**** BEGIN LOGGING AT Thu Feb 08 03:00:02 2018
Feb 08 10:53:36 <lukma> Dear All,
Feb 08 10:53:42 <lukma> I do have a question regarding wic
Feb 08 10:53:55 <lukma> is there any way to use wic to produce binary image?
Feb 08 10:54:12 <lukma> (I mean to replace hand crufted "dd
Feb 08 10:54:20 <lukma> in a shell script
Feb 08 10:54:52 <lukma> Now I do have my own image.bbclass, which provides support for my image
Feb 08 10:55:20 <lukma> but I think that wic will pose less maintanence in the future......
Feb 08 10:56:03 <lukma> so is it possible to ask wic to place u-boot.bin at offset X
Feb 08 10:56:11 <lukma> and some rootfs at offset Y
Feb 08 10:56:25 <lukma> and some other binary data (like first bootloader) to offset Z
Feb 08 11:18:45 <mcfrisk> hi, is yocto 2.5 release on schedule? Are there updates coming soon to milestone 2 of https://wiki.yoctoproject.org/wiki/Yocto_Project_v2.5_Status ?
Feb 08 11:20:28 <rburton> mcfrisk: in qa now iirc
Feb 08 11:20:36 <rburton> m2 was a little delayed thanks to meltdown
Feb 08 11:21:25 <mcfrisk> rburton: understood. I'm planning a yocto update so would like to jump to master/2.5 directly
Feb 08 11:22:58 <berton> lukma (IRC): see this wks file https://github.com/Freescale/meta-freescale/blob/master/wic/imx-uboot-spl.wks
Feb 08 11:59:06 <parthi> berton: "--source rootfs" this points to which file in tmp/deploy directory?
Feb 08 11:59:39 <parthi> berton: is there a way to specify specific file to "--source" ?
Feb 08 12:00:03 <lukma> berton: It looks like a solution
Feb 08 12:00:22 <berton> parthi (IRC): What file do you need to add?
Feb 08 12:00:28 <lukma> it is just a pity that it has only 1KiB granularity (with --align XX)
Feb 08 12:00:55 <berton> You can write a recipe to install files in rootfs
Feb 08 12:01:15 <berton> so this files are part to "--source rootfs"
Feb 08 12:01:36 <lukma> berton: I also would need to specify the exact size -> like 16 MiB
Feb 08 12:01:56 <parthi> berton: ok, my question is not about including some binaries or files into the rootfs.
Feb 08 12:02:29 <parthi> berton: "part / --source rootfs --ondisk mmcblk --fstype=ext4 --label root --align 4096" which rootfs file will be used in this case?
Feb 08 12:03:42 <parthi> berton: I have a rootfs image file named "beld-image-base.ext4.signed". I want this to be flashed.
Feb 08 12:03:53 <berton> parthi (IRC): if you're building a core-image-minimal, is a core-image-minimal rootfs file foung in deploy directory
Feb 08 12:11:05 <parthi> berton: ok, the "rootfs" plugin of wik looks directly into my rootfs directory and builds the ext4 image. It's looking into the deploy directory I believe.
Feb 08 12:11:28 <parthi> berton: Correct me if I am wrong?
Feb 08 12:12:28 <parthi> berton: s/wik/wic/
Feb 08 12:16:10 <lukma> berton: Do you maybe know how to make the exact size of the generated image?
Feb 08 12:16:17 <lukma> (like 16 MiB, sharp)?
Feb 08 12:17:08 <berton> parthi (IRC): Do you want to set a size for a partition?
Feb 08 12:18:12 <berton> part / --source rootfs --ondisk mmcblk0 --fstype=ext4 --label rootfs --size 16M --align 8192 --extra-space 0
Feb 08 12:24:10 <parthi> berton: no, I need to sign the image after it prepares the ext4 file
Feb 08 12:26:58 <LetoThe2nd> oO( i just read: "i need to sing the image" )
Feb 08 12:36:19 <parthi> LetoThe2nd: Yeah, I have bbclass functions written to append the sign using openssl to the image.
Feb 08 12:37:02 <parthi> LetoThe2nd: wic picks the rootfs image, but not the signed one.
Feb 08 12:38:36 <LetoThe2nd> parthi: hehe i know. all is well. (in case you wonder why i typed it, read again *very* precisely.)
