**** BEGIN LOGGING AT Wed May 01 02:59:57 2019 May 01 07:15:44 New news from stackoverflow: Is there a command to got to WORDIR of a recipe? May 01 07:16:20 cd? :) May 01 07:27:07 armpit: https://autobuilder.yocto.io/pub/non-release/20190430-13/testresults/testresult-report.txt - success with ltp! May 01 09:28:54 Hello #Yoctoians. May 01 09:29:24 Does anything exist that spits out a docker file for a toolchain? May 01 09:31:29 I currently deploy onto a build server, but there's always a need for clients (of varying platforms) to also have a toolchain. Spitting out something for Docker would be neat. May 01 10:59:55 Thought I might ask my question again, I don't know if that's bad practice though?.. May 01 11:00:32 Does anything (recipe?) exist that spits out a docker file for a toolchain? I currently deploy onto a build server, but there's always a need for clients (of varying platforms) to also have a toolchain. Spitting out something for Docker would be neat. May 01 11:37:20 JoeR: The system can build standalone toolchains in the form of the sdk (bitbake -c populate_sdk). Its not a docker image. May 01 11:37:37 RP: Yeah, that's what I do now. May 01 11:37:58 It gets spat out of Jenkins and forms each new "release" of toolchain for the dev team. May 01 11:38:20 I'd like to remove the need for the one build server though, so I want to give that toolchain to the devs for their client machines. May 01 11:38:37 The client machines are a bit varied and I have no control over them. Being able to give them a docker image would resolve that issue. May 01 11:39:15 So I'm left creating that from the toolchain as its spat out. So I wondered if this was something that could be done already. Or whether it'd be worth doing it as a class in its own right. May 01 11:39:17 JoeR: I'm not convinced since they likely have other things on their machines alongside the toolchain and your docker image won't contain those things May 01 11:39:55 Exactly. I want a one stop clean toolchain for their cross builds. Isolated from whatever else they get up to. May 01 11:40:28 They can still have whatever they want on their machine, but builds are remoted to their spun-up docker image. May 01 11:41:35 I'd like to get rid of my "build server" too, just have it as something quite simple and just have a docker image for the toolchain. It hosts other key bits of tooling that way. May 01 11:42:44 It's the wild west for client installs (windows, mac, ubuntu & mints over a five year set of releases) May 01 11:51:54 RP: btw, I'm updating vim right now to latest and folding in the locale thing, I should post it in a few hours May 01 11:53:25 Tartarus: sounds good thanks. I seem to have my hands full with various failures... May 01 11:55:41 The life of the head maintainer, never dull :) May 01 15:03:07 RP: there's some seriously funky stuff with all the kernel package manipulations that prevent a kernel-image postinstall script from working. I'm lost in it all at the moment. Hopefully I'll have a coherent answer by end of day. May 01 15:04:51 Tartarus: I think my head probably does need maintenance ;-) May 01 15:05:03 ha May 01 15:07:56 * zeddii is questioning his sanity as well today. May 01 15:08:21 someone needs to take elfutils's makefile tests away from me May 01 15:08:36 eee. May 01 15:09:16 "It's just a makefile", I say that everytime I wade into a perf issue. then I just want to curl up and become a stone mason. May 01 15:09:26 whoever thought silently continuing if it can't find its modules was a good thing was wrong... May 01 15:09:59 heh. May 01 15:10:22 zeddii: practical real world work is sometimes attractive May 01 15:10:42 I can chose from two postinst failures. one, the postinst doesn't make it into the package. the other, rootfs fails to assemble since the package isn't available to the rootfs .. but it does have the postinst in it :D May 01 15:10:47 * zeddii nods. May 01 15:10:56 maybe more coffee will clarify things for me. May 01 15:11:32 I should just revert the elfutils ptests May 01 15:11:48 I'd vote for that option :D May 01 15:12:24 Is elfutils even used much? May 01 15:12:50 JPEW: I was just more worried if there was some hidden issue with prelink or something