**** BEGIN LOGGING AT Fri Dec 02 02:59:56 2022 Dec 02 08:12:57 1 Dec 02 08:13:26 1/join libera Dec 02 12:34:33 OEDEVM has started (34 mins ago) Dec 02 12:34:56 https://www.openembedded.org/wiki/OEDvM_2022.11 Dec 02 17:06:16 paulbarker: hi, you are the one who did some rust attempts on Yocto, right? I remember seeing a presentation of yours Dec 02 17:06:58 Jin^eLD: I know I looked at some of it a while back. Other folks have done more and better work on Rust support since I last touched it Dec 02 17:07:57 do you know whom I should rather ask? basically I am wondering how much of a PITA it would be to integrate a rust application into my rootfs on kirkstone? Dec 02 17:08:30 I googled out a rust layer, but it was not clear how good the support is and if it wasnt only meant for the master branch for the moment Dec 02 17:08:31 You can ask the room at large :) Dec 02 17:08:37 right hehe Dec 02 17:09:46 * paulbarker wonders what Randy McLeod's nick is Dec 02 17:10:46 vmeson Dec 02 17:10:53 That's the one Dec 02 17:11:08 vmeson knows more about the current state of Rust support Dec 02 17:11:13 vmeson: ping :> Dec 02 17:11:14 You can also email the Yocto list Dec 02 17:12:39 let's see if I can get a quick answer here first, so far I am trying to get an idea if the topic is worth pursuing at all Dec 02 17:13:45 although I heard that newer kernels will have some rust stuff in some way, so I guess the suffering will be inevitable at some point :) Dec 02 17:14:18 Jin^eLD: in a meeting, i'll read later... Dec 02 17:14:25 ack, thx Dec 02 18:11:40 fray: I am still fiddling with this build dependency stuff you helped me with a few days ago; I am trying to figure out - is it possible to vardepsexclude a whole task? or does it only work for variables? Dec 02 18:11:51 Jin^eLD: it used to be quite easy to use 'cargo bitbake' to generate a recipe for most, but not all, rust apps. Dec 02 18:12:28 I did that for ripgrep and fd-find when the rust merge to oe-core was happening. Dec 02 18:12:44 vmeson: actually I did not quite understand what "cargo bitbake" was, it seemed like an alternative btibake, so it is instead a .bb recipe generator? Dec 02 18:13:27 Jin^eLD: yes but it may be obsolete depending on what release/branch you are using. Dec 02 18:13:34 I am on kirkstone Dec 02 18:13:36 The new approach is: meta/classes-recipe/cargo-update-recipe-crates.bbclass Dec 02 18:13:50 let me have a look.. Dec 02 18:14:12 that is part of meta-rust? Dec 02 18:14:19 meta/classes-recipe/cargo-update-recipe-crates.bbclass is only on master. Dec 02 18:14:30 no, let me find the repo... Dec 02 18:15:02 err, kind of part of meta-rust in that it's here: https://github.com/meta-rust/cargo-bitbake Dec 02 18:15:12 but not in the meta-rust layer itself. Dec 02 18:16:53 ok you got me cofused now, so where do I finnd meta/classes-recipe/cargo-update-recipe-crates.bbclass ? the above link to meta-rust/cargo-bitbake does not seem to have that bbclass, unless I missed it Dec 02 18:17:39 for master, look into: meta/classes-recipe/cargo-update-recipe-crates.bbclass Dec 02 18:17:48 for earlier releases, use cargo bitbake Dec 02 18:17:52 eg: Dec 02 18:17:58 cd .../ripgrep.git Dec 02 18:18:29 git pull; git checkout ; cargo bitbake --> Wrote: ripgrep_13.0.0.bb Dec 02 18:18:48 ok wait.. meta/classes-recipe/cargo-update-recipe-crates.bbclass is from master, meta-rust? or from master of cargo-bitbake? Dec 02 18:19:02 Jin^eLD: and that bbclass works fine in kirkstone if you backport it Dec 02 18:19:04 or master on poky? Dec 02 18:19:10 poky/master Dec 02 18:19:11 JaMa: that would have been my next question :) Dec 02 18:19:23 oh, its on poky master, ok got it Dec 02 18:19:49 * vmeson tries the new ripgrep recipe... Dec 02 18:19:52 https://github.com/webosose/meta-webosose/commit/9bdfae7988077d0211eeee79cc339d0770cd86b4 Dec 02 