**** BEGIN LOGGING AT Tue Jul 02 02:59:56 2019 Jul 02 04:01:48 ynezz: I agree with this. Just so I can mer-I mean help out with maintenance :) Jul 02 04:02:36 It seems buildbots don't like the new Lua version Jul 02 04:57:12 blogic: that src-include stuff does not appear to protect against recursions Jul 02 05:40:17 jow: ok Jul 02 05:40:35 blogic: nvm, noticed that its not your patch Jul 02 05:54:33 blogic: actually nevermind, that duplicate name check will prevent recursions, so all fine Jul 02 06:55:23 jow: yep, was about to say the same Jul 02 08:06:52 ynezz: Thanks! :-) Jul 02 08:07:55 ynezz: It would be really great if you can send it via mailing list as a summary from the Hamburg meeting. I'm not in any position to decide what's best or wrong as this is OpenWrt decision. If you say that I should move it there, I don't have any problem with that. Jul 02 08:22:32 Pepe: yeah, almost month has passed since that Hamburg meeting, so it would be really nice to get it out :) Jul 02 08:22:57 people are busy, holidays... Jul 02 08:43:34 Question - if going to drop the out of tree CAKE build and use the in-kernel version, should we make CAKE part of default traffic schedulers package or keep it as an optional module. Jul 02 08:48:29 I'm beginning to favour the latter because of packages having a 'kmod-sched-cake' dependency Jul 02 08:55:52 and it doesn't have a 'default size' implication Jul 02 08:56:07 comments/thoughts? Jul 02 09:12:11 naive question of the day - what is difference between autoload & autoprobe and when should I use load over probe? Jul 02 09:22:34 autoload will always load the module, autoprobe only if the hardware was detected, I think Jul 02 09:24:00 for things like bme680, where you have to manually assign the driver to an i2c bus / address, you can't use autoprobe Jul 02 09:24:05 and for a module that isn't dependent upon hardware ? Jul 02 09:24:39 you mean for your conntrack thing :) Jul 02 09:25:01 I'm not sure we support loading kernel module when userspace tries to use the feature it offers Jul 02 09:25:10 better wait for an adult ;) Jul 02 09:25:29 hey that's my excuse :') Jul 02 09:25:58 😂 Jul 02 09:26:07 stintel: is that a feature in the kernel or what do you mean ? Jul 02 09:26:11 use autoprobe Jul 02 09:26:22 jow: yeah i was going to sugget exactly that Jul 02 09:26:35 adults are like buses - none around, then two all at once Jul 02 09:26:44 making autoprobe work in kmodloader was a pita so might aswell make use of it Jul 02 09:26:59 ldir: where ? Jul 02 09:27:09 i just consulted the bus schedule Jul 02 09:27:09 I dunno how this works really Jul 02 09:27:27 but for bme680 autoprobe no worky so autoload it will be Jul 02 09:27:29 the bus schedule? its quite simple actually Jul 02 09:27:37 imho diffserv should never be present and we should always use cake Jul 02 09:27:38 my advice is don't go to Berlin airport Jul 02 09:27:46 ldir: dont go to berlin Jul 02 09:27:56 my advise is: don't use buses? Jul 02 09:28:11 stintel: says the petrolhead Jul 02 09:28:38 blogic: i'm trying to follow sysupgrade which seems quite big by now Jul 02 09:28:40 blogic: i see that /sbin/sysupgrade sets COMMAND to a const value, then.... Jul 02 09:28:41 1) it gets passed to the ubus system sysupgrade Jul 02 09:28:42 2) it gets passed to the /sbin/upgraded Jul 02 09:28:44 3) it gets passed to the /lib/upgrade/stage2 Jul 02 09:28:46 unrelated. try public transport in Belgium for a month and you'll know why :) Jul 02 09:28:48 do we need that? Jul 02 09:29:00 rmilecki: