**** BEGIN LOGGING AT Mon Feb 01 02:59:57 2021 Feb 01 06:47:31 opkg has a small reproducibility issue: https://autobuilder.yocto.io/pub/repro-fail/oe-reproducible-20210201-r996d8mk/packages/diff-html/ if anyone has a moment... Feb 01 07:39:43 good morning Feb 01 08:44:43 hello guys ! Feb 01 08:45:09 is it possible to have thoe source code of u-boot, in the shared folder like it's for kernel ? Feb 01 08:47:18 is anyone here involved in meta-rust? i tried to build the included rustfmt example, but it fails because rustfmt needs an unstable toolchain. Feb 01 08:49:38 thekappe: why do you want that? what exactly do you want to do? Feb 01 08:54:14 yo dudX Feb 01 09:00:39 LetoThe2nd: greetings Feb 01 09:17:57 qschulz, u-boot-fw-utils relies on u-boot {UBOOT_MACHINE}. The file related to this variable is added in the do_configure task of u-boot-xlnx (my actual virtual/bootloader). So u-boot-fw-utils fails because the defconfig is not available Feb 01 09:18:40 that is beacaust u-boot-fw-utils also compile u-boot. Feb 01 09:20:29 For now, I've created two bbappend in the same folder, one for u-boot and one for u-boot-fw-utils. Moreover I've created a "files" directory where I've put my defconfig. Finally I've added a SRC_URI to both the previously listed recipes so that they both have the desired config Feb 01 09:21:12 One other oprion was to have the source file of u-boot in the shared folder and add a proper CFLAG to u-boot-fw-utils Feb 01 09:24:25 thekappe: http://cgit.openembedded.org/openembedded-core/tree/meta/recipes-bsp/u-boot/libubootenv_0.3.1.bb?h=master Feb 01 09:25:25 that should probably remove your need for the defconfig and it's now the official way of interacting with u-boot environment from usersapce IIRC Feb 01 09:28:51 qschulz, that's good to know ! thanks ! Feb 01 09:29:35 hello. I wanted to try fedora for yocto development,yocto manual says it supports fedora 30, is fedora 33 not recommented then? Feb 01 09:31:25 shoragan: vmeson may have more recent work but won't be up yet or maybe paulbarker would know where it is Feb 01 09:32:18 RP, ah, thanks! Feb 01 09:32:19 kayterina: depends which version of the project you're using. We do run master builds on f33 Feb 01 09:38:42 * LetoThe2nd slaps RP PCly around with a large mastery trout! Feb 01 09:38:49 Feb 01 09:40:37 LetoThe2nd: er, what? :) Feb 01 09:41:58 RP: i thought all the cool kids would change their branch names to main to be super-PC? why u no kewl kid? ;-) Feb 01 09:42:01 qschulz, in thud there is no such libubootenv by the way Feb 01 09:42:08 amd I need to stick to it Feb 01 09:43:43 LetoThe2nd: ah, I'd probably go for devel just to be different ;-) Feb 01 09:43:57 RP: YAY! DIVERSITY! Feb 01 09:44:13 in case nobody noticed, i feel strange today. Feb 01 09:44:34 LetoThe2nd: It is noticable! :) Feb 01 09:44:51 phew! Feb 01 09:48:44 by the way: https://twitter.com/TheYoctoJester/status/1355134996403650561 Feb 01 09:51:22 LetoThe2nd: I thought one of those was a given! Feb 01 09:52:22 well i could also just (beer) and babble without "real" content :) Feb 01 10:10:30 thekappe: just backport it Feb 01 10:14:13 qschulz: thats evil. needs more döts. Feb 01 10:14:22 "bäckpört" Feb 01 10:22:01 LetoThe2nd: are you having a stroke? Feb 01 10:23:05 qschulz: nah. i'm just unusially awake, I guess. no interruptions this nights' sleep. and: https://en.wikipedia.org/wiki/Metal_umlaut Feb 01 10:24:03 LetoThe2nd: there's really a wikipedia page for anything Feb 01 10:24:13 yep. Feb 01 10:24:20 rburton: Feb 01 10:24:31 rburton: wake up i'm being annoying with metal again! Feb 01 10:28:05 qschulz: oh and for the record, i just went with a cheapo synology + 2 2.5 spinning rust disks for now. Feb 01 10:34:42 * RP thought he had a really neat idea