**** BEGIN LOGGING AT Thu Dec 10 02:59:56 2009 Dec 10 03:06:17 awong: ffmpeg builds for arm as of a week or so ago. there was a small preprocessor issue that landed a few days ago, since then it builds from tot Dec 10 10:02:01 i got a little problem with my beagle boar; i put a ubuntu 9.10 on my board and now i got the following problem: if i plug in a keyboard i get the following error: "hub 1-0:1.0 unable to enumerate USB device on port 2", if i plug in any other device, nothing happens Dec 10 10:02:23 is this known... or has anybody experiences with this Dec 10 10:34:32 hansdampf: This is an issue with your kernel; we don't provide beagleboard/OMAP kernels, I would check with whoever provided your kernel Dec 10 11:44:54 nice, seems thumb2 already saves us 20MB in image size Dec 10 11:52:30 ogra_: but does the image work? ;) Dec 10 11:53:37 amitk, imx51 boots (which is my main target for an A1 image) but seems to have issues to install on B3.0 (test on 2.5 is just running) ... dove has issues Dec 10 11:53:59 niceeee Dec 10 11:54:56 yeah Dec 10 13:17:17 hey folks Dec 10 13:17:25 Just talked to a French Freescale guy Dec 10 13:17:55 who told me that thumb 2 is mostly slower than regular arm code, I was aware the diffence was so significant Dec 10 13:18:04 Let's hope that the space savings in cache make up for it Dec 10 13:18:11 and he also told me that thumb 2 was bugged on TO2 Dec 10 13:18:24 So it might cause faultsevery 4 Kb or something he didn't have any detail on Dec 10 13:18:42 => we should check for erratas from FSL on thumb 2 issues -- it might break random programs Dec 10 13:18:45 lool, dusing image install i cant see *any* difference in speed on my bababge with lucid Dec 10 13:18:49 *during Dec 10 13:19:07 i was actually expecting to see a speedup Dec 10 13:19:20 but its as slow/fast as karmic was Dec 10 13:25:11 yey Dec 10 13:26:43 ogra: Ok Dec 10 13:27:31 lool, we have lucid babbage images, feel free to take a loo ;) Dec 10 13:27:39 and a look as well :P Dec 10 13:28:16 I don't have any babbage 2.x Dec 10 13:28:18 only 1 Dec 10 13:28:51 ah, yeah, you sent them to jamie Dec 10 13:29:09 and to asac Dec 10 13:29:20 Well I left the 2.0 in london and think it was sent to asac Dec 10 13:29:21 right Dec 10 13:29:29 no, he has a 3.0 Dec 10 13:29:40 Bah where is that 2.0 then I wonder Dec 10 13:29:46 no idea Dec 10 13:30:03 been sold so they can pay your bonus :) Dec 10 13:30:10 mine! Dec 10 13:30:24 (note tha past tense :) ) Dec 10 13:30:26 *the Dec 10 13:32:21 ogra: You've got a bonus? I thought we were all sitting on our bonus this year due to the financial crisis! Dec 10 13:33:52 lool: ogra: i got the 2.0 and 3.0 at the same time Dec 10 13:34:10 which happened because i wasnt aware i was getting the 3.0 any time soon Dec 10 13:34:16 so i have two atm Dec 10 13:34:26 which is not really enough itlr Dec 10 13:34:43 heh Dec 10 13:34:46 well. but i would give the 2.0 away if i can get a dove ... which currently is a deep whole in my hardware setup Dec 10 13:35:02 though i would like to have a dove and two bbg Dec 10 13:35:16 or two doves .. etc. if i do porting stuff like the chromium thing i cannot do anything else Dec 10 13:35:17 ogra: I was just kidding BTW Dec 10 13:35:37 lool, indeed :) Dec 10 13:35:39 I hope you felt bad! :) Dec 10 13:36:06 well, *i* will get my bonus ... *you* will have to wait until asac sold your babbage :) Dec 10 13:36:10 :P Dec 10 13:36:29 i am holiding back lool bonus until he releases his dove board ;) Dec 10 13:36:52 I'm holding back my babbage until asac's dove gets sold to ogra! Dec 10 13:36:58 haha Dec 10 13:37:04 i'm getting all boards Dec 10 13:37:15 lol Dec 10 13:37:42 soo, how can we end up with 16M initrds now Dec 10 13:37:44 armin76: Does Gentoo ship/plans to ship pre-build ARM binaries? Dec 10 13:38:05 armin76: I wonder whether you build for hard float; I think that's not well supported with EABI Dec 10 13:38:12 *pre-built Dec 10 13:38:19 lool: we do already, but we don't support them Dec 10 13:38:34 http://tinderbox.dev.gentoo.org/default-linux/ Dec 10 13:38:38 nope, no hardfloat Dec 10 13:38:38 armin76: what's your optimization level for your public binaries? Dec 10 13:38:55 v5/6/7? thumb2? Dec 10 13:39:07 armv5tel-softfloat-linux-gnueabi Dec 10 13:39:14 Hmm that's old Dec 10 13:39:26 I guess you miss hardware for higher than v5 Dec 10 13:39:34 yep Dec 10 13:39:58 why do you think i'm always saying that stuff i say? :) Dec 10 13:40:02 hehe Dec 10 13:42:13 i got an efika mx now, though Dec 10 13:42:23 The one in a nice case? Dec 10 13:42:34 * asac runs update-initramfs in verbose mode Dec 10 13:42:35 yep, but its a lange board anyway Dec 10 13:42:52 armin76: Did you get a BSP / linux source code with it? Dec 10 13:43:02 lool: yes Dec 10 13:44:09 Linux efikamx 2.6.28-ER1-efikamx #4 PREEMPT Sat Nov 28 16:47:52 CST 2009 armv7l GNU/Linux Dec 10 13:44:26 armin76: Oh nice, so you can easily redistribute these patches right? Dec 10 13:44:36 asac, ogra: ^ we could at least start from these patches Dec 10 13:44:46 AFAIK we still have no public patch stack from pegatron Dec 10 13:45:17 lool: you should ask neko @ #efika Dec 10 13:45:32 He came around here once Dec 10 13:45:39 there's no patch, only patched source Dec 10 13:45:45 That's "ok" Dec 10 13:45:51 The only thing I need is a public source for it Dec 10 13:45:54 s/I/we Dec 10 13:46:05 Until now we only had non-publishable patches because of foo or bar Dec 10 13:46:11 Isn't that why we have diff? Dec 10 13:46:40 http://pastebin.com/f1b2e2ab5 Dec 10 13:47:27 The Netwalker sources (another i.MX51) are