heo :)
2016-09-15 14:52:51 <^7heo> jirutka: review of 300 done.
2016-09-15 14:53:00 <^7heo> jbguerraz: sorry, got sidetracked, doing it now.
2016-09-15 14:53:34 <^7heo> gosh, travis failed on my caddy update
2016-09-15 14:53:37 <^7heo> it works on my host tho.
2016-09-15 14:53:58 <^7heo> jbguerraz: awesome :)
2016-09-15 14:54:02 <^7heo> jbguerraz: it's much more tidy this way.
2016-09-15 14:54:04 <^7heo> jbguerraz: I like it a lot.
2016-09-15 14:54:09 <^7heo> jbguerraz: let me see the content now ;D
2016-09-15 14:54:30 <^7heo> jbguerraz: is bash explicitely needed for that build?
2016-09-15 14:54:54 <jbguerraz> yep
2016-09-15 14:55:01 <jbguerraz> for the gradlew stuff
2016-09-15 14:55:06 <^7heo> ok
2016-09-15 14:55:24 <^7heo> why do you download https://github.com/${pkgname}/${pkgname}/archive/v${pkgver}.zip and then use builddir="$srcdir"/$pkgname-$pkgver?
2016-09-15 14:55:29 <skarnet> I hope the bash dep isn't as silly as the findutils dep was :P
2016-09-15 14:55:36 <^7heo> I mean, it's gonna be extracted fine I guess
2016-09-15 14:55:41 <^7heo> but it's not clean ;)
2016-09-15 14:56:05 <jbguerraz> Hello skarnet  ;)
2016-09-15 14:56:20 <jbguerraz> https://github.com/rundeck/rundeck/blob/master/gradlew
2016-09-15 14:56:29 <jbguerraz> that's the damn bash dep script ;)
2016-09-15 14:56:43 <skarnet> #!/usr/bin/env bash
2016-09-15 14:56:46 <skarnet> stopped reading there
2016-09-15 14:57:07 <skarnet> KILL THE /USR/BIN/ENV PEOPLE NOW
2016-09-15 14:57:11 <^7heo> jbguerraz: I've left comments on your PR
2016-09-15 14:57:26 <^7heo> skarnet: well, it's better than /bin/bash
2016-09-15 14:57:31 <skarnet> no it's not
2016-09-15 14:57:33 <^7heo> skarnet: heck, anything is better than /bin/bash
2016-09-15 14:57:39 <jbguerraz> skarnet: starting line 143 .. it's good looking :p
2016-09-15 14:57:41 <skarnet> #!/bin/sh is better
2016-09-15 14:57:46 <^7heo> the ONLY thing that works anyway is /bin/sh
2016-09-15 14:57:46 <skarnet> but if you have to use bash
2016-09-15 14:57:48 <^7heo> genau
2016-09-15 14:57:54 <skarnet> then #!/bin/bash it is
2016-09-15 14:57:58 <^7heo> meh
2016-09-15 14:57:59 <^7heo> I dunno
2016-09-15 14:58:02 <^7heo> I never use bash
2016-09-15 14:58:15 <^7heo> so the question doesn't bother me.
2016-09-15 14:58:19 <skarnet> there is NO MORE GUARANTEE that env will be in /usr/bin than there is a guarantee that bash will be in /bin
2016-09-15 14:58:29 <^7heo> skarnet: why not /usr/bin/env btw?
2016-09-15 14:58:30 <skarnet> and probably less
2016-09-15 14:58:38 <^7heo> yeah maybe.
2016-09-15 14:58:39 <^7heo> I dunno
2016-09-15 14:58:39 <skarnet> because it only displaces the problem
2016-09-15 14:58:51 <^7heo> I've seen /usr/bin/env sh tho
2016-09-15 14:58:52 <skarnet> nothing guarantees that env will be in usr/bin
2016-09-15 14:58:54 <^7heo> that really killed me.
2016-09-15 14:59:13 <^7heo> nah but I mean, for example, if you have a python script, why not /usr/bin/env python?
2016-09-15 14:59:57 <skarnet> for python it MAY make sense, because the python interpreter location is probably less standardized than the env location
2016-09-15 15:00:04 <skarnet> but for shells, it's definitely not the case
2016-09-15 15:00:28 <skarnet> and anyway, the only way to make a correct shebang is to patch the script at installation time with the correct bindir for the machine
2016-09-15 15:01:10 <skarnet> so if you're providing a python script, ask the user the location of the python interpreter, and "make install" should write the correct shebang
2016-09-15 15:01:31 <^7heo> skarnet: yeah and that only works if the machine's setup stays unchanged.
2016-09-15 15:01:47 <skarnet> just like any hardcoded path
2016-09-15 15:01:50 <^7heo> if python moves from /usr/bin/ to /bin then...
2016-09-15 15:01:52 <^7heo> yeah.
2016-09-15 15:02:21 <^7heo> ok
2016-09-15 15:02:25 <skarnet> the chances that env moves from /usr/bin to /bin are much greater than the chances that python does the same
2016-09-15 15:02:35 <^7heo> my go build failed because gopkg.in was offline at that time.
2016-09-15 15:02:36 <^7heo> <3 go.
2016-09-15 15:02:44 <skarnet> on embedded environments with busybox with the "don't use /usr" option checked, env is /bin/env
2016-09-15 15:03:04 <skarnet> so, screw the /usr/bin/env assumption
2016-09-15 15:03:30 <jbguerraz> ^7heo: I dont' get it much but I've tried it and it breaks the build
2016-09-15 15:03:36 <skarnet> especially for shells - the convention to have shells in /bin is pretty universal
2016-09-15 15:03:42 <jbguerraz> I probably should RFTM about that prepend in source
2016-09-15 15:04:11 <jbguerraz> source="
2016-09-15 15:04:13 <^7heo> jbguerraz: wat?
2016-09-15 15:04:13 <jbguerraz>         $pkgname-$pkgver::https://github.com/${pkgname}/${pkgname}/archive/v${pkgver}.zip
2016-09-15 15:04:15 <jbguerraz>         rundeck.initd
2016-09-15 15:04:15 <^7heo> wait wat?
2016-09-15 15:04:17 <jbguerraz>         rundeck.confd
2016-09-15 15:04:19 <jbguerraz>         "
2016-09-15 15:04:21 <jbguerraz> y
2016-09-15 15:04:26 <jbguerraz> then
2016-09-15 15:04:27 <jbguerraz> /usr/bin/abuild: cd: line 1: can't cd to /home/jbguerraz/aports/testing/rundeck/src/rundeck-2.6.9
2016-09-15 15:04:30 <^7heo> jbguerraz: dude, answer on IRC comments from IRC
2016-09-15 15:04:38 <jbguerraz> sorry)
2016-09-15 15:04:39 <^7heo> answer on github to comments from github
2016-09-15 15:04:44 <jbguerraz> got it
2016-09-15 15:04:48 <^7heo> or we're gonna blow our brains out :D
2016-09-15 15:04:52 <^7heo> I mean at least for me
2016-09-15 15:04:56 <jbguerraz> roger that ;)
2016-09-15 15:05:54 <jbguerraz> anyway, I'm RTFM now ;)
2016-09-15 15:06:10 <^7heo> jbguerraz: I'm gonna fetch your PR and test it.
2016-09-15 15:06:55 <skarnet> jbguerraz: the rest of the script looks pretty much like standard shell to me
2016-09-15 15:07:16 <skarnet> if you replace the shebang with #!/bin/sh, maybe with 1 or 2 other modifs tops you could drop the bash dep
2016-09-15 15:07:38 <skarnet> ah, no, array at the end -_-
2016-09-15 15:07:50 <skarnet> KILL ALL THE BASH PEOPLE
2016-09-15 15:07:52 <skarnet> sorry
2016-09-15 15:07:55 <jbguerraz> )))
2016-09-15 15:08:04 <fcolista> :)
2016-09-15 15:11:22 <^7heo> jbguerraz: dude, your package has rotten deps
2016-09-15 15:11:35 <^7heo> let's see if it builds now.
2016-09-15 15:12:51 <jbguerraz> ^7heo: rundeck didn't let me any choice...
2016-09-15 15:13:03 <^7heo> s/let/give/
2016-09-15 15:13:08 <^7heo> yeah let's ee.
2016-09-15 15:13:10 <^7heo> see*
2016-09-15 15:13:16 <^7heo> dude that stuff is heavy shit.
2016-09-15 15:13:24 <^7heo> it's filling my term with dots for some reason
2016-09-15 15:13:38 <jbguerraz> I know the reason ;)
2016-09-15 15:13:45 <jbguerraz> gradle use some "scroll trick"
2016-09-15 15:13:45 <^7heo> why?
2016-09-15 15:13:52 <jbguerraz> which I disabled with the TERM=dumb
2016-09-15 15:13:56 <jirutka> ^7heo: no, I wrote some script in Perl few years ago, so I know some basics, but don’t use it
2016-09-15 15:14:01 <jbguerraz> otherwhise travis won't like it
2016-09-15 15:14:03 <^7heo> jbguerraz: ok
2016-09-15 15:14:06 <^7heo> rafuckyouirssi
2016-09-15 15:14:09 <^7heo> jirutka: ok
2016-09-15 15:14:16 <^7heo> jbguerraz: yeah I see.
2016-09-15 15:14:23 <^7heo> jbguerraz: is this gonna blow my laptop? :)
2016-09-15 15:14:35 <jbguerraz> heh. there is a chance.
2016-09-15 15:14:37 <jbguerraz> )
2016-09-15 15:14:59 <^7heo> also, are those 'http://repo1.maven.org/maven2/com/googlecode/javaewah/JavaEWAH/0.7.9/JavaEWAH-0.7.9.jar' and other files pinned in some way?
2016-09-15 15:15:07 <jirutka> ^7heo: jbguerraz: please don’t remove newlines between statements, it’s good for better readability; however I don’t know which PR
2016-09-15 15:15:10 <^7heo> like, with a versoin file
2016-09-15 15:15:13 <^7heo> s/oin/ion/
2016-09-15 15:15:28 <^7heo> jirutka: that was full of newlines.
2016-09-15 15:15:33 <^7heo> jirutka: really, really full
2016-09-15 15:15:41 <^7heo> jirutka: I found it much more readable without.
2016-09-15 15:16:23 <jirutka> skarnet: /usr/bin/env is still better than /usr/bin/bash, isn’t it?!
2016-09-15 15:16:47 <^7heo> jirutka: /usr/bin/bash is gonna crash mostly everywhere.
