**** BEGIN LOGGING AT Thu Nov 08 03:00:02 2012 Nov 08 03:28:08 mjrosenb: Erm, that looks like a local libgdbm, rather than one from the archive... Nov 08 03:30:35 infinity: it was installed with dpkg Nov 08 03:30:44 is there a way to find out where it is getting the .deb from? Nov 08 03:30:56 iirc, I have the omap overlay installed and nothing else? Nov 08 03:32:17 mjrosenb: 'apt-cache policy libgdbm1' might give you a hint. Nov 08 03:32:32 and I have no clue where that information is stored, since the only things I have in sources.list are the ubuntu repos Nov 08 03:33:02 mjrosenb: Given that libgdbm1 doesn't actually exist in the Ubuntu archive, it's definitely not from us (nor is the thing that was linking to it) Nov 08 03:33:15 this is gbm not gdbm. Nov 08 03:33:27 Oh, I can't read. :P Nov 08 03:33:50 http://pastebin.mozilla.org/1923601 Nov 08 03:33:54 looks legit to me. Nov 08 03:34:38 and http://pastebin.mozilla.org/1923602 for unity-2d-panel, which has a dep on libgbm.so.1 Nov 08 03:35:55 mjrosenb: Right, that package came from the PPA, though. Nov 08 03:36:19 oh. Nov 08 03:36:31 so it did Nov 08 03:36:38 * mjrosenb has no clue what he's looking for. Nov 08 03:36:39 http://paste.ubuntu.com/1341643/ <-- The archive package. Nov 08 03:36:59 perhaps downgrading to 8.0.4-0ubuntu0.2 would fix the problem? Nov 08 03:37:00 Version 8.0.4-0ubuntu0.2 is from Ubuntu. Nov 08 03:37:34 well, I made a symlink, if that ever stops working, I'll downgrade. Nov 08 03:37:37 ndec: The mesa in your PPA for precise is busted. Nov 08 07:43:32 good morning Nov 08 11:39:09 janimo, for your ambient sensor, probably taking a look at https://android.googlesource.com/device/asus/grouper/+/7fce7d56cb077e869b9509cd2770d92e8cf29dcc/liblights/lights.c would be helpful :) Nov 08 11:39:24 ogra-nx7, yes I saw that Nov 08 11:39:31 ah, k Nov 08 11:40:04 but first I wanted to see if it works at all on android, to get some dmesg or logger prints Nov 08 11:40:14 and it does not seem to work at all Nov 08 11:40:15 i wonder if we should switch to the interactive power governor by default Nov 08 11:40:37 seeing all these fine tuned bits in the android init scripts it feels like that would save us some power Nov 08 11:40:45 I may need to take a look some android ambient light apps from google play, maybe some extra twiddling that is not in default is required Nov 08 11:41:31 well, theoretically you shouldnt need any userspace for it apart from some app to adjust it probably Nov 08 11:41:32 ogra_, no idea. Probably someone should try the same power tests (or others) victor did with various governors and then decide based on the numbers Nov 08 11:41:44 yep Nov 08 11:42:21 well, i bet victor would just do them, he already has a setup, but we would need to give him a kernel with ondemand disabled or help him to make sure ondemand doesnt get forced Nov 08 11:42:56 ogra_ - what's the correct way to make sure the N7 is up-to-date? Nov 08 11:43:09 update-manager :) Nov 08 11:43:19 we push updates to the PPA Nov 08 11:43:38 just dont enable -security or -updates, they will trash your desktop Nov 08 11:43:47 ogra_ - that's what i was getting at. which ppa? Nov 08 11:44:19 the ubuntu-nexus7 one, enabled in /etc/apt/sources.list.d ;) Nov 08 11:56:09 fwiw kinteractiveup spams dmesg a lot (interactive governor) Nov 08 11:58:03 well, i guess you can quieten that Nov 08 11:59:17 victorp, showing EULAs isnt really acceptable for a proper ubuntu image Nov 08 11:59:30 victorp, i would prefer a proper solution rather Nov 08 12:12:34 ogra_, what is a proper solution? Nov 08 12:12:47 ogra_ and why is it not acceptable? Nov 08 12:12:56 victorp, licenses that dont require the user to click anything Nov 08 12:13:15 that is exactly what we do for the business desktop image available to download from Ubuntu.com Nov 08 12:13:32 ogra_, is not a license Nov 08 12:13:46 if we really want to go down that rathole you need a) EULA support in usb-creator and b) all packages that ship any of the firmware will need it too (you can install an image, but yu could as well also install the single packages) Nov 08 12:13:48 it is just a note letting users know what is the ROM Nov 08 12:14:13 * victorp srugs Nov 08 12:14:27 if the license then needs to be accepted *before* downloading that gets all kinds of complicated Nov 08 12:14:55 sinbce we have no mechanism for that at all apart from doing something as ugly as the flash downloader package Nov 08 12:15:17 (which cant work for i.e. network driver frimware) Nov 08 12:15:21 ogra_, I think you are assuming that this is an EULA and that people have to click on it Nov 08 12:15:36 we can't pass the license to the users Nov 08 12:15:42 victorp, well, its some kind of popup people will have to agree to, no ? Nov 08 12:15:43 if says in the original license Nov 08 12:15:44 ... Nov 08 12:16:05 nope, it is just a notice saying this are the licenses in the rom Nov 08 12:16:17 you are not asked to agree to gplv3 every time you download ubuntu Nov 08 12:16:30 still, if the license requires us to show it somehow all of the above applies Nov 08 