Feb 08 12:40:00 <parthi> LetoThe2nd: :-)
Feb 08 13:25:02 <parthi> berton: wic command either in raw/cooked mode takes the path to rootfs files and prepares the fstype'd wic image.
Feb 08 13:26:43 <parthi> berton: What I wanted was to prepare ext4 image and append signature to it. Then use wic to prepare wic image. Is this possible?
Feb 08 13:57:21 <eduardas_m> hello, I am trying to use bmaptool as described in the Yocto manual for the first time and failing: https://pastebin.com/vPzA3nng
Feb 08 13:57:39 <eduardas_m> What am I doing wrong?
Feb 08 13:59:39 <eduardas_m> I have run bitbake bmap-tools-native as described in the documentation, have the appropriate fod-testing-image-fod-som.wic.bmap in my deploy directory too
Feb 08 14:01:16 <nayfe> don't know if it could help someone, i made usbmount recipe to automount usbstick https://github.com/nefethael/meta-random/blob/master/recipes-support/usbmount/usbmount_git.bb
Feb 08 14:55:54 <LetoThe2nd> behanw: ping concerning ypdd beginner track
Feb 08 15:33:41 * armpit doesn't think LetoThe2nd needs beginner track
Feb 08 15:34:04 <jaeckel> I'm on the latest krogoth branch and I can't build because of the same error as reported here https://www.mail-archive.com/yocto@yoctoproject.org/msg31427.html does someone know how to fix that?
Feb 08 15:35:40 <mckoan> jaeckel: I'd ask to meta-udoo people
Feb 08 15:37:37 <LetoThe2nd> armpit: in fact, i will be doing something comparable to the beginner track on a conf in june, and have some material licensing questions :)
Feb 08 15:37:50 <LetoThe2nd> armpit: but thanks for your confidence!
Feb 08 16:08:41 <T_UNIX> is UPSTREAM_CHECK_URI the intended way to force bitbake to refetch source code using the same SRC_URI?
Feb 08 16:18:21 <rburton> T_UNIX: no
Feb 08 16:18:47 <rburton> T_UNIX: if you want bitbake to fetch latest srcrev every build just set SRCREV="${AUTOREV}"
Feb 08 16:26:35 <samoud> Hello
Feb 08 16:27:00 <sgw> Morning all
Feb 08 16:32:46 <behanw> LetoThe2nd: pong
Feb 08 16:33:16 <LetoThe2nd> behanw: \o/
Feb 08 16:33:43 <behanw> What’s your question?
Feb 08 16:33:55 <ulf`> behanw: Are you THE Behan Webster?!?!?!
Feb 08 16:33:56 <ulf`> OMG OMG!
Feb 08 16:33:57 <ulf`> ;)
Feb 08 16:34:16 <LetoThe2nd> behanw: guess for you its rather simple to answer: what is the licensing on the ypdd materials? is it encouraged, tolerated, discouraged to base other workshops off it?
Feb 08 16:42:01 <T_UNIX> rburton: problem is, it's a `artifacts.zip` that contains the version info file.
Feb 08 16:42:38 <rburton> T_UNIX: well that sounds like fun. how do you know when it changes?
Feb 08 16:42:51 <LetoThe2nd> behanw: (i need to afk now, maybe laters)
Feb 08 16:43:10 <T_UNIX> rburton: I won't. At least by the url
Feb 08 16:43:31 <T_UNIX> it's kind of like `https://foo.bar/download/latest.zip`
Feb 08 16:43:52 <rburton> well thats non-trivial to know when it changes unless you want it to rebuild *every* time
Feb 08 16:44:13 <T_UNIX> rburton: it's bin_package.bbclass derived
Feb 08 16:44:42 <rburton> if youre happy for that to rebuild *every* time then you can probab;y force that
Feb 08 16:44:59 <T_UNIX> for some context: gitlab's artifacts api allows to fetch the most recent succesful build artifacts for branches and tags.
Feb 08 16:45:53 <T_UNIX> so I'm using that to compose a 'nightly'  image, containing the binaries genrated by the various teams.
Feb 08 16:49:31 <T_UNIX> the `latest.zip` contains a file `versions.txt`. AFAIK there is some sort of "I'll provide the version on my own" thing, right? But how do I force bitbake to refetch by the recipe's means?