sort of yes Jul 02 09:29:05 that is the staged sysupgrade Jul 02 09:29:10 blogic: could COMMAND be just put in /lib/upgrade/stage2 directly? Jul 02 09:29:11 originally introduced for nand flash Jul 02 09:29:17 then we had 2 code paths Jul 02 09:29:36 ermmm Jul 02 09:30:38 rmilecki: its to allow targets to override the process Jul 02 09:31:18 so no, lets keep it in the sysupgrade call Jul 02 09:31:23 i get nothing for grep -R "COMMAND" target/linux/*/base-files/* Jul 02 09:31:33 yeah i just noticed Jul 02 09:32:07 i'm wondering if all of that can be simplified... Jul 02 09:32:16 rmilecki: a5428244d969fda2b32e24b7cda589c2c65cbf7c Jul 02 09:32:17 stateless sysupgrade causes some problems Jul 02 09:32:28 i am assuming that its either for code style or if freifunk uses it Jul 02 09:32:42 i must admit its become a behemoth Jul 02 09:33:34 blogic: i don't like how options are passed between /sbin/sysupgrade and /lib/upgrade/stage2, e.g. SAVE_PARTITIONS being passed as touch/rm /tmp/sysupgrade.always.overwrite.bootdisk.partmap Jul 02 09:34:35 i understand variables can't be shared but that is qute hacky to guess earlier user choices depending on existance of some files Jul 02 09:35:23 rmilecki: its organically grown mothball code Jul 02 09:35:37 I see no reason not to pass it as env variables Jul 02 09:35:52 blogic: i was thinking about soltutions... i see two I think Jul 02 09:35:57 1) do some "export -p > /tmp/sysupgrade.state" in /sbin/sysupgrade and then ". /tmp/sysupgrade.state" in /lib/upgrade/stage2 Jul 02 09:35:59 2) rewrite most of sysupgrade to some deamon (put it in procd?) Jul 02 09:36:07 could be sent as dict along with the ubus system sysupgrade call Jul 02 09:36:19 which thne translates it to clearenv() plus a series of setenv() calls Jul 02 09:37:17 jow: that would be probably even better than "export -p > /tmp/sysupgrade.state" Jul 02 09:38:19 while being at it I would also rename the system/upgrade ubus call to system/reexec or similar Jul 02 09:38:35 because that is what it actually does, make procd exec() itself into a user supplied ocmmand Jul 02 09:38:56 which - in case of sysupgrade - happens to be upgraded which eventually calls stage2 Jul 02 09:39:02 but it is useful for other tasks as well Jul 02 09:39:12 i see Jul 02 09:39:14 pivot_root trickery, initramfs debugging etc. Jul 02 09:39:47 what about moving more of /sbin/sysupgrade logic to the ubus method? Jul 02 09:40:04 i have a feeling it's getting too much for a shell script Jul 02 09:40:21 I would keep the ubus method lean Jul 02 09:40:40 other tasks in the shell script, like the process kill loop could be moved to separate ubus procedures Jul 02 09:41:03 basically we need an "init 1" in procd Jul 02 09:41:05 jow: what does "lean" mean? as in "simple"? Jul 02 09:41:31 rmilecki: slim, simple ... Jul 02 09:41:34 yeah. Our sysupgrade process is not yet formalized enough to be put in C code imho Jul 02 09:41:35 tanks Jul 02 09:41:54 also we'd lose the ability to upgrade sysupgrade itself before installing updates Jul 02 09:42:10 jow: correct Jul 02 09:42:18 i'm still not sure how much I'd like to move to C Jul 02 09:42:33 does anyone does that? Jul 02 09:42:42 upgrading sysupgrade? Jul 02 09:42:55 rmilecki: at least I did on several occasions, during development Jul 02 09:42:59 ok Jul 02 09:43:42 what would make sense to be moved to procd: Jul 02 09:44:06 - an "shutdown everything except pid 1" mode Jul 02 09:44:44 or at least "terminate all services" (similar to loop calling /etc/init.d/x stop an all