12:16:48 we need to show a notice, yes Nov 08 12:16:52 if the license doesnt require us to show it, lets just not show it :) Nov 08 12:17:09 does is not what our legal team suggest Nov 08 12:17:10 :) Nov 08 12:17:12 right, so it needs to be shown by the packages as well as by the installer Nov 08 12:17:18 ogra_ I have a solution!!! Nov 08 12:17:27 we stick to the installer that we have now ;) Nov 08 12:17:47 victorp, what does that solve Nov 08 12:17:51 ogra_ the packages that contain the firmware Nov 08 12:17:57 ogra_ j/k Nov 08 12:18:03 you still need EULA popups in all the packages Nov 08 12:18:20 can you define all Nov 08 12:18:24 *all*? Nov 08 12:18:27 which is massively messy and adds a lot of mainbtenance overhead Nov 08 12:18:43 every package that ships any of the firmwares Nov 08 12:18:57 right, why dont you put them all in 1 package? Nov 08 12:19:11 thats ugly but would work Nov 08 12:19:23 or dont put them in a package Nov 08 12:19:25 at all Nov 08 12:19:39 how would you ship them properly then ? Nov 08 12:19:44 meaning that we could add the to the images, and put the notice in the installer Nov 08 12:19:54 or apply any updates in case there is newer firmware that fixes issues Nov 08 12:20:05 but if someone wants to use a package per package they can get them from the google site Nov 08 12:20:34 mmm Nov 08 12:20:36 generally we dont put any non packaged files into images Nov 08 12:20:46 ok - put them in 1 package then Nov 08 12:20:57 there are exceptions but it would be good to not go with that Nov 08 12:20:58 I think that is ugly from one sense, cleaner from another Nov 08 12:21:21 1 package + installers with notice? Nov 08 12:21:22 well, it wont help any other devices with the same HW ... thats the issue Nov 08 12:21:25 WFM Nov 08 12:21:34 i.e. the bcm4330 is a very common wlan chip Nov 08 12:21:42 not really ubuntu :) Nov 08 12:21:52 but yeah, would work for the device Nov 08 12:21:59 ogra_ you cant use those blobs in other devices Nov 08 12:22:05 sure i can Nov 08 12:22:08 it is against the license of the blobs Nov 08 12:22:15 oh no you cant :) Nov 08 12:22:30 i can use the bcm4330 blob for every bcm4330 welan card out there that works with the same kernel driver Nov 08 12:22:53 you can, and you can also get a call from the blob owners lawyers :) Nov 08 12:22:55 wether thats legal by that specific license is another story indeed Nov 08 12:23:03 talking about technical bnits here Nov 08 12:23:15 well.. now a days you can separate it Nov 08 12:23:19 what I mean is Nov 08 12:23:28 the proper way to do what we want would be to contact the manufacturers and get proper licenses so ubuntu benefits as a whole Nov 08 12:23:34 do one package called nexus7-blobs-from-other-people Nov 08 12:23:38 that wfm Nov 08 12:23:49 and if is not useful for re-using in other devices Nov 08 12:24:02 even better as the license does not allow you to do that anyway Nov 08 12:24:04 :) Nov 08 12:24:22 well, if the license says you cant use it on other devicews we cant make it available in the archive Nov 08 12:24:23 ogra_, If you give a list of the manufacturers , I can do that Nov 08 12:24:39 i can for sure install nexus7-blobs-from-other-people on my ac100 Nov 08 12:24:43 ogra_ why? Nov 08 12:24:45 so i would break it Nov 08 12:25:02 yes Nov 08 12:25:10 but that is not how it works Nov 08 12:25:17 the list should be at https://developers.google.com/android/nexus/drivers#grouper Nov 08 12:25:25 this is how ubuntu works Nov 08 12:43:03 we have an ancient version of smem in quantal Nov 08 12:55:34 achiang, in raring I added the latest one Nov 08 12:55:50 janimo: ah great! Nov 08 12:55:52 achiang, we won't do much quantal profiling I guess? Nov 08 12:55:55 no Nov 08 12:56:00 janimo: wonder if it should go into debian though Nov 08 12:56:32 achiang, well 1.2 I got an hour after filing a bug in Debian :) The new bugfix I still have to upstream Nov 08 12:56:49 janimo: awesome, you are a step ahead of me Nov 08 12:56:51 thanks Nov 08 12:57:12 achiang, well I discovered smem while at the memory profiling session, remembering that top & co simply lie Nov 08 12:57:22 https://wiki.ubuntu.com/Nexus7/MeasuringMemoryUsage Nov 08 12:57:23 and added it to etherpad and then went on testing it Nov 08 12:57:35 i think we should write a tutorial Nov 08 12:57:41 or not even that, just usage Nov 08 12:58:03 well I am still in the figuring out phase, i.e don;t know much more than what the package manpage says Nov 08 12:58:26 no worries, not asking you to do it Nov 08 12:58:30 i'm just playing with it too Nov 08 12:58:34 what I want to do though is have a setup where we smemcap locally (this is missing in quantal) which is quick and then analyse on a more powerful machine Nov 08 12:58:52 it may use a setup similar to how the bootchart dailies are to be set up Nov 08 13:38:01 * ogra_ thinks that the nexus display even when setting the brightness to 1 is way to bright Nov 08 13:42:29 ogra_: I too find it quite amazing that these devices that consume a total of 5W of power can emit 50W worth of light from their panels Nov 08 13:42:51 yeah, you can