**** BEGIN LOGGING AT Thu Jun 28 03:00:02 2018 Jun 28 03:13:03 dagnabbit, i was probably hosting doom games at milliways Jun 28 03:32:50 build #39 of rb532/generic is complete: Success [build successful] Build details are at http://release-builds.openwrt.org/18.06/images/builders/rb532%2Fgeneric/builds/39 Jun 28 05:09:03 pie: what! I missed a doom lan party? how disappointing. Jun 28 06:17:59 Hi jow It's just popped in to my head that I should remind you to not forget to put a line in the announcement for 18.0.6 that small changes were made to the user interface for better support for screen readers. Jun 28 06:18:58 the changes are not yet in 18.06 Jun 28 06:21:23 lined up for rc2? Jun 28 06:22:00 if one placed his default config under files directory in buildroot and flashed the image.. can he modify the rom (say change wireless SSID) so that a "factory reset" would return the device to the modified state not the original flashed state? Jun 28 06:36:06 *yawn* Jun 28 06:57:08 blogic:ping Jun 28 06:59:29 dedeckeh: hi Jun 28 08:00:14 ldir: ping Jun 28 08:33:36 Can anyone backport it to 18.06 branch, please ? https://github.com/openwrt/packages/pull/6062 Jun 28 08:34:07 or its freezed ? Jun 28 08:38:45 Would it make sense to document ubus calls generally with something like doxygen, so it's automatically updated, instead of manually updating the wiki? Jun 28 08:51:30 using the latest imagebuilder for ar71xx/tiny, I'm getting " * opkg_download: Failed to download http://downloads.openwrt.org/snapshots/targets/ar71xx/tiny/packages/base-files_191-r7336-d898113_mips_24kc.ipk, wget returned 8." when trying to build an image Jun 28 08:52:09 I'm guessing whatever script populates the snapshot directory with new builds failed to put everything in its place, as the base-files package /is/ listed in /snapshots/targets/ar71xx/tiny/packages/Packages, but the file itself is missing Jun 28 08:52:23 jspiros: seems the upload failed Jun 28 09:32:11 jow: any idea when it should run again? Jun 28 09:36:12 muhaha: let's wait for hnyman to see if he replies... that's why packages should have maintainers in general :) Jun 28 09:38:41 jspiros: 8-12 hours Jun 28 09:39:13 hi, looking at https://openwrt.org/toh/start?dataflt%5BBrand*~%5D=tp-link&dataflt%5BModel*~%5D=1043 it shows ath10k driver for all versions. shouldn't that be ath9k? Jun 28 09:40:39 rmilecki Jun 28 09:40:39 rmilecki: do not know how is release possible without this commit :D like we will disable usbip package and release a 18.06 stable? :] Jun 28 09:42:37 philipp64: https://patchwork.ozlabs.org/cover/881776/ Jun 28 09:43:25 muhaha: it could happen Jun 28 09:43:30 of course would be bad Jun 28 09:44:03 but commiters can't possibly handle all the problems with extra software Jun 28 09:44:19 i believe that's why packages were moved to the github in the first place Jun 28 09:44:41 i never used usbip, so I don't really want to push/cherry-pick that Jun 28 09:44:52 i'd prefer someone who knows it, can test it, to do it Jun 28 09:45:29 seems that snapshots were builded also without usbip package I guess Jun 28 09:48:50 . Jun 28 09:48:51 muhaha: it's almost like usbip is, has been, and likely will continue to be problematic and often broken, even upstream Jun 28 09:52:27 are you disabling it, please? CONFIG_PACKAGE_kmod-usbip=n CONFIG_PACKAGE_kmod-usbip-server=n CONFIG_PACKAGE_kmod-usbip-client=n does not work for me. Jun 28 09:52:30 *How Jun 28 09:52:40 # CONFIG_PACKAGE_... is not set Jun 28 09:53:04 Is it difference? Jun 28 10:04:07 use menuconfig if you don't know how to edit .config files directly. Jun 28 10:04:20 there's a line in the top of the file.... # Automatically generated file; DO NOT EDIT. Jun 28 10:09:04 I am using make defconfig ... So changes are perserved. I asked what is difference between #CONFIG_PACKAGE_kmod-usbip is not set and CONFIG_PACKAGE_kmod-usbip=n Jun 28 10:10:01 the former is the disabling of a symbol Jun 28 10:10:08 the latter is invalid syntax Jun 28 10:10:20 it will likely get ignored and repalced with defaults Jun 28 10:11:24 ah.. did not know. It tried it in kernel .config -> works ok Jun 28 10:12:26 as jow said, it was probably ignored, and you got the defaults, which was not enabled.... Jun 28 10:12:30 doesn't mean it "worked" Jun 28 10:15:03 but.... Then my builds fail to hunk failed in usbip package... Weird. I am using x86. Jun 28 10:16:14 I said that have tried to write a syntax CONFIG_=n in kernel .config -> works ok. So I tried to apply also in general .config Jun 28 10:16:22 to add to my previous question, are there any 802.11g/n chipsets