**** BEGIN LOGGING AT Thu Apr 01 02:59:56 2021 Apr 01 03:19:41 khem, I will try Apr 01 04:07:28 khem, yep, that did it. I will push a fix later tonight Apr 01 06:07:13 thanks armpit and sorry about that I forgot about gcc10 :) Apr 01 07:28:04 Is it possible to remove a dependency of the `do_patch` function? Apr 01 07:28:04 I have this in a recipe and would like to remove this dependency in another layer: `do_patch[depends] += "xenomai:do_patch"` Apr 01 07:28:04 Something like the following did not work: `do_patch[depends]_remove = "xenomai:do_patch"` Apr 01 07:40:23 good morning Apr 01 07:40:31 good morning Apr 01 07:55:40 yo dudX Apr 01 07:56:19 yo yo :) Apr 01 07:59:06 ho ho Apr 01 07:59:10 and a bottle of rum! Apr 01 08:07:06 sound like a plan for easter Apr 01 09:11:55 Everytime i startup, the network link is not ready. I have to set an IP, then it takes some seconds and the link becomes ready... Apr 01 10:12:49 couple of questions relating to SWAT Apr 01 10:13:49 1) for https://autobuilder.yoctoproject.org/typhoon/#builders/83/builds/2017 the quoted master-next commit seems no longer fetchable from git, making it a bit difficult to assess any link to patches that were incorporated - any suggestions? Apr 01 10:14:37 2) https://autobuilder.yoctoproject.org/typhoon/#builders/120/builds/4/steps/13/logs/errors relates to meta-agl, does this get any particular special handling? Apr 01 10:20:47 meta-agl should have or be getting a bugzilla entry Apr 01 10:21:19 but afaik its still in testing so i'm guessing RP is still kicking at it? Apr 01 11:24:23 rburton: still interested in mesa-gl? Apr 01 11:44:57 i agree that anyone touching mesa should be verifying mesa-gl still works Apr 01 11:45:03 i should throw a test case into oe-selftest Apr 01 11:53:46 rburton: thanks Apr 01 12:14:20 bluelightning: do you have any devtool patches pending? (wishful thinking mode) https://bugzilla.yoctoproject.org/show_bug.cgi?id=14333 Apr 01 12:46:59 hello guys ! Apr 01 12:47:25 Is it possible to modify the install task of a recipe based upon the image being compiled ? Apr 01 12:48:40 I have two images (A and B ), if A is being built only softwareA must be copied in ${D}, otherwise softwareB. The recipe building softwareA and softwareB is the same Apr 01 12:51:05 thekappe: as usual, yocto chant #1 applies. Apr 01 12:51:56 besides that, whats the problem with softwareA and softwareB being different packages that are being pulled in? Apr 01 12:59:52 mmmm Apr 01 13:00:05 LetoThe2nd, it sounds like arab to me :D Apr 01 13:00:49 but I'll follow your package hint Apr 01 13:00:53 thanks ! Apr 01 13:01:04 LetoThe2nd: where's the page on your website with all the Yocto chants :) ? Apr 01 13:01:23 I'm not very familiar with the system Apr 01 13:01:41 thekappe: recipe data is local, so you can't change a recipe from another (an image recipe is a recipe( Apr 01 13:04:26 thekappe: i think i did a session on package splitting. #4 or #5. essentially thats what you want. Apr 01 13:06:18 qschulz: substantial parts of my website are only existing in my head (yet), unfortunately. Apr 01 13:41:39 hi I'm dwayne Apr 01 13:42:15 How can I boot into uboot command line when running runqemu? Apr 01 13:42:34 certain key combination? Apr 01 13:51:09 I cannot locate this as a known issue with the raspberrypi bsp.  I am building in a container - does anyone recognize the pattern?  It's probably something simple- search isn't giving me any meaningful results either. Thank you for any direction you can give - I will start working with another bsp to see if I can get the same signature Apr 01 13:51:09 https://gist.github.com/rpcme/808f79afd42abdc0d3245acc7063f491 Apr 01 13:53:24 we know it is good weather, but someone still here? Apr 01 13:53:25 dwaynesohnson: qemu doesn't run through u-boot (unless you're on a very specific setup). it can go to kernel directly. Apr 01 13:54:04 LetoThe2nd seriously? '=( Apr 01 13:54:15 so it skips the boot stage and directly loads the kernel you say? Apr 01 13:54:25 dwaynesohnson: um, yes? Apr 01 13:54:40 dwaynesohnson: look at the cmdline it uses. Apr 01 13:55:07 yes, that