**** BEGIN LOGGING AT Mon Sep 24 02:59:57 2007 Sep 24 03:01:40 hm, if i do bitbake some piece of sw it will result in .ipk and maybe some .ipks of runtime dependencies. but if i want to have packages of build-deps then i need to exlicitly bake them Sep 24 03:01:47 is there any way to do that automagically? Sep 24 05:58:30 03cwixon 07org.oe.dev * r46e38633... 10/ (4 files in 2 dirs): abiword: ship glade files, should fix #3062 Sep 24 06:14:58 03koen 07org.oe.dev * r1c938b19... 10/ (4 files in 3 dirs): tasks: update to 0.12 and add owl patch from poky Sep 24 06:57:30 * Bernardo is back. Sep 24 07:24:10 03koen 07org.oe.dev * r709567d6... 10/ (4 files in 3 dirs): pointercal: remove broken n800 one Sep 24 08:41:47 yay! Sep 24 08:41:51 mmc_spi in 2.6.24 Sep 24 08:53:13 koen: nice :) Sep 24 08:53:18 morning all Sep 24 08:53:26 good morning RP Sep 24 09:12:27 03rwhitby 07org.oe.dev * rdeaf1334... 10/ (1 packages/meta/slugos-packages.bb): slugos-packages: Added bluez-hcidump Sep 24 09:12:35 03rwhitby 07org.oe.dev * rb2924f32... 10/ (1 packages/redboot-utils/fis_1.0.bb): fis: Update to svn rev 6 Sep 24 09:12:45 03rwhitby 07org.oe.dev * r69a79125... 10/ (1 packages/linux/ixp4xx-kernel.inc): ixp4xx-kernel: Ensure zd1211-firmware is built, so it can be included in the rootfs if the kernel module is included. Sep 24 09:16:01 RP: how is udev 115 working out? Sep 24 09:18:28 is there anyone who can help me?? i have a problem with bitbake Sep 24 09:18:51 don't ask to ask, ask Sep 24 09:20:10 runnning bitbake, it only states that a build has been started, then immediatly gives a traceback which ends with these lines: Sep 24 09:20:44 File "tmpHandler(e)", line 31, in tmpHandler ... ValueError: substring not found Sep 24 09:22:44 anyone? Sep 24 09:24:27 ~pastebin Sep 24 09:24:28 methinks pastebin is a place to paste your stuff without flooding the channel - try http://pastebin.ca, or http://channels.debian.net/paste, or http://rafb.net/paste/, or http://pastebin.com is usually painfully too slow and unresponsive to use, use one of the other pastebin sites, or dpaste.com is a very nice pastebin as well Sep 24 09:24:29 roelvp_: is there anything special with your config? Sep 24 09:25:14 special meaning what? Sep 24 09:25:38 i have the standard config with the necessary variables set Sep 24 09:26:36 hey zecke & XorA Sep 24 09:26:43 morning Sep 24 09:30:54 morning Sep 24 09:31:19 hey rwhitby Sep 24 09:32:09 nobody that has any idea? Sep 24 09:33:08 koen: I've not tried it yet. Its there for evalutation Sep 24 09:37:51 does anyone remember where the "tar: *control: Not found in archive Sep 24 09:37:52 " tar wildcards problem was fixed? I need to back-port it to a frozen old release. Sep 24 09:39:14 mtn log packages/ipkg-utils | more Sep 24 09:39:57 thx Sep 24 09:40:03 roelvp_: Something in an anonymous function or event handler is seeing a value it doesn't like. Since nobody is seeing it that suggests your configuration has something unique that is breaking it Sep 24 09:40:35 or mtn log packages/ipkg-utils --no-merges --no-graph --diffs | more Sep 24 09:42:31 koen: I have something you will hate me for :) Sep 24 09:43:29 zecke: I'm trying to figure out how to build a sane qtopia phone image without doing ljp style with a 20MB ipk Sep 24 09:44:11 koen: I will figure that out as well... Sep 24 09:44:33 koen: in theory you could configure each subdir... Sep 24 09:44:53 zecke: and have it link to OE libs instead of internal ones, etc Sep 24 09:44:57 koen: I might go as far as writing a script to split the tarball up, so you can build each piece... Sep 24 09:45:36 * zecke converts mtn to git :} Sep 24 09:49:17 rwhitby: what's up with the excessive merges you keep doing? Sep 24 09:49:57 8 merges for 2 commits is a red flag Sep 24 09:50:12 how do you do that? Sep 24 09:50:41 Crofton|home: I have no clue Sep 24 09:52:42 Crofton|home: especially considering we have an OE automerger running on monotone.openembedded.org Sep 24 09:54:07 I try not make sure not to push multiple heads Sep 24 09:54:24 and I do not think it will merge if there is nothing to merge ... Sep 24 09:55:21 it happens when you commit, pull, merge, pull, merge, pull, merge, pull, merge, push Sep 24 10:00:38 koen: I had a couple of unpushed changes for a number of days Sep 24 10:00:43 nothing unusual Sep 24 10:01:05 03koen 07org.oe.dev * r87770be3... 10/ (1 packages/networkmanager/networkmanager.inc): networkmanager: clean up packaging a bit Sep 24 10:01:07 that doesn't magically create merges Sep 24 10:01:37 my script to update does an auto-merge Sep 24 10:01:41 03koen 07org.oe.dev * r9290de30... 10/ (1 packages/ppp/ppp_2.4.3.bb): pppd: stage headers (from poky) Sep 24 10:02:05 so make that script do a sync before and after a merge Sep 24 10:02:17 03koen 07org.oe.dev * r87770be3... 10/ (1 packages/networkmanager/networkmanager.inc): networkmanager: clean up packaging a bit Sep 24 10:02:23 03koen 07org.oe.dev * r9290de30... 