init scripts) Jul 02 09:45:02 + shutdown wifi, unmount mounts, unload modules, flush caches... Jul 02 09:45:22 right, these make sense for sysupgrade in general, but not sure how much of that belongs into procd Jul 02 09:45:39 umount mounts should be handled by /etc/init.d/fstab stop Jul 02 10:06:55 heiho ... I'm thinking about implementing some kind of fair queung for a small network (25 - 50) clients having Internet-Access (IPv4) using a OpenWRT-Router. I'm not that experienced about scheduling. Naivly thinking, I'd like to create a bunch queues and schedule data based on the senders IP-Adress. I'd great to have one queue per sender Adress (network can be /24 or /25) and fair scheduling. Are you aware of a configuration / script for doing so? Jul 02 10:06:55 Most configurations I found do scheduling on a per-flow basis, honoring users with many, TCP-connections (e.g. bittorrent). I'd like to avoid that. Jul 02 10:08:17 Hi yanosz you mite beable to do that with Gargoyle Jul 02 10:08:30 thx - I'll check sec Jul 02 10:08:41 https://www.gargoyle-router.com/index.php Jul 02 10:09:32 It's built ontop of OpenWrt Jul 02 10:10:27 It's QOS is all about one queue per sender Adress Jul 02 10:11:20 ah - sounds helpful Jul 02 10:11:39 I'm checking the docs atm... looks exhaustive. Thanks Jul 02 10:13:02 Hmm.. https://www.gargoyle-router.com/wiki/doku.php?id=qos doesn't say anything about IPv4 vs. IPv6. Jul 02 10:13:15 Do you know, if this is a dualstack setup? Jul 02 10:28:10 Tapper: aha! Just the person. Jul 02 10:29:22 Tapper: your device has arrived, unlike my wire/connectors. But an initial opening inspection reveals you haven't broken a wifi antenna lead. just the connector disconnected. Jul 02 10:30:14 have tightened all the heatsink screws and re-fitted the board to the base assembly. Jul 02 10:30:53 once the connectors arrive I'll make a serial lead up and start talking to it over the serial port and find out what's going on. Jul 02 10:31:06 so far, all is positive. Jul 02 10:31:30 ldir: Can have one with you this week if you want to avoid waiting for China Jul 02 10:32:35 let me check the dispatch email 'cos it has been dispatched.... not sure from where :-) Jul 02 10:33:40 Somewhere in Guangdong province, I'll bet ;) Jul 02 10:35:40 I think you could be right :-). Ermm, yes, may I take you up on the offer of a UK based delivery :-) Jul 02 10:36:59 Send me your address and I'll make one up this afternoon Jul 02 10:37:20 Let me know what you want on the other end Jul 02 10:41:52 Well if ultra cheeky a 3.5mm inline jack socket - but more realistically standard .1" pitch header pins would be equally fab Jul 02 10:42:16 Sorry, no TRS sockets on hand today :) Jul 02 10:42:30 Especially not those evil 3.5mm inventions Jul 02 10:42:39 1/4" or bust Jul 02 10:43:38 lol - tinned bare ends would be just fine Jul 02 10:43:55 or even nothing... I can do the stripping & tinning :-) Jul 02 10:45:15 Male or female pins? :) Jul 02 10:48:22 jst end to go on https://openwrt.org/_detail/media/linksys/wrt3200acm/wrt3200acm_serial_header.jpg?id=toh%3Alinksys%3Alinksys_wrt3200acm other end...not fussed Jul 02 10:48:40 if honest confused by 'male or female pins' Jul 02 10:48:59 male pins are pins, female pins are sockets surely. Jul 02 10:50:02 Well yes, but they're crimps and the terminology can get awkward Jul 02 10:50:15 http://eidupont.scene7.com/is/image/eidupont/DPP_IMG_142_NonHalConnectors_hero_690x345_465x232?