also open and available to the public. Dec 10 13:47:28 persia: I mean a public location to retrive some form of source Dec 10 13:47:39 persia: Good luck splitting the imx51 stuff apart with diff though Dec 10 13:47:57 Well, true, but it's better than nothing, no? Dec 10 13:47:59 lool: they aren't public though, they are included in the ssd Dec 10 13:48:20 asac, looks dfine to me Dec 10 13:48:32 ogra: the produced initrd is better: Dec 10 13:48:38 weird Dec 10 13:48:46 oh no Dec 10 13:48:47 lool: i'll exchange them with a dove *g* Dec 10 13:48:49 sorry misread Dec 10 13:48:56 -rw-r--r-- 1 root root 1196802 2009-12-10 14:43 initrd.img-2.6.31-601-imx51 Dec 10 13:49:04 s/with/for Dec 10 13:49:05 well Dec 10 13:49:11 11M is still better than 16 Dec 10 13:49:27 ack Dec 10 13:49:28 asac, i really think we need to compare the contents Dec 10 13:49:32 that's 1 MB? Dec 10 13:49:39 1196802 => 1 196 802 Dec 10 13:49:40 ogra: the contents i already posted Dec 10 13:49:45 http://pastebin.com/f98ce43d Dec 10 13:49:49 those are all the modules added Dec 10 13:50:19 hmm Dec 10 13:50:21 i mean unpack both initrd files Dec 10 13:50:35 and then run diff on a find output Dec 10 13:50:36 ogra: i have those unpacked Dec 10 13:50:44 ogra: thats what the paste is about Dec 10 13:50:45 just for the modules Dec 10 13:51:03 oh, i was confused because it only shows the kernel dir Dec 10 13:51:27 yes. because the kernel modules dir has a different name Dec 10 13:51:29 so i ran inside Dec 10 13:52:30 ah Dec 10 13:52:40 lets compare the two configs Dec 10 13:52:54 ogra: http://pastebin.com/f7538b9e5 Dec 10 13:53:03 thats the diff for all (ignore the modules) Dec 10 13:53:53 so for me it feels like this crypt* stuff is a good target Dec 10 13:54:34 thats all new at least Dec 10 13:54:59 ah ! Dec 10 13:55:01 usr/share/initramfs-tools/hooks/cryptopenct usr/share/initramfs-tools/hooks/cryptpassdev Dec 10 13:55:04 usr/share/initramfs-tools/hooks/cryptopensc usr/share/initramfs-tools/hooks/cryptroot Dec 10 13:55:05 /lib/frimware ... Dec 10 13:55:12 how big is that Dec 10 13:55:33 # This list needs to be kept in sync with the one defined in passdev.c Dec 10 13:55:33 for fs in ext3 ext2 vfat reiserfs xfs isofs udf; do manual_add_modules "$fs" > /dev/null 2>&1 || true Dec 10 13:55:37 done Dec 10 13:55:39 thats a bunch of new fs modules Dec 10 13:55:40 i guess Dec 10 13:55:49 let me check Dec 10 13:55:57 ogra@osiris:/var/build/images$ du -hcs /media/bb4fdb28-11f1-4c72-a345-cc7006aa4898/lib/firmware/ Dec 10 13:55:57 12M /media/bb4fdb28-11f1-4c72-a345-cc7006aa4898/lib/firmware/ Dec 10 13:56:00 thats karmic Dec 10 13:56:30 2M difference in lucid Dec 10 13:56:32 13700 lib/firmware/ Dec 10 13:56:50 ogra: all the fs above are definitly one part Dec 10 13:57:08 yeah, that adds up :/ Dec 10 13:57:25 though i dont get why the casper initrd isnt that big then Dec 10 13:57:35 image building should have failed Dec 10 13:58:14 the casper initrd.lz is 2.8M Dec 10 13:58:16 cryptroot also adds a bunch of get_device_modules Dec 10 13:58:31 yes thats odd Dec 10 13:58:45 that might shrink by 1M due to lzma Dec 10 13:58:47 +./kernel/net/netfilter/xt_esp.ko Dec 10 13:58:52 but all that netfilter module stuff Dec 10 13:58:54 but then it would still be only 3.8M Dec 10 13:58:54 where got that added? Dec 10 13:59:10 kernel config Dec 10 13:59:34 hmm Dec 10 14:00:34 Hi all— for those who didn't spot it, I raised a bug on squashfs: https://bugs.launchpad.net/ubuntu/+source/squashfs/+bug/494667 Dec 10 14:00:39 Launchpad bug 494667 in squashfs "[armel] non-ISO-C misaligned pointer punning causes slowness and SIGILLs" [Undecided,New] Dec 10 14:00:56 ogra: but the "net" modules are explicitly listed in hook-functions Dec 10 14:01:03 so those are pulled in from somewhere else Dec 10 14:01:12 you say that hook-functions also looks somewhere else? Dec 10 14:01:26 you mean the NIC drivers or netfilter ? Dec 10 14:01:39 ogra: no ... i mean all the netfilter/* modules Dec 10 14:01:45 they are now added Dec 10 14:01:49 http://pastebin.com/f98ce43d Dec 10 14:01:56 +./kernel/net/netfilter/xt_esp.ko Dec 10 14:01:57 ... Dec 10 14:02:10 also Dec 10 14:02:11 +./kernel/net/bridge Dec 10 14:02:14 ... Dec 10 14:02:20 generally modules are added from hook-functions Dec 10 14:02:22 +./kernel/net/ipv4 Dec 10 14:02:23 ... Dec 10 14:02:23 Since we are stuck with the 2.6.28 kernel on the buildds for now, I think the quickest workaround is probably to avoid the mksquashfs crashes is for __packed__ attributes to be added. This isn't ideal, because it will result in slower execution, but it should avoid the SIGILLs. Dec 10 14:02:32 but every hook script can put additional modules in Dec 10 14:02:33 ogra: i am going to build a BabbageImageFromScratch, but how to get the latest initrd.img? Dec 10 14:02:40 ogra: really look in this: http://pastebin.com/f98ce43d tats really a huge amount of new modules Dec 10 14:02:48 ogra: yes. but i dont see netfilter anywhere Dec 10 14:03:01 feels like it was previously individual files and now its just the full "net" dir Dec 10 14:03:01 cooloney, use an old one, boot and run update-initramfs in the running system Dec 10 14:03:22 ogra: ok, thanks, downloading. heh Dec 10 14:03:39 dmart: thanks. let me read ;) Dec 10 14:03:50 cooloney, though if you have anything as module that you need to get the rootfs up, that wont work indeed Dec 10 14:04:30 ogra: Why not, as long as the module is in the initramfs? Dec 10 14:04:51 persia, because you likely have a version mistmatch Dec 10 14:05:55 ogra: Ah, using an old initramfs with a new kernel. I thought