**** BEGIN LOGGING AT Fri Jan 31 03:00:42 2020 Jan 31 06:33:04 *yawn* Jan 31 07:33:15 morning blogic Jan 31 07:33:25 hi Jow Jan 31 07:33:38 u80211d has taken shape Jan 31 07:34:04 little more ground work today and then add the feature detla to iwinfo Jan 31 07:34:15 and then I'll look into multi-psk and such Jan 31 07:35:33 great Jan 31 07:39:23 nl80211 is weird Jan 31 07:39:29 are there really no session ids or am i just blind Jan 31 07:39:49 looks like i need to open an extra socket for scans just so i can map results to requests Jan 31 07:40:03 or i just make scanning atomic and lock it Jan 31 07:40:16 so that you can't scan on multiple radios at a time Jan 31 07:50:09 blogic: there are session ids at least in generic netlink Jan 31 07:50:23 I think the kernel is supposed to increase it by one Jan 31 07:51:01 you can use wireshark to dissect netlink Jan 31 08:21:06 does xback's master branch contain the latest kernel 5.4 code? Jan 31 08:21:19 Hauke: yup Jan 31 08:21:26 Hauke: its our current dev master Jan 31 08:21:30 ok Jan 31 08:25:23 I'm wondering how long should the voting go Jan 31 08:25:32 wrap it up Jan 31 08:25:44 question is who of the people voting for 4.19 will do the 4.29 work Jan 31 08:25:53 i certainly wont invest time into 4.19 Jan 31 08:27:34 mirko: Noltari: [florian]: I'm planning to close the voting in the next few days, pls cast your vote soon :) http://lists.infradead.org/pipermail/openwrt-adm/2020-January/001253.html Jan 31 08:27:57 ynezz: just send a mail that you will close the vote tomorrow Jan 31 08:28:17 I think this will anyway not change much any more Jan 31 08:28:24 and over 50% votes Jan 31 08:28:52 3 potential 5.4 votes are a lot :) Jan 31 08:29:45 * ldir manages to resist mentioning brexit Jan 31 08:30:02 dammit Jan 31 08:30:27 ldir: can you please move to the channel for UK ? Jan 31 08:30:37 :D Jan 31 08:30:38 :-D Jan 31 08:30:55 and we will impose a tax of 1 euro per mail you send Jan 31 08:31:03 of course :-) Jan 31 08:31:11 and if you wanna pay in pounds there is an additional exchange fee Jan 31 08:31:12 lmfao Jan 31 08:31:14 :-D Jan 31 08:31:26 but for real, postage went up Jan 31 08:31:41 I sent a parcel to my sister for her boys bday and postage fee tripled already Jan 31 08:32:00 we can also restrict dual citizen ship, blogic ;-) Jan 31 08:32:20 and I'm wondering if ldir would be still able to access our codebase from behind of the great UK firewall Jan 31 08:33:45 ynezz: of course not... we've taken back control, apparently. Jan 31 08:34:45 I cannot tell you how angry the 51.9% of the turkeys who voted for Christmas make me. Jan 31 08:35:55 ldir: its all glitter and sunshine from now on! Jan 31 08:36:36 and it looks like the £350million per week for NHS might have come just in time for this Chinese virus. Jan 31 08:36:38 now scotland needs to become independent and join the eu Jan 31 08:37:14 jow: milk & honey mate, milk & honey! Jan 31 08:37:34 beer and kidney pie Jan 31 08:38:11 ah, the Farage diet :-) Jan 31 08:39:54 ldir hahahahaha I hate that bloke! Jan 31 08:40:40 I think we will become the 51fst state of the USA Jan 31 08:41:01 farage, boris & gove should all be shot... and not necessarily in that order. Jan 31 08:41:19 The PM will be trumps bitch! Jan 31 08:41:51 lol Jan 31 08:41:53 phwaaarrrr! You said bitch :-) Jan 31 08:44:02 Tapper: no we shall crumble into insignificance because the USofA won't touch us because of the Chinese virus... by which I mean Huawei. Jan 31 08:44:56 of course US based kit definitely doesn't have any backdoors in it. No Sir. No. Jan 31 08:46:26 haha Then we will all mutate