=== zyga is now known as zyga-afk [00:30] does anyone have a favorite version of ubuntu that works reasonably well on pandaboard? [02:13] MrCurious: natty works fine on Panda. [02:27] infinity: This thing is wonderfully designed. HDMI adaptor puts the cable exactly on top of the power/reset buttons. [02:27] Quickstart, that is. === asac_ is now known as asac [02:35] GrueMaster: I'm no longer sad that I didn't spring for the HDMI module for mine. [02:36] Meh. I have the power of remote control via serial relay (same power as Panda). [02:36] * GrueMaster wishes he could figure out why there is a 1.5s lag in irc. [02:37] Reconnect to a US server? [02:37] It is. Always. [02:37] (I think). [02:37] You're on a Bulgarian server right now. [02:38] hitchcock.freenode.net [Sofia, BG, EU] [02:39] GrueMaster: You might want to use irc.us.freenode.net instead of irc.freenode.net (the latter includes EU in the round-robin) [02:39] Switching. [02:40] Better? [02:41] Am I here? [02:41] Am I here? [02:41] Not sure. [02:41] Gah. 2.5s lag. [02:41] And this is a server in Corvallis, OR. [02:42] Weird. I'm on the same server as you with 0 issues. [02:42] Hunt down the torrenting kids? :) [02:42] They're at the cost. Pandora is the only consumer here. dsl-modem showing low traffic. [02:43] Whoa. desktop loadavg: 5.84. hrm. [02:44] grrr. Firefox (Lucid). [02:45] 25% CPU. [02:46] Must be Jenkins. This is going to be a fugly cycle. [02:46] I blame flash. [02:46] Or Jenkins, sure. ;) [02:46] Nope. Not running flash. [02:46] Even my server is feeling stressed. [02:47] Sigh. Guess I will need to dedicate my spare core2duo w/ 4G ddr2-800 to Jenkins. It requires Natty anyways and most of my infrastructure here is LTS (Lucid). [02:49] I do love that every time someone re-invents the CI wheel, they make it even less efficient in the name of elegance and/or abstraction. [02:50] Cl wheel? [02:50] CI. Continuous Integration. [02:51] Hrm. Bandwidth is good. Not stellar, but 5.6M/750 isn't bad (with pandora running on my chumby). [02:51] infinity : Heh.. yeah [02:51] infinity : I deal with bamboo all the time at work .. Atlassian. It works, but it's clunky [02:52] and it's support for git is very basic [02:54] Quickstart seems to be drawing power from either my USB, ethernet, or hdmi. weird. Not enough to do anything, but enough to make the power LED glow. [02:54] I just seem to be going backwards with CIs, as far as system overhead. [02:55] Tinderbox (perl) -> Buildbot (python) -> Jenkins (java) [02:55] Not that the implementation language means anything, but in this case, the systems seem to match the stereotypes usually leveled at people who use them. [02:56] Buildbot is horrible [02:56] The sooner I have to stop caring about it, the better [02:56] Tinderbox is simple, gets the job done, quick. Buildbot is a bit more modular, a bit more complex, reasonably quick about it. Jenkins is massively complex, and... Everything that goes with it. [02:57] Personally, I don't care what it is running, as long as it WORKS. [02:57] StevenK: I despised buildbot as a tinderbox replacement, but it kinda does the job. And doesn't bring machines to its knees just by existing. [02:57] steven@hudson:~$ uptime [02:57] 01:57:36 up 311 days, 7:04, 2 users, load average: 0.00, 0.00, 0.00 [02:57] My Jenkins VPS ^ [02:58] StevenK: "bring to knees" doesn't always imply downtime. :P [02:58] No, it implies heavy load [02:58] Is it doing anything at all currently? [02:58] Sure, it uses slaves to do the building :-) [02:58] I dunno. I've not used it locally, to be fair. Just heard horror stories from others about CPUs spinning out of control. [02:58] infinity: https://lpci.wedontsleep.org/ if you care [02:59] I understand you may not :-) [02:59] I'll care when someone tells me I have to care about it. [03:00] Haha [03:00] Next you'll tell me I don't count. [03:00] You count when you send me large bags of money. [03:00] Nice. Now to figure out how to use it on my panda farm. [03:00] s/money/Australian lamb/ ? [03:01] Money doesn't spoil in the mail. [03:02] That's a solvable problem [03:14] possible recursive locking detected? anyone encounter the same problem on pandaboard? [03:14] http://paste.pocoo.org/show/429588/ [03:15] I have problem to use pvr driver [03:17] I'm using http://ports.ubuntu.com/ubuntu-ports/pool/main/l/linux-ti-omap4/linux-ti-omap4_2.6.38-1208.11.tar.gz [03:20] Openfree`: Try using our toolchain too... [03:21] Openfree`: No guarantees (in fact, it's pretty unlikely) that Gentoo's 4.5.2 contains all the fixes from recent upstream, Debian, Ubuntu, and Linaro GCC 4.6.x [03:22] Openfree`: (Same argument goes for binutils) [03:30] infinity, so ,you suggest use GCC 4.6.x? it would be to use ubuntu toolchain for gentoo user, you know... so I would like to know which patches I may need, or just upgrade [03:32] Openfree`: I'd suggest using either Oneiric's toolchain or Linaro's (both GCC and binutils), yes. I wouldn't be the man to ask to dissect which patches you might need on top of an upstream checkout. [03:34] infinity, but looking at the log, it shouldn't be the problem of toolchain side.. well I will try 2.6.38-1309.13, then see what happens [03:34] infinity, anyway, thanks for your help [03:35] Ok, I will try to ask at #linaro, see if any helps [05:09] or do i misinterpret [05:10] oops === zyga-afk is now known as zyga [10:10] everytime , I had to call "sudo lxdm" to start desktop session, how do I start automatically at the boot [10:11] make an upstart file? [10:11] woglinde: how do I do it? :(, [10:13] woglinde: is this the file /etc/init.d/rc? [10:14] no [10:14] look into /etc/init [10:14] and learn about how to write an upstart file [10:16] woglinde: is this an useful material to learn? [10:16] woglinde: http://upstart.ubuntu.com/getting-started.html [10:26] one other thing when I start lxde in tv:ntsc mode, nothing is displayed on the screen :( [10:26] which screen? [10:26] your tv? [10:26] kunguz yes for upstart [10:26] and there are already enough files to get the clue at /etc/init [10:27] woglinde: I am using a pico-projector with a limited resolution 848x480 [10:27] ? [10:27] woglinde: I configured the startup thing, I copied a lxde.conf into the /etc/init [10:27] woglinde: it seems to work now [12:15] here is my boot script [12:15] http://sudrap.org/paste/text/15678/ [12:15] I can not see lxdm but I can display the non-gui terminal [12:16] whenever I call sudo lxdm in the terminal, the display shows nothing with this boot script [12:16] but if I modify it to run with dvi then I can display lxdm [12:16] any ideas? [12:17] problem is that the tv has to support the resolution [12:19] woglinde: how do I set a new resolution to lxdm? === zyga is now known as zyga-food [12:26] dont know [12:26] woglinde: thanks anyway, I think I found some way. === hrww is now known as hrw === zyga-food is now known as zyga === prpplague is now known as prpplague^2 [15:06] ogra_: i'm looking at the ac100 marvin kernel [15:07] i'll likely switch to .38 soon with the package [15:07] ogra_: in case i compile a new one, install it on the ac100 but it doesn't boot [15:07] ogra_: how do i recover? [15:07] (currently alpha2 keeps me busy, i plan a new package after that) [15:07] ogra_: i would like to rebase everything on top of 3.0 [15:07] ppisati, there is an SOS image if you hold down "home" on boot and press 1 [15:08] ogra_: ah yes, you told me [15:08] ppisati, we only have graphics drivers against .38 atm [15:08] ogra_: you mean the nvidia-graphics-drivers-tegra? [15:09] from the SOS image, mount your root disk under /root ... bind mount /dev to /root/dev ... chroot into /root, mount proc and sys ... then you can replace kernel and modules to your liking and just run flash-kernel to put them in place [15:09] yes, the nvidia driver [15:09] alberto just tries to make it work for us [15:10] i'm not sure if it would work with 3.0 [15:10] *having* 3.0 is surely not bad ... but i wont defualt to it unless i know the driver works [15:11] ogra_: ok, i'm fine with it being a fallback [15:11] not fallback ... the future ;) [15:11] ogra_: if it works and it's good, ok, else go with what we have [15:12] you might want to dropy by in #ac100 ... thats where the fun happens ;) [15:12] yep [15:12] as soon as i start, i'll be there [15:57] to get output from a server image for pandaboard do i need serial hooked up? [15:57] yes [15:58] dang [16:03] you could hack up the cmdline in boot.scr [16:03] then it would use the console [16:04] brendand: You could also use the netboot-fb image to install. [16:04] yeah ! [16:04] Requires more bandwidth though. [16:04] oh i have a ttyS0, just more setup [16:04] doing it now [16:04] and more patience :) [16:05] Use a straight serial cable. Not null modem. [16:07] err, i never did know how to tell the difference [16:07] is it just both ends have different connections? [16:07] 1 male/1 female? [16:07] Null modem cables have crossed-over pins. [16:08] its about the wiring of the pins [16:08] If you didn't buy or make one specifically for the purpose, you don't have one. [16:08] anyway - is 115200 8N1 okay for settings? [16:09] Yup. [16:11] getting stuff through minicom now [16:11] thanks folks [16:11] was