**** BEGIN LOGGING AT Mon Mar 04 02:59:59 2013 Mar 04 04:07:34 how do i get my nexus 7 back to normal? Mar 04 04:18:58 is there any plan to put strace 4.7 to raring Mar 04 09:26:13 hi, can I find a list of supported platforms for Ubuntu's ARM work? Mar 04 09:26:36 I'm wondering if it would work, with graphics acceleration, on an OMAP 4470 platform Mar 04 09:27:50 it definitely does on 4460, not sure about 4470 Mar 04 09:28:04 the pandaboard was for years our reference platform Mar 04 09:29:25 * infinity has never even seen a 4470... Mar 04 09:30:14 Looks like the big difference is a PVR SGX544 instead of a SGX540. Mar 04 09:30:39 So, entirely possible that the PVR binary blobs in precise wouldn't know about the new revision. Mar 04 09:33:16 yeah Mar 04 09:33:47 and since TI put OMAP to death for mobile we likely wont get any updated drivers anymore Mar 04 09:39:21 infinity, ogra_: I'm asking specifically because I'm looking for a cheap but usable platform to run ubuntu on, and I came across the Archos 101 XS Mar 04 09:40:04 which seems like a nice match, except for the low RAM. I was only worrying about the SGX544 Mar 04 09:40:24 sspiff: Well, I honestly can't say one way or the other. The last SoC *I* tested on was a 4460, but someone may have played with a 4470. Mar 04 09:40:32 do the binary drivers for Linux and Android differ greatly? Mar 04 09:40:34 well, if you would be fine moving to the ubuntu phablet edition (which HW wide depends on a minimal android layer) you should be fine with that HW Mar 04 09:40:42 sspiff: They're completely different, yes. Mar 04 09:41:05 if you expect to fully natively run ubuntu on it (xorg, ubuntu kernel etc) i would go with a nexus device Mar 04 09:41:05 infinity: hence the ubuntu phablet architecture leveraging the Android platform Mar 04 09:41:20 right Mar 04 09:41:32 ogra_: I considered that, but they're considerable more expensive and no decent good keyboard docks Mar 04 09:41:39 sspiff: The phablet/Android madness is more of a stopgap than an end goal. :) Mar 04 09:41:49 But it works for enablemnet on devices where we'll never see native drivers. Mar 04 09:41:53 For some value of "works". Mar 04 09:41:59 well, the android layer will stay Mar 04 09:42:16 it wil; shrink and bits will move into the distro, but it wont go away Mar 04 09:42:24 Erm. Define "stay". Mar 04 09:42:38 If we get native drivers and have no need for bionic, we have no need for said layer. Mar 04 09:42:52 infinity, the higher layer is designed to use libhybris and friend and to rely on HAL Mar 04 09:43:04 See above. Mar 04 09:43:18 Nothing depends on hybris, except that some things depend on bionic, thus we need hybris. Mar 04 09:43:24 well, i doubt we want to go back to "we only support one device" Mar 04 09:43:33 Kill the bits that need bionic (binary blobs), and we don't need any of it. Mar 04 09:43:34 while we already support over 30 after a week Mar 04 09:43:39 with the new image Mar 04 09:43:46 The option for the Android layer is always there. Mar 04 09:43:54 The end goal for a branded device isn't that, though. Mar 04 09:43:55 Just sayin'. Mar 04 09:44:05 android makes porting extremely easy ... Mar 04 09:44:21 * ogra_ produced that http://people.canonical.com/~ogra/phablet/i9100/ withing a few hours yesterday Mar 04 09:44:27 *within Mar 04 09:44:41 and i have *zero* android experience Mar 04 09:45:17 i guess the layer might go away for devices we will preinstall on, but not for the unwashed masses Mar 04 09:45:27 I'm not sure I call it "porting", when you're just sucking in device support that's already there. Mar 04 09:45:48 (if thats even possible with that design, the userspace apps make a lot of use of android bits) Mar 04 09:46:07 Erm, they do? Mar 04 09:46:10 true, its not actual porting :) Mar 04 09:46:11 The system is glibc-based. Mar 04 09:46:27 but needs the underlying API to talk to the HW Mar 04 09:46:42 That