**** BEGIN LOGGING AT Thu Mar 07 02:59:58 2013 Mar 07 03:33:55 hi Mar 07 03:58:48 arr Mar 07 04:04:28 did someone got to situation when bitbake says that it runs wget to fetch tarball from PREMIRROR, then it goes to another one and then trie to unpack not-existant archive? http://pastebin.com/zgN6Qz78 is log.do_fetch Mar 07 04:10:14 nope, when my git recipes fail it's usually because of transient network glitches Mar 07 04:10:46 i've seen it do the mirror cascade thing on normal recipes but not git ones... Mar 07 04:21:13 bom dia otwieracz Mar 07 04:21:25 bom dia otavio Mar 07 04:29:04 hrw: does your linaro kernel recipe use a tag/commit-hash, or is it autorev? Mar 07 04:30:06 if you want it to always look locally for things not in the local downloads then you might want a different setting Mar 07 04:31:25 i have SOURCE_MIRROR_URL set to my local apache server and INHERIT += "own-mirrors" in local.conf Mar 07 04:32:12 that seems to keep it from trying to resolve src_uri's that are inside TI's network Mar 07 04:32:32 we use a lot of TI davinci software from their sdk... Mar 07 04:34:23 i've also retrieved a few things manually and made the md5s, etc Mar 07 04:37:16 nerdboy: auto Mar 07 04:37:52 nerdboy: the problem is that it runs wget but do not fetch or sth like that. Mar 07 04:38:04 sure you don't want to build something a little more stable? Mar 07 04:38:06 I have to catch some experienced devs. Mar 07 04:38:15 but being in HK does not help ;D Mar 07 04:38:29 nerdboy: policy is to have it HEAD D: Mar 07 04:38:52 i'm not sure i would do that unless it was my own branch Mar 07 04:39:19 it only changes when we decide it should Mar 07 06:49:28 ok, looks like PREMIRRORS stuff affected my fetching Mar 07 09:44:51 hi woglinde Mar 07 09:48:44 Hello #oe Mar 07 09:49:07 JaMa: hi Mar 07 09:53:43 JaMa: still waiting for tom king's feedback for the paches I sent last saturday. do you have any news ? Mar 07 09:58:50 (he may be busy or on vacations, just asking in case) Mar 07 10:04:04 hi hrw Mar 07 10:04:09 hi apelete Mar 07 10:04:35 woglinde: hi Mar 07 10:16:30 morning apelete, hrw, woglinde, all Mar 07 10:17:22 hi bl Mar 07 10:17:36 bluelightning: hi Mar 07 10:20:03 I hate assumptions ;( Mar 07 10:20:23 BB_FETCH_PREMIRRORONLY == BB_NO_NETWORK is hrdcoded ;( Mar 07 10:26:44 ech Mar 07 10:33:13 apelete: yes, he is busy Mar 07 10:33:25 apelete: I'll ping him Mar 07 10:34:06 JaMa: ok, thanks Mar 07 10:35:04 DEBUG: Trying PREMIRRORS Mar 07 10:35:05 NOTE: fetch http://snapshots.linaro.org/openembedded/sources/git2_git.linaro.org.kernel.linux-linaro-tracking.git.tar.gz Mar 07 10:35:23 then wget is "started", file is not there Mar 07 10:35:29 DEBUG: Trying PREMIRRORS Mar 07 10:35:30 NOTE: fetch http://snapshots.linaro.org/openembedded/sources/git2_github.com.facebook.hiphop-php.git.tar.gz Mar 07 10:35:36 JaMa: I'm not in a hurry though, been reading about how to port the jlime distro from oe to oe-core Mar 07 10:35:40 wget is started, 404, git clone works Mar 07 10:37:09 hiphop php oh my god Mar 07 10:37:26 woglinde: ;) Mar 07 10:37:37 what comes next Mar 07 10:37:40 http://pastebin.com/U56WkETA shows log.do_Fetch for linux-linaro and for hiphop-vm Mar 07 10:37:57 who will find out why linux-linaro does not fetch? Mar 07 10:38:11 woglinde: ;) Mar 07 13:13:23 hello friends, i have a @ in my src_uri - how can i mask this @? Mar 07 13:15:53 ? Mar 07 13:16:04 paste your src_uri Mar 07 13:16:16 or an example Mar 07 13:16:59 :-) SRC_URI = "git://git@example.com/username@mail/testrepo" Mar 07 13:17:03 sorry Mar 07 13:17:42 It is abit tricky - the repos are maintained by gitolite and the user have access with her mail-addresses Mar 07 13:18:11 depends... is the recipe intended to be public or shared privately between developers? Mar 07 13:18:41 its shared privately Mar 07 13:19:10 ok, in that case I would suggest using a variable which each user has to set in their local.conf Mar 07 13:19:59 ah okay - and if its more public style? Mar 07 13:20:13 The repo is shared between more developers tha only one Mar 07 13:21:29 for a public recipe I think you'd need to have the base recipe