**** BEGIN LOGGING AT Wed Sep 04 03:00:01 2019 Sep 04 05:19:35 Hauke: My iPhone is similar. Does not connect to WPA3 networks Sep 04 05:34:30 mangix: also with mt76 as AP? because that seems to be Hauke's issue (mt76 working, ath9k/ ath10k not working) Sep 04 05:44:40 pkgadd: ath9[10]k Sep 04 05:44:51 I have no mt76 AP Sep 04 05:45:47 ah, I thought you had one (faintly remembering gnubee, but no idea if that has wireless capabilities) Sep 04 05:46:01 that's only a NAS device Sep 04 05:46:13 not a good one either Sep 04 05:47:05 I've since upgraded to an mvebu based one. Very satisfied Sep 04 05:53:57 I cheat in that regard, 8 big disks directly in the main workstation - only available over the network if that one is running, but that's the system that most needs lots of storage. and only the important stuff that (really) needs to be accessible 24/7 in a pentium j1900 with a single 500 GB 2.5" HDD Sep 04 06:11:29 right. x86 is king when it comes to NAS Sep 04 06:20:03 UMDNS seems to be crashing in Master. Unfortunately there is not much debug info available. Sep 04 06:20:04 procd: /etc/rc.d/S80umdns: Command failed: Request timed out Sep 04 06:20:04 procd: Instance umdns::instance1 s in a crash loop 6 crashes, 0 seconds since last crash Sep 04 06:53:35 Edg9vAXYD: if you execute it on the cli, does it crash as well? Sep 04 06:54:24 jow: yes, producing /Request timed out/ Sep 04 06:55:58 jow: the exact output from cli exec /Command failed: Request timed out/ Sep 04 06:59:37 that is a ubus error message Sep 04 07:00:00 Edg9vAXYD: just call umdnsd without the init.d wrapper Sep 04 07:01:08 blogic: that is what I did -> /usr/sbin/umdns/ Sep 04 07:01:18 ok Sep 04 07:01:20 weird Sep 04 07:44:59 i'm planning some cleanups for sysupgrade + handling config backups Sep 04 07:45:00 right now a lot of code depends on /tmp/sysupgrade.tgz path (hardcoded) Sep 04 07:45:01 i'd like "sysupgrade" ubus method to pass that path as UPGRADE_BACKUP env variable Sep 04 07:45:03 then I'd like all relevant generic & targets code to use $UPGRADE_BACKUP Sep 04 07:45:04 I think it's better than assuming / hardcoding some specific path Sep 04 07:45:06 any objections to that? Sep 04 07:50:01 no, its fine Sep 04 08:04:01 the most tricky part will be about calls like: Sep 04 08:04:04 cp -af "$CONF_TAR" /mnt/ Sep 04 08:05:05 if we one day decide that $CONF_TAR (or $UPGRADE_BACKUP) can be e.g. /tmp/foobar.tgz then it will result in /mnt/foobar.tgz while some other code expects /mnt/sysupgrade.tgz Sep 04 08:05:08 nothing impossible to fix Sep 04 08:05:14 just something to be careful about Sep 04 08:06:04 well since the tar archive is appended to the filesystem eventually it has to stay on a fixed location Sep 04 08:06:31 and keep in mind that space (in flash and ram) is scarce, so if the backup is large, we might not be able to affort to copy it Sep 04 08:10:22 ok Sep 04 08:20:33 rmilecki: are you the one working on the new sysupgrade stuff? Sep 04 08:20:42 rmilecki: apparently ;) Sep 04 08:20:57 aparcar[m]: apparently ;) Sep 04 08:21:11 A did not scroll up far enough bit kinda remembered your name from -dev Sep 04 08:21:30 Cool stuff Sep 04 08:21:55 not yet ;) Sep 04 08:21:59 but soon :p Sep 04 08:23:28 jow: are the buildbot automatically on the latest git status? Meaning, will they run prepare? Sep 04 08:51:19 aparcar[m]: no, I need to redeploy them Sep 04 08:53:09 jow: sorry for the extra work Sep 04 11:12:10 hmm. that's unfortunate. it seems that uboot-envtools and uboot on mr24 disagrees about the CRC computation of the environment and consequently fight over the contents of the uboot-env (erasing/replacing). Sep 04 11:14:57 i was hoping to "factory_boot" them via the reset button to force tftp booting, load my own custom built initramfs image, setup the uboot environment from openwrt and flash, without having to tediously disassemble to get the serial console. Sep 04 11:19:30 russell--: something like this can happen if u-boot and uboot-envtools