**** BEGIN LOGGING AT Fri Jul 05 02:59:56 2019 Jul 05 03:23:22 New news from stackoverflow: FILES Variable in Yocto Recipe Jul 05 03:25:45 Hello, Does Poky has it own certificates trust store? How can I add certificate without re-build OS? Jul 05 06:53:47 New news from stackoverflow: How can I make a yocto layer compatible with different releases that come with incompatible lib versions Jul 05 07:50:58 In poky I used to include kernel-devsrc in my TOOLCHAIN_TARGET_TASK variable by _append and it contained every single *.c csource file. In warriror I only see Makefiles and KBuild files but canÄt find anything on the mailinglist for this issue. anyone familiar? Jul 05 09:12:25 cpo: this commit changes src symlink of kernel-devsrc maybe there is a problem with it? https://git.yoctoproject.org/cgit/cgit.cgi/poky/commit/meta/recipes-kernel/linux/kernel-devsrc.bb?id=e3b22f384932626cfa09e239a3fe2bfe043730cf Jul 05 09:24:28 cpo: this commit removes sources from kernel-devsrc https://git.yoctoproject.org/cgit/cgit.cgi/poky/commit/meta/recipes-kernel/linux/kernel-devsrc.bb?id=9af0f1a46bbb6ad9ee8b35957251f4aa826b023f Jul 05 10:01:49 nayfe: thanks will hava a look Jul 05 13:30:52 I'm trying to setup a site configuration based on the "user-configuration" tag in the yocto manual such that the site.conf sets up the base layer configuration with individual projects extending BBLAYERS as needed. Is that supported? So far I'm getting an error that it can't find my bitbake.conf file from the poky core layer. My site.conf sets BBLAYERS to contain that layer path first, and my project-level bblayers.conf Jul 05 13:30:52 file sets BBPATH to ${TOPDIR}/../shared:${TOPDIR}. I've tried that variable with a space vs. colon, which also doesn't work. Jul 05 13:31:51 I've also tried flipping the order for BBPATH to ${TOPDIR}:${TOPDIR}/../shared and that also fails. Jul 05 13:38:32 Never mind, I think I found the answer as "no" over in the bitbake multiconfiguration section, stating BBPATH isn't constructed until after layers are parsed. Jul 05 13:40:26 It would figure I would find the answer after I type all that. LOL Jul 05 13:41:52 Hi community, i have an issue about corrupted files. I have some files https://paste.linux.community/view/ec6440e5 . I apply the following script https://paste.linux.community/view/daef9bd9 and i get sometimes (1time/10) the following result https://paste.linux.community/view/d5695f66 . In fact there is some null charaters appended at the beginning Jul 05 13:41:52 of the echo string. Is there something that i missed? Thank you ! Jul 05 13:42:46 uname -a : Linux tx6s-8035 4.13.0-karo-tx6+ #3 SMP Mon May 6 17:26:02 CEST 2019 armv7l GNU/Linux Jul 05 13:44:19 tgoodwin, BBPATH = "${TOPDIR}:${HOME}/.oe" Jul 05 13:44:23 is what I ahve Jul 05 13:44:47 i precise that i'm using yocto kroght branch and this script is a systemd service. This issue happen at the fist boot Jul 05 13:45:42 Crofton: Thanks. Glad to know I had the syntax right. But I think the ultimate issue was trying to have site.conf define a base set of layers. Jul 05 13:46:32 For example I was going to have it setup the base poky ones, then at the project level append bblayers. Jul 05 13:47:02 If that multiconfig comment is right though, it can't find conf/bitbake.conf (over in the poky core layer) because it's parsing bblayers.conf's BBLAYERS first. Jul 05 13:53:18 I guess by extension then if you're trying to use multiconfig to crank off loads of builds, your bblayers config has to be a superset of all configurations whether it has anything to do with your config's machine target or not. Jul 05 13:55:49 ah yes Jul 05 13:56:18 yeah, that is why it is importnat layer control what they enable for machines Jul 05 13:56:29 and not jsut enable random machine spciic stuff globally Jul 05 13:57:25 Right. Do you happen to know if this also means an sstate-cache cannot be re-used between different layer configurations? Jul 05 14:24:56 New news from stackoverflow: FILES Variable in Yocto Recipe [duplicate] Jul 05 14:26:05 tgoodwin, I don't know for sure, but check with RP maybe Jul 05 14:39:32 Thanks. What I'm aiming for is setting up a shared state mirror internally after running our common builds (I assumed only 1 was necessary). After this experience though, it seems like I would need multiple, one for either each machine or layer configuration...I'm not sure. Jul 05 14:53:11 I'd work toward a config with all alyeers and report problems with conflicts to maintainers Jul 05 14:53:33 unfortunately, it feels like many BSP venders fell like you sould only use one bsp at a time, which is wrong Jul 05 14:56:30 yeah, and then there's the issue with xilinx branches where some boards practically require the -tools layer even with the recent addition of the standalone layer. Jul 05 14:56:49 tgoodwin: sstate should always be shareable. If that breaks, its probably a bug Jul 05 14:57:54 RP: alright, so the idea of having a build server that maintains a shared state cache that is then exposed basically read-only via https is the way it should work? Jul 05 14:58:16 shareable, but if you have multiple