**** BEGIN LOGGING AT Fri Jul 08 23:59:56 2005 Jul 09 08:55:19 <[g2]> [cc]smart, NAiL feel like talking about openslug packages for 10-15 minutes ? Jul 09 08:55:35 I always feel like talking ;) Jul 09 08:55:45 <[g2]> others are always welcome to join in :0 Jul 09 08:55:46 <[g2]> :) Jul 09 08:56:43 talking is fun :) just dont know that much on the topic Jul 09 08:56:58 Me neither, but I pretend I do! ;) Jul 09 08:57:07 oh... did I say that out loud? Jul 09 08:57:12 ah, so thats how you should do it.. Jul 09 08:57:26 <[g2]> NAiL, too long ago to remember [cc]smart graciously offered to supply some computer resources for testing openslug Jul 09 08:57:56 wow.. What kinda resources? Jul 09 08:58:43 <[g2]> with the latest monotone switch I think all the pieces are in place where ppl can test packages and possibly report back to a centralized place Jul 09 08:59:11 <[g2]> as not everyone had monotone Jul 09 08:59:20 <[g2]> doh! BK Jul 09 08:59:33 yeah Jul 09 08:59:44 * NAiL finally pushes the stable feed Jul 09 09:00:35 <[cc]smart> in Jul 09 09:00:54 <[g2]> so one possibility would be having something like a LiveCD that ppl just need to boot and run which can pull/build/possibly test packages Jul 09 09:01:39 <[g2]> another is something like nightly builds/tests Jul 09 09:02:27 yeah, I'm thinking of doing daily builds on my build box. Just need more memory, and it can go make itself busy ;) Jul 09 09:02:43 <[cc]smart> my oroginal thought was, that there would be some signal on the net the machine could query and when it's flagged it would start building Jul 09 09:02:57 <[cc]smart> the box is exactly what i have Jul 09 09:03:39 like make CIA tell the buildbox something changed? Jul 09 09:03:44 <[g2]> I see two parts, 1 -- the build box, 2 -- a target Jul 09 09:04:02 <[g2]> actually 3 the mechanism to trigger stuff Jul 09 09:04:20 <[cc]smart> a possibility i thought would be to use ssh Jul 09 09:04:47 <[cc]smart> you can configure a key so it's usage will execute a script, and nothing but that Jul 09 09:04:52 <[g2]> I think just polling a web page every 15-30-60 minutes would work fine Jul 09 09:04:57 <[g2]> it's a simple wget Jul 09 09:05:00 <[cc]smart> and it can return a value Jul 09 09:05:12 <[cc]smart> would also be possible Jul 09 09:05:27 the value doesn't need to be anything but the current head does it? That would tell if something has changed. Jul 09 09:05:29 <[cc]smart> but it should not be a wiki :) Jul 09 09:05:41 <[cc]smart> i thought it shuld be a target Jul 09 09:05:59 <[cc]smart> so what would happen is, it would poll, wherever that is for now and read a target Jul 09 09:06:09 <[cc]smart> if there is a target it would update it's pacakges Jul 09 09:06:15 <[cc]smart> then start building the target read Jul 09 09:06:22 <[cc]smart> and upload the result somwhere Jul 09 09:06:32 <[g2]> nod Jul 09 09:06:37 <[g2]> something like that Jul 09 09:06:53 kindof the way debian builds its stuff Jul 09 09:07:08 <[cc]smart> and that target cold possibly be delivered by a script accessed/run by an ssh session Jul 09 09:07:48 <[g2]> [cc]smart, are you talking just NSLU2 targets or Development hosts also ? Jul 09 09:08:11 <[cc]smart> don't understand Jul 09 09:08:39 <[cc]smart> with target i meant like aisngle bitbake package or a list thereof Jul 09 09:08:43 <[g2]> I think the concept works for both development hosts (the x86 build systems) and NSLU2 targets the device under test Jul 09 09:09:26 <[g2]> Ah.. Ok... I think of a snapshot of the OE repo like a tagged version Jul 09 09:09:40 <[cc]smart> fine just as well Jul 09 09:10:08 <[cc]smart> ah yes Jul 09 09:10:15 <[cc]smart> iirc "make" now builds the whole lot Jul 09 09:10:17 Aha... the plot thickens. Alternatively, I'm starting to grasp what you're really meaning ;) Jul 09 09:10:19 <[g2]> for the target(NSLU2) side one can thing of testing a package via ipkg install foo Jul 09 09:10:35 <[g2]> that's why it's good to talk :) Jul 09 09:11:07 <[g2]> [cc]smart, yes, but we could just bb foo package-index Jul 09 09:11:25 [g2]: yes, I was about to say that Jul 09 09:12:08 <[cc]smart> you know better than me what to do here. what i can offer is just the principle and the HW Jul 09 09:12:19 <[g2]> IMHO I think a full rebuild only needs to happen every several days like 3-4 so twice a week would be plenty Jul 09 09:12:22 <[cc]smart> and some help once the plot is laid Jul 09 09:13:05 <[cc]smart> i don't like the predefined approach, since what people usually want is to be able to make the change visible soon Jul 09 09:13:08 <[cc]smart> which makes sense Jul 09 09:13:20 <[g2]> [cc]smart, THX so much. So we a PC available and an NSLU2 avaiable for maybe 8 hours a day ? Jul 09 09:13:20 <[cc]smart> while at the same time, there are periods when there is not much going on Jul 09 09:13:30 <[cc]smart> so a way to trigger seems to me the right thing to do Jul 09 09:13:44 <[cc]smart> any time, any amount Jul 09 09:14:08 <[cc]smart> it can be stressed whenever needed Jul 09 09:14:32 <[cc]smart> and yes, both systems can be used Jul 09 09:14:43 <[cc]smart> PC & NSLU Jul 09 09:16:06 <[cc]smart> so what's the way to realise it ? Jul 09 09:16:36 * NAiL envisions g2 doing the hard labour ;) Jul 09 09:17:05 <[cc]smart> the PC does have other jobs though, which means it cannot boot off a CD :) Jul 09 09:17:17 [cc]smart: btw, you saw what I said in #nslu2-linux? Jul 09 09:17:19 <[cc]smart> all that is needed will be integrated/installed on that system Jul 09 09:17:22 <[g2]> it doesn't need to boot from CD Jul 09 09:17:23 <[cc]smart> yes, i did Jul 09 09:19:02 <[g2]> I'd think that if we have several systems doing full builds no a staggered basis we'd get nearly full coverage with 3.5 systems Jul 09 09:19:16 3.5 systems? Jul 09 09:19:53 <[g2]> example.. Lets say [cc]smart you and me setup systems that listen for the updates Jul 09 09:19:54 My mirror doesn't have much to do but idle either. It can do builds. Jul 09 09:20:28 I could set one up too if you need Jul 09 09:20:55 <[g2]> So, [cc]smart does a full build on Monday/Thursday, You do Tuesday/Friday/ I do Wedsneday/Saturday and Daka2 does Sunday Jul 09 09:21:09 aaah, yes Jul 09 09:21:14 * NAiL gets it Jul 09 09:21:18 <[g2]> We more ppl, we get even more coverage Jul 09 09:21:26 The more the merrier Jul 09 09:21:57 but.. just building on one/two days? why? Jul 09 09:21:59 <[cc]smart> would there be a way to parallelise instead of serialize, too Jul 09 09:22:21 <[cc]smart> fast & on demand Jul 09 09:22:40 DaKa2: To lessen the load on the box. If one box does full builds every day, it'll barely be usable for anything. Jul 09 09:23:13 I have a few boxes doing basically nothing all the time Jul 09 09:24:23 <[g2]> I think a daily sanity check and report or even twice a day would be plenty coverage for the full build Jul 09 09:25:30 <[g2]> I think we could just check for availability of the sources and probably wouldn't need to pull the source Jul 09 09:25:48 <[cc]smart> ? Jul 09 09:26:21 <[g2]> the full build would wipe the tmp dir and rebuild everything Jul 09 09:26:51 <[g2]> we wouldn't need to wipe the downloads dir, but I do think we should check if the sites are up and the files are available Jul 09 09:27:02 <[g2]> we don't necessarily need to download them Jul 09 09:27:21 <[g2]> also producing md5/sha1 sigs would be helpful Jul 09 09:27:24 hmm.. yes Jul 09 09:27:39 that is a very good idea Jul 09 09:28:12 <[g2]> I think having a slug dedictated to testing 24/7 is very useful Jul 09 09:28:19 <[cc]smart> available Jul 09 09:28:58 <[g2]> nod.. I my case it'll probably be both Jul 09 09:30:49 dual xeon 2,4 with 1gb ram, could that be something? Jul 09 09:31:50 dual-boxes are always a nice resource ;) Jul 09 09:32:13 <[g2]> I think the really important items to come out of this are the promotion of the