**** BEGIN LOGGING AT Thu May 05 02:59:57 2011 May 05 07:57:04 hello. I have a problem with building angstrom-minimal-image. http://pastebin.com/mF2862r7 I don't understand how to get it fixed. My coworker claims he has absolutely the same settings and it runs smoothly on his system May 05 07:58:07 slider: do you have mercurial installed? May 05 07:58:14 on your host May 05 07:58:19 yes May 05 07:59:03 then try with -v -D to see more info May 05 07:59:38 hi all. i tried build image with uclibc. if i set USE_NLS="yes" i have message "preferred version 0.9.31 of uclibc not available (for item virtual/libiconv)". how to fix it? May 05 08:04:52 JaMa|Off: this is the log I got: http://fpaste.org/v2aQ/ It doesn't make anything clear to me May 05 08:08:18 JaMa|Off: should I repost the log so that the lines would be wrapped ? May 05 08:09:06 slider: try bitbake -e -b /home/tehnik/Workspace/tmp/oe/openembedded/recipes/cacao/cacao-native_hg.bb > /home/tehnik/Workspace/tmp/oe/openembedded/recipes/cacao/cacao-native_hg.bb.env and check SRC_URI there May 05 08:09:18 maybe it will fail too May 05 08:11:03 there is only one line there: ERROR: Please set the 'PERSISTENT_DIR' or 'CACHE' variable. May 05 08:30:08 XorA: ??? May 05 08:30:31 ant_work: what? May 05 08:30:41 why not fly to mars? May 05 08:30:45 :) May 05 08:31:12 we just want to be one better than mark shuttleworth May 05 08:31:34 because there is terrible delay when you're running builds on Earth over ssh... May 05 08:32:19 ant_work: the explanation email was just slightly later arriving than the poll I think May 05 08:32:33 btw the apostrophe b0rks email-sender May 05 08:34:16 ant_work: not sure what ypou mean May 05 09:10:24 XorA: nice poll May 05 09:20:00 heh, yeah May 05 09:20:03 * mickey|office voted May 05 09:20:44 * likewise wonders what moon exactly May 05 09:20:53 * likewise but voted anyway May 05 09:21:33 likewise: well its narrowed by the fact that the moon of choice needs to have solid surface :-D May 05 09:22:56 does earth moon have a name other than "The Moon"?? May 05 09:23:04 seems I've missed that poll :) May 05 09:23:14 Jay7: are you a member? May 05 09:23:21 ah, no :) May 05 09:23:26 that's reason :) May 05 09:23:33 it went out to eV members, see become a member :-D May 05 09:23:53 I don't know why to become.. May 05 09:24:12 I'm not active :) May 05 09:24:19 Jay7: then you can choose the board and TSC and change OE directions May 05 09:24:49 XorA: Please vote on which member of the TSC is most likely to moon first. :) May 05 09:25:08 likewise: hehe May 05 09:25:10 I'm sure there exists more experienced people to become members :) May 05 09:25:16 *membership May 05 09:25:40 but I'll consider :) May 05 09:26:03 Crofton: just did a call for members May 05 09:29:31 morning all May 05 09:35:27 Could anybody tell me why I get the message "ERROR: SRCREV was used yet no valid SCM was found in SRC_URI" when trying to invoke bitbake angstrom-minimal-image ? May 05 09:35:53 03Florian Boor  07org.openembedded.dev * rebb64cf105 10openembedded.git/recipes/lsb/lsb-base_3.2.bb: May 05 09:35:53 lsb-base: Move to a newer, fetchable source. Add checksums May 05 09:35:53 Signed-off-by: Florian Boor May 05 09:36:06 03Florian Boor  07org.openembedded.dev * r085dc5e99a 10openembedded.git/recipes/fakeroot/ (3 files in 2 dirs): (log message trimmed) May 05 09:36:06 fakeroot: Move from not fetchable 1.14.5 to 1.15.1 May 05 09:36:06 Fakeroot 1.14.5 was not fetchable any more. Updating to 1.15.1. May 05 09:36:06 Test environment: May 05 09:36:06 BB_VERSION = "1.10.2" May 05 09:36:06 METADATA_BRANCH = "org.openembedded.dev" May 05 09:36:06 TARGET_ARCH = "arm" May 05 09:37:22 hi bluelightning May 05 09:52:00 florian: heh, you dumped an old branch? funny to see all those 2010 commits May 05 09:54:37 mickey|office: sort of... you can see that it's hard for me to find time to clean up May 05 09:54:46 *nod* May 05 09:55:25 slider: it means exactly what it says - SRCREV_pn-XXX was specified for a package but that package in your configuration has a SRC_URI that does not specify a