10/ (1 packages/ppp/ppp_2.4.3.bb): pppd: stage headers (from poky) Sep 24 10:02:25 no, cause sometimes (like this one), it was not appropriate to push untested changes Sep 24 10:02:32 creating so much merges is plain anti-social Sep 24 10:02:43 why? Sep 24 10:02:58 untested changes shouldn't be committed Sep 24 10:02:59 isn't that part and parcel of distributed development with monotone? Sep 24 10:03:09 depends on the level of testing. Sep 24 10:03:26 they worked on my machine, but hadn't been tested on the autobuilder yet (different os) Sep 24 10:03:54 why are merges anti-social>? Sep 24 10:04:10 they're invisible to me Sep 24 10:04:23 I said "creating so much merges is plain anti-social" Sep 24 10:04:23 (i.e. I never notice merges that others do) Sep 24 10:04:30 not "merges are anti social" Sep 24 10:04:38 yes, and I'm still asking why you said that. Sep 24 10:05:06 read my mail to oe-devel Sep 24 10:05:24 don't answer then. see if I care. Sep 24 10:05:52 * rwhitby shrugs and continues to commit locally and push when tested properly Sep 24 10:07:06 I answered it in my mail Sep 24 10:07:16 good for you. Sep 24 10:07:31 while we are waiting, how do you notice this? Sep 24 10:07:56 Crofton|home: that's what I'm wondering, merges that others do are invisible to me Sep 24 10:08:02 git pushed people down the rebase route for this problem. I don't think there is a works for all solution Sep 24 10:08:20 Linus also disliked having too many merged making it into the mainline tree Sep 24 10:08:43 a merge is a revision, so every merge takes up space in the db, pollutes the log, creates more pull time and hence more server load Sep 24 10:09:45 so if you push 100 merges, I get to wait ten minutes while pulling and the mtn server is busy with 10minutes times N users for basically no reason Sep 24 10:09:46 thx for the answer. I will change the script so it doesn't do a merge if there is an outstanding commit which hasn't been pushed. Sep 24 10:10:50 and you can do edits in merges and sneak in changes since noone reviews merges Sep 24 10:10:56 this business of having lots of people developing against the same codebase is insane Sep 24 10:11:03 (see mickeyl's mail on regressions caused by bad merges) Sep 24 10:11:22 Crofton|home: that's why guidelines exist :) Sep 24 10:11:53 and they should be unpublishged so we an act like a secret society :) Sep 24 10:12:57 I thought it was pretty obvious what merges do and what their implications are Sep 24 10:13:04 to you Sep 24 10:13:16 I assume you watch the load average on the server? Sep 24 10:13:27 it's not the load average Sep 24 10:13:36 it's the connection multiplexing that is the problem Sep 24 10:13:56 thinking about the parallel paths and merging just makes my head hurt Sep 24 10:14:01 sqlite (what mtn uses as backend) doesn't have fine grained locking, so you can only serve a few people at a time Sep 24 10:14:33 I have no idea how you can keep so many thoughts in your head at one time Sep 24 10:14:40 that's my a pull just sits there after 'connecting to server' Sep 24 10:14:47 I can't track more than two or three things at once :) Sep 24 10:15:12 so it waits for the lock? Sep 24 10:15:33 basically, yes Sep 24 10:16:25 the story with the mtn lock manager is more complicated and over my head, though :) Sep 24 10:17:27 * Crofton|home 's brain explodes Sep 24 10:21:31 should I plane on bringing hardware to play with to OEDEM? Sep 24 10:24:31 Crofton|home: the gumstix would be nice (if the wallwart can do 230V) Sep 24 10:24:38 it can Sep 24 10:25:03 I have a couple of OSK's also Sep 24 10:25:30 the gumstix has the new wired/wireless/miniSD card Sep 24 10:25:48 I need to get the wireless going Sep 24 10:44:01 XorA: pigs can fly: http://www.flickr.com/photos/koenkooi/1431789221/ :) Sep 24 10:44:47 koen: :-D Sep 24 10:45:15 koen, did you see my comment about the backwads negative? Sep 24 10:59:13 03pH5 07org.oe.dev * r51bb8f34... 10/ (1 conf/distro/include/preferred-xorg-versions-X11R7.3.inc): preferred-xorg-versions-X11R7.3.inc: fix xserver preferred version (1.4, not 1.4.0) Sep 24 11:05:37 Crofton|home: I didn't Sep 24 11:09:25 http://www.flickr.com/photos/koenkooi/1288600286/ Sep 24 11:10:27 ah that one Sep 24 11:10:34 yes, it is backwards :) Sep 24 11:18:39 another yay! http://blogs.gnome.org/pbor/2007/09/24/delivering-the-killing-blow-to-libgnome/ Sep 24 11:20:20 http://www.openembedded.org/bonsai/view/rev/4774/ 0 changed files for a propagation... Sep 24 11:34:33 http://lists.maemo.org/pipermail/maemo-developers/2007-September/011781.html Sep 24 11:37:59 wow Sep 24 11:48:34 rwhitby: http://www.openembedded.org:1081/revision/filechanges/30b35d26d67618622ba1fd8fffabd5be4d9508a4/packages/ipkg-utils/ipkg-utils-native_1.6+cvs20050404.bb Sep 24 11:48:49 rwhitby: I just remembered viewmtn can track changes per file now as well :) Sep 24 11:49:37 oh god, big brother is watching even more :) Sep 24 12:05:04 koen: nslu2-linux and openmoko Makefile now only merge directly before a push, never on an update. Sep 24 12:05:22 thanks Sep 24 12:07:38 03koen 07org.oe.dev * r489882b5... 