**** BEGIN LOGGING AT Tue Feb 18 02:59:57 2020 Feb 18 05:34:20 LetoThe2nd: It's officially sanctioned :) https://lists.yoctoproject.org/g/yocto/hashtags Feb 18 07:54:26 erbo: OMG Feb 18 07:54:55 erbo: its not even 9AM and i want to die. Feb 18 08:00:04 LetoThe2nd: You can even use lists.yoctoproject.org to post to the lists now, and then you have a dropdown menu for adding tags. Feb 18 08:00:31 is there a way to whitelist GPLv3 packages in one image recipe? I tried something like WHITELIST_GPL-3.0-pn- in local.conf but it does not work Feb 18 08:01:22 erbo: *facepalm* Feb 18 08:03:42 LetoThe2nd: It doesn't bother me that much, but I did cringe a bit when I saw "Issue while adding the support for TLS1.3 in existing krogoth yocto #yocto #yocto #yocto #yocto #apt #raspberrypi #yocto" Feb 18 08:04:04 erbo: exactly that one. Feb 18 08:06:10 it seems that i've finally eached "old grumpy neckbeard" state. Feb 18 08:27:29 Hello. I need to know if there is a specific command that let me know the compiler and its version used by the SDK toolchain in yocto project, after installing it Feb 18 08:27:56 and its path Feb 18 08:31:15 gaston53: maybe there is something hidden in the environment setup sh that the sdk brings. Feb 18 08:42:02 LetoThe2nd I found it, first you must source the environement script and then type : ${CC} --version Feb 18 08:42:12 but I don't know how to find its path Feb 18 08:42:32 gaston53: which ${CC} Feb 18 08:43:43 LetoThe2nd arm-poky-linux-gnueabi-gcc Feb 18 08:44:16 gaston53: no, thats not what i meant. the "which" command tells you the path of the binary. Feb 18 08:44:49 LetoThe2nd okey i will make a try Feb 18 08:45:21 LetoThe2nd I can find it now thank you Feb 18 08:59:35 LetoThe2nd under this path : " /opt/poky-atmel/2.5.3/sysroots " I can find 2 directories which are : " cortexa5hf-neon-poky-linux-gnueabi " and " x86_64-pokysdk-linux " Feb 18 09:00:02 what is the difference between those 2 directories and what are they made for ? Feb 18 09:01:02 gaston53: IIRC then x86 is the sysroot where all the host tools are in, and the cortex is the one that the toolchain uses to build against. but no guarantees. Feb 18 09:02:20 LetoThe2nd I see .. Feb 18 09:03:06 I can find the compiler used by the SDK in the " x86_64-pokysdk-linux " directory Feb 18 09:03:11 is that logic ? Feb 18 09:03:31 gaston53: on which computer are you compiling? Feb 18 09:03:59 my intel x86_64 Feb 18 09:04:08 gaston53: think again what i just explained, please..... Feb 18 09:04:11 and execute the code on the ARM arch Feb 18 09:04:44 gaston53: cross-compilation ;) Feb 18 09:05:47 LetoThe2nd qschulz Yes than the compiler is on my host machine ( intel ) Feb 18 09:06:18 see :) Feb 18 09:06:46 LetoThe2nd qschulz thanks Feb 18 12:16:33 is it valid to first inherit for example go and then inherit useradd? Feb 18 12:17:33 ruru4143_: what's the actual question? you can have many inherits on the same line or on different lines Feb 18 12:18:14 i want to build a go package and add users in one bb file Feb 18 12:20:31 you can Feb 18 12:21:12 I don't know if it's been fixed yet or not, but there are some issues with using a group or user in one recipe that is defined in another (just FYI, that's not exactly your question) Feb 18 12:21:44 ruru4143_: should be fine. Feb 18 12:22:42 so just inherit go, then modify do_install_append(), then inherit useradd and then modify do_install() ? Feb 18 12:23:36 ruru4143_: https://www.yoctoproject.org/docs/current/mega-manual/mega-manual.html#ref-classes-useradd Feb 18 12:23:52 hello, Is there any relationship between SDK and CMake ? maybe I can include CMake in my SDK toolchain and re-build the SDK or something like this in order to use my SDK toolchain in a CMake project Feb 18 12:24:27 thanks qschulz and LetoThe2nd Feb 