$Full-Width-M$ or https://www.bc-robotics.com/wp-content/uploads/2018/12/3p-dupont.jpg Jul 02 10:52:53 ah, ok - male pins. Then I can directly attach the female sockets on my serial cable as required :-) Jul 02 10:52:57 what's wrong with 3.5mm jacks? Jul 02 10:53:32 They're small and fiddly and flimsy :) Jul 02 10:53:40 Also I've worn the plating right off too many to count Jul 02 10:54:17 * karlp boggles Jul 02 10:54:47 What? I ride them hard and put them away wet. :) Jul 02 10:55:24 ewww Jul 02 10:55:51 cheapest I could find is still rated for 5k insertions, what are you doing with them? dipping them in sand first? Jul 02 10:56:18 What they're rated for and what they survive can be two different things - they tend to rotate in use Jul 02 10:56:34 I've had far more luck with proper 1/4" ones. Jul 02 10:56:40 I mean, usb-a ports are only nominally 1500 Jul 02 10:56:53 "proper" becuase you're a ham fisted guitarist :) Jul 02 10:57:09 I've never played a note of Stairway in my life. :P Jul 02 10:57:19 1/4" are just so huge. Jul 02 10:57:23 They're properly sized Jul 02 10:57:35 I'll note I rarely use such connectors for.. alternate purposes Jul 02 10:57:45 what's your opinion on the "modern" ethernet connectors then? :) Jul 02 10:58:09 8P8Cs? Cheap and disposable, thankfully.. Jul 02 10:58:31 no, I mean the new ones .) Jul 02 10:58:37 Which new ones? Jul 02 11:00:00 * ldir wonders what sort of war he has accidentally started Jul 02 11:00:14 ldir: I don't know, but if someone says DB-9 they're getting shot Jul 02 11:00:24 * karlp digs up a link Jul 02 11:00:31 hang on, you like 1/4", but not db9? :) Jul 02 11:00:38 * Monkeh shoots karlp Jul 02 11:00:40 I don't want one of those :-) Jul 02 11:00:42 It's a DE-9, you barbarian Jul 02 11:00:49 And I love d-sub Jul 02 11:00:50 yeah it's a DE-9 :P Jul 02 11:01:22 darn i was late because i fact checked just in case Jul 02 11:02:24 Monkeh:these ones: compact ethernet port Jul 02 11:02:50 blah https://www.te.com/usa-en/products/connectors/modular-jacks-plugs/rj-point-five-connectors.html?tab=pgp-story Jul 02 11:03:17 there's another one from harting iirc. Jul 02 11:03:36 wargh Jul 02 11:03:38 Kill it Jul 02 11:03:40 Non-standard Jul 02 11:04:00 Unacceptable Jul 02 11:04:26 phwoar connector porn https://cpc.farnell.com/neutrik/ne8mx6/cat6a-ethercon-cable-connector/dp/CN20292?mckv=nvhejcRR_dc%7Cpcrid%7C74629370918234%7Ckword%7C%7Cmatch%7C%7Cplid%7C%7Cslid%7C%7Cpid%7CCN20292%7C&CMP=KNC-MUK-CPC-SHOPPING-M-400-Shopping-Adaptors+Cable+Connectors+Leads-Connectors&msclkid=57e7ce9c5ac811fee05b9387633dda43&utm_source=bing&utm_medium=cpc&utm_campaign=M-400-Shopping-Adaptors%20Cable%20Connectors%20Leads&utm_term=457822898294764 Jul 02 11:04:26 9&utm_content=Connectors Jul 02 11:04:45 https://www.hirose.com/product/series/IX too. Jul 02 11:04:48 well that's atleast a long link Jul 02 11:05:07 Hi yanosz Gargoyle has no IPv6 for now. Ask in the forums about it to get more info there are a nice bunch Jul 02 11:05:08 ldir: Mmmm, Neutrik... *drools* Jul 02 11:05:13 lol - yes, https://www.neutrik.com/en/products/data/ethercon better Jul 02 11:05:20 O ldir cool glad it got there Jul 02 11:05:37 ahrting/hirose hve tht IX one, that's some IEC "next gen ethernet" thing Jul 02 11:06:05 karlp: I just call it fibre, but that's me Jul 02 11:06:17 k, so what fibre connector is ok :) Jul 02 11:06:36 Well it's mostly LC these days, which works okay. Jul 02 11:06:42 At least in normal scale stuff. Jul 02 11:06:46 ldir sounds good thanks. Jul 02 11:07:05 But they don't rely on friction to prevent rotation and the loss of plating :P Jul 