right sources/fixes to the tagged stable version Jul 09 09:32:13 never seen to load on it go over 0.02 Jul 09 09:32:43 <[g2]> and also great awareness as to what works/needs fixing on the unstable branch Jul 09 09:32:51 yes Jul 09 09:33:20 I'm slowly figuring out which packages are broken on stable. Got fixes for several. But unstable is a moving target ;) Jul 09 09:34:18 basically, upload of build-log for every package somewhere, file bug somewhere if failed? Jul 09 09:34:54 that couldn't be so hard to script? Jul 09 09:35:28 <[g2]> So I see the following progression/status: 1) OE attempt, 2) Builds in OE, 3)Runs on Target with patches, 4)Runs on target no patches, 5)Tested/Production ready Jul 09 09:36:04 <[g2]> so like strace is now a 4 or 5 Jul 09 09:36:14 <[g2]> but mysql or gcc is a 3 Jul 09 09:36:40 <[g2]> and then we've go versioning of indidual packages Jul 09 09:37:04 yeah Jul 09 09:38:01 so something cannot be 5 when it has to be patched? Jul 09 09:38:29 <[g2]> DaKa2, right Jul 09 09:39:12 <[g2]> I'm not saying we don't put 2s 3s and 4s in the stable feed, just we mark them accordingly Jul 09 09:39:30 but some things just have to be patched, even if it works great after, because of dead upstream Jul 09 09:40:05 <[g2]> DaKa2, I'm talking about patches that are not in the OE/ipkg systems Jul 09 09:40:07 having that look bad with a 3 seems a bit.. well.. bad Jul 09 09:40:13 ahh Jul 09 09:40:25 then I see Jul 09 09:41:42 <[g2]> I'd say we start tacking the system builds and testing and then rapidly turn to individual packages Jul 09 09:41:58 <[g2]> but I'm all ears about what you guys think Jul 09 09:42:21 I can't say I've got any more input right now, actually... Jul 09 09:42:48 where would you have the 1-5? on a webpage somewhere? Jul 09 09:43:33 <[cc]smart> sounds good to me Jul 09 09:43:55 <[cc]smart> i'm just wondering how the testing on the slug can be automated Jul 09 09:44:00 <[g2]> DaKa2, nod. I think we can generate automatic feedback for a bunch of stuff Jul 09 09:44:13 <[g2]> [cc]smart, I was just thinking about that :) Jul 09 09:44:24 [g2]: Is it possible to make bitbake output what packages it'd build, and in what order? Jul 09 09:44:41 1 and 2 can be automatically updated by the builds Jul 09 09:44:42 like.. bb --pretend openslug-packages Jul 09 09:45:08 <[g2]> NAiL, I'm missing the point there Jul 09 09:45:24 <[cc]smart> the slug would have to have a possibility added, so that it can be hard reset from the pc Jul 09 09:45:45 <[cc]smart> plus a dedicated NIC maybe so that testlan is available Jul 09 09:45:46 [g2]: So it would be possible to single-step through the build, and do a status for each package thataway. Jul 09 09:46:06 <[g2]> NAiL, bb already produces status like that Jul 09 09:46:23 <[g2]> a simple perl script (I've done one already) produces the results Jul 09 09:46:28 neat Jul 09 09:46:41 You've got it all laid out, haven't you? ;) Jul 09 09:47:04 <[g2]> it's a work in progress and many others add great ideas and resources Jul 09 09:47:08 <[g2]> it's a team effort Jul 09 09:47:27 <[g2]> things just work way faster when ppl work together Jul 09 09:47:32 yup Jul 09 09:47:49 which reminds me... anyone *tested* sane-backends? Jul 09 09:48:00 <[g2]> [cc]smart, I think having the PC being able to control the Slug is a great idea Jul 09 09:48:00 NAiL: I was going to Jul 09 09:48:02 <[cc]smart> on slug ? Jul 09 09:48:09 [cc]smart: yes Jul 09 09:48:37 NAiL: but the slug here at work with my scanner is openslug 2.0 beta Jul 09 09:48:49 ah Jul 09 09:48:51 <[cc]smart> er, no... had two scanners in my life, both scsi Jul 09 09:48:52 I could bring the scanner home to my unstable ofcourse Jul 09 09:49:10 or bring the unstable to work Jul 09 09:49:21 Either way, that'd be great. I'd love to see if it works Jul 09 09:49:27 me too Jul 09 09:49:45 I could just reflash the one at work Jul 09 09:49:58 but it is working... better bring the scanner home Jul 09 09:50:05 yeah Jul 09 09:50:31 should probably get going now anyway, not good to be at work on a saturday at 18:50 when you are on vacation Jul 09 09:50:38 <[g2]> On the test slug I'd think a NFS boot may be the easiest way Jul 09 09:50:39 rofl Jul 09 09:51:11 <[g2]> DaKa2, thx for the help Jul 09 09:51:25 <[cc]smart> NFS root by means of providing the root in second stage, when testing the packages, yes Jul 09 09:51:40 <[cc]smart> first though, you need to get the image on and verify that Jul 09 09:51:46 neat Jul 09 09:52:01 * DaKa2 bb in 35 mins Jul 09 09:52:31 <[cc]smart> for that, hardreset from pc and serial console is needed Jul 09 09:53:11 <[g2]> [cc]smart, I was thinking that after a turnup nfs (with the nfs root being on the development box) would have us in a position for testing Jul 09 09:53:20 <[cc]smart> yes Jul 09 09:53:31 <[cc]smart> getting there is the tricky part :) Jul 09 09:53:43 <[g2]> what do you mean ? Jul 09 09:54:13 <[cc]smart> it should be possible to autmated - hardreset the slug - set it to upograde mode- falsh an image - verify boot Jul 09 09:54:19 <[cc]smart> do turnup Jul 09 09:55:25 <[g2]> Ok... How about a custom Redboot or APEX that allows scripting from the serial terminal Jul 09 09:55:33 <[cc]smart> over serial console and a hardreset thingy which somehow connexted to the pc Jul 09 09:55:54 <[cc]smart> er, i'm cluerupted Jul 09 09:56:10 <[g2]> ??? Jul 09 09:56:34 <[cc]smart> you got to a point where i'm at the edge and fell of the planes of clue Jul 09 09:57:34 <[cc]smart> dunno apex Jul 09 09:58:42 <[cc]smart> prolly the hardreset capability is still needed though Jul 09 09:58:50 <[cc]smart> should be possible though Jul 09 09:58:51 <[g2]> yes Jul 09 09:59:06 <[g2]> and external power box fixes the hard reset Jul 09 09:59:31 <[g2]> and maybe we'll need the "always on" mod Jul 09 10:01:11 <[g2]> or maybe just a little PIC controller with a push button to turn on the box Jul 09 10:01:42 <[g2]> [cc]smart, you've got serial right ? Jul 09 10:03:12 <[cc]smart> that slug for that purpose not yet, but will Jul 09 10:03:25 <[g2]> ok Jul 09 10:03:34 I've got an extra serial cable, if there's a need Jul 09 10:03:40 <[cc]smart> no Jul 09 10:03:44 <[cc]smart> will do it Jul 09 10:03:47 ok Jul 09 10:04:04 I've got two plXXXX-cables, both works :) Jul 09 10:04:06 <[cc]smart> plus MHZ mod while we're at it :) Jul 09 10:04:20 <[g2]> I may run 2 slugs Jul 09 10:04:33 <[g2]> one turbo and one non-turbo Jul 09 10:07:58 <[g2]> Ok... so I think we need a couple things now Jul 09 10:08:21 <[g2]> 1) An updated way to grab the results of the openslug build Jul 09 10:08:44 <[g2]> 2) A way to roll up those results and push them back to the wike/web somewhere Jul 09 10:16:25 <[g2]> I think either apache/php or thttpd/php on a slug would work well Jul 09 10:17:49 There's a bunch of servers available for serving the status pages, aren't there? Jul 09 10:22:36 * NAiL fixes a bunch of outstanding stable issues Jul 09 10:23:57 <[g2]> NAiL, what's the latest with the 2.6.12 kernel ? Jul 09 10:24:58 Haven't heard anything. Jul 09 10:25:54 <[g2]> I'm concerned that there might be flash issues Jul 09 10:26:52 I know. Jul 09 10:27:09 I'm not sure if I should test it for a while Jul 09 10:28:38 It won't touch the other mtdblock, will it? Like redboot? Jul 09 10:29:09 mtdblocks Jul 09 10:30:20 <[g2]> it's a flash corrupting dunno Jul 09 10:30:28 <[g2]> I don't know if it will even happen Jul 09 10:30:47 Exactly... But I don't have jtag, so if it touches RB I'm screwed Jul 09 10:30:58 <[g2]> right Jul 09 10:31:12 <[g2]> that's why I think we should be on the 11.2 kernel Jul 09 10:32:07 yeah.. That's done with PREFERRED_PROVIDER or something, innit? Jul 09 10:32:21 <[g2]> yes Jul 09 10:34:30 Mind changing that? Jul 09 10:34:55 <[g2]> I don't have write access setup on the monotone repo :( Jul 09 10:35:06 you don't? Jul 09 10:35:13 <[g2]> not yet Jul 09 10:35:19 <[g2]> it's been on my "todo" list Jul 09 10:35:54 gimme a key and I'll fix ;) Jul 09 10:36:16 ok, but where does that nice change go? Jul 09 10:36:19 <[g2]> I got to find my key again Jul 09 10:37:36 <[g2]> looks like it may already be there Jul 09 10:37:54 what file? Jul 09 10:38:01 <[g2]> PREFERRED_VERSION_openslug-kernel ?