is why I asked because I thought it was strange, the qemu limitations I guess then Apr 01 13:55:41 (y) Apr 01 13:56:06 dwaynesohnson: you can technically make it u-boot. its just not the standard, and probably will require some good tinkering. Apr 01 13:56:36 https://pandysong.github.io/blog/post/run_u-boot_in_qemu/ first google hit, essentially. Apr 01 13:58:15 cool, thanks! Apr 01 14:39:25 LetoThe2nd, qschulz, thanks for the hints Apr 01 15:23:59 So I am seeing that the base hash is changing on my recipes that use externalsrc because the value of S is changing (when I build from different directories). Is this expected? Apr 01 15:30:44 rabbit9911: the value of S is changing? Apr 01 15:32:52 Yes I use externalsrc to point to some code in my repo Apr 01 15:33:21 So I have two checkouts of the same repo and thus the absolute paths to that source is different Apr 01 15:33:59 I am expecting cache hits because the repos are exactly the same just at different locations Apr 01 15:35:23 rabbit9911: worktrees? Apr 01 15:42:14 I have a layout like    /repo/root/source  /repo/root/yocto  and /repo2/root/source /repo2/root/yocto    and a recipe in yocto that uses externalsrc with a path /repo2/root/source or /repo/root/source Apr 01 16:01:57 rabbit9911: are /repo2/root/source and /repo/root/source complete git repo or is one a git worktree of the other? Apr 01 16:04:42 rabbit9911: sounds like you might need to add S to the basehash whitelist to exclude it from checksums globally, or you could presumably arrange for externalsrc to set the vardepvalue of S to the old non-externalsrc value Apr 01 16:08:36 armpit: rustup https://www.rust-lang.org/tools/install then https://github.com/meta-rust/cargo-bitbake Apr 01 16:20:57 Is there any alternative to network-manager, if i want to configure the network via dbus? Apr 01 16:25:45 GeneralStupid: connman has a dbus interface Apr 01 16:26:14 smurray: thanks i will try that Apr 01 16:27:06 GeneralStupid: note that it has some limitations if you want to do things like multi-homing or fancier routing Apr 01 16:27:43 GeneralStupid: or at least we found that in AGL (Automotive Grade Linux) Apr 01 16:28:52 smurray: i only need to configure the IP, automatically, dynamically and have a fallback to AutoIP. I manage MDNS via avahi's dbus interfacce Apr 01 16:29:02 smurray: so i think it could be enough Apr 01 16:29:49 GeneralStupid: yeah, that sounds within it's capabilities Apr 01 16:30:22 GeneralStupid: I suspect systemd-networkd may also have a dbus interface, but I've not looked Apr 01 16:31:28 smurray: i just found some mail listing posts that they are planning a dbus interface but they dont have one ATM (but they are from 2016) Apr 01 16:31:40 I would expect some kind of documentation if they would have one Apr 01 16:31:42 GeneralStupid: heh, that suggests it didn't happen Apr 01 16:32:39 GeneralStupid: if you ignore the AGL app framework side of things, the dbus interfacing code in this might be of some use: https://git.automotivelinux.org/apps/agl-service-network/ Apr 01 16:32:41 i think so Apr 01 16:34:11 vmeson, thanks for the info. i will check it out later tonight Apr 01 16:34:40 smurray: yes, as a starting point it looks good for me thanks al ot Apr 01 16:35:18 GeneralStupid: no worries, good luck Apr 01 16:59:06 if a package expects a user-provided "foo" dir, how does it work? Should I bbappend the recipe and add the dir to SRC_URI *and* copy it myself to ${WORKDIR}? Apr 01 17:05:22 it looks like FILEEXTRAPATHS_prepend + SRC_URI_append is enough Apr 01 17:07:09 the file poky/meta/classes/kernel-arch.bbclass has a function map_kernel_arch that searches valid_archs for the passed architecture Apr 01 17:08:08 i put test code in to iterate through valid_archs and I get this: http://paste.ubuntu.com/p/7yptVTJ2cr/ Apr 01 17:09:34 the header of kernel-arch.bbclass also states that this valid_arch's are from the kernel's arch folder. i cloned the kernel and it has "csky" as an arch directory, yet i don't see it in my enumerated list Apr 01 17:09:49 is this because yocto is using an older kernel? Apr 01 17:11:03 because