=== bjf is now known as bjf-afk === cbrake is now known as cbrake_away [09:06] ogra, NCommander: So livefses are building for dove? [09:06] yes [09:06] lool, indeed, successful spin [09:06] they look fine [09:06] * NCommander isn't sure why he's awake but *shrug* [09:06] I certainly see one now, cool [09:06] imx51 images are still not, because someone thought it would be a good idea to add FLAVOUR=babbage to the boot scripts :P [09:07] You mean in the imx51 specific script? [09:07] lool, do you have an hour or so, I'd *LOVE* to get the d-cd stuff merged [09:07] lool, that's from find-live-filesystems I think [09:07] lool, yeah, indeed i forgot about your change [09:07] i banged my head against that until 11pm last night [09:07] ogra: So the kernel changed the flavour back from babbage to imx51 (which makes sense IMO) and we just need to follow; do I need to do this? [09:08] i have already committed the change in my branch, cjwatson asked me to commit directly in the future ... he never told me where the master branch lives though [09:08] (see #distro, waiting for a response) [09:08] as soon as i know where i will merge it [09:11] lool, i think i will have to postpone the rootstock gui parts, all that cdimage stuff and kernel back and forth have costed to much time [09:11] i'll get the rootstock cli stuff in today though [09:18] lool, so on d-cd ... [09:21] ogra: Hmm I need to remove that DI_FLAVOUR now I think [09:23] lool, why ? [09:23] it might serve a purpose some day [09:25] we should probably make it FLAVOUR=${DI_FLAVOUR} for now [09:25] No it's clutter [09:25] There's no valid reason for using a different flavour name in DI and main kernels [09:25] Except creating work :) [09:25] ok, rip it out then if you feel like [09:33] ogra: Pushed DI_FLAVOUR removal [09:33] thanks again [09:38] ogra, lool so that means imx51 images will spin now? [09:39] NCommander, build is running, lets see [09:39] i dont see why they wouldnt [09:39] ogra, it will be nice to get one image out the door [09:39] NCommander: Looking at your debian-cd changes, could you rename boot.script to boot.scr which I think is the name used for beagle? It avoids using long filenames under vfat [09:39] yes, better would be even two :) [09:39] lool, its boot.scr on the output image, boot.script is the temporary file [09:39] Ok [09:40] unless I made a mistake [09:40] yeah, mkimage spits out the right name [09:40] NCommander: Why do I see the same code twice, one time for alternate installer and another time for live CDs? [09:40] (In the same file) [09:41] lool, its not identical, the file paths of where to load the image are different, as are the bootargs [09:41] make one function and use variables ? [09:41] what ogra said [09:42] That sounds extremely messy vs. four lines of code which are written into a file, then turned into a boot script [09:42] its eight lines if you double the code :) [09:43] ogra, actually, its only two lines since only the bootm line is shared between the two [09:43] All the other lines are different [09:43] NCommander: Come on, you're hardcoding kernel load address multiple times... [09:44] OH [09:44] That code [09:44] sorry [09:44] I was thinking the output boot script [09:44] ARGH ! [09:44] It does have the addresses... [09:44] fatload usb 0:1 0x00200000 /cdrom/uImage [09:44] fatload usb 0:1 0x01100000 /cdrom/uRamdisk [09:44] I realize the pathnames are different [09:44] * ogra wonders why he makes changes to stuff if others re-upload the packages and revert them without looking [09:44] sigh [09:44] lool, I'm not object to doing it, but its going to be even uglier than the way it currently looks [09:45] Since now you need multiple variables for the path, bootargs, mkimage lines, etc. [09:45] NCommander: Ok well I'll do it then [09:46] lool, I don't want to unload work onto you, but maybe your bash foo is better than mine [09:48] lool, actually, i see a second change I need to make for consistantly with d-i, so let me roll that together right now if you don't mind [09:48] if thats the case you should learn from lool :) [09:48] NCommander: Sorry but no, I cant pick it up and at the same time wait for changes frm you [09:48] I think I'll just go out enjoy my leave day [09:48] lool, ack, today is your day off? [09:49] * ogra points NCommander to gcal [09:51] lool, ack, sorry, I didn't know. I can ask ogra to look at it if he's game [09:59] lool, ogra I take it back, it is cleaner this way [10:00] ogra, does this look sane: http://paste.ubuntu.com/259733/ (I can't test it at the moment) [10:03] console=tty1,115200 ??? [10:04] drop FLAVOUR, its not used anywhere [10:05] move the first COMMANDLINE above the "if [ "$CDIMAGE_INSTALL_BASE" = 1 ]; then", make the second COMMANDLINE look like: COMMANDLINE = "${COMMANDLINE} casper" [10:06] also is the video stuff really needed ? [10:06] setenv bootargs $COMMAND_LINE .... remove the underscore [10:09] ogra, rechecking right now if the video stuff is needed with the latest kernels [10:10] your here document writes to boot.scr [10:11] i assume that should go to boot.script [10:17] ogra, no, it should go to boot.scr, (vfat filesystem, you want to make sure its 8.3 compliant) [10:17] ?? [10:17] ogra, the bootscript is written out to boot.scr [10:17] so what about your mkimage call then ? [10:18] -d in the input file [10:18] your here document (the input for mkimage) isnt [10:18] ogra, huh? [10:18] Which document are you referring to? [10:19] cat >> boot.scr < the here document you use [10:19] Oh [10:19] Yeah, thats a mistake [10:19] from moving things around [10:20] * NCommander needs to fix phoenix's power supply so I can test d-cd changes again [10:20] we can test them after merging [10:20] just get the scripts right [10:22] NCommander: I factored boot script generation in a function; I noticed that you use /cdrom pathnames in the cmdline and boot script, but copy the kernel + initrd to /install; is