to improve warning/error handling and sstate but log message events aren't delivered to class event handlers and even if they are, don't include the datastore :( Feb 01 10:35:14 * RP feels he's a few too many levels of different issues deep now Feb 01 10:37:56 LetoThe2nd: running mainline distro or synology's? Feb 01 10:38:42 qschulz: run of the mill synology. i just don't feel like investing any time. Feb 01 10:42:57 LetoThe2nd: do you at least have the opportunity to flash something decent on it the day you decide otherwise? Feb 01 10:44:02 qschulz: the day i decide otherwise i'll get something else and put it on ebay. because thats actually what i did the other way round a couple of years back, when i moved from a qnap to a hp microserver. Feb 01 10:44:20 LetoThe2nd: microserver gen8? Feb 01 10:44:31 maybe we can switch to private chat :p Feb 01 10:46:28 qschulz: its kind of interesting :) I've wondered what to do about this in the past but built a small server in the end Feb 01 10:50:40 RP: I also have a microserver gen8 back in the days when it was still dirt cheap Feb 01 10:50:45 +bought Feb 01 10:51:38 * mcfrisk wonders why meta-selinux enables UsePAM for openssh unconditially... Feb 01 10:51:50 only thing I don't like about it is the hack for booting from the 5th sata that breaks every now and then and your homeserver does not reboot anymore until you go into the BIOS and do the things by hand allover again Feb 01 10:52:43 (but you can boot from USB/SD card instead if you want, still haven't managed to put grub on either and then boot the distro from SSD and it's painful to debug on a machine that takes 10min for each try :p) Feb 01 10:54:13 qschulz: I ended up with a crazy xeon processor in my mini server for 'reasons'... Feb 01 10:54:17 I'm following a bit on what's next for AMD with either their mobile low-power SoCs or Zen3 architecture. Feb 01 10:54:28 RP: hehehehe I was looking to upgrade to a Xeon too :p Feb 01 10:54:44 qschulz: it was a staff purchase at the time ;-) Feb 01 10:55:03 :) Feb 01 10:55:31 had to buy a cheap processor to reflash the motherboard to a firmware that supported it :/ Feb 01 10:58:42 RP: :sobing: Feb 01 11:06:21 hrhr Feb 01 11:48:22 RP, JPEW: Apologies, my patches got completely mangled. No idea how, I used git-send-email Feb 01 11:53:11 paulbarker: they look ok here? Feb 01 11:54:07 RP https://lists.openembedded.org/g/bitbake-devel/topic/rfc_patch_1_8/80282574 Feb 01 11:54:21 I've just resent them all Feb 01 11:54:53 * paulbarker removes postfix from my VM and fixes gitconfig Feb 01 11:55:22 paulbarker: interestingly my mail client displays them ok, the raw format does look broken Feb 01 11:56:13 paulbarker: and they did save and apply ok. Very strange Feb 01 11:56:18 RP: It looks like my .gitconfig wasn't present on this VM. That usually results in an error message but instead patches got sent via the local postfix instance Feb 01 11:56:40 Local postfix instance is now removed. I'll add that to my list of things to do on new opensuse installs! Feb 01 11:59:09 paulbarker: patches look good at a quick glance. Bit worried we still have logger calls that are broken like that! Feb 01 11:59:23 paulbarker: I'll let JPEW take a look later Feb 01 12:00:39 RP: Thanks! Feb 01 12:00:49 Just in time to get some lunch then start looking at the prserv side Feb 01 12:08:32 paulbarker, does the meta-rust rustfmt-example work for you? it fails for me and complains about the toolchain not being nightly Feb 01 12:37:36 RP : You asked the other day about maintainership of meta-gplv2 (I only follow IRC while at work, and with the current rules due to the pandemic that only happens occasionally). From my point of view, I co-maintain meta-gplv2 with Ross as the main maintainer. It's