18 12:26:47 gaston53: no relation by definition. you can probably add cmake into the sdk without much trouble, but i doubt that it auto magically does the full sdk dance then. Feb 18 12:28:15 If I may ask - why recipes are reparsed? (and for each parse the base hash value is checked) ? Feb 18 12:28:56 Can I add the " MAKE " build tool to my SDK ? Feb 18 12:29:02 how ? Feb 18 12:29:12 lukma: AFAIK to detect metadata that changes during the build and therefore is non deterministic in some way. Feb 18 12:29:13 LetoThe2nd ^ Feb 18 12:29:22 gaston53: have a look at https://stackoverflow.com/questions/41964891/yocto-sdk-with-cmake-toolchain-file Feb 18 12:29:40 gaston53: and at https://www.yoctoproject.org/docs/latest/mega-manual/mega-manual.html#var-TOOLCHAIN_HOST_TASK Feb 18 12:30:12 erbo LetoThe2nd thanks for the documents Feb 18 12:30:51 LetoThe2nd: Is there any way to see the exact cause of this reparsing? (bitbake ... -DDD is not very helpful) Feb 18 12:31:12 Or even better just ask OE/Yocto to don't do this? Feb 18 12:31:37 lukma: what's your issue Feb 18 12:32:05 qschultz: IMAGE_NAME·=·"update-${DATETIME}" Feb 18 12:32:15 IMAGE_LINK_NAME·=·"update" Feb 18 12:32:25 python·do_swuimage_append·()·{$ Feb 18 12:32:25 ····os.system("cd·"·+·imgdeploydir·+·";·mv·"·+·d.getVar('IMAGE_LINK_NAME',·True)·+·'.swu·'·+·d.getVar('IMAGE_LINK_NAME',·True))$ Feb 18 12:32:25 }$ Feb 18 12:32:49 the ${DATETIME} causes issues, as the TIME increases Feb 18 12:33:02 each time the recipe to build image is reparsed Feb 18 12:33:23 lukma: so your *REAL* problem is that your recipe is throwing errors now that you've updated to a newer state of poky, am i guessing right? Feb 18 12:33:24 IMAGE_NAME[vardepsexclude] = "DATETIME"? Feb 18 12:33:51 (I don't know if that is the expected way to handle this case but DATETIME throwing errors is not a new thing for sure) Feb 18 12:34:40 It seems so - List of dependencies for variable DATETIME is {'TIME', 'DATE'} Feb 18 12:36:32 I'm just wondering why time is not excluded by default from BB_HASHBASE_WHITELIST Feb 18 12:44:41 Another little question : When I add " TOOLCHAIN_HOST_TASK += "nativesdk-cmake" " to my local.conf file, should I build the all image with bitbake core-image-minimal command and then re-build the SDK , or only re-build the SDK ? Feb 18 12:45:12 gaston53: you add it to the image recipe, not the local.conf file Feb 18 12:45:32 gaston53: what erbo said. and rebuilding the sdk implictly rebuild everything from the image that it needs. Feb 18 12:45:33 And then you need to re-run "bitbake -c populate_sdk " Feb 18 12:49:45 erbo Just to be sure, the image recipe is supposed to be : poky/meta/recipes-core/images/core-image-minimal.bb Feb 18 12:50:50 gaston53: i massively suggest that you watch episode #2 of this https://www.youtube.com/playlist?list=PLD4M5FoHz-TxMfBFrDKfIS_GLY25Qsfyj Feb 18 12:50:55 gaston53: right now. Feb 18 12:52:24 LetoThe2nd (y) Feb 18 12:52:28 gaston53: whenever you interpret something as "modify this in poky", then its an absolutely certain sign that you are doing it wrong. Feb 18 12:52:31 LetoThe2nd thx for videos ! Feb 18 12:52:42 gaston53: and you are obviously missing a custom image, however basic it is.+ Feb 18 12:52:50 PinkSnake: \o/ Feb 18 13:14:51 LetoThe2nd: re:GSoC what's up? Feb 18 14:06:47 I don't know whats more upsetting; I have to keep using krogoth, or I keep misspelling it kergoth Feb 18 14:09:06 JPEW: LOL Feb 18 14:25:29 JPEW: which perl patch did you think fixed the compiler host vs target issue? Feb 18 14:26:45 JPEW: wondering if we've missed one Feb 18 14:30:47 hi. lately I've got tons of host contamination warnings from my local builds when files owned by the local user end up in sstate cache and then back to build and cause QA errors: do_package_qa: QA Issue: coreutils: /coreutils/bin/echo.coreutils is owned by uid 1005, which is the same as the user running bitbake. This may be due to host contamination Feb 18 14:31:27 this in sumo/sumo-next (+ large number of backported patches). Is there any patches in warrior, zeus or master which could help? Feb 18 14:32:47 hmm, building on Debian buster, maybe I need to update pseudo for glibc 2.28.. Feb 18 14:56:58 RP: just sent a new patch for the skipTests concurrency issue Feb 18 15:08:47 tgamblin: that looks much more like what we need, thanks! Feb 18 15:21:41 mcfrisk: I'd look at pseudo to start with, yes Feb 18 15:21:59 mcfrisk: new libc calls are a favourite way to encounter that Feb 18 15:22:43 armpit: I took my bitbake mem usage fix and merged to 1.44/1.42 in bitbake since I think its a key fix people will want Feb 18 15:22:58 * RP is curious if anyone else sees parsing mem usage improvements Feb 18 15:23:48 that fix looks awesome, hven't had a chance to test it here yet Feb 18 15:23:52 but nicely done Feb 18 15:24:48 kergoth: tracemalloc is rather neat :) Feb 18 15:25:13 kergoth: two days of head scratching until I tried that, then 10 mins and fixed Feb 18 15:30:37 JPEW: thanks for the write-up. from the SDWire website it's not obvious how to go about obtaining said item Feb 18 15:32:40 RP: yes, thanks found the patches to cherry-pick. And thanks everyone for maintaining warrior! Patches there help a lot. Feb 18 15:34:37 tlwoerner: yesterday some student asked about submitting an application @gesoc Feb 18 15:35:22 LetoThe2nd: somewhat premature. we find out this thursday which orgs are chosen, then we can worry about applications Feb 18 15:35:33 (if YP makes the list) Feb 18 15:35:33 tlwoerner: k Feb 18 15:36:21 LetoThe2nd: thanks for fielding in my absence! Feb 18 15:36:52 should YP get chosen, maybe it could be a topic for an upcoming twitch? Feb 18 15:37:37 tlwoerner: like, talking about what? Feb 18 15:39:11 i assume twitch is popular with the youngings, they're the people we're targetting for GSoC, so maybe there could be a sort of "hey, if you qualify for GSoC and are interested..." Feb 18 15:40:12 tlwoerner: it definitely is Feb 18 15:40:17 tlwoerner: and then show an up-to-date getting started? i mean i'm happy to also do an extra run for gsoc, but i don't completely understand what i would be talking about. Feb 18 15:42:00 talk about what GSoC is, who can apply, what the benefits to the project and student are, provide links, next steps, what to do Feb 18 15:42:19 tlwoerner: feed me the necessary info, then i can do that. :) Feb 18 15:42:42 tlwoerner: or.... do you have a webcam? ;-) Feb 18 15:42:47 LetoThe2nd: awesome! (we'll wait to see if we're picked) Feb 18 15:42:52 LetoThe2nd: maybe... Feb 18 15:43:25 * tlwoerner was afraid LetoThe2nd would ask that question ;-) Feb 18 15:43:31 tlwoerner: I'm having to build my own Feb 18 15:43:58 JPEW: out of scraps found in the garage?! Feb 18 15:44:02 lol Feb 18 15:44:28 tlwoerner: Well, I probably should have been more clear that I am building that piece from scratch :) Feb 18 15:44:43 i mean seriously, its not like i would do that stuff from a studio or such. i'm a dude with a laptop and a cheap headset, who happens to babble about $STUFF while drinking a beer. Feb 18 15:45:34 tlwoerner: It will be my first time soldering 402 sized components with solder paste & hot air... should be fun Feb 18 15:47:16 JPEW: did you see my question? Feb 18 15:47:21 LetoThe2nd: understood. i'm up for it Feb 18 15:47:26 :) Feb 18 15:47:39 JPEW: ah, future tense. good. when you get around to it, make me a dozen ;-) Feb 18 15:47:43 JPEW: Good luck. I recommend a good pair of tweezers with a rubber handle Feb 18 15:47:55 tlwoerner: if you know specifics, then let me know. we can certainly do a short extra