version control system to which that SRCREV could be applied May 05 09:55:37 slider: which recipe is it actually complaining about? May 05 09:55:47 Jay7: became a member May 05 09:57:00 mickey|office: I have been working with oe much more than it looks like. Building large feeds for multiple architectures is a good thing to identify issues. May 05 10:00:01 florian: that's right May 05 10:01:09 bluelightning: thank you for your answer. I'm new to OE and thus don't know how to fix it. I have a build log at: http://pastebin.com/mF2862r7 and more detailed one at http://fpaste.org/v2aQ/ May 05 10:02:00 slider: ok so it seems to be "cacao-native" that it's complaining about May 05 10:02:34 bluelightning: the interesting fact is that with the same settings my coworker doesn't experience this problem. May 05 10:02:47 what am I to do to fix it? May 05 10:03:52 I'm impressed... noone complaining about my patches I sent to the list May 05 10:05:10 slider: hmm, I can't see why it's complaining, it has an hg:// URI May 05 10:05:33 slider: also you can't have the same setup as your coworker or you'd be getting the same results May 05 10:06:06 bluelightning: I have his configuration cloned. May 05 10:07:26 so he suspects me in intentional sabotaging the work ("everything works on my side!") :( May 05 10:08:03 you're using the exact same version of bitbake? May 05 10:09:57 not sure May 05 10:10:25 he isn't talking to me anymore. Thus I'm with this bug on my own May 05 10:20:39 slider: are you using any customisations or just a local.conf and OE release/master? May 05 10:22:21 Jay7, you are overqualified to be an eV member :) May 05 10:22:34 hey hey Mr Crofton May 05 10:22:40 gm May 05 10:22:43 oh :) May 05 10:22:48 I'll vore after coffee May 05 10:23:00 where to leave request then? :-D May 05 10:23:04 seriously, we need people active in the community to join May 05 10:23:06 Crofton: site is really easy to drive May 05 10:23:15 email me and I will add you to the list May 05 10:23:55 bluelightning: it's just a local.conf.sample copied to local.conf and slightly modified (to choose x86-angstrom-2010.x). And that's all. Nothing other than that was changed. May 05 10:24:40 slider: ok, can you pastebin your local.conf just so I can have a look? May 05 10:24:47 sure. just a moment May 05 10:26:48 Crofton: done May 05 10:27:08 I suspect that now I should subscribe to some ML? :) May 05 10:27:30 no, we have to vote for you first :) May 05 10:27:34 bluelightning: here it is http://fpaste.org/rA1N/ May 05 10:27:38 ah :)) May 05 10:28:41 well, about the local.conf.example May 05 10:29:00 It's me or there is unnecessary cruft? May 05 10:29:49 # BBMASK = "/(nonworking|obsolete)/" -> in bitbake.conf May 05 10:30:20 ASSUME_PROVIDED += "help2man-native" that's wrong explanation imho May 05 10:30:47 # Uncomment this if you are building Linux 2.4 Embedix kernels. -> no comments about 2.4 May 05 10:31:53 yeah, there is loads of cruft in there May 05 10:33:38 ant_work: check for angstrom-2008.1 May 05 10:33:49 iirc it was still in local.conf.sample May 05 10:34:04 slider: what are you actually trying to build? there's no such thing as angstrom-minimal-image as an image target... May 05 10:35:10 bluelightning: a minimal image of the anstrom distribution for x86 May 05 10:35:45 slider: so minimal-image then? May 05 10:35:59 * Jay7 is gone May 05 10:36:08 my coworker uses the command "bitbake minimal-image" to build it. I've seen people use angstrom-minimal-image parameter. I tried both. Both have the same result and the same mistakes May 05 10:36:13 Jay7, can you send a little on what you do with OE May 05 10:36:23 your background May 05 10:36:51 Crofton|work: ah, sure May 05 10:38:07 Crofton|work: even easier: http://wiki.openembedded.org/index.php/User:Jay7 May 05 10:38:31 I'll use that :) May 05 10:39:10 add there * trying to create new buildstats collecting and reporting infrastructure together with Yocto project May 05 10:39:37 I just cut and pasted May 05 10:39:39 well, I've updated that page May 05 10:42:09 slider: ok so I just tried that conf