easily use the nexus as a torch at night Nov 08 13:42:56 even with brightness set to 1 Nov 08 13:42:57 heh Nov 08 14:49:13 oh sigh Nov 08 14:49:22 * ogra_ just looked at the plymouth issues Nov 08 14:49:51 the bootloadert sets console=none ... we override that to console=tty1 in our boot options ... Nov 08 14:50:19 plymouth sadly picks the first console= option from the cmdline and tries to connect to /dev/none Nov 08 14:56:31 yay, chromebook Nov 08 15:01:34 ogra_: me too :) Nov 08 15:02:22 ogra_: /win 9 Nov 08 15:07:38 ogra_: any inside news on when they'll be generally available? I'm burning to order one, just to use as a "nicer" pandaboard :) Nov 08 15:08:10 highvoltage, i just bought mine at amazon.co.uk Nov 08 15:09:08 nice, and they still have! Nov 08 15:09:41 mfisch: im thinkin mark this as wishlist, right? https://bugs.launchpad.net/ubuntu-nexus7/+bug/1076199 Nov 08 15:09:42 Launchpad bug 1076199 in ubuntu-nexus7 "Feature request for Dual Boot" [Undecided,New] Nov 08 15:10:03 cwayne, wontfix rather :) Nov 08 15:10:17 ogra_: lol, im gonna go with opinion / wishlist Nov 08 15:10:41 if anyone from the community wants to figure it out and write a howto on the wiki thats fine, but we wont work on such a feature for the official images Nov 08 15:10:55 right Nov 08 15:11:01 ill comment that and mark opinion Nov 08 15:11:08 ++ Nov 08 15:14:13 ogra_: eek, it costs more than $100 extra to order it from amazon.co.uk, I guess I'll wait until they have stock back in the US afterall then Nov 08 15:14:35 cwayne: +1 Nov 08 15:14:53 mfisch: also just marked the gnome-shell bug wontfix with a comment explaining why Nov 08 15:15:22 I can't wait to see the response Nov 08 15:23:30 im excited Nov 08 15:23:38 i wrote it as nicely as possible though :) Nov 08 15:37:47 mfisch: hmm, what about this guy? https://bugs.launchpad.net/ubuntu-nexus7/+bug/1075565 I think that the screen dimming is normal when no clicks are being registered Nov 08 15:37:48 Launchpad bug 1075565 in ubuntu-nexus7 "Auto dimming on brightness with spikes" [Undecided,New] Nov 08 15:41:25 cwayne, m,y screen flashes quite a lot if the device is attached to power and in DPMS suspend mode Nov 08 15:41:42 i assume at the same frequency that victorp sees Nov 08 15:42:14 ogra_: his wasnt suspended, or plugged into power Nov 08 15:42:37 my thinking is that he just happens to be logging the brightness after a click, when it is un-dimmed Nov 08 15:42:42 oh, right, he tested battery time :) Nov 08 15:42:52 cwayne, the spikes where on battery yes Nov 08 15:43:01 I was a sleep Nov 08 15:43:13 victorp: right, i think what you're seeing is the normal auto-dim to save power Nov 08 15:43:15 so I didnt see it,but got logged on the test results Nov 08 15:43:26 cwayne, ugh Nov 08 15:43:34 not normal that is resets every hour Nov 08 15:43:54 http://victorpalau.files.wordpress.com/2012/11/chart_11.png Nov 08 15:43:57 see grey line Nov 08 15:43:58 victorp: i think it may just be getting that data after a click (when the brightness is brought back up) Nov 08 15:44:20 I dont think so Nov 08 15:44:32 I run it again later with 49 Nov 08 15:44:35 sorry Nov 08 15:44:36 40 Nov 08 15:44:49 40 minutes between logging? Nov 08 15:44:51 and it did not spike because is below the level of auto dim Nov 08 15:44:58 oh 40% Nov 08 15:45:02 40 % Nov 08 15:45:12 victorp: was 'dim screen to save power' on when you did it at 99? Nov 08 15:45:27 yes Nov 08 15:45:41 and it immediately went down to 99 Nov 08 15:45:46 from 99 Nov 08 15:45:48 to 76 Nov 08 15:45:54 victorp, btw, i would like to prepare some power mgmt settings for you stealing bits from the android defaults so we can check if they improve the system in any way beyond using ubuntu defaults Nov 08 15:46:08 (after i got images up) Nov 08 15:46:27 ogra_, cool I can do that Nov 08 15:47:01 * ogra_ grumbles about plymouth Nov 08 15:47:06 silly software ! Nov 08 15:47:50 intrestingly i cant make the kernel crash anymore by removing all the override files Nov 08 15:48:11 but i cant get a splash either because plymouth conncts to a non existing /dev/none Nov 08 15:48:12 :P Nov 08 15:50:19 i wonder what happens if i create a udev rule to link /dev/none to tty1 Nov 08 16:03:35 infinity, can you please remove the nexus7 kernel from raring NEW? I just found it does not build in the PPA. thanks Nov 08 16:03:53 I agains assumed that if it builds in quantal it should in raring Nov 08 16:05:10 which si weird as gcc does not seem to be that different. (4.7.2-2 vs 4.7.2-5) Nov 08 16:07:53 janimo, python2 vs 3 ? Nov 08 16:08:09 Makefile:509: The path 'python' is not executable. Nov 08 16:08:14 from your build log Nov 08 16:08:19 ogra_, it fails while compiling a c file though Nov 08 16:08:24 hmm Nov 08 16:08:45 ogra_, same in the previous build, but that warning is harmless (python) Nov 08 16:08:50 https://launchpadlibrarian.net/122192174/buildlog_ubuntu-quantal-armhf.linux-nexus7_3.1.10-7.11_BUILDING.txt.gz Nov 08 16:08:53 well, yeah, it should just skip that bit gracefully Nov 08 16:09:06 maybe some extra Werror checks were added Nov 08 16:09:30 you