10/ (4 files in 3 dirs): linux-ezx: add patches for a1200 flip switch and keys Sep 24 12:45:47 if OE had not chosen to use bitbake (say, if it had not existed), what would have been the (2nd best) alternative? Sep 24 12:46:30 I'm trying to get a (non-vague) grasp on what role bitbake has in openembedded Sep 24 12:46:41 Skapare: thats kinda hard as bitbake and OE went hand in hand until they split Sep 24 12:47:12 ah ... maybe that's why some of the docs are vague Sep 24 12:47:13 What part did gentoo have in this? Sep 24 12:47:35 well it did say bitbake was "derived" from portage Sep 24 12:47:35 Skapare: Fun nick btw (I'm swedish...) Sep 24 12:47:35 CM: portage was the starting base Sep 24 12:47:48 CM: but its very different now Sep 24 12:47:53 *nod* I thought it was something along that line Sep 24 12:48:00 CM: check the title on my blog then :) ... skapare.blogspot.com Sep 24 12:48:35 Skapare: Hehe Sep 24 12:48:50 I'm not getting the impression that bitbake has the same role relative to openembedded as portage has for gentoo Sep 24 12:49:32 CM: Swedes (and other Scandinavians) would see the irony in that title ... the rest of the world (English reading) would not Sep 24 12:50:24 CM: I am not Swedish ... I got that nick around 16 years ago on a MUD game hosted in Sweden Sep 24 12:50:41 * XorA guses Destructive Creations or Creation Creations as the title :-) Sep 24 12:50:57 XorA: "Creator" Sep 24 12:51:28 Skapare: Which mud? :D Sep 24 12:51:55 I can't even remember the name of it now ... it disappeared around 1992 Sep 24 12:52:08 Oh, ok. I've played a few Sep 24 12:52:18 Nice c library btw Sep 24 12:52:57 C is my strong language ... though I am looking at picking up Python at some point Sep 24 12:53:36 I know a lot of languages, but I suck in all of them.. ;) Sep 24 12:54:19 pick one (except Perl) and focus on it ... develop anything you can in it ... you'll get better Sep 24 12:54:43 Why isn't bitbake like portage btw? Because it's only used for building images and packages, and not for installing? Sep 24 12:54:51 * Skapare always takes swipes at Perl when he can Sep 24 12:54:55 Hehe Sep 24 12:55:09 That's one language I've ever only written on script in Sep 24 12:55:20 Mostly I've used C, Java and C# Sep 24 12:55:52 CM: it was portage, then we needed to make it suck less and add missing functionality and suddenly nothing of portage was left :) Sep 24 12:55:53 Q: is bitbake meant to be something to build packages individually, or meant to be the overall wrapper around a set of packages to build the whole distribution? Sep 24 12:56:02 koen: Hehe Sep 24 12:56:12 bitbake knows nothing of packages Sep 24 12:56:16 Skapare: bitbake is a task executor Sep 24 12:56:19 it's a generic task executor Sep 24 12:56:30 Skapare: it knows tasks, dependencies of tasks and will execute them Sep 24 12:56:36 OE adds tasks which end up producing packages Sep 24 12:56:47 then what role are all those configuration files for bitbake that I saw (which I have not yet figured out what I need to put in them) Sep 24 12:57:20 I think it's a bit tricky to figure out, but I think I'm slowly getting it Sep 24 12:57:29 Skapare: this could be buying/ordering pizza, doing graph based image manipulation, or build and package software (in case of OE) Sep 24 12:57:35 * CM realises he thinks too much :( Sep 24 12:57:41 does it just decribe what dependencies exist (e.g. build this package X only after package B is built) ? Sep 24 12:58:10 Yes, how is the order things should build constructed? Sep 24 12:58:20 but is that all it does? Sep 24 12:58:29 Skapare: BitBake defines the file format for .conf/.ini/.bb/.inc and .bbclass files Sep 24 12:58:42 seems way more complicated than just listing dependencies Sep 24 12:59:04 zecke: but what are those files for? what is their purpose? Sep 24 12:59:20 Skapare: describing tasks, their actions and their dependencies? Sep 24 12:59:29 Skapare: check the BitBake manual to see what BitBake is doing Sep 24 12:59:42 and each task in bitbake is what? to build one package? Sep 24 12:59:54 Skapare: read the manual Sep 24 12:59:57 I was checking the manual ... my head is spinning Sep 24 13:00:13 Skapare: then take a walk and afterwards continue reading Sep 24 13:00:21 maybe a good idea Sep 24 13:01:05 someone (on openwrt) suggested bitbake ... but I'm not sure it's what I need ... things like dependencies are just a side factor for me Sep 24 13:01:07 ~lart git-fast-import Sep 24 13:01:07 * ibot declares git-fast-import a moron Sep 24 13:01:39 I need something that does stuff like scheduling machines to participate in building a distro Sep 24 13:01:52 Skapare: scheduling machines? Sep 24 13:01:55 but I don't see that in bitbake Sep 24 13:01:57 right Sep 24 13:02:09 Skapare: what should scheduling machine sbe? Sep 24 13:02:16 20 machines running in parallel to speed the build Sep 24 13:02:50 someone might request one package, or a set of packages, or in extreme cases, the whole distribution, to be rebuilt Sep 24 13:03:03 Skapare: then you want something like distcc, icecc (http://en.opensuse.org/Icecream) or teambuilder (http://trolltech.com/products/qt/addon/teambuilder) Sep 24 13:03:05 INHERIT += "icecc" ? Sep 24 13:03:21 the schedular should hand off work (tasks) to the various machines to optimize the building of whatever is requested Sep 24 13:03:35 distcc is already a component in the