file here and it got past the parsing phase... so it's something amiss on your system May 05 10:42:36 slider: suggestions: check that BBFILES is correct for your system (also that OEBASE is correctly set) May 05 10:42:41 bluelightning: I have installed all the packages required.. (as specified on wiki) May 05 10:42:52 also what version of bitbake are you using? May 05 10:43:25 BitBake Build Tool Core version 1.8.18, bitbake version 1.8.18 May 05 10:43:34 printenv |grep BB BBPATH=/home/tehnik/Workspace/tmp/oe/build:/home/tehnik/Workspace/tmp/oe/openembedded BB_ENV_EXTRAWHITE=OEBASE May 05 10:44:39 so what is OEBASE set to? May 05 10:45:43 it's set to /home/tehnik/Workspace/tmp/oe May 05 10:46:25 should I try a newer version of bitbake? May 05 10:46:27 hmm, ok so that would appear to match your system May 05 10:46:33 I'm just testing with 1.8.18 here May 05 10:47:10 slider: are you using OE master? or a release version? May 05 10:47:56 slider: ok I have just reproduced the issue with 1.8.18 May 05 10:48:08 perhaps it would be a good idea to upgrade May 05 10:48:33 what version were you using known to work fine? May 05 10:50:36 slider: I was using 1.12.0 May 05 10:51:13 bluelightning: thank you for helping me. I'm updating bitbake May 05 10:52:28 this might be a good time to up our recommended bitbake version everywhere... I see the OE manual says 1.8.18 is still the current version (!) May 05 10:54:29 slider: no problem May 05 11:32:08 bluelightning: now the image is being built successfully with bb-1.10.2. thank you one more time May 05 11:43:29 03Simon Busch  07master * r3b7d833620 10openembedded.git/classes/cmake.bbclass: May 05 11:43:29 cmake.bbclass: set the rpath for native binaries build with cmake May 05 11:43:29 In the default configuration cmake does not set a rpath for its builded binaries. This May 05 11:43:29 leads to errors at runtime when the binaries not find their needed runtime libraries. May 05 11:43:29 Please note that this is only done for recipes inheriting the native class! May 05 11:43:30 Signed-off-by: Simon Busch May 05 11:43:31 Acked-by: Phil Blundell May 05 12:41:04 slider: cool, glad things worked out :) May 05 12:43:07 bluelightning: almost... it failed at: package bison-native-2.4.3-r9.0: task do_compile: Failed May 05 12:43:30 slider: what's the failure? May 05 12:44:18 bluelightning: http://fpaste.org/B9Zo/ here is the log May 05 12:44:42 can anyone help me how do I configure Idesk..... when I execute Idesk .... I getting this errors .... http://pastebin.mozilla.org/1219055 May 05 12:45:15 slider: so my first guess would be that this is caused by ASSUME_PROVIDED += "help2man-native" that you have in local.conf May 05 12:46:10 but my host provides this tool. I have it installed May 05 12:46:22 maybe I should update it too May 05 12:48:18 slider: it seems that the host provided one is not the version that bison expects at least May 05 12:49:59 I've updated it. If building fails again, I'll comment that line May 05 12:50:44 ok, fair enough May 05 12:51:07 hmm.. strangely I have added it (and linux-libc-headers too) to A_P May 05 12:51:29 I remember Tartarus talking about help2man-native May 05 12:51:30 NOTE: package bison-native-2.4.3-r9.0: task do_compile: Succeeded May 05 12:51:42 help2man: Add native recipe, drop check for being provided. May 05 12:53:32 help2man has been in the 'required-software-on-buildhost' since years May 05 13:22:24 I want my embedded device to have a netBIOS name. I'm using busybox. How could i do that ? May 05 13:23:59 busybox is irrelevant, you need samba for that May 05 13:24:44 did some update to the latest samba releases? May 05 13:28:07 pb_: So i'm on the good way, i just cross compiled samba, but there are 30 binaries which have been built, which one should i use ? May 05 13:28:44 nmbd is the one you need to advertise a name; smbd is the one which provides services. May 05 13:29:06 pb_: thanks May 05 13:29:25 have someone noticed qt checksum change? May 05 13:31:12 no May 05 13:31:15 4.7.3 is out May 05 13:31:54 woglinde: in any case oe-core fails since