**** BEGIN LOGGING AT Tue Mar 10 02:59:57 2020 Mar 10 07:25:42 hello =) Mar 10 08:07:46 hi *! Mar 10 08:11:21 hi how do i disable the installing of an application in the sdk ( so i don't get te application source code and debug packages in the sdk ) Mar 10 08:12:09 can somebody provide me WIC output log from creation of EFI bootable image? Target/architecture might be whatever. Ideally will be to have log from WIC launched with --debug option. Reason: I want to understand step-by-step how WIC creates EFI bootable image (especialy how WIC writes GRUB's kernel.img into image) Mar 10 08:12:13 hmw1: TOOLCHAIN_TASK_HOST_remove or whatsitcalled doesn't do the trick? Mar 10 08:14:41 It's probably TOOLCHAIN_TARGET_TASK_remove if it's a package for the target, but that what I would try first too. Mar 10 08:14:50 LetoThe2nd: tnx :D :D Mar 10 08:26:13 erbo: darn going to try that second if the sloution of leto fails. btw tnx for the feedback :D Mar 10 08:26:16 erbo: i just can't keep those TOOLCHAIN_THINGS apart. Mar 10 08:26:34 hmw1: try mine first :) Mar 10 08:26:44 It's basically what LetoThe2nd meant Mar 10 08:26:53 yeah try erbo's one first! Mar 10 08:27:01 LetoThe2nd: I cheated and looked it up :D Mar 10 08:27:13 erbo: nothing wrong with that. Mar 10 08:27:52 had to many files open so i think its TOOLCHAIN_TARGET_TASK_remove first Mar 10 09:04:30 erbo: tnx it worked Mar 10 10:27:25 Hey folks, I was trying to explore the environment using `bitbake -e nginx` (for example). What I get back is my terminal being flooded with some python code. Is this what I should exepect, is this a bug or did I do something wrong? Mar 10 10:28:15 erbo: consider it an expectation fulfilled :) Mar 10 10:28:43 emrius: ^^^^ Mar 10 10:29:01 emrius: the environment includes all the code Mar 10 10:29:07 | less is your friend Mar 10 10:29:36 Hmm... okaaaay... Mar 10 10:29:59 all the code in the recipe and classes it pulls in, that is Mar 10 10:30:21 ok thanks Mar 10 10:43:25 Is there some method to easily switch between a recipe using externalsrc and a normal recipe? I would use it to test that the code actually compiles fine with yocto build system before doing the commit. Currently I have a recipe like my-recipe_9999 and then the normal recipe with the normal version number but it's pretty ugly. Mar 10 10:44:57 droman: you can do the externalsrc dance via local.conf, for example Mar 10 10:47:51 LetoThe2nd: I tried to do it following the example of the dev manual but it didn't worked as expected. It was trying to pull the source code from the normal recipe anyway Mar 10 10:52:56 droman: or use devtool? Mar 10 10:54:16 thats an option too. Mar 10 10:54:40 yet i am pretty certain that INHERIT and EXTERNALSRC_pn-mystuff should do the tricl Mar 10 10:55:55 qschulz: thanks, will play with it and see how it works Mar 10 10:59:10 droman: devtool modify, go to the sources, modify whatever you want, build your image or your SW with bitbake (or devtool build?). If you want to switch between recipes, remove/add the devtool layer to your bblayers.conf (I think devtool reset without removing the sources and then doing a devtool modify does not work but not 100% sure) Mar 10 11:22:26 LetoThe2nd: It seems that it tries to pull the source from the normal SRC_URI but if you use externalsrc in local.conf it uses that source tree. The problem was that the src_uri of my normal recipe was invalid. I don't know if it's supposed to work this way Mar 10 12:25:32 Hi, I have an easier question today :) Mar 10 12:26:27 When I build recipe foo.bb, which "DEPENDS +=" on bar.bb, but both are cached in the sstate cache Mar 10 12:27:09 foo.bb explodes because it inherits from adduser.bbclass, which dynamically adds the user but tries to add it to the group which bar.bb creates Mar 10 12:28:04 So as I understand it, the regular DEPENDS chain does not apply to