disagree about the size of the env Sep 04 11:20:31 * russell-- will experiment Sep 04 11:55:21 aparcar[m]: I think the buildinfo things are produced now Sep 04 15:01:02 I think someone once told me that there is a kernel option you can enable which causes stack traces to be written ... somewhere and survive a reboot. Have I been dreaming or does anyone know which option this is/where logs are stored? Sep 04 15:25:42 kristrev: active by default, and should in /sys/kernel/debug/crashlog if available ( https://forum.openwrt.org/t/crashlog-retrieval-mips/19731 ) Sep 04 15:50:22 KanjiMonster: Thanks! And good to know I have not gone crazy imagining things :) Sep 04 16:42:55 mangix: with which wifi chip do you seee the WPA3 probelms with the iPhone? Sep 04 16:43:12 and which OpenWrt version did you use? Sep 04 16:48:17 Hauke: When I upgraded my Pixel 3 from P to Q, I had to delete the saved WiFi network and reconnect to get it to connect. I also found that it would not connect at all if 802.11r was enabled. Additionally, my old Pixel XL will not connect at all if the network supports WPA3. I tried on both ath9k and ath10k. Sep 04 18:59:27 i need some bash help with escaping user's input Sep 04 18:59:37 can someone tell me if there is a way to make it work: https://pastebin.com/PBrL4gmt ? Sep 04 19:00:06 right now it prints every word separately Sep 04 19:17:55 rmilecki: you probably have to use eval, or arrays somehow Sep 04 19:18:06 i just found this: http://mywiki.wooledge.org/BashFAQ/050 Sep 04 19:18:13 is suggests eval or arrays... Sep 04 19:18:24 kab-el: so you're probably right Sep 04 19:18:34 (busybox has no arrays) Sep 04 19:25:54 rmilecki: busybox's awk has arrays (and hashes)... Sep 04 19:27:35 wow? Sep 04 19:43:43 nmrh: root@OpenWrt:/# arr=(Hello World); echo ${#arr[*]} Sep 04 19:43:48 /bin/ash: syntax error: unexpected "(" Sep 04 19:44:19 another one: root@OpenWrt:/# arr=(Hello World); echo ${arr[0]} Sep 04 19:44:28 /bin/ash: syntax error: unexpected "(" Sep 04 19:44:28 i don't think ash supports arrays Sep 04 19:44:38 busybox - 1.31.0-1 Sep 04 19:50:53 mamarley: when I change the SSID it also works, looks like a bug in Android, I created a bug report: https://issuetracker.google.com/issues/140508475 Sep 04 19:52:08 Hauke: Does it work then with 802.11r enabled as well, or is this just for the issue of having the existing WPA2 saved network? Sep 04 19:56:18 mamarley: 802.11r is not changed with sae Sep 04 19:56:37 e 802.11w Sep 04 19:57:04 when you select sae-mixed, 802.11w will be mandatory for sea connections and optional for PSK conenctions Sep 04 19:57:12 Well, my Pixel 3 won't connect to a WPA3 network with 802.11r enabled, but it works fine without. Sep 04 19:57:32 mamarley: I think that is a difefernt problem ;-) Sep 04 19:57:54 OK. I guess it sounded like someone above may have been having that problem. Sep 04 20:02:54 rmilecki: ash does not, try anouther tool in busybox like awk... Sep 04 20:03:14 nmrh: ah, you said "awk" i missed that Sep 04 20:03:26 nmrh: i need that for sysupgrade bash scripts Sep 04 20:03:31 so no awk there Sep 04 20:03:41 i'll just solve it differently Sep 04 20:04:10 awk should be in busybox, but yea, you'll need a different tool Sep 04 20:05:48 ash is a pita... Sep 04 20:46:49 isn't ash just a posix shell implementation? Sep 04 20:47:17 i often find a lot of 'shell' i learned is bashisms rather than SH :) Sep 04 21:18:36 Borromini: ash doesn't support arrays (it treats strings delimited with spaces similar to arrays) Sep 04 21:19:54 to work around this, I've had to call awk from ash and do a mix of strings feed to awk Sep 04 21:21:44 the op above has to rework both how the array items are encoded and the while loop that outputs them Sep 04 21:21:57 its a pita Sep 04 21:24:57 yeah i have had similar issues Sep 04 21:41:55 <_lore_> blogic: I understood why mtk-thermal is not detected on bananapi r64, mtk-fuse is compiled :) **** ENDING LOGGING AT Thu Sep 05 03:00:58 2019