layer configs, aren't you likely to end up with multiuple sets of the same thing? Jul 05 14:58:19 tgoodwin: well, you could do read/write nfs too, but yes Jul 05 14:58:55 RP: I'm looking for a way to accelerate builds at other engineer's desks by leveraging a "master build" server. Jul 05 14:59:05 tgoodwin: that would work then, yes Jul 05 14:59:18 its what its designed for Jul 05 14:59:18 I was under the impression giving write access in that scenario would corrupt the cache. Jul 05 15:00:13 tgoodwin: only if they wrote something malicious Jul 05 15:01:10 RP: what I mean is having multiple builds happening at the same time at user's desks vs. the build server's shared state cache. If I made the build server's cache writeable, that would cause problems (at least that's what I got from the Yocto training class last year). Jul 05 15:01:56 tgoodwin: I don't see how it would cause problems as long as your method of sharing writes is safe Jul 05 15:02:05 tgoodwin: our autobuilder does this over NFS Jul 05 15:03:10 I see. It was about december when I took that class and it came across that having multiple builds running at the same time with a single shared state cache would somehow break it. I forget who my trainer was. Jul 05 15:03:15 It was down in Austin. Jul 05 15:03:40 But that's good to know, thanks. Jul 05 15:03:45 tgoodwin: I can't comment on that, all I can tell you is how it was designed and how we use it Jul 05 15:04:07 tgoodwin: in case its not clear I wrote large chunks of it ;-) Jul 05 15:04:22 RP: :) Jul 05 15:05:31 I suppose in the autobuilder environment you don't have to worry about a user running cleaansstate though, eh? Jul 05 15:09:29 i'm checking to create a hybrid boot for image using grub legacy + grub efi, but seems grub legacy is not supported by any wic plugin, is there any approach for grub legacy boot without having to do manual mount and grub-install? some layer or wic plugin sitting somewhere? Jul 05 15:13:56 tgoodwin: that is true, cleansstate can cause problems Jul 05 15:14:32 tgoodwin: cleansstate is nearly always avoidable with --force though Jul 05 15:14:49 I agree relying on developers to remember that is harder Jul 05 15:19:46 Thanks y'all. I'm sure I'll have more questions as I work through this. Jul 05 15:27:58 rburton_: Just found the runqueue code I'm trying to emulate isn't actually needed at all in the first place :/ Jul 05 15:50:24 ha oh well Jul 05 16:17:06 RP, two new failures on 2.7 rc1.. so we need an rc2 once those issues are fixed? Jul 05 16:17:25 armpit: which failures? Jul 05 16:17:52 armpit: one was arm ptest which has clearly never been done there? Jul 05 16:18:27 RP, from the QA report Jul 05 16:21:45 armpit: ah. mdadm is known and supposed to be getting fixed, I know not when but its never worked Jul 05 16:22:26 gstreamer is the other new one Jul 05 16:22:39 armpit: that is a single test failure in the suite, right? Jul 05 16:23:10 yes Jul 05 16:23:15 would be good to know the cause of that :/ Jul 05 16:25:26 armpit: I guess we should look at the test log and see if that helps us understand it Jul 05 16:27:50 armpit: log is useless :( Jul 05 16:28:00 says it failed Jul 05 16:28:29 the warrior-nmut full build I did prior to QA had 3 failures listed Jul 05 16:28:43 the QA report shows a 4th Jul 05 16:28:50 armpit: intermittently bad test? Jul 05 16:28:58 armpit: we try another build? Jul 05 16:32:17 let me look at the -2 warrior-nmut Jul 05 16:34:49 looks like 3 failures are the norm for previous warrior builds.. Jul 05 16:40:13 RP, I responded to the QA report. I will take the gstreamer bug Jul 05 17:14:33 * zeddii books ELCE, but skips the yocto summit part Jul 05 17:14:54 * Crofton frowns at zeddii Jul 05 17:15:03 clearly whoever booked oct 31st, does not have kids Jul 05 17:16:12 fair enough Jul 05 17:16:23 only so much we can do with timing though Jul 05 17:16:57 yah. understood. I always preferred the before timeslots, since I'm always cutting out early. Jul 05 17:17:00 I assume the Satanic rituals take precedence? Jul 05 17:17:07 yup :D Jul 05 17:17:24 plus, by the time ECLe is over, I'm baked. not going to sit and yak for two more days. Jul 05 17:17:48 this begs the question, can we do something outside an LF event? Jul 05 17:18:04 and have people show up? Jul 05 17:18:23 I know I could. but that's probably the minority, since getting funding could/would be harder. Jul 05 17:18:47 I ended up waitlisted for plumbers, so I'm trying to tie up ELCe earlier Jul 05 17:19:12 * Crofton got a ticket for Camp, so he is happy Jul 05 17:46:35 RP: If a build has an SSTATE mirror setup, does the local build create a local copy of whatever it used from the remote cache? Jul 05 21:37:18 tgoodwin: depends if its a file or not.... Jul 05 21:52:26 tgoodwin: I mean that http would have to be downloaded, locally available files wouldn't Jul 05 22:45:22 JPEW: rpurdie/t222 now has something which runs setscene and normal tasks in parallel Jul 05 22:45:44 JPEW: needs cleanup and testing as I suspect its got bugs but it nice to see it vaguely work :) **** ENDING LOGGING AT Sat Jul 06 02:59:56 2019