plan Sep 24 13:04:04 I have icecc on my plate to have a look at later today ... teambuilder I will now add to that Sep 24 13:04:07 thanks Sep 24 13:04:07 Skapare: It is hard to judge what you are trying to do and why you are obsessed by bitbake... Sep 24 13:04:39 I'm not obsessed by it ... someone suggested it ... and I wanted to just be thorough before I dismiss it Sep 24 13:04:44 Skapare: OE is used to build packages and distributions, Sep 24 13:05:15 Skapare: OE instructs how software is to be build and packages Sep 24 13:05:45 zecke: does it in any way alter how an original source package builds itself? Sep 24 13:05:54 Skapare: OE features native support for icecc and supports building multiple packages/tasks/providers in parallel Sep 24 13:06:25 zecke: or does it just build that package the way the package says (usually "./configure && make && make install")? Sep 24 13:07:01 Skapare: one .bb file describes the process of building a packaging (from downloading, unpacking, patching, configuring, compiling, staging, installing, packaging) Sep 24 13:07:16 zecke: one of the issues with multiple machines participating in the build in parallel is to be sure each machine has in its build environment all the build-time dependencies Sep 24 13:07:31 Skapare: to be efficient common code is in the .bbclass but a .bb file has full control Sep 24 13:07:55 zecke: OK ... so the .bb file could be like what I have now in the form of a bash wrapper script that carries out those same steps Sep 24 13:08:17 Skapare: with icecc the toolchain gets distributed to other machines... Sep 24 13:09:09 Skapare, what are you trying to do? Sep 24 13:09:13 zecke: what I'm working on is building a kind of workplan graph which says which machines should build which packages so that I don't end up asking machine 12 to build something which depends on a packages that machine 12 does not have in its environment Sep 24 13:09:33 Skapare: you are totally confused... Sep 24 13:09:59 zecke: running machines in parallel is not as free as just handing it a task to do ... the build sequence dependencies have to be considered Sep 24 13:10:03 zecke: in what way? Sep 24 13:10:26 Crofton|home: building a scheduling system for a distribution builder Sep 24 13:10:38 Crofton|home: multi-machine schedular Sep 24 13:11:56 isn't that way icecc.bbclass does in OE? Sep 24 13:12:21 stelios did icecc, right? Sep 24 13:12:30 zecke: if package X needs package D to be present to build package X then I need to choose a machine to do the build of package X on which has previously built package D or has acquired the results of that package D build Sep 24 13:12:39 Skapare: icecc, distcc and teambuilder just distribute compile jobs... Sep 24 13:12:50 zecke: not all machines will have every needed dependency on each of them Sep 24 13:12:54 Crofton|home: well, I did back at ROAD and Stelios and Laibsch fixed it again Sep 24 13:12:58 I wonder how they return results? Sep 24 13:13:06 :) Sep 24 13:13:33 zecke: I need more than just plain "spread the work load around" ... I need to make sure that the machine chosen to do a task has on it what that task needs Sep 24 13:13:56 obviously icecc.bbclass must have some way of dealing with this ...... Sep 24 13:14:28 zecke: if package B needs package A and I just built package A on machine 3 then machine 3 can be used to build package B ... but machine 2 cannot because it does not have the dependency Sep 24 13:15:01 zecke: but if machines 3 and 2 both become idle ... then machine 3 can transfer what it has built to machine 2 then machine 2 is now eligible for package B, too Sep 24 13:15:42 zecke: the objective is to optimize that whole sequence Sep 24 13:18:19 koen|away: ixp4xx npe microcode 3.0 is under a new license which allows redistribution in binary form. Sep 24 13:18:29 http://downloadcenter.intel.com/license_agr.aspx?url=/13757/eng/IPL_ixp400NpeLibrary-3_0.zip&ProductID=2100&agr=Y&sType=&PrdMap=&DwnldId=13757&strOSs=All&OSFullName=All+Operating+Systems&lang=eng Sep 24 13:19:56 rwhitby: nice ... I like the ixp4xx boards Sep 24 13:20:46 Someone here should be on LUGRadio... ;) Listen at 52:20 or so. They want to know why linux embedded devices don't work, from someone who works with it. Sep 24 13:21:31 why they _don't_ work? uh ... Sep 24 13:22:01 ... but they do work Sep 24 13:22:23 well I need to run (home now, going to work) Sep 24 13:30:57 rwhitby: great, and they work with the new opensource driver? Sep 24 13:31:13 koen: building to test that now Sep 24 13:32:01 damn, this board has some bad caps Sep 24 13:32:11 morning Sep 24 13:32:15 this high pitched whine is driving me nuts Sep 24 13:32:16 ~lart ps3 mplayer Sep 24 13:32:16 * ibot puts ps3 mplayer into a headlock and administers a mighty noogie, rubbing half of ps3 mplayer's hair of Sep 24 13:37:34 'allo Sep 24 13:48:29 03tmbinc 07org.oe.dreambox * radcee12b... 10/ (3 files in 3 dirs): initscripts-opendreambox: add compatibility symlink for oled Sep 24 13:48:35 03tmbinc 07org.oe.dreambox * radcee12b... 10/ (3 files in 3 dirs): initscripts-opendreambox: add compatibility symlink for oled Sep 24 13:48:55 03tmbinc 07org.oe.dreambox * radcee12b... 10/ (3 files in 3 dirs): initscripts-opendreambox: add compatibility symlink for oled Sep 24 13:48:58 03tmbinc 07org.oe.dreambox * radcee12b... 10/ (3 files in 3 dirs): initscripts-opendreambox: add compatibility symlink for oled Sep 24 13:49:05 03tmbinc 07org.oe.dreambox * radcee12b... 10/ (3 files in 3 dirs): initscripts-opendreambox: add compatibility symlink for oled Sep 24 13:49:08 03tmbinc 07org.oe.dreambox * radcee12b... 10/ (3 files in 3 dirs): initscripts-opendreambox: add compatibility symlink for oled Sep 24 13:50:22 03tmbinc 07org.oe.dreambox * radcee12b... 10/ (3 files in 3 dirs): initscripts-opendreambox: add compatibility symlink for oled Sep 24 13:50:22 03tmbinc 07org.oe.dreambox * r39bf37b9... 