**** BEGIN LOGGING AT Wed Oct 05 02:59:58 2016 Oct 05 07:14:00 Hi. Is it possible to disable the generation of sstate artifacts? The scenario in case is builds that use SSTATE_MIRRORS (in a CI environment). These builds use sstate artifacts from SSTATE_MIRRORS, but still generate sstate artifacts for tasks they have to process (i.e., those not found in SSTATE_MIRRORS). The sstate artifacts generated by those builds are not reused -- they are just discarded. So, there is no point in generating Oct 05 07:14:00 sstate artifacts in those cases, as far as I can see. Oct 05 07:20:20 mario-goulart: I'm not sure I quite understand what you mean - why would there be no point in generating those artifacts? what are they? Oct 05 07:23:22 bluelightning: I mean the files in $SSTATE_DIR. They are not reused, as the CI system removes everything for each build. So, as far as I can see, there's no point in generating them. Oct 05 07:26:16 mario-goulart: they aren't generated though - they get downloaded from the mirror to SSTATE_DIR and then unpacked from there Oct 05 07:27:30 bluelightning: but what about the tasks that get processed, as they are not found in SSTATE_MIRRORS? Oct 05 07:28:07 why wouldn't you want to be putting those artifacts in the sstate cache? Oct 05 07:28:40 I.e. copy them to the mirror Oct 05 07:28:52 Because those builds are kind of "test builds", and we only put release artifacts into SSTATE_MIRRORS. Oct 05 07:29:51 "development builds" would be a better term to describe them. Oct 05 07:45:00 mario-goulart: ok... well there is no way to turn off generation of those artifacts, at least there is no simple switch to do so Oct 05 07:45:24 bluelightning: ok. Thanks for your comments. Oct 05 13:06:08 i am having an issue in my build Oct 05 13:06:43 i am getting "The recipe sysvinit-inittab is trying to install files into a shared area when those files already exist" Oct 05 13:07:19 Those files and their manifest location are: /home/build/oxt-setup/build/tmp-openxt_main-glibc/sysroots/xenclient-dom0/pkgdata/runtime-reverse/sysvinit-inittab Matched in manifest-xenclient-dom0-sysvinit.packagedata /home/build/oxt-setup/build/tmp-openxt_main-glibc/sysroots/xenclient-dom0/pkgdata/runtime/sysvinit-inittab.packaged Matched in manifest-xenclient-dom0-sysvinit.packagedata /home/build/oxt-setup/build/tmp-openxt_main- Oct 05 13:10:22 prefer not to just cram them into SSTATE_DUPWHITELIST, would like to understand why and fix the real issue Oct 05 13:17:01 JaMa, hi Oct 05 13:18:53 JaMa, so this time klibc built fine for qemux86...it's a race Oct 05 13:20:17 yes many of failures I usually report are some kind of race Oct 05 13:20:38 your fault having a powerful build server Oct 05 13:20:42 ;) Oct 05 13:21:28 maybe disabling parallel make or such could be a band-aid Oct 05 13:23:47 JaMa, anyway, thanks for the great work. Now the culprits are blacklisted and world build is almost clean Oct 05 15:37:14 armpit: could you ack the pypi.bbclass fix for meta-oe? it is completely broken at this moment Oct 05 15:40:14 hehe, if only I could spell my last name Oct 05 15:40:46 otavio, done Oct 05 15:41:03 hehe, any combination of "a" "c" "k" in single word will be accepted now Oct 05 15:46:18 * armpit away time to commute to sj Oct 05 15:55:34 is it expected to see python resource warnings with master? e.g. http://paste.debian.net/857598/ Oct 05 15:56:42 ntl: are you using icecc by any chance? Oct 05 15:57:10 JaMa: ah, yeah. Oct 05 15:57:30 try with https://patchwork.openembedded.org/patch/124887/ Oct 05 16:09:55 JaMa: thanks, seems to work Oct 05 18:34:39 moto-timo: back in an hour or so; did you decide what you want me to do with the evdev patch? Oct 05 19:21:39 nerdboy: haven't had time to runtime test my version. Mainly it needs the headers patch to have an upstream status and signed-off by. Then just send it to the ML with prefix [krogoth][meta-python] Oct 05 19:37:48 moto-timo: want me to use pypi instead of tarball? Oct 05 19:38:21 nerdboy: now that otavio has a patch, yes it's preferred Oct 05 19:38:45 otavio has patch for what? Oct 05 19:39:22 nerdboy: fix the fetch Oct 05 19:39:23 krogoth pypi.bbclass Oct 05 19:39:33 :) Oct 05 19:39:43 thank you again otavio :) Oct 05 19:41:27 moto-timo: is tis patch in your repo yet? Oct 05 19:41:33 *this even Oct 05 19:41:53 i did see it on the ml i think... Oct 05 19:49:26 nerdboy: my github repo is just a convenient sandbox to collaborate, I have not pushed local to it since I got the build to work Oct 05 19:50:06 nerdboy: this is on-my-own-time stuff, not my work responsibility, so I have other stuff for the release pressing Oct 05 19:54:07 uh huh Oct 05 19:54:28 * nerdboy is juggling at least 2 "full-time" projects Oct 05 19:55:17 nerdboy: are you trying to get muted? seriously dude Oct 05 19:55:35 chill, man Oct 05 19:55:40 we're all busy Oct 05 19:55:56 i just asked if you had it, if not i'll get it... Oct 05 20:00:29 i should be able to back to fixing the fix later today Oct 05 20:00:32 mtg time Oct 05 20:04:11 sounds good, thanks Oct 06 02:45:13 fray: heh, we both replied at the same time re openssl 1.1 Oct 06 02:47:31 I was trying to find out if any of the major distros have moved yet, I can't see that they have **** ENDING LOGGING AT Thu Oct 06 02:59:58 2016