you were suggesting to boot the old initramfs and old kernel, and then install the new kernel and run update-initramfs. Dec 10 14:06:09 right, that works for sure Dec 10 14:06:34 ogra and persia, yeah, if that works, i can build a new disk with my new kernel installed right? Dec 10 14:06:36 just drop the modules into /lib/modules and copy the kernel binary over vmlinuz Dec 10 14:06:52 then run update-initramfs Dec 10 14:06:58 that will just put them into fis Dec 10 14:07:15 ogra: ok, got it. Dec 10 14:07:28 ogra: where can i find the armel rootfs? Dec 10 14:07:33 dmart: hmm ok. Dec 10 14:07:43 how laborious is fixing the C code? Dec 10 14:07:47 cooloney, you dont have one installed yet ? Dec 10 14:08:18 ogra: i dd the karmice imx51 release image to a SD card. Dec 10 14:08:39 oh, you didnt install it ? Dec 10 14:08:49 ogra: so i guess this SD already has the rootfs, right? Dec 10 14:09:29 ogra: i installed it on a sata driver only, Dec 10 14:09:40 it only has the livefs Dec 10 14:10:25 ogra: ok, do i need to reinstall i on a SD card or USB disk Dec 10 14:10:41 i would recommend USB Dec 10 14:11:00 and that installed usb stick contains the rootfs, right? Dec 10 14:11:21 cooloney: And a copy of the kernel and the modules, etc. Dec 10 14:11:44 It's a full installed system (although there is an extra copy of the kernel and initramfs in some boot location) Dec 10 14:11:45 persia: yeah, understand. Dec 10 14:12:24 so how big usb stick is enough for that? Dec 10 14:12:59 Recommended minimum secondary storage for an Ubuntu installation is 4G. Dec 10 14:13:18 I know some people have managed it with far less, but there's no promises it works with less. Dec 10 14:14:47 persia and ogra ok, i got it. thanks a lot Dec 10 14:30:05 JamieBennett: what did we find so far for casper speed? anything we could fix easily? Dec 10 14:42:29 asac: To me, it looked very laborious to convert the squashfs C code to do aligned access only, unless someone has a good idea about how to do it— it might require quite a lot of code to be rewritten, and the logic of how the output data is built up might have to change. This doesn't feel feasible for now :( Dec 10 14:42:49 kk Dec 10 14:43:19 we should at least carry tha tupstream though Dec 10 14:43:34 dmart: why did this start to happen in lucid only? and only on that hardware? Dec 10 14:43:58 ... hang on ... Dec 10 14:44:36 Building with -marm might work. The Jaunty kernel can't cope with all unaligned accesses in Thumb, but should cope if the userland code is ARM. Dec 10 14:45:15 I've added that suggestion to the bug, and I'll give it a try. Dec 10 14:45:16 dmart: Would it just be building mksquashfs with -marm, or would the contents matter? Dec 10 14:45:17 ok. i think -marm feels acceptable. its not something that impacts user machines. Dec 10 14:45:41 asac: Doesn't it affect rootstock users? Dec 10 14:46:00 persia: Do you mean the squashfs contents? I think that shouldn't matter Dec 10 14:46:04 yes. but thats development tool ;) Dec 10 14:46:18 so ... developer machines ... or am missing something ? Dec 10 14:46:37 Well, most users *are* developers today :) Dec 10 14:46:50 Once we have true Karmic/Lucid buildds, we can switch to -mthumb and it should all work. So -marm would be a temporary measure for this package. Dec 10 14:53:23 I just rebuild squashfs-tool with -marm and re-ran my test on one of the nettops, and it ran without SIGILL, so I think this may work as a solution. Dec 10 14:58:34 thanks dmart Dec 10 15:02:47 targetted bug ... assigned to dyfet to prepare the debdiff Dec 10 15:02:54 * persia gets confused by http://packages.qa.debian.org/s/squashfs/news/20091202T163929Z.html Dec 10 15:03:34 hmm Dec 10 15:03:55 I found the issue. cf. http://packages.qa.debian.org/s/squashfs-tools.html Dec 10 15:04:18 But since we have source/binary namespace issues, we run into the same sort of issues why `apt-get source linux` doesn't actually work. Dec 10 15:04:20 ack Dec 10 15:04:44 should get removed from our arch too then Dec 10 15:04:53 Looks like we need to chase that transition in Ubuntu for other reasons, and so would want to do the -marm stuff at the same time. Dec 10 15:04:59 hmm Dec 10 15:05:12 yeah Dec 10 15:05:21 Might need a poke, as it's source removal but not binary removal (and stuff like that tends to confuse busy archive admins) Dec 10 15:05:21 dyfet: also look at the package rename at that time maybe Dec 10 15:05:23 linux ? Dec 10 15:05:39 you cant remove linux from armel, it builds versatile Dec 10 15:06:14 ogra: apt-get source linux pulls the linux-meta source. We now have the same issue in that apt-get source squashfs pulls the squashfs source and apt-get source squashfs-tools pulls the squashfs source and apt-get source won't pull the squashfs-tools source without being hit about the head. Dec 10 15:06:22 ah Dec 10 15:06:31 i missed the meta reference :) Dec 10 15:06:35 Okay...adding -marm itself for arm arch seemed simple enough for squashfs-tools... Dec 10 15:06:53 apt-get --source-only source sounds like it should work to solve this, but it never works for me; I usually have to resort to wget Dec 10 15:06:58 guess there is no way to fire off ubiquity at install stage ? Dec 10 15:07:18 dyfet: thanks. check if we also want to get the new source from debian or something Dec 10 15:07:19 (where hitting about the head is the --only-source parameter) Dec 10 15:07:19 asac, can you elaborate Dec 10 15:07:22 maybe two steps Dec 10 15:07:35 (except this fails for some reason as well :( ) Dec 10 15:07:36 asac: is there already a bug for this? Dec 10 15:07:40 ogra: well. want to trial and error change hooks etc. to see what is adding all those modules Dec 10 15:07:59 asac, https://wiki.ubuntu.com/ARM/BabbageInstallVariants only ubiquity ? Dec 10 15:08:00 ogra: and update-initramfs in chroot yielded a 1.6M initrd only Dec 10 15:08:19 ah, thats what you want Dec 10 15:08:28 ogra: i am in a live sesssion. ... wanted to rerun just the last step that triggers that Dec 10 15:08:29 no, it needs to get to the end of the install sadly Dec 10 15:08:47 1.6M seems to small Dec 10 15:08:52 ogra: so what is different from just running update-initramfs in chroot? Dec 10 15:09:00 good question Dec 10 15:09:05 which is where i had the verbose output Dec 10 15:09:10 ogra: how to best mount dev? Dec 10 15:09:18 seems like ubiquity doesnt mount /dev completely Dec 10 15:09:28 and the crypt hooks complain about /dev/null non-existing Dec 10 15:09:37 oh Dec 10 15:09:42 that sounds buggy Dec 10 15:09:57 did you check /dev in your initramfs ? Dec 10 15:09:59 *g* Dec 10 15:10:16 hmm. not inside it. i checked it in the /target mounted chroot Dec 10 15:10:20 probably it copies any virtual fs Dec 10 15:10:25 and there only /dev/pts is mounted Dec 10 15:10:31 yeah, for apt Dec 10 15:10:43 it should mount dev and friends dynamiocally Dec 10 15:10:47 are you saying that the hooks are run even in the initramfs? (e.g. not in the /target) ? Dec 10 15:10:54 no Dec 10 15:10:57 ok Dec 10 15:11:30 asac: Are you trying to update a live image and then reboot it? Dec 10 15:11:55 no. i am trying to run the update-initramfs that is run at end of ubiquity before the flashing without running the whole ubiquity Dec 10 15:11:59 persia, he tries to find out how update-initramfs is run from ubiquity Dec 10 15:12:19 asac, grep the ubioquity code is my best suggestion Dec 10 15:12:21 i see that its just run. but not sure what env it is run it ;) Dec 10 15:12:49 i know i added stuff to mount certain dirs that were missing in karmic, so i know it mounts unmounts dynamically for each command that needs it Dec 10 15:13:03 also i wonder why there is no /dev/null in /target ;) Dec 10 15:13:35 ok had to run the install anyway. seems that failing flash busts the SD card somewhat Dec 10 15:13:38 /dev gets bind mounted dynamically Dec 10 15:13:40 so i had to re dd that Dec 10 15:13:47 ogra: the full /dev? Dec 10 15:13:53 odd ... i will check that Dec 10 15:14:02 only saw devpts mounted Dec 10 15:14:08 but if its always unmounted that might explain it Dec 10 15:14:13 will just do it then Dec 10 15:14:25 check the code, it should have some function like "prepare_chroot()" or some such Dec 10 15:14:30 yep Dec 10 15:14:52 that calls the mount commands dynamically right before update-initramfs grub-install and other commands that need it Dec 10 15:15:44 might be named differently ... prepare_target or so ... i dont remember the actual name Dec 10 15:16:23 right it bind mounts /dev Dec 10 15:16:34 if not os.path.exists(os.path.join(self.target, 'proc/cmdline')): Dec 10 15:16:34 self.chrex('mount', '-t', 'proc', 'proc', '/proc') Dec 10 15:16:34 if not os.path.exists(os.path.join(self.target, 'sys/devices')): Dec 10 15:16:34 self.chrex('mount', '-t', 'sysfs', 'sysfs', '/sys') Dec 10 15:16:34 misc.execute('mount', '--bind', '/dev', os.path.join(self.target, 'dev')) Dec 10 15:16:38 its called chroot_setup Dec 10 15:16:44 ah, right Dec 10 15:17:20 ok great. odd that it it leaves devpts mounts behind though ;) Dec 10 15:17:25 will check in a few minutes when install failed Dec 10 15:17:42 devpts is used by dpkg/apt for logging Dec 10 15:17:58 they might be needed at that stage Dec 10 15:18:01 ogra: To console? Dec 10 15:18:24 to nothing ... i dont even think they are actually used Dec 10 15:18:39 but dpkg spills errors if it cant find them at the start of a package install Dec 10 15:18:52 i think its just kept around to quieten that Dec 10 15:18:54 * persia thought logs went to /var/log/* and to Dec 10 15:19:37 i think d-i and ubiquity only log to /var/log/* ... d-i just runs a tail -f on tty4 Dec 10 15:19:40 asac: just got back, see https://wiki.ubuntu.com/ARM/CasperSpeedup Dec 10 15:19:45 persia: what is universe-contributor about? Dec 10 15:20:02 my nm student wants to apply there, but i am not sure if he really should do that rather than motu Dec 10 15:21:28 JamieBennett: the more detailed analysis is the script content? Dec 10 15:21:33 like for 10adduser? Dec 10 15:21:36 asac: #ubuntu-motu or #ubuntu-devel are better places to ask that question, but it's a recognition of significant and sustained work within the Ubuntu development community. Dec 10 15:22:15 asac, i think debconf-communicate generally has speed issues Dec 10 15:22:16 asac: https://wiki.ubuntu.com/UbuntuDevelopers#Ubuntu%20Contributing%20Developers Dec 10 15:22:28 asac: yeah, the timings are there for each slow part (i.e. a chroot or a locale generation e.t.c) Dec 10 15:23:05 (and feel free to have someone talk to me directly, either here or on oftc about that stuff) Dec 10 15:23:09 asac: next step is to look in more detail exactly what each of these code paths does and see if we can optimise them Dec 10 15:23:10 persia: but is that really something developers should apply for? Dec 10 15:23:23 i mean developers that already do development for ubuntu? Dec 10 15:23:26 asac: ues Dec 10 15:23:33 (after that i will go to some other channel ;)) Dec 10 15:23:50 asac: Well, it depends on what they are doing. Contributing Developers get @ubuntu.com email and authority to represent the Ubuntu project. Dec 10 15:24:10 There's no related specific upload rights. Dec 