do_populate_sysroot_setscene tasks (at least in my version) Mar 10 12:28:50 Now I've seen some cryptical definitions of dependencies of tasks, and I'm wondering, how do I make foo.bb's do_populate_sysroot depend on bar.bb's do_populate_sysroot ? Mar 10 12:28:57 There is a syntax for this I think yes ? Mar 10 12:32:47 Inferring from https://www.yoctoproject.org/docs/1.6/bitbake-user-manual/bitbake-user-manual.html#inter-task-dependencies... perhaps: do_populate_sysroot[depends] = "bar:do_populate_sysroot" would work ? Mar 10 12:33:29 gtristan48: I'd look in the bugzilla as there are open bugs about this kind of issue :( Mar 10 12:33:39 gtristan48: i am not really involved with the topic, but a 1.6 manual is certainly not authorative any more. Mar 10 12:34:01 LetoThe2nd, the 1.6 manual is probably too new for me yeah hehe Mar 10 12:34:46 gtristan48: i doubt that. and if, then you are so deep in trouble that you shouldn't IRC, but crawl into a corner and weep. Mar 10 12:36:32 LetoThe2nd, sounds like the majority of people contracted to fix their clients yocto mess of layers provided by various vendors... who have cluelessly/frantically patched their recipes without following much guidelines Mar 10 12:36:50 Still, I've found IRC quite helpful in this struggle :) Mar 10 12:38:04 gtristan48: i take it that you have looked up which release 1.6 really is? Mar 10 12:39:00 Not really hehe, you got me :) Mar 10 12:39:54 I've got a commit sha somewhere telling me the poky this whole thing is based on, and I've yet to figure out where the "Version number" is encoded into the git repository Mar 10 12:40:13 gtristan48: git tag --contains [sha] Mar 10 12:41:55 rburton: probably warrior :P Mar 10 12:43:18 gtristan48: if you have poky then meta-poky/conf/distro/poky.conf Mar 10 12:45:22 Thanks, it looks like they bastardized their git pretty bad (I don't find any tags with that, although I do have a sha from upstream history) Mar 10 12:45:26 DISTRO_VERSION = "2.1.2" Mar 10 12:45:33 I guess that's much more recent than 1.6 :) Mar 10 12:46:23 gtristan48: krogoth, yes much more recent but predates per-recipe sysroot Mar 10 12:46:41 Yeah I'm aware of the per-recipe sysroot change Mar 10 12:46:49 That we're behind of :) Mar 10 12:47:11 So there be races here, still I've mostly revived builds from sstate cache Mar 10 12:49:44 how can we stop a build if the user is using debug keys, one way is comparing md5sum for debug key and current Mar 10 12:49:47 So https://www.yoctoproject.org/docs/2.1/bitbake-user-manual/bitbake-user-manual.html#inter-task-dependencies shows me fairly the same thing... but before I pull the trigger... *should* I be careful about defining setscene tasks vs non-setscene tasks ? Mar 10 12:50:24 Like, if I say do_populate_sysroot[depends] = "bar:do_populate_sysroot", is that supposed to be an sstate cache agnostic statement ? Mar 10 12:50:51 gtristan48: i would say that this is rather a sympton of a bad architecture/decision somewhere else. Mar 10 12:52:45 LetoThe2nd, Ummm sure, I'm not here to correct bad architecture hehe, just trying to find the right workaround Mar 10 12:53:10 gtristan48: did you have a look in bugzilla? Mar 10 12:53:26 RP, worried it will open a whole new can of worms :-/ Mar 10 12:53:57 gtristan48: using a release that old its too late for that :( Mar 10 12:54:43 Yeah, I can't argue an upgrade here, just trying to explore the possibility of enabling sstate cache in an otherwise "working" build Mar 10 12:55:03 that statement alone gives me the creeps. Mar 10 12:55:05 (it was a horrid mess downstream as well) Mar 10 12:55:06 gtristan48: https://bugzilla.yoctoproject.org/show_bug.cgi?id=13419 Mar 10 12:55:06 Bug 13419: normal, Medium+, 3.1 M3, richard.purdie, NEW , recipes that add users to groups cannot rely on other recipes creating those groups (when population from sstate happens) Mar 