one to aid gsoc Feb 18 15:48:32 LetoThe2nd: great, we'll see how the selection goes Feb 18 15:49:40 on the almost same topic. does anybody know offhand if devtool or recipetool can get started off a local src tree? or do they always need some form or upstream repo first? (will need this for tomorrows session) Feb 18 15:52:49 or, recipetool to get started, then devtool modify to basically re-route? Feb 18 15:53:39 LetoThe2nd: "devtool add" can take a path to a local source tree as an argument Feb 18 15:54:11 tlwoerner: thx Feb 18 15:55:50 LetoThe2nd: https://pastebin.com/M2NxbMMs Feb 18 15:56:24 tlwoerner: awesome, thanks. Feb 18 15:56:30 Hi. IIRC rburton gave me a hint that instead of running "bitbake -c cleansstate && bitbake ", one could do it in a single bitbake invocation.. was it "bitbake -C fetch " ? Feb 18 15:57:02 tlwoerner: i'll just do a little getting started with cmake, and then need a good line of explanation to get into a recipe. this should do the trick nicely. Feb 18 15:57:04 To clarify, that doesn't *clean* the recipe, but it *does* force it to build from scratch Feb 18 15:57:13 sometimes you'd want to clean then -C fetch Feb 18 15:57:34 kergoth: though this taints the recipe right? Feb 18 15:57:48 -C is superior to cleansstate in most cases as cleansstate will remove a local sstate package, but if you use sstate mirrors, it'll just re-fetch the sstate package and build from that again Feb 18 15:58:28 taint marks a task as needing to run from scratch, that's the point :) but yes, there's no clean way to wipe the taint without cleaning the recipe, so bitbake will keep warning you about it Feb 18 15:59:07 kroon: bitbake X -C fetch is equivalent to "bitbake X -c fetch -f; bitbake X" Feb 18 15:59:56 tlwoerner: Heh, I plan to make a post on the blog about it Feb 18 16:01:38 YPTM, armin is on Feb 18 16:02:26 YPTM, Jan-Simon is on Feb 18 16:03:12 YPTM: there soon Feb 18 16:04:59 we should setup a yp/oe "planet" for consolidating yp/oe-related blog posts :-) Feb 18 16:06:35 tlwoerner: can we ban hashtags there? Feb 18 16:06:35 rp: [OE-core] [PATCH] glibc-package.inc: fix multilib headers conflict Feb 18 16:07:02 kergoth, RP, thanks Feb 18 16:13:15 With all reproducibility work going on, are there any tools available in OE for creating reproducible ext4 images ? I'm getting differences in file system creation dates and "directory hash seed"'s Feb 18 16:16:03 LetoThe2nd: i think it works by only accepting certain tags Feb 18 16:17:00 tlwoerner: that was a joke :) Feb 18 16:17:41 tlwoerner: as i'm turning more and more into a grumpy old neckbeard who scoff at lines like #this #that #something #somethingelse Feb 18 16:19:09 LetoThe2nd: #hastags in #subject #lines are #pure #evil!!! Feb 18 16:19:45 With the rocko it was possible to build a custom image with IMAGE_CMD_foo = "bash commands" Feb 18 16:20:01 with zeus it is not possible anymore as by default "do_rootfs" task is performed Feb 18 16:20:07 vmeson: can we set up the ground rule that anybody who uses a hashtag other then #freebeer on this channel has to bring #freebeer to the next OEDEM/YPS? Feb 18 16:20:09 (or at least a bit annoying since subject lines and hashtags are functionally equivalent) Feb 18 16:20:14 is there any way to mimic the old behaviour? Feb 18 16:21:48 To say it in the other way - I would like to _just_ execute a few bash commands put into IMAGE_CMD_foo Feb 18 16:22:45 LetoThe2nd: works for me, but I may never attend OEDEM/YPS ... Feb 18 16:23:11 vmeson: do so and i will do my best to annoy you in real life! Feb 18 16:24:28 Hello, little issue during samba.do_configure : ERROR: System library pyldb-util.cpython-37m-aarch64-linux-gnu of version 1.5.6 not found, and bundling disabled, Any advice to fix that ? Seems really strange for me and totally