= "2.6.11.2" Jul 09 10:38:12 * NAiL found it the other day, but lost it again ;) Jul 09 10:38:26 <[g2]> onf/distro/openslug.conf Jul 09 10:38:30 <[g2]> conf/distro/openslug.conf Jul 09 10:38:56 yeah Jul 09 10:39:29 does it build openslug-kernel or nslu2-kernel? Jul 09 10:39:46 <[g2]> dunno Jul 09 10:40:04 <[g2]> I thougt last time I built nslu2-kernel at 2.6.12 Jul 09 10:40:36 yeah Jul 09 10:40:47 still does Jul 09 10:41:27 so if I change the preferred provider from openslug-kernel to nslu2-kernel? Jul 09 10:42:23 <[cc]smart> 2.6.12 ? Jul 09 10:42:41 <[cc]smart> i'm currently using 2.6.12.2 Jul 09 10:44:19 <[g2]> [cc]smart, the other day someone in #OE pb_/mickely/?? mentioned jffs2 corruption on an ARM target with 2.6.12 at the mtd level Jul 09 10:44:22 which may, or may not, have a mtd isue Jul 09 10:44:39 <[g2]> I don't know if it applies to us, but better to be safe than sorry Jul 09 10:45:37 man... svn up takes ages Jul 09 10:46:24 Think it was mickeyl, btw Jul 09 10:48:09 <[cc]smart> i see, i don't understand monotone really yet Jul 09 10:48:26 <[cc]smart> i should be able to be in the openslug directory and do montone sync... Jul 09 10:48:27 <[cc]smart> or not Jul 09 10:48:37 <[cc]smart> to make my changes available Jul 09 10:48:50 <[cc]smart> or would i have to be in openembedded directory Jul 09 10:49:28 openembedded dir Jul 09 10:49:38 but you have to commit first Jul 09 10:49:45 Did you have your key added? Jul 09 10:49:52 <[cc]smart> yes Jul 09 10:49:56 good Jul 09 10:50:05 <[cc]smart> and this then merges to nslu openembedded though, correct ? Jul 09 10:50:12 <[cc]smart> i don't want to mixup anything Jul 09 10:50:17 Added the PREFERRED change, should show up real soon now Jul 09 10:50:24 <[cc]smart> and then there is the policy question Jul 09 10:50:30 yes, then propagates to openembedded after a while Jul 09 10:50:32 <[cc]smart> should one be working on aprvate branch or on trunk ?= Jul 09 10:51:21 <[cc]smart> i always get htis: monotone: read from fd 4 (peer monotone.nslu2-linux.org) failed, disconnecting Jul 09 10:52:37 hmm.. the same one I got I think.. Jul 09 10:52:39 rwhitby: ping? Jul 09 10:56:33 NAiL: something strange happened to sane-find-scanners in your build of sane-backends... Jul 09 10:56:43 neat Jul 09 10:56:45 What? Jul 09 10:56:53 relink_command="(cd /home/repvik/unstable/openslug/tmp/work/sane-backends-1.0.15-r0/sane-backends-1.0.15/tools; { Jul 09 10:57:13 even some libgphoto stuff in there.. Jul 09 10:57:17 gah, libtool, innit? Jul 09 10:57:24 some .la file? Jul 09 10:57:30 maybe.. Jul 09 10:57:39 but this is in a sh-script Jul 09 10:57:56 # install mode needs the following variable: Jul 09 10:57:56 notinst_deplibs=' /home/repvik/unstable/openslug/tmp/work/libgphoto2-2.1.5-r3/libgphoto2-2.1.5/libgphoto2_port/libgphoto2_port/libgphoto2_port.la' Jul 09 10:58:04 gah Jul 09 10:58:10 bad .bb Jul 09 10:58:35 ahh... Jul 09 10:58:41 # sane-find-scanner - temporary wrapper script for .libs/sane-find-scanner Jul 09 10:58:41 # Generated by ltmain.sh - GNU libtool 1.5.10 (1.1220.2.131 2004/09/19 12:46:56) Jul 09 10:58:55 libtool :-) Jul 09 10:59:53 yeah Jul 09 11:00:02 Dunno how to fix libtool stuff Jul 09 11:00:12 <[cc]smart> is ldd available in one of the ipkgs ? Jul 09 11:00:26 you put the wrong file in the ipk Jul 09 11:00:36 you should have used the file in .libs/sane-find-scanner Jul 09 11:00:51 I think.. but I don't know bb.. Jul 09 11:01:23 I haven't put anything anywhere ;) Jul 09 11:01:40 well, then you should :) Jul 09 11:02:07 Yeah, I'll look at that later. I think I need to basically do everything manually Jul 09 11:02:18 :( Jul 09 11:03:17 that scanimage -L doesn't even find the scanner might not be good either Jul 09 11:03:25 DaKa2: You might wanna look into do_install() in the .bb file Jul 09 11:03:56 me? learn something? ok, I will Jul 09 11:04:03 haha ;) Jul 09 11:04:18 *very* useful resource: http://oe.handhelds.org/cgi-bin/moin.cgi/bb_20file Jul 09 11:05:58 <[g2]> nod Jul 09 11:06:07 Found it earlier today ;) Jul 09 11:07:01 <[g2]> NAiL, check out http://www.nslu2-linux.org/wiki/OpenSlug/PackageStatus Jul 09 11:09:31 [g2]: nice. Is it possible to move the revision away from the package name? Makes it a tad easier to read Jul 09 11:09:50 <[g2]> NAiL, one sec. check again Jul 09 11:10:49 nice Jul 09 11:11:23 <[g2]> Ok... I'll start by pulling off the r# Jul 09 11:12:05 Cool Jul 09 11:12:28 Rebuilding the stable feed meanwhile Jul 09 11:12:30 ;) Jul 09 11:12:33 [g2]: how about green background on Compiled? :-) Jul 09 11:13:02 <[g2]> DaKa2, I spend an hour or two looking at that Jul 09 11:13:26 <[g2]> meaning how to do colors in PmWiki Jul 09 11:13:36 what, not possible? cant be.. Jul 09 11:13:48 <[g2]> it upcoming revisions Jul 09 11:13:50 <[g2]> in Jul 09 11:13:57 <[g2]> and possibly with HTML tagging Jul 09 11:14:02 oh well... Jul 09 11:18:00 <[g2]> Ok... how about now ? Getting better ? Jul 09 11:18:17 so %green%Complete%% does only work in later versions? Jul 09 11:18:35 <[g2]> DaKa2, try Jul 09 11:19:30 http://www.nslu2-linux.org/wiki/Main/WikiSandbox Jul 09 11:19:37 seems to work... Jul 09 11:20:18 but then, background, I dont know.. Jul 09 11:20:50 <[g2]> Hey ! Jul 09 11:21:32 ahh, should be %bgcolor=green%Complete%% Jul 09 11:21:34 <[g2]> actually I think we want to set the background Jul 09 11:21:55 yes Jul 09 11:22:01 like that *updated* Jul 09 11:22:24 <[g2]> on sec. Jul 09 11:22:30 yes, only brighter green if that's possible Jul 09 11:22:46 <[g2]> is there a lightgreen ? Jul 09 11:22:49 you can use and #20aa10 you want Jul 09 11:23:28 <[g2]> lightgreen looks nice Jul 09 11:23:37 <[g2]> what do you thing ? Jul 09 11:23:59 Good Jul 09 11:24:07 hm. good Jul 09 11:24:56 actually I like #30cc20 better, but its up to you Jul 09 11:25:18 <[g2]> check now Jul 09 11:25:35 thats good Jul 09 11:25:52 Good :) Jul 09 11:26:01 just need a color for Bad build or something Jul 09 11:26:05 <[g2]> let me see what the #30cc20 looks like Jul 09 11:26:13 <[g2]> lightred Jul 09 11:27:03 <[g2]> hmmm no lightred Jul 09 11:27:25 try #cc3020 for lightred Jul 09 11:27:35 or look: http://www.nslu2-linux.org/wiki/Main/WikiSandbox Jul 09 11:27:59 <[g2]> check now Jul 09 11:28:21 fantastic Jul 09 11:28:29 but shouldnt status be 1-5? Jul 09 11:28:44 Hmm, a bit dark that red, innit? Jul 09 11:28:54 <[g2]> well we can also put statges in there Jul 09 11:28:58 <[g2]> like do_xxxxx Jul 09 11:29:12 <[g2]> but maybe on a different page Jul 09 11:29:25 hm.. not on my monitor, but try #dd5544 Jul 09 11:29:31 <[g2]> do you know what yellow is ? Jul 09 11:29:37 Do that on a different page, it'll only confuse "users" Jul 09 11:29:50 <[g2]> NAiL, agreed Jul 09 11:29:57 * NAiL is kinda proud Jul 09 11:29:58 <[g2]> for the dev's page Jul 09 11:30:01 yeah :) Jul 09 11:30:08 For us "leet" guys ;) Jul 09 11:30:26 #ffff00 <-- really yellow Jul 09 11:30:28 <[g2]> s/leet/l337/ Jul 09 11:30:43 s/l337/1337/ Jul 09 11:30:54 <[g2]> just testing you :) Jul 09 11:31:03 There's a pixel difference! :P Jul 09 11:31:10 <[g2]> ok... let me try to pull off the package version Jul 09 11:31:39 #eeee10 <-- good yellow Jul 09 11:32:14 I'm a bit fan of a more orange-ish "inbetween-states" Jul 09 11:32:29 * NAiL is getting picky :P Jul 09 11:32:38 well... lets see then Jul 09 11:32:54 a bit less green then Jul 09 11:34:00 http://www.nslu2-linux.org/wiki/Main/WikiSandbox Jul 09 11:34:05 how about that? Jul 09 