[00:05] <desrt> RAOF: at the end of the day, if i don't agree with your power policy i can always for (;;) ;
[00:06] <RAOF> desrt: But that won't spin up > 45W worth of GPU :P
[00:06] <desrt> also: unless you use proper access control at the kernel level, i can always ship my own nvidia driver blob and hit up the device node for myself
[00:07] <RAOF> Well, you can try. :)
[00:08] <RAOF> I think we *do* have proper access control, actually.
=== duflu_ is now known as duflu
[05:59] <mlankhorst> god morning
[06:45] <pitti> Good morning
[06:51]  * olli waves at pitti
[06:54] <pitti> hey olli
[07:31] <didrocks> morning
[07:37] <pitti> bonjour didrocks
[07:38] <pitti> didrocks: btw, if/when you plan to work on the enabled(preset) or similar things, you know src/test/test-install.c? with those unit tests this is much easier and faster to reproduce/iterate/fix
[07:41] <didrocks> pitti: oh nice, I'll work with that. Thanks to the pointer :)
[07:41] <pitti> didrocks: and it's rather important to have tests for this behaviour, too :)
[07:42] <didrocks> pitti: It would be good to have some other feedbacks anyway
[07:42] <didrocks> yeah :)
[08:00] <didrocks> pitti: so, for alternatives, I was thinking about using systemd alias rather than handling ourselves the symlinks
[08:00] <didrocks> that removes a little bit of internal systemd knowledge that we put in postinst
[08:01] <pitti> didrocks: not sure what you mean?
[08:02] <didrocks> pitti: here is an example for lightdm: http://bazaar.launchpad.net/~didrocks/lightdm/systemd-alias/revision/2099
[08:02] <pitti> didrocks: oh, for the display-manager; I thought for wants/ :)
[08:03] <didrocks> ah no, for the thing I was thinking yesterday evening :)
[08:03] <didrocks> so, how to treat alternatives, basically
[08:03] <pitti> didrocks: oh, I like that!
[08:03] <didrocks> the benefits is that it's compatible with existing ones, so we don't need to migrate everything in one shot :)
[08:04] <didrocks> if you +1, I'll propose patches to debian (and maybe upload right away in ubuntu?)
[08:04] <pitti> didrocks: but please note that e. g. systemctl reload won't work under upstart/sysv
[08:05] <didrocks> pitti: that's the thing I need to test, will it error out?
[08:05] <didrocks> (need to start a vm to test)
[08:05] <pitti> fairly sure
[08:05] <didrocks> let me have a try
[08:05] <didrocks> if so 2>/dev/null || true
[08:06] <pitti> didrocks: is-enabled/enable etc. also work without systemd, but this one might not -- so it at least needs to be guarded
[08:06] <pitti> didrocks: [ -d /run/systemd/system ]
[08:07] <didrocks> pitti: you prefer to guard that only call? like [ -d /run/systemd/system ] && systemctl daemon-reload
[08:08] <didrocks> yeah, it's failing
[08:11] <pitti> didrocks: I'm not 100% sure about is-enabled; if that works without systemd, use it; otherwise we need to check the file system directly as previously
[08:11] <didrocks> pitti: it does, just tested
[08:15] <didrocks> pitti: ok, so that would look like this in the end: http://bazaar.launchpad.net/~didrocks/lightdm/systemd-alias/view/head:/debian/lightdm.postinst#L72
[08:17] <pitti> didrocks: why do you need the daemon-reload at this point? you didn't change anything yet?
[08:18] <didrocks> pitti: we just dropped a new .service file from the package installation, no?
[08:18] <pitti> didrocks: oh, for this one
[08:18] <didrocks> and on upgrade, we added the alias
[08:18] <didrocks> yeah
[08:19] <didrocks> I was hoping that we would have some option like --force-unless-mask
[08:19] <pitti> didrocks: so please run this by Josselin Mouette
[08:19] <didrocks> but seems a little bit overkill :)
[08:19] <didrocks> ok, doing
[08:20] <pitti> didrocks: I worked with him on the sprint on that general scheme, and has a great understanding of that
[08:20] <pitti> didrocks: and we should hten have the same approach in gdm
[08:20] <didrocks> right :)
[08:20] <pitti> didrocks: so to be clear, this is no functional change, just simpler/fewer file system assumptions, right?
[08:20] <pitti> (on the way -- possibly -- to reducing symlinks in etc)
[08:21] <didrocks> pitti: exactly, it's the same mechanism as of today and the results in term of symlinks are the same
[08:22] <pitti> purrfect
[08:22] <didrocks> but we don't rely on internals
[08:22] <didrocks> so, hoping that if we can convince on clearer /etc… :)
[08:22] <pitti> even if not, using the official is-enabled/enable API is better anyway
[08:23] <didrocks> yeah
[08:38]  * willcooke -> sick.  Ping if you need anything, but I'll be mostly not here
[08:39] <didrocks> urgh, good luck willcooke :/
[08:39] <willcooke> :)
[09:02] <Laney> yo
[09:02] <didrocks> hey Laney
[09:03] <LocutusOfBorg1> morning desktoppers
[09:03] <didrocks> morning LocutusOfBorg1
[09:03] <Laney> what's up
[09:37] <pitti> hey Laney
=== vrruiz_ is now known as rvr
[09:58] <pitti> didrocks: FYI, offline for ~ 2 h to reinstall my laptop; just got a replacement for the dying hard drive
[09:59] <didrocks> oh ok! ttyl
[10:05] <Laney> "I can reproduce this", such sweet words when said by an upstream ;-)
[10:20] <didrocks> Laney: heh, which ones in that case? :)
[10:21] <Laney> libv4l
[10:22] <didrocks> ok ;)
[10:22] <Laney> it's a crash in some asm inside libjpeg
[10:22] <Laney> yeah ...