relatively painless [16:11] We like it that way. :P [16:12] :) [16:16] why's the text all weird? [16:16] full of question marks [16:17] Because minicom is a lousy terminal emulator. [16:17] thought so [16:17] recommend? [16:17] screen? [16:18] PuTTY :) [16:18] screen works well. [16:20] Should "just work" with something like "screen /dev/ttyS0" (or whatever your serial port is) [16:21] You could tack a 115200 on the end of that if it seems to be doing the wrong thing. [16:23] last question [16:23] should i be able to do a server install with just wifi? [16:23] or do i need ethernet (which i can get if needed) [16:24] The server image has enough glue to get your wifi going, probably. But ethernet's always less hassle (except for the cable) [16:25] i dont think we have wireless-tools seeded in minimal [16:25] It's minimal and standard. [16:25] Most of the glue should be in standard. [16:26] Although... Maybe not anymore. I might be living in the past. [16:26] yeah [16:26] brendand: Either way, real servers have wires sticking out of them. ;) [16:26] lol [16:26] glorious change of mind [16:27] (wireless-tools are seeded in server, so this goes away when I get the pool on the server image) [16:27] yeah [16:28] though pool doesnt mean installed ... that might need some preseed glue [16:28] We don't want it installed by default anyway, do we? [16:28] Just there, "in case someone wants it". [16:29] well, its would be nice to install it on certain platforms ... [16:29] i.e. the panda ... [16:29] The panda has perfectly functional wired ethernet... [16:29] Maybe if you find something that ONLY has wireless, I'll hear your argument. [16:30] k :) [16:30] It's not like wireless in D-I is any more pleasant an experience. [16:30] Not if WPA is involved (which it usually is these days) [16:30] as long as we use oem-config life is a bit easier ;) [16:31] we could use nm-cli :) [16:31] * ogra_ ducks in case a server guy reads that [16:32] * infinity smacks ogra. [16:33] ICS to the rescue [16:33] Nevermind the other reasons that NM on minimal systems smacks of icky (most of which go away when dbus moves to kernel space some day), but nm-cli is also lacking enough features to be useful. [16:34] there is a fully working cli client [16:34] probably called differently [16:34] nm-cli definitely isn't "fully working", but maybe someone's written something better. [16:34] (I mean, it works, for certain simple use-cases, but it's not "fully working") [16:34] there was something in universe ... not sure it still exists [16:36] cnetworkmanager iirc [16:36] grmbl ... [16:36] * ogra_ curses the nvidia driver ... corrupted fonts again [17:32] ogra_: wifi is being a bloody whore on the transformer :( [17:33] heh [17:33] seems a tegra illness [17:33] ac100 had longstanding probs too [17:34] how did you solve your problems [17:34] I can get the occasional scan to work but never connect [17:35] better drivers [17:35] heh [17:35] porting bits and pieces from compat-wireless etc ... [17:36] ask woglinde and marvin24 ... they did all of that [17:36] what chipset are you doomed with ? [17:37] bcm4329 [17:37] ouch [17:37] at least one that leaves you multiple driver options to tyr [17:38] driver is identical to a working one [17:38] I don't get it [17:40] try playing with the power options [17:40] (i.e. in iwconfig) [17:40] and check for rfkill [17:41] rfkill is only on the bluetooth part of the chipset [17:42] I must be overlooking something small, cause the guy who couldn't even figure out the mountpoints is claiming to have found the problem [21:09] anyone noticed that USB ports are broken on panda with the Alpha2 candidate image? [21:10] brendand, GrueMaster just did, yeah [21:11] GrueMaster - got a bug number? [21:11] Pffft. Who needs usb? [21:11] me? [21:11] Bug #791552 [21:11] Launchpad bug 791552 in linux-ti-omap4 "No USB support on beagle/beagleXM" [Undecided,Fix committed] https://launchpad.net/bugs/791552 [21:11] also ethernet seems broken [21:11] just use a GPIO and morse code [21:11] who needs keyboards [21:12] brendand: eth0 is on the usb bus. [21:12] used to work with natty [21:12] Hey, we have two leds that can be reprogrammed. Those plus a web cam can be engineered to act as a data transmission system. [21:12] ah, the good old times, yeah [21:13] brendand: New gcc exposes this issue. [21:13] natty ... when we were young [21:13] natty kernel fails when built with this gcc. [21:15] Also, the 20110705 server images work (as they have an older kernel). [23:45] gruemaster: whats the gcc version that works for the natty kernel and un-exposes some of the bugs?