**** BEGIN LOGGING AT Tue Jul 31 03:00:01 2018 Jul 31 04:55:53 morning Jul 31 05:02:26 Good moaning Jul 31 05:19:54 hello Jul 31 05:20:01 * blogic checks buildservers Jul 31 05:20:25 they are still building Jul 31 05:20:55 blogic: http://phase1.builds.lede-project.org/waterfall -> some fallout Jul 31 05:21:06 but not really more or less compared to the last weeks Jul 31 05:21:09 jow: i know, could not fix it all last night Jul 31 05:21:31 already fixed the brcm2708 Jul 31 05:21:47 gemini is a parallel build issue that i fail to understand Jul 31 05:22:39 x86 failed at netifd oO Jul 31 05:22:41 cc1: error: unrecognized command line option '-Wno-unknown-warning-option' [-Werror] Jul 31 05:22:51 fixed that already aswell Jul 31 05:23:03 ah Jul 31 05:23:06 * jow is lagging behind Jul 31 05:23:17 no worries Jul 31 05:23:27 i was at it for 16.5 hours yesterday Jul 31 05:24:19 still, the unknown flag error above looks fishy and unrelated to your fixes Jul 31 05:24:32 as if the cross clfags got tested against the host cc Jul 31 05:24:41 jow: /var/lib/buildbot/slaves/slashdirt-02/MAIN/build/build_dir/target-mipsel_mips32_musl/netifd-2018-07-30-75ee7905/interface-ip.c:724:11: error: unused variable 'macaddr' [-Werror=unused-variable] Jul 31 05:24:47 this is the actual error Jul 31 05:25:03 ah Jul 31 05:25:05 indeed Jul 31 05:25:20 i'll let that build for a sec and see when they are all done Jul 31 05:25:41 btw, I've seen that chris5560 also stopped maintaining ca-certificates Jul 31 05:25:43 any takers? Jul 31 05:26:31 I also want to get rid of ca-bundle and simply generate that on the fly upon opkg install ca-certificates Jul 31 05:26:36 maybe I should take a look Jul 31 05:28:32 after an upgrade to 17.01.5, Luci is reachable on port 80, but I can't connect on 443 anymore.. any ideas? Jul 31 05:28:55 I've confirmed the uhttpd config appears correct.. I don't think there is a luci-ssl package anymore? Jul 31 05:29:06 is there a new way to enabled SSL/TLS? Jul 31 05:29:29 witten: there still is luci-ssl which you need to install Jul 31 05:29:53 actually, assuming you retained your config you likely still have proper certs so you don't need a generator Jul 31 05:30:04 huh Jul 31 05:30:06 in this case installing one of the ustream library should be enough Jul 31 05:30:17 libustream-mbedtls Jul 31 05:30:25 or libupstream-openssl Jul 31 05:30:32 giving that a shot.. Jul 31 05:31:05 blogic: pci wifi on ath79 was not supported yet, right? Jul 31 05:31:29 blogic: at least it seemed common knowledge in the forum that units like the wndr3800 do not have working wifi yet on ath79 Jul 31 05:31:58 there was a bug in one of the pcie drivers rending it dysfunct Jul 31 05:32:06 it should work, if not it'll be minor Jul 31 05:32:16 i have a 3800 here and did make wifi work on it Jul 31 05:32:22 huh, I think maybe the problem was I hadn't done an "opkg update" and so luci-ssl package didn't show up in the list? Jul 31 05:32:22 but we merged a pile of patches Jul 31 05:32:23 asking because I've just seen 1714 Jul 31 05:32:26 FS#1714 Jul 31 05:32:37 does "opkg update" cache not persist across restarts? Jul 31 05:32:42 witten: no Jul 31 05:32:55 okay, I guess that was it.. thanks for the help! Jul 31 05:33:05 I mistakenly assumed it worked like apt-get update Jul 31 05:33:08 the package lists are several MB large, most devices could not store them outside of the ram Jul 31 05:33:41 jow: that bootlog does not load any wifi drivers Jul 31 05:34:00 i'd claim windows n00b that built a bad image Jul 31 05:34:00 makes sense Jul 31 05:34:19 blogic: for me its also one of the "meh *shrug*" tickets Jul 31 05:34:28 nothing actionable about it, little info Jul 31 05:34:39 effort to understand is larger than effort to write it Jul 31 05:34:48 That makes a psyborg bug report look good Jul 31 05:34:57 yeah i don't see wifi drivers @ all :/ Jul 31 05:35:04 jow: i shagged psyborg and he is pregnant now Jul 31 05:35:13 the pr i submitted yesterday fixed it on rspro for me Jul 31 05:35:14 blogic: yeah, seen it on the billboard Jul 31 05:35:20 (merged pr) Jul 31 05:35:28 :P Jul 31 05:37:00 ldir: the mvebu breakage stems from your kernel bump Jul 31 05:37:06 ldir: Speculative Store Bypass Disable (ARM64_SSBD) [Y/n/?] (NEW) aborted! Jul 31 05:37:19 Certainly his PCI is working so it's not that Jul 31 05:37:28 ldir: same as mt7622 Jul 31 05:37:46 jow: ok, so i'll wait another 2 hours before fixing build errors Jul 31 05:37:49 blogic: blame upstream for sneaking in new options in minor versions :P Jul 31 05:37:54 want the current pendings out the way Jul 31 05:38:04 jow: no i blame ldir Jul 31 05:38:07 its all his fault Jul 31 05:38:31 :D Jul 31 05:38:46 make kernel_olddefconfig Jul 31 05:40:57 FS#1682 - wtf? How'd he manage that? Jul 31 05:51:15 why does my AFD bell ring when i see this -> https://github.com/openwrt/openwrt/pull/1229 Jul 31 05:51:30 bunkershild.de, @verbotene.zone, ... Jul 31 05:57:31 blogic: I see no political remarks in the changes, so it should be fine Jul 31 05:59:03 me neither Jul 31 05:59:11 but still my alarm bell rings Jul 31 05:59:58 at least the changes to target/linux/ipq40xx/Makefile (DEFAULT_PACKAGES) look wrong Jul 31 06:00:10 I should fire up the ath79 gubbins and see how that's all gone Jul 31 06:00:19 pkgadd: the whole PR is fully of crappyness Jul 31 06:00:34 blogic: It looks like copy-pasta. Jul 31 06:00:44 yep, and I didn't even look into the individual patches - just the aggregated diff Jul 31 06:00:48 * Monkeh has been guilty of this in the past Jul 31 06:00:48 Monkeh: the n00bery is strong with this one Jul 31 06:01:03 blogic: We must all start somewhere Jul 31 06:03:56 interesting that the dts is also trhee times bigger than that of any other ipq40xx device (and I think it contains a lot of stuff that should be inherited from the generic dtsi instead) Jul 31 06:04:16 Wow, a thousand-line dts file. It looks like it was extracted from a .dtb and added some comments Jul 31 06:04:26 haha, I was reading the same file Jul 31 06:04:59 luaraneda: maybe it even was Jul 31 07:12:42 if anyone decides to dig into the slab code, think twice Jul 31 07:12:52 you'll know stuff after that you dont want to know Jul 31 07:13:08 day job work sucks Jul 31 07:13:39 sounds like every time i want to-do "something cool" done by linux :p Jul 31 07:15:25 btw i'm looking for a samba/timemachine