from eating all the steroid laste meat and chlorine chicken we are Forst to eat by are overlords! Jan 31 08:46:52 anyway, let's change the topic 'cos it's kinda depressing. Jan 31 08:47:22 The future is brite my friend. Jan 31 08:47:25 Yeah OK Jan 31 08:49:25 * Tapper Stretches out his arms in a all encompassing hug for all the OpenWrt family. Jan 31 09:01:21 * ldir moves to #openwrt-uk-devel Jan 31 09:27:08 ynezz: thanks for the procd feedback. I agonised over the log message and it's still not quite right. The real problem is that the use of jails is an optional 'use if available' feature. If unavailable is that an error? a warning? a 'do I look like I care?' Jan 31 09:32:30 maybe a new procd service parameter, jail_mandatory Jan 31 09:36:47 * ldir goes to collect the dog Jan 31 10:07:52 ldir: yah, that's what I was wondering yesterday too, there's no way of the intent difference between must jail and may jail. Jan 31 10:55:51 Hauke: blogic: I received some critical fixes which were merged yesterday and this morning Jan 31 10:56:02 * ldir returns with the dog Jan 31 10:56:19 * ldir goes back to pondering jails Jan 31 10:56:26 as soon as x86 builds (and runs) with all kmods enabled. It should be in a respectable state for merging to master Jan 31 10:57:03 when that happens, I would issue a mail to devlist for RFT Jan 31 11:11:51 ynezz: can it be, that i'm not even on that mailing list? Jan 31 11:12:05 double-checked, never received anything from openwrt-adm@ Jan 31 11:18:46 xback: I am currently looking at the kernel 5.4 integartion and have some patches Jan 31 11:19:03 should I just push them to my staging branch and you integarte them? Jan 31 11:19:27 Hauke: yes please. if you could create a branch for_koen or so Jan 31 11:19:48 I'm also not found of keeping the main track in my tree .. but I only want to merge to master if the base is working fine Jan 31 11:19:52 fond* Jan 31 11:21:01 xback: ok, I will do some testing and push in about 1 hour Jan 31 11:21:12 I do not have device sto test here as I am in a train Jan 31 11:21:19 *devices to Jan 31 11:21:30 Hauke: thanks. I'll merge it to my staging after lunch here then Jan 31 11:41:46 build #64 of rb532/generic is complete: Failure [failed updatefeeds] Build details are at http://buildbot.openwrt.org/openwrt-18.06/images/builders/rb532%2Fgeneric/builds/64 Jan 31 11:44:44 mirko: ok, I've send you the email off-list, can you subscribe to openwrt-adm@ where all the voting takes place currently? Jan 31 11:48:57 ldir: procd init script helpers should know if there is ujail available, so they shouldnt send jail params in the service json to procd in the first place Jan 31 11:49:39 ldir: so as I see it, this should be handled at the origin, in the procd init scripts Jan 31 11:50:33 perhaps procd_close_instance should filter out jail stuff if the ujail binary is missing? Jan 31 11:51:08 or those procd_adD_jail* should handle that Jan 31 11:51:58 but you still need a way for them to let the init script writer say "required" for "don't care if it isn't available" right? Jan 31 12:06:00 karlp: yes agreed. Jan 31 12:07:23 am fiddling with something like that now... I got distracted by a dishwasher spare part arriving and decided to fit it... dishwasher bit me.. sharp edged sheet steel & skin, blood spurt artery etc :-) Jan 31 12:20:14 ldir criste! I hope you have stopped leaking! Are you OK? Jan 31 12:21:36 didn't really hit an artery - but I do have a nasty slice along the top of my left hand.. just below the knuckles. Jan 31 12:22:46 and bonus... the dishwasher now has a working rinse aid dispenser. Jan 31 12:25:09 