this correct? [10:22] yeah, thats wrong [10:23] http://paste.ubuntu.com/259741/ - current version [10:23] * NCommander looks [10:23] "current"? I can only review a single version at a time [10:23] lool, you said you'd be out so i worked with him on the pastebin version [10:23] lool, you said you were leaving so I refactored it taking in ogra's advice [10:24] NCommander, curly brackets FTW (see my above comment about COMMANDLINE) [10:25] lp:~lool/ubuntu-cdimage/debian-cd_dove-soc has my changes [10:25] I think I said I'd review these debian-cd changes this week [10:25] * ogra steps back then [10:26] lool, he asked me too several times, i didnt know :) [10:26] NCommander: Stop asking everybody [10:26] "This branch has not been pushed to yet. " ... /me twiddles thumbs [10:26] NCommander: You're jumping from reviewer to reviewer all the time; I know you're eager to see that deployed, but we're busy and cant always make time on the spot [10:27] NCommander: You asked cjwatson, ogra and myself; I will look at them, I already pushed some refactoring which you didnt want to do [10:27] ogra: Pushed now [10:27] well, its fine to do so, as long as you notify the former reviewer [10:27] It's fine if everybody agrees [10:27] there might actually be someone who currently has more spare time than others ... [10:28] but the others need to know you asked someone else :) [10:28] Ok I'm really going out for lunch now; I might finish today but more likely tomorrow [10:28] Please be patient... [10:28] lool, feature freeze is less than 24 hours away [10:29] lool, i dont think 115200 for tty0 makes sense [10:31] looks a lot better overall though [10:36] lool, thank you for doing the refactoring (I already took back about refactoring in the backscroll), I made a few slight changes and adjustments based off Oliver's comments, and I removed the video line as we don't need it with the current inarchive kernel (the framebuffer comes up correctly; I wasn't getting video output with that video line on the new kernel) [10:36] lool, lp:~mcasadevall/ubuntu-cdimage/debian-cd_dove-soc [10:39] lool, I was also left with the understanding from david that we need an image before feature freeze, if I'm mistaken, and the due date is alpha 5 itself, then I'm operating off a false assumption to try and get this in. In any case though, I goofed w.r.t. to handling getting this properly revied. [10:40] due date is A5 [10:41] and debian-cd isnt a package ... it wont be affected by FF [10:41] at least what we use on antimony [10:41] ogra, ok, so I was operating under a false assumption [10:41] though if you want A5 images you need to test them before indeed [10:42] The installer isn't going to be fully functional until probably A6 due to the unresolved issue with flash-kernel/update-initramfs [10:43] we just need to make a decision [10:43] nor will the images until the modified u-boot exists [10:43] they should [10:43] ogra, what should? [10:43] you should have a wikipage with a step by step guide to make it work with the board as-is [10:44] the images should [10:44] ogra, oh, images yes, installer no [10:44] who cares :P [10:44] our target is a live image ... [10:44] alternate is a nice to have thing [10:44] ogra, one would think installation of that image to a SATA HDD would be important [10:44] ogra, when I say installer, I mean ubiquity [10:45] yeah, that should be made working indeed [10:45] but you wont be able to make it work without having images [10:45] ogra, I'm not expecting to have ubiquity working until A6 [10:45] right [10:45] lets do one thing at a time [10:45] extactly! [10:45] for A5 all that counts are live images that boot fine [10:46] s/fine//g [10:46] s//fine/ [10:46] :) [10:46] meaning you get a usable desktop [10:46] when booting the thing following the wiki [10:46] Well, when I think of boot fine, I think of stick USB key in, turn power on [10:46] Ok [10:46] so we're operating under different definitions of fine [10:47] i'm working under the definition of doing one step at a time ;) [10:47] I'm starting to think I've reached my limit without sleep [10:47] fine == everything thats possible technically and under the given constraints [10:47] go to sleep then [10:47] fine == what we want the final image to look like [10:48] you are jumping ahead of the "lets do one thing at a time" [10:48] ogra, I'm having insommia again. its why I was up at 4am local time to begin with [11:36] NCommander: Merged your code in the deployed debian-cd [11:37] * lool will look at cdimage bits later === cbrake_away is now known as cbrake [12:31] oh, my latest image has only 695MB [12:48] ogra@babbage2:~$ uname -a [12:48] Linux babbage2 2.6.31-100-imx51 #7-Ubuntu Mon Aug 24 13:59:02 UTC 2009 armv7l GNU/Linux [12:48] \o/ [12:49] I WIN! [12:49] ogra@babbage2:~$ dpkg -l |grep linux-image [12:49] ii linux-image-2.6.31-100-imx51 2.6.31-100.7 Linux kernel image for version 2.6.31 on I.M [12:49] ii linux-image-imx51 2.6.31.100.1 Linux kernel image for imx51 machines [12:49] phew, finally [12:50] GrueMaster, new armel image ready for you :) [12:58] ogra: say, we were talking about beagleboard kernels and stuff recently. if I want to do my own kernel for the bb the ubuntu way, do I build on the ARM box or crossbuild? [12:59] well, if you only have a BB to build on i guess a cross build is a lot faster [13:02] true [13:02] so I need the codesourcery cross-build stuff, anything else? [13:03] http://idlethread.blogspot.com/2009/01/recipe-of-day-cross-compiling-armel.html [13:03] that should help [13:03] (using an ubuntu git tree though, you will need to fiddle a bit there) [13:04] well, I take I'll have to carry over the omap patches and stuff, I might have a look at the sources used to build my current kernel... [13:06] but that blog post looks really helpful, thanks :) [13:06] ipv6tables, here I come :) [13:06] heh === bjf-afk is now known as bjf === bjf is now known as bjf-afk