**** BEGIN LOGGING AT Tue Apr 27 02:59:56 2021 Apr 27 07:25:17 good morning Apr 27 08:54:35 alicef: from Japan? Apr 27 08:59:32 mckoan: long time no see ! Apr 27 08:59:45 yes from japan Apr 27 09:59:15 i wonder why the layer check code adds the same layer repeatedly when setting up Apr 27 10:02:35 RP: there's no reason why layer-check-tool can't mine for machines on its own right. conf/machines/*.conf in the primary layer unless told otherwise seems like a good design Apr 27 10:02:54 including a failure path if it detects a machine layer but doesn't test with any machines, as that is cheating Apr 27 10:03:13 rburton: isn't there an open bug for that? :) Apr 27 10:03:20 maybe Apr 27 10:04:26 https://bugzilla.yoctoproject.org/show_bug.cgi?id=13004 Apr 27 10:04:53 rburton: right, just found that Apr 27 10:05:02 should be easy enough... Apr 27 10:05:24 rburton: I think the question was always "compare to what" Apr 27 10:05:58 our internal check-layer call does a glob to get all machines in the layer to pass to --machines Apr 27 10:06:09 compare to qemux86-64 i guess Apr 27 10:13:04 good morning, new project and it runs fine until it tries to compile enllightenment foundation libraries and fails at do_configure with  ERROR: Problem encountered: extension Xcomposite not found for X11 would be extermeley grateful for any hint Apr 27 10:14:49 sounds like missing DEPENDS Apr 27 10:18:36 rburton: it always DEPENDS! unless it RDEPENDS! Apr 27 10:23:34 the bb file comes directly from efl git Apr 27 10:24:13 well its broken then Apr 27 10:24:19 it might also be a missing DISTRO_FEATURE, maybe? Apr 27 10:26:46 https://pastebin.com/fqwvNFsq Apr 27 10:26:59 is the bb files which exists on the git Apr 27 10:27:17 thats an xserver recipe Apr 27 10:27:20 not efl Apr 27 10:27:26 you said it fails building efl Apr 27 10:27:38 gosh so sorry Apr 27 10:27:43 one sec Apr 27 10:27:47 can you point to the actual file not a pastebin Apr 27 10:28:57 https://pastebin.com/gn3j9xTM Apr 27 10:30:14 @rburton: one sec I will see if I can do that Apr 27 10:30:15 rburton: here, have a drink. Apr 27 10:31:12 * rburton panic-breathes into a paper bag Apr 27 10:31:20 the SRC_URI is in the bb file Apr 27 10:31:32 I actually wanted the url of the recipe Apr 27 10:31:38 as you said it was in efl git and i couldn't find it Apr 27 10:32:54 rburton: a nice brown paper bag with a bottle of cheapo liquor inside? Apr 27 10:35:21 I would like to rebase a misaligned patch file with 'devtool modify ', but then I get an error telling me the patch file could not be found. It seems to look in the wrong location. Apr 27 10:36:11 it is in meta-variscite which is the board am developing for thud branch Apr 27 10:36:30 The patch file comes from an append recipe which has the patch in files and has 'FILESEXTRAPATHS_prepend := "${THISDIR}/files:"', so that looks good Apr 27 10:37:07 The patch file is specified as SRC_URI_append = "file://0010-Add-skip-to-avoid-mender-auto-update-on-startup.patch;patchdir=src/${GO_IMPORT}" Apr 27 10:37:23 intera_91: still can't see a link. i'm guessing you need to add missing DEPENDS, such as libxcomposite. Apr 27 10:37:53 intera_91: note that as per https://wiki.yoctoproject.org/wiki/Releases thud is *very* end of life Apr 27 10:37:56 And it seems to look for that file within GO_IMPORT Apr 27 10:38:21 Exception: FileNotFoundError: [Errno 2] No such file or directory: '/home/redacted/repos/yoc/meta-redacted/build/tmp/work/aarch64-poky-linux/mender-client/2.5.0-r0/devtooltmp-qbx8nnck/workdir/git/src/github.com/mendersoftware/mender/patches/0010-Add-skip-to-avoid-mender-auto-update-on-startup.patch' Apr 27 10:38:23 manuel1985: don't forget to add the leading space when using _append Apr 27 10:39:21 You're right, but that just slipped in when pasting it into IRC. In the recipe the SRC_URI stretches over multiple lines and has a leading space. Apr 27 10:39:26 qschulz Apr 27 10:40:47 manuel1985: what's the value of S? Apr 27 10:41:10 ${WORKDIR}/git I assume? Apr 27 10:41:41 your patchdir is probably very wrong, but I don't understand exactly what