goes right back to the binary blobs thing. Mar 04 09:46:45 But yeah. Mar 04 09:46:51 we dont have alsa, there is vold, no graphics stack etc Mar 04 09:46:52 We'll see how it pans out. Mar 04 09:47:17 You're right that for mass porting to unsupported devices, we'll likely keep the hybris/android solution indefinitely. Mar 04 09:47:29 i.e. my image above has no wifi because samsung decided to have a dalvik based tool to bring up the driver Mar 04 09:47:33 You know, until we replace Android as the #1 phone/tablet OS. Mar 04 09:47:35 In 2023. Mar 04 09:47:48 so i end up without wlan0 on the device and there is no way from userspace to activate it Mar 04 09:48:07 i will need to hack something together on the android side for it Mar 04 09:49:52 the longer that "interim" situation persists the harder it will get to get rid of the android side, community is actively contributing bits that will closer tie us into that Mar 04 09:50:46 infinity: I understand the fact that it's a stopgap and that it's needed for easy device compatibility Mar 04 09:52:31 regardless, I'm not in the market for having to create a device branch for an ubuntu phablet, I'm looking for a minimal effort to get an ultramobile notebook-tablet hybrid experience Mar 04 09:52:55 if I'm going to have to hack support for the device, I know I'll never get around to it, to many hobby projects queuing in the backlog :) Mar 04 09:53:19 well, we currently support the nexus7 natively Mar 04 09:54:02 if the 4470 isn't going to work, a nexus solution would give me an inferior keyboard experience and with a decent dock etc the price of a nexus 10 becomes high enough to even consider Intel Core i3 solutions, which have good open source driver support Mar 04 09:54:11 and the panddaboard for desktop installs, the the future here is blurry .... i would ratrher expect that to die due to lack of drivers Mar 04 09:54:28 nexus7 is really to small for my needs, I'm looking for 10" Mar 04 09:54:49 well, we can onlye easily get our hands onto the tegra drivers Mar 04 09:54:50 yeah I assume the panda will go the way of the dodo soon, it's been removed from the AOSP as well Mar 04 09:55:14 ogra_: open source tegra drivers? or binary tegra drivers for Linux for a specific distro? Mar 04 09:55:16 so if you want Xorg, tegra HW is your best bet Mar 04 09:55:25 binary indeed Mar 04 09:55:35 I thought tegra/nvidia were completely unsupportive? Mar 04 09:55:42 GLES isnt well suppported in the open ones Mar 04 09:55:51 uhu, I really need GLES :P Mar 04 09:56:19 it might get there (if you belive nvidias marketing) Mar 04 09:56:39 but surely isnt yet Mar 04 09:56:49 I do some GLES coding, and the main goal of the device I'm looking for is a tablet testing environment, and a way to do some light coding/compiling/... on the train/plane/somewhere else where my 15" notebook isn't an option Mar 04 09:57:06 experience learns me to never expect support post-launch Mar 04 09:57:23 SoC obsolete so fast that you really want on-launch support Mar 04 09:57:37 well, i know for sure that nvidia works on an update for the drriver for the next Xorg ABI bump Mar 04 09:58:01 so these devices should be safe for a little while still (if you find an unlockable one like the nexus line) Mar 04 09:58:08 uhu Mar 04 09:58:29 the Transformer TF300 would be a Tegra3 based option, but not sure if it's unlocked Mar 04 09:58:33 * ogra_ needs to give the cat her insuline shot, brb Mar 04 09:58:46 huh, I didn't know they did that for cats Mar 04 10:05:51 re Mar 04 10:06:18 yeah, they luckily do ... Mar 04 10:45:36 infinity, one for you :) https://plus.google.com/112266164281670850856/posts/RuPVvyrPBtU Mar 04 13:27:47 So, can anyone tell me how ubuntu makes their rootfs.img? Mar 04 13:28:31 we use live-build with some slight changes that are hidden in the livecd-rootfs source package Mar 04 14:35:47 ogra_: do you know if it is possible to make a rootfs.img