10 12:55:51 RP: in another twist of irony, I'm having a hard time recreating the gcc reproducibility issues with poky :/ Mar 10 12:56:13 RP, interesting, fwiw I had already *patched* shadow so that useradd/groupadd choose to allocate UIDs/GIDs in far ranges (because this build has some hard coded UIDs) Mar 10 12:56:52 tgamblin: good news is I think rburton may have some insight Mar 10 12:57:25 tgamblin: For the gcc issue you likely need to compare a build with an old compiler with a build done with a recent compiler Mar 10 12:57:43 Thanks for the pointer to that specific bugzilla entry ! Mar 10 12:57:47 RP: ah, ok Mar 10 12:58:04 an old host gcc will given different results to a newer host gcc in this case :( Mar 10 13:01:56 RP: do you want me to incorporate your coreutils hack into the original patch and re-send? Mar 10 13:02:16 or is it easier for you to just squash them? Mar 10 13:02:25 tgamblin: I can just sort that thanks Mar 10 13:02:30 RP: k Mar 10 13:17:15 hello, does anyone build under arch linux? Trying to bitbake like described on https://wiki.archlinux.org/index.php/Yocto ends for me in "ERROR: Task (/home/hopser/projekte/spielwiese/yocto/poky/meta/recipes-devtools/binutils/binutils-cross_2.31.bb:do_unpack) failed with exit code '1'" Mar 10 13:17:45 RP: Did the logging work on the AB? Mar 10 13:17:48 hopser: in a nutshell, just us a docker container. Mar 10 13:18:23 LetoThe2nd: I really need to schedule with you to give a live stream on pyrex/containers Mar 10 13:18:35 JPEW: yes :) Mar 10 13:18:36 I'm new to yocto. What is AB? And why do I need docker here? Mar 10 13:18:55 hopser: that guide is outdated in the scale of years. plus, we have an excellent quick start guide ourselves. Mar 10 13:19:17 JPEW: absolutely agreed. Mar 10 13:19:20 hopser: AB -> AutoBuilder: The automated builds used to do testing of patches run by the Yocto project Mar 10 13:19:59 hopser: i would strongly suggest to watch at least #1 at https://www.youtube.com/playlist?list=PLD4M5FoHz-TxMfBFrDKfIS_GLY25Qsfyj Mar 10 13:20:04 JPEW: thanks Mar 10 13:20:05 hopser: archlinux isn't a supported distro. Docker so that you can use one of the officially supported distros (e.g. ubuntu/debian) Mar 10 13:20:23 LetoThe2nd: did you mean https://www.yoctoproject.org/docs/3.0.2/brief-yoctoprojectqs/brief-yoctoprojectqs.html ? Mar 10 13:20:31 hopser: yep, basically. Mar 10 13:20:58 allright, thank you. I'll try docker next. Mar 10 13:21:18 hopser: before wasting (cpu) time, really, watch the video :) Mar 10 13:21:43 hopser: For docker, I recommend https://github.com/garmin/pyrex, but I'm a little biased because I wrote it :() Mar 10 13:21:46 LetoThe2nd: which one? Mar 10 13:21:54 JPEW: nothing exploded, I've not checked the details yet Mar 10 13:22:08 hopser: 13:19 < LetoThe2nd> hopser: i would strongly suggest to watch at least #1 at https://www.youtube.com/playlist?list=PLD4M5FoHz-TxMfBFrDKfIS_GLY25Qsfyj Mar 10 13:22:39 hopser: the problem might actually be trivial to fix but you're not showing us the actual errors Mar 10 13:23:07 LetoThe2nd: thanks, i've seen that one allready. it worked fine on ubuntu VM ;) Mar 10 13:23:36 hopser: see. as an arch user myself, i'd suggest to just stick with ubuntu in a container :) Mar 10 13:24:25 good, i'll use ubuntu in a container. thanks for advise Mar 10 13:26:15 thank you all! on IRC are good people :) Mar 10 13:26:32 * LetoThe2nd interprets that as a personal insult. Mar 10 13:26:58 I'm https://youtu.be/_7VsoxT_FUY Mar 10 13:27:30 lol Mar 10 13:27:40 sorry for that :-D Mar 10 13:29:56 LetoThe2nd: when my brain is working too slow... You made the yocto videos :) Mar 10 13:30:14 LetoThe2nd: thank you very much for these vids!! Mar 10 13:30:25 hehe yes, indeed. Mar 10 13:30:36 you're welcome :) Mar 10 13:51:29 yay, just read the 3.1 LTS announcement Mar 10 14:00:58 mcfrisk: I think the LTS