tester, anyone here wanna volunteer? Jul 31 07:15:27 so turns out that calling kmalloc will check if the size is a builtin constant Jul 31 07:16:04 and if it is and if it is >= a cache'able page it'll run into the node allocator which has pre allocated regions Jul 31 07:16:41 but as we are allocating large chunks that will eventually fragment badly. normally you use the CMA region for this but not if its a __builtin_constant() Jul 31 07:17:28 so, my guess is that if i dont allocate 1MB but 1MB+1byte the oops will go away as the alloc will hit the normal kmalloc path and thus grab a CMA region Jul 31 07:17:51 now how ugly is that as a production work around Jul 31 07:18:05 * blogic grabs the super glue and hotglue gun and duct tape Jul 31 07:18:25 if it works.... have seen uglier stuff done in c++ Jul 31 07:18:43 c++ is always ugly Jul 31 07:19:41 yeah but i have seen oop overkill so many times, like having 100+ classes all with 2 accessor functions and a single member... just because "everything needs its own class" and "we need to protect everything"... Jul 31 07:21:41 so for some masochistic reason everytime i look at c code it kinda like it :) Jul 31 07:29:47 blogic:thx for fixing the netifd compile issue; sorry Jul 31 07:37:17 mhh whats up with this: "ERROR: "gcc" either does not exist or does not work" for host build on "aarch64_generic" ? Can this happen via buildbot? Jul 31 07:42:48 dedeckeh: sure Jul 31 07:50:08 dedeckeh: does this change make sense to you? https://pastebin.com/csZ60zas Jul 31 07:54:06 KanjiMonster:are you hitting an issue by using the local endpoint for outgoing communication ? Jul 31 07:54:50 dedeckeh: yeah, the upstream router seems to be dropping outgoing connections not going to the AFTR Jul 31 07:57:04 KanjiMonster:would it make sense to make this configurable ? Jul 31 08:03:18 dedeckeh: maybe, would need to be done at the dhcpv6 level then, though. probably should take a look at the RFCs if they say anything about the usage of the local address Jul 31 08:18:12 not fallout from my nightly kernel bump yet? Jul 31 08:18:14 dedeckeh: the map rfc has under security considerations ( https://tools.ietf.org/html/rfc7597#section-10 ) "Furthermore, the MAP CEs SHOULD use a DNS transport proxy [RFC5625] function to handle DNS traffic and source such traffic from IPv6 interfaces not assigned to MAP." can't find anything for lightweight 4 over 6 in the rfc Jul 31 08:18:15 no* Jul 31 08:18:54 stintel: some, iirc a new kconfig symbol crept in on arm Jul 31 08:19:05 oh :/ Jul 31 08:19:38 Speculative Store Bypass Disable (ARM64_SSBD) [Y/n/?] (NEW) aborted! Jul 31 08:20:28 did anyone take care of it already or shall I fix it Jul 31 08:20:47 appears to affect mediatek/mt7622, mvebu/cortexa53 and mvebu/cortexa72 Jul 31 08:23:48 stintel: would be great if you could look into it Jul 31 08:43:33 stintel: go ahead, all aarch64 targets are effected Jul 31 08:47:20 blogic: only the ones on 4.14 ;-) Jul 31 08:47:44 considering a change to the refresh script to avoid that in the future Jul 31 08:48:37 stintel: yep Jul 31 08:49:03 stintel: anything that makes it more ldir proof would be a good thing ;-) Jul 31 08:49:45 ldir: ;D Jul 31 08:53:04 man these hardware bugs have really caused headaches this year Jul 31 08:53:16 I'm seeing many new symbols on ARM actually Jul 31 08:53:18 broken inside Jul 31 08:53:23 😂 Jul 31 08:53:24 stintel: yep Jul 31 08:53:34 broken arm Jul 31 08:53:57 sounds painfull Jul 31 09:00:07 hmmm, make kernel_oldconfig now asks about CC_OPTIMIZE_FOR_PERFORMANCE vs CC_OPTIMIZE_FOR_SIZE everywhere Jul 31 09:02:24 mkresin: could you look into ^ Jul 31 09:18:29 make kernel_oldconfig seems to add a shitload of unrelated stuff Jul 31 09:18:47 going to manually add ARM64_SSBD Jul 31 09:19:01 and push, fix further fallout when it becomes visible on the buildbots Jul 31 09:57:57 stintel: not promptly. looks like kernel_oldconfig ignores config/Config-kernel.in for some reason Jul 31 10:04:35 probably because it just calls "make oldconfig" in the kernel dir? Jul 31 10:11:33 maybe. but it should copies generic/config target/config at least before doing so Jul 31 10:27:11 mkresin: stintel: the reason is probably to have a defined stable state for the kernel config; without any user changeable symbols enabled (or disabled if they default to y). similar how we ignore the kmod selection state for kernel_*config Jul 31 10:31:08 makes sense. so far I'm lost in the makefiel magic without an idea how to fix the regression Jul 31 10:36:32 is there a way to forcefully expire a dnsmasq DHCP lease (without rebooting)? Jul 31 10:37:03 remove it from the lease file and restart dnsmasq :D Jul 31 10:37:08 otherwise, no Jul 31 10:37:49 ok :) Jul 31 10:38:27 static leases are ignored as long as the previous (auto) assignment remains valid. Hence my question Jul 31 10:39:21 heh Jul 31 10:40:13 f00b4r0: tell the client device to release the previous lease and get a new one Jul 31 10:40:41 KanjiMonster: IP camera. Not an option I'm afraid ;) Jul 31 10:41:08 just install openwrt on it as well ;) Jul 31 10:41:18 that will teach you to use long lease times Jul 31 10:41:48 6h Jul 31 10:42:16 heh Jul 31 10:42:30 https://i.imgur.com/Zykc35j.png Jul 31 10:42:34 may actually die Jul 31 10:44:52 23C here. Nice and comfy Jul 31 10:45:27 my bed is now a leaky water mattress Jul 31 10:45:33 so so damp :D Jul 31 10:46:54 hmm. 5GHz is completely useless here. Might as well disable it Jul 31 10:54:17 jwh: 32 here currently, and forecast said going up to 38 Jul 31 10:54:44 so 30 sounds nice from my standpoint ;P Jul 31 10:55:10 heh Jul 31 10:55:42 move North. Jul 31 10:55:43 but remember, climate change is a hoax invented by the chinese Jul 31 10:55:54 Global warming says it's a smart move :) Jul 31 10:56:16 s/Global warming/Chinese hoax/ ;^) Jul 31 10:56:23 trump did it Jul 31 11:16:06 jow: ping? Jul 31 11:28:46 KanjiMonster:had also a lokt to the map rfc but could not find a clear statement Jul 31 11:28:53 *look Jul 31 11:29:53 KanjiMonster:the patch makes sense but I would propose to make it configurable Jul 31 11:49:16 q! Jul 31 11:49:22 -ewin Jul 31 11:50:09 :) Jul 31 11:50:35 wigyori: there is a notion that you ban spacerat Jul 31 11:51:07 yes Jul 31 11:51:13 someone just do it Jul 31 11:58:40 wigyori: thanks Jul 31 11:58:58 could you also ask imre to hand over