10 15:24:37 Some development teams require Ubuntu Membership prior to application, and since all Contributing Developers are also members, this can be one way to get it. Dec 10 15:25:06 persia: kk Dec 10 15:25:08 thx Dec 10 15:26:26 asac: By the way, I suspect you're a member of the team (likely through inheritance). Dec 10 15:26:36 he slacks Dec 10 15:26:38 * armin76 runs Dec 10 15:27:56 feels like i keep my old stuff for ever Dec 10 15:28:09 at least for the next 6 month i dont see any change :( Dec 10 15:28:19 just stop doing it :) Dec 10 15:28:42 ignorance ftw ! Dec 10 15:28:57 Well, both of you are not likely to care, really, because this is all stuff that matters more to people who don't already have root on tens of millions of systems :) Dec 10 15:28:59 for NM i can stop and things will survive. for mozilla we all will get fired once the world notices that we are without security updates for year ;) Dec 10 15:29:25 who cares, in one year we'll all use armel devices with chromium Dec 10 15:29:41 yeah. still distro reputation can go down quickly Dec 10 15:29:43 * persia will probably stick to epiphany Dec 10 15:29:46 also mozilla sues us ;) Dec 10 15:30:17 in worst case making our brand damage reponsible for loosing their google money at some point ... that would be expensive ;) Dec 10 15:30:18 heh, i would like to see them actually sueing anyone Dec 10 15:30:36 did they ever enforce their policy at court ? Dec 10 15:30:45 trademark? Dec 10 15:30:51 yep Dec 10 15:30:54 no. but they managed to get all distros not using it Dec 10 15:30:59 but three: ubuntu/suse/redhat Dec 10 15:31:08 yep Dec 10 15:31:17 so i am sure they will be able to enforce it Dec 10 15:31:33 trademark law has enough precedence that as long as they complain about it properly they don't need to have taken it to court. Dec 10 15:31:42 right Dec 10 15:31:54 (well, perhaps not in some jurisdictions, but at least for most of the places they really care about) Dec 10 15:32:00 and since they are really running around forcing folks to stop doing that you cannot say they handled it to openly Dec 10 15:32:50 * ogra goes to call sudo make sandwich ... Dec 10 15:32:54 enjoy Dec 10 15:44:49 dyfet: thanks for drafting the ooffice spec. in the work items i added a few options (from minimal to almost-perfect integration) Dec 10 15:44:58 do those make sense? can you add them to the implementation section too? Dec 10 15:47:29 hmm. noticed that bindmounting /dev is slightly different from the original /dev Dec 10 15:47:42 like /dev/shm has different attributes in /dev than in the bindmounted location Dec 10 15:47:45 odd Dec 10 15:47:53 bindmounting /dev/shm somewhat fixes this Dec 10 15:47:56 shouldnt cause our issues though Dec 10 15:47:59 asac: I realized when reviewing specs that it got lost from when we did the original split.... Dec 10 15:48:09 ogra: no. but it cauess issues for chromium when run in chroot Dec 10 15:48:14 it uses /dev/shm for sandboxing Dec 10 15:48:18 ah Dec 10 15:48:46 dyfet: right. thanks a lot Dec 10 15:49:07 dyfet: to make it perfect putting one long sentence for the three integration variants i put into the work items would be great Dec 10 15:49:21 put that to implementation section Dec 10 15:49:27 ok Dec 10 15:49:30 problem is that its pretty hard to do the integration compared to mailto: Dec 10 15:49:37 thats why i want three potential solutions to be in there Dec 10 15:49:45 yea...I was thinking about that issue :) Dec 10 15:49:46 1st. simple bookmark+ open browser integration Dec 10 15:49:58 2nd. auto import when double clicking (but no back and forth synch) Dec 10 15:50:10 3rd. gvfs gdocs mount for Documents folder Dec 10 15:50:27 i think those are the approaches i put into the work items ... having a sentence for each would be good Dec 10 15:54:47 asac: so your suggesting we turn the document into a browser bookmark entry? Dec 10 15:55:26 dyfet: no. 1st. approach is simple. it just allows you to open the borwser with gdocs if you click on the "office" launcher icon Dec 10 15:55:31 no import etc. Dec 10 15:55:42 also adding a bookmark to the browser might be ok Dec 10 15:55:59 2nd. is about importing office documents if you want to open them Dec 10 15:56:06 using libgdata Dec 10 15:56:27 and then opening a browser with just gdocs opened and the right folder maybe or even file Dec 10 15:56:45 if we are really smart we could rename that file at that time and replace it with a link to it in gdocs Dec 10 15:56:51 but that can get out-of-sync Dec 10 15:56:52 so i am not sure Dec 10 15:57:12 rather would wnat to have 3. which would import it on first click and move it to the mounted Documents folder Dec 10 15:57:31 so starting there it would work to just open it in browser and also we wont get out of sync if you move files in gdocs Dec 10 16:01:13 yes, each has some ugly aspects... Dec 10 16:02:24 though a simple mime helper app could be written in python to do any of these things... Dec 10 18:02:32 dyfet: well. if you can think of a forth ... potential perfect option feel free to discuss that :) Dec 10 18:02:40 the gvfs mount in Documents feels for me like the best solution Dec 10 18:02:44 If I can I will Dec 10 18:02:45 almost perfect. Dec 10 18:02:59 when you click on a file outside of the moint you would get asked if you want to import it Dec 10 18:03:07 and then the file would be moved to some "backup" location Dec 10 18:03:16 so you usually just use the mounted location Dec 10 18:04:10 dyfet: but imo all close to perfect solutions are off what we can do for lucid (e.g. 3. needs a bunch of work