[08:01] <dholbach> good morning
[08:05] <dholbach> does anyone else have the problem that the global menu is hardly clickable on the nexus7 with current raring? and a corrupted background during installation?
[09:47] <iceroot> hi
[09:47] <iceroot> are there any plans to support the raspbian pi in the future?
[09:47] <iceroot> or will there be no support for Armv6
[10:21] <lilstevie> iceroot: no support for armv6
[10:21] <iceroot> lilstevie: ok
[10:23] <ogra_> see topic :)
[10:32] <infinity> iceroot: Regressing our platform support for the sake of exactly one board (no matter how cool and hip it is) wouldn't make much sense.
[10:33] <infinity> iceroot: But raspbian has worked out most of their kinks, and it's even the Pi's recommended OS these days, so that should do.
[10:34] <ogra_> dholbach, yo ... morning, when did we have the nx7 meeting again (time) ? i promised alex to run it today (we werent sure when you come back and were to lazy to dig through the calendar)
[10:41] <iceroot> infinity: thank you for that info, i am not very familiar with armv6 but as it seems its something "old" which does not have a future (like non-pae systems on x86)
[10:41] <infinity> iceroot: Yeah, pretty much.  armv7 (for 32-bit) and armv8 (for 64-bit) are the future.
[10:41] <iceroot> but what other arm devices are common for GNU/Linux instead of raspberry pi and nexus 7?
[10:41] <infinity> iceroot: And, frankly, armv7 is the present too, not just the future.
[10:41] <ogra_> it surely has a future, just not in the desktop or smartphone market
[10:42] <ogra_> i guess v6 will still be used for a long time in industrial automation etc
[10:42] <infinity> iceroot: There are a ton of fun armv7 devices out there.  Pandaboards, Chromebooks, i.MX dev boards, etc, etc.
[10:42] <ogra_> nexus7 :)
[10:43] <infinity> iceroot: But the Pi certainly has its place for educational use and other fun stuff.  It's just not something Ubuntu's targetting.  And Debian fills that gap nicely, so we don't have to.
[10:43] <infinity> (In fact, Debian fills the gap twice... Officially with Debian/armel and unofficially with Raspbian)
[10:43] <iceroot> infinity: debian is not supporting raspberry pi too
[10:43] <infinity> See above. :)
[10:44] <infinity> People who claim Debian doesn't support it are people who assume you MUST run hard-float for that tiny performance boost.
[10:44] <infinity> Debian/armel will work great on it.
[10:44] <iceroot> infinity: the official debian arm version does not include a special firmware, so something on the cpu is not worrking, because of that there is the hacked kernel on raspbian (imo)
[10:44] <iceroot> infinity: no debian is not working great on the pi
[10:44] <infinity> The userspace would.
[10:44] <infinity> I can't speak to kernels.
[10:45] <infinity> Anyhow, somewhat out of scope for this channel.
[10:45]  * ogra_ would prefer raspbian anyway, since its an armhf recompile 
[10:45] <iceroot> but i am fine with raspbian, i just need a apt-get distro :) but ubuntu would be nice too, specially the education part should be a big target for ubuntu but that is just my opinion
[10:45] <infinity> Honestly, I wouldn't recommend a Pi to anyone who wants to do serious ARM development for desktop/mobile/tablet type stuff.  It's far too slow.
[10:45] <ogra_> ++
[10:46] <lilstevie> and far too little ram
[10:46] <iceroot> infinity: xbmc is my target and that is fine
[10:46] <infinity> lilstevie: The too little RAM is a huge contributing factor to the too slow, yes. :/
[10:46] <ogra_> its good for toying around but surely not suited to run a desktop centric distro like ubuntu
[10:46] <infinity> I'm kinda hoping they switch to a v7 SoC in a future Pi edition and we can stop having this discussion.
[10:46] <ogra_> ++
[10:47] <iceroot> ogra_: but it has a nice price specially for countries which are not that rich like the one from europe
[10:47] <lilstevie> do broadcom even make a v7 soc
[10:47] <ogra_> sure, but you only get what you pay for
[10:47] <infinity> lilstevie: Probably.  Never checked.
[10:47] <iceroot> ogra_: then i would get nothing when using ubuntu :)
[10:47] <ogra_> cant make a porsche out of a beetle
[10:47] <infinity> lilstevie: The list of A8 and A9 licensees is too long to memorize. :P
[10:48] <lilstevie> I mean I have a tv box that uses the same soc as the rpi and it is fine with that, but that is an os that is tuned to limited resources
[10:48] <ogra_> iceroot, be sure someone pays a lot for ubuntu, just not you ;)
[10:48] <infinity> ogra_: Pfft, just add a whale tail to your beetle and you're done.
