**** BEGIN LOGGING AT Fri Jan 03 03:00:23 2020 Jan 03 04:41:13 ugh. why is libcxx failing to compile? Jan 03 08:05:31 Can someone take a look in the moderator queue of openwrt-devel? My two emails are stuck there and I think it will be faster if I re-send them. Jan 03 09:33:57 build #62 of sunxi/cortexa53 is complete: Failure [failed] Build details are at http://buildbot.openwrt.org/openwrt-18.06/images/builders/sunxi%2Fcortexa53/builds/62 blamelist: Koen Vandeputte , Henryk Heisig , Julien Rabier , Hauke Mehrtens Jan 03 09:35:37 build #62 of oxnas/ox820 is complete: Failure [failed fetchrefs] Build details are at http://buildbot.openwrt.org/openwrt-18.06/images/builders/oxnas%2Fox820/builds/62 blamelist: Koen Vandeputte , Henryk Heisig , Julien Rabier Jan 03 09:45:17 build #61 of lantiq/ase is complete: Failure [failed updatefeeds] Build details are at http://buildbot.openwrt.org/openwrt-18.06/images/builders/lantiq%2Fase/builds/61 blamelist: Koen Vandeputte , Henryk Heisig , Julien Rabier Jan 03 09:45:17 build #184 of ath79/generic is complete: Failure [failed updatefeeds] Build details are at http://buildbot.openwrt.org/master/images/builders/ath79%2Fgeneric/builds/184 blamelist: Rosen Penev , Adrian Schmutzler , David Bauer , DENG Qingfang Jan 03 11:13:10 Hello ! I'm trying to build a custom Python package (python3-netfilterqueue) with SDK (Docker: openwrtorg/sdk:ipq40xx-generic-18.06.5). Jan 03 11:13:45 Each time I launch the 'make' command, it seems that all the dependent package are always re-compiled. Jan 03 11:14:11 Is this expected ? Any way to only re-compile what has been modified and only those packages/files ? Jan 03 11:15:19 they shouldn't be recompiled, jsut run make in them each, which should jsut be checking they're uptodate Jan 03 11:17:27 Thanks karlp. I confirm that "make" or "make package/..../compile" always trigger a lot of recompiles. Trying to understand why and fixing it. Jan 03 11:20:02 More info: my package is a Python package. I created it with the dual variants (Python2 + Python3). The package name is "python-netfilterqueue". In "make menuconfig" I only selected the "Python3" variant, which is "python3-netfilterqueue". However when building I cannot use the python3 name (not recognized) Jan 03 11:20:18 so I use "make package/python-netfilterqueue/compile" Jan 03 11:30:26 maybe asdf Jan 03 11:30:50 maybe docker screwing up timestamps? Jan 03 11:30:59 maybe the dependent packages are busted Jan 03 11:31:40 you're not doing docker make thought right? you run your docker blob and then are running make multiple times there? Jan 03 11:39:57 llange: pastebin output of `scripts/diffconfig.sh` Jan 03 12:31:59 jow: in luci: 184ea6230076b you redid the dispatcher to not use node_childs anymore, is there any "compat" method so that the same theme can work on 19.07 and master? Jan 03 13:02:43 (would it be enough if the node_childs and related functions remain, even if they're not used anymore?) Jan 03 13:58:34 hi all, is there a way to downgrade the cfe of netgear r7000? Jan 03 14:06:36 ynezz: here it is: https://pastebin.com/rxbNTkiV Jan 03 14:09:00 karlp: you're right, I've a permanent session on this docker container, and run multiple times make, etc... there. My 'mypackages' folder is a shared folder between the docker guest and the host ; time is the same between those, and a "touch" in the guest (or the host) produce the proper timestamped file in the host (or the guest). Jan 03 14:16:59 is it perhaps rebuilding them to make them for each target? Jan 03 14:32:36 llange: ok, I've thought, that it's perhaps caused by CONFIG_AUTOREMOVE=y but its not set in your config Jan 03 14:34:40 you can run `make -d package/foo/compile` and see why has make decided to rebuild your package Jan 03 14:37:06 BTW have you tried to rule out the Docker's quite common overlayfs hiccups by using the --volume from your host? Jan 03 15:01:07 ynezz: Not fully knowing what should