10/ (1 packages/dreambox/dreambox-dvb-modules.bb): dreambox-dvb-modules: update dm7025 drivers Sep 24 13:50:22 03svolpe 07org.oe.dev * r9194b2f0... 10/ (1 packages/cairo/cairo-directfb_1.4.8.bb): cairo-directfb: r1: Added RCONFLICTS for cairo Sep 24 13:50:31 gm Sep 24 13:50:49 svolpe: thanks for directfb support Sep 24 13:50:53 no one ever did explain why OEDEM is in Berlin not a Greek isalnd Sep 24 13:51:07 good point Sep 24 13:52:29 Crofton|home: what makes you think it is not a greek island? Sep 24 13:52:49 well, I have never been to Berlin Sep 24 13:52:55 maybe it is Sep 24 13:53:47 of course, when I was growing up, Berlin was an island (parts at least) Sep 24 13:55:15 koen|away: v3.0 microcode is only for ixp465 and later. Does not work on ixp42x. However, it looks like the 2.4 microcode may now be under the same license anyway. Sep 24 13:56:44 likewise, no problem, I hope to add as many directfb packages as I can over the next month or two. Sep 24 14:01:14 svolpe: where do you test on? which directfb enabled h/w? Sep 24 14:02:30 likewise, I'm currently testing it on a custom PXA270 board... well right now I'm waiting on new hardware so I won't be doing testing for about 2 weeks. I also hope to purchase a N800 in November and I will start developing on it as well. Sep 24 14:04:59 likewise, how about yourself? I'm assuming your using directfb as well, what are you using it on? Sep 24 14:12:22 svolpe: I tried to use it on a framebuffer. Sep 24 14:14:36 svolpe: custom FPGA hardware. But currently, I use cairo on a image buffer, which we then tranfer ourselves. Sep 24 14:14:59 svolpe: is there nv800 support in directfb yet? Sep 24 14:15:06 s/nv800/n800 Sep 24 14:34:10 likewise, directfb is hardware independant.. it just uses the kernel fb drivers. Sep 24 14:35:08 likewise, so it should work for you if your display has a driver that provides framebuffering through the kernel. Sep 24 14:36:31 omapfb was merged in to mainline in 2.6.23 Sep 24 14:38:28 koen: thanks for the heads-up. Sep 24 14:39:28 we shoudl get that working with the OSK, even though the display is absrudly expensive Sep 24 14:41:04 kergoth, you here man? Sep 24 14:41:27 * koen gets prompt on dht-walnut Sep 24 14:41:29 svolpe, could you access a list of those that got commit rights into the tslib? Sep 24 14:42:06 kristoffer: that info is on the berlios project page Sep 24 14:42:26 koen, doh, being lazy as always :) thnx Sep 24 14:42:29 kristian-m: http://developer.berlios.de/project/memberlist.php?group_id=3809 Sep 24 14:42:35 ehm Sep 24 14:42:39 kristoffer: http://developer.berlios.de/project/memberlist.php?group_id=3809 Sep 24 14:43:04 oki, so basicly kergoth then Sep 24 14:43:55 looks like it Sep 24 15:14:39 hey HopsNBarley Sep 24 15:17:05 Jin^eLD, I added a patch for dv git kernel to shrink kernel to sane size Sep 24 15:17:07 hi Jin^eLD Sep 24 15:17:23 Crofton: I think the size was just a binutils version problem? Sep 24 15:17:29 yeah Sep 24 15:17:46 I added a patch to the git kernel to avoid the issue Sep 24 15:17:54 HopsNBarley: anything new on openprotium? I did not have time to continue my hacking yet Sep 24 15:17:57 hopefully it flows back from upstream eventually Sep 24 15:18:06 Crofton: ah, I see, nice Sep 24 15:18:15 well, I still did not have time to reflash the board with your uboot version Sep 24 15:18:26 small steps :) Sep 24 15:18:29 :) Sep 24 15:58:36 who maintains the pango package? Sep 24 16:23:44 Jin|away, i'm working on it (-; Sep 24 17:07:37 I think I'm getting a u-boot git tarball from from the source server, then unpacking and updating it Sep 24 17:07:41 but this dies Sep 24 17:07:46 sound familiar? Sep 24 17:15:11 re Sep 24 17:15:36 hi Jin^eLD Sep 24 17:15:37 Crofton: for me uboot always failed in do_fetch, and yes, I believe it was also during the git update Sep 24 17:15:40 hey HopsNBarley Sep 24 17:16:28 HopsNBarley: when doing a "shutdown -h now" on the storcenter, the leds stay on, is there a specific powerdown procedure to shut it down completely, what I am missing? Sep 24 17:20:43 I'm wondering if there is a bum tarball on the angstrom source cache machine? Sep 24 17:21:39 kergoth: have you played with git-fast-import yet? Sep 24 17:22:04 how can I stop OE from checking for tarballs of git repos? Sep 24 17:22:31 Jin^eLD, they stay on? red or blue? Sep 24 17:22:57 HopsNBarley: I think red Sep 24 17:22:58 zecke: not yet, haven't had a repo that needed it to convert yet, only used git-svn so far for conversions Sep 24 17:23:11 HopsNBarley: but then again - its my angstrom build, not openprotium, I was just wondering if I missed something Sep 24 17:23:22 of it there was something special that had to be done Sep 24 17:23:25 Jin^eLD, it should power down, off, no lights. etc (-; Sep 24 17:23:37 hmm Sep 24 17:23:37 :) Sep 24 17:23:42 hold on... Sep 24 17:24:25 kergoth: I'm writing a bidirectional git<->mtn importer/exporter but git-fast-import gives me a hard time... Sep 24 17:26:31 Jin^eLD, if i recall correctly - to power off the board - you have to send a message to that same controller chip (the lights and fan). You can do this with scc -p off Sep 24 17:27:26 HopsNBarley: aaah! ok, thanks, I guess that is what I am missing Sep 24 17:27:45 yeah, just get this in the right place in the RC scripts, and it'll CLICK off. Sep 24 17:28:12 ... a better solution would be to get this into the kernel, but... (-; Sep 24 17:28:18 :) Sep 24 17:35:42 I keep running into RRECOMMENDS="gdk-pixbuf-loader-png" but I can't find gdk-pixbuf-loader-png bb file anywere, were os the recipe for this? Sep 24 17:36:23 what image are you building? Sep 24 17:36:40 zecke: ah, cool Sep 24 17:37:12 * kergoth was actually toying with the idea of seperate git repositories for different oe packages, with somethign like git-submodule to create metadata repositories using them.. but itd probably add too much overhead :) Sep 24 17:38:16 Crofton, gtk+-directfb well I'm working on it and the gtk-2.10.inc has it in there but because of it it builds a bunch of X11 packages. If I remove the RRECOMMENDS all is good and no X packages build but that seems like something I would want for gtk so I was try to trace it to figure out if I can use it with the directfb build. Sep 24 17:38:26 kergoth: fast-import is currently just dropping a commit :} Sep 24 17:39:25 svolpe, I thought you were running into EXTRA_RRECO versus ESSENTIAL_RRECO Sep 24 17:39:41 sorry, I don't know about that problem Sep 24 17:43:58 do_split_packages and PACKAGES_DYNAMIC Sep 24 17:45:13 koen, was that comment for me? Sep 24 17:45:24 yes Sep 24 17:47:37 koen, thanks I will need to look those up first. I'm injoying this directfb excersize as it is really making me learn more about OE. Normally I just add very simple linear packages. Sep 24 17:56:37 Crofton|home: http://dev.openbossa.org/trac/mamona/browser/qemu-omap/trunk Sep 24 17:56:40 arg Sep 24 17:56:46 I need to hide and edit chapter 4 Sep 24 17:57:47 hmmm, emulator for OSK? Sep 24 17:58:09 yes Sep 24 17:58:28 "e are doing this as part of Mamona project (dev.openbossa.org/trac/mamona). Sep 24 17:58:28 Nowadays, we are implementing the OMAP 16xx (1610, 1710, 5912) Sep 24 17:58:29 emulation. The project is in early stage but you can see the code in Sep 24 17:58:29 http://dev.openbossa.org/trac/mamona/browser/qemu-omap/trunk." Sep 24 17:58:49 hmm Sep 24 17:58:50 Crofton: one of the mamona dudes is attending OEDEM, so you can ask him for details :) Sep 24 17:58:55 I wonder if they are doing to DSP? Sep 24 17:58:58 yeah Sep 24 17:59:08 I should bring th eOSK with the LCD Sep 24 17:59:22 do we have hacking goals for the weekend? Sep 24 17:59:36 not AFAIK Sep 24 18:00:15 The gumstix is easy to bring Sep 24 18:00:19 03rwhitby 07org.oe.dev * r610444b5... 10/ (1 packages/slugimage/slugimage.bb): slugimage: Updated to latest version which can accept the version 3.0 intel microcode. Sep 24 18:00:25 03rwhitby 07org.oe.dev * re5dff199... 10/ (5 files in 3 dirs): ixp4xx-npe: Updated to version 2.4 microcode - already well tested in OpenWrt for some months, and now under a redistributable in binary form license. Sep 24 18:00:28 the OSK's aren't too bad either Sep 24 18:00:47 oops Sep 24 18:01:01 * mwester hopes the auto-merge will work Sep 24 18:01:13 it does :) Sep 24 18:02:49 ok, I am going to hide Sep 24 18:04:21 Ah, not to worry. CIA is just really slow. Sep 24 18:09:25 03mwester 07org.oe.dev * red5684c6... 10/ (3 files in 3 dirs): Sep 24 18:09:25 /etc/init.d/gsmd: Disable echo before powering on the gta01 GSM modem so that Sep 24 18:09:25 the modem doesn't talk to itself during startup - fixes gsmd startup problem. Sep 24 18:09:25 Also re-order disable of HW flowcontrol on shutdown to eliminate a crash window. Sep 24 18:09:30 03mwester 07org.oe.dev * ra0edf3d1... 