Feb 08 16:52:52 <behanw> ulf`: you laugh. But I’ve heard that before. And having a unique name I always inwardly have to answer “duh”
Feb 08 16:54:10 <behanw> LetoThe2nd: as it states on the e-ale.org website all material is CC by-SA and all related lab code GPL/MIT
Feb 08 16:54:45 <behanw> So as long as attribution is given the Yocto training bits are freely usable.
Feb 08 16:55:45 <behanw> LetoThe2nd: oh sorry. Missed you. I was driving when you asked your question. (At destination now of course)
Feb 08 17:19:02 <khem> rburton: I am sending a v2 of binutils patch, I saw aarch64 build regressions which should be addressed there
Feb 08 17:20:01 <rburton> cool
Feb 08 17:21:41 <khem> T_UNIX: are you dealing with maven ?
Feb 08 17:22:03 <T_UNIX> khem: no, just gitlab's Ci
Feb 08 17:22:54 <khem> once upon a time, I dealt with maven which had similar setup where latest.zip was what contained the version
Feb 08 17:23:30 <needmoarram> hi, does anyone know how I can pass in SSH_AUTH_SOCK variable to a recipe from my environment?
Feb 08 17:23:39 <khem> solved it in fetcher
Feb 08 17:23:58 <khem> needmoarram: for doing what ?
Feb 08 17:24:05 <needmoarram> I tried to `export SSH_AUTH_SOCK` in my conf/local.conf, but that triggers a rebuild of all packages everytime my agent socket name changes
Feb 08 17:24:28 <needmoarram> I need to scp some files in addition to installing them
Feb 08 17:25:23 <needmoarram> khem: ^
Feb 08 17:25:55 <T_UNIX> khem: yeah, I'm just trying to stick to the default as much as possible
Feb 08 17:36:42 <khem> needmoarram: add it to bb env whilelist
Feb 08 17:36:54 <khem> rburton: patch is on ml now and here too https://github.com/kraj/openembedded-core/commit/4c6549bee0c0ba748f37901232c8eb6a1433fe85.patch
Feb 08 17:37:08 <rburton> cheers khem
Feb 08 17:37:22 <khem> just finished a aarch64/rpi3 build
Feb 08 17:37:57 <khem> this should also fix the glibc issue you were seeing with bintutils 2.30/glibc 2.26
Feb 08 17:38:26 <khem> so next issue was meta-gpl2 make and glibc locale issue in eSDK right
Feb 08 17:38:42 <rburton> yes
Feb 08 17:39:06 <rburton> i've just done a tweak to the bbappend in meta-mingw so that can get exercised next build run
Feb 08 17:40:28 <khem> ok
Feb 08 17:55:19 <T_UNIX> is there a recommended way to add systemd drop-ins?
Feb 08 17:56:05 <T_UNIX> I assume automounting via udev rules+script is broken (https://unix.stackexchange.com/questions/152485/mount-is-not-executed-when-called-by-udev) for many rocko users currently?
Feb 08 17:58:44 <needmoarram> khem: thanks for the suggestion
Feb 08 17:59:14 <needmoarram> adding it to BB_ENV_WHITELIST didn't work b/c an empty path was somehow added to the PATH variable
Feb 08 17:59:44 <needmoarram> adding it to BB_ENV_EXTRAWHITE generates a lot of complaints about metadata not being deterministic
Feb 08 19:09:30 <xtron> how to ensure that a specific  image feature can't be added in variable like IMAGE_FEATURES, I do remove a value but later another recipe add it again, any way ?
Feb 08 19:14:33 <khem> IMAGE_FEATURES are not global, they just define what options you want to select to include in image
Feb 08 19:20:15 <xtron> khem: is it something done at parsing time, I don't how value is still there
Feb 08 19:20:47 <xtron> so If I remove a value at parsing time, it shouldn't be added any way
Feb 08 19:21:04 <khem> which feature are you looking for
Feb 08 19:22:04 <xtron> I'm removing a feature like read-only-rootfs
Feb 08 19:25:02 <khem> ok and how are you doing it
Feb 08 19:27:07 <xtron> IMAGE_FEATURES_remove = "read-only-rootfs" should work, but in environment -e, it do exist
Feb 08 19:28:09 <khem> rburton: I am trying to run bitbake core-image-minimal -c testsdkext and its complaing about python being python3 however in env I set the right paths so bitbake <imahge> works fine
Feb 08 19:28:27 <khem> rburton: I wonder where is testsdkext poking for python
Feb 08 19:28:34 <khem> it seems to be ignoring my env
Feb 08 19:29:34 <LetoThe2nd> behanw: ah kay. /me ponders actually tpying something into the e-ale chat now :)
Feb 08 19:30:15 <behanw> LOL. It just gets to me.
