**** BEGIN LOGGING AT Thu Nov 07 02:59:58 2013 Nov 07 08:46:40 morning all Nov 07 08:46:58 morning silvio Nov 07 09:36:23 morning all Nov 07 09:38:10 morning Nov 07 09:38:25 hi jama and bl Nov 07 09:40:13 hi JaMa, woglinde Nov 07 09:47:48 JaMa: the patch for make 4.0 support from the mailing list works for me Nov 07 09:47:59 I only built gstreamer with it so far though. but eglibc builds fine. Nov 07 09:48:08 next, I am building an entire rootfs Nov 07 09:48:31 you mentioned in bugzilla that the build fails later with your patch (I am using jonathan liu's). what does this mean? Nov 07 09:54:25 dv_: ssp is somehow incorrectly enabled in my build, maybe it's related to my patch changing autoconf as well (because I had 2.69 as well) Nov 07 09:55:11 hmm. I also have autoconf 2.69 on my system Nov 07 09:56:34 I've already downgraded make, because I need to use the same system also for dora builds Nov 07 09:59:46 btw this is already fixed in glibc and eglibc repos, so upstreaming would be inappropiate Nov 07 10:01:44 bluelightning, Hey. Nov 07 10:01:56 hi Stygia Nov 07 10:02:49 bluelightning, I haven't fully reviewed your pull request, but I'm going to have to disagree with the liburi one. Nov 07 10:02:53 bluelightning: btw thanks for the danny<->dora links Nov 07 10:03:02 bluelightning, I base my RPROVIDES on this: https://metacpan.org/release/URI Nov 07 10:03:06 the arguments were quite compelling Nov 07 10:03:28 It's possible liburi-perl doesn't provide all the things it should, but liburi-escape-perl (URI::Escape) certainly is an RPROVIDE of this package, I think. Nov 07 10:03:37 dv_: np Nov 07 10:03:41 I can't help but think Dora the Explorer whenever anyone says Dora. Nov 07 10:04:35 bluelightning, So in this case I think the OE-core is wrong. :) Nov 07 10:05:01 Stygia: it depends on whether you see it as a requirement to have an RPROVIDES matching all the perl modules provided by the package Nov 07 10:05:10 Stygia: I'm fairly sure we haven't done that in the past Nov 07 10:05:12 bluelightning, I sorta do. Nov 07 10:05:13 bluelightning, Oh. Nov 07 10:05:21 bluelightning, Well, I'd argue we should have that. Nov 07 10:05:26 bluelightning, I've been trying to do so in newer recipes. Nov 07 10:05:34 we could patch the recipe in OE-Core so it does Nov 07 10:05:40 bluelightning, Because, IMO, it's neat, and it RPROVIDES exactly what the actual package does. Nov 07 10:06:47 I can see some benefits Nov 07 10:08:18 It would be an issue in the future if I want to specifically depend on URI::Escape. Nov 07 10:08:29 I'd have to reduce my recipe to a package-wide dependency and not module-wide. Nov 07 10:08:40 I hope the difference between a perl package and module is obvious, right? Nov 07 10:10:54 bluelightning, The OE convention of course wins, but if it isn't a strong convention I'd lobby in favour of changing it. Nov 07 10:11:00 dv_: right, it wasn't there when I was preparing my patch (or I was blind) Nov 07 10:12:08 liu's patch is the same Nov 07 10:12:18 perhaps thats why it is working for me Nov 07 10:12:36 anyone fluent in German that could understand and make a measured response to the recent posting on oe-devel? Nov 07 10:13:12 understand, yes. make a measured response, no. :) Nov 07 10:13:54 I'm assuming the person is having trouble removing themselves from the list, but I can't be sure Nov 07 10:14:15 I thought it's spam :) Nov 07 10:14:29 it certainly mentions spam Nov 07 10:15:10 yeah, he tried to remove himself, but he still gets mail Nov 07 10:15:56 Stygia: well, if you prefer to keep liburi-escape-perl then we'll need to patch the liburi-perl recipe in OE-Core before we can proceed Nov 07 10:16:07 and you know the volume of emails the mailing list produces.. Nov 07 10:16:17 Stygia: I'm slightly more concerned by the other issues I found Nov 07 10:17:04 bluelightning, The rest seem like genuine problems with the recipes. Nov 07 10:17:17 The nonsense in SQL::Abstract I hope you'll pardon, that one was hard to get to compile at all. Nov 07 10:18:12 And I thought I'd added those .