**** BEGIN LOGGING AT Mon Mar 14 02:59:57 2011 Mar 14 08:44:01 morning Mar 14 10:04:32 guys, anyone know how to preserve a working serial console after framebuffer start? Mar 14 10:27:14 ogra: do you know how i can preserve a working serial console even after the framebuffer start? Mar 14 10:28:17 preserve ? Mar 14 10:32:22 flag_: yep, when the framebuffer starts, my console dies and i have to switch to my monitor to see what's going on Mar 14 10:33:13 you should get all boot messages if you set the cmdline to point to serial and drop the "quiet" from the cmdline Mar 14 10:33:32 just edit /boot/boot.script and run sudo flash-kernel Mar 14 10:39:04 ogra: and what if one the variable in my uboot environment is longer than 80 chars? on do i print it in its entirety? Mar 14 10:39:41 you editor should display it fine Mar 14 10:40:16 no no Mar 14 10:40:32 i mean, i'm in uboot and i try to "printenv $var" Mar 14 10:40:50 but the content of var is longer than 80chars Mar 14 10:40:56 thus it's truncated Mar 14 10:43:16 there is only the hardcoded env (which just loads the boot.scr file) you will have to change a lot if you dont just edit boot.scr Mar 14 10:43:33 so printenv wont get you very far Mar 14 10:43:55 if it cuts off after 80 chars thats an issue with your terminal program Mar 14 10:44:08 or with the TERM variable you use Mar 14 10:45:13 ogra: i'm using minicom Mar 14 10:47:23 well, i'd really recommend to rather change your boot.scr than to poke around in the uboot env Mar 14 11:28:20 ogra: ok, deleted quiet from bootargs in boot.scr, but still my console dies after: Mar 14 11:28:30 ogra: Starting kernel ... Mar 14 11:28:31 Uncompressing Linux........................................................................................................................................................................................... Mar 14 11:28:55 you should get all boot messages if you set the cmdline to point to serial and drop the "quiet" from the cmdline Mar 14 11:28:56 ogra: setenv bootargs root=UUID=e18764f5-7859-4509-gj48-bbcc01a1ff77 ro splash Mar 14 11:29:12 did you set the cmdline to point to the serial terminal ? Mar 14 11:30:22 ogra: bootargs=console=ttyS0,115200n8 root=UUID=54ec8187-d25e-413a-ac5b-99af9cc379fc (from u-boot) Mar 14 11:30:46 is it like that in your boot.scr ? Mar 14 11:32:12 ogra: bootargs in u-boot and boot.scr should match, right? Mar 14 11:32:24 ogra: ah, got it Mar 14 11:37:05 ogra: my bad, there was no console= in /boot/boot.scr but only in u-boot bootargs and i thought the kernel picked the console variable from u-boot Mar 14 11:37:45 no, as i said above the u-boot stuff is totally irrelevant Mar 14 11:38:03 its overriden as soon as boot.scr was loaded Mar 14 11:42:00 ogra: k, thanks Mar 14 12:19:05 heh.. I forgot how long can arm system need to upgrade.. Mar 14 12:19:31 heh Mar 14 12:19:46 i did an ac100 update-manager upgrade over the weekend Mar 14 12:19:49 took 36h Mar 14 12:20:02 though on a very slow SD card Mar 14 12:20:07 apt-get dist-upgrade from alpha3 to current - started >1h ago on sd Mar 14 12:20:52 ogra: I reinstalled all packages on my dualcore atom in ~30 minutes. on 60MB/s sata Mar 14 12:30:20 hmm Mar 14 12:30:29 who fired off a netbook buid Mar 14 12:30:33 *build Mar 14 12:31:26 oh, not netbook, headless Mar 14 12:49:06 http://www.computerworld.com.au/article/379549/calxeda_arm_chips_designed_480-core_servers/ - want one ;D Mar 14 12:53:23 hrw: you can try convincing Rob or Martyn in Budapest Mar 14 12:53:54 amitk: calxeda people? Mar 14 12:55:05 hrw: yeah, you might know them as smoothstone people from previous UDSes Mar 14 12:55:13 ah, right Mar 14 12:55:27 the company which changes names Mar 14 13:15:41 k, finally unity-2d on ac100 Mar 14 13:21:51 you did not had it before? Mar 14 13:22:05 we didnt have neon runtime detection in QT Mar 14 13:22:11 ah, right Mar 14 13:22:26 and after that i had to find out that the natty libc doesnt woork on tegra2 Mar 14 13:22:41 o.. hows that? Mar 14 13:22:51 http://gitorious.org/ac100/kernel/commit/227af643e6253e9a5c2a2f74468f855686c44117?diffmode=sidebyside Mar 14 13:23:00 see #linaro, was just discussed there Mar 14 13:23:17 time to blag a ac100 now then :-D Mar 14 13:24:09 unity-2d is really snappy here Mar 14 13:26:03 * hrw -> lunch Mar 14 14:27:37 janimo, poke Mar 14 14:28:47 ogra_: cool, so new qt really worked well for you? Mar 14 14:28:56 yep Mar 14 14:29:11 currently using it afer i spent the whole weekend to get natty running Mar 14 14:29:12 nice Mar 14 14:29:20 ogra_, hello Mar 14 14:29:24 ogra_: did you find what were the issues? Mar 14 14:29:28 libc is really badly broken on the ac100 Mar 14 14:29:33 you said you had many issues last week Mar 14 14:29:43 someone pointed me to http://gitorious.org/ac100/kernel/commit/227af643e6253e9a5c2a2f74468f855686c44117?diffmode=sidebyside Mar 14 14:29:56 i'm currently running with the maverick libc pinned Mar 14 14:30:04 janimo, so did you test headless ? Mar 14 14:30:12 ogra_, not yet Mar 14 14:30:22 seems we have a bunch of issues with oem-config/ubiquity Mar 14 14:30:23 ogra_: ouch, got it Mar 14 14:30:29 was trying tosee why images fail Mar 14 14:30:51 ogra_, I saw you made some changes around oem-config/upstrart last week Mar 14 14:31:01 janimo, the oem-config UI never comes up on serial Mar 14 14:31:12 unless no tty exists at all Mar 14 14:31:37 (i.e. it immediately works if i use a broken omapfb mode so ttys cant be started) Mar 14 14:32:05 i suspect ubiquity/oem-config simply has no code for that and uses the first console it finds Mar 14 14:32:20 the second big bug i see is that no user is created at all Mar 14 14:32:21 what would a fix be? We need to keep ttys I think Mar 14 14:32:29 no idea yet Mar 14 14:32:44 i would suggest looking at d-i how it determines the used console Mar 14 14:33:05 if werial is set we should defautl to use it i think Mar 14 14:33:22 ok Mar 14 14:33:30 effectively it doesnt seem like oem-config finishes its job Mar 14 14:33:53 the most funny part is that the debconf package removal that runs afterrrr oem-config defaults to serial Mar 14 14:33:56 so right now one can log in but even if oem-config is started by hand it does not run Mar 14 14:33:56 ? Mar 14 14:34:12 oem-config runs fine Mar 14 14:34:16 but on tty1 Mar 14 14:34:41 while it takes input from serial (since the kernel defaults to it as first console) Mar 14 14:35:01 i can use my laptop kbd to fiddle with oem-config on the panda dvi screen Mar 14 14:35:02 :) Mar 14 14:35:26 it seems to just not use the right output Mar 14 14:36:01 besides oem-config what else is broken? Mar 14 14:36:07 i havent found ut why we dont have build attempts since three days (ampty mails usually suggest that the connection to the image builder failed= Mar 14 14:36:30 well, oem-config doesnt create a user at all atm Mar 14 14:36:30 ogra_, telepathy-glib FTBFS (regression) which holds up empathy Mar 14 14:36:46 janimo, shouldnt affect headless at all Mar 14 14:37:05 ogra_, indeed Mar 14 14:37:08 we get empty build failure mails and i see no logs at all on the imagebuilder Mar 14 14:37:34 who can debug that? Mar 14 14:37:38 someone has changed livecd-rootfs or debian-cd ... or the ssh connection from cdimage to the image builder is broken Mar 14 14:37:45 * ogra_ will research that Mar 14 14:42:16 so there was one change in livecd-rootf that only affects edubuntu Mar 14 14:44:19 * ogra_ cant see any changes that could have made our builds fail Mar 14 14:44:24 lets try a manual build Mar 14 14:54:44 seems to run flawless Mar 14 15:34:55 http://cdimage.ubuntu.com/ubuntu-headless/daily-preinstalled/20110314/ Mar 14 15:34:59 hmm, worked fine Mar 14 15:44:01 janimo, ^^^ Mar 14 17:29:32 ogra_, janimo: It still has issue with oem-config. The program is displaying the user interactive screens on the attached monitor, but only taking input via serial console. Mar 14 17:29:40 ubuntu Mar 14 17:29:43 oops Mar 14 17:34:25 It also failed to create a user. May be due to no network. Not sure. Mar 14 17:37:15 GrueMaster, yes, as discussed above Mar 14 17:37:48 Bugs filed? Mar 14 17:38:11 GrueMaster, fi you give a broken frambuffer setting it works (i.e. if ttyS/O is the only console around) Mar 14 17:38:31 and the debconf bits after oem-config