from a nexus 7 tablet? basically I'm trying to take a functional modified tablet and copy it to another. Mar 04 14:36:38 zou can either just grab the nexus7 one and modify the tarball inside or start from scratch using an ubuntu-core tarball Mar 04 14:39:08 to be honest, I have a fully customized device now that I want to copy, I don't want to start over Mar 04 14:39:46 Zero_Chaos: nandroid backup in recovery should be enough in fact, if you can/want use that Mar 04 14:40:40 tassadar_: is that a custom recovery or the stock android one? forgive me I'm new ;-) Mar 04 14:41:31 custom recovery, it basically does .tar.gz from whole /data partition Mar 04 14:42:32 tassadar_: anything that does *everything*? I have a slightly custom android that has a custom ubuntu chroot in /data/local/ubuntu so I really wanted to grab all of rootfs Mar 04 14:43:11 well, what do you consider "all of rootfs"? Mar 04 14:43:34 if you have a custom chroot already, just apt-get what you want Mar 04 14:43:38 tassadar_: / Mar 04 14:43:50 tassadar_: minus the obvious /dev /proc and /sys Mar 04 14:44:01 uh Mar 04 14:44:14 that is "rootfs" afterall Mar 04 14:44:17 data partition is / for ubuntu Mar 04 14:44:34 tassadar_: it's android with an ubuntu chroot right now Mar 04 14:44:54 but I know the full ubuntu install has a real nice rootfs that is easy to flash, just didn't know if I could make one from a running device easily Mar 04 14:45:03 looks like the answer is "not easily" Mar 04 14:45:07 yeah Mar 04 14:46:08 damn you, school wifi network -.- Mar 04 14:47:36 school sucks Mar 04 14:49:56 yeah, and I have to go now, sorry, bye Mar 04 14:51:36 tassadar_: thanks for your help Mar 04 15:02:53 Could someone please tell me if this look likes cosmic rays: https://launchpadlibrarian.net/132280468/buildlog_ubuntu-raring-armhf.octave_3.6.4-1_FAILEDTOBUILD.txt.gz Mar 04 15:03:20 I don't think I'd blame cosmic rays no matter what that log shows Mar 04 15:03:27 so no, it's not cosmic rays Mar 04 15:04:15 micahg, give it back ... if it doesnt fail the same way its cosmic rays :) Mar 04 15:04:55 ogra_: please don't tell people that, most of them are dumb enough to believe you... Mar 04 15:05:06 ?? Mar 04 15:05:15 ogra_: it's not cosmic rays Mar 04 15:05:23 ogra_: thinking that is just embarassing Mar 04 15:05:50 ogra_: we aren't in space, we don't get enough random space radiation to have such issues down here Mar 04 15:05:58 well, it could be humidity in the datacenter ... or a fly sitting on a chip and shortening two contacts Mar 04 15:06:02 ogra_: hardware failure on the other hand is a very common issue. Mar 04 15:06:15 Zero_Chaos: being that the build machines are in a unique case, we refer to cosmic rays as random failures to due HW env Mar 04 15:06:18 ogra_: it could be aliens with mind control, but hardware failure is a bit more common don't you think? Mar 04 15:06:56 it could be cosmic ray induced HW failure ;) Mar 04 15:07:00 micahg: if you want to use an insane case to speak about a sane one that is your choice, but I'm still going to watch and assume you are stupid. Mar 04 15:07:15 * micahg is really feeling the love in here Mar 04 15:07:22 * ogra_ hugs micahg Mar 04 15:07:30 * micahg hugs ogra_ back Mar 04 15:07:44 * micahg should probably just break down and create a cross-build chroot Mar 04 15:07:52 micahg: I know a lot of people that think "cosmic rays" means "cosmic rays". using stupid terms like that causes more misunderstanding in an already misunderstood field. Mar 04 15:08:02 that wont get you far on the buildd though Mar 04 15:08:35 Zero_Chaos, its a commonly used term in here ... and please stop calling people stupid in ubuntu channels Mar 04 15:08:40 ogra_: well, it should tell me if there's an inherent failure in the build, it certainly won't catch all cases though Mar 04 15:08:51 yeah Mar 04 15:09:12 do you know if thats a virtualized or a devirtualized