[10:24] <Laney> http://techcrunch.com/2014/11/19/mozilla-partners-with-yahoo-which-will-become-the-default-search-engine-in-firefox-next-month/
[10:37] <mlankhorst> so we will use microsoft? :P
[10:39] <chrisccoulson> mlankhorst, short answer - no ;)
[10:40] <larsu> long answer?
[10:40] <larsu> (which I presume is the more interesting one)
[10:40] <chrisccoulson> larsu, the same :)
[10:41] <larsu> haha :)
[10:42] <larsu> short answer: no. Long answer: nein
[10:45] <chrisccoulson> lol
[10:45]  * didrocks remembers the buzz when we talked about switching to Yahoo on the netbook edition in 2010
[10:45] <didrocks> and all those hard words against us
[10:46] <didrocks> I think Mozilla won't have the same treatment "because"
[10:47] <ogra_> well, they are at least clever enough to do everything at once ... their advertisement tiles one month, the default search engine next
[10:48] <ogra_> we always waited til one issue settled to fire up the next :)
[10:48] <didrocks> indeed :)
[10:49] <larsu> didrocks: they're just changing providers, whereas it looked like we wanted to extract more money and are greedy
[10:52] <didrocks> larsu: well, they are doing that as their deal is closing, so they want to extract some money, which is understandable :)
[10:56] <larsu> didrocks: of course, but it doesn't seem like that to the public
[10:56] <larsu> at least that's the impression i get
[10:56] <larsu> or they just like mozilla better :P
[10:58] <ogra_> nah, the public just freaks more out over scientist shirts than over apps ... the fashion moved on ;)
[10:59] <larsu> that freak-out was justified though
[10:59] <ogra_> oh come on ... he wore a present his GF made for him on his big day ...
[11:00] <larsu> you mean its better that he and his girlfriend feel good than all the people he works with?
[11:01] <larsu> how about this: "my girlfriend once gave me a 'the people that make ubuntu are idiots' tshirt and I wore it to UDS - I don't know why the people there were offended"
[11:02] <ogra_> well, seeing how petty the web got over the last ten years i slowly start not feeling bad that i'm a german anymore
[11:02] <ogra_> he never claimed he didnt know why people are offended
[11:02] <ogra_> and he aplologized ...
[11:03] <ogra_> what i find horrible is that thw world got so humorless that he had to
[11:08] <didrocks> it's really true (and sad) that now, everything is a base of controverse
[11:08] <larsu> ogra_: ya, they certainly overreacted - just saying the sentiment was justified
[11:08] <larsu> didrocks: that's not true
[11:08] <didrocks> soon, we'll have everyone suited the same, with the same hair cut, in a very white room
[11:08] <ogra_> overreaction might be *the* word describing this decade :)
[11:09] <didrocks> larsu: I'm not talking about that particular case, but seeing the number of meaningless "buzz" we are seeing…
[11:09] <didrocks> buzz on $random polemics
[11:09] <ogra_> and i'm scared where that might go ... socially ...
[11:09] <didrocks> ogra_: completely agree
[11:10] <didrocks> the president has his tie knot never done right -> first world problem
[11:11] <didrocks> (and this really happened for months in France)
[11:13] <larsu> didrocks: obviously this is a different problem and I totally agree there
[11:13] <larsu> the press was talking about whether the 1st lady is allowed to have a tatoo in Germany for a while
[11:13] <larsu> ridiculous.
[11:14] <ogra_> well, that just the press being the press ...
[11:14] <ogra_> *that's
[11:15] <didrocks> the press wasn't like that before…
[11:15] <didrocks> or not that much
[11:15] <didrocks> and it's because the society asks for this
[11:16] <ogra_> its a two edged thing ... it is to generate more money due to more sensationalism ... the society gets more numb ... you need more sensationalism ... etc etc
=== ara is now known as Guest1529
=== m_conley_away is now known as m_conley
[17:18] <achiang> didrocks: maybe no one cares about switching to yahoo because everyone who does care already switched to chrome ;)
[17:18] <didrocks> achiang: ahah, I'm sure this is not untrue :)
[17:20] <achiang> btw, i understand there is still some bug in unity/compiz 14.04.1 where resuming on multi-monitor setup results in lost window placements...
[17:20] <achiang> is that stuck in the sru queue somewhere?