out of my scope :S Feb 18 16:33:39 if anyone has been building and using meta-rust for a while, I'd appreciate ideas on: https://github.com/meta-rust/meta-rust/issues/264 Feb 18 16:34:16 <-- rust newbie, trying to debug the compiler. Feb 18 16:35:45 ...and the pendulum swings... Feb 18 16:38:08 YPTM: minutes at https://docs.google.com/document/d/1ly8nyhO14kDNnFcW2QskANXW3ZT7QwKC5wWVDg9dDH4 Feb 18 16:39:38 lukma: it is possible but you need to add an image type now Feb 18 16:41:52 kroon: Not yet. It's on the TODO list Feb 18 16:42:34 JPEW: can you point me at the issue you're having, where you were asking if my fix would help there? Feb 18 16:42:49 JPEW, cool Feb 18 16:44:39 kroon: If you're interested I don't think it would be too difficult to do. The main thing thats needed is the actual support for building the images reproducibly, and extending the reproducible OEQA test to make sure it stays that way Feb 18 16:45:07 kroon: It will probably be a while until I would be able to get around to it ;) Feb 18 16:45:25 JPEW, I see what you're hinting at :-) Feb 18 16:46:29 JPEW, as I see it, it might require some special tools(of which I'm not aware currently), mkfs.ext4 doesn't seem to support specifying things like directory hash or creation date Feb 18 16:47:24 kroon: Possibly. I haven't looked at it at all TBH Feb 18 16:47:52 RP: There's currently an outSideTestaddError (which is the one you originally added in 8f7352ed9c1, that I've based my patch's outSideTestaddSkip on). There's no outSideTestaddFailure yet Feb 18 16:47:56 kroon: FWIW we've had pretty good dialog with upstream on those tools Feb 18 16:48:15 kroon: I think even explaining to them what our issues are could be productive Feb 18 16:48:26 JPEW, mkfs.vfat has an "--invariant" flag, which does precisely what we want Feb 18 16:48:39 tgamblin: in the context of the code, would one make sense? Feb 18 16:49:46 tgamblin: One of problems we has was that a failed subtest skipped all the others, there were also problems with the counting of the failures and reporting of errors in the results json file Feb 18 16:49:47 RP: Depends on whether or not we would see and/or can trigger a failure, as opposed to an error, in setUpClass() or whatnot. I'm not sure how that could be done off the top of my head Feb 18 16:49:59 RP, yeah Theodore Tso is sensible Feb 18 16:50:01 RP: JPEW: ah, maybe that's the answer Feb 18 16:51:11 tgamblin: Sorry I don't have a more specific answer. I've pretty much just been carrying around http://git.yoctoproject.org/cgit.cgi/poky-contrib/commit/?h=jpew/reproducible&id=625704f96b6157d580c388dad4623d31bdd42a76 and http://git.yoctoproject.org/cgit.cgi/poky-contrib/commit/?h=jpew/reproducible&id=9bdd781b7d98da49cef0504f73d7e7006b9865bd locally to fix it Feb 18 16:51:29 tgamblin: if you look at the reproducibility test it uses subtest. It might be worth having a look at experimenting with a simpler version which just genarated a few pass/fails Feb 18 16:52:01 RP: I've disabled the do_rootfs task with d.setVarFlag("do_rootfs", "noexec", "1") Feb 18 16:52:34 RP: And - yes - I do know that this may introduce some hacks .... but sometimes one needs to combine two images Feb 18 16:52:52 and MULTIARCH is not fitting the purpose Feb 18 16:53:04 lukma: multiconfig? Feb 18 16:53:14 RP: JPEW: Hmm ok, I'll have a look Feb 18 16:53:48 given the chatter over the weekend about the server configs and branches disappearing … Feb 18 16:53:59 is anyone else seeing permission denied when trying to push to poky-contrib ? Feb 18 16:54:02 lukma: or here we're using some_image_task[depends] = "otherimage:do_build" Feb 18 16:54:16 lukma: I'm lost track of when we did it but you know we split do_rootfs to then be followed by do_image, do_image_XXXX, do_image_complete ? Feb 18 16:54:42 