02 11:07:29 Don't make me get the XLRs out Jul 02 12:01:45 Hauke: ping Jul 02 12:59:04 * ldir abandons his attempt to drop the out of tree cake kmod - 4.19 v what is in 5.1 do not cherry-pick cleanly and since I got zapped by a migraine yesterday the brain cannot reconcile the two. Jul 02 14:25:50 does someone know more about atheros 803x phy and why nearly always it is selecting "at803x-disable-smarteee"? Jul 02 14:33:55 lynxis: smarteee is a feature for MAC that doesn't support EEE at all. Jul 02 14:34:24 I think most MAC Jul 02 14:34:35 I think most MAC's these days support EEE Jul 02 14:36:08 Rene__: right. But why is it disabled on most of the supported boards with an at803x phy? Jul 02 14:39:43 If smarteee is enable the PHY don't send any EEE data to the MAC. So the MAC doesn't do any powersaving. Jul 02 14:44:34 lynxis: I have vague memories of reading somewhere that EEE is "broken" in some way on the QCA switches Jul 02 14:44:57 (at least with the open-source drivers) Jul 02 14:48:09 Monkeh: ping - did you get my address PM ? Jul 02 15:00:22 ldir: I did, already on their way :) Jul 02 15:01:10 fantastic service - take a pay rise Jul 02 15:02:48 I already pay myself the maximum allowed for a volunteer - nothing. :) Jul 02 15:18:13 is that normal I don't get base-files_.ipk generated? Jul 02 15:18:19 this gives me nothing: find ./bin/packages/ -name "base-files*" Jul 02 15:35:40 goes under targets/xxx/generic/packges Jul 02 15:46:58 hmmm, wonder what slow blinking power led means on wrt 3200acm Jul 02 15:47:39 RUN! RUN! SHE'S GONNA BLOW!!!! Jul 02 15:51:15 unlikely ;-) Jul 02 15:53:13 jow you found anytime to evaluate the opkg patch with licence info? Jul 02 15:55:55 needs to speak serial at it to see how unhappy it really is :-( Jul 02 15:59:44 ldir having fun? Jul 02 16:00:11 :) Jul 02 16:00:15 yes, well ish. Jul 02 16:00:31 ynezz ping Jul 02 16:01:30 Some times I love to get my teeth in to a prob, but that one was just to mutch lol Jul 02 16:01:35 All antennas are plugged back in, the case is back together, even put 2 of the screws back. Thought I'd power it up, which other than the slowly flashing power led is a little disappointing. Jul 02 16:01:59 My wife says why do you mess around with stuff all the time! Jul 02 16:02:06 Did you try to take the heatsink off at one point ? Jul 02 16:02:15 no mate Jul 02 16:02:20 why? Jul 02 16:02:55 One of the screws holding the heatsink on was very undone. Jul 02 16:03:21 That router is the one that was sent to me by Linksys on the openwrt forums Jul 02 16:03:44 They gave 50 away to get people to use openwrt on them Jul 02 16:04:17 Curious. Oh well. Need to go down the serial adapter route to see what if anything is going on. That might arrive tomorrow. Jul 02 16:04:20 I definitely did not tuch the heat sink Jul 02 16:04:39 yeah Jul 02 16:04:55 I had flashed it a tunn of times Jul 02 16:05:00 At least it's all back in the box. Jul 02 16:05:21 then it just stopt booting to one of the partitions. Jul 02 16:05:42 and then the other ? Jul 02 16:05:52 cool did you like my packing? :) Jul 02 16:06:15 Old amazon boxs can come in handy Jul 02 16:06:42 yeah the other lol Jul 02 16:08:30 erm, now I wasn't going to say anything about the packing, but now you bring it up.... it looked like it had been packed by a blind man :-) Jul 02 16:09:13 or his dog. Hard to tell. ;-) Jul 02 16:09:38 hahahahah I will not tell my wife you said that! ;) Jul 02 16:09:47 lol Jul 02 16:10:12 I packed it really Jul 02 16:10:42 phew Jul 02 16:11:16 