will be a good thing... we've had trouble at work congealing around specific Yocto versions (used to have 5 in support, but was finally able to kill one to drop us to 4 :eyeroll:), so I hope this gives some leverage for that Mar 10 14:18:45 JPEW: fully agree. I'm supporting 2-3 yocto versions. hope to converge most to LTS and also push to contribute to maintenance. still need to polish my sumo patches, sigh.. Mar 10 14:22:31 hello all, first of all, i would like to thank the person making the livestream videos for the heads up of this irc channel and, of course, of the helpful videos :) Mar 10 14:22:59 LetoThe2nd: ^ new fan :) Mar 10 14:23:00 alexpires: that would be LetoThe2nd Mar 10 14:23:52 secondly, a technical question now haha, I have a provider that supplies their camera drivers only with .deb packages. How can I include them in my build and install them? Mar 10 14:25:04 alexpires: maybe there is a deb eating bbclass, but just write a recipe which unpacks the .deb and copies files to correct place in do_install() task. Mar 10 14:28:05 oh, of course. I will check if there is some example that I can follow for that and follow my work from there. Mar 10 14:28:11 thanks ! Mar 10 14:28:21 put the deb URL in SRC_URI and bitbake will expand it for you in unpack Mar 10 14:28:32 alexpires: also, you may want to inherit bin_package :) Mar 10 14:28:34 then you can just install the files into ${D} like usual, and voila Mar 10 14:28:57 alexpires: not gonna help with what you want but brings some slight build time benefit Mar 10 14:31:09 alrighty, thanks all :) seems pretty straightforward. I will try it out and break my build many times until I get it right Mar 10 14:32:46 JPEW: if you're around to do a reprod test, ross/gcc has two potential fixes for the gcc issues Mar 10 14:34:56 YPTM: is which timezone is the meeting? i.e. is the meeting in .5 hour or 1.5 hour? Mar 10 14:39:36 tlwoerner: I think it's usually in 1.5h? Not sure Mar 10 14:40:47 silly DST Mar 10 14:43:35 my calendar says its in 1.5h Mar 10 14:44:14 Oh, right... DST Mar 10 14:44:37 What TZ is the meeting schedule in? Mar 10 14:47:06 lets assume UK :) Mar 10 14:47:42 New news from stackoverflow: Include numpy's arrayobject.h into bitbake recipe - how to fix installation order? Mar 10 14:48:28 rburton: its run by Stephen so US time Mar 10 14:48:42 does all of the US shift at the same time? Mar 10 14:48:46 I know Stephen plans to run it in about 10 mins Mar 10 14:49:03 rburton: most but not all of North America does Mar 10 14:49:04 my calendar says 1 hour 10 Mar 10 14:49:08 rburton: yes, anything else would be silly. Ah, hmm :) Mar 10 14:49:16 so i suspect the invite is knackered? Mar 10 14:49:25 rburton: I have two on mine, one moved, one didn't Mar 10 14:49:40 so when is the TSC? :) Mar 10 14:49:49 after it? Mar 10 14:50:29 i hope to make that then! Mar 10 14:51:57 rburton: Yes, all of the US shifts at the same time Mar 10 14:53:06 rburton: It's a weird dynamic; individual states can decided if they want to observe DST, but the federal government sets what it means to observe it :) Mar 10 14:53:50 Thankfully, DST is ending so everyone's on the same time now Mar 10 14:54:24 Err, sorry thats wrong Mar 10 14:54:33 Even I can't get it right :( Mar 10 14:55:07 Hi, may I draw your attention to this post of mine? https://stackoverflow.com/questions/60619902/include-numpys-arrayobject-h-into-bitbake-recipe-hot-to...Feel free to throw in thoughts here. If it helps I can also phrase and post the answer if this helps in any way Mar 10 14:55:53 We should all just switch to UTC, globally. I'm down to refer to midday as 17:00 Mar 10 14:56:02 tgamblin: Works for me Mar 10 14:56:14 YPTM: armin is on Mar 10 14:56:44 JPEW: any idea on the warning in https://autobuilder.yoctoproject.org/typhoon/#/builders/87/builds/767 (looks logging related) Mar 10 14:57:39 