you want to do Apr 27 10:42:00 can you give us more context? like what's the path relative to ${WORKDIR} where you patch should be applied? Apr 27 10:42:02 @rburton: libxcomposite is already in the DEPENDS Apr 27 10:42:42 qschulz: S="/home/redacted/repos/yoc/meta-redacted/build/tmp/work/aarch64-poky-linux/mender-client/2.5.0-r0/git" yep seems to be ${WORKDIR}/git Apr 27 10:43:24 So theres the recipe mender-client.bb in meta-mender and we've got an append recipe mender-client.bbappend in our own layer Apr 27 10:43:26 intera_91: maybe it actually want the xorgproto Apr 27 10:43:48 And the append recipe contains a patch file which now has a fuzz Apr 27 10:43:49 added latest release *dates* to https://wiki.yoctoproject.org/wiki/Releases to emphasis that thud has been EOL for 18 months Apr 27 10:44:06 And I want to fix that Apr 27 10:44:48 RP: well i got meta-arm-bsp down to just the kernel failure Apr 27 10:47:29 And the patch file is referenced like so: SRC_URI_append = " file://0010-Add-skip-to-avoid-mender-auto-update-on-startup.patch;patchdir=src/${GO_IMPORT}" Seems the patchdir was added because the hunks are specified relative to ${GO_IMPORT} Apr 27 10:47:46 @rburton: the meta-variscite am using was created by the conpany  and is private Apr 27 10:47:52 damn Apr 27 10:49:25 rburton: nice :) Apr 27 10:49:40 rburton: is the kernel issue going to need deeper help? :/ Apr 27 10:49:47 yes Apr 27 10:49:58 the kernel code slurps all files it sees Apr 27 10:50:07 even if they're not picked up Apr 27 10:54:12 intera_91: fwiw that recipe is horrible so i'm not surprised its broken Apr 27 10:54:16 and thud is *ancient* now Apr 27 10:54:23 so, speak to your vendor Apr 27 10:54:42 if you've a private layer then presumably you paid for it, so they should be fixing it Apr 27 10:55:28 manuel1985: can't exclude a bug in devtool patching logic. You can remove the patch from the original recipe so that devtool works fine, then apply the patch manually with git am, reorder with git rebase if needed, then git format-patch and you're good to go Apr 27 10:56:43 qschulz: Alright, will look into that! Thanks! Apr 27 10:57:47 manuel1985: it'd be very helpful if you could reproduce on master of poky or openembedded-core with a simple recipe and patch (or mender's if you can only with this one) and open a bug on our bugzilla :) Apr 27 11:01:35 qschulz: Can't promise but I'll see what I can do :) Apr 27 11:43:52 Can I make devtool not delete its tmpdir? It looks for a patch in '${WORKDIR}/devtooltmp-w6f5w_u2/ but I can't dig deeper as that dir doesn't exist Apr 27 12:03:46 manuel1985: comment out shutil.rmtree(tempdir) in scripts/lib/devtool/standard.py I guess? Apr 27 12:53:56 I had a recipe with do_compile() { make FOO=1 } and do_compile_machine2() { make FOO=2 }, and they worked "standalone", but when building both machines at the same time they could get a prebuilt package from sstate cache of the other variant. is that expected? Apr 27 12:54:39 PACKAGE_ARCH was just the default TUNE_PKGARCH. But they must have had different taskhashes, so I was surprised! Apr 27 14:11:04 in raspberry pi 3, dt-blob.dts should be dt-blob.bin, but the whole RPI_KERNEL_DEVICETREE forces you into a .dtb extension. I tried to rename it in a do_install_append and add the file to files_$PN like this - install -m 0644 ${D}/${KERNEL_IMAGEDEST}/dt-blob.dtb ${D}/${KERNEL_IMAGEDEST}/dt-blob.bin - but it still only has dt-blob.dtb in the boot partition Apr 27 14:11:57 I also tried IMAGE_BOOT_FILES += "dt-blob.bin" Apr 27 14:38:35 https://pastebin.com/W0U8YSKs obviously, the problem is in timestamp. some packets for SDK are generated with an old timestamp: img-tps-base/1.0-r1/sdk/image/opt/tps/ 3.1.2-20201228/, 3.1.2-20210427/ Apr 27 14:38:55 how can I update the timstamp for packets used in SDK without recompiling everything from scratch, Apr 27 14:46:15 yay I made it work Apr 27 16:43:30 sgw1: OEQA_DEBUGGING_SAVED_OUTPUT=/srv/autobuilder/autobuilder.yoctoproject.org/pub/repro-fail/ DISPLAY=:1 oe-selftest -r reproducible -j 1'] Apr 27 16:43:42 sgw1: https://autobuilder.yoctoproject.org/typhoon/#/builders/115/builds/206 Apr 27 16:44:41 JPEW: empty layer https://github.com/JPEWdev/meta-phosh Apr 27 16:45:03 sgw1: see also meta/lib/oeqa/selftest/cases/reproducible.py -> 153: if 'OEQA_DEBUGGING_SAVED_OUTPUT' in os.environ: Apr 27 17:02:34 is there any way to build the eSDK without having to build an image? Apr 27 17:05:29 yes, build the eSDK Apr 27 17:05:38 it won't build an image Apr 27 17:05:48 There's huge overlap between the contents of the image and the SDK though Apr 27 17:06:49 how do you "build the eSDK"? Apr 27 17:09:29 i've been looking through https://www.yoctoproject.org/docs/2.6/sdk-manual/ and don't see it Apr 27 17:10:39 rather, https://docs.yoctoproject.org/sdk-manual/index.html Apr 27 17:11:45 rburton: how do you "build the eSDK"? Apr 27 17:12:19 https://docs.yoctoproject.org/sdk-manual/appendix-obtain.html#building-an-sdk-installer Apr 27 17:13:48 thanks Apr 27 17:22:43 moto-timo: Heh, sorry Apr 27 17:36:16 is yocto 3.2 supposed to work with python 3.6.9? Apr 27 17:38:12 nm, it's a python add-on i configured (pyenv) Apr 27 17:38:30 /home/randyy/.pyenv/libexec/pyenv: line 134: /home/randyy/.pyenv/libexec/pyenv-exec: Argument list too long Apr 27 18:02:52 Is there a way to cause yocto to error if it builds a package but the files listed under FILES_${PN} are not found? Apr 27 18:03:34 The installed-vs-shipped seems to do the opposite which causes an error if there are files installed under ${D} that are not assigned to a package. Apr 27 18:03:39 rabbit9911: no, because the files are potential globs Apr 27 18:04:05 files not matching is silently accepted by design, so that the default FILES_PN can be $bindir $includedir $libdir/*.so.* etc Apr 27 18:04:29 you can add your own post-install sanity check if you want to verify that specific files exist Apr 27 18:06:04 Right now on the legacy system I have a HUGE manifest file that verifies every file is where it should be on the rootfs. It's clunky but its also nice to know that you didn't miss a file (which happens).. Sounds like this would have to be added to yocto in some form then.. Apr 27 18:08:58 simplicity with globs vs reliability. Apr 27 18:11:43 It would be nice to verify all the files that are not globs. Apr 27 19:02:40 RP: When I first saw honsiter in an error message, I thought it was some string parsing problem that chopped of the front half of some word (not sure _what_ word per se)... :) Apr 27 19:15:43 RP: I signed with linutronix Apr 27 19:15:47 rburton, ^^^ Apr 27 23:05:21 was git.yoctoproject.org recently moved? Apr 27 23:06:42 i have a clone on my hard drive and attemptng to visit the "git config --get remote.origin.url" of https://git.yoctoproject.org/git/meta-arm gives a "page cannot be found" Apr 27 23:08:11 is it now https://git.yoctoproject.org/cgit/cgit.cgi/meta-arm/tree/? Apr 27 23:10:20 yates, if we did, we would have notified the community Apr 27 23:10:40 its been cgit as long as I have been cloning Apr 27 23:13:19 armpit: that clone is 6 mnoths old at the very outside. Apr 27 23:13:38 perhaps it was also available at the other url until recently. Apr 27 23:14:18 anyway, mystery solved Apr 27 23:15:43 kanavin_: congrats! Apr 27 23:15:50 why is there no "qemu" machine here? https://git.yoctoproject.org/cgit/cgit.cgi/meta-arm/tree/meta-arm-bsp/conf/machine Apr 27 23:15:59 JPEW: I'm sure we'll get used to it :) Apr 27 23:16:25 JPEW: this name series does get interesting Apr 27 23:20:06 JPEW: and it is honister, I did just double check I had the right spelling in the layer :) Apr 27 23:20:46 * RP was brought up with some of these names so I guess they're more natural to me! Apr 27 23:36:41 yates, you might want to ping the maintainers of the layer Apr 27 23:38:59 JPEW, I keep miss pronouncing it and I expect to get in trouble for it soon Apr 28 02:27:07 moto-timo: Ok, I push my meta-phosh changes Apr 28 02:35:47 JPEW, do you have the UK version? meta-posh Apr 28 02:36:18 for the fancy engineers Apr 28 02:51:32 Lol armpit Apr 28 02:51:43 Thank you JPEW **** ENDING LOGGING AT Wed Apr 28 02:59:56 2021