[03:29] <pitti> Good morning
[05:05] <airdrone> hi, can we do ubuntu touh dev on a desktop running raring with no touch on desktop
[06:46] <jibel> good morning
[06:51] <pitti> bonjour jibel
[07:01] <jibel> bonjour pitti, ça va ?
[07:05] <pitti> jibel: je vais très bien, merci ! et toi ?
[07:05]  * pitti manges beaucoup de fraises
[07:06] <jibel> pitti, ça va bien. Tu as trouvé des fraises en automne ? ;)
[07:06] <pitti> jibel: c'est l'été !
[07:07] <jibel> pitti, haha, when I watch outside, the calendar is lying
[07:07] <pitti> il ne pleut plus aujourd'hui
[07:07] <pitti> seulment hier
[07:08] <jibel> it's like 15°c with rain since last Sunday, happy July
[07:09] <pitti> we had a few days of summer since Saturday, I hope the weekend will be better; we'll go to a wedding
[07:09] <pitti> nice, just installed current saucy grouper build, it works again
[07:09] <jibel> \o/
[07:10] <jibel> pitti, have you seen rbasak's proposal of an lxc driver for adt ?
[07:11] <jibel> https://code.launchpad.net/~racb/ubuntu/saucy/autopkgtest/lxc/+merge/172856
[07:11] <pitti> no, I just saw your G+ message
[07:11] <jibel> I'll make a deeper review, but that's something that should go directly to git
[07:13] <pitti> lxc_container_name = 'adt-virt-lxc-' + ''.join([random.choice(string.ascii_lowercase) for i in range(6)])
[07:13] <pitti> hm, I don't understand that bit
[07:13] <pitti> don't you specify the container name as argument?
[07:14] <pitti> oh, that might be a backing file
[07:14] <pitti> mkstemp!
[07:14]  * pitti follows up in MP
[07:15] <jibel> pitti, I think that's because if you don't specify a name lxc-start-ephemeral generate a random name for the new container, and the only way to grab it is parsing stdout
[08:12] <DanChapman> Good Morning all :-)
[10:11] <rbasak> jibel, pitti: right. lxc-start-ephemeral's stdout is human-formatted. It would be very tricky to parse. And I could check for existing container names to avoid conflicts, but that would race because of the gap between the check and the call to lxc-start-ephemeral anyway. I could file a bug against lxc to make it machine parseable, and add an autopkgtest task so that it's fixed when lxc supports it if you like?
[10:12] <pitti> rbasak: sounds good; the race with that isn't very bad, I just wonder if one coudl quickly stat /var/lib/lxc/whatever/$name for that
[10:12] <rbasak> pitti: re: copyright, it's Canonical copyright, since I wrote it on company time as part of my work for the Canonical server team. But I can bump the copyright date I suppose. It started off as a copy of adt-virt-schroot anyway.
[10:12] <rbasak> pitti: so just wrap it in a while exists /var/lib/lxc/...: regenerate loop?
[10:12] <pitti> rbasak: ok; I don't mind much the copyright, it just stuck out (2007 is a while ago..)
[10:13] <pitti> and  having author names is always good for asking someone about stuff
[10:13] <rbasak> Ack. I'll update the MP.
[10:13] <pitti> rbasak: some test -d presumably, yes
[10:13] <pitti> rbasak: but this is all nitpicking, of course
[10:13] <rbasak> Sure. I'd rather get it right.
[10:13] <rbasak> I appreciate the review.
[10:14] <rbasak> No comment on the quality of the rest of the code base :-/
[10:14] <rbasak> Global variables everywhere, etc. :-(
[10:14] <rbasak> pep8? What's pep8? :-/
[10:23] <rbasak> pitti: how I about I use tempfile.mktemp? http://docs.python.org/3/library/tempfile.html#tempfile.mktemp
[10:24] <rbasak> pitti: it has a massive warning about a race between choosing a name and creation, but we have that anyway
[10:24] <pitti> rbasak: no, I didn't try it yet or get into details
[10:24] <pitti> rbasak: pep8> autopkgtest code is an absolute horror; no amount of pep8ification will rectify that
[10:24] <rbasak> Indeed
[10:24] <pitti> rbasak: I believe that's deprecated
[10:24] <rbasak> What's deprecated? pep8 or autopkgtest?
