**** BEGIN LOGGING AT Wed Feb 13 02:59:56 2019 Feb 13 08:20:44 Hi guys, I sent a patch on the mailing list, is it the correct way to do that ? Feb 13 08:21:30 hi, I sent a patch on the mailing list, is it the correct place to send it ? Feb 13 08:23:11 Hi guys, I sent a patch on the mailing list, I don't know if that's the correct way to do so ? Feb 13 08:23:44 yacar_: it probably is, but no need to ask 3 times in a row. IRC is a rather slow medium, usually. Feb 13 08:27:34 sorry I had connection problem, my jitsi is a crazy b x) Feb 13 08:29:10 yacar_: it depends a bit on what you actually want to patch - if its something that actually comes from oe-core, then the openembedded-core ML might be a better match. but generally the yocto ML is a good starting point. Feb 13 08:35:41 It's in poky/scripts Feb 13 08:36:49 The mail on the mailing list is called "Fixing devtool issue with kernel bbhappen do_patch recipe", I think that there's few stuff I didn't do properly since that's my first patch Feb 13 08:41:34 bluelightning: ^^^^^ maybe you can comment? Feb 13 10:06:00 yacar_: LetoThe2nd: will take a look tomorrow, thanks Feb 13 10:23:52 Thanks :) Feb 13 10:51:45 New news from stackoverflow: How to remove systemd's service in yocto? Feb 13 10:58:05 hi.. what is best way to compare recipes (package versions) between poky versions? Feb 13 10:59:24 diff? Feb 13 10:59:26 what do you want to compare Feb 13 10:59:47 in openembedded comes with layer index: https://layers.openembedded.org/rrs/recipes/OE-Core/2.5/All/ Feb 13 11:00:03 that is a good starting point.. something similar for poky? Feb 13 11:00:21 rburton: i need to know which recipes did change from 2.2 to 2.5 Feb 13 11:00:44 enuuts: poky *is* openembedded Feb 13 11:00:51 i can use git for this Feb 13 11:01:30 rburton: you mean the layer index page is for both? poky and meta-openembedded? Feb 13 11:02:00 the layer index is for the greater OE ecosystem Feb 13 11:02:11 poky is just oe-core + bitbake + meta-poky in a single git repo Feb 13 11:02:58 understood, thanks. can we compare releases with a web frontend (like layer index page)? Feb 13 11:03:14 or should i clone and use git to compare? Feb 13 11:05:43 well eg https://layers.openembedded.org/layerindex/branch/master/layer/openembedded-core/ has a export csv button, so you could download the csv for the two releases you care about Feb 13 11:06:26 very good! Feb 13 11:21:05 did nobody notice in the perf reports that images suddenly grew by 20mb Feb 13 11:21:12 * rburton has a fix Feb 13 11:38:10 anyone intimately familiar with swupdate? how would one write image using raw handler to mmc, but from image compressed with .xz ? there is archive handler, but i don't know how to link that with raw handler Feb 13 12:53:20 Has someone installed Sierra's QMI SDK on its device? Feb 13 14:01:30 anyone know where the recipe aintainers file hides? Feb 13 14:03:17 Crofton: behind the curtain. thats at least my guess. Feb 13 14:12:55 Crofton: its called maintainers.inc Feb 13 14:13:23 Where is it? Feb 13 14:13:41 I am looking around in the poky repo Feb 13 14:14:28 lol Feb 13 14:14:54 path: root/meta/conf/distro/include/maintainers.inc Feb 13 14:15:22 may i introduce you to 'find' ;) Feb 13 14:16:04 or if you're feeling adventurous, git ls-files|grep maintainers.inc Feb 13 14:17:14 well, I do not have anything checked out with poky in it Feb 13 14:20:30 surely you have an oe-core clone Feb 13 14:20:38 its in there Feb 13 14:22:06 hmm, I had figured it was hidden in poky Feb 13 14:22:40 nope Feb 13 14:22:56 poky over oe-core is basically a few bsps that only qa use, and a minimal distro config Feb 13 14:23:12 it *used* to be in poky a while ago, but moved to core Feb 13 14:23:26 this is what happens when you get old Feb 13 14:26:07 re find/git-ls-files, see also ack -g, ag -g, rg -g, and fd. i'm a fan of fd recently :) Feb 13 14:26:38 kergoth: fd? filedescriptor? Feb 13 14:29:56 https://github.com/sharkdp/fd Feb 13 14:32:30 fd is great Feb 13 14:34:34 ah i see. Feb 13 14:34:56 fd to find is like ag to grep, right? Feb 13 15:17:12 yeah Feb 13 15:17:16 like rw to dd Feb 13 15:17:21 i reallly love rw Feb 13 15:18:16 adelcast: are opkg-utils fine to work with? py2 only i assume? Feb 13 15:18:59 https://sortix.org/rw/ in case anyone wanted to know ;) Feb 13 15:25:03 * LetoThe2nd notes Feb 13 15:27:31 rburton: Is the recipe image size grow checked when building master-next, or only after all the patches are merged to master? Feb 13 15:27:48 Ugh, that was a terrible sentence Feb 13 15:30:38 Piraty: I have use some of the tools with Py3 too, but haven't done extensive testing on all of them Feb 13 15:30:52 maybe i will in the future Feb 13 15:31:13 and yes, opkg-utils is the preferred set of tools to create packages, build indexes, etc Feb 13 15:31:23 not super polished set of tools, but they work Feb 13 15:48:01 RP, do we still have weekly builds enabled? Feb 13 15:52:13 adelcast: thanks Feb 13 15:52:17 JPEW: the build perf machines are for master Feb 13 15:52:38 JPEW: i usually do periodic master-next builds and check stuff like that but haven't had the time recently :( Feb 13 15:53:07 Ya, I wasn't aware that oe.utils.packages_filter_out_system even existed :( Feb 13 15:53:22 does anyone happen