it doesn't match May 05 13:33:02 otavio didnt yet make a testrun with oe-core and oe-meta May 05 13:45:20 morning May 05 13:45:47 he kergoth May 05 13:51:53 g'day kergoth May 05 14:45:44 kergoth: morning :-) May 05 14:56:45 03Tom Rini  07master * r05143e9b5d 10openembedded.git/recipes/gcc/gcc-package-target.inc: May 05 14:56:45 gcc-package-target.inc: Switch to TARGET_PREFIX in symlinks May 05 14:56:45 When TARGET_PREFIX and TARGET_SYS didn't match up, the symlinks we made May 05 14:56:45 were invalid. TARGET_PREFIX is what we use when configuring so it's what we May 05 14:56:45 should be using here. May 05 14:56:46 Signed-off-by: Tom Rini May 05 14:57:11 03Tom Rini  07master * r79b497edc0 10openembedded.git/recipes/binutils/binutils.inc: May 05 14:57:11 binutils.inc: Switch to TARGET_PREFIX not TARGET_SYS May 05 14:57:11 When TARGET_PREFIX and TARGET_SYS didn't match up, the symlinks and May 05 14:57:11 update-alternatives weren't working. TARGET_PREFIX is what we use when May 05 14:57:12 configuring so it's what we should be using here. May 05 14:57:12 Signed-off-by: Tom Rini May 05 14:57:17 03Tom Rini  07master * r4a022cf3fe 10openembedded.git/recipes/sysstat/sysstat.inc: May 05 14:57:17 sysstat: Switch to inherit gettext May 05 14:57:17 This needs to both virtual/libnl as well as gettext-native in order to May 05 14:57:17 build. So we switch to using inherit gettext. This was compile tested May 05 14:57:17 with qemuarm+angstrom-2010.x+eglibc and qemuarm+minimal-uclibc. May 05 14:57:17 Signed-off-by: Tom Rini May 05 14:57:18 03Tom Rini  07master * r9bd73a0623 10openembedded.git/recipes/busybox/ (busybox-config.inc busybox.inc): May 05 14:57:18 busybox: Add PING6 when DISTRO_FEATURES has ipv6 May 05 14:57:19 Signed-off-by: Tom Rini May 05 14:57:19 03Tom Rini  07master * rf0652d96ff 10openembedded.git/recipes/bash/bash.inc: May 05 14:57:20 bash: Add bison-native to DEPENDS May 05 16:02:19 anyone know of some recipes in oe for tools to agregate network statistics; packet error rate; packet throughput; latency; roundtrip delay May 05 16:02:59 jconnly hm May 05 16:04:17 I would made something with collectd maybee May 05 16:38:27 what exactly is the relationship between the openembedded and the openembedded-core git repos at the moment? I'm still finding my way around, and hadn't been able to suss that out completely yet May 05 16:39:47 liam oe-core is the future May 05 16:40:25 in the future, the metadata is more broken up, and oe-core is the stable, minimal subset of OE which is the core of it all May 05 16:40:41 most stuff of oe master goes to oe-core and oe-meta May 05 16:40:57 s/most/most maintained/ May 05 16:41:08 :) May 05 16:42:49 i see - so if I'm starting a new project, is there a recommended way to get on board with oe-core? or is it easier to stick with the current setup, since that's what all the existing doc seems to support? May 05 16:43:20 oe-core comes with a basic env setup script, you can run that to set up a build directory, then modify to add meta-oe, etc to it May 05 16:43:27 or you can use angstrom's setup scripts May 05 16:43:32 which has a branch, iirc, for the new stuff May 05 16:43:49 i would definitely suggest using oe-core if you're starting something, personally May 05 16:44:05 yes angstrom does have oe-core based setup May 05 16:44:31 interesting. don't suppose there's any doc on this approach available? :) May 05 16:44:40 i should make a quick and dirty git repo with a minimal oe-core based setup using my personal way of doing things May 05 16:44:42 or just going through the angstrom scripts is the way to go? May 05 16:44:51 i can't even get oe-core to parse ;( May 05 16:45:01 heh May 05 16:45:04 liamMT: you don't *have* to use the new mechanisms, you could use all these layers just by adding them to BBPATH and BBFILES, the same way you use OE May 05 16:45:07 oe-core hates me May 05 16:45:15 it's just taht most are also moving to the new bitbake built in support for layering May 05 16:45:17 'bblayers' May 05 16:45:35 foerster: git clone https://github.com/kergoth/reloc-tests; cd