[10:25] <pitti> tempfile.mktemp()
[10:25] <pitti> rbasak: but with a check whether the machine exists, you coudl as well just try autopkgtest001, autopkgtest002 etc. until you find a free one :)
[10:26] <pitti> rbasak: but anyway, bikeshedding; I think your approach is quite okay, it just looks funny at first sight
[10:26] <rbasak> pitti: mktemp() will do that for me, since every machine that exists has an entry in /var/lib/lxc/
[10:27] <rbasak> Using mktemp() would admittedly reduce the size of the race
[10:27] <pitti> rbasak: again, mktemp() is deprecated since python2.3, it will cease to exist at some point
[10:27] <rbasak> I should have at least left a comment explaining the race
[10:58] <rbasak> pitti: MP updated. I'm finding a looping random name generation until I find one that's not taken. I've added a comment about a race, and comments at the top adding copyright and authorship.
[11:15] <DanChapman> pitti, when launching rhytmbox with autopilot i'm getting this output http://pastebin.ubuntu.com/5843179/ is this an autopilot bug?
[11:16] <DanChapman> pitti, vis works fine so I can introspect, just the memory leak part concerned me
[11:16] <pitti> DanChapman: I don't get that here -- does that only happen when you run it from autopilot?
[11:17] <pitti> actually, I do get it, RB just takes ages to start
[11:17] <DanChapman> pitti yeah, its slow starting
[11:17] <pitti> some saucy regression
[11:17] <pitti> DanChapman: anyway, I get the same message when not being under autopilot, so you can ignore it for now
[11:18] <DanChapman> pitti, ok thanks for that :-)
[11:53] <jibel> rbasak, you could also generate a unique name with uuid.uuid4(). performance is close to the random/range approach
[11:55] <rbasak> jibel: I have a dislike of UUIDs. They're useful for some things, but a pain when dealing with by hand.
[11:56] <jibel> rbasak, since it's for generating a temporary testing env, does it matter?
[11:56] <rbasak> Here, the string doesn't need to be universally unique. Just unique in /var/lib/lxc, so it doesn't have to be that long
[11:56] <rbasak> jibel: it matters when developing and testing the driver, since I have to examine/clean up containers by hand when doing that. And high development velocity always helps.
[11:57] <jibel> rbasak, okay :)
[11:57] <rbasak> eg. mktemp doesn't use UUIDs either :-)
[11:58] <pitti> indeed that seems to simplify things
[11:58] <pitti> no need for a polling loop or this fun loop construction
[11:59] <rbasak> The fun loop construction is pretty much exactly what mktemp does. So I think there's precedent here :-)
[12:04] <DanChapman> didn't mktemp get replaced with mkstemp? or was it mkdtemp? well it was one of them two anyway
[12:04] <rbasak> mktemp(3) did, yes. mktemp(1) is still there and not deprecated, AFAIK.
[12:28] <smartboyhw> BACK
[12:28] <DanChapman> hey smartboyhw
[13:22] <smartboyhw> DanChapman, SergioMeneses hmm I got two mouse pointers using Mir.
[13:23] <DanChapman> smartboyhw, really? thats wild. Ive not seen that
[13:23] <smartboyhw> DanChapman, that's on notebook
[13:24] <smartboyhw> When I connect to a mouse, both pointers move
[13:24] <smartboyhw> :O eh?
[13:24]  * smartboyhw will report a bug soon
[13:24] <DanChapman> strange!
[13:25] <DanChapman> smartboyhw, are they located in the same place or dofferent areas of the screen? just out of curiosity :-)
[13:25] <DanChapman> *different
[13:30] <smartboyhw> DanChapman, same
[13:30] <smartboyhw> Actually, no
[13:30] <smartboyhw> One doesn't clik things
[13:30] <smartboyhw> :P
[14:05] <smartboyhw> DanChapman, the problem seems like happening occasionall
[14:05] <smartboyhw> *occasionally
[14:05] <smartboyhw> Sometimes it appears, sometimes it doesn't
[14:10] <smartboyhw> balloons, do you know of that problem?
[14:30] <smartboyhw> Guys, seriously, no comments on the pilot system?
[14:36] <DanChapman> smartboyhw, sorry had to pop to shop. Ran out of coffee!!!! Has anyone filed a similar bug to your mouse prob?
[14:36] <smartboyhw> DanChapman, dunno
[15:05] <SergioMeneses> smartboyhw, jeje nice!
[15:05] <smartboyhw> SergioMeneses, which one are you meaning?
[15:06] <smartboyhw> chilicuil, have you looked at the email yet?
[15:10] <SergioMeneses> smartboyhw, about the mouse! jeje
[15:10] <SergioMeneses> smartboyhw, what email?