Well that's something lol Jan 31 12:25:51 I have cut my hands a lot on sheet steal when building computers. Jan 31 12:26:09 They sting like a bastard! Jan 31 12:26:19 build #65 of rb532/generic is complete: Success [build successful] Build details are at http://buildbot.openwrt.org/openwrt-18.06/images/builders/rb532%2Fgeneric/builds/65 Jan 31 13:19:07 karlp: https://git.openwrt.org/?p=openwrt/staging/ldir.git;a=commit;h=53fa5398f5e7bf7767eb5029062c8d9761c99fe7 Jan 31 13:21:21 loooks reasonable Jan 31 13:22:23 changing procd_add_jail dnsmasq ubus log to procd_add_jail dnsmasq ubus log mustjail will provoke the error and stop dnsmasq from starting. Jan 31 13:23:12 and the additional 'mustjail' option can be done in the dnsmasq startup script based on a uci config. Jan 31 13:46:30 * ldir sends a patch and waits for screaming Jan 31 13:47:11 * ldir ponders 'am I really messing with procd, the mother of all processes?' Jan 31 13:59:27 ynezz: subscribed, am i allow to post? Jan 31 14:09:07 xback: see my for_koen barch Jan 31 14:09:11 *branch Jan 31 14:09:17 Hauke: noted Jan 31 14:11:30 xback: this only got very minimal testing I hope Ifixed more bugs than I introduced ;-) Jan 31 14:11:48 Hauke: It can only improve ;-) no worries Jan 31 14:16:55 ynezz: will you be at the fosdem? Jan 31 14:20:47 heh, have a case of -a --> && replacement changing the behaviour. great :) Jan 31 14:27:30 Hauke: merged. thanks Jan 31 14:29:40 karlp: oooh go on, where? Jan 31 14:31:46 nto in openwrt code, but my own, after looking at the same things. Jan 31 14:32:04 ah Jan 31 14:32:19 [ -n "$bad" -a "$bad" = "true" ] || echo "doign the thing when it's not explicitly set to true" Jan 31 14:32:32 that doesn't work if you split the -a and replace with a ] && [ construct. Jan 31 14:33:10 could probably switch to the [ "x$bad" = "xtrue" ] || echo "do the thing" maybe, Jan 31 14:33:21 I guess that's the reason for that xform you see around the place. Jan 31 14:33:46 the x-form is just code for ancient shells Jan 31 14:34:14 [ -n "$bad" -a "$bad" = "true" ] is redundant Jan 31 14:34:34 no, $bad=false or $bad=blahblahblah.... Jan 31 14:34:54 you reckon just "$bad" is enough for the -n, because it resolves to just ""? Jan 31 14:35:17 [ "$bad" = "true" ] is sufficient Jan 31 14:35:26 [ "" = "true" ] is false Jan 31 14:35:33 that's probably right, but still, [ -n "$bad" ] && [ "$bad" = "true" ] || echo "do the thing" doesn't do the right thing. Jan 31 14:35:43 yes, that is true of course Jan 31 14:45:16 A && B || C is not the same as (A && B) || C Jan 31 14:45:24 which I'm not sure make sense in shell but you get the idea Jan 31 14:47:00 yes, I know, but shellcheck highlights it as as warning. Jan 31 14:47:10 because -a is not the same as && either. Jan 31 14:47:46 and () will make a subshell which isn't helping. jow's "just don't do that silly thing" is the best course in this case at least :) Jan 31 14:49:18 ah, didn't get your point, right :) Jan 31 15:30:00 Hauke: nope, maybe next year, heading for the week off tomorrow with the family Jan 31 15:30:20 Hauke: BTW is that BattleMeshOpenWrt place/date known already? :) Jan 31 15:33:24 mirko: honestly I don't know the setup of the openwrt-adm@ mailing list or who the admin/moderator is Jan 31 15:33:46 mirko: just send your vote there and add me ynezz@true.cz in the Cc: loop Jan 31 15:34:38 mirko: I simply assume, that if you're subscribed, it should go in directly without moderator approval (but this might have changed due to the recent spam posts?) Jan 31 16:16:34 jow: for js validation, "ufloat" works, but "float" doesn't, despit being listed in validation.js. I guess it'ðs a reserved word in js or something? If I change it to "anyfloat" in validation though, it still doens't work. any ideas? Jan 31 16:17:11 define "doesn't work" Jan 31 16:17:54 jsut doesnt' do any validation? Jan 31 16:18:18 I have a o.datatype = "ufloat" and it nicely gives me red validation errors on text, but "float" just does nothing and lets me save the form Jan 31 16:20:11 yeah, its a bug Jan 31 16:20:26 helpful property of JS NaN !