2016-09-15 15:17:15 <jirutka> skarnet: you’re wrong, it’s definitely more guarantee that there’s /usr/bin/env than that there’s /bin/bash… almost every unix and unix-like system has /usr/bin/env
2016-09-15 15:18:00 <jbguerraz> jirutka / ^7heo  : should I update it again with some newlines ? )
2016-09-15 15:18:07 <jbguerraz> jirutka: https://github.com/alpinelinux/aports/pull/296/files
2016-09-15 15:18:10 <jirutka> skarnet: “the chances that env moves from /usr/bin to /bin are much greater than the chances that python does the same” … sorry, but this is total bullshit
2016-09-15 15:18:32 <skarnet> jirutka: you're pretty antagonistic these days. Please give me a break.
2016-09-15 15:20:04 <jirutka> jbguerraz: no, leave it as it is now
2016-09-15 15:20:20 <skarnet> I gave you the example of busybox, which is real-life. At least I have env in /bin.
2016-09-15 15:20:36 <skarnet> I'm pretty sure I'm not the only one - it would be interesting to look at what OE does for instance.
2016-09-15 15:21:25 <skarnet> Whereas I don't know of a system that has python in /bin. I'm sure it exists somewhere, but I don't know it.
2016-09-15 15:21:25 <jirutka> ^7heo: jbguerraz: I will review that Java package later
2016-09-15 15:21:48 <jirutka> skarnet: what about finally killing this stupid /usr split and use just /bin?!
2016-09-15 15:22:12 <skarnet> there's a reason, and a good one, for the split.
2016-09-15 15:23:11 <skarnet> Having / on a separate partition - ultimately read-only, for guaranteed bootability - frees you of the need for a rescue system.
2016-09-15 15:23:21 <jirutka> we all know origins of /usr…
2016-09-15 15:23:32 <skarnet> Yes, for NFS mounts.
2016-09-15 15:23:39 <jirutka> no
2016-09-15 15:23:45 <skarnet> Now that's obsolete, but the idea of a small / is still good.
2016-09-15 15:24:07 <jirutka> what has read-only common with /usr? that’s about /var and others
2016-09-15 15:24:09 <skarnet> oh, because the fs didn't have support for all the binaries in /bin
2016-09-15 15:24:15 <skarnet> yeah, that was funny
2016-09-15 15:24:30 <skarnet> nothing, you can make /usr read-only if you want
2016-09-15 15:24:42 <jirutka> because not enough space on a disk… that was the original reason for /usr
2016-09-15 15:25:04 <^7heo> jbguerraz: that package takes AGES to build ffs.
2016-09-15 15:25:06 <skarnet> but the idea of having a small / with guaranteed bootability is still good, and including all the giant stuff you can install nowadays defeats that purpose
2016-09-15 15:25:19 <jirutka> and there’s no standard even for /bin/sh… on many system this is a symlink for bash
2016-09-15 15:25:45 <skarnet> of course it can be a symlink, so what?
2016-09-15 15:26:02 <skarnet> The only system that doesn't provide a POSIX shell in /bin/sh, however, is Solaris
2016-09-15 15:26:16 <jirutka> from what I’ve read few months ago, /usr/bin/env is less evil, supported on most of the systems; however, I definitely agree with /bin/sh, just don’t agree with hard-coding e.g. python binary path
2016-09-15 15:26:17 <skarnet> all the other ones follow the convention
2016-09-15 15:26:17 <jbguerraz> ^7heo: thx java ?!
2016-09-15 15:27:26 <^7heo> jbguerraz: I guess yeah
2016-09-15 15:27:30 <skarnet> if you want to #!/usr/bin/env python, I won't stop you. It makes reasonable sense. I'd prefer it if you found the location of "env" at install time though, and wrote #!$path_to_env python
2016-09-15 15:27:46 <skarnet> if you insist on using env to dynamically find python.
2016-09-15 15:28:20 <skarnet> but for bash? come on.
2016-09-15 15:28:42 <jirutka> skarnet: I disagree, in 99 % it would be /usr/bin/env… so you can write this to the script and if the actual location on target system is different, then just replace it with the actual location during installation… instead of forcing build to replace some placeholder everytime
2016-09-15 15:29:08 <skarnet> so instead of accommodating users, you force users to accommodate you?
2016-09-15 15:29:17 <skarnet> that's impolite.
2016-09-15 15:29:19 <jirutka> skarnet: yeah, I would also use /bin/bash for bash instead of env
2016-09-15 15:29:25 <jirutka> skarnet: no
2016-09-15 15:29:59 <jirutka> skarnet: I would just use actually working path instead of placeholder
2016-09-15 15:30:15 <jirutka> skarnet: that path can be replaced by the makefile or whatever
2016-09-15 15:30:52 <skarnet> it's the same thing: you need dynamic generation at install time. I don't care if you do it with cat, sed, or an ad-hoc Go program running in a Docker container shipped with the package just for this.,
2016-09-15 15:30:58 <jirutka> skarnet: it makes more sense for me than writing some arbitrary placeholder like #!$path_to_env to the scripts, so you can’t run it without replacing it for actual path
2016-09-15 15:31:42 <jirutka> skarnet: so we are basically in compliance
2016-09-15 15:31:51 <^7heo> jbguerraz: >>> WARNING: rundeck*: No arch specific binaries found so arch should probably be set to "noarch"
2016-09-15 15:31:55 <^7heo> jbguerraz: there's that.
2016-09-15 15:31:55 <jbguerraz> "or an ad-hoc Go program running in a Docker container" sounds so trendy :p
2016-09-15 15:32:01 <skarnet> if you get your rocks off having #!/usr/bin/env python in the script and running sed on it at installation time, instead of generating the shebang at installation time and just catting the rest of the script to it, then whatever floats your boat.
2016-09-15 15:32:31 <^7heo> jbguerraz: I got your last update with the changes I requested, and it built.
2016-09-15 15:32:47 <^7heo> jbguerraz: I don't get your issue with the tgz thing.
2016-09-15 15:32:57 <^7heo> jbguerraz: by the way, it's better to get the tgz than the zip
2016-09-15 15:33:27 <jbguerraz> ^7heo: I've fixed it by adding ".zip" as suffix of that prepend )
2016-09-15 15:33:41 <jbguerraz> will replace with tgz now (and move to "noarch"
2016-09-15 15:34:15 <^7heo> jbguerraz: no wiat
2016-09-15 15:34:16 <^7heo> wait*
2016-09-15 15:34:19 <^7heo> wtf fingers
2016-09-15 15:34:21 <^7heo> WHY
2016-09-15 15:34:25 <^7heo> anyway, wait
2016-09-15 15:34:28 <^7heo> tar.gz
2016-09-15 15:34:30 <^7heo> not tgz.
2016-09-15 15:35:09 <jbguerraz> yep, did it already
2016-09-15 15:35:11 <jbguerraz> going to push
2016-09-15 15:37:39 <jbguerraz> pushed
2016-09-15 15:39:00 <^7heo> jbguerraz: http://ix.io/1naS
2016-09-15 15:40:51 <jbguerraz> ?
2016-09-15 15:41:13 <^7heo> that's the diff I made, sorry if I lagged to paste it.
2016-09-15 15:41:35 <jbguerraz> oh ok
2016-09-15 15:42:01 <jbguerraz> on my side, I've replaced the first pkgname by github username (rundeck) as you've said in your right comment
2016-09-15 15:42:01 <^7heo> jbguerraz: for me this is good.
2016-09-15 15:42:07 <jbguerraz> that's why I was surprised
2016-09-15 15:42:12 <jbguerraz> like "wtf, it's back here?!"
2016-09-15 15:42:14 <jbguerraz> ;)
2016-09-15 15:42:16 <jbguerraz> great)
2016-09-15 15:42:30 <ncopa> anyone can help me find patches for tiff CVEs? https://dpaste.de/9oVM
2016-09-15 15:49:50 <jbguerraz> ncopa: https://git.centos.org/tree/rpms!libtiff.git/1ad9335dc0c1325262c62842eda01476243ec821/SOURCES maybe ?
2016-09-15 15:50:19 <jirutka> jbguerraz: what servlet container does it use to run?
2016-09-15 15:51:47 <jbguerraz> jirutka: I dunno
2016-09-15 15:52:02 <jbguerraz> jirutka: looks like jetty
2016-09-15 15:52:28 <jbguerraz> "Rundeck runs as a Spring-based webapp that runs in a servlet container (by default in Jetty on port 4440)"
2016-09-15 15:54:06 <jirutka> ^7heo: what about the caddy, why it fails?
2016-09-15 15:54:30 <^7heo> jirutka: I left a comment on the PR for that I think
2016-09-15 15:54:54 <^7heo> jirutka: "gopkg.in was unaccessible during the build, causing it to fail."
2016-09-15 15:54:58 <jirutka> ^7heo: huh, I thought that I commented on it… that Ijve already restarted it and it failed again
2016-09-15 15:55:11 <^7heo> jirutka: gonna read the thing agian.
2016-09-15 15:55:22 <^7heo> s/thing/log/; s/ian/ain/
2016-09-15 15:55:43 <^7heo> jirutka: that's another issue now.
2016-09-15 15:55:47 <^7heo> jirutka: I'll check it out.
2016-09-15 15:56:00 <^7heo> jirutka: "fatal: No names found, cannot describe anything."
2016-09-15 15:56:09 <^7heo> NO IDEA what binary is outputing that.
2016-09-15 15:56:10 <^7heo> wtf.
2016-09-15 15:56:54 <jirutka> ^7heo: IIRC this is abuild’s message that is in all builds and I have no clue what it actually mean
2016-09-15 15:57:26 <^7heo> jirutka: http://stackoverflow.com/questions/4916492/git-describe-fails-with-fatal-no-names-found-cannot-describe-anything#4916591
2016-09-15 15:57:37 <^7heo> algitbot: sure sure :P
2016-09-15 15:57:50 <^7heo> algitbot: I want the name of your provider, he seems to have good stuff ;)
2016-09-15 15:58:13 <^7heo> jirutka: it's gitsvn
2016-09-15 15:58:17 <jirutka> ^7heo: aha, so it’s from git
2016-09-15 15:58:27 <jirutka> ^7heo: does it work on your machine?
2016-09-15 15:58:49 <^7heo> jirutka: yes.
2016-09-15 15:59:06 <^7heo> $ ls -l packages/community/x86_64/caddy-0.9.1-r0.apk
2016-09-15 15:59:06 <^7heo> -rw-r--r--    1 theo     theo       3855329 Sep 15 16:35 packages/community/x86_64/caddy-0.9.1-r0.apk
2016-09-15 15:59:17 <^7heo> for all I know it does work.
2016-09-15 16:01:32 <^7heo> yaaay, gopkg.in is online again now!
2016-09-15 16:01:45 <^7heo> let's hope it doesn't crash again on that bs "cannot describe anything" message.
2016-09-15 16:02:09 <jirutka> weird, it works even on my machine
2016-09-15 16:02:35 <^7heo> jirutka: go.
2016-09-15 16:02:38 <jirutka> gopkg.in is a joke, right?
2016-09-15 16:02:40 <^7heo> jirutka: s/^/zal/
2016-09-15 16:03:01 <^7heo> jirutka: s/$/ He comes./
2016-09-15 16:03:53 <^7heo> jirutka: failed again, same error.