nasty May 01 15:12:58 turns out its just broken May 01 15:13:04 (the tests) May 01 15:13:57 huh, I guess it's used more often that I would have gussed May 01 15:14:08 JPEW: we use the lib for chrpath and a few other key things May 01 15:14:38 You can apparently even compile systemd with elfutils support... although I shudder to think what that might be used for May 01 15:14:45 JPEW: a lot of the tracing stuff sits on top of it too May 01 15:14:56 Ah makese sense May 01 15:15:12 the problem with instrumenting package.bbclass .. is the savage amount of rebuilding. May 01 15:15:55 well, that and I'm clueless about it all :D May 01 15:16:44 I'm almost convinced that the emit_pkgdata needs to check PKG_ before it goes and tries to process postinsts May 01 15:20:04 zeddii: there could be a problem like that quite easily May 01 15:21:13 yah. something is going on right in that area. I'll flail around and see what I can break :D May 01 15:27:13 * armpit zeddii a stone mason... watch out for flying stones May 01 15:27:31 * zeddii would break some windows May 01 16:01:20 RP: So, uh, mild apologies for bringing up the BB_MIN_VER thing again over in http://lists.openembedded.org/pipermail/openembedded-devel/2019-May/199442.html May 01 16:06:05 Tartarus: I really don't want to bump the version every 5 mins, we need to get better about the APIs May 01 16:06:16 (the min version) May 01 16:07:44 Tartarus: I did notice, its a frsutrating the issue didn't get properly fixed, I thought it had been :/ May 01 16:22:22 RP: I agree the vim direction is the wrong way to go :) May 01 16:22:39 RP, Tartarus: Sorry :( Looking at it now May 01 16:25:24 yeah, lets use "Notes" May 01 16:25:55 armpit: we still have selftest problems btw May 01 16:26:12 armpit: not least we installed puthon3-cairo on the AB with me not realising pybootchart is py2 May 01 16:26:25 armpit: probably need a bug to migrate it to py3 May 01 16:26:38 yep May 01 16:26:41 armpit: other tests failed too, but I had gcc9 in there so it was a mess May 01 16:26:50 (packageconfig tests) May 01 16:26:52 does anyone us pybootchart? May 01 16:27:05 armpit: it is useful for looking at task timings May 01 16:27:37 k May 01 16:27:59 its sad the need for cairo is not listed in the manual test May 01 16:28:45 RP, what about the other patches removing things that are already done @ runtime or selftest? May 01 16:29:54 armpit: need to go carefully through those again. I thought there was a dependency May 01 16:30:14 k May 01 16:30:40 I don't me the sdk one.. that need to be redone May 01 16:33:02 kergoth: when you're around I wouldn't mind a second opinion on the bitbake header/license changes May 01 16:34:13 it's fine with me. also good riddance to the editor modelines. we're using standard indenting conventions anyway, and even if we werent, editorconfig would be better May 01 16:34:29 kergoth: my feelings too :) May 01 16:35:27 kergoth: going to try and standardise on the SPDX license headers. The dilemma is recipes :/ May 01 17:00:22 https://github.com/mikeal/self-care/issues May 01 18:53:08 Hi, sorry for the novice question. For a init script (init.d) what is the best method to determine whether the service is working or not. The skeleton use "pidof -x xxx" but versus using a PIDFILE ? May 01 18:57:18 working=running May 01 19:00:54 is it possible to use the same bin_package reciepe for multiple rpms ? May 01 19:18:36 bl4ckb0ne: you mean, one recipe that provides several packages? May 01 19:19:46 i have several rpms from third party that needs to be extracted on the rootfs May 01 19:20:04 and I dont want to make one recipe for every one of them May 01 19:20:42 i came upon the bin_package class that looks nice, but if i understand it correctly, i need to feed directly the rpm May 01 19:20:52 my 3rd party rpms are on a git repo May 01 19:20:59 bl4ckb0ne: you'd probably have to replicate the bin_package logic and make it work with a list of packages instead of a single one May 01 19:21:20 bl4ckb0ne: i am not aware of an out-of-the-box way for doing that. May 01 19:21:34 currently doing that May 01 19:21:42 but I encounter a lot of warnings from "QA" May 01 19:21:52 can I shut them all at once ? May 01 19:22:51 I