reloc-tests; . ./setup.sh; bitbake -p May 05 16:45:38 see if that works May 05 16:46:40 kergoth: ok, sounds like I still have a fair amount of familiarization ahead of me :) thanks May 05 16:46:56 not sure if there's a good summary of bblayers.. May 05 16:46:57 kergoth: ok, let me try May 05 16:46:58 try googling May 05 16:48:05 liamMT: the cool thing about bblayers is it automatically finds a conf/bblayers.conf by searching up from your current location, and it + the config files in teh layers does all the setup May 05 16:48:17 this means you can avoid setting any env vars but adding bitbake to PATH, now May 05 16:48:21 hi khem May 05 16:48:23 just cd into your build area and go May 05 16:48:25 which is lovely May 05 16:49:11 kergoth: had to --recursive in there too. Cloning in progress. May 05 16:49:19 kergoth: ok, cool May 05 16:49:30 foerster: oh, yes, forgot about that, sorry May 05 16:49:47 * kergoth makes a version without the reloc-tests scripts May 05 16:50:11 well, it's parsing. So it hates something in my setup apparently. May 05 16:50:32 okay, thats what i was wondering. you can look at how that one is set up to compare against yours May 05 16:50:43 of course, if you use the build env setup script from oe-core it should work too May 05 16:50:47 * kergoth just doesn't like it much May 05 16:50:52 i did that - throws a stupid error May 05 16:50:59 hmm, odd May 05 16:51:17 throws a "error parsing, no split on NoneType" or similar May 05 16:51:33 hmm, okay, will try using it here too, see if it does that May 05 16:51:50 kergoth: it's probably something in my conf that's wrong/missing. May 05 16:52:04 it's basically my local.conf from oe.dev May 05 16:52:09 I'm sure it's missing something that's needed. May 05 16:52:32 i recommend doing the base bbpath/bbfiles in bblayers.conf, and only settnig machine/distro/etc in the local.conf May 05 16:53:00 yes, I have all of the bbpath/files stuff commented out May 05 16:53:35 cool. hard to say exactly what it is speculatively, would require a bit more investigation :) May 05 16:53:42 hopefully it's not terribly painful May 05 16:53:46 yea, let me fire that up again May 05 16:54:00 * kergoth really isn't a fan of how bblayers works, today, but it's a hell of an improvement over the old way, at least May 05 16:54:53 ERROR: Error parsing /home/foerster/dev/openembedded/dev/oe-core/meta/recipes-extended/lsof/lsof_4.84.bb: 'NoneType' object has no attribute 'split' May 05 16:55:18 That's all it tells me ;) May 05 16:56:33 I'll go through the local.conf.sample and see if there's something that I'm missing May 05 17:02:40 ugh, we really need to improve handling of parsing errors May 05 17:02:44 kergoth: if one wants to include items from other layers, how do you recommend doing that with oe-core? Previously, I munged FILESPATHBASE to search my overlay then oe.dev May 05 17:02:51 i don't want to show a traceback, but we do need *context* May 05 17:03:45 foerster: for include/require, just use the path relative to the top of the bbpath entry. require recipes/foo/bar.inc. for filespath stuff, you can either mangle filespathbase globally the way we used to do with collections, or can have a .bbappend add it May 05 17:03:45 foerster there are some examples (of course now I can't find them) on adding layers and setting up the layer configurations.. May 05 17:04:15 ahh if the question is includes, then kergoth is right.... May 05 17:04:49 So I'll have to do that in my overlays bblayer.conf then, right? May 05 17:05:00 s/overlay/layer/ May 05 17:05:21 usually May 05 17:05:47 * foerster kicks self for using amend instead of bbappend May 05 17:09:24 d'oh, recipes/tasks/task-sdk* => recipes-core/tasks/task-sdk* May 05 17:09:37 heh, indeed May 05 17:10:26 kergoth: fyi - it hates something in my local.conf. I replace with your minimal, and it parses (til it couldn't find recipes/tasks/*) May 05 17:10:41 ah, good that you have it narrowed down May 05 17:12:08 for reference, https://github.com/kergoth/basic-oe-core-setup is my preferred setup minus the reloc testing stuff May 05 17:12:12 * kergoth goes to take the dog out May 05 17:12:38 kergoth: thanks May 05 17:18:29 the main difference between how i do things and how the core setup scripts (and others) do things is that I do my day to day work on the metadata, and the changes are often invasive, so it makes a lot of sense to be able to keep my changes isolated. I don't want my prototype of this or that hosing my builds in my other project areas May 05 17:18:44 so the metadata is more first class, part of the project, rather than the project being config + build output May 05 17:19:00 plus, disk space is cheap May 05 17:19:02 :) May 05 17:20:18 damn, it still hates me :) May 05 17:20:22 damn May 05 17:20:26 Failure expanding variable METADATA_REVISION, expression was ${@base_detect_revision(d)} which triggered exception ValueError: substring not found May 05 17:20:44 hm. why is that looking for "packages" in BBPATH? May 05 17:20:52 hm, that stuff needs a rework. it should assemble a list of revisions of all the layers May 05 17:26:56 ok, order of BBLAYERS seems important (oe-core, meta-oe seemingly must be before my layer) May 05 17:27:55 what was the libc dependent _append_ again? May 05 17:28:33 yes, bblayers specifies the order in which to look for the recipe's, bbappends, etc.. May 05 17:29:36 technically, it only controls the layer.conf parse order. if a layer.conf prepends to bbpath rather than appending, you'll end up with something you likely didn't expect (part of why i didn't like the implementation). also, that only controls the bbpath May 05 17:29:56 it does *not* control the priority of the layers wrt recipes, only wrt config files. BBFILE_PRIORITY_* in teh layer.confs control that May 05 17:30:07 (the other reason i dislike it) May 05 17:30:27 ah. _append_libc_$thing. There are two (potential) bugs with that: recipes/lm_sensors/lmsensors-apps_2.10.8.bb:SRC_URI_append_uclibc May 05 17:30:31 and recipes/meta/oplinux-packages.bb:OPLINUX_BROKEN_PACKAGES_append_uclibc-linux May 05 17:30:37 and if a layer is prepending to bbpath I'd content it's implementation is likely broken for the "normal" use-case May 05 17:30:58 (just as a sidenote) May 05 17:31:28 meta-oe: BBPATH := "${LAYERDIR}:${BBPATH}" May 05 17:31:32 I agree. I'm just pointing out that the knowledge of layer priority/importance isn't held in a single place. May 05 17:31:38 so meta-oe should not be used as an example? May 05 17:32:09 yeah, that's wrong. also the := hasn't been necessary for quite some time (need to confirm 1.12 has that fix, pretty sure it does though) May 05 17:32:28 ya, I agree that does look wrong.. likely something to point out on the oe, oe-core, or another list.. May 05 17:32:56 I know there were some hacks in meta-oe to resolve some issues, but they were supposed to be going away last week.. (maybe they didn't?) May 05 17:32:59 BBPATH .= ":${LAYERDIR}" is likely what it should be May 05 17:33:11 * foerster thinks if the main repos are wrong, we're all doomed to follow down the path to pain May 05 17:33:31 perhaps someone wants to fix those two, like: http://paste.debian.net/116098/ May 05 17:33:34 agreed, they're the role models, as it were May 05 17:33:43 foerster intention is that meta-oe is supposed to be a "good" example on how to do various things.. so if it's wrong we need to fix it.. if there are workarounds for other issues, they were supposed to be noted... May 05 17:34:33 fray: noted. As such, perhaps it's too early for a mere mortal such as myself to embark on using oe-core/meta-oe :) May 05 17:35:00 they are still rough, but the TSC expects folks to start using them soon, if the announcement didn't already go out.. May 05 17:35:09 so any problems like this should be reported so they can be fixed.. May 05 17:35:23 I will say it won't be painless though, not at first.. May 05 17:35:31 hmm, wonder how much work it'd be to try to consolidate the knowledge of layer priority. maybe just something at ConfigParsed that (in a stable way) resorts BBPATH to match BBFILE_PRIORITY_*.. i think i'll try that May 05 17:36:02 