zeddii: no, talk to halstead Feb 18 16:54:44 kroon: https://bugzilla.yoctoproject.org/show_bug.cgi?id=13515 Feb 18 16:54:45 Bug 13515: normal, Medium, 3.1, unassigned, NEW , reproducible QA: Verify final images are binary identical Feb 18 16:54:50 zeddii: something probably got messed up Feb 18 16:55:24 yah. that’s what I guessed. I suppose I can point the AB at some other repo ? i.e. something on github ? I was going to test the 5.4 as the default kernel changes. Feb 18 16:55:38 zeddii: you can in theory, yes Feb 18 16:55:41 halstead: ping, whenever you get in. Feb 18 16:55:58 RP: I’ll just wait, I have other things to debug while I wait to launch that. Feb 18 16:57:09 FWIW I *just* pushed to poky-contrib Feb 18 16:57:32 less than 10 minutes ago Feb 18 17:00:37 zeddii: try again now? I might have access to fix it Feb 18 17:02:43 no luck. it gives me permission denied immediately. Feb 18 17:02:51 same cut & paste line that I’ve been using for years. Feb 18 17:03:08 zeddii: ok, probably a key issue. I'll let halstead work magic Feb 18 17:03:30 that was my thought as well, given how fast it bounced the push Feb 18 17:04:14 zeddii: I'm here. Feb 18 17:04:42 zeddii: at a quick look I think you need to send a new key, yours is very old Feb 18 17:06:11 kergoth: too if I'm right Feb 18 17:12:10 I want to use libsystemd in a project. I don't see anything in the systemd recipe which provides this.. any idea which package I need? Feb 18 17:19:09 fullstop: I'm fairly sure it is part of the systemd recipe even if not explicitly specified. Feb 18 17:19:37 systemd-dev: /lib/libsystemd.so Feb 18 17:19:44 seems that you are right.. Feb 18 17:22:20 fullstop: There is this in the recipe: PACKAGES_DYNAMIC += "^lib(udev|systemd|nss).*" I have to look up what PACKAGES_DYNAMIC does though. Feb 18 17:22:58 That explains why searching for libsystemd was not fruitful Feb 18 17:29:01 JPEW: I hacked up http://git.yoctoproject.org/cgit.cgi/poky-contrib/commit/?h=rpurdie/t222&id=ebc2d237254cead13d3922bd46d416476b09830b quickly Feb 18 17:29:45 PACKAGES_DYNAMIC tells the system that these packages may be created and this pattern can be used for DEPENDS/RDEPENDS Feb 18 17:46:09 RP: just email keys to halstead? Feb 18 17:46:58 kergoth: yeah. That works. Feb 18 17:47:59 k Feb 18 18:25:41 RP: Ya, thats what I was thinking Feb 18 18:25:44 Thanks Feb 18 19:20:43 khem: ping Feb 18 19:30:03 hi Feb 18 19:30:19 khem: send you a msg over in oe Feb 18 19:40:55 New news from stackoverflow: How to compile a Library with Yocto SDK toolchain? Feb 18 19:58:32 whats the trick with slirp in runqemu? Feb 18 19:59:43 LetoThe2nd: what do you mean? Feb 18 20:04:51 as Feb 18 20:08:44 tgamblin: sorry, session got messed up. Feb 18 20:09:23 i mean, if i use slirp i get qemu up and running without root privileges respectively tun/tap, which is good. but how can i use the resulting interface? Feb 18 20:12:42 hi. i know it is possible to use two git repos in one recipe in SRC_URI. how can i specify SRCREV for each of those individually? Feb 18 20:12:47 LetoThe2nd: my knowledge is limited, but I know that (provided you have something like dropbear in the image) you can scp stuff to/from the host from inside the image, e.g. scp tgamblin@10.0.2.2:/path Feb 18 20:15:41 tgamblin: nevermind. i'm an idiot Feb 18 20:16:22 tgamblin: i tried to ping, and the outright first thing once you google "qemu slirp" is basically "ICMP doesn't work, so don't try to ping" Feb 18 20:16:29 -> https://wiki.qemu.org/Documentation/Networking Feb 18 20:18:18 * LetoThe2nd calls it a day then. Feb 18 20:18:58 lol Feb 18 20:23:32 mischief: you can use "name" in SRC_URI and SRCREV, https://www.yoctoproject.org/docs/latest/mega-manual/mega-manual.html#var-SRC_URI Feb 18 20:24:37 thanks! Feb 18 20:25:16 mischief: example