already have `INSANE_SKIP_${PN} += "already-stripped ldflags arch staticdev"` May 01 19:23:45 bl4ckb0ne: nothing to add besides whats mentioned in https://www.yoctoproject.org/docs/2.6.2/ref-manual/ref-manual.html#ref-classes-insane May 01 19:26:15 ty May 01 19:26:23 alright, down to RDEPENDS trouble now May 01 19:30:02 it works! thanks LetoThe2nd May 01 19:36:00 have fun May 01 19:37:20 ty May 01 19:37:30 it feels so wrong though May 01 19:37:50 i feel that yocto has this need to compile everything, and here i'm just forcing prebuilt stuff May 01 19:39:38 it's certainly not ideal, but if you're dealing with vendors you don't always have a choice May 01 19:39:44 been there, done that May 01 19:42:17 its the intel compiler May 01 19:42:32 its strange intel doesnt provide it May 01 19:49:01 oh well, now I have another dependency missing May 01 19:49:06 > nothing provides ld-linux-k1om.so.2()(64bit) needed by icc May 01 20:15:47 can I completely ignore the dependencies of one recipe ? May 01 20:29:51 * RP wonders if he should complain at himself on the mailing list about his broken patches May 01 20:30:38 RP: I find a stern talking to in person gets the point across much better May 01 20:36:17 RP, Tartarus: I send the OE core patch that I beleive will fix rm_work with older bitbake. I tested with 1.40 before the hash equive changes were added to bitbake. May 01 20:36:42 https://lists.yoctoproject.org/pipermail/yocto/2015-November/027597.html May 01 20:36:49 im clearly going the wrong way May 01 20:37:04 JPEW: just saw that, looks good thanks May 01 20:37:18 np. Really sorry I forgot about it May 01 20:38:12 I think we'll need that "backported" to warrior also? May 01 20:42:54 JPEW: yes, we need armpit for that, unless he wants to tell me just to do it May 01 20:43:41 RP: I'll ping him after the patch is on master.... I'll remember this time :) May 01 20:44:12 JPEW: I get tempted to merge obvious things like this, then things break in weird and wonderful ways May 01 20:45:32 remembering is something for the young May 01 20:45:48 Ya.... thats hard. I suspect that we will *probably* be OK doing the proper QA on this one. AFAIK, there has only been the one complaint May 01 20:46:07 armpit: Indeed. I must be getting older... it keeps getting harder May 01 20:47:15 RP just don't get into a flame war with yourself ; ) May 01 20:47:37 armpit: that could get entertaining May 01 20:48:03 I wonder if that is covered in the code of conduct ? May 01 21:18:44 * moto-timo envisions an RP<->RP flamewar May 01 21:19:13 * moto-timo thinks it would be a top selling video game May 01 21:21:15 like RP on a mountain bike racing RP on a dirt bike May 01 21:38:27 Crofton: RP on a road bicycle had a top speed of 41.4mph earlier this evening... May 01 21:39:40 hu, the RP flamewar would do 60 mph May 01 21:40:37 armpit: not sure the dirt bike's gearing would do 60mph, mountain bike certainly wouldn't May 01 21:57:07 RP: How did you hit 40 on a bicycle? May 01 21:58:01 JPEW: small hill, got to lead the group down it tonight :) May 01 21:58:29 JPEW: record is about 49mph, that was a bigger hill! May 01 21:59:08 RP: Crazy. Last time I went that fast (years ago) I braked poorly in the wet and went skidding on my side for about 20 feet. It wasn't fun May 01 21:59:56 And by "that fast" I think I was *maybe* doing 35 May 01 22:00:25 JPEW: I probably shouldn't mention it was wet tonight, we'd been warned it was greasy and that was 40mph into a corner. Braking when doing this would be potentially problematic May 01 22:01:02 Ya, I knew I was in toruble when my backwheel started passing me.... that said, I wish I could bike every day still :( May 01 22:01:09 JPEW: I was surprised it was that speed, only saw it from the trace afterwards from the garmin May 01 22:02:13 JPEW: wheels overtaking you tends not to end well. Back wheel of the mountain bike going over my head on Saturday was one of those moments :/ May 01 22:02:48 This is making me sound crazy, it wasn't that bad, I'm sure I'm exaggerating :) May 01 22:03:20 JPEW: I do the cycling to get me away from the computer and get some exercise May 01 22:41:35 warrior landed.. I am behind **** ENDING LOGGING AT Thu May 02 02:59:57 2019