When my wife packs things she uses a hole rool of sticky tape! Jul 02 16:11:22 roll* Jul 02 16:12:03 The kids have all xmas day to get there gifts haha Jul 02 16:12:09 is that to cover the bits that you did? Jul 02 16:13:09 * Tapper displays a sheepish grin! Jul 02 16:14:08 I am going give up trying to build for thes dam c7 Jul 02 16:14:36 I am just going to flash snapshots from the site and install luci after the fact Jul 02 16:19:14 * ldir takes the dog for a walk before getting clawed to death Jul 02 16:21:26 w00t our new Procerra Packetlogic is here PL8840, 10Gb baby, shape all the packetz Jul 02 16:32:59 [Tue 2019-07-02 08:46:55 AM PDT] hmmm, wonder what slow blinking power led means on wrt 3200acm <------ last night, someone else in #openwrt was also asking this Jul 02 16:47:48 KanjiMonster: ping Jul 02 16:57:17 jow: did you see my message about nf_reject vs. nft_reject in 4.19? Jul 02 17:04:41 philipp64: uhm no. which message? Jul 02 17:20:14 is python3 already a requirement for the buildroot? Jul 02 17:33:56 aparcar[m]: buildroot is a different project. you mean the build system of openwrt, right? Jul 02 17:34:08 lynxis: si Jul 02 17:35:12 aparcar[m]: pong Jul 02 17:35:33 KanjiMonster: do you have more thoughs on the device title split? Jul 02 17:35:44 lynxis: IIRC our build system is still called buildroot in some places Jul 02 17:35:55 I always have many thoughts Jul 02 17:36:17 KanjiMonster: nice, share some :) Jul 02 17:37:34 KanjiMonster: ohh. I did not know Jul 02 17:39:18 aparcar[m]: so far I only found a python2 dependency, but not 3. Jul 02 17:39:46 lynxis: ack Jul 02 17:39:54 ldir: replacing fq_codel with cake would need a size comparison I think Jul 02 17:40:21 lynxis: actually we mostly do in the git commit logs Jul 02 17:41:07 aparcar[m]: btw. I would love to see the python2 dependency replaced with v3 Jul 02 17:43:22 lynxis: will work on that if you check the buildinfo PR ;) Jul 02 17:44:02 lynxis: btw I think ynezz already did some work on that, the problem was some target using it for complex foobar. I'll try to find the mail Jul 02 17:44:44 aparcar[m]: my thoughts are probably now at the nitpick/bikeshedding stage, in general it looks okay to me Jul 02 17:46:56 KanjiMonster: can you elaborate? Jul 02 17:47:28 just did Jul 02 17:48:42 KanjiMonster: cool thank, will test this in a few hours Jul 02 17:48:47 *thanks Jul 02 18:00:02 jow: subj "Issues with 4.19 and nft_reject_ipv4.ko after rebasing/updating" Jul 02 18:02:40 got the feeling it's something trivial but I've not been able to sort it out myself… not clear what the difference is between nft_reject.ko and nf_reject.ko, and how they relate to each other… is this something new with 4.19? Jul 02 18:02:57 or post 4.14, in any case… Jul 02 18:28:23 hmm, isn't nft 'nftables' and 'nf_foo' netfilter aka x_tables ? Jul 02 18:34:56 Wow... just noticed that there is no /dev/stderr in OpenWrt! How do I send something to stderr? Jul 02 18:38:15 Apparently -Werror=implicit-function-decleration is very good at finding bugs. Jul 02 18:40:36 oh wow. It causes problems with C++ Jul 02 18:42:44 ldir: not following. Jul 02 18:43:26 mangix: there was a time when I didn't build anything without -pedantic -Wall -Wformat=2 -Wformat-security Jul 02 18:45:25 I'm wondering if nft_reject.ko is the reject operation for nftables, whereas nf_reject.ko is the reject operation for iptables Jul 02 18:45:34 ah, got it. Jul 02 18:47:39 it's a wild stab in the dark guess Jul 02 18:49:27 in that case, the definitions for nf-core