10/ (1 packages/glibc/glibc_2.3.5+cvs20050627.bb): glibc: ldconfig patch applies to big-endian (armeb) as well as arm. Sep 24 18:14:36 mwester: I bet $10 that harald finds a way to break gsmd within 2 days Sep 24 18:15:14 I take bets on wether it will be the kernel, uboot or gsmd itself where he will break it ;) Sep 24 18:16:08 * koen takes the apple pie out of the oven and has some coffee Sep 24 18:17:33 Harald is too busy! :) Sep 24 18:18:13 not for breaking gsmd! Sep 24 18:18:46 I still have the kernel patch that would get rid of the cruft in the gsmd startup file waiting in the wings for someone (Harald, I think) to say yay/nay on... Sep 24 18:20:14 Frankly, it's looking more and more like the entire GSM modem switching stuff needs to be moved into the serial driver. Sep 24 18:20:29 Which just makes sense if you think about what it does. Sep 24 18:22:02 i don't remember ... if i have a package FOO anche a machine 'bar' ... and i define FOO = "a b c d" and FOO_bar = "a b" ... when i'm building for machine 'bar' what's the value of FOO variable ??? Sep 24 18:22:16 anche->and Sep 24 18:32:44 koen, what actually is going on with gdk-pixbuf-loader-png? its not a bb package, is it some module generated by gtk? Sep 24 18:35:09 yes Sep 24 18:35:14 in do_split_packages Sep 24 18:37:10 koe any help for me ? ;) at least the manual to read :) ... Sep 24 18:37:13 koen Sep 24 18:37:38 bitbake.berlios.de Sep 24 18:38:30 i'm ytying to define DISTRO_FEATURES_h3600 in angstrom.inc but seem to not work Sep 24 18:39:06 koen so if I set PACKAGES_DYNAMIC to the module names and then call do_split_packages split_packages (looking at the python script) parses the modules into seperate packages, correct? Sep 24 18:41:39 do_split_packages splits them into seperate packages and PACKAGES_DYNAMIC tells OE that that recipe provides that packages, since do_split_packages doesn't add to PACKAGES Sep 24 18:45:34 koen, the way gtk-2.10.inc does it is it uses RRECOMMENDS then calls do_split_packages but by doing this it causes other packages not in the depends list to be built such as libxfixes. Sep 24 19:41:15 koen, well I think the real source of my woes is a typo in the gtk-2.10.inc, it looks like FILES_${PN} does not have a closing " I'll fix it and push it. Sep 24 19:41:44 hmm, I have not received an email from the hotel regarding my attempt at making an OEDEM reservation .... Sep 24 19:50:30 Crofton, were are you from in the US? Sep 24 19:51:43 yeah Sep 24 19:51:49 Blacksburg Virginia Sep 24 19:52:04 Once upon a time I was English Sep 24 19:53:14 Crofton, ok, your just right down the road.. well at least compared to everyone else :-) I'm in Pittsburg, PA. Sep 24 19:53:27 heh Sep 24 19:53:36 yeah, practically neighbors Sep 24 19:53:44 4-5 hours away :) Sep 24 19:56:08 could someone kindly remind me what the problem is here: NOTE: preferred version 2.5 of glibc not available (for item virtual/arm-linux-gnueabi-libc-for-gcc) Sep 24 19:56:19 Crofton, well cbrake is 3-4 hours west of me :-) Sep 24 19:56:36 gm Sep 24 19:56:53 hi likewise! Sep 24 19:57:07 hmm, don't remember what that problem was about anymore... Sep 24 19:57:34 it's just an annoyance, but i'd like to understand it. Sep 24 20:01:52 koen, if I use do_split_packages and PACKAGES_DYNAMIC correctly for a module: example gdk-pixbuf-loader-png, I should see an ipkg package called gdk-pixbuf-loader-png after I build correct? Sep 24 20:12:47 yes, even without PACKAGES_DYNAMIC Sep 24 20:13:16 PACKAGES_DYNAMIC is only there to tell OE that gtk+_2.10.48.bb generates dk-pixbug-loader-* packages Sep 24 20:14:06 time to zzzz now Sep 24 20:14:21 koen" trustu Sep 24 20:14:29 s/"/: :-) Sep 24 20:16:34 gn Sep 24 20:43:29 good nite all Sep 24 20:46:18 Irregularity, unpredictability, caprice, and miscellaneous Sep 24 20:46:18 strangeness have been my constant companions in attempting to work Sep 24 20:46:18 with the Gumstix buildroot. Sep 24 20:47:49 At least malice is absent. Sep 24 20:48:31 heh Sep 24 20:53:21 Crofton, I started out with using a gumstix and buildroot, I was never happier than the day I found OE and could move away from buildroot :-) Sep 24 21:02:44 03svolpe 07org.oe.dev * r6f95a0ee... 10/ (5 files in 3 dirs): gtk+-directfb: r0 Added verion 2.10.14, deleted 2.10.9, fixed gtk-2.10.inc Sep 24 21:02:49 03svolpe 07org.oe.dev * r6f95a0ee... 10/ (5 files in 3 dirs): gtk+-directfb: r0 Added verion 2.10.14, deleted 2.10.9, fixed gtk-2.10.inc Sep 24 21:03:02 03svolpe 07org.oe.dev * r6f95a0ee... 10/ (5 files in 3 dirs): gtk+-directfb: r0 Added verion 2.10.14, deleted 2.10.9, fixed gtk-2.10.inc Sep 24 21:03:07 wtf, how can we have a revision without a branch :} Sep 24 21:03:10 03svolpe 07org.oe.dev * r6f95a0ee... 10/ (5 files in 3 dirs): gtk+-directfb: r0 Added verion 2.10.14, deleted 2.10.9, fixed gtk-2.10.inc Sep 24 21:03:16 03svolpe 07org.oe.dev * r6f95a0ee... 10/ (5 files in 3 dirs): gtk+-directfb: r0 Added verion 2.10.14, deleted 2.10.9, fixed gtk-2.10.inc Sep 24 21:03:24 03svolpe 07org.oe.dev * r6f95a0ee... 10/ (5 files in 3 dirs): gtk+-directfb: r0 Added verion 2.10.14, deleted 2.10.9, fixed gtk-2.10.inc Sep 24 21:20:26 zecke, I thought that was the idea of pulling -- merging -- pushing to avoid branches. Sep 24 21:36:27 03svolpe 07org.oe.dev * r8cab20c1... 