are on serial too Mar 14 17:38:56 oem-config was simply never used on serial i think Mar 14 17:39:08 so we are the first ones to face these issues Mar 14 17:39:18 no bugs filed yet, no Mar 14 17:39:33 I do see something interesting in the /var/log/oem-config.log. It has an error open: /dev/tty not found. Mar 14 17:39:58 oh really ? Mar 14 17:40:08 I'll file a bug soon. Flashing another SD to try on beagleXM. Mar 14 17:40:12 k Mar 14 17:40:29 it worked fine for me with the old kernel btw Mar 14 17:40:30 But it is near the end of the log (after other activity). Mar 14 17:40:44 because the omapfb setting completely broke the framebuffer Mar 14 17:40:46 Also, it failed to create a user, but that may be due to lack of network. Mar 14 17:40:55 unlikely Mar 14 17:41:30 wither we dont call all modules from the debconf frontend or what i suspect is my hack to rename the machine to localhost by default is bad Mar 14 17:41:39 Hey, I find unlikely bugs all the time. I'm good at it, remember? :P Mar 14 17:41:47 heh, yeah Mar 14 17:42:03 but i really doubt missing network has anything to do with it Mar 14 17:42:44 Same here, but I have seen weirder issues. Mar 14 18:33:23 rsalveti: ping Mar 14 18:34:41 rsalveti: i've begun cleaning xbmc's packaging and install scripts. it would be nice to have a mentor that i could speak with about best practices. know of anyone who might be willing/able to help? Mar 14 18:35:05 TheUni: I can try to help on that Mar 14 18:35:14 persia would be the best guy around for that Mar 14 18:35:21 he knows a lot about packaging Mar 14 18:35:36 TheUni: is it still the packaging tree? Mar 14 18:35:43 yes Mar 14 18:36:22 i have some work done locally to begin consolidating packages, but i'd like to know that i'm making things better and not worse Mar 14 18:36:25 TheUni: and how is the daily builds, did you manage to get it to work? Mar 14 18:36:47 rsalveti: yes, daily builds are up. but they're not publicized yet because i'd like to rework the packages first Mar 14 18:36:56 oh, ok Mar 14 18:37:10 incase we end up causing some conflicts while shifting things around Mar 14 18:37:32 https://launchpad.net/~team-xbmc/+archive/unstable Mar 14 18:38:12 TheUni: at the moment it seems you're building for arm Mar 14 18:38:18 at least Architecture: i386 amd64 powerpc ppc64 Mar 14 18:38:35 is that because of the daily builds? Mar 14 18:38:46 cool Mar 14 18:39:05 rsalveti: simply because we haven't tackled packaging it yet Mar 14 18:40:14 but afaik current git is building/working on beagle Mar 14 18:40:33 TheUni: cool, nice to know Mar 14 18:41:07 rsalveti: ok, i'll work up a list of questions and hand them off to you. you can pass me along if you'd like :) Mar 14 18:41:25 some XBMC_CONFIG_OPTIONS would be arch specific, as for ARM we'd like to have gles support, and not gl and vdpau Mar 14 18:41:28 TheUni: sure Mar 14 18:42:15 right Mar 14 18:42:21 gles is default for arm Mar 14 18:42:22 but it does seems to be better Mar 14 19:49:25 Not sure what changed, but unity-2d-places is less than usable in the 20110311 image. Can't even get to a terminal without crashing. Mar 14 19:49:55 s/hanf/hang/ Mar 14 19:50:35 pmathews: ??? Mar 14 19:52:51 GrueMaster: wrong wide post Mar 14 19:52:54 GrueMaster: that was for me Mar 14 19:52:56 ah Mar 14 20:00:24 oops Mar 14 20:01:27 kgilmer: you still working for bug? Mar 14 22:12:28 hi prpplague, yep sure am. Mar 14 22:14:23 kgilmer: just ran across a post on linuxdevices, and just reminded me to ask Mar 14 22:15:18 kgilmer: http://www.linuxfordevices.com/c/a/News/Bug-Labs-BugSecure/ Mar 14 22:17:08 yep, that's our next bug version. Mar 14 22:17:45 kgilmer: don't suppose you guys are looking into omap4? Mar 14 22:17:53 the pb chip is pretty cool Mar 14 22:18:09 kgilmer: i was curious about the pb chip Mar 14 22:18:34 kgilmer: i have heard that pb does a lot of cool designs that they keep pretty quiet about Mar 14 22:20:10 sorry prpplague, in support crisis mode for a demo atm, got to get back to you later. Mar 14 22:20:25 kgilmer: np, later bud Mar 14 22:20:33 kgilmer: i understand all too well **** ENDING LOGGING AT Tue Mar 15 02:59:56 2011