the contact to freenode ? Jul 31 11:59:11 yup Jul 31 11:59:12 i spoke with the freenode admins and only imre can have people added to the cloaking Jul 31 11:59:21 so we would like him to hand over ownership Jul 31 11:59:31 freenode would prefer him to voluntarily give it over Jul 31 11:59:48 otherwise we need to reapply and they need to make a comitee decision, yadiyadiya Jul 31 11:59:54 and we really want to avoid that Jul 31 12:00:14 ah, so democracy is tough Jul 31 12:00:15 :) Jul 31 12:00:20 (pinged him now) Jul 31 12:00:23 cool Jul 31 12:00:37 wigyori: democracy is easy if those with the key participate Jul 31 12:06:04 getting weird "Uncaught ReferenceError: cbi_init is not defined" in luci today... Jul 31 12:06:20 how does that even happen when all the sources loaded. Jul 31 12:06:44 karlp: tried cleaning the cache ? Jul 31 12:07:05 yeah, and adding ?fixme to the url to reload, and all sorts. Jul 31 12:07:10 some of them work, some don't Jul 31 12:07:17 main admin status page is just busted complete. Jul 31 12:07:31 I wonder if chromium update decided to start blocking xhr's ina weird way or soemthing Jul 31 12:08:03 didn't they start enforcing "http==insecure" or somesuch? Jul 31 12:10:02 I'm only using http anyway. Jul 31 12:11:05 https://www.theregister.co.uk/2018/02/08/google_chrome_http_shame/ Jul 31 12:11:28 dunno if they also enforce some more restrictions on what you can do over http, along with that Jul 31 12:18:39 wigyori: can you remove that ban? I am talking to the guy - he is fixing it Jul 31 12:20:08 so apparently it's xchat that doesn't handle freenode channel redirects Jul 31 12:20:46 how has he not noticed for 3 months Jul 31 12:20:51 hes not said a word Jul 31 12:29:56 31|14:23:09 I've seen some networks/chanops basically abusing the fakejoin feature, that's why I wrote a script to defeat it. Jul 31 12:29:57 heh Jul 31 12:30:24 stintel: using xchat (on osx), it doesn't have any trouble with the redirect Jul 31 12:30:51 * #lede-dev #openwrt-devel :Forwarding to another channel Jul 31 12:30:54 abusive behaviour, check Jul 31 12:31:18 and as jwh mentioned he hasn't said anything and never reacted when pinged Jul 31 12:32:56 kinda convenient tbf, mystery logger perhaps! Jul 31 12:33:34 well I sent him a notice once Jul 31 12:33:53 but either way trying to subvert the forwarding stuff is probably reason enough Jul 31 12:33:55 the thing is pinging him in the channel is pointless Jul 31 12:33:56 * karlp just ignores parts/joins. Jul 31 12:34:29 and last night I got fed up with it so I pasted him 40 lines of him join/parting Jul 31 12:35:33 stintel: I just find it somewhat dubious that he suddenly finds interest in staying on the channel now that he's effectively banned from it, considering his typical presence lasted a couple seconds every 10mn... Jul 31 12:35:40 for the past what, few months? Jul 31 12:36:20 there's currently 405 people in here and maybe less than 15-20 of them chat on a regular basis. Jul 31 12:36:34 thats what i meant by convenient Jul 31 12:36:35 heh Jul 31 12:36:49 salcedo: none of them trash the channel with their broken client ;) Jul 31 12:36:49 sometimes folks forget they have clients running Jul 31 12:37:18 (or purported broken client, even ;) Jul 31 12:37:56 stintel: just to clarify: so you PM'd him last night, and you got ignored until he was banned? Jul 31 12:38:16 nobody forgets they are running a gui client for months Jul 31 12:38:23 sorry Jul 31 12:38:29 ridiculous notion Jul 31 12:38:57 stintel: if so I'd be even more suspicious Jul 31 12:39:05 heh Jul 31 12:40:29 ya pretty easy to forget you have a gui client running on osx Jul 31 12:40:49 not if you use it Jul 31 12:40:51 ;) Jul 31 12:41:25 i left a znc on another network for like 2 years. none of my clients were connecting to it so i had no idea i was on the network. Jul 31 12:41:46 digital messy Jul 31 12:46:10 xchat is noy znc Jul 31 12:46:11 not Jul 31 12:46:18 znc is a bouncer Jul 31 12:46:58 not even remotely similar Jul 31 12:49:30 is samba considered safe? Jul 31 12:50:52 i'm looking for a nas solution with cross platform support for my windows and linux computers Jul 31 12:51:10 safe for what? Jul 31 12:51:20 karlp: for using it on LAN Jul 31 12:51:23 i guess Jul 31 12:51:26 yes Jul 31 12:51:31 its as safe as nything on LAN Jul 31 12:51:40 it gets dangerous when you do wan side stuff Jul 31 12:53:04 from what I see samba 4 is new in openwrt Jul 31 12:53:42 so? Jul 31 12:54:12 I just have a nasty habit of stating things Jul 31 13:16:34 stintel: sure Jul 31 13:18:22 wigyori: thanks - feel free to add it again in case he didn't properly fix it Jul 31 13:20:35 default procd init handling for "reload" is stop/start right? config has changed, I see start getting called again, but it's not stopped Jul 31 13:21:14 correct Jul 31 13:22:11 gonna be one of those days then by the look of things. Jul 31 13:22:17 * karlp turns it off and on again Jul 31 13:30:57 I suspect I'm running into https://developers.google.com/web/updates/2018/07/site-isolation being turned on by default from chormium 67 whicih I just got Jul 31 14:18:49 Hi any one fancy bumping the mwlwifi driver in master? Jul 31 14:46:06 wigyori: for future reference, a ban suffixed with $##fix_your_connection will forward joins from that user to the channel with that name, which is the customary bitbucket on freenode for exactly these situations Jul 31 14:48:03 not that it would've made a difference in this case Jul 31 14:57:02 gninrom Jul 31 14:57:11 Congraaaaaaaaaaaaaaaaaaaaaaaaats for 18.06 Jul 31 14:57:23 my routers will never be the same again :-))))) Jul 31 14:59:02 I should update see if it makes ath10k usable Jul 31 14:59:45 ath10k-ct seems to work Jul 31 14:59:47 I will have loooooooog nights this week. A lot of routers to update Jul 31 15:21:49 blogic: any chance the e1700 test is still on for today? Jul 31 15:24:22 dangole: Ntemis was looking for you in reference to a certain oxnas board Jul 31 15:25:25 DonkeyHotei: i'd be glad for more people testing oxnas stuff, esp. also ox810se which is supposedly working but i don't have hardware to test Jul 31 15:26:43 he's obviously not here atm, but maybe he might respond to a ping on github, idk Jul 31 15:33:50 dangole: i also noticed your commits regarding lantiq vectoring firmware. what's the plan with that? Jul 31 15:34:29 what plan do