could just set do_tools = false for now Nov 08 16:11:54 victorp: hey, how do i install that extension with your changes? do i just drop that extension dir somewhere? Nov 08 16:15:27 cwayne, branch into the nexus 7 Nov 08 16:15:33 then open chromium Nov 08 16:15:39 go to extensions Nov 08 16:15:47 set it to developer mode Nov 08 16:16:04 that will give you an option to do "unpack extension" Nov 08 16:16:55 victorp: ah, great, thank you Nov 08 16:17:08 click on that and will let pick a folder Nov 08 16:17:17 cwayne, let me know if that works Nov 08 16:24:17 victorp: seems to, thanks Nov 08 16:24:31 victorp: i'll get you some more results overnight tonight as well Nov 08 16:26:27 cwayne, cool Nov 08 16:52:24 slangasek, another thing ... i played with make_ext2fs vs img2simg and ext2simg ... ext2simg produces about 100M bigger images than make_ext4fs, the latter apparently compresses the unused fs metadata or something along these lines so i think we should rather go with that Nov 08 16:53:27 (we are constrained wrt image size by the system ram of the device, images need to stay around 680M) Nov 08 16:55:12 (fastboot has an -S option to actually flash bigger images in chunks, but sadly thats not 100% reliable and sometimes you end up with a corrupted fs) Nov 08 16:56:40 ogra_: xnox: janimo: i'd like to get a sense of the WIs that need to be done before we can start producing dailies. i understand it as 1) kernel+fw bits uploaded into R, 2) cdimage infrastructure working (aka ext2simg, etc.) 3) usb-creator patches. anything else? Nov 08 16:57:18 achiang, nope, thats largely it Nov 08 16:57:30 achiang: usb-creator should not block "production of dailies" Nov 08 16:57:33 for *building* i dont even need the firmware Nov 08 16:57:35 achiang, yes. I am working on 1) Nov 08 16:57:43 xnox: well, "installation of dailies" :) Nov 08 16:57:48 even an empty package would do for that Nov 08 16:58:00 achiang, fastboot flash foo.img ;) Nov 08 16:58:12 I have packaged the utilities, and now will be flashing my nexus with them..... i hope i will not brick it =) Nov 08 16:58:14 oh, so continue using the nexus7-installer then? Nov 08 16:58:20 usb-creator is really not important yet :) Nov 08 16:58:35 well, its four fastboot commands, not really hard to do Nov 08 16:58:41 ok, so let's strike out #3 Nov 08 16:58:46 I need to fix the kernel build with raring gcc Nov 08 16:58:50 ogra_: right, I think it makes sense to stick with make_ext4fs for now in that case. Would be good at some point to unpick why the fs is different (e.g., how does 'make_ext4fs+simg2img' differ from 'mkfs.ext4') and improve things Nov 08 16:59:04 right, usb-creator will happen, but yioz can always install by just using fastboot Nov 08 16:59:07 in that case we just need to get cdimage.uc. building images, and we can patch nexus7-isstaller Nov 08 16:59:11 or maybe it's how does 'make_ext4fs' differ from 'make_ext4fs+simg2img+img2simg' Nov 08 16:59:23 yxeah Nov 08 17:00:02 slangasek: i'm getting cross-eyes from all the img2img2sigm2igm2sigm+ Nov 08 17:00:20 i also think i'll add the make_ext4fs part to livecd-rootfs/live-build instead of debian-cd Nov 08 17:00:32 will make everything easier for everyone Nov 08 17:01:09 * ogra_ goes afk for a while Nov 08 17:03:05 hrw, do you usually follow gcc versions with the corresponding cross-gcc package or the releases are more or less independent? Nov 08 17:15:30 janimo: is git clean -xdf going to remove my code and config changes in the kernel? Nov 08 17:15:34 cwayne: re: dual-boot bug, it would be nice to document why it's difficult, which is to say, you'd have to flash into the recovery partition, etc. etc. Nov 08 17:15:58 cwayne: can you track down the details of that and update the bug? Nov 08 17:16:03 achiang: sure Nov 08 17:16:42 thanks Nov 08 17:17:58 cwayne: re: gnome-shell bug, we can do better. desrt has 2 upstream bugs enabling gnome-shell. can you track those down and link them in our bug report? Nov 08 17:18:30 cwayne: he's already opened the upstream bugs, so all you need to do is link them Nov 08 17:26:12 cwayne: I've got some cycles while this builds, I'll do the gnome shell one Nov 08 17:28:09 janimo: You probably just want to drop do_tools until you sort it out (as I did for ac100). Nov 08 17:28:44 janimo: Rejecting for now, though. Nov 08 17:30:12 achiang: re: desrt/gnome-shell: did you mean patches or bugs? I dont see any upstream bugs that seem to fit Nov 08 17:30:33 mfisch: he filed bugs, that much i know. he *might* have attached patches too, not sure. Nov 08 17:30:42 mfisch: and they might already be closed, unknown to me... Nov 08 17:30:52 hmm, I'll go search his lp bug list Nov 08 17:30:53 mfisch: i only have this knowledge from a 5 minute hallway conversation with him Nov 08 17:30:59 mfisch: upstream gnome bugs Nov 08 17:31:04 ah Nov 08 17:39:14 ogra_: ping Nov 08 17:49:09 cwayne, have a look at this - https://wiki.ubuntu.com/Nexus7/BrowsingSimulation and please improve if you think anything is missing Nov 08 17:49:39 victorp_: will do, i've got it running now just for some preliminary info Nov 