on the gdocs backend for gvfs) Dec 10 18:04:27 but writing the ideas down makes sense still Dec 10 18:04:44 yes, there is always the option of deferring it, too, in this case Dec 10 18:05:04 dyfet: let me know when you wrote the 3 ideas down so i can just approve the spec for now ;) Dec 10 18:05:09 that would make it done imo Dec 10 18:05:35 I initially added it to the spec already, but I think I need to expand on the third one Dec 10 18:08:06 dyfet: just write one sentence for each outlining the implementation/integration approach Dec 10 18:08:15 thats what I did Dec 10 18:08:18 ok Dec 10 18:08:29 * asac checks wiki again Dec 10 18:09:06 dyfet: maybe make bullet points out of the three approaches Dec 10 18:09:17 that improves visual recognizability in the spec Dec 10 18:09:17 okay :) Dec 10 18:09:42 dyfet: the third would also open the browser with the right location after import Dec 10 18:09:55 also if you click on the mounted documents the browser would be opened too of coures Dec 10 18:10:07 (or openoffice if user chooses to open with openoffice) Dec 10 18:11:28 cool. so what i will do is send a request to design team to make a design suggestion for the wizards ... i see three wizards atm: a) select preferred mail handler, b) select preferred office handler, c) import wizard (if we implemnt that) Dec 10 18:11:41 also we need icons like outlined in the work items Dec 10 18:11:59 for generic mail ... and office (with option to overlay webservice branding based on users selection) Dec 10 18:34:38 ogra: so yeah. cryptroot adds 13.5m of modules Dec 10 18:34:45 basically all the drivers in the world Dec 10 18:34:51 now i need to find that package ;) Dec 10 18:34:59 dpkg -S cryptroot on lucid anyone? Dec 10 18:42:44 bug 495161 filed Dec 10 18:42:45 Launchpad bug 495161 in cryptsetup "initramfs cryptroot hook bloats armel initrd by adding >13M of compressed modules" [Critical,Triaged] https://launchpad.net/bugs/495161 Dec 10 18:44:00 * persia stops trying to install initramfs-tools on lucid to verify Dec 10 18:45:45 yeah ;) Dec 10 18:46:33 I only had another 27 minutes to wait for the download to complete :) Dec 10 18:46:39 (including dependencies, etc.) Dec 10 18:48:12 persia: heh. keep on going Dec 10 18:48:19 having lucid ready will be handy ;) Dec 10 18:49:04 I have lucids available for about half the architectures I use. I should get to 75% this weekend, and probably 100% next week. Dec 10 18:49:45 It's just that I don't currently run any lucid live systems, so need to install the low-level stuff (rather than just forwarding X from a chroot). Dec 10 18:50:03 (again, I should have at least two running lucid environments by next week) Dec 10 19:38:36 asac: re the aac.c arm patch for chromium, hclam apparently has a fix open for it, and will be submitting it soon. The -fPIC change should also land sometime today. So, things should be good probably by later today or tomorrow. Dec 10 19:42:05 awong: great. thatnks for the update Dec 10 19:42:25 awong: so he will land more fixes than for ffmpeg? Dec 10 19:42:33 no prob. Thanks for bringing it up. Dec 10 19:43:25 fbarchard had the -fPIC change yestertday but it broke other parts of our build I think, so he had to revert. hclam is working on various arm things (I actually don't know exactly what) and one of them is that compile fix. I just asked him to try and submit that compile fix early. Dec 10 19:45:50 FYI, video performance on arm is probably going to be pretty dismal for chromium right now. I saw some tests of the performance here, and it super slow/choppy. There's actually some bugs in our rendering logic that doesn't handle super slow processers well for video and leads to exteremly bad behavior in the frame dropping code. hclam was working on that last, but don't know his status exactly. Dec 10 20:58:38 asac, you rock ! Dec 10 21:01:37 awong: thx. first we need a working build though ;) Dec 10 21:16:19 asac: completely agree. :) Dec 10 21:45:57 hmm the installer is quite slow imo Dec 10 22:01:34 plars: so reformatting existing partitions seems to be buggy. crashes ubiquity. but deleting partition and creating works though Dec 10 22:07:06 Adjusting partitions on flash tends to do wonky things with the FTL anyway (assuming you're looking at issues installing to USB flash) Dec 10 22:13:10 plars: hmm, mine crashed even when repartitioning Dec 10 22:15:53 plars: I tend to have better luck with much abused USB flash doing an erase rather than a write every once in a while. How this really happens depends on the firmware, but at least for the Elecom keys I prefer, dd if=/dev/zero of=${flash device} bs=${flash eraseblock size} seems to do the right thing. Dec 10 22:16:05 Then it needs to have a new partition table written from scratch, etc. Dec 10 22:37:01 plars: yes. reformatting is what makes it choke for me Dec 10 22:37:09 if i really add parition or something its ok Dec 10 22:37:19 just if i select previously used partition and set it to / or something Dec 10 22:43:31 hmm. bad issues ;) Dec 10 22:43:37 on reboot Dec 10 22:43:41 asac: were you able to get around the install failure by removing the hook from /target Dec 10 22:43:45 also usb errors about bad device descriptors Dec 10 22:43:55 plars: yes. after the "copying files ..." stage Dec 10 22:43:58 but before "flashing kernel" Dec 10 22:44:05 you need to remove the cryptroot stuff Dec 10 22:44:09 or actually comment out one line Dec 10 22:46:12 gotta reboot, brb Dec 10 22:53:55 asac: where would the line need to be commented from? Dec 10 22:55:22 plars: can you paste that file ;) Dec 10 