2016-09-15 16:05:07 <jirutka> ^7heo: I’ve merged it, it seems that it works on build server, so it’s just a Travis-related issue
2016-09-15 16:06:28 <^7heo> jirutka: yeah so, fine.
2016-09-15 16:06:30 <^7heo> jirutka: thanks ;)
2016-09-15 16:14:04 <jbguerraz> ncopa: does https://git.centos.org/tree/rpms!libtiff.git/1ad9335dc0c1325262c62842eda01476243ec821/SOURCES helps ? ;)
2016-09-15 16:50:04 <^7heo> I think that perl is starting to suck for me.
2016-09-15 19:31:25 <^7heo> jirutka: I just had the "No names found cannot describe anything" locally.
2016-09-15 19:31:30 <^7heo> jirutka: I think it's a make error.
2016-09-15 19:36:28 <coredumb> jirutka: seems like depends in _pyX doesn't work as expected
2016-09-15 19:36:45 <coredumb> pygit2 while having py2-cffi as dep doesn't pull it in
2016-09-15 19:37:26 <coredumb> ah wait that's in edge ... my bad
2016-09-15 20:00:04 <^7heo> jirutka: actually the makefile I found is using git describe
2016-09-15 20:00:09 <^7heo> so the output is the output of git describe
2016-09-15 20:02:30 <^7heo> git-describe - Describe a commit using the most recent tag reachable from it
2016-09-15 20:02:33 <^7heo> if there's no tag
2016-09-15 20:02:37 <^7heo> it won't be able to "describe"
2016-09-15 20:02:44 <^7heo> hence the "no names found cannot describe anything"
2016-09-15 20:11:28 <^7heo> so I got a coredump saying:
2016-09-15 20:11:32 <^7heo> (bt)
2016-09-15 20:11:33 <^7heo> #0  0x00006cf418659b6f in pthread_timedjoin_np () from /lib/ld-musl-x86_64.so.1
2016-09-15 20:11:36 <^7heo> #1  0x0000000000000000 in ?? ()
2016-09-15 20:11:52 <^7heo> I suspect that it's because there was a NULL pointer passed to the pthread_timedjoin_np function in musl?
2016-09-15 20:11:59 <^7heo> (since #1 is null)
2016-09-15 20:13:11 <^7heo> okay I found the faulty patch at least.
2016-09-15 20:18:36 <clandmeter> ^7heo: many go projects use git tags as versions, so if you download a tarball from github, it doesnt include the .git dir.
2016-09-15 20:18:49 <clandmeter> if you are still doing go stuff that is.
2016-09-15 20:20:04 <skarnet> clandmeter: it's not a real vacation if you don't unplug!
2016-09-15 20:20:28 <clandmeter> skarnet: tomorrow morning
2016-09-15 20:20:30 <clandmeter> :)
2016-09-15 20:20:37 <skarnet> I see :)
2016-09-15 20:26:35 <^7heo> clandmeter: yeah I see.
2016-09-15 20:26:39 <^7heo> clandmeter: but the makefile relies on it
2016-09-15 20:33:27 <nidan_> Does anyone know how apk behaves if two packages include the same file, with the same checksum?
2016-09-15 20:34:14 <nidan_> Does it still create a .apk-new file even though the files are the same? (Let's assume same u/g and perms too.)
2016-09-15 20:59:44 <kaniini> no
2016-09-15 20:59:46 <kaniini> it errors
2016-09-15 21:01:57 <coredumb> jirutka: both libcloud and pygit2 works OK with salt \o/
2016-09-15 21:03:40 <coredumb> but yeah I confirm the statement earlier
2016-09-15 21:03:54 <coredumb> depends in _py2 seems to not work correctly
2016-09-15 21:10:45 <^7heo> ncopa: found the bug with nlplug-findfs
2016-09-15 21:10:49 <^7heo> ncopa: it's using two devices
2016-09-15 21:11:04 <^7heo> ncopa: therefore it's calling the start_cryptsetup twice.
2016-09-15 21:11:22 <^7heo> ncopa: and because of that, the first one joins() fine, the second one, not so much.
2016-09-15 21:15:30 <^7heo> okay maybe not actually
2016-09-15 21:15:34 <^7heo> maybe it's because of something different.
2016-09-15 21:15:51 <^7heo> but it's weird that it calls start_cryptsetup twice
2016-09-15 21:20:55 <^7heo> okay, it's calling once with the action 'add', once with the action 'change'.
2016-09-15 21:24:38 <jirutka> coredumb: what do you mean that it does not work correctly? how it behaves?
2016-09-15 21:26:10 <coredumb> not pulling any deps
2016-09-15 21:26:13 <coredumb> :(
2016-09-15 21:26:26 <jirutka> coredumb: what PR?
2016-09-15 21:26:43 <coredumb> py-pygit2
2016-09-15 21:27:56 <jirutka> coredumb: when does it not pull any deps?
2016-09-15 21:28:23 <jirutka> coredumb: I mean, do you install pyX-pygit2 or py-pygit2?
2016-09-15 21:28:33 <^7heo> ncopa: did you ever test that binary (nlplug-findfs) before?
2016-09-15 21:28:40 <^7heo> I mean, out of using it at boot
2016-09-15 21:28:52 <^7heo> because, I am wondering if it's a behavior we introduced or not.
2016-09-15 21:28:55 <coredumb> jirutka: I install py2
2016-09-15 21:28:56 <^7heo> it might be tht not.
2016-09-15 21:29:01 <^7heo> s/ht/hat/
2016-09-15 21:29:05 <^7heo> raaah
2016-09-15 21:29:10 <^7heo> I'm too tired
2016-09-15 21:29:13 <^7heo> wrote mighat
2016-09-15 21:29:15 <^7heo> v_v
2016-09-15 21:29:26 <coredumb> it doesn't pull at install indeed
2016-09-15 21:29:28 <^7heo> s/hat/ht/; s/tht/that/
2016-09-15 21:30:22 <coredumb> deps from _py() are OK
2016-09-15 21:30:34 <^7heo> oh wait, it could be a problem introduced in search_cryptdev.
2016-09-15 21:31:29 <jirutka> coredumb: I’ve just tried it and it installs dependencies http://haste.fit.cvut.cz/notafen.log
2016-09-15 21:33:02 <jirutka> coredumb: I does not when you install py-pygit2 and don’t have any python installed… that’s known limitation of apk that is not fixed yet :(
2016-09-15 21:33:22 <coredumb> I only install py2*
2016-09-15 21:33:38 <coredumb> your log is incorrect
2016-09-15 21:33:47 <coredumb> it's not pulling py2-cffi nor py2-six
2016-09-15 21:34:04 <jirutka> coredumb: my log is correct, it’s from running system, I didn’t faked it ;)
2016-09-15 21:34:34 <coredumb> and it's not pulling deps as stated in depends in _py2()
2016-09-15 21:35:03 <jirutka> py2-pygit2 does not depend on py2-cffi or py2-six https://pkgs.alpinelinux.org/package/edge/community/x86_64/py2-pygit2
2016-09-15 21:35:15 <jirutka> ahaa
2016-09-15 21:35:18 <jirutka> wait a moment
2016-09-15 21:35:37 <coredumb> wut
2016-09-15 21:35:46 <jirutka> ah yes, I’m an idiot
2016-09-15 21:36:10 <coredumb> oh you didn't merge the correct one ?
2016-09-15 21:36:22 <jirutka> of course there must be depends="$depends …" in _py(), because we need to inherit depends from _py2()/_py3()
2016-09-15 21:36:44 <coredumb> ahhhhhhhhhhhhhhhhhhhh
2016-09-15 21:36:46 <coredumb> dang
2016-09-15 21:36:52 <jirutka> as I said, I’m idiot
2016-09-15 21:37:03 <jirutka> how can I overlook this
2016-09-15 21:37:35 <coredumb> you fix or I PR ?
2016-09-15 21:38:29 <coredumb> once fixed salt is fully functional with this and salt-cloud pulling py2-libcloud
2016-09-15 21:38:59 <coredumb> deployed successfully an instance on EC2 and fetched files from git repo
2016-09-15 21:40:02 <jirutka> it’s my fault, so I’ll fix it
2016-09-15 21:41:48 <jirutka> pushed
2016-09-15 21:42:05 <coredumb> cool
2016-09-15 21:43:16 <coredumb> if you don't have any other comment, salt is finished for me
2016-09-15 21:43:34 <^7heo> we have salt now?
2016-09-15 21:44:30 <jirutka> ^7heo: and what about sugar?
2016-09-15 21:45:02 <coredumb> ^7heo: salt with apk support ^^
2016-09-15 21:46:49 <^7heo> coredumb: cool
2016-09-15 21:46:55 <^7heo> jirutka: sugar is the new salt.
2016-09-15 21:52:56 <^7heo> ncopa: segfault fixed.
2016-09-15 21:53:18 <jirutka> ^7heo: great work! \o/
2016-09-15 21:55:47 <jirutka> coredumb: btw Salt does not support python 3?
2016-09-15 22:01:23 <^7heo> jirutka: thanks
2016-09-15 22:01:56 <^7heo> jirutka: that was a hairy one tho.
2016-09-15 22:05:46 <^7heo> however my code isn't working yet.
2016-09-15 22:05:53 <^7heo> the state is the same as before now
2016-09-15 22:05:57 <^7heo> but with no header support.
2016-09-15 22:14:00 <coredumb> jirutka: Python 2.6 >= 2.6 <3.0
2016-09-15 22:14:03 <coredumb> so no
2016-09-15 22:16:25 <jirutka> coredumb: there should be move instead of copy, shouldn’t be? https://github.com/alpinelinux/aports/pull/282/files#diff-6a454f92f63557b8a626be4f06d32e81R61
2016-09-15 22:17:24 <jirutka> coredumb: currently /usr/bin/salt would be both in base pacakge and master subpackage
2016-09-15 22:18:39 <coredumb> jirutka: you right I missed it
2016-09-15 22:18:48 <jirutka> coredumb: I’ll fix it before merging
2016-09-15 22:19:02 <coredumb> I can push if you want
2016-09-15 22:49:42 <coredumb> thx jirutka
2016-09-15 22:49:52 <jirutka> coredumb: your welcome!
2016-09-15 22:50:00 <jirutka> s/your/you’re/
2016-09-15 22:50:07 <coredumb> :)
2016-09-15 23:02:38 <coredumb> jirutka: https://github.com/alpinelinux/aports/pull/307 :D
2016-09-15 23:02:44 <coredumb> forgot that ^^
2016-09-15 23:03:25 <jirutka> coredumb: does fcolista know about it?