== NaN == true Jan 31 16:21:35 try this: http://sprunge.us/a18NcU Jan 31 16:22:29 related, o.required = true or o.optional = false doesn't seem to prevent saving either? Jan 31 16:23:04 try o.rmempty = false Jan 31 16:25:22 NaN fix is great, working, thanks. Jan 31 16:26:38 rmempty works too, was just comparing to lua model properties. Jan 31 16:26:45 GridSection is rad by the way. Jan 31 16:27:25 I've had a few things that were just DynamicList with me entering field:field:field text, now they can automatically get an editor :) Jan 31 16:30:21 I plan a treewide rmempty/optional fix soon Jan 31 16:30:47 rmempty has been misused for years as "is optional" while the actual .optional sematics were broken Jan 31 16:30:52 and that carried over into JS Jan 31 16:32:05 my own stuff has 10 .optionals, and 3 .rmemptys, no idea which ones are "correct" usages :) Jan 31 16:37:13 dump1090 only current user of "float" and dump1090, nut, ntpc, mwan all use "integer" and those are all lua. noone in js land is running into that NaN issue. Jan 31 17:37:50 ynezz: the BattleMeshOpenWrt place is known but not announced Jan 31 17:37:57 the date not yet Jan 31 19:06:47 ynezz: thank you. i do not feel like i can provide much valuable input on any of the survey points in question though. i don't have strong opinions on those (and i think i shouldn't have, considering i'm mainly passive atm) Jan 31 20:19:34 mirko: can you reply with "no opinion" please ? Jan 31 20:19:51 and now ladies and gents, its time to watch picard episode 2 Jan 31 20:19:55 been waiting all week Jan 31 20:20:11 kid just fell asleep, I wiped the floors and hung up the washing Jan 31 20:20:19 time to treat myself Jan 31 20:25:47 4.19: 12 votes, 5.4: 12 votes Jan 31 20:25:49 :D Jan 31 20:28:26 hm, thats strange as I don't see Yousong's vote counted, can't find his vote on openwrt-adm archives either Jan 31 20:33:47 yes, January 25th Jan 31 20:33:57 he is for 5.4 Jan 31 20:33:59 :) Jan 31 20:38:58 "Is being held until the list moderator can review it for approval." Jan 31 20:39:28 Message has a suspicious header Jan 31 20:39:44 http://lists.infradead.org/mailman/confirm/openwrt-adm/ebbb7684d5e84eebd20b6fddfa6174cf8f16502f Jan 31 20:40:46 So 5.4 wins then? Jan 31 20:41:29 ynezz: all the addons for thunderbird which would allow me to set custom headers are incompatible with my verison of thunderbird Jan 31 20:42:05 Is there any thing in 5.4 that is speshel for OpenWrt? Apart from being more up to date. Jan 31 20:43:29 I am reading: https://www.omgubuntu.co.uk/2019/11/linux-5-4-kernel-release-features Jan 31 20:45:40 mirko: then just simply reply to that email and don't care about the In-Reply-To: header :) Jan 31 20:46:05 mirko: openwrt-adm@ is low traffic list, so it won't lost in the noise Jan 31 20:46:16 s/lost/be lost/ Jan 31 20:48:15 mirko: thanks! Jan 31 20:48:26 This is better: https://lkml.org/lkml/2019/11/24/187 Jan 31 20:48:26 the last two 18.06.7 builds are about to finish Jan 31 20:48:46 I guess we can announce then Jan 31 20:49:33 last one is uploading now ETA ~9min Jan 31 20:49:47 build #80 of mvebu/cortexa53 is complete: Success [build