you expect. they are in master. Jul 31 15:35:56 mkresin: you had reservations about that approach, so i figured things may yet change Jul 31 15:36:28 DonkeyHotei: the only slight problem is that t-com's download server now requires https Jul 31 15:36:51 the ca-certificates maintainer just quit Jul 31 15:36:59 hence you need ustream-ssl (and ca-certificates) Jul 31 15:37:04 and wolfssl... Jul 31 15:37:14 (maintainer quit as well) Jul 31 15:37:28 DonkeyHotei: yes, I'm not the biggest fan of the approach.But I don't have anything better either. Jul 31 15:37:46 i consider taking over wolfssl because it's too useful for wpa_supplicannt/hostapd's SAE support and the only way to get that working on a 4MB flash device Jul 31 15:38:04 though having support for mbedtls in wpa_supplicant/hostapd would be nicer... Jul 31 15:38:18 xdarklight's page says that dsp firmwares released after a certain date are redistributable Jul 31 15:38:43 DokeyHotei: but only those without vectoring support (supposedly for patent reasons) Jul 31 15:38:51 ah. Jul 31 15:39:20 for licence fee reasons. ltq charges customers extra for vectoring enabled firmware Jul 31 15:39:37 DonkeyHotei: and on some recently deployed VDSL2 DSLAMS the link might not even go up without vectoring support, maybe or maybe not fallback to ADSL2+... Jul 31 15:40:25 All you lucky people with vectoring Jul 31 15:40:26 :P Jul 31 15:40:26 apparently ltq even sold the xRx200 platform chipset and licenced it for ADSL2+ only (eg. for the TL-W8970B, which happends to be the most popular VDSL router we support...) Jul 31 15:40:30 that would give your fw downloader a chicken and egg issue Jul 31 15:40:33 dangole: https://github.com/openwrt/packages/blob/master/utils/cryptsetup/Makefile#L32 - you can drop the @KERNEL_DIRECT_IO dependency, the symbol was removed over a year ago Jul 31 15:41:11 KanjiMonster: cheers. i'll try to find time for some package updates (cryptsetup, lvm2 and all that) in the next couple of days Jul 31 15:42:11 DonkeyHotei: we used to have an extra partition for the VDSL firmware back in the days when there was no redistributable firmware at all Jul 31 15:42:25 only the first hardware rev of the tp-link was lantiq, the later ones broadcom Jul 31 15:42:32 that had the advantage that at least once downloaded, the firmware would survive OpenWrt sysupgrade Jul 31 15:43:13 DonkeyHotei: yes, but they are easy to distinguish, because only the LTQ-based version got 2 USB ports, the BRCM-based ones only got 1 USB port Jul 31 15:43:47 and i thought the most popular was the hh5a Jul 31 15:44:20 Sort of hard to buy off the shelf Jul 31 15:44:26 And limited to one country. Jul 31 15:44:31 DokeyHotei: that's definitely the most feature-complete and best-supported device with VDSL2 support, but i guess the TL-W8970B is more popular (i've seen it way more often) Jul 31 15:44:48 Also, relative four letter word to install Jul 31 15:45:04 Monkeh: the lantiq variants of the tp-link have been unobtainium for some time Jul 31 15:45:14 DonkeyHotei: But they were obtanium for long enough Jul 31 15:45:27 Monkey: that applies to the TD-W8970B just as well, you also need to briefly attach a serial console for the initial flashing Jul 31 15:45:41 because the stock firmware does RSA signature tests and nobody cared to look for exploits Jul 31 15:45:45 dangole: Does it involve cracking a really tough case and soldering to vias? Jul 31 15:46:01 And then spending an hour dealing with a UART loaded bootloader? Jul 31 15:46:52 Monkeh: easy to open, soldering required and stock u-boot can be used Jul 31 15:47:10 So pretty normal, then Jul 31 15:47:11 :P Jul 31 15:47:15 the case is easy to open, the serial is unpopulated, but rather easy to spot. i usually only tape cables to it for those 4 lines of uboot console input required to flash the box, so the board remains unaltered and after re-assembly there are no obvious traces... Jul 31 15:47:16 yes :-) Jul 31 15:47:34 The HH5A is a bit of a step up in annoyance level Jul 31 15:47:36 in .us and .au, the easiest to find xrx200 device is the dm200 Jul 31 15:48:11 I even managed to rip a via out of one Jul 31 15:48:43 while we are talking about, anyone aware that hauke send a DSA driver for the xrx200 switch upstream? Jul 31 15:49:00 * dangole is very aware of that Jul 31 15:49:01 Nope, don't watch upstream.. :) Jul 31 15:49:15 *The incredible Hauke of course Jul 31 15:49:36 that PSB switch is picky and not the most stable thing at all when it come to long/shitty ethernet cables... Jul 31 15:50:11 and doesn't do full-size jumbo-frames afaik Jul 31 15:50:15 dangole: Consumer hardware, what were you expecting? :) Jul 31 15:50:18 * takimata is secretly hoping we will someday get support for the Fritz!Box 3390 ... as far as feature-completeness it is the Annex-B version of the HH5a. Jul 31 15:50:21 Standards compliance? Jul 31 15:51:02 takimata: Well, send one of us one. ;P Jul 31 15:51:24 Monkey: ADMtek IP recycled by Infenion, passed on to Lantiq... their FastEthernet switches were already a nightmare, needed bigger heatsinks than anything else on the boards back then... Jul 31 15:51:46 dangole: Oh good grief, ADMtek live on? Jul 31 15:51:56 Monkeh: I might just do that. :) Jul 31 15:52:20 dangole: you forgot the passed on to intel in your list :-) Jul 31 15:52:21 Monkey: i wouldn't exactly call that 'living' Jul 31 15:52:37 mkresin: but they don't produce those Ethernet switches any longer, do they? Jul 31 15:52:48 dangole: It's alive enough, it moves and causes pain Jul 31 15:53:13 dangole: hard to say what is new intel stuff or renamed ltq stuff Jul 31 15:54:57 Monkeh: I'm not a dev (not an OpenWrt dev anyway) ... the 3390 looks supportable without much pain, no? vrx288 SoC, AR9381 for 2.4G, AR9590 on 5G wifi? Jul 31 15:56:25 Shouldn't be a problem? Not sure about a factory image, never worked AVM, but the hardware should be fine Jul 31 15:56:52 Monkeh: factory install via ftp recovery in teh bootloader Jul 31 15:57:03 mkresin: That works. Jul 31 15:57:30 isn't AR9590 802.11a/n only? In compare to 802.11ac in the HH5a. Jul 31 15:57:46 for all the world, it looks like a 3370 plus an additional AR9590 for dualband. Jul 31 15:58:16 oh, right. it might not do ac. hm. Jul 31 15:58:20 mkresin: Yeah, it's N, no AC Jul 31 15:58:49 The 3490 has a .11ac radio Jul 31 15:59:08 I just noticed that