2016-09-15 23:04:30 <coredumb> yep I asked him here and he said it was OK
2016-09-15 23:06:51 <jirutka> okay
2016-09-15 23:08:40 <coredumb> thx
2016-09-15 23:13:39 <^7heo> fuck
2016-09-15 23:13:41 <^7heo> I'm still at work
2016-09-15 23:13:43 <^7heo> I'm tired af
2016-09-15 23:13:47 <^7heo> My dog's sleeping
2016-09-15 23:13:53 <^7heo> and I can't seem to fix that sh*t
2016-09-15 23:28:32 <coredumb> then do it tomorrow
2016-09-15 23:28:50 <coredumb> a good sleep is always better when you can't fix something ;)
2016-09-15 23:32:37 <^7heo> yeah well
2016-09-15 23:32:39 <^7heo> I think I fixed it.
2016-09-15 23:33:04 <coredumb> then => sleep :P
2016-09-15 23:33:13 <^7heo> I'm at work.
2016-09-15 23:33:17 <^7heo> I need to go home first.
2016-09-15 23:33:20 <^7heo> ...
2016-09-15 23:34:07 <^7heo> I FUCKING FIXED IT.
2016-09-15 23:34:19 <skarnet> grats
2016-09-15 23:34:44 <^7heo> thanks man.
2016-09-15 23:34:51 <^7heo> I feel like getting wasted now.
2016-09-15 23:34:53 <^7heo> like
2016-09-15 23:34:55 <^7heo> getting a beer
2016-09-15 23:34:57 <^7heo> smoking a joint
2016-09-15 23:35:04 <^7heo> eating shitloads of pizza
2016-09-15 23:35:26 <skarnet> I'd advise doing just one of those three
2016-09-15 23:35:44 <skarnet> more will make you have a shitty night :P
2016-09-15 23:36:32 <^7heo> beer + pizza is okay.
2016-09-15 23:36:37 <^7heo> the point is
2016-09-15 23:36:41 <^7heo> the real issue is beer + joint
2016-09-15 23:37:11 <skarnet> if you eat shitloads of pizza and drink over it, your digestion will be slow af
2016-09-15 23:37:20 <skarnet> and you'll feel the pizza all night.
2016-09-15 23:37:54 <^7heo> true
2016-09-15 23:37:58 <^7heo> but if I drink and smoke
2016-09-15 23:38:07 <^7heo> I'm gonna have a hard time with a lot of dizziness
2016-09-15 23:38:10 <^7heo> and spinning
2016-09-15 23:38:16 <^7heo> UNLESS I EAT AF
2016-09-15 23:38:20 <^7heo> :P
2016-09-15 23:38:29 <skarnet> then just eat and smoke
2016-09-15 23:38:32 <^7heo> I guess yeah.
2016-09-15 23:38:34 <^7heo> So
2016-09-15 23:38:41 <^7heo> I'm gonna make a real life test
2016-09-15 23:42:38 <^7heo> okay
2016-09-15 23:42:42 <^7heo> Wish me luck
2016-09-15 23:42:46 <^7heo> let's GOOOO
2016-09-15 23:44:34 <^7heo> IT FUCKING WORKS!
2016-09-15 23:44:35 <^7heo> YES!
2016-09-15 23:44:37 <^7heo> YEEEEEES!
2016-09-15 23:44:41 <^7heo> OMG.
2016-09-15 23:44:42 <skarnet> \o/
2016-09-15 23:44:57 <^7heo> tomorrow I'm gonna setup my deported header on my work laptop
2016-09-15 23:45:04 <^7heo> and I'm gonna have the best system EVAR.
2016-09-15 23:45:10 <^7heo> MOST SECURE SHIT EVAR.
2016-09-15 23:45:20 <skarnet> famous last words
2016-09-15 23:45:26 <^7heo> true,.
2016-09-15 23:45:32 <^7heo> if someone steals the laptop, they won't ever have the possibility of descripting the hard drive
2016-09-15 23:45:37 <^7heo> they will have a 2k laptop tho.
2016-09-15 23:45:43 <^7heo> but NOT company secret.
2016-09-15 23:45:46 <^7heo> YAY
2016-09-15 23:45:53 <^7heo> s/NOT/NO/
2016-09-15 23:46:03 <skarnet> ah, that's what you were working on.
2016-09-15 23:46:07 <^7heo> eyah
2016-09-15 23:46:08 <^7heo> oops
2016-09-15 23:46:09 <^7heo> yeah
2016-09-15 23:46:21 <^7heo> deported LUKS header.
2016-09-15 23:46:30 <^7heo> in nlplug-findfs
2016-09-15 23:46:33 <^7heo> now works.
2016-09-15 23:46:45 <^7heo> alpine can have some secure setups now.
2016-09-15 23:46:57 <^7heo> it costs 50 bucks to get a usb key with a pinpad on it
2016-09-15 23:47:02 <^7heo> but then you kinda have something secure.
2016-09-15 23:47:11 <^7heo> ok
2016-09-15 23:47:16 <^7heo> switching to phone
2016-09-15 23:47:19 <^7heo> going home by foot
2016-09-15 23:47:20 <^7heo> at 2 am
2016-09-15 23:47:21 <^7heo> o
2016-09-15 23:47:23 <^7heo> o/
2016-09-16 07:08:28 <ScrumpyJack> morning, and Happy Friday!
2016-09-16 07:09:43 <coredumb> morning ScrumpyJack
2016-09-16 07:34:57 <jbguerraz> morning!
2016-09-16 07:35:04 <jbguerraz> damn, it's friday already :s
2016-09-16 09:32:25 <jirutka> already friday?! what happened to all the week :/
2016-09-16 09:38:10 <nidan_> kaniini: Ok, thanks.
2016-09-16 10:05:53 <^7heo> moin
2016-09-16 10:05:55 <^7heo> skarnet: you about?
2016-09-16 10:07:40 <skarnet> what's up? I need to run some errands.
2016-09-16 10:13:25 <^7heo> well
2016-09-16 10:13:42 <^7heo> Do you have any clue where to start to find those leaking descriptors in nlplug-findfs?
2016-09-16 10:13:52 <^7heo> because I do not think it comes from there.
2016-09-16 10:25:57 <^7heo> I like the fix: https://bugs.launchpad.net/ubuntu/+source/lvm2/+bug/591823
2016-09-16 10:26:12 <^7heo> "If you see messages like these:"
2016-09-16 10:26:49 <^7heo> "File descriptor X (...) leaked on lvm invocation. Parent PID Y: /..."
2016-09-16 10:26:57 <^7heo> "You can set LVM_SUPPRESS_FD_WARNINGS to suppress these warnings."
2016-09-16 10:27:02 <^7heo> \o/
2016-09-16 10:27:12 <^7heo> IT'S NOT A PROBLEM IF I DO NOT SEE IT!
2016-09-16 10:27:17 <^7heo> Huurrayyyy
2016-09-16 10:40:38 <skarnet> they're right
2016-09-16 10:40:52 <skarnet> the existence of the message is a bug
2016-09-16 10:41:06 <skarnet> it's perfectly valid to execute a program with more open fds than 0, 1 and 2
2016-09-16 10:41:46 <skarnet> now, granted, if nlplug-findfs invokes lvs and doesn't *intend* to keep some other fd open, that's also a problem
2016-09-16 10:42:33 <skarnet> but intrinsically, lvs just shouldn't care - we can thank it for bringing the fd to our attention, but it's not its job to do so
2016-09-16 10:44:46 <skarnet> http://git.alpinelinux.org/cgit/mkinitfs/tree/nlplug-findfs.c#n626 not sure if it's a good idea to have a "searchdev" parameter inside the "searchdev" function
2016-09-16 10:44:50 <skarnet> it may work, but it's confusing
2016-09-16 10:47:23 <^7heo> skarnet: thanks for reviewing
2016-09-16 10:48:01 <^7heo> skarnet: will fix.
2016-09-16 10:48:54 <^7heo> skarnet: also the dev git tip, since I am working on it atm, and I do not have access to g.a.o is: https://github.com/7heo/mkinitfs/blob/master/nlplug-findfs.c
2016-09-16 10:49:20 <^7heo> so I'll fix https://github.com/7heo/mkinitfs/blob/master/nlplug-findfs.c#L653
2016-09-16 10:49:34 <skarnet> https://github.com/7heo/mkinitfs/blob/master/nlplug-findfs.c#L115
2016-09-16 10:49:51 <^7heo> skarnet: I know.
2016-09-16 10:49:52 <skarnet> if you execve() right after fork()ing
2016-09-16 10:50:00 <skarnet> without doing anything in the meantime
2016-09-16 10:50:02 <^7heo> ah good point.
2016-09-16 10:50:05 <^7heo> I did not notice.
2016-09-16 10:50:08 <skarnet> use posix_spawnp() instead
2016-09-16 10:50:22 <^7heo> well, yeah.
2016-09-16 10:50:36 <^7heo> skarnet: would it be fine for you to create issues on the github repo?
2016-09-16 10:50:51 <^7heo> skarnet: or do you prefer me to take your comments here and create issues there?
2016-09-16 10:51:09 <^7heo> also, you can send in PRs if you wish, I'll pull.
2016-09-16 10:51:11 <skarnet> the less I use github, the better I fare
2016-09-16 10:51:14 <^7heo> okay
2016-09-16 10:51:17 <^7heo> then I'll do it.
2016-09-16 10:51:21 <^7heo> I thank you for your feedback tho.
2016-09-16 10:51:28 <^7heo> should I mention you in the tickets I open, or would you rather not?
2016-09-16 10:52:04 <skarnet> idc either way. Also, those are nitpicks - I'm just having a quick look to see if I can find the fd leak.
2016-09-16 10:52:14 <^7heo> You're awesome.
2016-09-16 10:52:17 <skarnet> I don't have time for a thorough review right now.
2016-09-16 10:52:24 <skarnet> Will do one in a few months. ;)
2016-09-16 10:52:36 <^7heo> imho the fd leak comes from the fact that we're using the same spawnmgr for nlplug-findfs and lvm
2016-09-16 10:53:26 <^7heo> https://github.com/7heo/mkinitfs/blob/master/nlplug-findfs.c#L768
2016-09-16 10:53:27 <^7heo> and
2016-09-16 10:53:40 <^7heo> https://github.com/7heo/mkinitfs/blob/master/nlplug-findfs.c#L349
2016-09-16 10:54:10 <skarnet> honestly the quickest way to find the fd leak would be to strace it, or stop it while it's running and look at /proc/$blah/fd/
2016-09-16 10:54:12 <^7heo> nlplug-findfs uses the cryptsetup bindings to manage/open the cryptfs
2016-09-16 10:54:29 <^7heo> yeah but given when it's executing it's delicate to do.
2016-09-16 10:54:31 <^7heo> I thought about that.
2016-09-16 10:54:45 <^7heo> the only way I can replicate that is by setting up lvm over cryptsetup in loop devices.
2016-09-16 10:54:50 <^7heo> not sure I want to :P
2016-09-16 10:54:53 <skarnet> printf debugging.