would be wrong in include/netfilter.mk … which might explain why "make -j1 world V=s" just prompted me for NF_TABLES_IPV4 and NF_TABLES_IPV6… something is definitely hinky. Jul 02 18:51:20 valku: /proc/self/fd/2 Jul 02 18:51:42 or echo foo >&2 Jul 02 18:52:11 jow : thank you, I have figured it out but I wonder why not have /dev/stderr in the first place? Jul 02 18:52:52 echo foo >&2 is a lot better and easier to memorize, thank you for the great tip Jul 02 18:53:38 For one thing, /dev/stderr is not portable Jul 02 18:55:28 I think the openwrt /dev population logic is simply not propgrammed to create the /dev/stderr -> /proc/self/fd/2 symlink Jul 02 18:56:00 jow: ping opkg license patch Jul 02 18:56:17 aparcar[m]: yeah, didn't get around to it yet Jul 02 18:56:46 ack Jul 02 18:57:40 gdbm was not updated since 2014 Jul 02 18:57:41 oh boy Jul 02 18:58:03 jeffsf : I always thought /dev/stderr is one of those things that always exist, like /proc and /sys Jul 02 18:58:57 Maybe on Linux systems... Jul 02 18:59:17 Isn't OpenWrt a Linux system? Jul 02 18:59:57 True, but Linux isn't the entire world of POSIX systems Jul 02 19:00:15 It's akin assuming that bash is the shell Jul 02 19:00:23 Oh, that's what you mean... now I get it Jul 02 19:00:48 (as evidenced by the kerfuffle of switching to sh/dash/ash on some major distros) Jul 02 19:02:08 * ldir grins - kerfuffle Jul 02 19:10:53 Yeah kerfuffle is a grate word. Jul 02 19:12:57 For some one who cant spell, like me. I do like nice sounding words. Jul 02 19:59:26 KanjiMonster: okay updated the device-title-split Jul 02 20:40:23 Anyone hankering for a GL.iNet GL-AR750S, Amazon US has a 20% off coupon running on it right now. Jul 02 20:41:22 ynezz: you worked on the python2 to python3 tansition right? Jul 02 20:44:13 * mangix learned today the PCEngines ALIX is extremely slow Jul 02 20:45:07 * jeffsf boxes up his SE440BX behemoth and ships it to mangix Jul 02 20:45:47 lol Jul 02 20:46:06 I use it as a space heater in the winter Jul 02 20:46:27 I was referring to this: https://github.com/openwrt/openwrt/commit/7c23f741e97f6645bb5cd662a4943796a344b26a Jul 02 20:50:10 mangix: slow relative to what? Jul 02 20:50:49 beats me Jul 02 20:51:28 xdarklight: nice work Jul 02 20:52:33 hmm Jul 02 20:52:55 ill have to try openconnect to connect to our anyconnect service Jul 02 21:27:44 What's the objective of the PR spam? Anyone figured that out? Jul 02 21:28:17 (pull-request, not the PR announcement of BattleMesh support) Jul 02 21:29:26 There's PR spam? Jul 02 21:30:19 Looks like a noob mistake. He wants to rebase master onto 18.06. Jul 02 21:30:24 Yeah, been a few cases of PR of a couple hundred commits Jul 02 21:30:45 Oh, yeah, that's just people not grasping how github works. Jul 02 21:31:06 Everyone pokes blindly a bit at first Jul 02 21:41:59 Hauke: :). too bad my patches don't seem to show up on any mailing list yet... I have no idea why though :/ Jul 02 21:42:01 we get teh same type of prs in othher projects, it's just git fuckups Jul 02 22:03:05 karlp: is that a comment on my "patches don't show up on mailing lists yet"? Jul 02 22:03:16 LOL Jul 02 22:15:14 cotequeiroz: there have been several PR's submitted with that exact mistake Jul 02 22:15:29 they just get closed Jul 02 22:29:12 xdarklight: no, the "pr spam" Jul 02 22:31:54 speaking of PR spam, can someone maybe uhm take a look ;) https://github.com/openwrt/openwrt/pull/2156 Jul 02 22:32:33 fixes ability to specify openvpn tls-cipher **** ENDING LOGGING AT Wed Jul 03 03:00:28 2019