11:34:21 That's good Jul 09 11:35:12 %bgcolor=#33cc22%Compiled%% Jul 09 11:35:12 %bgcolor=#eeee11%Half Compiled%% Jul 09 11:35:12 %bgcolor=#eecc11%Half Compiled 2%% Jul 09 11:35:21 %bgcolor=#dd4433%Not Compiled%% Jul 09 11:35:23 <[g2]> Ok check out the page now Jul 09 11:36:44 portmap/e2fsprogs is b0rked, but other than that, it looks really nice Jul 09 11:36:57 thats nice, but.. could you make the cellborders 1px solid black? :-) or maybe not Jul 09 11:37:38 <[g2]> I'm just using borders=1 Jul 09 11:38:06 <[g2]> you seem to know a bunch more about this than me :) Jul 09 11:38:08 hm, can one use css markup in the wiki? Jul 09 11:38:24 <[g2]> DaKa2, are you a web guru ? Jul 09 11:38:41 Im a php-coder with some webdesign skills.. Jul 09 11:39:15 I do everything in php, even shellscripts :-) Jul 09 11:39:47 haha Jul 09 11:40:05 *finally*... mysql is finished compiling Jul 09 11:42:47 <[g2]> DaKa2, excellent Jul 09 11:44:15 but this wiki is a bit limiting Jul 09 11:51:51 I'm considering writing a request for status reports on packages to the mailinglist... I'm not getting much feedback on what's broken (Hopefully that's good) Jul 09 11:52:59 [g2]: well.. (i think) border=1 cellpadding=2 cellspacing=0 looks a bit better atleast, and the top should be ||! Package ||! Version ||! OE Rev level ||! Status || Jul 09 11:55:07 <[g2]> DaKa2, thx look again :) Jul 09 11:55:12 much better Jul 09 11:55:30 Very nice. Good work you two :D Jul 09 11:56:05 <[g2]> it's a team effort Jul 09 11:56:06 would ofcourse like to loose the shadows on the table, but doing that means using AdvancedTables == almost unreadable source for the page Jul 09 11:56:09 :-) Jul 09 11:56:26 Nah, looks good enough for me ;) Jul 09 11:57:11 <[g2]> Ok... Jul 09 11:57:39 main thing is beeing informative, and it is that Jul 09 11:58:00 but add a blank line between "This is snapsnot from a 1 July 05 build by [g2]." and the table Jul 09 11:58:10 <[g2]> DaKa2, sure Jul 09 11:58:10 then I'm happy Jul 09 11:58:47 Wish me luck :-) Jul 09 11:58:58 luck with what? Jul 09 11:59:15 I've set up a maze of Makefiles and stuff, to build and push to the feeds automagically Jul 09 11:59:22 I'm testing it now Jul 09 11:59:27 :) it'll work :-) Jul 09 11:59:38 <[g2]> and if it doesn't... he'll fix it :) Jul 09 11:59:43 haha Jul 09 11:59:45 exactly :-) Jul 09 11:59:47 <[g2]> :) Jul 09 11:59:54 yeah, I'm kinda hoping it'll work the first time around Jul 09 11:59:59 I know half (unstable) works Jul 09 12:00:49 <[g2]> NAiL, are you doing a "time make openslug-build 2.&1 | tee openslug-build-`data`.log ? Jul 09 12:01:10 <[g2]> that should be 2>&1 Jul 09 12:01:46 <[g2]> cat the perl script is Jul 09 12:01:48 no Jul 09 12:02:00 <[g2]> cat manifest.pl Jul 09 12:02:00 <[g2]> #!/usr/local/bin/perl Jul 09 12:02:00 <[g2]> print "\n||border=1 cellpadding=2 cellspacing=0\n"; Jul 09 12:02:00 <[g2]> print "||! Package ||! Version ||! OE Rev level ||! Status ||\n"; Jul 09 12:02:00 <[g2]> while (<>) { Jul 09 12:02:01 <[g2]> if (/NOTE: package (\S+)-([\d|\.]+)-r(\d+):\s+task\s+(\w+):\s+(started|completed)/o) { Jul 09 12:02:03 <[g2]> $packages{"$1|| $2 || $3 "} = 1; Jul 09 12:02:05 <[g2]> } Jul 09 12:02:07 <[g2]> } Jul 09 12:02:09 <[g2]> foreach $i (sort keys %packages) { Jul 09 12:02:11 <[g2]> print "||$i||%bgcolor=lightgreenompiled%||\n"; Jul 09 12:02:15 <[g2]> } Jul 09 12:03:38 <[g2]> that's been running on the slug btw Jul 09 12:03:51 nice :) Jul 09 12:04:10 I'll look into doing that tomorrow. Right now I'm just making sure it works properly. Jul 09 12:04:30 <[g2]> I mean all those updates were cut-n-paste from the slug running that script on the output file Jul 09 12:04:53 * NAiL just got a mail that stable push failed :P Jul 09 12:06:44 <[g2]> Ok.. one more thing... I lose all the -natives Jul 09 12:07:06 NAiL: and... the backends didn't end up in the sane-ipk.. I just do not understand this... Jul 09 12:08:04 DaKa2: ah... Well. Did you file a bug on sane/gphoto? Can't remember if there was any. Just include the info you can remember ;-) Jul 09 12:08:22 I can take a look later. Now I'm preparing to party (carefully) ;) Jul 09 12:08:51 :) Jul 09 12:08:54 <[g2]> Ok... natives are gonn Jul 09 12:09:01 I'll do that later Jul 09 12:09:16 <[g2]> I'll do a pull and full build then put up the results Jul 09 12:09:23 <[g2]> in many hours :) Jul 09 12:10:03 Everything in unstable and stable should build, at the moment :P Jul 09 12:21:27 [g2]: http://projects.thestuffguy.com/Repvik/ToDo <-- It's getting better :D Jul 09 12:23:32 <[g2]> NAiL, Cool... you may want to mark that date on the done sutff Jul 09 12:23:33 <[cc]smart> nice list Jul 09 12:23:37 <[cc]smart> i'm working on cyrus Jul 09 12:23:49 <[cc]smart> it seems to me sasl is not ok so far Jul 09 12:23:53 NAiL: btw, gphoto2 works Jul 09 12:24:00 <[cc]smart> which seems to be due to directories not set up Jul 09 12:24:04 DaKa2: sure? Even for getting pics of the cam? Jul 09 12:24:05 <[cc]smart> but might be wrong Jul 09 12:24:07 <[cc]smart> will see Jul 09 12:24:16 [cc]smart: cool, keep me posted ;) Jul 09 12:24:19 <[cc]smart> i'll continue on that tomorroe Jul 09 12:24:52 <[g2]> [cc]smart, did you see http://www.nslu2-linux.org/wiki/OpenSlug/PackageStatus Jul 09 12:25:00 <[cc]smart> maybe somebody knows the package where ldd is in ? Jul 09 12:25:02 Detected a 'Canon:PowerShot S410 Digital ELPH (normal mode)'. Jul 09 12:25:02 Downloading 'IMG_0365.JPG' from folder '/DCIM/103CANON'... Jul 09 12:25:02 Saving file as IMG_0365.JPG Jul 09 12:25:12 * NAiL hugs DaKa2 Jul 09 12:25:29 What'd you have to do to make it work? Jul 09 12:25:32 guess what was needed? installing rev2 of libusb... Jul 09 12:25:37 I was so shure I did that Jul 09 12:25:49 had the old one installed :-( Jul 09 12:25:49 hahahah Jul 09 12:25:56 What about the cam libs? Jul 09 12:26:04 yeah, the bb needs fixing Jul 09 12:26:11 ok Jul 09 12:26:15 Im trying to figure out how right now Jul 09 12:26:24 dammit... *** Error (-3: 'Out of memory') *** Jul 09 12:26:45 <[cc]smart> btw... somebody workin on girlfriend.bb ? just broke mine... Jul 09 12:26:55 seems to fetch stuff into main memory.. should use some swap on this thing... Jul 09 12:27:03 [cc]smart: :-( Jul 09 12:27:04 [cc]smart: sorry, no can do. I've got three revs, all broken here Jul 09 12:30:24 <[g2]> OK.. rm -rf tmp done and the 09-July OpenSlug packages builds is kicked off Jul 09 12:30:34 nice Jul 09 12:30:57 <[g2]> in several hours I'll run the script and post the results on the wiki Jul 09 12:31:03 <[g2]> probably 4-5 Jul 09 12:31:53 <[g2]> [cc]smart, DaKa2 THX! Jul 09 12:31:55 <[cc]smart> and let us know when it's ready for scheduled execution Jul 09 12:32:37 let us know, and Ill set it up here Jul 09 12:32:51 <[g2]> [cc]smart, will do Jul 09 12:33:21 And I'll set it up on this box Jul 09 12:34:19 (when it's ready) Jul 09 12:34:25 <[g2]> NAiL, I think we should package up the native built perl 5.8.7 and Python 2.4.1 and put them in the stable release Jul 09 12:34:57 <[g2]> I can give you the recipe and will have to look at the install for creating the ipkg Jul 09 12:34:59 [g2]: yeah, as noted in my ToDo. Dunno how to package it up thouggh. Jul 09 12:35:30 <[g2]> have you played with udev ? Jul 09 12:35:45 <[g2]> DOH! unionfs Jul 09 12:35:52 nope Jul 09 12:36:03 <[g2]> is that nope on the unioinfs ? Jul 09 12:36:04 I have a faint idea of how it owrks Jul 09 12:36:08 yeah Jul 09 12:36:23 <[g2]> well be able to figure it out Jul 09 12:36:47 <[g2]> because should only be a simple copy of a bunch of files Jul 09 12:37:31 <[g2]> we should talk to jacques and perlguru about too Jul 09 12:37:50 <[g2]> and we can start the working/patched/non-working packages stuff Jul 09 12:38:08 <[g2]> well GREAT Work guys! Jul 09 12:38:15 <[g2]> thx for *all* the help Jul 09 12:38:26 np :) Jul 09 12:38:41 <[g2]> see you guys later Jul 09 12:40:03 hm.. sane rebuilding.. maybe fixed the "copy wrong and not all files needed" Jul 09 12:40:14 cool Jul 09 12:40:14 probably not Jul 09 12:40:19 this thing is strange Jul 09 12:40:34 I just tried adding: Jul 09 12:40:34 do_install () { Jul 09 12:40:34 oe_runmake 'DESTDIR=${D}' install Jul 09 12:40:34 } Jul 09 12:41:19 wtf... "NOTE: the following files were installed but not shipped in any package:" and then lists all the files I want.. Jul 09 12:41:57 HOW DO I SHIP THEM IN A PACKAGE.. but no.. it couldnt tell me that.. just that it isn't shipped Jul 09 12:42:04 then fix that in do_package () ;) Jul 09 12:42:15 hm... Jul 09 12:42:51 Don't know how to fix it as of yet. And I'm not gonna start any new projects today. Jul 09 12:43:55 hm.. it creates sane-backends.package in work/sane.../install .. but from what? oh well.. back to the docs Jul 09 12:44:14 * NAiL goes off to "party" Jul 09 12:44:20 :) have fun Jul 09 12:44:48 Oh yeah :) Jul 09 14:40:39 hm.. a much nicer sane package made.. http://david.thg.se/saker/sane-backends_1.0.15.bb, that actually works, only thing missing is some preconfiguration of saned Jul 09 14:41:06 next up, gphoto2 Jul 09 14:54:22 there, put in your .bb Jul 09 14:54:41 :) Jul 09 14:54:47 did you look at it? Jul 09 14:54:54 yeah Jul 09 14:55:02 I dont know if what I did is allowed, but it works Jul 09 14:55:50 Looks right by me Jul 09 14:55:55 good Jul 09 14:56:07 I'm no bb guru though :P Jul 09 14:56:20 libgphoto2 on the way, just have to testbuild Jul 09 14:57:19 great work :) Jul 09 14:57:50 :) Jul 09 15:02:58 dammit.. Jul 09 15:03:08 there died the connection to work.. Jul 09 15:03:33 DaKa2 should die soon... Jul 09 15:04:16 dammit.. was going to point to http://david.thg.se/saker/libgphoto2_2.1.6.bb but I kindof guess thats dead Jul 09 15:20:20 oh well.. the stupid idiots taking care of that network will probably notice that a large chunk of stockholm city network is down in a week or two Jul 09 15:22:07 on the happy side, sane works perfect, gphoto2 works perfect (with 128mb nfs swap) Jul 09 15:23:31 cool Jul 09 15:23:37 I'll look at it tomorrow Jul 09 15:23:44 Now I'm off to town :D Jul 09 15:23:44 sane works now? Jul 09 15:23:59 Ill post the updated bb-file for libgphoto somewhere else Jul 09 15:24:10 dyoung-zzzz: according to legend, yes ;) Jul 09 15:24:13 wow, if sane works now, I can actually use this USB Camera for something. Jul 09 15:24:45 I made pwc and spca5xx drivers work with no way to grab the images. :-) Jul 09 15:24:51 yeah, I'm hoping on getting my IXUS to work with gphoto2 Jul 09 15:24:52 other than UBE. Jul 09 15:24:56 or, NAiL, in the libgphoto2 bb, bump revision and add to the end: Jul 09 15:24:56 PACKAGES =+ "libgphoto2-camlibs" Jul 09 15:24:57 FILES_libgphoto2-camlibs = "/usr/lib/gphoto2_port/ /usr/lib/gphoto2/" Jul 09 15:24:57 RDEPENDS_libgphoto2 = "libgphoto2-camlibs" Jul 09 15:25:05 DaKa-home: ok Jul 09 15:25:09 NAiL: my ixus works with gphoto :-) Jul 09 15:25:56 but you need to have atleast as much ram+swap free as the stuff you are transfering from it Jul 09 15:29:31 in contrast to SANE, that hardly used any ram when scanning an A4 in 600dpi (100mb output) Jul 09 15:38:39 WOHO! Jul 09 15:51:08 <[g2]> DaKa2, NAiL http://www.nslu2-linux.org/wiki/OpenSlug/PackageStatus Jul 09 15:51:17 <[g2]> that's today's build status Jul 09 15:52:04 does ctrlproxy still build? Jul 09 15:53:34 <[g2]> it's not in the list Jul 09 15:53:45 Yeah, I know thats why I asking Jul 09 15:53:52 so I can have someone put it in the list. Jul 09 16:11:00 ah hell Jul 09 16:11:06 its *really* not in the list. Jul 09 16:11:23 guess it didnt make it through the upgrade. Jul 09 16:11:32 <[g2]> dyoung-zzzz, it probably just fell out of openslug packages Jul 09 16:11:42 No, its *really* not there anymore. Jul 09 16:12:01 like not in openembedded/packages . Jul 09 16:12:37 <[g2]> yeah.. you are right about that Jul 09 16:13:06 oh well. Jul 09 16:23:19 <[g2]> rwhitby, did you see all that chatting in here earlier today ? Jul 09 16:28:18 <[g2]> NAiL, you are *so* close with samba Jul 09 16:31:56 [g2]: saw, scanned, and decided that I'll leave it up to you guys :-) Jul 09 16:32:26 It seems that the openslug packages have now officially been taken over by NAiL, and I never have to push to the feed again :-) Jul 09 16:36:01 <[g2]> rwhitby, it's setting up an automated job Jul 09 16:36:03 Ok so I dont mess this up... mt pull; mt add foo; mt push; mt sync; check heads? Jul 09 16:36:26 * [g2] is an still and mt idiot Jul 09 16:36:51 dyoung-zzzz: just use the convenience targets in the master makefile Jul 09 16:37:48 <[g2]> rwhitby since unslung/optware packages are more mature, I don't know of you guys plan to track their status Jul 09 16:37:59 <[g2]> other than "I should just work" Jul 09 16:38:33 [g2]: make the scripts generic enough so that jp30 and jf can use them too, and put up similar pages Jul 09 16:39:12 <[g2]> the script currently just parses the OE ouput Jul 09 16:39:33 <[g2]> which works for the unslung build Jul 09 16:39:43 dyoung-zzzz: make update-openembedded ; cd openembedded ; mt add foo ; mt commit -m "fooo" ; cd .. Jul 09 16:39:44 push master? Jul 09 16:39:57 <[g2]> the optware stuff is a different animal Jul 09 16:40:01 ah that more famliar. Jul 09 16:40:20 dyoung-zzzz: is this for master, or ctrlproxy? Jul 09 16:40:38 ctrlproxy. Jul 09 16:40:39 there are core team members targets at the end too. you can use one of them to push after the commit Jul 09 16:41:08 okay Jul 09 16:41:12 the important thing is to remember to cd into openembedded before the add, otherwise it will be added to the master makefile repo Jul 09 16:41:22 right, I remember that bit well. Jul 09 16:41:57 I guess now that [ccsmart] has write access, I need another section for non-core write access targets. Jul 09 16:47:35 we should probably address that. Jul 09 16:47:55 BTW, we need to come up with a planned schedule for migration to monotone 0.20 Jul 09 16:50:11 <[g2]> dyoung-zzzz, you don't use samba do you ? Jul 09 16:50:42 Not if I can help it.... Jul 09 16:50:57 I might need to for a "special project" though. Jul 09 16:56:28 <[g2]> hey cool I think samba is running on Openslug Jul 09 17:00:09 sigh Jul 09 17:04:34 dyoung-zzzz: we will migrate when vanille.de migrates Jul 09 17:07:25 NAiL: I restarted the server - perhaps that is required after adding [ccsmart]'s key Jul 09 17:10:30 [g2]: Close?! What's not working now? Jul 09 17:11:27 rwhitby: Ok, we'll see if that fixes it. What was the problem when I couldn't write again? Jul 09 17:11:42 IIRC, you had added my key, but that didn't cut it Jul 09 17:12:39 yes, the packages are now pushed. It even sends me mail if it doesn't update/build/push ;-) Jul 09 17:13:33 NAiL: can you add a target to the Master makefile ? Jul 09 17:14:00 to upload to the feed? Jul 09 17:14:09 I would like to have all infrastructure stuff in the Master Makefile (in the core-only section), so that there is no single point of failure by relying on only one person who knows how to do something Jul 09 17:14:17 yes, the upload to the feed Jul 09 17:15:10 there is an upload-openslug-cross target there now. feel free to modify it - it's now yours Jul 09 17:16:11 I'd also like to put the update-mirrors script in the master repo too. Jul 09 17:16:57 I want to make sure that there are no dependencies on stuff which is not in the repos. We should be able to take a new machine, install the master makefile, run some core targets, and have a replica of nudi (including the stuff which manages uploads) Jul 09 17:17:23 Create a top-level "scripts" directory for the misc scripts. Jul 09 17:18:00 <[g2]> hey NAiL ! Jul 09 17:18:12 Note that there should not be any passwords in any of the scripts - our security is via ssh keys Jul 09 