it's just a list. that was input at one point. new additions are done if someone cares about an arch. Apr 01 17:11:35 zeddii: yeah, well it's not just a list cause it's preventing my new architecture from being accepted. Apr 01 17:11:43 so patch it Apr 01 17:11:46 * zeddii shrugs Apr 01 17:12:26 why would the function not just dynamically create the list from the kernel's arch directories ? Apr 01 17:12:29 or override the bbclass in your layer with your own. It is literally just a list we whacked in in the past, it has changed as ppc became powerpc, and x86 unified, etc. Apr 01 17:12:46 I mean it could. but why add that complexity. The source isn't available when it runs. Apr 01 17:17:23 zeddii: doesn't the linux kernel source always have to be downloaded for any linux image? what is so complex about parsing a directory? the answer to your question "why add that complexity" is "so that yocto core can handle new/changed kernel architectures dynamically without patching .bbclass files" Apr 01 17:17:44 no, the kernel source doesn't always need to be downloaded. Apr 01 17:17:52 kernel-arch is at the very bottom of the depenency stack. Apr 01 17:18:06 and sorry, but adding new architectures is extremely uncommon. Apr 01 17:18:35 i am so lucky... Apr 01 17:19:00 zeddii: how can a linux image be built with linux source? Apr 01 17:19:04 so you can't really start a download of the kernel source, to get an architecture, for a class that is part of the kernel build. there's a cycle in that graph almost immediately. Apr 01 17:19:07 s/with/without/ Apr 01 17:19:13 uboot arch is also in that same file. Apr 01 17:19:19 kernel-dummy can be used. Apr 01 17:22:01 zeddii: by "override the bbclass in your layer with your own," did you mean my own kernel-arch.bbclass? Apr 01 17:22:46 unfortunately yes. only temporarily of course. Since you can't bbappend a bbclass. that's your option. But your class can actually include the base one, and then only override the part that it needs, etc. Apr 01 17:23:53 I'd just do that to prove the architecture, etc, and then obviously send a patch to oe-core to add it in. A new arch in that list isn't going to cause any drama on the list (it isn't like it is implying official support or anything just by being there). Apr 01 17:27:17 so if i put a "INHERIT += kernel-arch-bbclass" in my layer.conf and then provide a mylayer/classes/kernel-arch.bbclass, it will override the one in meta/classes? Apr 01 17:28:04 kernel-arch.bbclass rather Apr 01 17:28:37 I don't think so. I've only done it via require/include from my the bbclass itself. But it has been years since I've done it, so the semantics are blurry. Apr 01 17:29:54 i should probably just tias Apr 01 17:30:34 thanks zeddii Apr 01 17:39:26 that didn't work Apr 01 17:40:47 zeddii: How is Hurd support coming along? Apr 01 17:41:04 "almost ready" Apr 01 17:41:23 GNUStep will be the graphical shell of choice as well. Apr 01 17:59:25 but Fuschia is a realitty or redox OS Apr 01 18:15:37 khem: will test that mesa-gl patch later, now builder is busy building the built-time-benchmark with qemuarm (to compare with qemux86-64 times) as you mentioned in YPTM Apr 01 19:16:23 question about image-live.bbclass, is there a build dependency towards the image defined in INITRD_LIVE? Can't find anything to that effect. Or is INITRD_LIVE assumed to be already built? Apr 01 19:21:46 ^ kk spoke too soon, I see that do_bootimg depends on {initrd}:do_image_complete Apr 01 19:57:19 JaMa: I think so far it looks like my issue is crufty VM but have to validate more Apr 01 22:16:44 hey everyone. I'm really struggling to get the docker engine up and running on a reference `qemuarm64` build Apr 01 22:17:08 I keep getting met with: failed to start daemon: Error initializing network controller: Error creating default "bridge" network: Failed to Setup IP tables: Unable to enable NAT rule: (iptables failed: iptables --wait -t nat -I POSTROUTING -s 172.17.0.0/16 ! -o docker0 -j MASQUERADE: iptables v1.8.4 (legacy): Couldn't load target `MASQUERADE':No such file or directory **** ENDING LOGGING AT Fri Apr 02 02:59:57 2021