use a publicly-accessible read-only version of the code by default Mar 07 13:22:15 then developers who need to use their own repo could use an extra layer with a bbappend to change SRC_URI for the recipe Mar 07 13:25:54 thank you for your fast response bluelightning! Mar 07 14:57:40 Is there a way to force bitbake to skip tasks or only run the task you request? Mar 07 14:58:55 I'm testing changes to a custom image recipe I'm creating and just want it to take whatever do_rootfs is laying around. But do_rootfs[nostamp] is set so it's running every time. Mar 07 15:13:36 rcw: you may find incremental image building useful for that... unless you're using master it's only supported for rpm though Mar 07 15:15:22 hmm actually maybe the ipk/deb patches never landed for that Mar 07 15:16:41 bluelightning: I would be interested in those for ipk :) Mar 07 15:17:28 bluelightning: do you know if incremental rootfs is calling postrm scripts? e.g. between package upgrades or when some package was part of older rootfs but is not pulled to newer? Mar 07 15:18:52 hmm actually it looks like my memory is faulty, the recent patchset was just for deb Mar 07 15:19:05 wouldn't be too difficult to implement also for ipk though I wouldn't think Mar 07 15:19:38 JaMa: I'm not sure, I'm not very familiar with the feature since I've never found reason to use it Mar 07 15:27:15 bluelightning: fair enough, I had some task to improve do_rootfs time and found out that we were using tar.bz2 in IMAGE_FSTYPES and were spending 2mins in bzip2 when pbzip2/lbzip2 wasn't installed on host Mar 07 15:27:25 bluelightning: so probably good enough improvement for now Mar 07 15:28:16 bluelightning: and I was thinking about postrm, because in one recipe I've noticed "rm -rf ${sysconfdir}/foo" without $D prefix so if incremental rootfs calls it, host's foo will be removed Mar 07 15:28:45 or rather it will fail Mar 07 15:28:56 yes depends on env Mar 07 15:29:05 but it would be a bit sad to have to run postrms at first boot on the target... Mar 07 15:29:07 in my minimalistic chroot everything is owned by bitbake Mar 07 15:29:11 so I will notice it :) Mar 07 15:29:39 bluelightning: not on target, only in incremental rootfs construction Mar 07 15:30:06 bluelightning: opkg calls them when it finds orphan packages and decides to remove them Mar 07 15:32:52 I guess postrms running at do_rootfs hasn't been something that has been considered in the past Mar 07 16:00:33 bluelightning: Thanks, I'll check out the incremental image building. I had noticed it, but was wary about using it with IMAGE_FSTYPE="live". Mar 07 16:02:23 is there any documentation discussing how to port an oe-classic distro to oe-core (apart from http://www.openembedded.org/wiki/Migrating_metadata_to_OE-Core) ? Mar 07 16:04:54 apelete: I don't believe there is... you may be able to compare the angstrom distro configs between OE-Classic and meta-angstrom as far as the distro config goes Mar 07 16:05:43 args still the rpmatch problem in linux-utils with uclibc Mar 07 16:06:11 I should sent the patches Mar 07 16:17:21 bluelightning: thought so, I guees I'll be reading a few distro configs to figure it out then. thanks. Mar 07 16:18:33 apelete: it all depends how much special stuff is in your distro config Mar 07 16:18:46 apelete: if you could point me at it I might be able to make suggestions Mar 07 16:40:51 bluelightning: here is it: Mar 07 16:40:51 distro -> http://cgit.openembedded.org/openembedded/tree/conf/distro/jlime-2010.1.conf?h=2011.03-maintenance Mar 07 16:40:51 machine -> http://cgit.openembedded.org/openembedded/tree/conf/machine/ben-nanonote.conf?h=2011.03-maintenance Mar 07 16:40:51 image recipe -> http://cgit.openembedded.org/openembedded/tree/recipes/images/muffinman-image.bb?h=2011.03-maintenance Mar 07 16:42:17 bluelightning: that's my starting point, maybe something else will be needed, don't know yet Mar 07 16:43:29 apelete: ok there are a bunch of things in the distro config Mar 07 16:43:43 we probably should have a page listing the needed changes... Mar 07 16:43:55 if I get a moment I will try to create one Mar 07 