**** BEGIN LOGGING AT Mon Dec 09 04:43:06 2019 Dec 09 08:51:08 hi, just want to ask posted 2 weeks ago patch to mailinglist with recipe for janus webrtc sw but didn't get any response Dec 09 10:31:25 PY2 IS NOT LONGER A HOST REQUIREMENT THIS IS NOT A DRILL PY2 IS NO LONGER A HOST REQUIREMENT Dec 09 13:17:33 rburton: what is py2? Dec 09 13:18:29 I am waiting for 1st January to be able to just pretend that I do not even know what py2 is and that there was python <3.6 Dec 09 14:03:40 JPEW: how do i wrap /../bitbake/bin/* and bind /.. on the next branch? I seem to have to alter capture.sh for the former, the latter i can't use {builddir:oeroot} as a ref in the binds section anymore Dec 09 14:10:46 kergoth: Ah, OK. Sorry about that. For the bind, PYREX_OEROOT is in the environment, so you should be able to use that. For the commands: I moved the commands to be specified by the container, but I see that is wrong now. I'll move them back to the pyrex.ini Dec 09 14:15:24 kergoth: Hmm... unless there is some env var that oe-init-build-env sets that gives the path to bitbake... Dec 09 14:18:09 There is BITBAKEDIR, but it's unset after use. Too bad Dec 09 15:04:12 hrw: what a glorious future Dec 09 15:08:02 JPEW: hmm, yes, unfortunate. you could actively check for where `bitbake` is coming from in the PATH, but that might be more magic than you'd prefer in the pyrex design.. Dec 09 15:08:33 kergoth: I have an idea... perhaps it's terrible but I'd be curious to see what you think. j/s Dec 09 15:08:42 JPEW: might want to add at least one unit test using a bitbake+oe-core nodistro layer configuration rather than a poky setup where bitbake lives in oeroot Dec 09 15:08:54 kergoth: Ya, I was thinking that Dec 09 15:08:54 no rush, just finally got around to testing earlier this morning :) Dec 09 15:11:21 JPEW: how do i access PYREX_OEROOT from pyrex.ini? ${env:PYREX_OEROOT} says it's undefined.. do i need to add it to envimport? Dec 09 15:11:43 * kergoth tries that Dec 09 15:11:54 kergoth: Yes, you do. It should probably be added by default Dec 09 15:12:18 kergoth: The joys of beta-testing ;) Dec 09 15:12:59 huh, even after adding to envimport, it still says it's not a valid option name.. that's.. odd Dec 09 15:13:11 it's certainly never boring :) Dec 09 15:16:02 kergoth: envimport is for inside the contanier. If you want to reference a variable in the config file, add it to [env] with a default value Dec 09 15:16:23 e.g. [env]\nPYREX_OEROOT="/foo/bar" Dec 09 15:16:57 Hmm... the documentation there might be off :) Dec 09 15:17:20 ah, it needs a default even if it should be set by pyrex already? i do use the 'env' sectionf or default sfor a number of vars. where the host vs container boundary is isn't always clear, and i think has changed, so.. :) Dec 09 15:17:50 if i set a default value for PYREX_OEROOT, then it just uses that when i reference it from bind, unless i hardcode the full absoulte path to my oe-core clone that's not ideal :) Dec 09 15:17:54 kergoth: No. I lied. You are correct, you should only have to add it to envimport Dec 09 15:18:52 kergoth: Ah, I think PYREX_OEROOT isn't set in your build environment? Dec 09 15:19:40 Because pyrex-init-build env unsets it :) Dec 09 15:19:52 ah, that'd explain it Dec 09 15:21:15 hm, i was going to suggest just adding /../bitbake/bin/* to the default commands, but having it in pyrex.ini is still probably better, in case the user has other bitbake-related scripts to run in container context Dec 09 15:22:51 * kergoth tweaks his setup script to export COREBASE and adds that to envimport to work around it for now :) Dec 09 15:23:17 kergoth: Works for now. Dec 09 15:28:53 hmm, adding $PYREX_OEROOT/../bitbake/bin/* to capture.sh didn't work, even after ensuring $PYREX_OEROOT/.. is in bind via a different env var.. i think i'm close to unblocked though, between those two workarounds Dec 09 15:28:56 afk, food Dec 09 15:38:12 JPEW: i'm getting the bad value subst due to undefined PYREX_OEROOT *when pyrex.py capture is run*, not when running bitbake, and pyrex.py capture is passing -a to explicitly set PYREX_OEROOT. Even removing the unset of PYREX_OEROOT in pyrex-nit-build-env doesn't fix that one :) Dec 09 15:38:33 that is, that error is before pyrex_cleanup is run Dec 09 15:40:26 (figured i'd try removing the unset instead of keeping the workaround, but reverting back now, since it wasn't *quite* so trivial :) Dec 09 15:41:16 * kergoth back to work tasks for now Dec 09 17:52:19 kergoth: OK, here's my (crazy?) idea: https://github.com/garmin/pyrex/commit/0ac3e346b848f03872b96928de8b1b84d843810a **** ENDING LOGGING AT Tue Dec 10 02:59:58 2019