not something we've discussed though. Feb 01 12:41:54 shoragan: Is that the one in `recipes-example/rustfmt` in meta-rust? In my (brief) testing I didn't get that to work no Feb 01 12:42:28 shoragan: https://gitlab.com/pbarker.dev/rust/meta-rust-demo has better demo recipes Feb 01 12:42:32 paulbarker, yes, that one. Feb 01 12:43:36 i also tried to create some test recipes, but failed with https://github.com/meta-rust/cargo-bitbake/issues/22 Feb 01 12:45:05 i'm now tring the referenced PR Feb 01 12:46:59 Saur: Ok, thanks for the info. I'm not sure how rburton feels about that layer at this point :/. Feb 01 12:47:50 RP, Saur: I considering myself very much backseat maintainer applying patches if needed. My official position is: burn it with fire. Feb 01 12:48:09 rburton: Ok, nice to know. ;) Feb 01 12:48:51 Hi folks , i m new to Yocto's world, i ve a quick question about packages splitting:D Feb 01 12:49:16 rburton: only fire? :) Feb 01 12:49:29 Alright Feb 01 12:52:40 for example , let's take libraries, how can i  only install (with cmake or any tool) the library SRCs and BINs and not the DOC ,DEV or Version with Debugging symbols (sorry for my explanation, i ve weak English  ) , i m in a context where i really really have to make a tiny image Feb 01 12:54:15 medaliyou: if the library build normally, then just install libfoo and not libfoo-dev libfoo-doc libfoo-dbg Feb 01 12:55:17 i mean how to ? editing the recipe and defining PACKAGE & FILES_XXX variables ? is that the correct way ? Feb 01 12:55:42 the default rules do it for you Feb 01 12:55:51 medaliyou: if your library does not do dirty tricks, this works out of the box. if you have to manually take care of it, you have to adjust PACKAGES and FILES, yes. see also https://www.youtube.com/playlist?list=PLD4M5FoHz-TxMfBFrDKfIS_GLY25Qsfyj especially #3 Feb 01 12:55:55 OR in the IMAGE_INSTALL variable ? Feb 01 12:57:00 Alright , Thank you guys  ! Feb 01 12:59:52 In general , what steps and/or approaches to consider when i want to Optimize Boot Time let's say from 25-27 sec to 5-8 sec ? Feb 01 13:01:48 medaliyou: It needs a mix of hardware and software optimisations usually Feb 01 13:02:01 medaliyou: schedule a lot of time and a exponential workflow Feb 01 13:02:58 medaliyou: https://www.youtube.com/watch?v=gIK1he6Ocpg was a good presentation about the sw side Feb 01 13:03:22 Does DMESG makes a good start ? Feb 01 13:04:39 medaliyou: well you can certainly look at dmesg to find some ideas... and there is more than enought presentations on boot time already on youtub.e Feb 01 13:05:23 medaliyou: but! 5-8s from power on to something useful is pretty hard. you might really want define your usecase properly first. Feb 01 13:05:41 I m so Grateful guys =D =D Feb 01 13:06:23 yeah i want to load the kernel and start the sys, and when it's done i have to load my QT & Python apps Feb 01 13:07:00 "you might really want define your usecase properly first." Feb 01 13:09:55 paulbarker, ok, i've now successfully build ripgrep and some other stuff. thanks Feb 01 13:55:18 paulbarker: Ya, I was a little worried about the logging call change... I don't think we are using the bitbake logger in the hash server? Feb 01 13:56:08 hello all. I don't speak Python but some of my collegues do. They want to install a few python application on their yocto system, but as I understand their dependencies differ in versions. I heard there is pyenv for that. How can I put multiple python apps on a system which run in different pyenvs? Someone good a good writeup on that? Feb 01 13:56:39 JPEW: I got backtraces instead of log output before I made that change. I can give it a closer look though Feb 01 14:13:13 paulbarker: Other than the logging, I had a few minor comments, but