I missed an eintire day of backlogs because I didn't scroll to the end of the screen buffer with irssi Jul 31 15:59:25 I'm sure you didn't miss much Jul 31 15:59:29 tohojo: pong Jul 31 16:01:18 3490 has QCA9558 + QCA9880 for wifiless. Jul 31 16:01:53 It has an SoC for the wifi? Jul 31 16:02:28 ugh, that's what the hardware table says here. hrm. Jul 31 16:02:30 karlp: did you manage to figure out your cbi_init() issue? Jul 31 16:02:46 256 MB RAM, 512 MB flash. Jul 31 16:03:06 I would expect the 2.4GHz to be the AR9381 or similar again Jul 31 16:03:32 I do sort of have access to a 3490 to look now, but I can't really steal it to port Jul 31 16:03:48 jow: not yet. I _suspect_ at present that I may have an incompatbiility with my theme. Jul 31 16:04:22 consider it a non-issue unless someone else sees it. Jul 31 16:05:07 karlp: due to ... reasons the xhr.js and cbi.js scripts are now included by default on every page in the header Jul 31 16:05:14 karlp: you should do the same in your theme Jul 31 16:05:24 since map.htm does not include cbi.js anymore Jul 31 16:05:46 Monkeh: a funky italian forum says: "2.4GHz Wireless Chipset: Qualcomm QCA9558-AT4A operating at 720Mhz frequency and up to 450Mbps in MIMO 3T3R configuration. It is equipped with a dedicated RAM Winbond W9751G6KB-25 de 128MB DDR2 800Mhz." Jul 31 16:05:54 yeah, sounds like something like that. Jul 31 16:06:02 takimata: That seems strange. Jul 31 16:06:04 that sounds ... weird. Jul 31 16:07:02 I mean, we have seen weird for wifi (Easybox 904) ... but I can't imagine why AVM would do that. Jul 31 16:07:03 Monkeh: considering the ralink SoCs iNIC mode, not that strange Jul 31 16:07:27 mkresin: It seems a little overkill though Jul 31 16:07:42 Unfortunately the only photos of the 3490 I've seen have a can on. Jul 31 16:08:03 well, rt3883 used on the easybox 904. but yes, something I wouldn't expect to see in AVM hardware Jul 31 16:08:22 (the EB904 would be another lovely device, if it wasn't for the RT3883 iNIC wifi.) Jul 31 16:08:51 This article: https://www.hwupgrade.it/articoli/periferiche/4297/3490-il-miglior-fritzbox-di-sempre-per-il-videogiocatore_6.html Jul 31 16:08:57 Says QCA9880 - which is what I was expecting Jul 31 16:08:58 takimata: and the custom bad block magic and so on Jul 31 16:09:23 But no mention of the other chip Jul 31 16:09:24 yeah, it's a shame. (I actually have two of those lying around here.) Jul 31 16:10:36 wifi is not weird unless it requires the lantiq wave driver Jul 31 16:11:41 Monkeh: apparantly they did the same on the 7490. QCA9558 for 2.4GHz Jul 31 16:12:25 in fact, the 7490 seems like pretty much the same device. Jul 31 16:13:42 takimata: That may be a damper to 3490 porting ideas Jul 31 16:13:46 .. not that the price wasn't already Jul 31 16:14:46 ugh, yes. just browsed ended ebay listings. :/ Jul 31 16:15:12 €60-70 ... meh. Jul 31 16:15:19 Yeah, try the new prices. Jul 31 16:16:10 yeah, big fat "nope" there. Jul 31 16:16:58 so the EB904 pretty much remains the only "spare change" modem+ac wifi for annex B. Jul 31 16:17:29 aren't the annexes largely interchangeable on vdsl? Jul 31 16:17:49 I'm sure we just went over this one Jul 31 16:18:16 yes they are. not everyone has vdsl yet, though. Jul 31 16:18:22 The annexes are.. nothing but annexes in a document. Jul 31 16:19:05 i used an annex b device on adsl here for a while Jul 31 16:19:07 ... and a set of resistors on the board. Jul 31 16:20:25 Which don't matter for VDSL operation Jul 31 16:20:51 (afaik, but hey, no docs) Jul 31 16:21:12 In either case VDSL does not use annexes to describe different band maps like that Jul 31 16:21:34 Really confusing abuse of terminology. Jul 31 16:24:53 hmm ... the strip club down the street says they have air conditioning. hmm. Jul 31 16:25:21 So you can go sit in a cool room while you remain hot and bothered? Jul 31 16:25:29 lol Jul 31 16:52:25 seems like the build system does have some trouble coping parallel builds Jul 31 16:52:32 +with Jul 31 16:57:19 dangole: the oxnas device in question was something called the xtreamer etrayz Jul 31 16:59:05 yikes. naming devices by throwing a keyboard at the wall. Jul 31 17:00:16 I'll have a dgytujwryj sthrwyrjhwr please Jul 31 17:00:36 takimata: and that's after i did a tolower() on the name Jul 31 17:00:52 ah. trekstor. that explains it. Jul 31 17:00:57 https://en.wikipedia.org/wiki/List_of_Battlestar_Galactica_characters Jul 31 17:01:03 currently using ^ for hostnames :) Jul 31 17:02:13 stintel: I hope you have a couple twin machines called Boomer and Athena Jul 31 17:02:21 corporal komma? ensign semicolon! lieutenant parenthesis! Jul 31 17:02:32 DonkeyHotei: hrm, Ihave an old xtreamer that has an etrayz option. Jul 31 17:02:46 never found anything on what it was meant to be, I always presumed it was like virtual dvds or something Jul 31 17:03:03 from the icons on the menus at least. Jul 31 17:03:14 karlp: if you have a log of #openwrt from july 13, take a look Jul 31 17:03:27 * karlp peers Jul 31 17:03:37 stintel: I'm currently using station names on the Yamanote line. Jul 31 17:03:52 (which is also a popular drinking game.) Jul 31 17:04:35 hrm, not, I'm only in -devel sorry Jul 31 17:05:17 ldir: ping? Jul 31 17:06:00 tl;dr was that Ntemis wanted to add support for the device and was looking for dangole to coordinate Jul 31 17:06:43 Monkeh: actually apollo and athena ==> web/application servers Jul 31 17:07:00 he provided the following links: Jul 31 17:07:00 [Fri 2018-07-13 08:14:22 AM PDT] https://oldgoodstuff.weebly.com/store/p4/Xtreamer_eTRAYz%2C_2Bay_RAID_NAS.html Jul 31 17:07:00 [Fri 2018-07-13 08:14:57 AM PDT] https://nas-tweaks.net/tag/etrayz/ Jul 31 17:07:04 Monkeh: deadbolt and dipper -> database servers Jul 31 17:07:14 Monkeh: ninja & nuke -> nameservers Jul 31 17:07:16 etc Jul 31 17:08:47 hmm Jul 31 17:08:48 Applying ./patches-2.0/100-musl-compat.patch using plaintext: Jul 31 17:08:48 patching file src/usbipd.c Jul 31 17:08:48 Hunk #1 FAILED at 453. Jul 31 17:08:50 running suricata on openwrt, publishing eve to redis, then having logstash with redis input plugin throw the logs in elasticsearch Jul 31 17:08:51 cool stuff Jul 31 17:08:55 trying to build v18.06.0 Jul 31 17:09:08 f00b4r0: thats known, use IGNORE_ERRORS=m to skip Jul 31 17:09:15 ah Jul 31 17:09:37 "make IGNORE_ERRORS=m world" ? Jul 31 17:10:23 not sure if