08 17:50:29 cool, the main thing is the extension really, rather than the script Nov 08 17:50:33 infinity, thanks. Was the issue the same for ac100? I did not even know about that Nov 08 17:51:26 victorp_: yep. did you set it to never turn the screen off when you ran overnight? Nov 08 17:53:21 infinity, I wonder if we could just copy the binaries from PPA to raring until we have a raring package. Would that be bad? Nov 08 17:56:23 janimo: Just upload with do_tools = false. :P Nov 08 17:56:25 infinity, ok I saw the recent ac100 fix in raring Nov 08 17:56:41 janimo: Copying from the PPA for a NEW package would still require the same review, so buys you nothing. Nov 08 17:56:46 achiang: no joy in finding ryan's upstream bugs Nov 08 17:57:22 mfisch: did you just ask him? :) Nov 08 17:57:54 achiang: he's not around, we will have to email Nov 08 17:58:25 yep, sorry for asking for the hoop jumping, but i think we should link those bugs Nov 08 17:58:38 I did find a bunch of other upstreams that I linked Nov 08 17:58:42 many of which predate our work Nov 08 17:58:46 like rotation not working Nov 08 17:59:07 yeah, watching the mails come into my inbox now Nov 08 18:00:14 janimo: Once you have something that builds, I'll try to fasttrack reviewing it. If the source is shockingly similar to ac100, it won't take too much effort. Nov 08 18:03:05 achiang: the main difficulty on dual-boot is either hacking fastboot or switching to u-boot and then figuring out how to reflash the bootloader? Nov 08 18:04:53 mfisch: i'm fuzzy on the details, it was like, resize the userdata partition, flash over the recovery, then something in picking which userspace to boot Nov 08 18:06:41 you'd have to manually enter the bootloader Nov 08 18:07:03 that method is safer than messing with fastboot Nov 08 18:07:31 cwayne: I forgot about that option ^^ Nov 08 18:07:43 * cwayne didnt know about that option Nov 08 18:07:58 so ubuntu would go into recovery Nov 08 18:08:04 so like, booting into android you would just use the recovery partition? Nov 08 18:08:05 not sure how userdata would get divided up Nov 08 18:08:08 ah Nov 08 18:08:56 cwayne: I've upstream linked the two bugs to the gnome-shell bug, are we re-opening it? Nov 08 18:08:59 https://bugs.launchpad.net/gnome-session/+bug/1072509 Nov 08 18:09:00 Launchpad bug 1072509 in ubuntu-nexus7 "Gnome Shell/Classic is Unable To Provide Accelerated Experience" [Low,Won't fix] Nov 08 18:09:34 mfisch: i suppose we can confirm it, but def leave it low Nov 08 18:09:45 cwayne: done Nov 08 18:10:01 mfisch: cwayne: i don't think it needs to be open, just a pointer to the upstreams are ok Nov 08 18:10:15 imho Nov 08 18:10:22 but do what makes sense. :) Nov 08 18:10:36 we can leave it open Nov 08 18:10:51 we'll leave it open for now, and see what happens in those upstreams i suppose Nov 08 18:12:25 * cwayne is not seeing the auto-dimming on browsing yet Nov 08 18:13:12 achiang: if we get the tegra3 bugs into good shape can you poke nvidia? If so, I'd like to prioritize that effort Nov 08 18:13:25 mfisch: i've been, in the background Nov 08 18:13:36 mfisch: just ping me if you need help on specific bugs Nov 08 18:13:57 achiang: okay, we'll get them into shape Nov 08 18:14:10 cwayne: can we do a review of them after you finish up the current issue? Nov 08 18:14:12 why was this marked as affecting nexus7? it doesn't seem quite right https://bugs.launchpad.net/ubuntu-nexus7/+bug/974260 Nov 08 18:14:13 Launchpad bug 974260 in gnome-settings-daemon (Ubuntu) "Brightnes of laptop panel is set to minimum when system starts" [Medium,Confirmed] Nov 08 18:14:19 mfisch: sure Nov 08 18:14:23 cwayne, yes I did Nov 08 18:14:26 cwayne: I duped the nexus7 one to that one Nov 08 18:14:53 cwayne, not many users browse with their screens turn off :) Nov 08 18:15:12 cwayne: the original: https://bugs.launchpad.net/ubuntu-nexus7/+bug/1075473 Nov 08 18:15:12 mfisch: it isnt minimal brightness though, else it'd be completely black Nov 08 18:15:12 Launchpad bug 974260 in gnome-settings-daemon (Ubuntu) "duplicate for #1075473 Brightnes of laptop panel is set to minimum when system starts" [Medium,Confirmed] Nov 08 18:15:21 victorp_: hah, true :) Nov 08 18:15:35 cwayne: the bug is more that the system doesn't remember where you left the brightness setting after a reboot Nov 08 18:15:43 mfisch: ah, okay Nov 08 18:15:46 that makes more sense Nov 08 18:15:47 cwayne: my laptop, for example, always starts at a blinding 100% Nov 08 18:17:16 mfisch: right, ok Nov 08 18:18:20 mfisch: wanna do a quick g+ and go over the tegra3 bugs? Nov 08 18:18:29 cwayne: IRC, I'm at the library Nov 08 18:18:43 mfisch: alright: https://bugs.launchpad.net/ubuntu-nexus7/+bugs?field.tag=tegra3 Nov 08 18:26:19 fwiw, i expect we're finding normal usability bugs in core ubuntu that have simply been ignored forever Nov 08 18:26:25 good thing is, upstream can no longer hide from us :) Nov 08 18:27:21 achiang: thats what it seems like to be honest Nov 08 18:27:31 most if not all of these have an older bug that's been a bit ignored Nov 08 18:27:39 yep Nov 08 18:40:09 ogra_: mfisch: i have recreated the rootfs image using packaged fs-utils, and flashed my nexus7 and it did boot and everything is ok. Nov 08 18:40:26 ogra_: mfisch: I'm planning to upload them into the raring archive now. Nov 08 18:41:00 do I need to backport them back to quantal & precise? or is ppa sufficient for cdimage? Nov 08 18:42:03 achiang: ^^? I think raring was the goal Nov 08 18:42:33 xnox: no backport needed at all Nov 08 18:42:40 xnox: Raring or nothing!