22:55:24 i just lost it Dec 10 22:55:29 (and the image even) Dec 10 22:57:47 plars: are you good at launchpadlib writing ;)? Dec 10 22:58:02 (by coincident) Dec 10 22:59:17 asac: good? I wouldn't say so, I've done some basic things though Dec 10 22:59:25 ;) Dec 10 22:59:29 asac: if you have ideas for scripts, I'd love to work on them though Dec 10 22:59:57 for the relesae team report we basically need a script that gives info about lucid targetted bugs added/fixed/triaged in last week with armel tag i guess Dec 10 23:00:11 https://wiki.ubuntu.com/MobileTeam/ReleaseStatus/Lucid Dec 10 23:00:12 so that Dec 10 23:00:20 the RC bugs section basically ;) Dec 10 23:00:22 somehow automized Dec 10 23:02:03 not sure if one can easily query if something got closed/etc. in last week though Dec 10 23:02:43 I don't think that can be queried, but we can get a list e.g. daily or 6-hourly and then compare them periodically. Dec 10 23:02:49 Needs some client-side storage. Dec 10 23:02:59 asac: http://src.chromium.org/viewvc/chrome?view=rev&revision=34297 Hopefully that fixes the compile issues Dec 10 23:03:41 rock on Dec 10 23:03:55 yeah client side storage is bad ;) Dec 10 23:04:04 but if thats the only way its better than having to look through all bugs ;) Dec 10 23:04:37 client-side storage doesn't need to be bad. There are enough folk about with always-on servers. Dec 10 23:04:50 awong: great. so fPIC is still waiting for some issues? Dec 10 23:05:06 I'm not sure I really want to code it, but I certainly could host it if nobody else wished to do so. Dec 10 23:05:15 thanks persia Dec 10 23:05:47 so getting a weekly dump of targetted bugs with its state in a filterable/diffable form Dec 10 23:05:51 would be good i think Dec 10 23:06:48 * persia wonders what "thanks" is for Dec 10 23:08:01 for the offer and for caring ;) Dec 10 23:08:05 asac: nope, that should be fixed too. http://src.chromium.org/viewvc/chrome?view=rev&revision=34284 Dec 10 23:08:35 Oh, good. Let me know if nobody else will write it, as it would be an excuse to learn lplib, but hosting is easy. Dec 10 23:09:07 awong: hmm. there is no PIC mentioned there Dec 10 23:10:01 asac: yeah, it's int he changelog, but not in the diff. However, if you look at the ffmpeg.gyp file, it has -fPIC in the cflags section for ['target_arch=="arm"', {. Dec 10 23:10:28 great Dec 10 23:10:29 thanks!! Dec 10 23:10:52 we switched to 4am UTC for daily snapshots so i can copy that to builders for arm right away tomorrow morning Dec 10 23:11:27 no prob. I didn't do the work. :) Let me know if things still don't work for some reason. Dec 10 23:11:39 hmm. i had one patch for chromium i gave fta Dec 10 23:11:41 let me check Dec 10 23:11:49 its basically about brokenness for non armv7=0 builds Dec 10 23:11:52 ah right Dec 10 23:11:54 its skia.gyp Dec 10 23:12:06 the SSE source is not excluded if its not armv7=1 Dec 10 23:12:11 and also the opts are not added Dec 10 23:12:53 http://paste.ubuntu.com/338225/ Dec 10 23:12:59 Okay. If fta has the patch, he can submit it for review and the skia guys can have a look. Dec 10 23:13:00 thats the new block in skia.gyp needed Dec 10 23:13:07 replacing it in skia.gyp Dec 10 23:13:19 * asac should get a git tree of chromium i guess Dec 10 23:13:24 or svn checkout at least ;) Dec 10 23:13:44 fta: do you need anything from me for the skia.gyp changes? Dec 10 23:13:44 heh. seeming like it. :) And we should get you into the authors file to clear all the hurdles. Dec 10 23:13:53 right. Dec 10 23:14:00 i will take that as an action ;) Dec 10 23:14:07 to sign contributor agreement etc. Dec 10 23:14:16 and setup svn Dec 10 23:15:15 i don't really have much time atm Dec 10 23:15:24 all fine ;) Dec 10 23:15:55 busy with work Dec 10 23:17:58 plars: what timezone are you in? Dec 10 23:21:03 ( UTC -6 ) Dec 10 23:21:22 weather clock is sorted by name Dec 10 23:21:26 not UTC offset :/ Dec 10 23:21:33 what is that? chicago? Dec 10 23:21:45 I use Morelia for that timezone. Dec 10 23:21:51 what is morelia Dec 10 23:21:57 is that in US? Dec 10 23:22:31 first search result says Mexico. Dec 10 23:22:42 Yep, that's it. Dec 10 23:23:06 the region selection in weather clock can be ... erm... improved ;) Dec 10 23:23:18 its like a zillion scrollable menu ;) Dec 10 23:23:36 I know. That's part of why I try to pick odd city names. Dec 10 23:24:33 morelia is not in there :( Dec 10 23:24:54 hmm i have dallas Dec 10 23:24:56 thats enough ;) Dec 10 23:24:58 Hrm. Dallas works. Dec 10 23:25:14 yeah ;) Dec 10 23:25:25 clearly the right solution woudl be to reduce it down to one choice. That way the users can't be confused with having too many options. And ofcourse, that one choice should be something htat's 30minutes off of UTC. Dec 10 23:25:44 awong: Why choose 30 minutes when 15 and 45 are available? Dec 10 23:26:26 i think we should just use the ubiquity timezo9n selector Dec 10 23:26:29 thats great thing Dec 10 23:26:38 though some cities could deserve a point ;) Dec 10 23:26:54 i clicked on berlin and got prague or something similar close Dec 10 23:27:05 timezone is the same though Dec 10 23:27:19 Except weather doesn't work. Dec 10 23:27:38 I was very annoyed last week because I couldn't pick a location near where I was, and so couldn't see the weather. Dec 10 23:27:42 ;) Dec 10 23:27:51 weather was too broken so i only use it for time now ;) Dec 10 23:27:56 the old weather thing worked at least Dec 10 23:28:21 So it's not 5 degrees, feels like 0.4 there? Dec 10 23:28:24 not sure if they moved to an "open" serive to get weather datra now ;) Dec 10 23:31:04 