RP: I'll take a look Mar 10 14:58:06 tgamblin: 'cuz that's when the drinkin starts? Mar 10 14:58:25 i hear drinks?!? Mar 10 14:59:15 tlwoerner: yeah, and all you gotta do is rent some fancy limo to follow the terminator across the continent - constant happy hour! Mar 10 14:59:35 except not at all because UTC Mar 10 14:59:44 * tgamblin apparently needs coffee Mar 10 15:01:20 alexpires: you're welcome! throw money! Mar 10 15:05:56 Wasn't there a tool in yocto for direct r/w access to physical memory ? Mar 10 15:06:04 (on target) Mar 10 15:06:15 kroon: devmem? Mar 10 15:06:25 Maybe to abbreviate my question: How do I make sure `python3-numpy` is installed before my recipe is compiled which required `numpy/arrayobject.h`? Mar 10 15:06:28 LetoThe2nd, right, thank you monkeyman Mar 10 15:06:56 * LetoThe2nd doesn't like bananas. Mar 10 15:07:10 LetoThe2nd: how do you feed your 1200+ monkeys then? Mar 10 15:07:38 kroon: there's a devmem2 also I think. Don't know the difference though Mar 10 15:07:38 qschulz: 1725. Mar 10 15:07:53 LetoThe2nd: technically that's 1200+ :p Mar 10 15:07:55 qschulz: properly with beer and steak, ofcourse. Mar 10 15:08:49 qschulz: but awful tolerance. Mar 10 15:11:04 Hi all, what is the optimum way to add a license to a recipe which only installs systemd services? I am currently using `LICENSE = "CLOSED"` to avoid the `LIC_FILES_CHKSUM` error. Mar 10 15:11:26 My recipe structure is as follows: Mar 10 15:12:06 .├── files│   ├── button│   │   └── button.service│   ├── gps│   │   └── gps.service│   ├── imu│   │   └── imu.service│   ├── influx│   │   └── influxdb.conf│   ├── nmea2k│   │   └── nmea2k.service│   └── umg│ Mar 10 15:12:06 └── conf.json└── umg_1.0.bb Mar 10 15:13:20 shan1 unfortunately copy-paste doenst work here. You can use https://pastebin.com/ instead Mar 10 15:15:06 emrius https://pastebin.com/di82i1zu Mar 10 15:15:55 shan1: Include an SPDX-License-Identifier: line in the service file, and then use LIC_FILES_CHKSUM = "file://button.service;beginline=3;endline=3;md5=..." (assuming the SPDX-License-Identifier is on line 3 in the button.service file). Mar 10 15:15:55 shan1: https://git.yoctoproject.org/cgit.cgi/poky/tree/meta/recipes-core/systemd/systemd-machine-units_1.0.bb#n4 Mar 10 15:15:55 shan1: or what Saur said, that's nicer IMO Mar 10 15:18:41 Saur do you have a recipe example somewhere? Mar 10 15:19:52 shan1: build/poky/meta/recipes-rt/rt-tests/rt-tests_1.1.bb Mar 10 15:19:58 remove build :) Mar 10 15:20:11 shan1: have a look at poky/meta/recipes-rt/rt-tests/files/rt_bmark.py Mar 10 15:24:00 shan1: meta/recipes-devtools/icecc-toolchain/nativesdk-icecc-toolchain_0.1.bb is probably the closest. It doesn't use an SPDX-License-Identifier, but rather a full license boilerplate in "icecc-env.sh". Mar 10 15:25:57 Would it be dumb to add my Layer's License in the `files` folder and then point its md5 checksum in the recipe? Mar 10 15:28:47 shan1: layer's license? Mar 10 15:30:37 I will push the recipes to my meta layer which is anyways under the MIT License. Since every recipe requires `LICENSE = ""` within it. I just added the MIT License into the `files` folder and added the `LIC_FILES_CHKSUM = "file://${WORKDIR}/LICENSE;md5=..."` and it seems to work Mar 10 15:32:55 shan1: just use LIC_FILES_CHKSUM = "file://${COMMON_LICENSE_DIR}/MIT;md5=0835ade698e0bcf8506ecda2f7b4f302" then? Mar 10 15:33:12 no need for adding the license file manually Mar 10 15:34:55 Ah this will add the poky common license file ? Mar 10 15:38:04 shan1: it will refer to the one in openembedded-core or poky Mar 10 15:42:08 @qs Mar 10 15:42:25 qschulz worked like a charm! thanks :+1: Mar 10 15:42:52 I've this badly coded third party SW which funnily enough does not work if there's anything else than -O0 passed in CFLAGS. As -O2 is set by default in it, what is the advised way to remove it or switch it