[17:23] <didrocks> hum, I didn't hear about it, bregma? ^
[17:23] <didrocks> or Trevinho? ^
[17:24] <Trevinho> achiang: never heard of that...
[17:24] <Trevinho> achiang: get a crash also?
[17:24] <Laney> I have similar
[17:24] <achiang> no crash, it's just unity/compiz losing its mind
[17:24] <Trevinho> weird... and i *only*suspend... :o
[17:24] <bregma> achiang, I haven't heard of a problem on simple resume
[17:24] <Laney> if I turn one of my monitors off then it gets disabled as far as the system is concerned
[17:24] <achiang> bregma: it's one of my worst papercuts :(
[17:24] <Laney> this messes up the window placement
[17:25] <Laney> i.e. they all get put on one monitor
[17:25] <bregma> oh, unless it's the timing-related problem where a monitor does not come back online right away and Compiz thinks it's been removed
[17:25] <Trevinho> ah, well... that's because one monitor gets turned off, and thus it's what we have to do..
[17:25] <Trevinho> it might be the case where you unplugged it
[17:25] <bregma> Trevinho, it shouldn't move the windows to a different workspace
[17:26] <achiang> in my case, chrome on laptop display (maximized) before suspend, will resume maximized on external monitor
[17:26] <Trevinho> mh ChrisTownsend did some work on that
[17:26] <bregma> which is what it does for me when I lose my monitor (damn that loose connection)
[17:26] <achiang> i also have workspaces turned on
[17:26] <achiang> and this is over vga, not anything fancy like displayport or hdmi
[17:27] <bregma> either way, there's no queued fix for that at the moemnt
[17:27] <achiang> how can i collect debug information?
[17:27] <achiang> to file a bug?
[17:29] <bregma> "ubuntu-bug compiz" should do the trick
[17:29] <achiang> ok
[17:30] <achiang> also, is there a way to restart a unity session from the command line?
[17:30] <achiang> i have another hard to reproduce on demand (but happens about 2x a week) where resume from suspend results in compiz and parts of unity (like unity-panel-service) going crazy
[17:31] <achiang> and no new windows can be opened
[17:31] <achiang> the only way i can recover is to reboot the entire machine
[17:31] <achiang> would be nice to just restart unity if possible
[17:35] <bregma> achiang, 'setsid unity >/dev/null 2>&1 &' should work
[17:36] <achiang> bregma: awesome, thanks
[17:36] <achiang> bregma: i can do that from VT1, and it'll restart on VT7?
[17:36] <bregma> you'd want to set DISPLAY=:0.0
[17:38] <bregma> ideally you could just type 'restart unity7' from the command-line, but doing it from a VT means needing to set your DBUS_SESSION and UPSTART_SESSION_ID and that can get complicated
[17:38] <Trevinho> achiang: doing that from VT1 you'd also need dbus...
[17:39] <achiang> hm
[17:39] <Trevinho> ouch bregma preceded me :P
[17:39] <achiang> i mean, my goal in this scenario is to preserve my session, like all my open editor windows, etc.
[17:39] <Trevinho> achiang: but, if you've a terminal open in your session, once yuou load it you can restart from there
[17:39] <bregma> I guess I should write a quick script just for this purpose, except it would be tricky on a multi-seat system
[17:40] <achiang> Trevinho: ah, so if i have a terminal open in the session, i can just type that "setsid" command without worrying about anything else?
[17:41] <Trevinho> achiang: if you have a terminal, and you get a compiz crash... then you can go to vt1, launch "unity", and then go to vt7.... There launch "unity" again from the terminal
[17:41] <Trevinho> in this way it will load all the proper dbus and upstart env vars
[17:41] <Trevinho> (and many others :))
[17:42] <bregma>  eval "$(tr '\000' '\012' </proc/$(pidof compiz)/environ | grep DBUS_SESSION) $(tr '\000' '\012' </proc/$(pidof compiz)/environ | grep UPSTART_SESSION) restart unity7"
[17:42] <bregma> that should work
[17:42]  * bregma wishes he could get more obscure in public
[17:42] <achiang> Trevinho: what if unity is still mostly running, but can't do interesting things like open new windows?
[17:42] <achiang> bregma: ok, run that from an open terminal in existing unity session, or run from VT1?
[17:43] <Trevinho> achiang: mh, that's a weird case
[17:43] <bregma> achiang, either one should work
[17:43] <achiang> Trevinho: well, i hit it about 1.5x per week
[17:43] <achiang> bregma: thx
[17:45] <achiang> seems exacerbated by running spotify desktop client + hipchat desktop client
[17:45]  * bregma wonders if rewriting that command in Perl could obfiuscate it more
=== alan_g is now known as alan_g|EOD
[18:02]  * Laney goes away
[18:02] <Laney> ttyl!
[18:05] <didrocks> see you guys
[18:21] <humbletoad> can someone help me with ssh? im having a hell of a time
[18:21] <humbletoad> im trying to delete my keys from my server, but it isn't letting me log into the thing now
[18:24] <mitya57> humbletoad: this channel is for development, user support is on #ubuntu
=== m_conley is now known as m_conley_away