17:21:01 [g2]: What's broken with samba? Jul 09 17:21:12 rwhitby: Taking notes, gonna look at it tomorrow. Jul 09 17:21:37 <[g2]> NAiL, actually I don't think anything is broke Jul 09 17:21:41 NAiL: no hurry - just stating medium term goals Jul 09 17:22:10 <[g2]> we just need to copy the /etc/samba/smb.conf.default to /etc/samba/smb.conf Jul 09 17:22:19 <[g2]> then /etc/init.d/samba start works Jul 09 17:22:35 NAiL: Our philosophy has always been that if it's not in a repo, then it doesn't exist (ssh keys excluded). Jul 09 17:22:36 <[g2]> but I don't think I've got the net all configured properly Jul 09 17:23:18 [g2]: Well... Installing a config files doesn't make sense to me with samba. Installing an example does make a bit sense. Jul 09 17:24:10 rwhitby: Ok, no hurry. But I'm not sure how to set it up though.. It's easy enough with the unstable feed, but the stable feed pulls off berlios and builds in a different directory. Jul 09 17:24:23 a config file with home directories served by user passwords makes sense to me Jul 09 17:24:56 yes.... Jul 09 17:24:59 NAiL: it's probably ok at the moment if the stable feed is done differently Jul 09 17:25:16 just make sure that the smb.conf is listed in CONFFILES Jul 09 17:25:45 BTW, I can't build anything on nudi at the moment - bitbake fails: Jul 09 17:26:41 Has bitbake been updated? Jul 09 17:27:03 (source setup-env ; bitbake unslung-packages) Jul 09 17:27:03 Environment set up for Unslung development. Jul 09 17:27:03 Traceback (most recent call last): Jul 09 17:27:03 File "/home/slug/unslung/bitbake//bin/bitbake", line 27, in ? Jul 09 17:27:03 from bb import utils, data, parse, debug, event, fatal Jul 09 17:27:05 ImportError: cannot import name utils Jul 09 17:27:19 using the Master Makefile in /home/slug on nudi Jul 09 17:27:53 go back to rev 269? I *know* that rev works ;) Jul 09 17:28:38 * NAiL notices that the current head is not nail@nslu2-linux.org Jul 09 17:30:22 Gotta fix that ;) Jul 09 17:30:39 272 has problems? Jul 09 17:30:49 272 has problems. Jul 09 17:30:57 :-( Jul 09 17:30:58 278 was closer to working, IIRC Jul 09 17:31:14 dang what are they up to now? Jul 09 17:31:37 Don't know exactly. Fixing stuff I guess... Unfortunately breaking stuff in the process ;-) Jul 09 17:32:08 for a while they were adding tons of new features Jul 09 17:32:30 They started breaking it two days ago, I think Jul 09 17:32:44 Then there was a bunch of revisions to fix stuff Jul 09 17:33:02 I was at 269 at that time. I think I got stuff working again with 278 Jul 09 17:33:35 * [g2] notes the BB version should be added to the package manager status script Jul 09 17:33:45 yes, not a bad idea Jul 09 17:34:02 There... now the head is back to nail@nslu2-linux.org. Order is restored :P Jul 09 17:34:32 ok, can someone update the master makefile to pull a bitbake from svn that works? Jul 09 17:34:33 <[g2]> I ran with 270 Jul 09 17:34:47 rwhitby: fair nuff Jul 09 17:35:01 <[g2]> which is awfully close to 269 Jul 09 17:36:36 <[g2]> http://www.nslu2-linux.org/wiki/OpenSlug/PackageStatus Jul 09 17:39:25 sigh Jul 09 17:39:31 what is the magic to checkout of svn with key? Jul 09 17:39:41 svh+ssh://user@ ...? Jul 09 17:39:43 svn+ssh Jul 09 17:39:59 and /svnroot/repos/openslug I think Jul 09 17:40:17 [g2]: oh well... I set it to 269 Jul 09 17:40:23 <[g2]> that's fine Jul 09 17:40:38 NAiL: push-master is the target to hit after you commit that change Jul 09 17:40:59 actually, publish-master, cause that puts it up for wget too Jul 09 17:41:51 Nail, what is the URL for the symlinks? Jul 09 17:41:53 the weird thing is that bitbake works on my local machine, but is broken on nudi Jul 09 17:42:00 can we remove the installed version on nudi? Jul 09 17:42:33 dyoung-zzzz: openslug/trunk/openslug/nslu2-linux/packages$ Jul 09 17:42:34 ? Jul 09 17:42:35 <[g2]> python dep ? Jul 09 17:42:44 dyoung-zzzz: can you add setup-foo-developer targets and a top-level setup-developer target in the Master Makefile, which does all that stuff correctly? Jul 09 17:43:30 rwhitby: Well... publish-master tries to push with my local username, which is repvik. My sf.net username is orepvik :( Jul 09 17:43:32 so developers would type "make setup-developer" instead of "make setup" and it would check stuff out with write access Jul 09 17:43:58 Yeah, if I ever figure out how to push. Jul 09 17:44:07 been having pebcak today Jul 09 17:45:17 So we need some environment variable for sf.net username, which defaults to username but can be overridden if it is in the environment already. Jul 09 17:45:30 and berlios_user Jul 09 17:45:35 yeah Jul 09 17:47:04 NAiL, what is the complete URL you're using? Jul 09 17:47:18 svh+ssh://nail@svn.berlios.de/ .... ? Jul 09 17:48:57 svn+ssh://repvik@svn.berlios.de/svnroot/repos/openslug/trunk/openslug/nslu2-linux/packages, I think ;) Jul 09 17:49:07 for only the symlinks, that is Jul 09 17:49:51 we need to put out a bounty of USD$100 for someone to get symlinks working in monotone Jul 09 17:50:11 I should actually write that down somewhere, because I have to figure it out every time Jul 09 17:50:30 NAiL: that's what the master makefile is for - write it in there Jul 09 17:51:20 then everyone can stop working it out each time Jul 09 18:02:49 we'll need a update-foo-developer too. Jul 09 18:05:07 is 'make update-openembedded' supposed to work? Jul 09 18:09:14 <[g2]> NAiL, is there a way to check the smb.conf locally ? Jul 09 18:09:24 check it? like testparm? Jul 09 18:10:09 <[g2]> you are just *full* of useful info :) Jul 09 18:10:30 <[g2]> I think I need to change my workgroup name Jul 09 18:11:11 holy bejesus Jul 09 18:11:48 jacques: yes Jul 09 18:11:55 what on earth is ediff, and how on earth does it work? Jul 09 18:12:40 dyoung-zzzz: most repos remember, so update doesn't need to be different, does it ? Jul 09 18:13:15 ok, then it's not working for me because of my monotone version Jul 09 18:17:23 <[g2]> NAiL, I"m getting *really* close now :) Jul 09 18:18:06 <[g2]> My XP box the kids PC can see the samba server (but I'm not really sharing stuff yet) Jul 09 18:19:16 good :) Jul 09 18:19:16 NOTE: preferred version 2.6.11.2 of nslu2-kernel not available Jul 09 18:19:23 not so good :P Jul 09 18:20:29 <[g2]> 2.6.12 built for me Jul 09 18:21:10 NAiL, you would need to change the provider from nslu2 to openslug . Jul 09 18:21:47 dyoung-zzzz: well... PREFERRED_VERSION_openslug_kernel is already set to 2.6.11.2, but it still builds 2.6.12.2 Jul 09 18:22:05 and it builds nslu2-kernel, not openslug-kernel, iirc Jul 09 18:22:15 Right. Jul 09 18:22:25 virtual/kernel? Jul 09 18:22:30 right, the line above those is a prefrered provider line. Jul 09 18:22:41 saying the preffered provider is nslu2-kernel. Jul 09 18:23:14 What's the difference between nslu2-kernel and openslug-kernel? Jul 09 18:23:44 the former should be removed, and the latter should be used Jul 09 18:24:16 jbowler was using the former as a testbed for 2.6.12 Jul 09 18:24:46 but that should be done using oe directives to make it non-preferred for testing Jul 09 18:25:42 the nslu2-kernel files are rather extensive. Jul 09 18:26:42 or maybe nslu2-kernel is a dir for common files between unslung and openslug, but never built as a kernel in it's own right Jul 09 18:26:50 I hope you don't mean remove the files from oe Jul 09 18:28:01 since there are two kernels for nslu2, then nslu2-kernel is confusing Jul 09 18:28:19 nslu2-kernel is 2.6.12 stuff only, iirc. Jul 09 18:28:49 so that should be in a 2.6.12 subdir of openslug-kernel Jul 09 18:29:35 I agree its very confusing, and I did not spend the time to look at it in detail. I just know theres a lot of stuff in there. Jul 09 18:29:48 oe has provisions for new non-preferred versions - we should use it properly Jul 09 18:30:20 (unless it's common between openslug