make is smart enough to treat it as varable when in between Jul 31 17:10:26 make it last or fist Jul 31 17:10:31 first Jul 31 17:10:45 hey guys, anyone online that can help me unbrick a tp-link router? Jul 31 17:10:50 make target V=s IGNORE_ERRORS=m Jul 31 17:10:57 ok thanks Jul 31 17:14:07 DonkeyHotei: thanks. Jul 31 17:15:04 oh right, so etrayz was another hardware altogehter, and the etrayz icon in my media box thing is for a pseudo private networking discovery thing for the nas portion. Jul 31 17:15:05 he hasn't been active on irc since that i have seen, but his nick is apparently also his github username Jul 31 17:16:25 I'm gonna be kinda sad when it eventually dies honestly, Jul 31 17:17:01 it's got a nice simple clean interface, just select file and play, no faffing around with requirng tv/movies to be in separate places, no attempts to scrobble or fetch posters or any garbge, works real well. Jul 31 17:22:12 DonkeyHotei: that etrayz thing seems to be OX810SE, so yes, that'd be interesting to support and see if the (existing) drivers for ox820 sata and usb may be modified to work also on ox810se Jul 31 17:27:03 hi guys Jul 31 17:46:08 Is there a way to get the br-lan MAC address from a ubus cmd? Jul 31 17:46:35 jow: yeah, updated my theme and most of my problems are gone. Jul 31 17:48:07 just a little css I need to fix again now. Jul 31 17:48:44 riproute: ubus call network.device status '{ "name": "br-lan" }' | jsonfilter -e '@.macaddr' Jul 31 17:49:59 karlp: do you want me to notify you when I change such things? Jul 31 17:50:26 will probably do some larger markup cleanups again in the future, this time uncluttering css classes Jul 31 17:50:50 if you remember, but not really a huge deal. Jul 31 17:50:55 jow: from #openwrt: [Tue 2018-07-31 10:27:35 AM PDT] my system log is full of "Tue Jul 31 16:41:08 2018 daemon.err uhttpd[1312]: sh: /usr/sbin/nvram: not found" Jul 31 17:51:08 I'd not updated the base since what appears to be luci 0.11 and it's only now that I've got things actaully stopped working :) Jul 31 17:51:08 the device is brcm63xx Jul 31 17:51:28 most of it is now just much better, so this is just good things to fix, even if will be a bit fiddly Jul 31 17:51:58 we were only really trying to add a banner at the top anyway for the most part. Jul 31 17:52:33 nvram pkg has DEPENDS:=@TARGET_brcm47xx||@TARGET_bcm53xx||@TARGET_ar71xx Jul 31 17:54:11 DonkeyHotei: no `nvram` in the entire luci repo Jul 31 17:54:20 DonkeyHotei: maybe he installed a 3rd party package or something Jul 31 17:54:32 [Tue 2018-07-31 10:30:04 AM PDT] the only file containing that strng in my firmware is libiwinfo.so Jul 31 17:54:40 ah Jul 31 17:55:06 yeah, uhm. WIll fix it. Until then bad luck. I just ln -s /bin/true /usr/bin/nvram to make it silent for now Jul 31 17:55:19 *I suggest just Jul 31 17:55:41 in one to two days an updated libiwinfo will be available via opkg Jul 31 17:56:48 in 18.06? Jul 31 17:57:28 yeah Jul 31 18:15:46 hey guys has anyone ath79 running on the tl-wr1043nd v2 already? Jul 31 18:16:16 i know support was merged into master, just wondering if there's any loose ends (like loosing access, no hero with serial :D ) Jul 31 18:16:21 s/loosing/losing/ Jul 31 18:17:23 [Tue 2018-07-31 11:12:43 AM PDT] "ln -s /bin/true /usr/sbin/nvram" doesn't work because it is busybox, it tries to launch the applet based on the link name Jul 31 18:18:20 I am getting hi pings with mwlwifi Jul 31 18:19:20 I have wrt3200 as my router and a wrt1900ac-v2 as a repeter bridge to get internet to a pc i have upstairs pings are 100+ Jul 31 18:20:46 DonkeyHotei: well, be creative. echo -en "#!/bin/sh\nexit 0\n" > /usr/sbin/nvram && chmod +x /usr/sbin/nvram Jul 31 18:25:05 whenever I try to compile a build with nfs-kernel-server it fails. the bug has apparently been around for over two months: https://bugs.openwrt.org/index.php?do=details&task_id=1694 Jul 31 18:25:11 can anyone look into that? Jul 31 18:25:15 any fancy bumping the mwlwifi driver to latest? Jul 31 18:25:50 Hauke: I've fixed the Orange Pi R1 U-Boot :D :D Jul 31 18:26:39 There's a different driver in the u-boot-spi git tree that seems very similar to the StephanVanSchaik one. I managed to make that one work (it has a very minor bug, going to tidy it all up and send patch upstream tonight :D) Jul 31 18:43:11 fahadsadah: nice Jul 31 18:43:22 thanks for working on that Jul 31 18:43:33 I haven't found the time for that in the last months Jul 31 18:43:51 are there already parts of it integarted in upstream u-boot? Jul 31 18:44:56 dangole: you have to change some settings for jumbo frames, I think the driver for the gswip in openwrt doesn't do that Jul 31 18:48:56 Hauke: there are sun4i and sun6i SPI drivers in a staging tree, but the sun6i one at least is buggy (and there isn't a DT for the R1) Jul 31 18:49:51 My new issue is that linux can't boot (can't mount root because can't find the MTD partition), but this is presumably/hopefully simple. I wonder whether it has a DT that says this board has SPI and flash Jul 31 18:54:55 fahadsadah: I added the dt for the r1 Jul 31 18:55:07 to activate spi you just have to remove a comment Jul 31 18:55:21 in upstream u-boot Jul 31 18:59:48 Hauke: Sorry, I couldn't find the SPI driver in upstream U-Boot; I had to copy it from http://git.denx.de/?p=u-boot/u-boot-spi.git;a=blob;f=drivers/spi/sun6i_spi.c;h=3192548e83bbf8fe0414de933afeb038f77b4aab;hb=refs/heads/allwinner-spi (branch in a staging) Jul 31 19:00:09 I've found the DT in Linux you added in , and removed the comment from the DT you added in U-Boot Jul 31 19:00:10 going craze Jul 31 19:00:21 trying to fix some of the reported bugs Jul 31 19:00:29 nfs-kernel-server -> builds fine Jul 31 19:00:34 usbip -> builds fine Jul 31 19:00:42 I should have pastebinned the Linux boot error Jul 31 19:01:58 jow: different toolchains? Jul 31 19:01:59 jow: thanks for taking care of the release, you did a very good job Jul 31 19:02:45 could you tell in the release notes of 18.06 that openwrt 15.05 is EOL and will not get any support not even for severe security problems. Jul 31 19:02:56 Hauke: please update them Jul 31 19:03:00 wanted to announce tomorrow morning Jul 31 19:03:05 just do you have a link? Jul 31 19:03:52 jow: found it: https://openwrt.org/releases/18.06/notes-18.06.0 Jul 31 19:04:08 f00b4r0: hm yeah, will try x86 Jul 31 19:04:57 