**** BEGIN LOGGING AT Sun Oct 06 02:59:57 2019
Oct 06 06:52:17 <mangix> swalker: I saw no real problem with it. It's nicely organized FWIW
Oct 06 08:36:44 <aparcar[m]> dhewg: hey did you do any more testing with gcc8?
Oct 06 08:37:06 <aparcar[m]> from my side I don't see what could go wrong
Oct 06 08:41:22 <dhewg> aparcar[m]: looks good to me, i run gcc8 builds on sunxi and omap by now
Oct 06 08:41:43 <dhewg> both arm though, dunno about other archs
Oct 06 08:42:11 <aparcar[m]> have you seen my v2 patch?
Oct 06 08:44:06 <dhewg> that's just with my fixup, right?
Oct 06 08:44:15 <aparcar[m]> dhewg: if you find the time please comment on the v2 that everything works. I thinks it's hard to motivate people to recompile the whole toolchain
Oct 06 08:44:44 <dhewg> sure thing
Oct 06 08:45:49 <aparcar[m]> dhewg: thank you
Oct 06 08:46:12 <aparcar[m]> jow: do you know why the json files don't end up on the server? It looks like they're generated in the buildlogs http://buildbot.openwrt.org/master/images/builders/ath79%2Fgeneric/builds/107/steps/images/logs/stdio
Oct 06 08:50:09 <ynezz> borken rsync-fu ?
Oct 06 08:59:24 <stintel> so on gentoo musl the rtl_sdr or dump1090 just gets killed right after starting
Oct 06 08:59:29 <stintel> weird stuff
Oct 06 08:59:40 <stintel> let's try again on the glibc one
Oct 06 09:04:47 <aparcar[m]> ynezz: possible, but the json files don't even appear in the checksum
Oct 06 09:16:01 <KanjiMonster> aparcar[m]: are you sure they are successfully generated? I see a lot of tracebacks/errors, assuming that's the python script generating them (e.g. "json.decoder.JSONDecodeError: Extra data: line 25 column 2 (char 590)")
Oct 06 09:17:19 <aparcar[m]> KanjiMonster: I saw that and figured some are failing (I'm looking into that also) but not all
Oct 06 09:26:03 <stintel> geez. killed on the glibc install as well
Oct 06 09:27:51 <stintel> starting to think this might be arm64 related. there's probably a reason raspbian does not use 64bit at all
Oct 06 09:28:22 <stintel> russell--: did you try brcm2708/bcm2710 or bcrc/bcm2709 on your pi3 ?
Oct 06 09:32:41 <aparcar[m]> KanjiMonster: so I count three tracebacks and 233 json_add_image_info.py calls
Oct 06 09:45:53 <KanjiMonster> aparcar[m]: yeah, got confused and thought it's a x86 build (which doesn't build that many images)
Oct 06 09:51:32 <russell--> stintel: i haven't tried openwrt on it at all, just arch (both 32 and 64-bit) and raspbian
Oct 06 09:51:43 <stintel> ah
Oct 06 09:56:15 <KanjiMonster> aparcar[m]: completely unrelated (and my python-foo is lacking), but won't https://github.com/openwrt/openwrt/commit/881ed09ee6e23f6c224184bb7493253c4624fb9f#diff-d0124e209ecca288f504fd5a3774df1fR52 cause the list of images growing with each invocation (since you load the json file if it exists, and then append), despite there only ever being one image?
Oct 06 09:57:57 <KanjiMonster> ah wait, just saw you delete the file in Device/Build
Oct 06 09:59:09 <stintel> russell--: oh well, catching a flight later today, raspbian it is, for now
Oct 06 09:59:51 * russell-- is annoyed at feeling boxed in, but ... it works
Oct 06 10:08:08 <russell--> i've been seeing 3-second switch port downage on ubnt-erx. goes down, three seconds later comes up.
Oct 06 10:08:11 <russell--> [ 9414.844294] mtk_soc_eth 1e100000.ethernet eth0: port 0 link down
Oct 06 10:08:14 <russell--> [ 9417.830470] mtk_soc_eth 1e100000.ethernet eth0: port 0 link up
Oct 06 10:16:19 <ynezz> aparcar[m]: http://buildbot.openwrt.org/master/images/builders/ath79%2Fgeneric/builds/107/steps/targetupload/logs/stdio
Oct 06 10:16:46 <ynezz> aparcar[m]: rsynclist is probably of interest?
Oct 06 10:18:59 <ynezz> aparcar[m]: https://git.openwrt.org/?p=buildbot.git;a=blob;f=scripts/sha2rsync.pl;h=de286a1e505cf159bab92d5e3acb2986bc457aea;hb=HEAD#l100 probably needs some tweaks
Oct 06 10:22:13 <aparcar[m]> ynezz: yea I'm looking into it, don't fully get the idea. So only files are transfered which existed before?
Oct 06 10:23:21 <aparcar[m]> ynezz: nevermind, wrong understanding
Oct 06 10:24:10 <ynezz> rsync --files-from=rsynclist
Oct 06 10:24:26 <aparcar[m]> so I still don't see the problem. target/install generates the json, afterwards it runs make checksums so the json files should end up in the rsynclist
Oct 06 10:27:00 <ynezz> curl -s http://downloads.openwrt.org/snapshots/targets/ath79/generic/sha256sums | grep -c json
Oct 06 10:27:03 <ynezz> 0
Oct 06 10:29:46 <aparcar[m]> 🙂 I know
Oct 06 10:30:17 <aparcar[m]> ynezz: I rebuild the master branch wit BUILDBOT=y and the files appear just fine
Oct 06 10:30:56 <aparcar[m]> but I'll look more into it
Oct 06 11:12:33 <stintel> russell--: interesting. booted raspbian via tftp/nfsroot (using my kernel but rootfs copied from the rpi4 that is happily running dump1090 for a few days), and dump1090 gets killed there too
Oct 06 11:12:45 <stintel> must be some kernel option I disabled
Oct 06 12:34:01 <danitool> hi, if I disable the onboard wifi on an ar71xx device and install an external minipciE wifi (hardware mod)
Oct 06 12:34:18 <danitool> will openwrt automagically detect it?
Oct 06 12:34:51 <danitool> I mean the ART data in the flash partition for the old wifi chip will be ignored?
Oct 06 12:36:03 <danitool> I need to do this because the radio part became unusable, with very poor signal reception
Oct 06 12:40:58 <danitool> maybe a fried RF amplifier, or who knows, my plan is to solder a new minpcie wlan card and disable the old onboard radio
Oct 06 16:43:22 <swalker> mangix: packages with SECTION:=libs or SECTION:=utils inside a net/dcwifi folder is "different", generally SECTION is the packages parent dir
Oct 07 02:11:58 <mangix> swalker: sure. there are certain packages in both libs and utils that defines both libraries and utilities though. lz4 and zstd as an example.
**** ENDING LOGGING AT Mon Oct 07 02:59:57 2019