@ Nov 08 18:42:45 ack =) Nov 08 18:43:16 uploading then. Nov 08 18:43:44 I think I will do a ppa and block for general public =) Nov 08 18:43:50 s/block/blog/ Nov 08 18:45:18 xnox: curious... what ppa? Nov 08 18:46:06 achiang: a new one, unless there is one already. Nov 08 18:46:21 achiang: do we have one under ~ubuntu-nexus7 ? Nov 08 18:46:22 xnox: sorry, i mean, what would the ppa have? Nov 08 18:46:48 the fs-utils package? Nov 08 18:46:55 achiang: android-tools package patches to produce android-tools-fs-utils which has simg2simg, simg2img, img2simg, make_ext4fs Nov 08 18:46:58 yeah =) Nov 08 18:47:12 and ext2simg I think.... Nov 08 18:47:20 xnox: maybe our Nexus7 public PPA? Nov 08 18:47:28 mfisch: sure. Which one is that? Nov 08 18:47:36 xnox: https://launchpad.net/~ubuntu-nexus7/+archive/ubuntu-nexus7-installer Nov 08 18:47:45 achiang: not that one Nov 08 18:47:58 ynot? Nov 08 18:48:03 * mfisch is thinking Nov 08 18:48:07 I guess those are image build tools Nov 08 18:48:08 it should indeed be that one Nov 08 18:48:09 right Nov 08 18:48:12 ok Nov 08 18:48:18 unless i misunderstand Nov 08 18:48:22 which is probably the case Nov 08 18:48:29 xnox: let me add you to the group Nov 08 18:48:49 ah, I lack the power Nov 08 18:49:08 * mfisch asks for the power Nov 08 18:49:10 xnox: can you apply for the ppa? Nov 08 18:49:20 https://launchpad.net/~ubuntu-nexus7 Nov 08 18:49:30 "Dmitrijs Ledkovs" (xnox) pending approval. Nov 08 18:49:43 xnox: gotcha, you're in Nov 08 18:49:48 thanks =) Nov 08 18:50:23 xnox: ping me when those are available, I'd like to update my blog post and remove the bianries from people.canonical.com Nov 08 18:50:42 mfisch: ack. Nov 08 18:50:44 achiang: I have a fix for the brightness issue Nov 08 18:51:57 * achiang adds a blurb about team membership to LP Nov 08 18:55:26 mfisch: which one, the 0 one? Nov 08 18:55:45 cwayne: ? Nov 08 18:56:06 mfisch: which brightness issue do you have a fix for Nov 08 18:56:12 cwayne: ah yes, that one Nov 08 18:56:18 more testing first Nov 08 19:16:30 anyone seeing the screen go black even when its not supposed to on the n7? Nov 08 19:17:57 <[mbm]> nope; worked fine when I tried it Nov 08 19:18:15 <[mbm]> screensaver? dpms? Nov 08 19:22:22 [mbm]: hmm, not sure. its disabled in brightness and lock, but that's all i did Nov 08 19:36:09 infinity, I did a reupload with tools disabled Nov 08 19:45:28 <[mbm]> mfisch: there was some talk of using kexec for dual booting, although it should also be possible to place the ubuntu kernel on the recovery partition, at the expense of losing android recovery Nov 08 19:46:28 [mbm]: well, there is more to do in the system itself, a kernel update would currently wipe your android kernel Nov 08 19:47:18 ogra-cb: [mbm]: can we get these issues listed on the dual-boot bug? Nov 08 19:47:55 <[mbm]> ogra-cb: the flash-kernel scripts would need to be tweaked Nov 08 19:48:21 well, they would be overwritten with the next flash-kernel update Nov 08 19:48:39 flash-kernel would need to learn to accept more than one bootdevice Nov 08 19:48:46 <[mbm]> there's also the question of where to put the ubuntu root Nov 08 19:49:02 which is a functionality that was just removed on purpose with the rewrite in debian Nov 08 19:49:42 the bug is here btw: https://bugs.launchpad.net/ubuntu-nexus7/+bug/1076199 Nov 08 19:49:42 Launchpad bug 1076199 in ubuntu-nexus7 "Feature request for Dual Boot" [Wishlist,Opinion] Nov 08 19:49:44 <[mbm]> thinking that the ubuntu root can just be a subdirectory of userdata, avoiding partitioning problems and free space segmentation Nov 08 19:50:47 the ubuntu root currently is userdata Nov 08 19:51:07 mmmcblk0p9 is / Nov 08 19:51:32 <[mbm]> right, but you don't want to see android files in / Nov 08 19:51:37 and we need to completely wipe it on install, else we get filesystem corruption over time Nov 08 19:51:49 i.e. fastboot erase Nov 08 19:52:09 (we had that issue) Nov 08 19:52:38 <[mbm]> thinking of a ubuntu directory on p9 Nov 08 19:53:09 well, then you will need to provide a modified image Nov 08 19:53:13 <[mbm]> combination of mount -o bind and pivot_root can be used in a similar manner to chroot Nov 08 19:53:37 the default image will wipe and use the full partition Nov 08 19:53:55 <[mbm]> right, the current install clobbers android Nov 08 19:53:55 cant android get along with that ? Nov 08 19:54:57 <[mbm]> well, android mounts userdata as /data and then makes various directories like /data/data, /data/app, /data/system .. Nov 08 19:55:26 ah, right Nov 08 19:55:31 <[mbm]> you don't want those directories showing up in ubuntu attached to / Nov 08 19:56:00 <[mbm]> and it's easily avoided with a few initrd tweaks in ubuntu Nov 08 19:56:03 well, dual boot isnt our focus, our focus is to get the cleanest ubuntu install possible onto teh device in the given situation Nov 08 19:57:00 we wont hack teh initrd nor change teh image design, so if you want dual boot, either get it working with the image as is (happy to accept patches to flash-kernel if they are proper) or roll a modified image Nov 08 19:57:58 <[mbm]> installation could be a debootstrap from an adb while running android Nov 08 19:58:23 <[mbm]> some packages to handle the dual boot config changes Nov 08 19:58:43 <[mbm]> no worse than the existing -nexus7 packages Nov 08 19:59:01 they will go awaty mostly Nov 08 19:59:11 *away Nov 08 19:59:42 hrw: did you actually use teh chromebook with chromeos before installing ubuntu ? Nov 08 20:00:06 feels really weird if your world is restricted to a browser window Nov 08 20:01:54 <[mbm]> how dors the nexus7 handle partitions? is there a gpt or is it just hardcoded? Nov 08 20:02:35 i think its a gpt hardcoded in the bootloader Nov 08 20:02:59 we dont touch the partitioning Nov 08 20:12:29 hmpf Nov 08 20:12:51 that webchat thingie is unstable Nov 08 20:14:46 <[mbm]> dual booting would be cleanest with u-boot and a new set of partitions, but that just wastes space Nov 08 20:34:47 ojn: seems clicking inside flash on teh chromebook only works sometimes, is that a known bug ? Nov 08 20:35:30 ogra-cb: under chrome os? Nov 08 20:35:44 It's not something I have noticed myself Nov 08 20:35:48 NekoXP: so I fixed the issue, but I didn't need your patch, your note about brightness of 13 was spot-on Nov 08 20:36:19 NekoXP: http://paste.ubuntu.com/1343515/ Nov 08 20:36:32 ojn: well, not a single game works for me, in some videos i have play/pause but cant go fullscreen Nov 08 20:37:02 ogra-cb: hm, odd. Let me check here. Nov 08 20:37:10 youtube works though Nov 08 20:37:22 including fullscreen Nov 08 20:37:30 youtube might be html5 and not flash though (in some cases) Nov 08 20:37:36 ah, indeed Nov 08 20:38:40 ojn: try www.myvideo.de (if you can, not sure they block foreign countries) i can play/pause but none of the bottom controls work Nov 08 20:38:46 vimeo works for me, i can switch to full-screen without problems Nov 08 20:39:00 it's slow due to no codec offload with their player though Nov 08 20:39:01 ok one sec Nov 08 20:39:10 (works fine on x86) Nov 08 20:39:57 Sorry! Dieses Video darf in deinem Land nicht angezeigt werden. Nov 08 20:39:59 BUT Nov 08 20:40:07 I can make the commercial go full screen and adjust the volume on it Nov 08 20:40:23 weird Nov 08 20:40:30 what's your chrome os version (from about:chrome)? Nov 08 20:41:09 23.0.1271.83 Nov 08 20:41:30 i see a green checkmark Nov 08 20:41:37 I'm on .84 Nov 08 20:41:45 but I'm on dev-channel, not stable or beta Nov 08 20:41:57 and it actually did an upgrade before i could log in Nov 08 20:41:59 I don't think there were any flash fixes between the two anyway Nov 08 20:42:09 very strange Nov 08 20:42:27 yes, that's quite odd Nov 08 20:42:38 I can ask my coworkers when I am back in California next week, if you still see it Nov 08 20:42:54 it's not something silly like a stuck alt/ctrl key or the like? Nov 08 20:43:14 for youtube you need to go to http://youtube.com/html5 and enable the Html5 beta Nov 08 20:43:17 k, not sure i'll keep using chromeos tthough Nov 08 20:43:43 but i definitely wont wipe it so i can go back for verifying etc Nov 08 20:44:03 mfisch: I think it turns on by default from chromebooks Nov 08 20:44:12 ogra-cb: ok Nov 08 20:44:24 i Nov 08 20:44:38 i'll surely go on using it for a few days though Nov 08 20:45:11 * ogra-cb is actually intrested if he can get used to live in a browser only Nov 08 20:46:06 the last bit for me was an irc client, and I got that through irccloud.com now. Browser + terminal window for SSH access is all I really need Nov 08 20:46:07 oh, wait ! Nov 08 20:46:34 * ogra-cb uses teh touchscreen extension on teh nexus, seems that was copied over to this setup Nov 08 20:46:44 Oh! that'd do it Nov 08 20:47:19 yeah, switching it off makes it work Nov 08 20:47:21 heh Nov 08 20:47:40 phew Nov 08 20:47:43 ok, calling it a day here Nov 08 20:47:54 lots of crap to take care of tomorrow before going back to the us. :) Nov 08 20:48:40 good luck with that Nov 08 20:48:44 and a safe flight Nov 08 20:51:07 tnx Nov 08 21:09:19 <[mbm]> touchscreen on the nexus is buggy Nov 08 21:12:12 <[mbm]> in some ways it'd make more sense to run the touchscreen in relative mode Nov 08 21:13:48 <[mbm]> treat the entire surface like a laptop touchpad Nov 08 21:19:23 does someone know how to debug a segfault in a armhf pbuilder (qemu)? I'm obviously doing something wrong: http://paste.kde.org/600374 Nov 08 21:24:27 I have ubuntu-server on my beagleboard-xm. I would like to do remote-debugging. what package do I need to install on my host machine (ubuntu 12.04) to have the toolchains ? Nov 08 21:24:51 apt-get install build-essential? Nov 08 22:20:49 mfisch, excellent.. 