[10:48] <ogra_> lol
[10:49] <lilstevie> can't imagine running a desktop class os on it though
[10:49] <infinity> The alternate counterpoint would be "You obviously never owned a Golf GT".
[10:49] <ogra_> heh, no, i never owned a VW in my life
[10:50] <dholbach> ogra_, 16 utc I think
[10:50] <ogra_> dholbach, thx
[10:50] <dholbach> ogra_, not sure if you saw my question earlier:
[10:50] <dholbach> does anyone else have the problem that the global menu is hardly clickable on the nexus7 with current raring? and a corrupted background during installation?
[10:50] <ogra_> oh, kind of overread it
[10:50] <iceroot> but thank you all for the usefull answers.
[10:51] <infinity> iceroot: Good luck, and happy hacking.
[10:51] <ogra_> dholbach, the background is known, not sure xnox works on a fix already
[10:51] <ogra_> it happens since we switched the installer to compiz
[10:51] <infinity> iceroot: You may find that some software has a whole lot of assumptions that "Hey, you're on ARM, that means you're v7 with thumb2 and neon and, and, and..."
[10:51] <ogra_> my global menu is fine apart from the times where it isnt ... due to the xinput bug
[10:51] <infinity> iceroot: (At build time that is)
[10:52] <xnox> currently not working to troubleshoot / fix that corruption.
[10:52] <infinity> iceroot: If you run into stuff like that, bug reports and patches welcome to both Debian and Ubuntu to fix broken upstreams. :P
[10:52] <dholbach> ogra_, I just opened firefox and tried to click on "Help" and then on "About" or something
[10:52] <ogra_> oh, i see what you mean
[10:52] <ogra_> the selection vanishes after the first menu item if you try to scroll down
[10:53] <ogra_> thats definitely a new one
[14:13] <dholbach> ogra_, mentioned your announcement in a couple of places :)
[14:15] <ogra_> thx !
[14:15] <smartboyhw> what annoucement?;P
[14:16] <dholbach> smartboyhw, we'll have the nexus7 meetings again
[14:17] <smartboyhw> dholbach, ah. Got it in email
[14:26] <wookey> anyone else noticed that netbase in raring doesn't install under sbuild because it tries to replace /etc/services and barfs?
[14:28] <ogra_> shiny
[15:07] <infinity> wookey: That's not netbase's fault, it's schroot's.
[15:07] <infinity> wookey: Comment out services and protocols from /etc/schroot/default/nssdatabases
[15:08] <infinity> wookey: It's (incorrectly, IMO) copying in those files from the base system, then netbase wants to install them (since it, y'know, owns them), and boom.
[15:08] <wookey> aha. that's what's going on. cheers.
[15:08] <wookey> is there  abug reported on that? it's breaking a load of my builds
[15:09] <infinity> Probably not.  I should just fix it in schroot upstream and let it trickle down.
[15:09] <wookey> please do.
[15:09] <infinity> I've been a bad man and been fixing it locally instead.
[15:09] <wookey> yeah, I've founda  few of those I did in quantal and fogot to file...
[15:10] <wookey> come and bite me in the bum a second time
[15:16] <hrw> infinity: can I overuse your SRU powers?
[15:17] <infinity> hrw: If it's for the alsa-* stuff, I'll get to it after I've slept.
[15:17] <hrw> infinity: thank you very much
[16:03] <dholbach> ogra_, meeting time?
[16:03] <dholbach> ogra_, in a different call myself, so can't join
[16:03] <ogra_> oh, right
[16:41] <wookey> doko_: crosbbuild-essential-arm64 does not bring in pkgbinarymangler so clean sbuild builds end up with mismatch between build arch changelog and host arch changelog, so build packages won't install
[16:41] <wookey> did you leave that out on purpose for some reason?
[16:42] <wookey> I put it in because otherwise this happens.
[16:47] <wookey> maybe I sent you a version with that missing? I which case apologies
[16:48] <infinity> wookey: I removed it on purpose.
[16:48] <wookey> Ah yes. I see the changelog comment. Yes it's not essential but without it nothing works
[16:49] <infinity> wookey: Build your chroots with mk-sbuild.
[16:49] <infinity> wookey: And "nothing works" is wrong.  If you build BOTH your arches in the same sort of chroot, they both work together fine.