https://git.yoctoproject.org/cgit/cgit.cgi/poky/tree/meta/recipes-core/glibc/cross-localedef-native_2.31.bb Feb 18 20:32:27 I have a question, if a library is installed into my Yocto image , is it supposed to be automaticaly generated into the SDK which I build ? or shoud I add it in a way or another manualy ? Feb 18 20:35:26 Does anyone know if meta-ti will support DRA80xM SoC and eval boards? Or do I need to look into meta-arago, or some more closed deliveries from TI.. Feb 18 20:38:10 mcfrisk: perhaps ask on meta-ti mailing lists Feb 18 20:39:40 mcfrisk: DRA80xM == AM65xx Feb 18 20:41:44 denix: ah! that helps! I'll try setting up meta-ti and am65xx based machine then. Feb 18 20:55:08 mcfrisk: this one - http://www.ti.com/lit/ml/sprw323b/sprw323b.pdf Feb 18 20:56:32 mcfrisk: feel free to subscribe to meta-ti list for any specific questions Feb 18 21:00:01 zeddii: turning the world pink! Feb 18 21:00:05 ahaha. my old pokylinux reference was no longer updating after the weekend. Feb 18 21:00:15 RP: yah, it was failing in a way it shouldn't Feb 18 21:00:20 and I didn’t have khem’s patch for db Feb 18 21:00:33 and I finally realized it was due to my master no longer updating. Feb 18 21:00:38 I’m pushing v4 to build now :D Feb 18 21:02:03 zeddii: fair enough Feb 18 21:02:14 zeddii: its usually me doing this :) Feb 18 21:03:12 I was muttering just a few mintues ago to not understand how it wasn’t working, when it was over the weekend. Feb 18 21:03:28 different machine, very old clone, old url now. hopefully I have it now. Feb 18 21:30:43 RP: I'm still around, just not having any linux machines at work since I started on the 1st of Feb. It'll get sorted :) Feb 18 21:30:58 meanwhile I am doing yocto on a nuc :D Feb 18 21:42:41 kanavin_home: no problem, I knew you were having machin challenges! Feb 18 21:44:23 woot: https://autobuilder.yoctoproject.org/typhoon/#/builders/64/builds/1583 Feb 18 21:46:14 kanavin_home: I guess my main question would be that gstreamer reproducibility issue - do you have any time to look at that or should I? I haven't really dived into meson yet but I probably should... Feb 18 21:46:50 RP: I'll try to look into it, as I can do that on the AB, where builds are fast Feb 18 21:47:01 anything involving qemu I probably can't do Feb 18 21:47:12 zeddii: https://autobuilder.yoctoproject.org/typhoon/#/builders/45/builds/1587 though :/ Feb 18 21:47:33 kanavin_home: I was hoping that one might be an easy one to pick off since its relatively easy to see the issue Feb 18 21:47:38 (I've done and sent some patches, but it's a very unhurried experience building things on a haswell nuc) Feb 18 21:48:02 kanavin_home: I can imagine. My parsing memory fix may help there! :) Feb 18 21:49:51 RP: yah, it’s odd. I *fixed* that already. I wonder if I’ve managed to nuke a fix. Feb 18 21:49:55 * zeddii looks Feb 18 21:55:02 * RP notes green selftests in -next Feb 18 21:56:54 zeddii: btw, I fixed that source fetching error the AB would show so that shouldn't happen now Feb 18 22:06:29 clearly I’ve somehow lost that fix shuffling between all my builders, I’m going to have to redo the change. Feb 18 22:06:45 now i I can figure out what the exact config is, and start a build before I have to flee for a bit. Feb 18 22:06:46 * zeddii looks Feb 18 22:06:53 I hate when that happend Feb 18 23:50:34 JPEW: ahhrgg, https://autobuilder.yocto.io/pub/repro-fail/oe-reproducible-20200218-hlur22jp/packages/diff-html/ - ccwarnflags and ccstdflags in Configure_heavy.pl :/ Feb 19 00:05:25 kanavin_home: ^^^ Feb 19 00:05:53 not sure if anyone has insight into what cflags.SH is doing :/ Feb 19 00:05:59 * RP -> Zzzz Feb 19 02:42:13 New news from stackoverflow: How to build debug libraries for Qt in Yocto SDK? **** ENDING LOGGING AT Wed Feb 19 02:59:57 2020