http://gitorious.org/angstrom/meta-angstrom/blobs/master/conf/layer.conf May 05 17:36:08 angstrom prepends BBPATH as well May 05 17:36:18 heh, copied the wrong bits from meta-oe May 05 17:36:26 (and the same person setup both angstrom and meta-oe.. meta-oe came from Angstrom) May 05 17:36:36 ah, right May 05 17:36:38 ah May 05 17:37:08 point is, we need to get it fixed.. if angstrom needs to do something in a strange order -- then they can as a distro.. but oe-core and meta-oe should not have to do that, other then as a temporary (docuemtned) workaround May 05 17:37:08 my oe-fu is growing, but far from good yet May 05 17:37:37 since i'm unknowing - what exactly is wrong with prepending? what problems does it cause? May 05 17:38:12 foerster prepending is fine -- other then it changes the expected ordering of things.. May 05 17:38:38 the expected order if someone looks at the main bblayer variable is the order listed there.. by prepending we've violated the users expectations May 05 17:38:49 (again, there may be valid reasons to do that -- but not in OE) May 05 17:38:59 ok, thanks. May 05 17:42:50 oh, i remember now why i couldn't do bbpath adjustment to consolidate that knowledge, it was because by the time config parsed is hit we've already parsed bitbake.conf, wuold have to reexec bitbake the way collections did to implement in the metadata May 05 17:43:11 hmm, maybe could teach bitbake to add a layer to bbpath + bbfiles + bbfile_*, only if the layer doesn't do so explicitly May 05 17:43:34 then we'd have BBLAYERS + expressed layer dependencies via whatever mechanism we have for that, influencing the priority May 05 17:44:46 foerster: heh, https://github.com/kergoth/OE-BBLayers was an example of a correct layer.conf, uploaded april 2010 :P May 05 17:44:48 * kergoth rolls eyes May 05 17:45:12 the poky manual has it properly defined too it seems. May 05 17:45:18 good good May 05 17:45:19 Just the current in-use references are wrong. May 05 17:46:15 ok, well I'm at least getting closer to oe-core working for me. May 05 17:46:28 Seems the thing is mostly parsing properly, so my setup must be close. May 05 17:46:55 but things moved/migrated/changed, and I'll have to figure out how to update my stuff (eg: bootstrap-image seems gone). May 05 17:47:13 * kergoth nods May 05 17:47:20 see the core-image-* recipes May 05 17:47:50 I see those, but they don't look to use MACHINE_TASK_PROVIDER, so I'm unsure how they map. May 05 17:47:56 ah May 05 17:48:47 nonetheless, it's getting closer. I will probably have to shelve this for a while, but I think it's something I can hopefully do soon. May 05 17:50:13 It looks like angstrom has some example images I can learn from. May 05 18:10:43 kergoth: /scratch/src/oe/openembedded/bin/install -c --mode 755 foo /tmp/bar May 05 18:10:51 yields: /scratch/src/oe/openembedded/bin/install: illegal option -- - May 05 18:11:00 it does, yes May 05 18:23:14 kergoth: don't we have an automatic fixup in autotools.bbclass before autoreconf'ing ? May 05 18:23:29 you'll need to elaborate May 05 18:23:35 fixup of what? May 05 18:23:47 s/--mode/-m/ May 05 18:24:05 no, we don't May 05 18:24:25 should we have it? May 05 18:24:31 in general i'm not a big fan of magic sed's of tons of files, personally. we'd probably be better off teaching the install wrapper to do it May 05 19:50:40 how does one get a oe wiki account? May 05 19:58:08 wmat: send email to ka6sox-work May 05 19:58:21 anonymous edits are no longer allowed May 05 19:58:26 due to spamming May 05 20:16:29 khem: thx May 05 20:41:58 hmm... this was an odd error: /usr/bin/cp: No such file or directory May 05 20:45:11 risca: it's usually /bin/cp. What's that error from? May 05 20:47:12 It's from xproto-native May 05 20:47:23 I'm researching it right now May 05 20:47:45 I just didn't think anything would try to use cp from /usr/bin May 05 20:47:54 weird May 05 20:47:58 Totally May 05 20:49:20 Build log: http://pastebin.com/L1SR2NVf May 05 21:33:43 blindvt`: hmm, wonder how much work itd be to get the wrapper to handle --mode.. currently it uses getopts, which only handles short options, but which is more portable than getopt May 05 21:34:10 for meta-toolchain.bb - what is the LIBTOOL_SYSROOT_PATH intended for thats exported from the environment-setup script? May 05 21:34:57 it's for libtool.. May 05 21:35:09 makes sure /usr/lib gets corrected to be relative to the sysroot, etc.. May 05 21:36:20 is it for a specific version of libtool? May 05 21:47:17 Probably belongs here: Is anyone fixing fakeroot? 