Feb 08 19:30:47 <LetoThe2nd> guessed that.
Feb 08 19:31:09 <behanw> LetoThe2nd: For those who don't know how to ask questions on IRC or $WHATEVER
Feb 08 19:31:14 <LetoThe2nd> hehe
Feb 08 19:32:01 <khem> xtron: bitbake -e should tell you where its being added
Feb 08 19:32:54 <LetoThe2nd> behanw: nah, seriously. especially the "what is yocto" bits, the data flow diagram of the build and such are things that i don't want to re-invent. i feel its better if the attendants can remember those if they hopefully check the docs then later
Feb 08 19:33:27 <behanw> Right
Feb 08 19:36:12 <LetoThe2nd> so those are the bits i'd actually like to take out of the existing material. for the contents, i'll probably pick up some ideas, re-vamp them to latest devtool and see how much i need to mix in. running time is 7hrs, so i'll totally have a sore throat afterwards.
Feb 08 19:37:23 <behanw> Yah big baby. I have to do this 4 or 5 days in a row most classes.
Feb 08 19:37:45 <behanw> And the worst is when people want us to squish a 4-5 day class down into 2-3 days
Feb 08 19:37:52 <LetoThe2nd> hrhr
Feb 08 19:38:27 <behanw> Or my personal favourite. With one customer I taught the entire 4-day LF yocto course in 1 day. (no labs however)
Feb 08 19:38:38 <behanw> Just me talking very fast and them taking notes.
Feb 08 19:38:46 <behanw> That was very unfun.
Feb 08 19:46:41 <batman_> Hello, Is there a C like struct "data structure" in bitbake?
Feb 08 19:50:32 <kergoth> not with the bitbake file format, anyway
Feb 08 19:55:03 <khem> behanw: 4 day course in 1 day, it seems that manager was performing some rituals
Feb 08 19:56:02 <behanw> khem: There were $REASONS. But indeed it was insane. It was meant as review.
Feb 08 19:56:26 <behanw> Whenever people do these sorts of rush jobs, they *ALWAYS* regret it.
Feb 08 19:57:31 <behanw> I've even had a VP apologize to me personally for not just doing the longer version.
Feb 08 19:57:57 <behanw> Which is one of the reasons I've learned never to complain about the people making the decisions about a short day. Because sometimes they're in the room. ;)
Feb 08 19:59:07 <fray> lol
Feb 08 19:59:31 <batman_> kergoth: Thanks
Feb 08 20:05:44 <LetoThe2nd> behanw: hehe. one thing i hope that i can push is the 50/10 scheme.
Feb 08 20:08:07 <LetoThe2nd> <- attended a 2-day seminar by alexei alexandrescu some time ago. he was super strict about talking 50mins straight, then 10min break. best thing i ever saw in respect to retainning attention by the audienc.e
Feb 08 20:12:30 <rburton> khem: damned arch
Feb 08 20:19:05 <khem> rburton: heh, I think we are just trailblazing for slackers
Feb 08 20:19:25 <rburton> khem: the cause of a world of pain by making python v3
Feb 08 20:20:15 <khem> rburton: I am sticking symlinks in openembedded-core/scripts but it still ignores
Feb 08 20:20:51 <khem> somewhere this testsdk is re-invoking setup script after pruning my env
Feb 08 20:20:55 <khem> second time
Feb 08 20:21:01 <khem> this is wierd
Feb 08 20:22:40 <khem> https://gist.github.com/e2f66735ee533b46b2dddda06c1b1dd1
Feb 08 20:23:23 <rburton> when you find the offending script that is using python instead of python2, at least that's an easy fix :)
Feb 08 20:28:44 <khem> i think the script is scripts/oe-buildenv-internal
Feb 08 20:28:55 <khem> but it depends on what one fed into it
Feb 08 20:29:21 <khem> there is a check for py2 and py3
**** ENDING LOGGING AT Fri Feb 09 03:00:01 2018