18:20:24 https://github.com/webosose/meta-webosose/commit/f45bc985f7da1ac4fda8dd1425f5f9ce48869591 Dec 02 18:21:21 notice the warning about python-3.11 in kirkstone backport (I've already forgot about that) Dec 02 18:21:40 yes, exactly, just wanted to ask whats up with python 3.11 as this may be a more complicated backport Dec 02 18:21:49 but you can still use it in master and than backport the resulting recipes to kirkstone (it's one-time task to generate SRC_URI list) Dec 02 18:22:25 aha.. so 3.11 is only needed to generate the actual recipe Dec 02 18:22:32 right Dec 02 18:22:33 I could try that Dec 02 18:22:42 old dog hates new tricks :) Dec 02 18:22:55 I wish my colleagute sticked to good old C ;) Dec 02 18:23:09 *colleague Dec 02 18:23:23 Jin^eLD: have you ever spent time fixing CVEs ? ;-) Dec 02 18:23:38 vmeson: no :P Dec 02 18:24:10 If you had, then your opinion of good old C/C++ might change. Mine did. Dec 02 18:24:28 well unfortunately we're all doomed, cargo is as bad as npm and all other fancy language-specific package managers Dec 02 18:25:30 vmeson: my issue is usually lesser with the language (maybe except C++), but more with - exactly what JaMa mentioned, once they also come up with some super fancy package management it gets ugly Dec 02 18:25:34 vmeson: once you start hunting CVE issues inside those old C/C++ components statically linked to something INSIDE various containers your product might use.. Dec 02 18:26:12 anyhow.. I guess I will need to set up a poky master build so that I can generate taht one recipe :) Dec 02 18:26:30 thank you for the pointer! let's see if I manage to do it Dec 02 18:26:46 fixing tiff is realativelly easy, fixing all of the tiff libraries involved in something is on another level Dec 02 18:27:34 Jin^eLD: good luck and it's probably worth sending an email to the yocto list with a summary of what you learned and if you encountered any problems. Dec 02 18:27:48 It may be ignored but it will be in the archives for people to search. Dec 02 18:27:57 *ignored initially Dec 02 18:28:02 :) Dec 02 18:28:38 you mean not ignored but blended out as "lets hope I don't have to go through this too"? :) ack, will do Dec 02 18:29:59 Jin^eLD: thanks! Dec 02 18:30:08 thank you for the pointers! you too JaMa Dec 02 19:06:16 Jin^eLD: I fixed the license info in the recipe generated by cargo bitbake and I'm bitbaking ripgrep 13.x on master. I'll checkout the new approach after some food. Dec 02 19:07:16 In the last few days, one of my machines has been failing in an interesting / abrupt manner. start bitbake and withing a few seconds: Dec 02 19:07:24 [Errno 11] write could not complete without blockingrmacleod@ala-lpggp3:/ala-lpggp31/rmacleod/src/distro/yocto/b/rusty$ echo $? -> 0 Dec 02 19:07:42 Has anyone else seen that happen on poky/master ? Dec 02 19:08:16 vmeson: cool, curious how it plays out, it will take a while till I get to it, most likely sometime next week Dec 02 19:08:24 It's hard to reproduce but I've seen it more than 4 times in the last few days. Dec 02 19:09:00 Jin^eLD: sure, we'll be waiting for your summary! Dec 02 19:11:48 :) Dec 02 20:14:11 if I wanted to conditionally pull a repository within do_compile(), how coudl I overcome pseudo? git clone complains about ssh permissions and I guess its pseudo related since the owner of /etc/ssh was changed to nobody Dec 02 20:15:56 it works in the devshell, but it does not work when just bitbaking the recipe Dec 02 20:17:23 doh, never mind, I forgot do_compile[network] = "1" Dec 02 21:23:46 fray: the vardepseclude way on DEPENDS did not do the trick, unfortunately; I ended up with a hack that works though, I got rid of the "A" recipe and am checking it out "manually" directly in do_compile() of "B", which ultimately solves my problem **** ENDING LOGGING AT Sat Dec 03 02:59:57 2022