to be familiar with meta-qt5, and how to get qmlcachegen on sdk? Feb 13 15:57:10 JPEW: its useful, if you ever see any other recipes trying that logic by hand then you know to fix it :) Feb 13 15:57:23 Ya, makes sense Feb 13 16:18:38 rburton, my understanding is the perf data can be used to make comparisons between branches as well Feb 13 16:22:24 * armpit time for a sumo base line Feb 13 16:22:45 yeah the machines could generate the data, the storage tracks the branch Feb 13 16:59:35 rburton: "rootfs size278.52 MiB+80.0 kiB+0.0 %" is the huge step or am I looking at the wrong part of that e-mail? Feb 13 17:00:13 the third chart down iirc shows rootfs size of core-image-sato, and about half way along is a big step Feb 13 17:00:38 i couldn't be bothered to find the exact mail which has the change as it happened :) Feb 13 17:01:14 armpit: no, we could define something though Feb 13 17:02:13 RP, k.. I can do it manually for the stable branches. Feb 13 17:02:17 rburton: ah I see now, thanks Feb 13 17:02:44 wednesday 30th jan 00:02 for centos7 shows the actual change, 53121 - 53107 Feb 13 17:02:51 thats when it hit my inbox anyway Feb 13 17:03:01 armpit: http://git.yoctoproject.org/cgit.cgi/yocto-autobuilder2/commit/?id=ee71955e5c516c0f2cedf534a7aaef842d312524 Feb 13 17:03:05 armpit: dead easy to add Feb 13 17:04:09 armpit: my plan is to have a perf baseline for the stable branches as well Feb 13 17:04:28 armpit: been ironing out issues before trying that but it should work now Feb 13 17:04:42 armpit: first build will give a error as "nothing to compare to" Feb 13 17:08:32 RP, thanks for getting that working Feb 13 17:09:31 armpit: next is to get reports for all QA, not just build perf :) Feb 13 17:09:42 armpit: and include perf builds in a-full Feb 13 17:09:51 one of those is easier than the other :/ Feb 13 17:28:26 I have a question on the yocto-check-layer script. Should I be running this with my layer already added to BBLAYERS? Feb 13 17:42:37 Looks like that's a no. yocto-check-layer expects to add the layer itself. That probably needs to be made clearer in the docs Feb 13 18:03:18 paulbarker, I think there is an option to have look at the BBLAYERS, I believe I use -n then call out all the layers I want checked Feb 13 18:04:40 Ok will re-read the help once I'm done with my current testing Feb 13 18:07:21 I can't understand what -n/--no-auto does Feb 13 18:08:25 I can see that yocto-layer-check needs to first capture signatures without the layer included, then add the layer and see what signatures have changed. So for that initial capture to work the layer has to be absent from BBLAYERS Feb 13 18:11:06 paulbarker: there is a patch about adding duplicated in master-next Feb 13 18:11:22 paulbarker: we really need some people helping tweak a few of these things :/ Feb 13 18:11:58 paulbarker: we should perhaps fail if the layer was already added. Could you file a bug please? Feb 13 18:12:22 RP: I'm still really limited on time sadly so can't do too much tweaking. I can file a bug though Feb 13 18:15:29 paulbarker: thanks, that in itself is a big help Feb 13 18:23:53 Done: https://bugzilla.yoctoproject.org/show_bug.cgi?id=13176 Feb 13 18:23:54 Bug 13176: normal, Undecided, ---, richard.purdie, NEW , yocto-check-layer can only check layers which aren't already in BBLAYERS Feb 13 18:30:17 RP: I've also filed a bug for the sstate file name lengths (https://bugzilla.yoctoproject.org/show_bug.cgi?id=13177). Hoping to get back to that but got a bunch of higher priority items to clear first Feb 13 18:30:18 Bug 13177: normal, Undecided, ---, ross.burton, NEW , sstate file names can exceed 255 char limit Feb 13 18:41:20 paulbarker: good thinking, this means we shouldn't forget Feb 13 21:23:50 I have a question on devtool sdk-update, We have a requirement that our esdk is extracted in publish mode as its on an nfs mountpoint and bitbake doesnt like nfs. There used to be two ways to update sstate, (1)through the install_sstate_objects function which manually copies sstate objects (for local updates) and (2) setting SOURCE_MIRROR in site.conf to the updateserver and running bitbake setscene (for remote updates), after this patc Feb 13 21:23:50 h https://gitenterprise.xilinx.com/Yocto/poky/commit/02986886fc40c0fcd33359a279ba9520fd5cac2a the former is no longer supported. What would be the best way for us to update sstate through sdk-update now? as we can't do use bitbake setscene.. seems this is a gap? Feb 14 00:03:39 is it possible to docker with yocto advice Feb 14 00:04:16 Hi, I've been struggling a bit with this question. I want to use Yocto to build two+ products with separate dev/prod images for each (dev including debug-tweaks, etc.). I've ruled out separate image recipes because my dev builds need ENABLE_UART on my RaspberryPi and that needs to be set at the conf level (I've got it set conditionally in my base dist conf). Multiconfig looked promising, but I'm not happy about how long the pars Feb 14 00:04:38 --postread looks nice though Feb 14 00:04:55 but it's barely mentioned anywhere and I'm worried that it's not well supported Feb 14 00:04:58 what do most people do? Feb 14 02:51:26 How can I have the kernel headers in my rootfs? **** ENDING LOGGING AT Thu Feb 14 02:59:57 2019