and unslung kernels) Jul 09 18:35:45 jacques: gphoto2 works :) Jul 09 18:35:45 Downloading 'IMG_1893.JPG' from folder '/DCIM/118CANON'... Jul 09 18:35:45 Saving file as IMG_1893.JPG Jul 09 18:36:21 NAiL, yay!!!!!!!!!! Jul 09 18:36:32 Indeed :D Jul 09 18:36:49 I just started an xorg emerge, but after that I will try it Jul 09 18:37:17 NAiL, what I'm really intersted in is the camera control Jul 09 18:37:27 being able to take pictures remotely Jul 09 18:37:49 ah Jul 09 18:38:02 <[g2]> to look at the birds in the woods right ? :) Jul 09 18:38:03 Dunno if that works or not. But it fetches my pics just fine though :) Jul 09 18:38:51 that alone is a big improvement Jul 09 18:38:58 it used to segfault for me immediately Jul 09 18:39:18 OMG Jul 09 18:39:27 It actually can take pics remotely with my IXUS Jul 09 18:39:35 ??? Jul 09 18:39:41 you just did it? Jul 09 18:39:44 yup Jul 09 18:39:57 which driver did you use? there are two which may work Jul 09 18:40:12 canon and another which I forget Jul 09 18:40:27 haha, ultra-cool Jul 09 18:40:27 New file is in location /DCIM/119CANON/IMG_1917.JPG on the camera Jul 09 18:40:50 NAiL, can you please document this on the wiki? ;-) Jul 09 18:41:50 <[g2]> that sounds about right Jul 09 18:42:19 Woo Jul 09 18:42:30 This is just *SO* cool Jul 09 18:42:34 <[g2]> hey NAiL I'm trying to log in with smaba from the XP box but it doesn't seem to works Jul 09 18:42:58 [g2]: Does the logs say anything? Jul 09 18:43:04 <[g2]> nope Jul 09 18:43:49 <[g2]> I get the user login prompt and I try to log in as normal user but it just shows up again Jul 09 18:44:10 uhm... have you added that user to smbpasswd? smbpasswd -a username Jul 09 18:44:11 <[g2]> I changed the smb.conf and restart with Jul 09 18:44:22 <[g2]> no Jul 09 18:44:26 Do that ;) Jul 09 18:44:59 It doesn't look up the users from /etc/passwd Jul 09 18:45:21 <[g2]> I'm aware of that *now* :) Jul 09 18:45:26 <[g2]> thx Jul 09 18:45:29 np Jul 09 18:45:47 I often forget that when installing samba :P Jul 09 18:45:55 <[g2]> WHOA... Jul 09 18:46:01 <[g2]> I'm in Jul 09 18:48:27 jacques: There's not exactly much to document. ipkg install gphoto2 ;) Jul 09 18:48:59 well.. atleast after the feed has been pushed again Jul 09 18:49:04 NAiL, what commands are you using to take pictures? Jul 09 18:49:22 how did you configure gphoto? Jul 09 18:49:38 <[g2]> NAiL, CONGRATS on the samba stuff Jul 09 18:49:40 <[g2]> very nice Jul 09 18:49:51 gphoto2 --capture-image Jul 09 18:49:54 [g2]: thx Jul 09 18:50:18 jacques: No configuration whatsoever. Just ipkg install, gphoto2 --capture-image and the camera switched on :D Jul 09 18:51:20 well.. "switched on" is an exacceration. It switched from transfer-pictures-mode to take-pictures mode and back. Jul 09 18:51:35 Now that it has actually switched itself off, I can't take pictures anymore :( Jul 09 18:51:53 is there gphoto2 --turn-the-camera-on or something? Jul 09 18:52:33 dyoung-zzzz: Doesn't look like it, no Jul 09 18:52:38 One strange thing though... Jul 09 18:52:38 Capture choices : Jul 09 18:52:39 : Capture not supported by the driver Jul 09 18:53:07 Capture not supported by the driver? I can capture just fine! :) Jul 09 18:53:40 <[g2]> NAiL, I'm getting like 5.8MB xfer with samba is that good/normal/bad ? Jul 09 18:53:56 That's rather good, I'd say :) Jul 09 18:54:18 well... with openslug it's normal, afaik Jul 09 18:54:30 That's what I'm getting Jul 09 18:54:36 <[g2]> I just pasted a 188.6MB file and it took like 32 seconds Jul 09 18:54:37 [g2] finally gets to use the slug for what it was intended by the manufacturer ... Jul 09 18:54:49 <[g2]> it's only been a year :) Jul 09 18:55:16 <[g2]> the *real* problem is I went Windowsless 2.5 years ago Jul 09 18:55:30 haha Jul 09 18:56:10 that's a problem? Jul 09 18:56:18 <[g2]> so I'm VNC'd into the kids PC testing virtually to the turbofat slug Jul 09 18:56:24 I thought that was a solution ... Jul 09 18:56:35 <[g2]> it has been for me :) Jul 09 18:57:08 <[g2]> but now I'll be able to start "communicating" with all my non-geek friends :) Jul 09 18:58:31 <[g2]> oh... just plug this device in and schedule a task to backup to this folder here Jul 09 18:59:47 <[g2]> so OpenSlug's got samba and nfs Jul 09 19:00:06 <[g2]> along with lots of other packages like gphoto Jul 09 19:00:54 sane-backends Jul 09 19:01:06 It's quickly becoming very useful :D Jul 09 19:01:16 <[g2]> OpenSlug ? Jul 09 19:01:33 yes Jul 09 19:01:51 <[g2]> I think it's that new OpenSlug package MGR :) Jul 09 19:02:01 <[g2]> I hear he's an ass-kicker Jul 09 19:02:15 yup Jul 09 19:02:15 I expect being able to connect my cam to my slug tomorrow, and have it automagically transfer all pictures to a folder in a shared folder. Jul 09 19:02:23 <[g2]> NAiL is THE MAN! Jul 09 19:02:37 Respect my authoritae! :P Jul 09 19:02:55 <[g2]> that's the kinda stuff that my wife would like Jul 09 19:03:46 <[g2]> now we just need to add some stuff to the LEDs and will be all set Jul 09 19:04:03 ah... that's a brilliant idea Jul 09 19:04:13 <[g2]> and setup hotplug Jul 09 19:04:28 plugging in the cam, watch the led blink, see that it finishes... Jul 09 19:04:47 <[g2]> pa-zactly Jul 09 19:04:57 dyoung-zzzz: were you going to send that disk btw? Jul 09 19:05:24 I hate to say it but the only thing keeping openslug from mass usage is a web interface Jul 09 19:06:09 http://webtools.wl500g.info/ ? Jul 09 19:06:18 yep Jul 09 19:06:42 <-- got it all figured out. Jul 09 19:06:57 (in a *really* rough sketch) :P Jul 09 19:07:14 NAiL: have you contacted Styno? Jul 09 19:07:54 Styno? no. Jul 09 19:08:10 Have I forgotten something again? Jul 09 19:08:33 he's the webtools dude, and he's waiting for us to give him some feedback Jul 09 19:09:05 oh, ah, ok :) Jul 09 19:09:42 I haven't gotten around to doing the webinterface yet, so it'll take a while. But it's a feature I really want :) Jul 09 19:09:53 dyoung-zzzz: does ctrlproxy work, and are you still asleep? Jul 09 19:10:01 Okay, I think I'm back in a state of sorta being able to commit stuff. Jul 09 19:10:40 * NAiL notices the head has been changed from nail@nslu2-linux.org *again* Jul 09 19:10:47 it builds, but theres something wrong with the source file's trailer, so you'll need to touch tmp/stamps/ctrlproxy-2.6.2-r0.do_unpack first. Jul 09 19:10:48 My head! ;) Jul 09 19:10:55 NAiL has head envy Jul 09 19:11:12 So its not ready for the package manager yet, but it sorta works. Jul 09 19:11:17 for now. Jul 09 19:11:21 Fix later. Jul 09 19:11:45 I like being a-head Jul 09 19:11:52 a what-head? Jul 09 19:11:58 It's funny. Laugh! ;) Jul 09 19:12:27 I work with a guy who's name is Richard Brain Jul 09 19:12:48 nice Jul 09 19:12:48 NAiL-head is ahead of us in monotone commits. Jul 09 19:13:10 Me must never heard the end of it. Jul 09 19:13:12 poor guy. Jul 09 19:14:36 dyoung-zzzz: NAiL is at 3.2 hours. I'm down to 5.21 hours, but over a longer sustained period. We'll see if NAiL is just a flash-in-the-pan .... Jul 09 19:15:09 BTW, should we filter out org.openembedded. stuff from our CIA messages in #nslu2-linux? Jul 09 19:15:10 rwhitby: It's just growing pains. It'll settle after a while Jul 09 19:15:41 Filtering them is good by me, I read most in #oe anyway Jul 09 19:16:30 done Jul 09 19:16:52 (and it makes my commits show more, since there is less clutter! ;)) Jul 09 19:17:29 I wonder how many other projects use four different CM systems .... http://cia.navi.cx/stats/project/nslu2-linux Jul 09 19:18:42 I should set up an RSS feed on the Treo from cia ... Jul 09 19:19:17 heh... like I'm gonna keep up >41 commits a week :P Jul 09 19:20:30 rwhitby: btw, why doesn't that page mention the svn commits I've made (Like the ones today?) Jul 09 19:21:39 I'm at a whopping 2.89 *days* Jul 09 19:22:42 aha Jul 09 19:22:51 that's because they're done by "repvik", not "nail" Jul 09 19:23:29 yep, you're diluting your own stats by having a split net peronality Jul 09 19:24:10 *sigh* Jul 09 19:24:20 I'll have to do the maths myself... Jul 09 19:24:24 Oh well Jul 09 19:25:58 I'm going to sleep now Jul 09 19:26:45 rwhitby, how do I get something frmo the environment in a Makefile ? Jul 09 19:27:07 it makes my head spin Jul 09 19:29:33 ${FOO}, where foo is not set in the makefile. Jul 09 19:31:47 FOO ?