jow: the failure I reported happened after getting config.seed from downloads for ar71xx/mikrotik, running feeds update; feeds install -a; make world in 18.06.0 Jul 31 19:05:18 (+1 backport patch for rb931 which touches nothing in packages) Jul 31 19:06:35 i.e. trying to build valid release of 18.06, if I understand this correctly. Jul 31 19:08:15 forgot a "make defconfig" in-between, but ykwim :) Jul 31 19:08:57 https://pastebin.com/62kJcZ0J Jul 31 19:09:10 does the fact that SPI doesn't appear in the kernel bootlog suggest that that driver isn't loading? Jul 31 19:11:08 fahadsadah: or plain missing from the kernel Jul 31 19:14:51 * f00b4r0 notices that 18.06 appears to be a lil more memory hungry than 17.01 Jul 31 19:18:01 Thanks f00b4r0. that's curious as CONFIG_SPI=y; CONFIG_SPI_SUN6I=y; CONFIG_MTD_SPI_NOR=y; etc Jul 31 19:21:40 fahadsadah: alternatively the DT node is missing/malformed? Jul 31 19:21:50 I'm merely speculating at this point, I don't know the hardware. Jul 31 19:22:03 DT node for the flash* Jul 31 19:24:40 hm, no luck Jul 31 19:24:42 usbip builds Jul 31 19:27:00 ah ha. It's possible to distinguish the EX3700 from the EX3800 (they both share the same image) from the content of the config on the flash Jul 31 19:28:25 ha Jul 31 19:28:28 ah even Jul 31 19:29:21 istr rmilecki mentioned it would be possible to update the model name at runtime but I don't remember the specifics and can't find it in my irc logs Jul 31 19:30:25 f00b4r0: does it need to appear in aliases with spi0 = &spi0;? Jul 31 19:31:14 phew, usbip fixed Jul 31 19:31:37 fahadsadah: i don't know. Jul 31 19:32:19 Thanks for your help! :) Jul 31 20:35:40 any news on 18.06? Jul 31 20:36:14 https://openwrt.org/releases/18.06/notes-18.06.0 Jul 31 20:36:59 seen the page Jul 31 20:37:15 release has been postponed a number of times Jul 31 20:37:22 curious if that's happening again Jul 31 20:37:50 https://openwrt.org/releases/18.06/start#roadmap Jul 31 20:38:34 latest date for final was July 31st Jul 31 20:43:10 anon^_^: .. it's out Jul 31 20:44:24 Monkeh, thanks Jul 31 20:44:28 https://downloads.openwrt.org/releases/18.06.0/ Jul 31 20:44:30 no announce on the front page Jul 31 20:44:40 yeah just checked the release dir Jul 31 20:48:55 I need people who use Perl to take PR #6623 for a spin before I merge it…. there's a CVE that this fixes, but it's also a major version just (from 5.26.2 to 5.28.0) and I'd like to make sure nothing is broken… so if you can build it, install it, and test it and report back on your success, that would help. thanks Jul 31 21:09:07 hmm Jul 31 21:09:14 I have a bridge problem with 18.06 it seems Jul 31 21:10:45 philipp64: will do Jul 31 21:24:46 russell--: Thanks! Jul 31 21:28:22 nbd: regarding issue 6382… how do you tell procd you want to wait until the last (boot-time configured) interface is up? Jul 31 21:28:57 or if you have a boot-time configured IPsec tunnel that never comes up, what's to stop that from services waiting on such an event from never coming up? Jul 31 21:36:31 philipp64: there was an outstanding runtime problem on ppc (apm821xx) with the old perl that i've been lazily carrying a patch for, any idea if that has been fixed? Jul 31 21:37:32 it appeared at a gcc bump, involved adding some CFLAGS iirc Jul 31 21:58:30 not ringing any bells… if you send me error messages, I can ask on perl5-porters. Jul 31 22:01:35 wigyori: blogic: thx for taking care of SpaceRat Jul 31 22:02:10 Poor rat. Jul 31 22:02:38 twas diseased, for the best. Jul 31 22:39:29 philipp64: i have a patch that works (literally just adds "-fwrapv -fno-strict-aliasing" to -Dowrt:cflags in the Makefile) Jul 31 22:40:21 without that patch, on apm821xx i get segfaults. it might just be fixed in 2.8, i'll try it. Jul 31 22:43:46 philipp64: i am not seeing any problems on ar71xx in my limited use case Jul 31 22:53:24 russell--: good! that's something. did you file an upstream bug report for the apm821xx? Jul 31 22:58:57 what instruction set/processor family is the apm821xx? Jul 31 23:01:10 guessing PPC 405 or so. Jul 31 23:01:57 https://openwrt.org/toh/hwdata/netgear/netgear_wndr4700_v1 --> powerpc_464fp Jul 31 23:04:40 okay, thanks. odd that this would affect only Perl and not be required as part of the generic target cflags. Jul 31 23:05:19 jow: thanks for you response on the ubus command! I must have missed it earlier Jul 31 23:17:17 philipp64: confirmed the problem still exists with your perl 2.8 patch on apm821xx. https://bugs.openwrt.org/index.php?do=details&task_id=1464 Jul 31 23:37:08 russell--: if you build with -fwrapv but omit -fno-strict-aliasing … what happens then? Jul 31 23:37:26 And are there any other PPC or apm821xx boxes this could be tested on? Jul 31 23:41:25 russell--: http://thiemonagel.de/2010/01/signed-integer-overflow/ Jul 31 23:45:56 philipp64: i only have these meraki mr24s Jul 31 23:48:50 trying now without the -fno-strict-aliasing Jul 31 23:52:00 Hi people just flashed a snapshot to my c7-v2 and cant ssh to it the wifi is still up but I cant get any pings back from it over ethernet Jul 31 23:52:21 can you get a ssh client for android? Jul 31 23:52:39 sure. I use JuiceSSH. Jul 31 23:52:50 thanks dude Jul 31 23:53:11 I have no idea how accessible it is. it doesn't seem too concerned with purely visual stuff though. Jul 31 23:53:22 BTW never never flash a snapshot just befor you have to go to bed lol Jul 31 23:53:59 yup. rule one: never bring something online before you want to leave your desk. that's why I don't publish websites on friday. Jul 31 23:54:00 "have to"??? Jul 31 23:54:03 It's ok on that frunt as I am going to get my wife to log in to it and do the opkg update and install luci Jul 31 23:54:25 ssh client i use for android is called VX-ConnectBot Jul 31 23:54:43 VX like the poison gas? Jul 31 23:55:02 that seems ominous. Jul 31 23:55:04 DonkeyHotei: you're missing out, juicessh Jul 31 23:55:11 WTF VX-ConnectBot sounds like it was built by the ruskies and will mod all your shit to eat your face Jul 31 23:55:27 tapper: one of these days I will have to ask you how you end up at words like "frunt" Jul 31 23:55:46 Tapper: I mean ... autocorrect and all, but what dictionary is that? :) Jul 31 23:55:49 that's what it sounds like Jul 31 23:55:55 it's a fork of a client called ConnectBot which i uninstalled in its favor from my previous phone because i