**** BEGIN LOGGING AT Fri May 28 03:00:00 2021 May 28 07:41:39 good morning May 28 08:37:56 morning May 28 09:33:23 morning folks May 28 10:26:22 morning all May 28 12:29:11 how do I append SYSTEMD_SERVICE_${PN} when I have set it generally SYSTEMD_SERVICE_${PN} = "xxxx", and I want machine specific ones in addition? enough with SYSTEMD_SERVICE_${PN}_machine += "yyyy" ? May 28 12:45:54 SYSTEMD_SERVICE_${PN}_append_machine = " yyyy" May 28 12:46:13 thanks. I am still confused between when to use _append and when to use += May 28 12:46:21 haha after all this time May 28 12:52:00 I guess I have to mount the image and see if the symlinks for systemd were created. but is there a simpler way to see if it actually worked? May 28 13:12:15 Ad0: paraphrasing Yocto Jester's chant that says "Recipe data is local, conf data is global ....", we could simplify the matter with "Recipe operator is + =, conf data is _append. :-D May 28 13:15:09 haha May 28 13:15:30 I thought that variable was recipe data May 28 13:19:37 I mean SRC_URI is recipe right? May 28 13:19:44 yes May 28 13:19:45 that uses += May 28 13:20:04 is it possible to infer which is which? May 28 13:20:07 it's all caps lock :) May 28 13:20:21 just by the names right? May 28 13:20:59 _append is like +=, but doesn't add a space by default, can be used together with overrides and is processed later (so _append will append to default value set with ?=, while += will just overwrite the default), it doesn't matter if it's local or global variable May 28 13:21:45 when in doubt always use bitbake -e to see if the resulting value is what you expect, or you can use bitbake-getvar with latest bitbake May 28 13:32:12 Ad0: https://pretalx.com/media/yocto-project-summit-2021/submissions/WTT3UV/resources/Demystifying_the_OVERRIDES_mechan_no6J6fb.pdf ;) May 28 13:34:20 wow great thanks ! May 28 13:57:15 wondering what's wrong... SYSTEMD_PACKAGES = "${PN}" and SYSTEMD_SERVICE_${PN} = " ovo-hostname.service " seem to look properly, yet "loaded (/lib/systemd/system/ovo-hostname.service; disabled; vendor preset: disabled)". I miss something... May 28 13:58:45 ah, right, inherit systemd :-) May 28 16:07:06 question on this file: https://git.yoctoproject.org/cgit.cgi/poky/plain/meta/recipes-support/boost/boost-1.76.0.inc May 28 16:07:11 UPSTREAM_CHECK_URI = "http://www.boost.org/users/download/" May 28 16:07:14 UPSTREAM_CHECK_REGEX = "release/(?P.*)/source/" May 28 16:07:48 should I be able to access https://www.boost.org/users/download/release ? May 28 16:21:58 chrysh: maybe it should use jfrog.io as SRC_URI does after https://git.openembedded.org/openembedded-core/commit/?id=6a76da15ece9d27fca20ace12db4978092e042b7 ? May 28 16:23:42 Fetching sources from my private github repo, I still get "Host key verification failed." May 28 16:24:23 but while running the git ls-remote command manually, it prompts me for "Are you sure you want to continue connecting (yes/no)?", and when typing yes, everthing works fine May 28 16:24:34 can this prompt break the bitbake fetcher? May 28 16:27:16 bitbake is non-interactive so I'd say yes, it's safe to assume it breaks it May 28 16:28:50 but what causes this? shouldn't bitbake do something about it? like setting -o UserKnownHostsFile=/dev/null or something? May 28 16:32:43 JaMa: Is that a bug? Should I report it to somebody? May 28 16:36:33 chrysh: or send a patch :) May 28 16:40:13 chrysh: you can test by reverting back to 1.75.0 and run auh (I think that's what uses those variables?) May 28 16:43:08 qschulz: like shouldn't the bitbake fetcher use at least StrictHostKeyChecking=accept-new for instance? May 28 16:50:30 v2d: that would probably be a reasonable enhancement, perhaps open a bug in the yocto bugzilla? May 28 19:01:24 chrysh: that downloads /users/download/ and looks for href elements that contain release/.*/source/ May 28 19:13:45 rburton: so not a bug? May 28 20:31:27 are the yocto summit 2021 videos available somewhere? May 28 20:44:55 pretty sure that's a question for ndec. I suspect the answer is not yet, as I've not seen an email to the ml May 28 20:45:51 Correct , not yet, but very soon ;) May 28 20:47:32 ok cool May 28 20:48:05 btw, great conference. i was very impressed by the quality May 28 20:48:28 some of you actually seemed to know what you were doing May 28 20:57:54 dl9pf: are/were you at Ericsson? May 28 21:08:30 no May 28 21:38:56 Any opinions on how much RAM is required to build Hardknott's nodejs (v12.0)? We threw 10 gigs at it and it still died. nodejs-native seems to build just fine. May 28 21:48:15 Ch^W: try limiting PARALLEL_MAKE May 28 21:48:23 it needs about 4GB/thread :( May 28 21:48:29 Yikes May 28 21:48:54 RP: Thanks, I will give that a go. May 28 21:48:55 we're in the process of deciding how to constrain it on our automated testing :( May 28 21:49:28 Speaking only for myself and the build I am responsible for, I am more than happy to trade build time for a build that does not break :) May 28 21:52:20 Ch^W: seems crazy to need that much to me :( May 28 21:53:31 Yeah, me too. I pinged it off of a professor friend (wrote one of the original Pascal compilers) and he has a hard time believing any abstract syntax tree is actually that large. May 28 21:53:56 "It is clearly doing something else..." **** ENDING LOGGING AT Sat May 29 03:01:08 2021