= override Jul 09 19:32:17 http://theory.uwinnipeg.ca/localfiles/infofiles/make/make_64.html Jul 09 19:32:29 <[g2]> re: OpenSlug web intf.... between my of apache/ssl/php5 and Daka the web guys we might just get the killer Web UI Jul 09 19:32:53 <[g2]> s/my/my build Jul 09 19:33:50 My Makefile mojo bites. Jul 09 19:34:03 I can check it whats there now if someone can fix my error(s). Jul 09 19:36:16 Okay, I just pushed that makefile, can someone look at it please Jul 09 19:43:25 damn monotone takes a long time to verify even a minor update Jul 09 19:44:11 [g2]: I'd like it to be possible to use thttpd as well.. It has a lower footprint, doesn't it? Jul 09 20:25:39 <[g2]> NAiL, yeah Jul 09 20:26:20 <[g2]> I think we'd have to tunnel the ssl though Jul 09 20:35:08 <[g2]> nite all Jul 09 20:35:15 nitey Jul 09 21:13:13 why is pcmcia-cs in openslug-pacakges ? Jul 09 21:14:54 bluez-utils seems to have a dependency Jul 09 21:19:39 I can't imagine pcmcia-cs doing much of anything on a slug Jul 09 21:25:03 ERROR: see log in /home/packages/nslu2/mtn/openslug/tmp/work/libgphoto2-2.1.6-r4/temp/log.do_compile.15070 Jul 09 21:25:03 NOTE: Task failed: /home/packages/nslu2/mtn/openslug/tmp/work/libgphoto2-2.1.6-r4/temp/log.do_compile.15070 Jul 09 21:25:03 NOTE: package libgphoto2-2.1.6-r4: task do_compile: failed Jul 09 21:25:06 dammit Jul 09 21:25:50 /usr/lib/libjpeg.so: could not read symbols: File in wrong format Jul 09 21:25:55 this actually worked for someone? Jul 09 21:27:15 NAiL, what did I do wrong? Jul 09 21:27:36 I'm giving it a go now. Jul 09 21:27:58 also I thought there was a patch needed for libgphoto2 Jul 09 21:29:47 dammit probably monotone is screwed up somehow with multiple heads Jul 09 21:31:58 It compiled for me. Jul 09 21:32:46 great Jul 09 21:32:53 when did you update? Jul 09 21:33:03 you do a clean build? tm tmp ? Jul 09 21:33:07 rm tmp? Jul 09 21:33:07 I du nno, maybe 90min ago? Jul 09 21:33:20 clean build? Jul 09 21:33:31 ??? Jul 09 21:33:32 I pulled shortly before pushing the top-level makefile change. Jul 09 21:33:39 I didnt clean build... Jul 09 21:34:10 which version did it build? Jul 09 21:34:44 md5sum openembedded/packages/libgphoto2/libgphoto2_2.1.6.bb Jul 09 21:37:55 which version of bitbake? Jul 09 21:38:07 e2b186910101e00d34520b8a4d2eafcd ../openembedded/packages/libgphoto2/libgphoto2_2.1.6.bb Jul 09 21:38:31 bitbake version 1.3.2 Jul 09 21:38:40 BitBake Build Tool Core version 1.3.0, bitbake version 1.3.2 Jul 09 21:39:10 I have no idea why it's not working for me Jul 09 21:39:52 * dyoung dunnos Jul 09 21:39:58 dunno if that's enough bitbake info - what does svn info in the bitbake dir say? Jul 09 21:40:03 do you have some funky envirnment vars set or something? Jul 09 21:40:19 279 Jul 09 21:40:42 I have 272 Jul 09 21:40:48 but it id's same as yours Jul 09 21:40:53 weird. Jul 09 21:41:06 I dunno why I have 279 either, since nail locked it to 269 earlier. Jul 09 21:41:26 ./bitbake/bin/bitbake --version Jul 09 21:41:26 BitBake Build Tool Core version 1.3.0, bitbake version 1.3.2 Jul 09 21:42:08 time to flail around Jul 09 21:42:13 updagint bitbake Jul 09 21:42:22 ok Jul 09 21:42:43 well, we know that 269 and 279 work, so maybe try one of those... Jul 09 21:43:36 which version of libtool do you have in tmp/work? Jul 09 21:43:46 (and cross) Jul 09 21:44:05 libtool-1.5.10-r2 libtool-cross-1.5.10-r2 libtool-native-1.5.10-r1 Jul 09 21:44:37 same versions as mine for cross and native Jul 09 21:45:32 I dont have libtool anything in tmp/cross/* Jul 09 21:46:52 this could be anything Jul 09 21:46:56 including a libtool issue Jul 09 21:48:53 I did another make openslug-update but it claimed there was nothing to get Jul 09 21:50:02 failed again Jul 09 21:50:58 this is actually in the link line: /usr/lib/libjpeg.so Jul 09 21:51:05 ccache armeb-linux-gcc -march=armv5te -mtune=xscale -shared .libs/blink2.o -Wl,--rpath -Wl,/home/packages/nslu2/mtn/openslug/tmp/work/libgphoto2-2.1.6-r4/libgphoto2-2.1.6/libgphoto2/.libs -L/home/packages/nslu2/mtn/openslug/tmp/work/libgphoto2-2.1.6-r4/libgphoto2-2.1.6/libgphoto2_port/libgphoto2_port/.libs -L/home/packages/nslu2/mtn/openslug/tmp/staging/armeb-linux/lib /usr/lib/libjpeg.so ../../libgphoto2/.libs/libgphoto2.so -march=armv5te -mtu Jul 09 21:51:05 ne=xscale -Wl,-rpath-link -Wl,/home/packages/nslu2/mtn/openslug/tmp/staging/armeb-linux/lib -Wl,-O1 -Wl,-soname -Wl,libgphoto2_sipix_blink2.so -o .libs/libgphoto2_sipix_blink2.so Jul 09 21:51:27 which is totally fucked Jul 09 21:57:37 so assuming I want to get past this and just build the rest of the openslug-packages, how do I do it? Jul 09 21:58:14 hmm, could cheat and touch the stamp. Jul 09 22:01:45 like 6 stamps you mean ? Jul 09 22:02:57 hmmm Jul 09 22:03:00 not sure if you can touch just 1, or you need all 7. Jul 09 22:03:12 did you build gphoto2 or libgphoto2 or both ? Jul 09 22:03:17 gphoto2 Jul 09 22:03:28 and it built libgphoto2 first. Jul 09 22:03:42 ok, well that's what failed - libgphoto2 Jul 09 22:03:51 and now of course gphoto2 failed Jul 09 22:04:07 what dist is your build system? Jul 09 22:04:31 mandrake (mand-break?) 10.1 Jul 09 22:04:38 do you have a /usr/lib/libjpeg.la ? Jul 09 22:05:15 Yes, its there. Jul 09 22:05:37 then I am at a total loss Jul 09 22:06:02 except I did a clean build - well with no tmp or ccache, I didn't start from a wget Makefile Jul 09 22:06:37 and make update doesnt get anything new? Jul 09 22:06:55 it didn't a few monutes ago Jul 09 22:07:47 if you cd openembedded; monotone heads what does it tell you? Jul 09 22:08:03 90985ad6d341c9f6bdacb4243ddf551bffb6c2d6 dyoung@nslu2-linux.org 2005-07-10T02:04:03 Jul 09 22:08:05 is what mine is Jul 09 22:08:22 monotone: branch 'org.openembedded.nslu2-linux' is currently merged: Jul 09 22:08:22 90985ad6d341c9f6bdacb4243ddf551bffb6c2d6 dyoung@nslu2-linux.org 2005-07-10T02:04:03 Jul 09 22:08:48 hmm Jul 09 22:11:56 the -L is the correct one in my compile log. Jul 09 22:12:01 I suppose you could hack this file.... Jul 09 22:13:04 tmp/work/libgphoto2-2.1.6-r4/libgphoto2-2.1.6/libgphoto2/.libs/libgphoto2.la Jul 09 22:13:16 but now we're getting UBE. Jul 09 22:17:07 my -L are correct too, but it's specifying an absolute patch to libjpeg.so Jul 09 22:17:11 path Jul 09 22:17:33 in the stupid sipix camlib that I don't even need Jul 09 22:18:06 it really looks like a libtool thing Jul 09 22:18:57 I dunno if it matters, but my host libtool is 1.5.6 Jul 09 22:20:11 mine is 1.5.189 Jul 09 22:20:13 mine is 1.5.18 Jul 09 22:20:33 shouldn't matter at all Jul 09 22:20:47 and yes I do have to touch about 6 stamps for each package that won't build Jul 09 22:25:20 in the case of gphoto2 it's not working Jul 09 22:26:52 ok, they just have to be touched in the right order apparently Jul 09 22:27:47 I wonder when the last time someone other than me tried a clean build was Jul 09 22:28:26 I'm about to right now. Jul 09 22:28:35 since I'm going out for a bit anyways. Jul 09 22:29:13 youa re going to rm tmp and ~/.ccache/* ? Jul 09 22:30:23 Yup, done. Jul 09 22:30:33 great Jul 09 22:30:52 I'm off, back later; hopefully with some results. Jul 09 22:38:24 ok ttyl Jul 09 23:08:22 NOTE: package openldap-2.2.24-r3: task do_compile: failed Jul 09 23:08:36 /usr/lib/libltdl.so: could not read symbols: File in wrong format Jul 09 23:08:40 I see a pattern **** ENDING LOGGING AT Sat Jul 09 23:59:56 2005