looks good otherwise! Feb 01 14:16:32 JPEW: Thanks for the review Feb 01 14:38:50 RP: Ah.... do_pacakge_qa has no inputs and outputs so it's going to be "equivalent" a lot? Feb 01 14:56:05 RP: I'm seeing reproducable builds breaking for (probably) every distro in the a-full from last night. Is this a known issue or should I open bugs? Feb 01 15:37:07 Morning folks Feb 01 15:37:16 is patchwork dead? Feb 01 15:50:46 JPEW: that is what is worrying me a bit :/ Feb 01 15:52:24 JPEW: the tasks fail so I think the issue is different to what I originally thought Feb 01 15:52:41 abelal: do you mean generally? ;-) Feb 01 15:53:42 jonmason: the opkg reproducibility #issue is a new bug Feb 01 15:57:52 @LetoThe2nd: hehehe talking about patchwork.openembedded.org Feb 01 15:58:03 abelal: hrhr Feb 01 15:58:26 I sent a patch and it did not show up, the last patch on the list is from Dec 2020 Feb 01 15:58:37 haven't had a response on the list either Feb 01 15:58:47 so I was worrying if I sent it correctly Feb 01 16:01:05 abelal: are you registered on the mailing list? Feb 01 16:01:08 (there's a new one) Feb 01 16:02:24 @qschulz: I am Feb 01 16:05:52 abelal: to which mailing list are you sending and what exactly? Feb 01 16:07:02 qschulz: sent a patch to openembedded-core@lists.openembedded.org titled "kernel.bbclass: fix deployment for initramfs images" Feb 01 16:12:48 abelal: https://lists.openembedded.org/g/openembedded-core Feb 01 16:13:01 there are many patches on it that were sent last month Feb 01 16:13:08 not yours AFAICT Feb 01 16:21:21 RP: Ya, I was curious how a failed task (which shouldn't be reporting any equivalencies anyway) would be causing a problem :/ Feb 01 16:40:34 JPEW: I was struggling to see how it could reproduce locally but not on the autobuilder, I think the answer is multilib Feb 01 16:53:49 you usually set IMAGE_FSTYPES in your distro conf, right? Feb 01 16:57:06 vdl: the machines often have a much better idea of what kind of image they might need Feb 01 17:51:41 What area should I open a reproducible builds bug in bugzilla? build system/oe-core? Feb 01 17:56:12 RP: ^^ Feb 01 17:58:10 jonmason: yes, we should document that if in doubt to use that Feb 01 17:58:21 jonmason: its possible we should file the bug against opkg! Feb 01 17:58:30 (in this specific case) Feb 01 17:59:38 RP: I'm still feeling my way around this. So expect lots (more) stupidity :) Feb 01 18:00:47 jonmason: its fine, we should add something to the page about it though Feb 01 18:15:46 Hi all, I would like to make my yocto to compile UBI image with config.ini file option. Maybe someone could point me to some example how to achieve that? Feb 01 19:27:14 JPEW, I went to look at getting the reproducible builds status set up. Can you check what's up with http://wattissoftware.com/yocto-r13y/ ? Feb 01 19:34:03 RP: should I triage all the other binutils autobuilder issues or have them all marked as "mail sent" as most are the kernel issue you already emailed khem about? Feb 01 20:09:03 jonmason: if they're all that issue they can be marked as such Feb 01 20:11:50 Errors look different than the linux issue, just wondering if they might be fixed by the same (and thus no reason to open bugs for them). Or if I should pile on in the email with 4-5 more things in it Feb 01 20:21:59 jonmason: do you have some examples I can look at? Feb 01 20:22:53 RP: https://autobuilder.yoctoproject.org/typhoon/#/builders/20/builds/3308 Feb 01 20:28:27 jonmason: in the email to khem I mentioned two issues, one was the kernel build issue, the other was a nativesdk parallel make install issue Feb 01 20:28:49 jonmason: that means nativesdk-binutils:do_install issues were also reported Feb 01 20:32:06 thanks Feb 01 20:34:07 RP: for the gatesgarth