to -O0? Mar 10 15:42:57 shan1: pleasure Mar 10 15:44:21 LetoThe2nd: ^ maybe you have some hindsight? Mar 10 15:45:48 qschulz: depends on the buildsystem in use, but technically CFLAGS is a variable like any other inside the recipe evaluation. Mar 10 15:46:21 qschulz: so you can modify it at will. maybe grep for it in poky or oe-core to find an example. Mar 10 15:47:12 YPTM: who has the patch for gcc-sylinks to make it multilib safe? Mar 10 15:47:15 LetoThe2nd: mmmm. sorry, badly explained. -O2 is passed in CFLAGS by Yocto Mar 10 15:47:20 qschulz: i know. Mar 10 15:47:51 qschulz: but its passed in as a variable jsut like PN, so you can modify it for your recipe. Mar 10 15:48:21 LetoThe2nd: The question is, is there any preferred way? Mar 10 15:48:44 qschulz: using a proper buildsystem and writing clean code is to be preferred. Mar 10 15:49:29 LetoThe2nd: thank you for stating the obvious :) I'm fixing it, but I can't be bothered to actually fix code which works with -O0 but not -O2 :) Mar 10 15:50:12 qschulz: hrhr. you're welcome. Mar 10 15:54:02 qschulz: you're asking for more trouble at runtime there :) been there, would not do again even with management pressure... Mar 10 15:54:21 Also, I already have a layer that is writing a configuration file to `/etc/influxdb/` directory. The recipe is called `influxdb_1.7.9.bb` but I wish to add my configurations to it. How do I go about appending my configuration to the recipe? Is it through a Mar 10 15:54:32 .bbappend? Mar 10 15:56:15 shan1: usually yes. Mar 10 15:57:19 shan1: yet, you often have a real application recipe that RDEPENDS on influx, so you could also create an influx-config recipe and also RDEPEND on that. it all depends a bit on your architecture. Mar 10 15:57:56 tlwoerner: I believe that was Jeremy from MV who volunteered that Mar 10 15:58:07 so I might need to take the original conf file. add my stuff to it and create a patch and then add it to the bbappend? Mar 10 15:58:10 smurray: okay, thanks Mar 10 15:58:48 shan1: it depends. if you want to patch, if you want to replace, of its another file, if its version dependent or not... Mar 10 15:59:50 I checked. I mostly append my stuff in the end because the default conf that is installed and it should work. Mar 10 15:59:55 mcfrisk: it's a signing tool used in one very specific process, so if it does its job, good enough. It's not running on the target as well... Mar 10 16:01:09 qschulz: and see poky/meta/conf/bitbake.conf for FULL_OPTIMIZATION etc flags. you can modify them. I used that to build a full system with Os once and also fixed all components which were abusing default optimization flags... Mar 10 16:01:53 mcfrisk: yeah I've seen that one, the question is what's advised? Mar 10 16:02:36 I'm not an expert but you can override those variables or _append|_remove in recipes.. Mar 10 16:03:11 mcfrisk: I could do a CFLAGS_remove = "-O2" for example, that would maybe work better than overriding the variable and avoid weird problems when upgrading to newer yocto releases Mar 10 16:10:19 qschulz: yep, could work. Mar 10 17:18:16 New news from stackoverflow: cache vs sstate-cache folder in Yocto Mar 10 18:43:05 RP: how soon do you need the coreutils-ptest patch? i.e. do you care about a cover letter updated with arm/arm64 results? Those will take a couple of hours to get Mar 10 18:48:35 New news from stackoverflow: add git to yocto image Mar 10 20:09:59 rburton: do you remember that couple of articles about implementing yocto-like features with buildroot? do you have the links? Mar 10 20:17:31 kanavin_home: yes, it was codethink Mar 10 20:17:31 kanavin_home: https://www.codethink.co.uk/tag/buildroot.html has the links Mar 10 20:18:39 rburton: thanks :) as chance would have it, we bought a teufel radio last week :D Mar 10 20:18:59 the reason I asked is that some people raise 'buildroot' at work