#filename# files to my gitignore Nov 07 10:18:16 Guess not. Nov 07 10:18:21 It's some emacs backup nonsense, so. Nov 07 10:19:44 bluelightning, Why did you delete the rdepends from Class::Accessor? Nov 07 10:20:48 Stygia: ah, I didn't... ? Nov 07 10:21:25 if you mean libclass-accessor-perl_0.34.bb, I only moved RPROVIDES down and dropped libclass-accessor-perl from its value Nov 07 10:21:27 bluelightning, Oh, wait, it first shows you deleting all of them, then adding everything but libclass-accessorr-perl which was reduntant. Nov 07 10:21:30 Never mind! Nov 07 10:21:35 It shows - for lines you slightly moved. Nov 07 10:21:50 right, but RPROVIDES not RDEPENDS Nov 07 10:22:07 Oh, right sorry. Nov 07 10:24:15 bluelightning, Am I reading this right if I read it as if you removed all the RPROVIDES from libdate-manip-perl? Nov 07 10:24:54 Stygia: er, ouch... that wasn't intended Nov 07 10:24:58 sorry Nov 07 10:25:08 bluelightning, Ah, no worries. :) Nov 07 10:25:32 look through the rest first to see if it's OK and then I can send a new pull request Nov 07 10:25:34 bluelightning, You're taken time out of your life to fix my silly recipes, I think we're probably not even close to even yet. Nov 07 10:25:46 bluelightning, Not to mention I'd have missed X number of deadlines without your advice, so. Nov 07 10:26:42 bluelightning, I'm going to assume you cut-pasted RDEPENDS Lines and such, so I'm not going to be comparing every single line for each change. Nov 07 10:26:58 yep, pretty much Nov 07 10:28:06 good morning Nov 07 10:28:17 bluelightning, What were we going to do about matching RPROVIDES to all modules for X recipe? Nov 07 10:28:17 Morning. Nov 07 10:28:31 hi mckoan Nov 07 10:28:48 Stygia: sorry, I'm not following Nov 07 10:29:17 bluelightning, Ah, you removed URI::Escape from liburi-perl, for example. Are we going to keep that in there since it's a provided module of the URI package? Nov 07 10:29:25 Or did you want to remove it since it wasn't an RPROVIDE of the meta-OE one? Nov 07 10:29:53 well, one or the other needs to happen Nov 07 10:30:13 I'm fine if you want to send a patch to fix the version of liburi-perl we already have Nov 07 10:30:22 I would rather have it patched into the OE-core one that whatever metaCPAN says the package provides w.r.t. modules we RPROVIDE Nov 07 10:31:51 Assuming nobody thinks that's a bad idea, I feel like that's more what's implied by RPROVIDES Nov 07 10:47:01 RP: after the override reversal both _arm and _armeb overrides get triggered for initscripts_1.0.bb + initscript_1.0.bbappend Nov 07 10:47:17 RP: but 'arm' isn't in OVERRIDES in the bitbake -e dump Nov 07 10:48:17 RP: the -e dump: http://dominion.thruhere.net/koen/arm-override.txt Nov 07 10:50:35 koen: remind me what the bbappend says? Nov 07 10:51:28 koen: hmm, # _append[arm] /build/linaro/build/openembedded-core/meta/recipes-core/initscripts/initscripts_1.0.bb:38 # " file://alignment.sh" # _append[armeb] /build/linaro/build/meta-linaro/meta-bigendian/recipes-core/initscripts/initscripts_1.0.bbappend:1 # " file://arm/alignment.sh" Nov 07 10:51:37 koen: so you're right, it is applying both :( Nov 07 10:56:59 RP: the bbappend is a one-liner: Nov 07 10:57:00 SRC_URI_append_armeb = " file://arm/alignment.sh" Nov 07 10:57:20 it took me a full day to spot that missing 'arm/' in the fetch error :/ Nov 07 10:58:14 the error: http://pastebin.com/7DS9vy4g Nov 07 10:58:55 the easy way out is just to add an armeb/alignment.sh, but I want to fix the real bug first Nov 07 11:01:35 suihkulokki: I found the oe-classic commits: Nov 07 11:01:37 https://github.com/openembedded/openembedded/commit/e0241d95b58de6c11be24f2aa0d90db3a5e44b50 Nov 07 11:01:43 and the fixup: Nov 07 11:01:44 https://github.com/openembedded/openembedded/commit/b48e4e380f359631131bebace6e768ebfdefcfd8 Nov 07 11:07:56 I see python-numpy is still broken Nov 07 11:08:27 "build/temp.linux-x86_64-2.7/numpy/linalg/lapack_litemodule.o: could not read symbols: File in wrong format" Nov 07 12:27:51 morning all Nov 07 12:37:13 hi pb_ Nov 07 12:37:33 hi ant_work Nov 07 12:39:06 you also bitten by overrides order? Nov 07 12:41:36 Hey. Nov 07 12:42:02 Does anyone know