patch queue autobuilder, is that what's queued in master-next? Feb 01 20:48:01 jonmason: I don't understand the question. gatesgarth's "next" is in poky-contrib Feb 01 20:48:14 jonmason: also, I believe the owner of the build would look after that? Feb 01 20:49:41 halstead: Thats the cross site scripting check I was talking about Feb 01 20:50:43 The browser won't let the js code read the JSON file from git.yoctoproject.org because the origin (wattissoftware.com) doesn't match Feb 01 20:54:02 halstead: Or... it could also be because the JSON file is missing Feb 01 20:57:10 halstead: Ya, it appears that http://git.yoctoproject.org/cgit.cgi/yocto-testresults/plain/oeselftest/testresults.json no longer exists... Feb 01 20:57:12 ^^ RP Feb 01 21:08:38 RP: Sorry, I'm looking at https://autobuilder.yoctoproject.org/typhoon/#/builders/83/builds/1813 and it has a title of "Reason: test gatesgarth patch queue; remove uri patch". So, my understanding is that I should try to find the patches in question, so that i can email the authors (or respond to the patch submissions) and tell them there is a problem. Feb 01 21:09:39 A quick look on the mailing lists only has CVE patches for gatesgarth and doesn't seem to be hitting the same places that are failing Feb 01 21:16:31 jonmason: if you click on "Build Properties" on that link you'll see a lot more information about that build Feb 01 21:17:11 jonmason: you'll notice that swat_monitor is false Feb 01 21:17:33 jonmason: that was anujm (the owner) saying they would triage it Feb 01 21:18:03 jonmason: haven't quite got that info to swatbot as some of the builds from auto triggers don't set it correctly :( Feb 01 21:18:30 thanks, lots of learning going on Feb 01 21:18:43 jonmason: better documentation needed I think :/ Feb 01 21:19:08 jonmason: that properties page also shows which branches and locations the code came from Feb 01 21:19:24 on Friday, I'll do a brain dump on the wiki with lessons learned Feb 01 21:19:53 jonmason: might be better to add as you learn? I'm worried some of this may be forgotten by then :/ Feb 01 21:21:00 this "owner" thing is significant. Hopefully some of the bugs I opened today aren't garbage Feb 01 21:21:17 jonmason: "owner" is who started the build Feb 01 21:21:19 I'll add that Feb 01 21:21:35 jonmason: that is supposed to be displayed by swatbot too Feb 01 21:21:37 I'm assuming that anything with an owner can be ignored Feb 01 21:21:53 after notifying them its broken Feb 01 21:24:20 jonmason: except master-next Feb 01 21:34:37 there's no way to install the rootfs with wic in a sub-directory via --source rootfs, right? Feb 01 21:36:12 jonmason: I did add the bit about swat_monitor on the wiki last Friday Feb 01 21:36:27 If it is not clear engouh, don't hesitate to rephrase Feb 01 21:37:18 I triaged a few build from anuj today so issues wouldn't pile up too much for you Feb 01 21:40:22 abelloni: I saw it this morning and did appreciate it Feb 01 21:40:53 I'm trying not to interfere too much so we can figure out what we need to do with the process... Feb 01 21:41:14 abelloni: but based on what RP said above, anujm's don't need to be triaged, correct? Feb 01 21:41:36 jonmason: ultimately we need better presentation in the UI Feb 01 21:42:08 the UI will improve over time, with feedback from people like me who have no idea WTF they are doing :) Feb 01 21:42:29 jonmason: yes, I put them as Handled, Not to be triaged Feb 01 21:42:53 BTW, di you see the "SWAT Pending Builds" at the top left ? Feb 01 21:42:54 abelloni: cool, there are a couple more from him. Just making sure I understand what is to be done Feb 01 21:43:28 abelloni: yes, I'm using that exclusively. Very nice Feb 01 21:44:20 well, I just ensured swat_monitor was set to false