2016-09-16 10:54:53 <^7heo> but maybe I will.
2016-09-16 10:54:56 <^7heo> yeah
2016-09-16 10:54:59 <^7heo> that's what I'm doing for now.
2016-09-16 10:55:00 <^7heo> but again.
2016-09-16 10:55:06 <skarnet> Make it list /proc/$$/fd.
2016-09-16 10:55:15 <^7heo> can't confirm, since I am *not* using LVM in the test script.
2016-09-16 10:55:20 <skarnet> ah, yes.
2016-09-16 10:55:27 <skarnet> but start_lvm2 doesn't look problematic.
2016-09-16 10:55:36 <skarnet> Chances are the fd is already open much earlier.
2016-09-16 10:55:37 <^7heo> the fact that it's reusing the same spawnmgr
2016-09-16 10:55:39 <^7heo> could be
2016-09-16 10:55:40 <^7heo> or not?
2016-09-16 10:56:02 <^7heo> I'm not saying that it's opening the fds in the start_lvm2 no.
2016-09-16 10:56:02 <skarnet> I don't know what a spawnmgr is. I saw a bunch of things with malloc, I passed.
2016-09-16 10:56:07 <^7heo> ok :D
2016-09-16 10:56:12 <^7heo> Well, I'll check later.
2016-09-16 10:56:23 <^7heo> for now the important part is that I fixed the deported luks header
2016-09-16 10:56:44 <^7heo> the rest isn't functionality, it's security, so it's important; but does not prevent me from booting.
2016-09-16 11:01:47 <^7heo> skarnet: created issues for your remarks.
2016-09-16 11:01:53 <^7heo> skarnet: I'll be sure to refactor.
2016-09-16 11:02:04 <^7heo> thanks again.
2016-09-16 11:08:25 <skarnet> I still don't understand why you need a spawn queue.
2016-09-16 11:09:03 <skarnet> https://github.com/7heo/mkinitfs/blob/master/nlplug-findfs.c#L115 please don't use signal(). Use sigaction() instead. signal() is obsolescent (if it's not yet, it should be).
2016-09-16 11:09:46 <jirutka> could please someone look at libtheora, why it fails on armhf? http://build.alpinelinux.org/buildlogs/build-edge-armhf/main/libtheora/libtheora-1.1.1-r11.log
2016-09-16 11:13:48 <jirutka> and also krb5 http://build.alpinelinux.org/buildlogs/build-edge-armhf/main/krb5/krb5-1.14.3-r0.log
2016-09-16 11:17:31 <skarnet> https://github.com/7heo/mkinitfs/blob/master/nlplug-findfs.c#L189 it would be nicer to have major and minor as uints, since it's always how they're used. The atoi conversion should be when filling the struct uevent, not when using major and minor.
2016-09-16 11:20:49 <skarnet> welp, no obvious fd leak in nlplug-findfs afaict. I'd look at the program that calls nlplug-findfs, maybe the open fd is already inherited.
2016-09-16 11:21:22 <skarnet> lots of stuff to simplify and refactor, but no huge bug at first sight.
2016-09-16 11:23:01 <^7heo> skarnet: thanks for checking.
2016-09-16 11:23:16 <^7heo> skarnet: you shall get a beer when we meet ;)
2016-09-16 11:23:18 <^7heo> or many.
2016-09-16 11:23:25 <^7heo> I will beer you.
2016-09-16 11:23:35 <skarnet> I'll give the beer to clandmeter
2016-09-16 11:23:40 <^7heo> you don't beer?
2016-09-16 11:23:46 <skarnet> I don't like it.
2016-09-16 11:23:56 <^7heo> m'kay
2016-09-16 11:23:59 <^7heo> Whiskey?
2016-09-16 11:24:16 <skarnet> Not really. I drink very little alcohol.
2016-09-16 11:24:50 <^7heo> ok
2016-09-16 11:24:52 <scadu> well, you can always smoke together :f
2016-09-16 11:24:55 <^7heo> skarnet: https://github.com/7heo/mkinitfs/blob/master/nlplug-findfs.c#L238-L247 about signal, or?
2016-09-16 11:25:23 <^7heo> (I didn't find a signal() on Line 115)
2016-09-16 11:25:34 <skarnet> scadu: I'll drink an entire cellar's worth of whiskey before I take a puff of smoke :P
2016-09-16 11:25:47 <^7heo> :D
2016-09-16 11:25:51 <^7heo> skarnet: what about vaping?
2016-09-16 11:26:20 <skarnet> ^7heo: uh, yes, those lines, not 115.
2016-09-16 11:26:26 <skarnet> vaping is -offtopic
2016-09-16 11:27:58 <jirutka> ^7heo: I’ve created mkinitfs mirror on GH, so you can send PR when you’re done https://github.com/alpinelinux/mkinitfs ;)
2016-09-16 11:30:23 <skarnet> jirutka: about krb5, off the top of my head, looks like there's a missing dep to gettext-dev
2016-09-16 11:30:40 <jirutka> skarnet: thanks for the tip, I’ll try it
2016-09-16 11:31:51 <skarnet> same thing with libtheora, it's also dependency hell, but since I know nothing about graphical libraries I can't help you :/
2016-09-16 11:31:53 <^7heo> jirutka: thanks ;)
2016-09-16 11:32:03 <^7heo> jirutka: I'll create the further issues there.
2016-09-16 11:38:55 <^7heo> jirutka: https://github.com/7heo/mkinitfs/tree/deported-header
2016-09-16 11:39:20 <jirutka> ^7heo: what about it?
2016-09-16 11:39:21 <^7heo> jirutka: I "forked" the alpinelinux repo, and put my work there, in a new branch.
2016-09-16 11:39:32 <^7heo> so it's closer to the workflow we have on gh usually.
2016-09-16 11:39:49 <^7heo> and we can see the diff with the alpinelinux one as reported by gh
2016-09-16 11:40:36 <jirutka> ^7heo: btw == is not in POSIX-sh https://github.com/7heo/mkinitfs/commit/60816fc61224654f6ab1c8d01e8bc4b45ae90533#diff-1b531c5e405e0bc95890d52cf2f2f5b5R43
2016-09-16 11:40:48 <jirutka> ^7heo: it’s weird, but POSIX-sh defines only =
2016-09-16 11:40:49 <^7heo> jirutka: now I have to manage the issues, and then I'll be done.
2016-09-16 11:40:54 <^7heo> jirutka: a friend already told me.
2016-09-16 11:41:07 <^7heo> jirutka: it's just a test script, it's not meant to stay there for now if we don't want it.
2016-09-16 11:41:17 <^7heo> jirutka: tbh I slapped it as fast as I could.
2016-09-16 11:41:26 <^7heo> jirutka: but thanks for reminding me.
2016-09-16 11:41:26 <jirutka> now, tests scripts are always useful, don’t remove it
2016-09-16 11:41:31 <^7heo> ok
2016-09-16 11:41:34 <^7heo> then I'll fix it.
2016-09-16 11:49:18 <ScrumpyJack> rabbitmq-server apk is broken :(
2016-09-16 11:50:21 <ScrumpyJack> anyone know the contributor/maintainer?
2016-09-16 11:50:30 <^7heo> isn't it writen?
2016-09-16 11:50:32 <^7heo> written*
2016-09-16 11:51:08 <^7heo> Maintainer: Nathan Johnson <nathan@nathanjohnson.info>
2016-09-16 11:51:21 <ScrumpyJack> ^7heo: as in, anyone *know* him  - is he *familiar* to anyone :)
2016-09-16 11:51:39 <^7heo> ah ok
2016-09-16 11:57:04 <skarnet> ScrumpyJack: he's the guy I bothered about the mailing-list server spamblocking me, so in a way, I know who he is ;)
2016-09-16 12:18:57 <ScrumpyJack> skarnet: that was another Nathan (Angelacos, aka nangel) :(
2016-09-16 12:32:23 <skarnet> ah, damn, my memory is so holy at times.
2016-09-16 12:32:33 <skarnet> Not because it's sacred. Because it has holes in it.
2016-09-16 12:39:54 <jirutka> XD
2016-09-16 15:27:16 <gena01> ^7heo, hi, not sure about those 3 lines, but it sorta works for me, if there's some specific way we should be doing Go apps, let me know. P.S. I sent in PR for consul-template
2016-09-16 15:29:03 <gena01> ^7heo, speaking of consul, noticed that one of the architectures is disabled (didn't feel like touching, so not sure if it's still needed)
2016-09-16 17:16:22 <fcolista> for all of you, shell wizard: i've a CSV contains a variable. Let's say <%CHANGEME%>. I want to change it according with a value i find in the row. So, if in the same row i match this regexp: [A-Z][0-9]{3}  then <%CHANGEME%> should be "foo"
2016-09-16 17:16:48 <fcolista> if i match another regexp, <%CHANGEME%> should be "bar"
2016-09-16 17:21:00 <skarnet> is your regex->value map fixed?
2016-09-16 17:21:08 <skarnet> do you read the CSV from a fd ?
2016-09-16 17:21:15 <fcolista> no
2016-09-16 17:21:17 <fcolista> yes
2016-09-16 17:21:35 <skarnet> is your number of fields known in advance?
2016-09-16 17:21:36 <fcolista> regex changes
2016-09-16 17:21:52 <fcolista> no, number of fields are different
2016-09-16 17:22:09 <skarnet> my advice would be "use awk"
2016-09-16 17:22:14 <fcolista> yes
2016-09-16 17:22:20 <skarnet> but I don't know awk XD
2016-09-16 17:22:21 <fcolista> correct.
2016-09-16 17:22:52 <fcolista> i can itherate for each field
2016-09-16 17:23:02 <fcolista> check the field against a pattern
2016-09-16 17:23:34 <fcolista> but i need to substitute the variable in the same row the pattern is matched
2016-09-16 17:24:02 <fcolista> if i could implement "sed" inside awk i've solved.
2016-09-16 17:24:09 <jbguerraz> 
ACTION says bye bye ! enjoy the weekend ;)

2016-09-16 17:24:10 <fcolista> umh
2016-09-16 17:24:19 <skarnet> iterate twice: one to find the pattern that is matched, one to find the variable to replace with the value you computed in the first pass
2016-09-16 17:25:00 <fcolista> umh...
2016-09-16 17:25:06 <fcolista> not very clean
2016-09-16 17:25:20 <fcolista> something like that:
2016-09-16 17:25:20 <fcolista> cat myfile|awk -F "*" '{print $1F |sed 's/#//g' >> "tfile1" ; print $NF >> "lastfile"} '
2016-09-16 17:25:27 <fcolista> (which is horrible, i agree)
2016-09-16 17:25:46 <skarnet> useless use of cat (tm)
2016-09-16 17:25:52 <fcolista> :)
2016-09-16 17:26:04 <skarnet> again, I can't speak awk
2016-09-16 17:26:05 <fcolista> just copied and pasted an example
2016-09-16 17:26:18 <fcolista> looks that gsub is the way
2016-09-16 17:26:22 <fcolista> in awk
2016-09-16 17:26:34 <skarnet> but it seems clean to do the first operation in one pass and the second operation in another
2016-09-16 17:26:45 <skarnet> since you have 2 different things to do
2016-09-16 17:27:04 <skarnet> 1. match a shitload of regexes to compute $value
2016-09-16 17:27:10 <fcolista> since i've three different regexp, i should do that 6 times.