16:45:24 would be great indeed Mar 07 16:45:39 image recipe should be ok with no changes Mar 07 16:46:09 hmm, do we not already have a nanonote machine config anywhere? Mar 07 16:46:21 if not, some tweaks to that might be needed Mar 07 16:50:30 not sure for the nanonote machine config, I'll browse oe-core and meta-oe just in case. Mar 07 16:56:34 apelete: my searches just now haven't turned anything up so I suspect there isn't one yet Mar 07 17:01:56 bluelightning: okay. I guess the machine config will be part of the "jlime layer" (or whatever it will be) Mar 07 17:03:28 apelete: well, we try to keep distros, non-distro specific software and machines separate these days Mar 07 17:03:43 apelete: so ideally there would be a layer for nanonote or machines in its class Mar 07 17:04:25 it could be added to meta-handheld, it might sort of fit in there Mar 07 17:04:35 unless you'd prefer to maintain it in a meta-nanonote layer Mar 07 17:06:30 bluelightning: I was thinking about meta-handheld too (saw it on http://www.openembedded.org/wiki/LayerIndex) Mar 07 17:07:40 it might be a better choice for maintainability. Mar 07 17:11:51 oh, and there is support for jornada machines in meta-handheld already. since jlime also runs on hp jornada and a few other devices, going meta-handled seems a good choice indeed Mar 07 17:12:35 s/meta-handled/meta-handheld/ Mar 07 17:12:54 indeed Mar 07 17:13:10 for my part as maintainer nanonote would be welcome there :) Mar 07 17:13:26 nice :) Mar 07 17:13:41 FYI re jornada, those machines have never been built since the move to OE-Core Mar 07 17:14:00 so they would need work if you're interested in them Mar 07 17:14:52 I don't have one, but if the jlime port goes well some more work may be done I think. Mar 07 17:15:41 I do happen to have a couple (and have run jlime on them :) but don't really have time to work on them unfortunately Mar 07 17:17:01 bluelightning: the more devices I play with, the funnier. I may get one at some point (the nanonote was really an excuse for me to play with oe, so a jornada will do just as good). Mar 07 17:17:55 I went through a phase of buying old devices on ebay... they're mostly just taking up space in my cupboards now Mar 07 17:18:49 I've been trying to whittle down the number of computers and gadgets in my house, slowly but surely. kind of nice for clutter reduction, anyway Mar 07 17:19:00 though I think I still may have a couple zauruses in the garage Mar 07 17:20:34 ending up with a gadget clutter is ok for me as long as I learn something from tinkering with them. Mar 07 17:20:53 * kergoth nods Mar 07 17:21:12 I need to finish setting up the FRI2. anyone done anything with theirs yet? Mar 07 17:23:09 bluelightning: I'll look into this oe-core stuff, starting with jlime and the nanonote, I'll let you know how it turns out. Mar 07 17:23:51 apelete: ok, great! Mar 07 17:27:02 kergoth: Nice, you still have Zaurii :) Mar 07 17:27:08 I have some pieces of my old one. Mar 07 17:32:28 hehe Mar 07 17:32:54 I have a sl-5600 (iirc) with a unapproved-by-the-fcc warning, and a c500 Mar 07 17:33:27 Anyone have a qemu machine layer handy for machines other than the versatile? e.g. one of hte cortex a8, a9, a15, arm1136 boards? Mar 07 17:34:02 the machine config is trivial, but linux-yocto is still voodoo to me Mar 07 17:34:23 * mr_science just gave away his 5000D Mar 07 17:35:06 I think i tossed mine after i cracked the screen Mar 07 17:35:07 hehe Mar 07 17:36:07 hmm, meta-ivi has a vexpressa9 config Mar 07 17:53:19 * mr_science signs kergoth up for SCaLE 12x Mar 07 18:02:08 kergoth: Wher are you working these days? Mar 07 18:03:19 mewyn: Mentor Graphics Mar 07 18:05:38 *nod* I used some of their stuff in school recently :) Mar 07 18:06:57 Some of the VHDL stuff for one of the capstone classes Mar 07 18:07:15 (Of note, I went back to school in '09 :) Mar 07 18:07:25 I've since then graduated Mar 07 18:51:40 kergoth: i'll ask you about that later Mar 08 01:06:58 the meta-raspberry seems to build for softfp floatin point, does someone know how to switch to hardfp ? **** ENDING LOGGING AT Fri Mar 08 02:59:58 2013