how I can see which package depends on which thin? I want to find out why ofono becomes part of my image, can I somehow search for things in my image that depend on it? Nov 07 12:42:24 I don't explicitly depend on it anywhere, and I don't have it in my image. Nov 07 12:42:35 Well, it's in the image when I build it, I mean the bb file for my image. Nov 07 12:44:54 I'll try in #yocto Nov 07 12:52:58 hi pb Nov 07 13:12:52 Stygia: bitbake -g -u depexp Nov 07 13:14:32 tobiash, Cheers. Nov 07 13:32:14 ant_work: I don't think so. Nov 07 13:32:18 hi woglinde Nov 07 13:55:29 bluelightning, everybody hates my indentation. :P Nov 07 13:57:23 bluelightning, Where you gonna undelete my liburi on your next pull request? Nov 07 13:57:43 Stygia: I just sent a new one with that removed also yes Nov 07 13:57:46 bluelightning, Otherwise everything checks out. Nov 07 13:58:19 bluelightning, But why do you want to delete liburi? Nov 07 13:58:32 bluelightning, Seems to me my recipe has a more complete RPROVIDES than the core one? Nov 07 13:58:51 well, as I keep saying one or the other needs to happen Nov 07 13:59:13 with my current patchset, the existing liburi-perl recipe must be fixed Nov 07 13:59:40 we don't want to add things to meta-perl that are already present in OE-Core Nov 07 14:00:46 bluelightning, Ah, alright. Nov 07 14:01:03 bluelightning, So we'll just keep it deleted in this repo, merge that to meta-perl, and I'll offer a patch for the one in OE-core Nov 07 14:01:21 yep, although I'd suggest sending the patch for OE-Core as soon as possible Nov 07 14:04:22 bluelightning, Hmm right. Nov 07 14:05:21 bluelightning, Might as well just do that now then. Nov 07 16:32:36 bluelightning, Ah, won't get it done today, sorry. Nov 07 16:32:54 bluelightning, Reinstalling my system didn't fix my problem, but re-reinstalling on a new harddrive did, so. Gonna be setting it up. Nov 07 16:33:09 bluelightning, I'll be sure to get it done before I try and merge my recipes into meta-perl, though. Nov 07 18:04:44 is there any reason why having my "/tmp" and "tmp-eglibc" on different filesysems would cause my image not to build? It works on another system with no extra filesystem mounts Nov 07 18:05:26 I get a lot of errors like "sh: /tmp/opkg-YGekpo/opkg-intercept-rqW4oH/depmod: Permission denied" Nov 07 18:06:20 i'm guessing you have noexec on /tmp Nov 07 18:06:24 the mount option Nov 07 18:06:54 I do, that's bad I take it? Nov 07 18:07:13 um, duh, yeah, makes sense Nov 07 18:07:41 well, clearly in this case it wants to write a temporary script and run it Nov 07 18:07:58 whether you want to allow that is up to you, but clearly oe/yocto requires exec on /tmp at this point :) Nov 07 18:08:50 Is it possible to brute force a hmac-sha256 sig with 64x AMD 7990? Nov 07 18:08:52 oooh, that also explains my issue with the TI packages building properly I'd never even looked at the mount args, I don't need 'em, just copy-pasted from somewhere Nov 07 18:09:09 i'm guessing its a potential security concern? *shrug* Nov 07 18:10:06 yeah, nothing I'm worried about Nov 07 18:10:36 * kergoth nods Nov 07 18:10:41 I had just switched to tmpfs on /tmp, so before I did that there wer no such restrictions Nov 07 18:10:49 indeed Nov 07 18:15:41 ungh, I'll bet dollars to donuts that's why all of my Jenkins builds on that slave were stalling... Nov 07 18:19:36 Maximus64: that's not really on topic. This channel for the openembedded project, not general embedded discussion, nor parallelism/concurrency :) Nov 07 18:20:56 oh sorry wrong channel missread oe instead of re Nov 07 18:24:17 :) Nov 07 19:23:34 hmm perhaps I should have bundled the orc migration patches together with gstreamer 1.2 Nov 07 19:24:13 now I'll have to hope the orc changes are applied before gstreamer is upgraded Nov 07 19:34:49 it will be short term pain though, right? Nov 07 19:48:06 yeah. I just pushed the gstreamer 1.2 patches, and I guess the patches are looked at in chronological order. Nov 07 22:45:22 koen: what do you set in local.conf to enable bigendian for genericarmv7ab ? **** ENDING LOGGING AT Fri Nov 08 02:59:58 2013