and what should'nt be rebuilt, let me share the "make -d" output with you: https://pastebin.com/1FBP2XZb Jan 03 15:01:35 Regarding docker, yes I use --volume to map drives. I don't know if it's overlayfs or not. Jan 03 16:18:19 Since I always seem to mess up submitting patches, I'd like some advice on a couple incoming patches I have (adding support for EdgeRouter4). Namely the git aspect of things. Since I'm not authorized to push, should there be a 'Signed-off-by' line in my commits? Jan 03 16:18:46 (I assume no, but I want to make sure) Jan 03 16:19:48 https://openwrt.org/submitting-patches Jan 03 16:22:37 Thanks, that should do. Jan 03 19:21:06 Hey, I am pretty new to this space and have been trying to take a stab at running a demo of openwrt on opensync found in this repo: https://github.com/plume-design/opensync-openwrt. I am wondering is anyone is familiar with that example, I am following all the instructions provided but the build ends up failing. I can go into more details about the Jan 03 19:21:07 failure case once I know this is the correct audience Jan 03 19:32:51 rmilecki: FYI https://patchwork.ozlabs.org/project/openwrt/list/?series=150989 in case it has ended in your spam folder again :) Jan 03 19:40:25 somji: you should probably try #opensync Jan 03 19:40:35 (if there's such channel) Jan 03 19:44:21 Appears to be, ill try that. Thanks Jan 03 21:43:02 so, about my mem leak: /proc/meminfo diff at https://paste.debian.net/hidden/708fbf23/. I'll try to upgrade over the weekend and then look again. Jan 03 21:49:44 build #63 of sunxi/cortexa53 is complete: Success [build successful] Build details are at http://buildbot.openwrt.org/openwrt-18.06/images/builders/sunxi%2Fcortexa53/builds/63 Jan 03 21:54:41 flip214: the difference seems to be insignificant? Jan 03 22:19:54 ynezz: thank you for your feedback and encouragement, much appreciated. And BTW, congrats with finally pinpointing that nasty EINTR bug! Unix is sometimes so confusing, I hope we'll be able to start using new languages and libraries soon that will really make the life easier without compromising performance. All of your static analysis fixes are nice showcases of how difficult it is to get Jan 03 22:20:00 anything written in C right. Jan 03 22:23:14 ynezz: what is the status of your procd changes? Jan 03 22:53:52 build #76 of ar71xx/tiny is complete: Failure [failed] Build details are at http://buildbot.openwrt.org/openwrt-19.07/images/builders/ar71xx%2Ftiny/builds/76 blamelist: Moritz Warning , David Bauer Jan 03 22:54:49 build #75 of rb532/generic is complete: Failure [failed] Build details are at http://buildbot.openwrt.org/openwrt-19.07/images/builders/rb532%2Fgeneric/builds/75 blamelist: Moritz Warning , David Bauer Jan 03 22:55:08 build #63 of oxnas/ox820 is complete: Success [build successful] Build details are at http://buildbot.openwrt.org/openwrt-18.06/images/builders/oxnas%2Fox820/builds/63 Jan 03 23:01:32 build #62 of lantiq/ase is complete: Success [build successful] Build details are at http://buildbot.openwrt.org/openwrt-18.06/images/builders/lantiq%2Fase/builds/62 Jan 03 23:11:02 Hauke: waiting for Reviewed-by: or Acked-by: :-) Jan 03 23:17:53 Hauke: https://patchwork.ozlabs.org/project/openwrt/list/?series=150989 Jan 03 23:32:43 hmm i was seeing Failed to parse JSON on sysupgrade as well. x86_64 Jan 03 23:34:58 e.g. https://paste.ee/p/ZVFG7#rHHJmTcUdZmFbiqn5Ipqf8gF3o1DzXLW Jan 03 23:35:05 * m4t rebuilds procd Jan 03 23:45:18 yeah, that error goes away. i can't test the other 3 in the series because they aren't applying cleanly to my 19.07 tree Jan 04 01:19:14 Hauke: you have any idea why the buildbots can't build libcxx? I can't reproduce locally. Maybe needs -flto=jobserver in LDFLAGS? Jan 04 01:21:15 https://downloads.openwrt.org/snapshots/faillogs/x86_64/base/libcxx/compile.txt Jan 04 01:22:37 wait a minute...I already added that **** ENDING LOGGING AT Sat Jan 04 02:59:57 2020