could not type a pipe character in it Jul 31 23:56:05 philipp64: without -fno-strict-aliasing i get the segfault Jul 31 23:56:10 spell like I talk Jul 31 23:56:24 russell--: interesting. thanks, that helps. Jul 31 23:56:42 Tapper: ah, so no dictionary at all, purely phonetic. interesting. Jul 31 23:56:47 so… with the -fno-strict-aliasing but without -fwrapv… Jul 31 23:57:02 I do wish I was a grate speller! I just make do with what I can do. Jul 31 23:57:10 takimata: not purely. he types "know" instead of "no" a lot Jul 31 23:57:43 I don't know wish one to use some times! Jul 31 23:57:44 philipp64: i'm one step ahead of you Jul 31 23:57:48 ssh and android is not really pretty, yes, it works, kind of... but it's everything but convenient Jul 31 23:57:49 building now Jul 31 23:57:58 DonkeyHotei: https://i.imgur.com/PHost3g.png - join the enlightened Jul 31 23:57:59 :D Jul 31 23:58:40 Tapper: oh, so you are typing that? I thought it would be some kind of voice recognition. Jul 31 23:58:45 i never got the hang of screen, but i've used tmux Jul 31 23:59:17 screen > tmux, because you can use it for serial consoles! Jul 31 23:59:50 DonkeyHotei: i meant the ssh client :p Jul 31 23:59:53 * takimata has used screen for 20+ years, not going to change now. Aug 01 00:00:19 ed rocks! Aug 01 00:01:11 i do all my editing on paper tape! Aug 01 00:02:00 So my c7 is working as a switch how would I find out it's IP it does not show up in the dhcp table on my mane router and when I plug in a cable to my computer form the lan on the c7 I don't get a IP on my computer Aug 01 00:02:01 I listen to podcasts on wire recordings. Aug 01 00:02:46 I download all my books from the audio book bay Aug 01 00:02:56 :-D Aug 01 00:04:20 Tapper: you can try to find it via packet captures (wireshark, tcpdump) or arp, if neither works - factory reset/ failsafe time Aug 01 00:04:21 I don't like podcasts much! People ramble to much Aug 01 00:05:26 I cant get the dumb thing to reset as I flashed a bootloader for over clocking and when i press the button it brings up the page for the bootloader. Aug 01 00:06:01 philipp64: no segfault with just -fno-strict-aliasing (no -fwrapv) Aug 01 00:06:15 I will have a go at ssh over wifi in the morning Aug 01 00:06:44 thanks for your help people and I am off to bed take care! Aug 01 00:08:28 you too. sleep well. Aug 01 00:12:06 hrm am i misunderstanding something about how DEVICE_PACKAGES is supposed to work? DEVICE_PACKAGES := kmod-usb-ohci kmod-usb2 -uboot-envtools | DEVICE_PACKAGES += swconfig | DEVICE_PACKAGES += swconfig Aug 01 00:12:26 DEVICE_PACKAGES := is inherited by the second 2 devices Aug 01 00:13:16 er Aug 01 00:13:23 one of them is += -swconfig Aug 01 00:13:45 * m4t digs Aug 01 00:14:51 basically im trying to do a shared stanza in generic-ubnt.mk that is utilized both by routerstation and routerstation pro Aug 01 00:15:10 routerstation has a dumb switch, rspro has an ar8366 or whatever Aug 01 00:15:37 *ar8236 Aug 01 00:24:37 russell--: thanks, that helps. Aug 01 00:32:46 russell--: can you pick up pprindeville/packages branch perl-5.28.0 and build perl from that? Aug 01 00:33:12 it's got this workaround: https://github.com/openwrt/packages/pull/6623/commits/c850aeff40e012d6b0e1099e0e254c70112258e9 Aug 01 00:36:02 oh, duh. TARGET_PER_DEVICE_ROOTFS lol. Aug 01 01:05:40 so, i found something that would make a good openwrt pkg that nobody has tackled, but upstream's build system is such an unmitigated disaster that he has to commit precompiled binaries of the end result into his github source tree: https://github.com/philippe44/AirConnect/blob/master/airupnp/Makefile Aug 01 01:09:03 it expects one to download a blessed but unstated snapshot of each of those other projects's source and just links directly against their internals Aug 01 01:09:47 huh Aug 01 01:10:08 i spent a while last night trying to find something that emulates a chromecast. so i could stream e.g. youtube from chrome to another device Aug 01 01:10:13 completely closed api :( Aug 01 01:10:33 this does the opposite -- it talks to a chromecast Aug 01 01:10:35 easy to cast, receiving is impossible it seems Aug 01 01:10:37 yeah Aug 01 01:10:56 imma just try pulseaudio or something i guess, idk if i even care enough Aug 01 01:11:46 so ath79 selects ath9k by default - in adding support for routerstation, it shipped as a bare board. i'm thinking SUPPORTED_PACKAGES := -kmod-ath9k -wpad-mini might be correct? Aug 01 01:12:04 i use ath9k personally but i know some people who dont use any wifi with it Aug 01 01:12:29 but i don't see anything else that emulates an airport express without requiring a separate server Aug 01 01:13:34 what build system does it use btw? Aug 01 01:14:13 something called Embarcadero C++ which does not use makefiles Aug 01 01:14:25 oh sounds kinda mesy yeah Aug 01 01:14:53 he *has* makefiles, but i linked directly to how messy that is Aug 01 01:16:39 hmm so it's some kind of IDE Aug 01 01:16:50 yes Aug 01 01:44:42 probably (ex-Borland) delphi? (the successor or tubro pascal) Aug 01 02:29:32 m4t: Yes, the google cast protocol has some parts open and others closed. The cast applications and devices use mutual authentication (I think is certificate based) to maintain control of who can play/stream content (DRM?) Aug 01 02:29:49 drm++ Aug 01 02:30:01 IMO that could be done with an open protocol too Aug 01 02:30:16 on non-chromecast devices, it is Aug 01 02:32:17 There are some efforts to reverse-engineer the protocol, but without a proper certificate signed by google they will fail Aug 01 02:32:39 luaraneda: you don't like security by obscurity? Aug 01 02:33:02 so you need a cert to cast. lovely. Aug 01 02:34:17 there was this thing https://github.com/dz0ny/leapcast Aug 01 02:35:41 philipp64: No, not at all, I believe the opposite actually. Open protocols provide better benefits on the long run Aug 01 02:36:07 at least apple's device can still be emulated Aug 01 02:36:59 I was being facetious… clever people invariable reverse engineer or disassemble code and figure it out. it only gives you a modest head start. Aug 01 02:46:25 Yeah, that's exactly why I don't like closed protocols, someone will eventually reverse engineer it. Now we have wasted time twice, to close it and the reverse-engineer it... Aug 01 02:46:46 s/the/then **** ENDING LOGGING AT Wed Aug 01 03:00:00 2018