successful] Build details are at http://buildbot.openwrt.org/openwrt-18.06/images/builders/mvebu%2Fcortexa53/builds/80 Jan 31 20:50:06 okay, faster than expected Jan 31 20:52:57 jow is there a link to the notes - changelog? I can post on twitter. Jan 31 20:58:37 Tapper: preparing them now, will take a while Jan 31 20:59:05 jow OK thanks. Jan 31 21:09:07 oh biyt that was amazing Jan 31 21:09:12 *boy Jan 31 21:09:47 so I decided to wath episode 3 and 4, then stall and wait for them to all be out and watch the rest in a bender with a bottle of montepulciano Jan 31 21:10:01 hope there are trekkie nerds here that feel for this Jan 31 21:10:31 #nohomo but if i were, sir patrick would be my type Jan 31 21:10:38 amazing to say the least Jan 31 21:12:12 earl grey, hot, decaf Jan 31 21:19:24 * karlp thinks it's time to queue up some watching.... Jan 31 21:39:59 karlp: it is truely amazing Jan 31 21:40:10 like best content i watched in 10 years Jan 31 21:40:22 but ehn i am a net-gen/picard fan-boy Jan 31 21:40:26 *then Jan 31 21:40:33 *next-gen Jan 31 21:40:49 android keyboards are too small Jan 31 21:44:45 I am thinking if in OpenWrt 18.06.7, which was announced there works procd as I can not start packages like transmission through procd, but manually they works. ;) Jan 31 21:51:05 jow, Hauke: I think it would be good to wait a bit after the build is done before announcing the release Jan 31 21:51:14 last time, the wiki updates (supported versions etc) had to be done in a rush *after* the release was announced Jan 31 21:55:03 tmomas has documented the process: https://openwrt.org/wiki/internal/toh_update_to_new_release Jan 31 21:55:33 (it requires shell access to the wiki though, so the number of people that can do that is limited) Jan 31 22:02:37 I just did https://openwrt.org/wiki/internal/toh_maintenance#add_new_release and I'm going to ping tmomas for the rest Jan 31 22:05:05 zorun: sorry, I wasn't paying attention to IRC, I already sent the announcements Jan 31 22:09:08 oh well Jan 31 22:10:23 after oscillating between "no release" and "rushed releases", at some point it will converge :) Jan 31 22:11:11 thanks for the release btw! Jan 31 22:12:15 there is also https://openwrt.org/advisory/start now Jan 31 22:12:22 not sure yet how to best promote it Jan 31 22:13:50 nice Jan 31 22:14:47 front page + adding a link from the release notes + from the security page Jan 31 22:15:04 this one https://openwrt.org/docs/guide-developer/security Jan 31 22:18:51 jow: it looks like https://openwrt.org/advisory/start is automatically generated? Jan 31 22:19:18 probably https://openwrt.org/docs/guide-developer/security#security_advisories should do the same Jan 31 22:21:07 just did that Jan 31 22:23:16 zorun: I think that page should be moved out of the docs/guide-developer namespace though Jan 31 22:24:32 jow: yes, or split, because the last parts look like internal documentation Jan 31 22:24:36 while the rest is communication Jan 31 22:32:43 https://hardforum.com/threads/the-openwrt-community-is-proud-to-announce-the-first-service-release-of-the-stable-openwrt-19-07-series.1992492/ Jan 31 22:33:42 Tapper: can you add a link to the release notes? Jan 31 22:33:58 this way, we can update them and people have a chance to read the updated version Jan 31 22:35:32 I don't have the link mate. Jan 31 22:35:43 I just got that from the ML Jan 31 22:36:13 https://openwrt.org/releases/19.07/notes-19.07.1 Jan 31 22:36:30 thanks Jan 31 22:37:07 I'm adding links to the advisories Jan 31 22:39:58 https://news.ycombinator.com/item?id=22206562 **** ENDING LOGGING AT Sat Feb 01 03:00:26 2020