as an alternative, and I want to write a blog post Mar 10 20:18:59 kanavin_home: one of the smart connected ones? Mar 10 20:19:05 kanavin_home: please do Mar 10 20:19:18 kanavin_home: the old rule about 'easy to start with' does appear to be true Mar 10 20:19:36 especially when you read those posts and see how much effort they put into custom infrastructure to speed up builds Mar 10 20:19:42 rburton: it's not a multiroom product, just an internet radio Mar 10 20:20:14 Hey guys, I'm having a fetch error on our pyro branch due to NXP deleting the git.freescale.com repositories. I tried using the PREMIRROR but it isn't working for me. Any suggestions? bbappend everything.... ugghh Mar 10 20:20:21 rburton: this https://www.teufelaudio.com/stereo/radio-3sixty-p16568.html Mar 10 20:20:25 rburton: it is way easier. alone Kconfig vs. image-recipes is a kind of killer-feature. Mar 10 20:20:36 rburton: but it just does not scale, IMHO Mar 10 20:20:50 nate02: upgrade :P Mar 10 20:21:13 kanavin_home: i've still got the multiroom speakers pre-teufel when they were raumfeld. the base and controller are dead but the speakers still work, they'll be spotify speakers in my study at some point Mar 10 20:25:23 so PREMIRRORS is broken, or was broken? Or am I misreading the yocto manual? Mar 10 20:26:38 nate02: no it should work if you already have everything. and if the signatures match etc. Mar 10 20:28:01 hmm, it seems like the git repo isn't an exact mirror, grrr Mar 10 22:23:18 tgamblin: sorry, had to step away with a headache Mar 10 22:23:27 tgamblin: how's it looking? Mar 10 22:24:17 RP: just churning through the arm target tests, should have it submitted tonight Mar 10 22:24:42 RP: ended up just removing that assert test (for now) Mar 10 22:25:29 * tgamblin wishes he had a beefy ARM server to kvm this stuff on Mar 10 22:50:19 tgamblin: Can you share it on a branch. Even if you don't send yet I'm tempted to queue for testing in -next Mar 10 22:57:21 RP: sent it directly to your email. Note that it includes your hack patch too Mar 10 22:58:48 tgamblin: ok, I was thinking I might just clean that up as a fix on top and explain it but I guess either way is fine Mar 10 22:59:53 RP: ah, oops Mar 10 23:07:32 RP: well, I can pull that chunk back out when I submit and you can still do the second patch Mar 10 23:07:43 I just added it in for my testing Mar 10 23:08:11 tgamblin: I'm tempted to do this mainly so we can have a commit message describing what we did and why Mar 10 23:08:20 RP: sure, no worries Mar 10 23:08:24 since its been a pain to figure out Mar 10 23:08:55 tgamblin: good news is the hack and other patches in -next have green builds apart from infrastructure issues Mar 10 23:09:11 RP: excellent :) Mar 11 01:03:59 hi. it's been a long time since i messed around with openembedded (sharp zaurus sl-5600 days) Mar 11 01:03:59 trying to gauge feasibility of building a modern kernel with dri graphics for kindle paperwhite 3 (i.mx6s) Mar 11 01:03:59 i'm not much of a low level dev but I'm pretty handy/adventurous. Mar 11 01:04:37 edrex: welcome back Mar 11 01:04:41 reading up but figured I'd seek guidance in channel too Mar 11 01:04:56 thanks khem Mar 11 01:05:01 you can definitely do what you want perhaps its better now a days than it was back then Mar 11 01:05:40 all imx stuff is in meta-freescale Mar 11 01:05:55 it no more a monolithic tree Mar 11 01:06:02 so you have layers Mar 11 01:06:12 which pile on core layer Mar 11 01:07:24 i think i was paying a little attention when the layer thing was starting to happen. can't recall what hardware I had then Mar 11 01:12:17 maybe I should target some more common hardware first, to get to used to yocto. I have an rpi3 handy. Mar 11 01:12:32 yeah now perhaps you need to look into oe-core+bitbake+meta-freescale for what you want **** ENDING LOGGING AT Wed Mar 11 03:02:27 2020