[15:10] <smartboyhw> SergioMeneses, the email about having testcase pillots
[15:10] <smartboyhw> SergioMeneses, that;s not good
[15:11] <SergioMeneses> smartboyhw, =/
[15:12] <smartboyhw> SergioMeneses, did you get the mail?
[15:14] <chilicuil> smartboyhw: not yet, I'm gonna look at it right now
[15:14] <smartboyhw> chilicuil, oh great:)
[15:14] <smartboyhw> balloons is OK with the idea
[15:14] <chilicuil> smartboyhw: any hint for what I'm looking for? =P
[15:14] <smartboyhw> chilicuil, what do you mean?
[15:15] <chilicuil> smartboyhw: the email about having testcase pillots ?
[15:15] <smartboyhw> chilicuil, well that's what I plan to have. The pilots should be testcase admins or Canonical QA people
[15:15] <smartboyhw> i.e. you will be one:P
[15:17] <SergioMeneses> smartboyhw, yes
[15:17] <smartboyhw> SergioMeneses, phew:)
[15:17] <chilicuil> smartboyhw: oh, yep I read it yesterday, I think pilots should be another group, I for instance don't have any clue about autopilot =P
[15:17] <smartboyhw> chilicuil, well, maybe we make it autopilot pilots and manual pilots.
[15:18] <smartboyhw> IDK
[15:18] <smartboyhw> It's just sheer copying from #ubuntu-devel (sorry devs)
[15:19] <SergioMeneses> smartboyhw, that would be awesome
[15:19] <SergioMeneses> btw I still have some testcase bugs :'(
[15:20] <smartboyhw> lol
[15:23] <chilicuil> smartboyhw: what are necessary two pilots groups?, AFAIK autopilot is all about automating testcases? is not enough autopilot pilots?, why in the first place we need them?, is not working as it's now?
[15:23] <chilicuil> why*
[15:31]  * SergioMeneses has the same question 
[15:39] <smartboyhw> chilicuil, try to ask in the mailing lists:P
[15:44] <smartboyhw> chilicuil, SergioMeneses well it's rather like a system for people to standby if new contributors want to work on new things:P
[16:27] <slickymaster> elfy: hi, elfy. As you didn't show up here, I've sent you a PM in the forums. Hope it's ok
[16:28] <elfy> slickymaster: just got in from work - the first thing I did was drink tea - the second was reply to you :)
[16:29] <slickymaster> elfy: the reason was because yesterday you mentioned that, like me, you are also new to bzr, and I found this mini-tutorial about it
[16:30] <elfy> yep
[16:30] <slickymaster> Most probably you're already aware of it, but just to to be safe I thought of sending the link
[16:30] <slickymaster> elfy: http://doc.bazaar.canonical.com/latest/en/mini-tutorial/
[16:31] <slickymaster> elfy: I see you already read it :)
[16:31] <elfy> :)
[16:35] <slickymaster> elfy: reagrding the settings test, I've done a tiny bit of work on it, last night, but now as a read your answer a doubt pop-up in my mind
[16:35] <slickymaster> elfy: I was not supposed to push the changes I've made?
[16:35] <elfy> not yet :)
[16:35] <slickymaster> was I not, sorry
[16:35] <elfy> slickymaster: you should see a window called elfy in your irc client
[16:37] <slickymaster> elfy: I'm not on XChat, I'm on my browser, still at work
[16:38] <elfy> yea - have a look across the top of it
[16:38] <elfy> maybe a tab or something
[16:38] <slickymaster> elfy: :) you're right. there she is
[17:07] <DanChapman> Has anyone got the time to help me setup an autopkgtest for evince using my autopilot test?
[17:38] <DanChapman> pitti, hey what does this mean? /home/dan/auto-package-testing/bin/../etc/config: distro-info: not found
[17:39] <pitti> DanChapman: a-p-t uses that to determine the default release if you don't specify one
[17:39] <pitti> DanChapman: install the distro-info package for it, or specify one with -r
[17:43] <DanChapman> pitti, great thanks.
[20:11] <phillw> Hi, I'm just having a 'blonde' moment... what is the name of the system that speeds up boot time which gets triggered each time something like a new kernel is installed? I'm not 100% sure if it is upstart.
[20:21] <Patrickdk> bootchart?
[20:22] <Patrickdk> oh, you want to speed up
[20:22] <Patrickdk> well, that is not always used, but ureadahead
[20:27] <phillw> Patrickdk: thanks, it is ureadahead ! ... I just assume it to be there now (I was there when it 1st was introduced) and had forgotten the little critter's name :)
[20:28] <Patrickdk> hmm, back in 9.04 I think