10/ (1 packages/pango/pango-directfb_1.18.1.bb): pango-directfb: Added directfb support to pango Sep 24 21:36:31 03svolpe 07org.oe.dev * rff7e7d4b... 10/ (1 packages/gtk+/gtk+-directfb_2.10.14.bb): gtk+-directfb: r1: Add gtk+ to conflicts list Sep 24 22:44:41 hi there Sep 24 22:45:09 anyone knows if there is any effort in using oe to build openwrt ? Sep 24 22:45:20 firmwares for that wrt54g Sep 24 23:00:50 bedboi: you can make your own software for openwrt then Sep 24 23:14:42 03pfalcon 07org.oe.dev * rb49e5f3c... 10/ (3 files in 3 dirs): Sep 24 23:14:42 sysvinit 2.86: ENABLE_ROOTFS_FSCK is referenced by checkroot.sh, so add it here. Sep 24 23:14:45 * The value is a bit arbitrary. Sep 24 23:14:47 03pfalcon 07org.oe.dev * rb49e5f3c... 10/ (3 files in 3 dirs): Sep 24 23:14:49 sysvinit 2.86: ENABLE_ROOTFS_FSCK is referenced by checkroot.sh, so add it here. Sep 24 23:14:51 * The value is a bit arbitrary. Sep 24 23:15:20 03pfalcon 07org.oe.dev * rb49e5f3c... 10/ (3 files in 3 dirs): Sep 24 23:15:20 sysvinit 2.86: ENABLE_ROOTFS_FSCK is referenced by checkroot.sh, so add it here. Sep 24 23:15:20 * The value is a bit arbitrary. Sep 24 23:15:22 03pfalcon 07org.oe.dev * rb49e5f3c... 10/ (3 files in 3 dirs): Sep 24 23:15:24 sysvinit 2.86: ENABLE_ROOTFS_FSCK is referenced by checkroot.sh, so add it here. Sep 24 23:15:26 * The value is a bit arbitrary. Sep 24 23:15:28 03pfalcon 07org.oe.dev * rb49e5f3c... 10/ (3 files in 3 dirs): Sep 24 23:15:30 sysvinit 2.86: ENABLE_ROOTFS_FSCK is referenced by checkroot.sh, so add it here. Sep 24 23:15:32 * The value is a bit arbitrary. Sep 24 23:15:37 03pfalcon 07org.oe.dev * rb49e5f3c... 10/ (3 files in 3 dirs): Sep 24 23:15:39 sysvinit 2.86: ENABLE_ROOTFS_FSCK is referenced by checkroot.sh, so add it here. Sep 24 23:15:41 * The value is a bit arbitrary. Sep 24 23:15:43 03pfalcon 07org.oe.dev * rb49e5f3c... 10/ (3 files in 3 dirs): Sep 24 23:15:49 sysvinit 2.86: ENABLE_ROOTFS_FSCK is referenced by checkroot.sh, so add it here. Sep 24 23:15:51 * The value is a bit arbitrary. Sep 24 23:15:53 03pfalcon 07org.oe.dev * rb49e5f3c... 10/ (3 files in 3 dirs): Sep 24 23:15:55 sysvinit 2.86: ENABLE_ROOTFS_FSCK is referenced by checkroot.sh, so add it here. Sep 24 23:15:57 * The value is a bit arbitrary. Sep 24 23:15:59 03pfalcon 07org.oe.dev * rb49e5f3c... 10/ (3 files in 3 dirs): Sep 24 23:16:03 sysvinit 2.86: ENABLE_ROOTFS_FSCK is referenced by checkroot.sh, so add it here. Sep 24 23:16:05 * The value is a bit arbitrary. Sep 24 23:16:07 03pfalcon 07org.oe.dev * rb49e5f3c... 10/ (3 files in 3 dirs): Sep 24 23:16:09 sysvinit 2.86: ENABLE_ROOTFS_FSCK is referenced by checkroot.sh, so add it here. Sep 24 23:16:11 * The value is a bit arbitrary. Sep 24 23:16:13 03pfalcon 07org.oe.dev * rb49e5f3c... 10/ (3 files in 3 dirs): Sep 24 23:16:15 sysvinit 2.86: ENABLE_ROOTFS_FSCK is referenced by checkroot.sh, so add it here. Sep 24 23:16:19 * The value is a bit arbitrary. Sep 24 23:16:21 03pfalcon 07org.oe.dev * rb49e5f3c... 10/ (3 files in 3 dirs): Sep 24 23:16:23 (16 lines omitted) Sep 24 23:19:11 dcordes: what? Sep 24 23:19:23 is wrt54g supported? Sep 24 23:19:32 i didn't find anything about this Sep 24 23:20:07 no idea Sep 24 23:20:38 03pfalcon 07org.oe.dev * ra493869a... 10/ (3 files in 3 dirs): Sep 24 23:20:38 initscripts 1.0: checkroot.sh: Die, die, die fsck interactivity! Sep 24 23:20:38 * Works around #3065. Sep 24 23:20:38 * Fixes infamous issue with hang during boot on h2200, etc. Sep 24 23:20:58 03pfalcon 07org.oe.dev * ra493869a... 10/ (3 files in 3 dirs): Sep 24 23:20:58 initscripts 1.0: checkroot.sh: Die, die, die fsck interactivity! Sep 24 23:20:58 * Works around #3065. Sep 24 23:20:58 * Fixes infamous issue with hang during boot on h2200, etc. Sep 24 23:21:34 bedboi: there is a wrt54 machine type in OE Sep 24 23:22:39 you might want to check around the openwrt site Sep 24 23:41:51 hello Sep 24 23:42:33 how do I set a variable from a bb recipe to be read from a makefile I'm calling in the recipe itself? Sep 24 23:58:44 Ghiottone, is it a standard variable like CFLAGS or a custom variable? Sep 25 00:01:15 I have to set the DESTDIR variable Sep 25 00:02:16 that is used by the makefile as a prefix to install things Sep 25 00:11:31 Ghiottone, use EXTRA_OECONF = "--prefix= " that passes it to the configure when it it is run. you can do this with any configure (configure -h to see all the options) option. Sep 25 00:11:55 unfortunately it has not a configure script Sep 25 00:12:10 it is just that that variable works as a --prefix Sep 25 00:15:03 Ghiottone, you should be able to just export the variable inside of the .bb file. Sep 25 00:15:33 Ghiottone, there is also oe_runmake. Sep 25 00:15:42 which is the syntax to export? Sep 25 00:16:06 I'm using oe_runmake but I can't see any difference from just using make Sep 25 00:16:26 Ghiottone, export VARIABLE_NAME=value Sep 25 00:17:20 Ghiottone, did you try oe_runmake 'VARIABLE_NAME=value' Sep 25 00:17:20 You could call it from do_compile{ } Sep 25 00:17:33 Ghiottone, I'm only guessing at that as I have not done it before. Sep 25 00:18:36 Ghiottone, look at the following .bb file "org.openembedded.dev/packages/cups/cups_1.2.7.bb" Sep 25 00:18:55 Ghiottone, I think the oe_runmake example under the do_compile is what you want to do. Sep 25 00:19:11 ok I'll look! thanks a lot Sep 25 00:19:21 Ghiottone, no problem. Sep 25 00:20:38 yes! it is exactly what I was trying to do Sep 25 00:20:53 he put the double quotes in it Sep 25 00:21:08 I did a similar thing but didn't work Sep 25 00:24:05 Ghiottone, I think that is the right way to do it. Sep 25 00:24:33 exporting a variable should work as well but that is more of a hack, I think. **** ENDING LOGGING AT Tue Sep 25 02:59:56 2007