2016-09-16 17:27:16 <skarnet> 2. match a variable name to replace with $value
2016-09-16 17:27:38 <skarnet> no, you do all the regexes in one go
2016-09-16 17:27:44 <skarnet> (for each line)
2016-09-16 17:27:52 <skarnet> so you know the value
2016-09-16 17:27:55 <skarnet> and THEN you sed
2016-09-16 17:29:31 <fcolista> mumble mumble..
2016-09-16 17:29:53 <fcolista> why is so difficult write in code what is clear in your mind ?
2016-09-16 17:32:33 <skarnet> because scripting languages suck
2016-09-16 19:29:20 <^7heo> gena01: the question isn't if it works.
2016-09-16 19:29:27 <^7heo> gena01: the question is about what we want ;)
2016-09-16 19:30:15 <gena01> well, what do we want?
2016-09-16 19:30:46 <^7heo> that's the question
2016-09-16 19:31:25 <^7heo> personally, because it is possible to get go to re-download the app we're building via git and screw up the versionning, if we use the method you've used in the current APKBUILD, I would go for the other way
2016-09-16 19:31:53 <gena01> well. do we need to go that route? we can grab the files and just build
2016-09-16 19:32:03 <gena01> with checksum and stuff, like any other package
2016-09-16 19:32:33 <^7heo> nah
2016-09-16 19:32:37 <^7heo> go sucks with dependencies
2016-09-16 19:32:47 <^7heo> and we *do* have to address that problem
2016-09-16 19:32:49 <gena01> well, govendoring is on and by default in 1.8
2016-09-16 19:32:53 <^7heo> or we won't get reproducible builds
2016-09-16 19:33:06 <^7heo> in 1.7 too normally.
2016-09-16 19:33:12 <^7heo> but that's still a problem
2016-09-16 19:33:19 <gena01> i don't think it's behind a flag anymore in 1.8
2016-09-16 19:33:27 <^7heo> because it's new and not everybody uses it, let alone uses it right
2016-09-16 19:33:37 <^7heo> gena01: in 1.7 neither.
2016-09-16 19:33:43 <gena01> well.. that problem is not specific to go ;-)
2016-09-16 19:33:49 <^7heo> it kinda is.
2016-09-16 19:34:04 <^7heo> because other languages have solutions to freeze deps
2016-09-16 19:34:14 <gena01> doesn't mean people are doing it right
2016-09-16 19:34:19 <^7heo> Go is like "includes are URLs, just get the tip and you're set"
2016-09-16 19:34:40 <gena01> well isn't the commit vendor folder the thing now?
2016-09-16 19:34:47 <^7heo> not "the" thing no.
2016-09-16 19:34:53 <^7heo> it's "a" thing at best.
2016-09-16 19:35:12 <^7heo> this works as long as people use git correctly
2016-09-16 19:35:16 <gena01> i've committed vendor folder in other projects and languages.. but yes
2016-09-16 19:35:19 <^7heo> i.e. develop on their branches
2016-09-16 19:35:23 <^7heo> and merge when the feature is mature
2016-09-16 19:35:29 <^7heo> but if they dev on master
2016-09-16 19:35:38 <^7heo> then it's fucked
2016-09-16 19:35:48 <^7heo> and so go is kinda fucked because it assumes best practices.
2016-09-16 19:37:48 <gena01> but projects will follow what works for them
2016-09-16 19:37:52 <^7heo> yeah
2016-09-16 19:37:55 <^7heo> and ymmv
2016-09-16 19:38:00 <^7heo> and that isn't a good thing.
2016-09-16 19:38:15 <^7heo> as a distribution, we can't afford to learn a new way of using anything for each new package
2016-09-16 19:39:12 <gena01> so then we don't package some projects?
2016-09-16 19:39:20 <^7heo> yeah no.
2016-09-16 19:39:24 <^7heo> that's a non-solution
2016-09-16 19:39:58 <^7heo> bbl, drinking
2016-09-16 19:46:56 <jirutka> I vote for not packaging craps
2016-09-16 19:47:43 <jirutka> even f*cking PHP has now better culture of versioning and dependencies management than Go… go is just complete disaster
2016-09-16 19:54:56 <coredumb> who cares there's no PHP app as good as consul or terraform
2016-09-16 19:55:03 <coredumb> so I vote for packaging
2016-09-16 19:55:12 <coredumb> :)
2016-09-16 19:55:27 <coredumb> even if it's fscking PITA
2016-09-16 19:55:33 <gena01> also hashicorp (at least few apps that I've compiled) tend to commit dependencies into vendor folder in git
2016-09-16 19:55:56 <coredumb> yeah now they do
2016-09-16 19:56:28 <gena01> coredumb, funny that you mention it. consul 0.7 PR that I opened started the whole discussion (was merged <24 hours ago)
2016-09-16 19:57:58 <coredumb> gena01: yeah I pushed the first version of the APKBUILD
2016-09-16 19:58:08 <coredumb> such a PITA
2016-09-16 19:58:34 <coredumb> and I seriously think that the problem is not Go per se
2016-09-16 19:58:59 <coredumb> it's the way all those fucking devops kikoolol hispter are handling their projects
2016-09-16 19:59:13 <gena01> i cleaned up the build, which removed quite a bit of the complexity, at least I tried
2016-09-16 20:01:40 <coredumb> I should give these new release a try
2016-09-16 20:01:52 <coredumb> I still run 0.5.4 in production :O
2016-09-16 20:03:47 <gena01> coredumb, you should. There's now a log file and config check before service start.. if you using openrc
2016-09-16 20:05:06 <coredumb> talking about hashicorp
2016-09-16 20:05:16 <coredumb> nomad is also a cool tool
2016-09-16 20:19:16 <jirutka> coredumb: maybe you’ve misunderstood me, I definitely do not prefer PHP, I don’t even consider it as a real language
2016-09-16 20:19:55 <jirutka> gena01: commiting dependencies to git repository is uterly stupid
2016-09-16 20:20:43 <jirutka> coredumb: interesting, I think that there’s even more hipster idiots in JS and still, they are able to handle dependencies…
2016-09-16 20:21:01 <gena01> i never said I didn't do stupid things (including PHP)
2016-09-16 20:51:55 <gena01> speaking of PHP, can somebody bump community/php7 to 7.0.11?
2016-09-17 09:15:04 <^7heo> I agree with all that was said
2016-09-17 09:17:08 <^7heo> PHP is horrendous, but does a better job at managing deps than go; the problem with go is that when designing go, google assumed best practices everywhere; and yes, while more hype and shitty, js is better at its job than go, because older.
2016-09-17 09:17:36 <^7heo> for example
2016-09-17 09:17:43 <^7heo> (about best practices)
2016-09-17 09:17:58 <^7heo> if you use the master branch of your repo as the 'release branch' and only develop in other branches
2016-09-17 09:18:24 <^7heo> go repos *can* be used as deps without a problem, because getting "tip" just means "getting latest stable"
2016-09-17 09:19:56 <^7heo> if you decentralize your git repo on the layer 3 (as opposed to 7 - i.e. by having multiple servers with the same IP address so you can do CDN right)
2016-09-17 09:20:05 <^7heo> you can use the urls in includes
2016-09-17 09:20:31 <^7heo> and that will ALWAYS work.
2016-09-17 09:21:23 <^7heo> now, when you have stuff like pkg.in or whatnot, that flickers like a f*cking christmas tree, commits to the master branch every other minute and people who never tag releases and use git as if it was a cheap CVS...
2016-09-17 09:21:28 <^7heo> then go sucks.
2016-09-17 09:21:43 <^7heo> so the problem with go isn't that go sucks, it's that people suck.
2016-09-17 09:22:04 <^7heo> or differently said, the problem with go is that it assumes intelligent people.
2016-09-17 09:23:17 <^7heo> and that's not surprising, from an organisation that hires the most intelligent people around, to have a biased POV on the average IQ of humanity.
2016-09-17 09:23:30 <^7heo> But yeah
2016-09-17 09:23:35 <^7heo> people suck, so go sucks.
2016-09-17 09:24:51 <^7heo> now, I have to work a bit on that nlplug-findfs, with some luck, I'll be done done by tonight.
2016-09-17 10:10:55 <jirutka> I strongly disagree about that google assumed best practices… they did exactly the opposite, they arrogantly ignores all best practices and enforces bad practices
2016-09-17 10:10:59 <jirutka> ^7heo: ^
2016-09-17 10:12:34 <jirutka> ^7heo: you’re wrong, go doesn’t assume intelligent people, exactly opposite, read the quote of Rob Pike http://nomad.so/2015/03/why-gos-design-is-a-disservice-to-intelligent-programmers/
2016-09-17 11:07:33 <^7heo> jirutka: he f*cking designed it!
2016-09-17 11:07:50 <jirutka> ^7heo: yes, I know
2016-09-17 11:09:16 <^7heo> also
2016-09-17 11:09:27 <^7heo> that does not contradict any of the points I made so far.
2016-09-17 11:09:32 <^7heo> He says its stupid for different reasons.
2016-09-17 11:11:44 <^7heo> I'm gonna keep your link tho, it's nice.
2016-09-17 11:12:01 <^7heo> Also, the segfault I solved in nlplug-findfs was there before I started working on it.
2016-09-17 11:12:11 <^7heo> according to my test program
2016-09-17 11:12:17 <jirutka> ^7heo: I haven’t sent it to you yet? then you must also read this one http://yager.io/programming/go.html
2016-09-17 11:18:52 <^7heo> dang
2016-09-17 11:19:08 <^7heo> the internet provider at the cafe I'm at is really shitty
2016-09-17 11:19:14 <^7heo> it's fast but the connections die all the time
2016-09-17 11:19:21 <^7heo> there's no access on port 22
2016-09-17 11:19:25 <^7heo> can't even push on gh
2016-09-17 11:19:28 <^7heo> dafuq seriously.
2016-09-17 11:19:56 <^7heo> I have to use tsocks to push
2016-09-17 11:20:00 <^7heo> that is SO sucky.
2016-09-17 11:20:55 <^7heo> jirutka: I sent a PR for the test script.
2016-09-17 11:21:37 <^7heo> jirutka: there's no way to automatically run it for now, since there is no way to pass the passphrase without a TTY
2016-09-17 11:21:48 <^7heo> but I feel like I should implement that.