builder ? Mar 04 15:09:22 ogra_: using the cross libc, it wouldn't tell me if there were an assembler failure though, right? Mar 04 15:09:22 could well be a qemu issue Mar 04 15:09:28 ogra_: I didn't call anyone stupid, I said I would sit back and assume he was because he is using a term that deliberately breed misunderstanding. Mar 04 15:09:30 ogra_: devirt (archive) Mar 04 15:09:31 yeah Mar 04 15:09:43 oh, ok Mar 04 15:10:17 * micahg will give back then as he doesn't have the proper hardware to do a test (and the buildds are relatively free) Mar 04 16:50:36 Zero_Chaos: in fact, computers on earth *are* affected by cosmic rays. in early days of one of the supercomputers i worked on, we had to install a large tank of water above the server room to absorb them, to prevent RAM corruption issues caused by random bit flipping due to cosmic rays Mar 04 16:50:58 Zero_Chaos: later on, we improved the shielding on the DIMMs Mar 04 16:52:17 Zero_Chaos: and of course, due to wave/particle duality at the quantum levels, i'd say that it is quite reasonable to use "ray" and "particle" interchangably in this context Mar 04 17:25:31 achiang: is this 1962? Mar 04 17:26:03 Zero_Chaos: is that relevant? Mar 04 17:26:56 achiang: yes, because "computer's are affect by cosmic rays" is very different from "computer's were affected by cosmic rays 30 years ago before they were properly shielded" Mar 04 17:27:24 Zero_Chaos: this was 2002. Mar 04 17:28:06 Zero_Chaos: so if you could stop calling people stupid, that would be great. thanks. Mar 04 17:28:18 achiang: during one of the longest solar minimums in history? I'm... shocked may be the word Mar 04 17:29:19 achiang: and again, picking the single least likely scenario is not good practice, and it perpetuates confusion in the industry, and that is stupid, which is what I said. Mar 04 17:30:27 can we drop thatr off-topic now ? Mar 04 17:30:46 achiang: when gcc compile fails over and over, in a different place, I don't say it's cosmic rays, I say it's ram. And you know what, wrapping the computer in 10 ft of tin foil can't fix what new ram can Mar 04 17:32:14 there is no computer to wrap ... its all dev boards without any casing Mar 04 17:32:32 but pretty please stop this topic now Mar 04 17:35:02 ogra_: it was my intent to say that using the term "cosmic rays" for seemingly random failures is confusing for new users and should be avoided. I'm sorry this point confused so many people in here. Mar 04 17:56:15 https://lists.ubuntu.com/archives/ubuntu-devel/2013-March/036776.html Mar 04 17:58:15 interesting Mar 04 19:05:33 * micahg shows Zero_Chaos the retry build has made it past where it was: https://launchpad.net/ubuntu/+source/octave/3.6.4-1/+build/4325207, hooray for Cosmic Rays :) Mar 04 19:06:06 micahg: yeah, random hardware failure does the same thing all the time. especially heat and RAM issues Mar 04 19:06:33 micahg: but hey, you are welcome to think whatever you like, I realized long ago that if you don't listen I shouldn't care either Mar 04 19:07:04 if you want people to listen to you, it doesn't hury to listen first ;) Mar 04 19:07:21 s/hury/hurt/ Mar 04 19:08:06 micahg: I did listen. you think cosmic rays messed up your build. I think it's just as likely that a ghost did it. Mar 04 19:08:55 [09:06] Zero_Chaos: being that the build machines are in a unique case, we refer to cosmic rays as random failures to due HW env Mar 04 19:09:58 micahg: yeah I read that, and I was trying to say that people reading that assume you mean real cosmic rays not using it as a euphemism for "something failed and I don't know what" Mar 04 19:10:11 micahg: thus, causing entropy and stupidity in the general populace Mar 04 19:10:40 micahg: and since like three people have "stood up to me" in the last hour to tell me how cosmic ray can actually affect building, I stand by my point. **** ENDING LOGGING AT Tue Mar 05 02:59:58 2013