for his builds before setting them to Handled, Not to be triaged Feb 01 21:45:36 I have it down to 3 now, and one is anujm, khem, and no owner field Feb 01 21:46:19 we discussed the warning with RP already Feb 01 21:46:29 I'm going to open a bug for this one Feb 01 21:46:58 this is a common issue you will see, a ptest timeouting Feb 01 21:47:58 https://git.openembedded.org/openembedded-core/commit/?id=5af9f32d9b12654793289f44366251f978f6378a was supposed to fix that Feb 01 21:48:11 * RP wonders why the owner info doesn't show up Feb 01 21:49:21 to be 100% clear, for those 2 I should email them and tell them to set the swat_monitor to false in their future builds, show them there are failures, and the clear the pending with "emailed them" Feb 01 21:50:20 jonmason: SWAT monitor field is there as a "does SWAT need to help with this". I'm not sure whether they set it or not intentionally or otherwise :/ Feb 01 21:51:23 at least Saul was not aware of the field last week Feb 01 21:51:24 RP: should default be false? Feb 01 21:52:22 jonmason: there are bugs on the buildbot UI side where it doesn't so on every form. There are then issues with it not being set correctly for automatic builds :/ Feb 01 21:52:42 jonmason: I'll try and figure out what is going on with it Feb 01 21:53:09 jonmason: I prioritised having swatbot working at all rather than chasing down this particular rabbit hole :/ Feb 01 21:59:17 RP: BTW, I looked around for the rpm/perl issue: Feb 01 21:59:20 error: /etc/rpm/macros.perl: line 35: Macro %global is a built-in (%define) Feb 01 21:59:29 I think we only see it on centos Feb 01 21:59:54 it is probably rpm-native taking that file from the host when running Feb 01 22:00:25 this is hidden on fedora because they fixed /etc/rpm/macros.perl Feb 01 22:03:55 abelloni: can we stop it looking at the host? Feb 01 22:04:25 jonmason, abelloni: some research suggests the owner data is being put into swatbot and its not displaying correctly :/ Feb 01 22:04:50 I don't know yet Feb 01 22:05:28 RP: https://autobuilder.yoctoproject.org/typhoon/#/builders/79/builds/1784 is an autoconf issue in rpm-native, on gatesgarth Feb 01 22:05:28 RP: all good, one more email and I'm done. Feb 01 22:06:36 yes, that is the last one :) Feb 01 22:07:09 yeah :) Feb 01 22:10:00 jonmason: you can add the new one to 14210 Feb 01 22:10:17 abelloni: sweet! Feb 01 22:10:26 caught up until tomorrow :) Feb 01 22:14:57 abelloni, jonmason: http://git.yoctoproject.org/cgit.cgi/swatbot/commit/?id=4e0020b2b0dc563017d425d9b2f5d2b04f8f08fd will either work for people or we'll hate it... Feb 01 22:15:54 ok, we'll see ;) Feb 01 22:17:22 jonmason: which build didn't have swat_monitor set? Feb 01 22:17:59 jonmason: or did it always have a true/false value? Feb 01 22:18:10 it didn't have an owner either Feb 01 22:18:14 looking for it now Feb 01 22:18:54 jonmason: lack of owner tells it it was auto-triggered, like https://autobuilder.yoctoproject.org/typhoon/#/builders/83/builds/1809 Feb 01 22:19:22 jonmason: that helps as I can track down why its not set, I think its the a-quick and -afull autotriggered builds Feb 01 22:21:56 I think it was https://autobuilder.yoctoproject.org/typhoon/#/builders/93/builds/800, but I see the field now. So either it was added since I looked or I didn't see it the first time. Feb 01 22:25:45 jonmason: interesting, its in the children but not the parent (https://autobuilder.yoctoproject.org/typhoon/#/builders/85/builds/1315) Feb 01 22:27:34 jonmason: going to guess at http://git.yoctoproject.org/cgit.cgi/yocto-autobuilder2/commit/?id=a20082d067e618c701e30e6d00916de423dad02e Feb 01 22:29:07 jonmason: cancelled a couple of perf builds and restarted the controller **** ENDING LOGGING AT Tue Feb 02 02:59:57 2021