13 is actually just 12.5% rounded up, btw ;) Nov 08 22:24:35 janimo: cross toolchain gets updates ~monthly Nov 08 22:25:01 janimo: raring one next week I hope - have some build problems which I did not had time to check during conferences Nov 08 22:25:21 ogra-cb: I used chromium os before switching to raring Nov 08 22:25:46 ogra-cb: check my G+ to get link to ppa with newer alsa-lib packages to get ucm profiles for chromebook Nov 08 22:26:12 ogra-cb: x11 driver will follow but not sure if it will be tomorrow or next week Nov 08 22:37:41 hrw: Please send those ucm profiles my way so I can include them in raring., Nov 08 22:38:07 hrw: Alternatively, where can I find your PPA? Nov 08 22:40:50 TheMuso: Looks to be https://launchpad.net/~hrw/+archive/my-own-packages/+files/alsa-lib_1.0.25-4ubuntu2.dsc Nov 08 22:40:54 hrw: Will your updated cross toolchain include an arm64 one too? Nov 08 22:42:16 persia: Thanks, will grab those and get them into raring. Nov 08 22:49:23 hrw: did you already try to fish out the flash plugin from the chrome partition ? Nov 08 22:49:31 i suppose it could work for us Nov 08 22:51:09 ogra-cb: ooh, would that enable g+ hangouts? Nov 08 22:51:35 not auew, but hangouts definitely work fine under chromeos Nov 08 22:52:18 hm... maybe they are not using flash anyway Nov 08 22:52:24 s/auew/sure Nov 08 22:52:51 right, i think its a separate plugin Nov 08 22:52:54 i had a flashblocker for a while, and hangouts still worked Nov 08 22:53:08 but chromeos is armhf so all binaries should kind of work Nov 08 22:53:26 something else in the ars technica review convinced me to not get this chromebook yet Nov 08 22:53:54 what was that ? Nov 08 22:54:50 oh maybe it was the screen Nov 08 22:55:21 or the fact that it's *still* thicker than a mac air! :) Nov 08 22:55:28 well, its fine for 720p movies ... surely not as impressive as the nexus screen though Nov 08 22:56:07 and yeah, the clamshell feels a bit cheap Nov 08 22:56:13 the fact that they shipped without getting hdmi working is... lame Nov 08 22:56:18 the kbd is great though Nov 08 22:57:14 it really cant cope with the ac100 frtom a coolness POV but the cpu and ram balance that out a bit :) Nov 08 22:57:23 ac100 still cooler? Nov 08 22:58:36 lighter, slimmer, i like the kbd a bit better and with teh upgraded screen it can easily cope with this one Nov 08 22:58:57 but its a tegra2 and only has 512M Nov 08 22:59:04 once ogra-cb confirms hangouts work in ubuntu... i'll consider it. :) Nov 08 22:59:34 haha Nov 08 22:59:54 hrm, this web irc client is odd, it doesnt auto-scroll Nov 08 23:01:48 but i'll go afk now anyway Nov 09 00:29:15 mfisch: what are your thoughts on this bug?https://bugs.launchpad.net/ubuntu-nexus7/+bug/1076793 Nov 09 00:29:16 Launchpad bug 1076793 in ubuntu-nexus7 "Bad multicore performance" [Undecided,New] Nov 09 00:30:37 cwayne: unsure really Nov 09 00:30:45 cwayne: I'm about eod, lets check it out tomorrow Nov 09 00:30:52 cwayne: I guess we should repro it Nov 09 00:31:03 I'd like to know how it works on a multicore x86 box too Nov 09 01:50:31 with regards to nexus7 and running android side-by-side can I suggest LVM, it is the solution I have gone down for the tf201 cause repartitioning on these tegra devices is a real pain Nov 09 01:54:03 lilstevie: Turn all the Android partitions into one big LVM volume? I think I suggested that to Oliver a couple of weeks ago. ;) Nov 09 01:54:57 infinity, well that too, I started with just UDA Nov 09 01:55:09 but lately yeah, the whole damn lot Nov 09 01:55:51 Would be nice if we could just repartition, like I did on my ac100, but that seems to be a non-option on the nexus7. Nov 09 01:56:14 you need nvflash for that Nov 09 01:56:26 Indeed I did, yes. Nov 09 01:56:27 and the nexus 7 is ODM secure Nov 09 01:57:27 and our nvflash stuff doesn't work with it Nov 09 01:58:09 not to mention I don't think canonical could endorse such hackery Nov 09 02:00:50 * persia considers mobiln-image-creator and smirks Nov 09 02:01:55 lol Nov 09 02:02:34 persia, do you have time at the moment? Nov 09 02:02:39 also hi Nov 09 02:02:39 :p Nov 09 02:02:54 Some, not lots. Nov 09 02:03:07 heh ok Nov 09 02:03:33 repo.lilstevie.geek.nz is where that kernel is if you get some time to have a look Nov 09 02:03:40 Give me a URL, and I'll hit it as soon as I have time, which likely saves scheduling issues (and apologies for not poking you earlier today) Nov 09 02:04:04 that's ok, you probably wouldn't have reached me Nov 09 02:04:13 I didn't get up until 10am and it is only 1pm now Nov 09 02:04:27 What did you have to change in casper? Nov 09 02:04:48 TheMuso changed something, I think he pushed that for quantal though Nov 09 02:04:48 Yeah, but I had more time when you got up than I do now :) Nov 09 02:06:11 it was something to do with not looking at any storage on the usb bus for the tf201 Nov 09 02:06:14 Anyway, unless something comes up, I should be able to look at it in ~6 hours. Nov 09 02:06:20 ok cool :) Nov 09 02:13:00 anyway with some minor work, the new kernel should support the entire tegra3 transformer range **** ENDING LOGGING AT Fri Nov 09 02:59:58 2012