2016-09-17 11:21:51 <^7heo> at least for testing purposes
2016-09-17 11:23:43 <^7heo> jirutka: don't pull my PR yet
2016-09-17 11:23:49 <^7heo> I have some corrections to make
2016-09-17 11:28:57 <^7heo> <3 tsocks
2016-09-17 11:29:09 <^7heo> now the PR can be pulled.
2016-09-17 11:31:07 <^7heo> jirutka: do you think I should implement a passphrase flag in nlplug-findfs so we can automatically test it?
2016-09-17 11:31:29 <^7heo> (obv. undocumented so people don't start using it)
2016-09-17 11:39:08 <^7heo> or I maybe can use socat
2016-09-17 11:40:07 <jirutka> ^7heo: I dunno, I don’t even know what nlplug-findfs actually do
2016-09-17 11:48:05 <^7heo> jirutka: it's for finding the filesystems to mount at boot time.
2016-09-17 11:48:26 <^7heo> jirutka: it handles raid, cryptfs (LUKS) and lvm
2016-09-17 11:48:30 <^7heo> jirutka: it's very handy.
2016-09-17 11:48:40 <^7heo> jirutka: do you think you can pull my PR?
2016-09-17 11:49:22 <jirutka> ^7heo: I’d perfer to let ncopa review it first
2016-09-17 11:49:41 <^7heo> it's not C, it's shell...
2016-09-17 11:49:44 <^7heo> it's just a test.
2016-09-17 11:49:54 <^7heo> and it's blocking me, because I'd like to base my other branches on that one.
2016-09-17 11:50:28 <^7heo> I mean it's not really blocking me
2016-09-17 11:50:33 <^7heo> but it's bugging me.
2016-09-17 11:50:43 <^7heo> (i.e. the fix for the current version of nlplug-findfs for instance)
2016-09-17 11:50:55 <^7heo> (that I'll push right after)
2016-09-17 11:50:58 <jirutka> ^7heo: that’s not a problem, you can base your branches on top of that commit and send PRs, just add note that that commit should be merged in previous PR
2016-09-17 11:52:11 <^7heo> jirutka: it IS a problem.
2016-09-17 11:52:14 <^7heo> jirutka: it is bugging me.
2016-09-17 11:52:25 <^7heo> jirutka: don't decide what's a problem to me please ;)
2016-09-17 11:52:25 <jirutka> ^7heo: I don’t understand how
2016-09-17 11:52:31 <^7heo> because it's not merged
2016-09-17 11:52:34 <^7heo> and that bugs me.
2016-09-17 11:54:22 <^7heo> it's not any technical problem
2016-09-17 11:54:25 <^7heo> it's more like an OCD
2016-09-17 11:54:52 <jirutka> ^7heo: it’s CDO, it should be sorted alphabetically! XD
2016-09-17 11:55:13 <^7heo> tssk
2016-09-17 11:56:58 <^7heo> jirutka: the reason why it's bugging me is: I modify that test.sh file later on with my other PRs
2016-09-17 11:57:06 <^7heo> jirutka: if anyone asks for a modification before its merged
2016-09-17 11:57:15 <^7heo> jirutka: I'll have a HARD time to merge all that again
2016-09-17 11:57:20 <^7heo> jirutka: and make sure all is correct
2016-09-17 11:57:30 <^7heo> jirutka: right now, I just did, so I'd like it set in stone.
2016-09-17 11:57:34 <^7heo> jirutka: here's why it bugs me.
2016-09-17 11:59:51 <^7heo> So yeah, I rebased my other branch on that one
2016-09-17 11:59:58 <^7heo> but still, I hope that there's no modification requested.
2016-09-17 12:00:20 <^7heo> and my battery is running out
2016-09-17 12:01:17 <^7heo> like, big time.
2016-09-17 12:01:33 <^7heo> and github's javascrhell doesn't help
2016-09-17 12:02:49 <jirutka> ^7heo: I’ve just reviewed it
2016-09-17 12:03:57 <^7heo> Awesome, thanks
2016-09-17 12:04:01 <^7heo> 1% battery left
2016-09-17 12:04:02 <^7heo> let's see
2016-09-17 12:04:14 <jirutka> ^7heo: reviewed, not merged yet ;)
2016-09-17 12:05:08 <^7heo> yeah but still, it's good.
2016-09-17 12:05:16 <^7heo> ok
2016-09-17 12:05:18 <^7heo> going home
2016-09-17 12:14:52 <^7heo> back
2016-09-17 12:30:09 <^7heo> jirutka: pushed the requested changes.
2016-09-17 12:53:35 <^7heo> okay, this is starting to look cleaner and cleaner.
2016-09-17 13:22:45 <^7heo> jirutka: we can now automatically test nlplug-findfs
2016-09-17 13:22:59 <^7heo> jirutka: should I add a .travis.yml file in that commit?
2016-09-17 14:52:54 <^7heo> jirutka: in case you're interested, we can have automated testing for nlplug-findfs now.
2016-09-17 14:53:28 <^7heo> jirutka: I wrote the script and the travis file.
2016-09-17 19:56:50 <t0vig> who -HERE *
2016-09-17 19:58:02 <t0vig> ^Theo:  I think I figured out how to create the cscope package with abuild
2016-09-17 19:58:16 <t0vig> ^7theo: that is....
2016-09-17 19:58:22 <t0vig> ^7heo
2016-09-17 19:58:26 <t0vig> crap
2016-09-17 19:59:06 <t0vig> ^7heo : anyway, I got a build, so I've submitted a patch against testing
2016-09-17 19:59:32 <t0vig> ^7heo: it's impressive how well the apk build system works :)
2016-09-17 20:01:06 <^7heo> t0vig: isn't it?
2016-09-17 20:01:21 <^7heo> t0vig: smart tech by smart people
2016-09-17 20:01:25 <^7heo> ftw
2016-09-17 20:01:28 <^7heo> \o/
2016-09-17 20:01:59 <^7heo> even algitbot agrees
2016-09-17 20:02:32 <t0vig> lol
2016-09-17 20:03:11 <t0vig> and I'm still not using 40MB RAM... while developing... in a VM... remotely via ssh
2016-09-18 00:42:17 <Donno> On a Pentium 2?
2016-09-18 09:17:37 <^7heo> clandmeter: the SSL cert on forum.a.o is expired.
2016-09-18 18:20:32 <coredumb> thx jirutka for merging
2016-09-18 18:20:55 <coredumb> I may have a bunch of them in the following weeks as I'm digging deeper in salt
2016-09-18 19:38:38 <Klowner> ^7heo: since you're fresh on the nlplug-findfs, am I reading it correctly that it attempts to mount root and then falls back to running mdadm or whatever until the desired root device appears?
2016-09-18 19:44:32 <Klowner> just read it yourself, Klowner, don't be lazy
2016-09-18 19:44:49 <nidan_> =)
2016-09-18 19:44:54 <coredumb> lazy is good
2016-09-19 06:38:36 <coredumb> morning
2016-09-19 07:49:19 <jbguerraz> Morning guys!
2016-09-19 07:59:29 <ncopa> morning
2016-09-19 08:07:49 <jbguerraz> morning ncopa !
2016-09-19 09:05:11 <^7heo> ncopa: moin
2016-09-19 09:06:15 <^7heo> ncopa: I have fixed the segfaults in nlplug-findfs, I also have implemented (incomplete) automatic testing, and I have fixed the deported header feature
2016-09-19 09:14:32 <ncopa> ^7heo: nice!
2016-09-19 09:29:33 <^7heo> ncopa: it's on https://github.com/7heo/mkinitfs if you wanna check.
2016-09-19 09:29:57 <^7heo> ncopa: otherwise there's PRs waiting for you anyway: https://github.com/alpinelinux/mkinitfs/pull/1 :)
2016-09-19 09:30:38 <^7heo> ncopa: I started with automated testing because Klowner is waiting for the PR to be in to add his zfs (w/ zpool) integration to the tests
2016-09-19 09:31:46 <ncopa> which branch?
2016-09-19 09:32:14 <^7heo> ncopa: the one I opened a PR with was "tests"
2016-09-19 09:32:25 <^7heo> or "test" maybe
2016-09-19 09:32:45 <ncopa> i see you have various branches
2016-09-19 09:32:49 <^7heo> once this is pulled, I will submit a PR fixing the current state (it currently segfaults)
2016-09-19 09:32:59 <ncopa> apparently for different fixes
2016-09-19 09:33:00 <ncopa> ok
2016-09-19 09:33:08 <ncopa> i will look at it asap
2016-09-19 09:33:08 <^7heo> yeah I made branches, based on each other, so I would just have to PR one after the other.
2016-09-19 09:33:13 <ncopa> need to do some other stuff first
2016-09-19 09:33:17 <^7heo> in the right order
2016-09-19 09:33:18 <^7heo> no worries
2016-09-19 09:33:26 <ncopa> ping me tm if i dont manage to do it today
2016-09-19 09:33:31 <^7heo> the cool part is that now, we have a travis test
2016-09-19 09:33:36 <ncopa> need to openssl fixes
2016-09-19 09:33:38 <^7heo> ok
2016-09-19 09:33:39 <ncopa> and uprgade kernel
2016-09-19 09:33:44 <^7heo> yeah makes sense.
2016-09-19 09:34:10 <^7heo> ncopa: and it's not like I cannot use it myself anyway.
2016-09-19 09:34:16 <ncopa> :)
2016-09-19 09:34:26 <ncopa> i definitively want this is
2016-09-19 09:34:31 <ncopa> this *in*
2016-09-19 09:34:46 <ncopa> i think detached header is a cool feature
2016-09-19 09:34:49 <^7heo> yeah
2016-09-19 09:34:54 <^7heo> and automated testing is a must have
2016-09-19 09:35:01 <ncopa> +1
2016-09-19 09:35:01 <^7heo> even tho it doesn't cover all the code yet.
2016-09-19 09:35:19 <^7heo> so we will have to be careful (but that's the case even with 100% code coverage)
2016-09-19 09:35:52 <^7heo> in the automated testing branch, I had to modify the code of nlplug-findfs to add a tty detection
2016-09-19 09:36:02 <^7heo> so we can pipe the password to it.
2016-09-19 09:36:05 <^7heo> s/word/phrase/
2016-09-19 09:36:54 <^7heo> also I did various fixes suggested by skarnet.
2016-09-19 09:37:28 <skarnet> yeah, I still need to take some time and understand that spawn-manager thing. I'm pretty sure it's entirely unnecessary
2016-09-19 09:37:53 <^7heo> skarnet: it's because we need parallelism for devices to popup independently
2016-09-19 09:38:02 <skarnet> so what
2016-09-19 09:38:13 <^7heo> skarnet: so if the keyboard isn't loaded, we don't have a deadlock situation.
2016-09-19 09:38:41 <^7heo> skarnet: but if you remove the complexity introduced by the threading, I'm all for it.