persia: are there actually timezones that are 15 and 45 mins off? I only knew of the 30-min ones... Dec 10 23:31:30 Yes. Dec 10 23:31:37 * persia hunts one up Dec 10 23:32:24 http://www.worldtimezone.com/wtz-names/wtz-wct.html is one example Dec 10 23:32:57 oh fun. Dec 10 23:33:27 I'm so glad icu exists. Dec 10 23:33:59 ICU? Dec 10 23:34:33 http://site.icu-project.org/ Dec 10 23:35:32 Oh, for the timezone calculations? Dec 10 23:35:35 yeah Dec 10 23:35:46 heh. Dec 10 23:36:53 Actually, I can't find any :15 timezones. Only :00, :30, and :45. Odd, that. Dec 10 23:38:32 is there a way to grab the pre-image livefs somewhere? Dec 10 23:40:06 what timezone is pacific e.g. what city? Dec 10 23:40:30 Santa Fe Dec 10 23:40:43 Or Portland Dec 10 23:40:47 its not in there Dec 10 23:40:51 even tokio isnt Dec 10 23:40:54 Los Angeles? Dec 10 23:40:58 neither Dec 10 23:40:59 nor sf Dec 10 23:41:04 let me check portland Dec 10 23:41:04 Try Tokyo. We don't spell it that way anymore. Dec 10 23:41:07 seattle neither Dec 10 23:41:12 Vancouver? Dec 10 23:41:13 persia: well i looked for Tok* Dec 10 23:41:15 nothing Dec 10 23:41:28 Mountain View? Redmond? Dec 10 23:41:45 Sacramento? Dec 10 23:42:16 ha Dec 10 23:42:25 there is a "United states" submenu ;) Dec 10 23:42:28 with pacific Dec 10 23:42:36 I'm surprised Tokyo isn't there. Japan used to have lots of timezones, and settled on Tokyo time (except for Okinawa) about 50 years ago. Dec 10 23:42:41 Aha! Dec 10 23:42:45 maybe there is a "japan" submenu Dec 10 23:42:46 ;) Dec 10 23:42:49 or asia Dec 10 23:42:52 Probably Asia/Tokyo Dec 10 23:42:59 (That's what LP calls my timezone) Dec 10 23:43:05 but given the length of the top level list thats ridiculous Dec 10 23:43:12 Indeed. Dec 10 23:43:17 antartica ;9 Dec 10 23:43:19 is a submenu Dec 10 23:43:21 australia Dec 10 23:43:22 too Dec 10 23:43:24 but no asia ;) Dec 10 23:43:26 wow Dec 10 23:43:35 Australia deserves a submenu: there's something like 10 timezones. Dec 10 23:43:42 even chile and congo are submenus ;) Dec 10 23:43:52 (because some states do DST and others don't, and there's WCT, etc.) Dec 10 23:44:29 Chile gets interesting, just because there's something like a 3-hour gap in the middle of the timezones it spans. Dec 10 23:44:46 ok added taiwan Dec 10 23:44:51 thats as close as i get to japan ;) Dec 10 23:44:52 That's not my timezone. Dec 10 23:44:58 It's off by one. Dec 10 23:44:58 i know Dec 10 23:45:02 OK. Dec 10 23:45:05 what city do you have in weather applet? Dec 10 23:45:21 Sydney, Tokyo, Berlin, London, Boston, Morelia, Portland Dec 10 23:45:33 morelia isnt there for me Dec 10 23:45:38 tokyo neither :( Dec 10 23:45:46 did you set that by latitude? Dec 10 23:45:54 Nope, but I set it in Jaunty. Dec 10 23:45:57 err long... Dec 10 23:46:06 * persia upgrades rather than reinstalling Dec 10 23:46:06 ok i give up Dec 10 23:46:16 enough time wsated scrolling an endless menu ;) Dec 10 23:46:21 I can give you lat/long if that helps. Dec 10 23:46:22 guess i am better in my head :) Dec 10 23:46:32 OK. About livefs images. Dec 10 23:46:35 no. i can remember its taiwan + 1 ? Dec 10 23:46:39 or -1 ? Dec 10 23:46:44 heh :) Dec 10 23:46:46 +1 Dec 10 23:46:49 ok Dec 10 23:46:54 so its 8:45 ;) Dec 10 23:46:57 quite early :) Dec 10 23:47:09 They do exist. They can be accessed, but I think they are only accessible to livefs-builder admins and ubuntu-cdimage by default. Dec 10 23:47:11 ok livefs is much better ;) Dec 10 23:47:15 hmm Dec 10 23:47:17 ok Dec 10 23:47:29 I tend to just pull them from the image, because that's easier. Dec 10 23:47:41 have a script ready for that? Dec 10 23:47:46 or monuted? Dec 10 23:47:55 (or construct them myself with livecd-rootfs) Dec 10 23:48:05 do you have a full mirror? Dec 10 23:48:10 i would like to have something that matches what we produce Dec 10 23:48:14 not something "close" Dec 10 23:48:47 `sudo mount src/images/foo.img /mnt; cp /mnt/casper/squashfs.img src/scratch/squashfs.img; sudo umount /mnt` Dec 10 23:49:24 As long as the archive is consistent, running livecd-rootfs locally will produce the same thing as the livefs-builders, unless there is something odd about them. Dec 10 23:50:07 (like the mksquashfs bug we were discussing earlier) Dec 10 23:50:26 persia: hm thought the image cannot be mounted because its has multiple partitions etc. Dec 10 23:50:32 oh Dec 10 23:50:46 i will figure ;) Dec 10 23:50:53 OOps. script error. Dec 10 23:51:11 `sudo mount src/images/foo.img /mnt; cp /mnt/casper/filesystem.squashfs src/scratch/squashfs.img; sudo umount /mnt` Dec 10 23:51:32 Oh, if you're working with the multipartitioned SD image, just use losetup. Dec 10 23:51:43 * persia hunts up kirkland's excellent blog post Dec 10 23:52:25 http://blog.dustinkirkland.com/2008/10/mounting-kvm-disk-image.html Dec 10 23:55:13 So, you run losetup and kpartx before my script, and mount /dev/loopN to /mnt Dec 10 23:55:31 yeah Dec 10 23:55:33 that will do i guess Dec 10 23:55:37 great Dec 10 23:55:59 bookmarked till i try Dec 11 00:33:09 hmm. those dvi things on babbage boards. are those compatible with VGA? or do they only support digital? Dec 11 01:17:30 asac: US Central Dec 11 01:17:53 yeah Dec 11 01:18:47 so with commenting out the lines Dec 11 01:18:54 i end up with a console login ;) Dec 11 01:19:03 and i cnanot log in because of these timestamp issues i suspect Dec 11 01:19:24 if you paste that file i can show it you ;) Dec 11 01:19:39 you need to edit that after "Copying Files ..." (i could change it at about 60% Dec 11 01:24:31 * asac out **** ENDING LOGGING AT Fri Dec 11 02:59:56 2009