1.14.5 was dropped from the debian site: fakeroot story, 1.14.5 went away http://packages.qa.debian.org/f/fakeroot.html. Testing a fix, but looking at the dependencies it's not a recipe for beginners. May 05 22:18:50 03Florian Boor  07org.openembedded.dev * r4e0d30c804 10openembedded.git/recipes/libtool/libtool-sdk_2.2.6b.bb: May 05 22:18:50 libtool-sdk: postprocess libtoolize and install additional files. May 05 22:18:50 Install some additional files that might be missed by autotools. May 05 22:18:50 Libtoolize has a path to build directory inside we can get rid of by postprocessing. May 05 22:18:50 Signed-off-by: Florian Boor May 05 22:18:53 03Florian Boor  07org.openembedded.dev * r994b4ede18 10openembedded.git/recipes/initscripts/initscripts-1.0/checkroot: May 05 22:18:53 initscripts: Improve parsing of mount parameters in order to fix remounting of rw jffs2. May 05 22:18:53 This change filters additional mount parameters which can cause jffs2 filesystems to May 05 22:18:53 get remounted even if mounted rw already. May 05 22:18:53 Signed-off-by: Florian Boor May 05 22:18:54 03Florian Boor  07org.openembedded.dev * r150edcc04f 10openembedded.git/recipes/initscripts/initscripts_1.0.bb: May 05 22:18:54 initscripts: bump 1.0 PR May 05 22:18:54 Signed-off-by: Florian Boor May 05 23:24:25 wmat ping? May 05 23:27:02 wmat: you about? May 05 23:28:59 ka6sox-farfarawa: yep, i'm here May 05 23:29:06 are you actually on a mountain? May 05 23:29:18 man, i can wait on the username change May 05 23:29:58 yes, I'm on the mountaintop May 05 23:30:03 and you should have a new email? May 05 23:30:06 did you get it? May 05 23:32:27 * wmat looks May 05 23:33:25 nothing yet May 05 23:34:20 whew, finally May 05 23:34:29 fixed the lack of useful tracebacks from parse errors May 05 23:34:34 been meaning to do that for a while May 05 23:36:52 wmat: resent May 05 23:38:28 got it, thanks! May 05 23:38:43 okay you should be able to edit now May 05 23:39:00 yep, thanks May 05 23:39:36 749W TPO, HD1 + Analog , HD2 May 05 23:39:51 uh wrong place May 05 23:47:13 kergoth: yay :) May 05 23:59:12 foerster|away: http://kergoth.pastey.net/149716 May 05 23:59:29 note: also adds 3 lines of context from each file, the way we do for errors from recipes, and limits the number of entries to 5 May 05 23:59:36 think this makes it much more useful May 05 23:59:45 also note that it shows the values of the arguments for the functions May 05 23:59:54 File "/home/kergoth/code/oe-parseerrors/bitbake/lib/bb/siggen.py", line 107, in _build_data(self=, fn='virtual:nativesdk:/home/kergoth/code/oe-parseerrors/oe-core/meta/recipes-devtools/unfs-server/unfs-server_2.1+2.2beta47.bb', d=) May 05 23:59:59 bit verbose, but.. :) May 06 00:00:34 this is the same code for making tracebacks pickleable which i'm going to use to shift the traceback formatting into the UIs, rather than having to pre-format it in the logger May 06 00:04:30 the next bits beyond that to add will be syntax highlighting with pygments for the tracebacks May 06 00:04:37 which i had in the standaloen version of this May 06 00:46:18 http://pioneerspacesim.net/ May 06 00:46:22 nice game to be ported May 06 00:46:44 Elite-like :) May 06 01:07:44 how do I specify a subdir for the fetcher to unpack an archive to? May 06 02:21:58 hm - which packages should I include in my image if I want to be able to build with gcc on device? I added 'gcc' to IMAGE_INSTALL for a working build, and see a handful of libraries installed in /usr/libexec/gcc/arm-angstrom-linux-gnueabi/4.5.3 and /usr/lib/gcc/arm-angstrom-linux-gnueabi/4.5.3 but no executables. **** ENDING LOGGING AT Fri May 06 02:59:57 2011