[16:50] <wookey> OK, but in practice at least some stuff is going to come from the real archive
[16:50] <wookey> and that all been mangled
[16:50] <infinity> wookey: All your build environments should match.  But there's no reason everyone else's need to match launchpad's.  Cause people building custom packages should change version numbers.
[16:50] <infinity> wookey: Your use case (using cross chroots to build archive versions) is, actually, a really tiny corner case.  Port bootstraps don't happen THAT often.
[16:51] <wookey> OK. so why doesn't sbuild-chreat-chroot put it in?
[16:51] <infinity> For normal people, there's nothing build-essential about the weird things Ubuntu buildds do.
[16:51] <infinity> Nobody uses sbuild-create-chroot.
[16:51] <wookey> I had assumed that this was required in ubuntu world. I didn;t put it in the debian version, of course
[16:51] <infinity> Even the Debian sbuild wiki docs tend to recommend mk-sbuild. :P
[16:51] <infinity> (Which is from ubuntu-dev-tools)
[16:52] <wookey> Hmm. I use sbuld-createchroot, seemed to make sense for doing sbuild builds
[16:52] <infinity> mk-sbuild is much better maintained.
[16:52] <wookey> we have far too many of these....
[16:52] <wookey> each with different set of useful features
[16:53] <infinity> I'm tempted to drop sbuild-create-chroot from sbuild entirely, adopt mk-sbuild again, and Replaces: ubuntu-dev-tools.
[16:53] <infinity> But that's a conversation for another day.
[16:53] <wookey> sbuild should have a set of matching tools, that work for debian and ubuntu
[16:53] <infinity> Anyhow.  The take-home message is "Ubuntu buildd chroots are weird, we do weird things, that doesn't make our weird stuff build-essential, it just makes it essential if you want builds identical to ours".
[16:53] <wookey> But I guess it's fair enough for ubuntu builds to require a tool from ubuntu-dev.
[16:54] <wookey> Yes. OK, point taken
[16:54] <wookey> And does avoid annoying diff between debian and ubuntu verions of build-essential
[16:54] <wookey> I shall go an revise a load of instructions
[16:55] <wookey> I guess a lintian check on crossbuilds for something aorund this might help avoid too much pain
[16:55] <infinity> Seriously, just tell people to use mk-sbuild.
[16:55] <wookey> that's what I mean by 'revise instructions'
[16:56] <infinity> https://wiki.ubuntu.com/CrossBuilding
[16:56] <wookey> But it's annoying if it's different for ubuntu
[16:56] <infinity> I use mk-sbuild for both Ubuntu and Debian.
[16:56] <infinity> And ubuntu-dev-tools is in Debian.
[16:57] <wookey> yes, but its kind of daft if you need to install ubuntu dev tools to build a standard debian sbuild chroot
[16:57] <infinity> Despite the name, it's got a lot of handy Debian tools. :P
[16:57] <infinity> Nah.
[16:57] <infinity> It's contentious for people who think the word Ubuntu is evil, but they can suck it up.
[16:57] <infinity> You're installing something from the Debian archive and running a script.
[16:58] <wookey> I don;t think it's evil, just rather illogical that the sbuild chroot creation tool is not the one in sbuild, but some other one from a package named as if it ought not to be relevant
[16:58] <infinity> (But, like I said, I'm going to talk to rleigh about replacing sbuild-create-chroot with mk-sbuild and shipping it in sbuild)
[16:58] <wookey> OK. I'll be perfectly happy if the good bits of those two are combined
[16:59] <wookey> That's the right answer
[17:00] <wookey> you are doing an excellent job of answering my questions today :-)
[17:04] <wookey> infinity: btw did you see my nice new build output giving actualy reasons for failure on the summary page: http://people.linaro.org/~wookey/buildd/raring-arm64/status-bootstrap.html#summary
[17:04] <wookey> major improvement in being able to see what still needs fixing
[17:05] <infinity> "something went wrong"... Very informative. :)
[17:06] <wookey> yes, room for improvement there, but grepping the line out of a build that is actually relevent is tricky
[17:06] <wookey> it used to say 'dpkg-buildpackage died' which was about equally informative
[17:07] <wookey> there is now nice structure for adding smarter regexes for typical build failures
[17:09] <infinity> Anyhow.  I should find a bed and attack it with some vigor.
[17:11] <janimo> ogra_, ah nice. I may have misunderstood what marvin24 said a few days ago. I thought ac100 images were gone and that is why he had to test the 3.8 kernel with netboot