2016-09-19 09:38:59 <skarnet> I understand the need for parallelism, I don't understand the need for spawnmgr.
2016-09-19 09:40:01 <^7heo> Oh, for the specifics of the implementation, you'd have to ask ncopa :)
2016-09-19 09:41:12 <ncopa> skarnet: iirc, that was to not spawn more threads in parallel than there are cores
2016-09-19 09:41:25 <skarnet> oh come on
2016-09-19 09:41:57 <^7heo> ncopa: are there cases when having more than one thread per code would be a problem to boot?
2016-09-19 09:42:06 <skarnet> do you have profiling results that show this was a problem in the first place?
2016-09-19 09:42:20 <skarnet> how many threads are you typically going to spawn? 2? 4 tops?
2016-09-19 09:43:06 <^7heo> skarnet: it depends on the device numbers I'd say?
2016-09-19 09:43:09 <skarnet> premature optimization is a young man's disease, favored by performance anxiety
2016-09-19 09:43:16 <skarnet> we have good cures against it now
2016-09-19 09:43:17 <^7heo> skarnet: since afaict we have one thread per device
2016-09-19 09:43:30 <ncopa> iirc "spawn thread" includes execute mdev, so it also prevents forking too many children
2016-09-19 09:43:42 <ncopa> i think we execute mdev for each even
2016-09-19 09:44:11 <^7heo> in any case, before we make rehauls like this
2016-09-19 09:44:14 <^7heo> we need better testing
2016-09-19 09:44:17 <^7heo> at least IMHO
2016-09-19 09:44:22 <ncopa> there are like 500? events on boot
2016-09-19 09:44:26 <ncopa> or 200
2016-09-19 09:44:30 <ncopa> dpends a bit on hw
2016-09-19 09:44:41 <ncopa> but i think it was a real problem on rpi
2016-09-19 09:44:47 <^7heo> my current testing just checks if cryptsetup worked as expected
2016-09-19 09:44:52 <ncopa> fabled implemented the spawnmgr
2016-09-19 09:45:00 <^7heo> (but doesn't fail if there's a segfault AFTER it worked)
2016-09-19 09:45:06 <ncopa> so he probably remembers the details
2016-09-19 09:49:13 <ncopa> http://git.alpinelinux.org/cgit/mkinitfs/commit/?id=0e2672eebb4efd725f0e906f07b3c1bae81dbde7
2016-09-19 09:50:08 <skarnet> how does that solve the blocking issue on single-core machines?
2016-09-19 09:54:35 <ncopa> that is a good question actually
2016-09-19 09:54:46 <skarnet> well thanks :P
2016-09-19 09:54:59 <ncopa> i am not sure it is a problem but potensially it is
2016-09-19 09:55:10 <ncopa> fix is trivial though, in case it is a problem
2016-09-19 09:55:24 <ncopa> mgr->max_running = sysconf(_SC_NPROCESSORS_ONLN);
2016-09-19 09:55:38 <ncopa> set it to 2, in case its a single core
2016-09-19 09:55:56 <skarnet> oh? so just like that, you'll break your rule of "don't spawn more threads than there are cores" ?
2016-09-19 09:56:02 <skarnet> then why is it there in the first place
2016-09-19 09:56:16 <skarnet> may as well use 2 threads and be done with it
2016-09-19 09:56:57 <ncopa> and let 46 cores idle on a 48 core arm box?
2016-09-19 09:57:23 <ncopa> mgr->max_running = sysconf(_SC_NPROCESSORS_ONLN) + 1;
2016-09-19 09:57:25 <ncopa> would also work
2016-09-19 09:57:27 <skarnet> uh, yes? this thing will be heavily IO-bound anyway
2016-09-19 09:58:32 <skarnet> if you think 2 is bottlenecked by CPU use, use 4
2016-09-19 09:58:38 <skarnet> but a dynamic queue is overengineered
2016-09-19 09:59:03 <skarnet> and I doubt the limiting factor would be the CPU even with 1 core
2016-09-19 09:59:44 <skarnet> so if you have N threads that may deadlock, spawn N+1 threads, done
2016-09-19 09:59:51 <ncopa> bug fabled with it when he is back
2016-09-19 10:00:17 <skarnet> 
ACTION sharpens his machete

2016-09-19 10:00:28 <ncopa> after all it was he who had the issue of some rpi based media player not booting fast enough
2016-09-19 10:02:05 <skarnet> a RPi can't boot fast enough anyway :P
2016-09-19 14:53:46 <Klowner> I wonder if the zfs package should be putting zpool in /sbin rather than /usr/sbin
2016-09-19 15:04:47 <mkumatag> Hi All, can someone help me understand the effort or the way we can build alpine for different archs like ppc64le. I'm looking for a instructions to build alpine for ppc64le
2016-09-19 15:27:03 <Klowner> ncopa: would you be interested in reviewing some nlplug-findfs additions?
2016-09-19 15:31:42 <ncopa> Klowner: additions?
2016-09-19 15:41:15 <^7heo> ncopa: Klowner did implement zfs/zpool support as I said.
2016-09-19 15:43:21 <Klowner> ya
2016-09-19 15:43:30 <Klowner> https://github.com/alpinelinux/mkinitfs/compare/master...Klowner:zfs
2016-09-19 16:39:41 <Guest1419> This please: Hi All, can someone help me understand the effort or the way we can build alpine for different archs like ppc64le. I'm looking for a instructions to build alpine for ppc64le
2016-09-19 16:41:56 <ncopa> Klowner: i think that looks ok
2016-09-19 16:43:44 <^7heo> ncopa: what about the tests?
2016-09-19 16:45:10 <^7heo> ncopa: jirutka insisted for you to review it, and if something that trivial takes that lomg to go in, I don't think I'll get my work merged this year :p
2016-09-19 16:47:18 <ncopa> :)
2016-09-19 17:06:08 <jbguerraz> does one feel to review https://github.com/alpinelinux/aports/pull/296 ? :)
2016-09-19 17:06:13 <jbguerraz> I'd love to get it into testing
2016-09-19 17:06:52 <jirutka> jbguerraz: uh, I haven’t done it yet… okay, this evening… I hope…
2016-09-19 17:07:06 <jbguerraz> that would be duper cool )
2016-09-19 17:07:14 <jbguerraz> btw I was wondering
2016-09-19 17:07:30 <jbguerraz> as I'm OK to maintain it, should it go to "community" directly ?
2016-09-19 17:07:40 <jbguerraz> so should I update that PR ?
2016-09-19 17:07:48 <jbguerraz> or will it go to community after some time in testing ?
2016-09-19 17:07:50 <jbguerraz> or.. dunno )
2016-09-19 17:07:50 <jirutka> no, new abuilds should go to the testing first
2016-09-19 17:08:14 <jbguerraz> ok great then
2016-09-19 17:11:52 <mkumatag> anyone can please help, Here is my original message - Hi All, can someone help me understand the effort or the way we can build alpine for different archs like ppc64le. I'm looking for a instructions to build alpine for ppc64le
2016-09-19 17:12:24 <jirutka> clandmeter: ^
2016-09-19 17:12:25 <skarnet> we read you fine the first and second times
2016-09-19 17:12:46 <skarnet> if we don't answer, it means we don't think we're qualified to answer and are waiting for someone who is
2016-09-19 17:13:35 <skarnet> btw, has anyone seen fabled? that's more his turf
2016-09-19 17:14:50 <mkumatag> ah, okay
2016-09-19 19:14:55 <jugs> How does one become a package maintainer?
2016-09-19 19:15:38 <jirutka> jugs: creates a package or claims maintainership on some existing package…
2016-09-19 19:17:01 <jugs> How can I claim existing packages?
2016-09-19 19:17:10 <jugs> and for new packages I just commit the APKG?
2016-09-19 19:20:03 <jugs> Also is commit access to testing public or do I need access?
2016-09-19 19:20:36 <jirutka> jugs: send pull request to https://github.com/alpinelinux/aports
2016-09-19 19:21:04 <jirutka> jugs: and this might be helpful https://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package
2016-09-19 19:24:36 <jirutka> jugs: each abuild contains the Maintainer line, like https://github.com/alpinelinux/aports/blob/master/main/tig/APKBUILD#L2; you can pick some abuilds that currently don’t have a maintainer and claim it by writing your name and email to Maintainer:
2016-09-19 19:28:07 <jugs> Yeah I saw the wiki doc, that's key
2016-09-19 19:28:36 <jugs> Thanks for the advice, hopefully I get some PRs in soon
2016-09-19 19:29:30 <gena01> jirutka: thanks for merging the vault PR this morning
2016-09-19 19:32:11 <jirutka> gena01: you’re welcome
2016-09-19 19:32:54 <gena01> jirutka, wasn't sure how much cleanup i should do, was trying to fix the reported issue as well
2016-09-19 20:35:15 <przemoc> jirutka: I see that you've become a full-fledged AL committer.  congrats!  (and thanks for pushing my patches :>)
2016-09-19 20:35:45 <^7heo> przemoc: s/comitter/contributor/
2016-09-19 20:38:36 <przemoc> he was a contributor even before, now I suspect he has rights to commit stuff to aports (or I was too quick in this conclusion and some other AL dev just pushed his PR)
2016-09-19 20:39:52 <coredumb> yes he has
2016-09-19 20:41:27 <^7heo> przemoc: he wasn't a full fledged (or rather "core") contributor
2016-09-19 20:41:43 <^7heo> I suspect that's what you meant
2016-09-19 20:44:16 <przemoc> for me contributor is anyone contributing to the project in any way.  you improved AL's wiki - you're contributor.  you send fixes to ML for aports - you're contributor.  I believe I'm not alone in that kind of thinking.  but project's committer (or developer) is something more than mere contributor, and his responsibility is also bigger.
2016-09-19 20:44:30 <^7heo> ohh seriously
2016-09-19 20:44:35 <^7heo> whatever
2016-09-19 20:46:44 <przemoc> not sure what you're getting irritated at, but it's getting OT, so explain in -offtopic if you're willing to
2016-09-19 20:54:32 <^7heo> I'm getting irritated by you discussing I wasn't
2016-09-19 20:55:26 <^7heo> you say maintainer != contributor, and I was never saying the contrary.
2016-09-19 20:55:51 <^7heo> I said contributor != core contributor
2016-09-19 20:56:24 <^7heo> which is what you were saying (with the wrong terms) in the first place
2016-09-19 20:56:29 <^7heo> EOS
2016-09-19 20:57:00 <^7heo> s/discussing/& something/
2016-09-19 21:02:31 <przemoc> committer is a known term for someone with rights to commit (and push) to the project's repository (BTW being committer isn't the same as being maintainer).  I don't know why you're insisting that I'm using wrong terms, but I agree, no need to dwell on it.
2016-09-20 07:40:48 <jbguerraz> morning guys!
2016-09-20 08:06:04 <^7heo> moin jbguerraz