that are ath10k supported? Jun 28 10:16:32 muhaha: don't use =n Jun 28 10:17:02 openwrt buildroot interprets .config directly and many make constructs will use expressions like $(if $(CONFIG_foo),do stuff) Jun 28 10:17:07 no no, they know better :) Jun 28 10:17:15 which evaluates to true for "=n" as well Jun 28 10:18:38 I kinda feel it's lazy and error prone scripting that people relied on that sort of thing, but can't change where we are now of course. Jun 28 10:18:40 the only valid use of =n is as seed input directly before running "make defconfig" as kconf will treat both "# CONFIG_FOO is not set" and "CONFIG_FOO=n" as equivalent and normalize it to "# CONFIG_FOO is not set" Jun 28 10:19:04 ah, thanks for clarification. Like I said before: I am not a developer, its really hard for my to get all necesary knowledge about openwrt build system. I just want to finally compile my own image :D Jun 28 10:19:32 karlp: its not just a matter of lazyness. A lot of inlien conditional expressions would be rather complex if you need to perform equality checks Jun 28 10:20:17 well, that what I am doing..... edit seed. use CONFIG_PACKAGE_kmod-usbip=n -> make defconfig Jun 28 10:25:09 jow: perhaps, make has pretty good functions for lists too though, I'm not really convinced it couldn't be done better, but it's a piddlingly small improvement to try and go back and do it now, I'm certainly not advocating that. Jun 28 10:26:15 sure, you could rewrite any $(if $(CONFIG_FOO),...) into $(if $(filter-out n,$(CONFIG_FOO)),...) for starters Jun 28 10:26:24 should be even possible semi-automaticly Jun 28 10:27:10 to be perfectly kcon compliant it should be the reverse though Jun 28 10:27:23 $(if $(filter m y,$(CONFIG_FOO)),...) Jun 28 10:31:11 you also need to instruct all package maintainers in all feeds to never ever test using $(if $(CONFIG_FOO),...) because someone *might* use acceptable but non.canonical .config format Jun 28 10:31:38 * f00b4r0 cues in 'Highway to Hell' ;P Jun 28 10:47:58 jow: does make defconfig perform any network operation? Jun 28 10:48:49 hello can someone help me in flashing my mikrotik b951g, I Jun 28 10:51:54 should I ask on #openwrt? https://openwrt.org/contact#irc_channels shows only #openwrt-devel and #openwrt-adm Jun 28 11:35:01 aversario: what do you mean by chipsets ? 802.11n is available in Z390 which is not releases afaik Jun 28 11:35:01 possibly insecure services should run as their own users on openwrt right? Can I just set the param in the procd file and the user is automatically created? **** BEGIN LOGGING AT Thu Jun 28 11:35:20 2018 Jun 28 11:45:30 will be rc2, or final release? Jun 28 12:27:46 jow: please consider https://github.com/openwrt/openwrt/pull/1126 and eventually to cherry-pick it for openwrt-18.06 -- if its two dependencies are not backported to 18.06, everything is fine as it is - however if they are (well, would be), this pull request should go along with them. Jun 28 12:31:14 it's needed for master, but -in its current state- not yet necessary for openwrt-18.06 (only if the ath10k LED patch actually gets enabled by the two follow-up patches from mkresin and stintel) Jun 28 12:31:43 I'm neither touching ipq nor kernel stuff Jun 28 12:31:51 o.k., then it's all fine Jun 28 12:32:33 I was just concerned that the two follow-up patches might end up as last minute backports Jun 28 12:32:44 jow: ping? Jun 28 12:32:53 f00b4r0: pong Jun 28 12:33:05 jow: did you see my messages? Jun 28 12:33:16 yes Jun 28 12:33:21 ok. Jun 28 12:33:27 something appears to have broken defconfig Jun 28 12:33:38 it went from a few seconds to about 3mn Jun 28 12:33:56 I hope that's not intended :) Jun 28 12:34:14 did not touch anything, dunno Jun 28 12:34:22 http://phase1.builds.lede-project.org/builders/pistachio%2Fgeneric/builds/781 before Jun 28 12:34:27 http://phase1.builds.lede-project.org/builders/pistachio%2Fgeneric/builds/782 after Jun 28 12:34:36 it's like that on every build Jun 28 12:35:00 no idea Jun 28 12:35:12 did not touch anything even remotely related to that Jun 28 12:35:33 jow: the only logical explanation is that one of the commits in between causes this Jun 28 12:35:38 but I can't figure out which Jun 28 12:36:19 a few commits touch directly the build files, the others appear to be unrelated, but none strike me as obvious suspects Jun 28 12:39:57 different builders, different tree modes Jun 28 12:40:12 clean feeds vs. updated feeds Jun 28 12:40:15 rmadisk vs. rust Jun 28 12:40:33 defconfig scans package metadata in background Jun 28 12:40:48 (prepare-tmpinfo target in buildroot) Jun 28 12:42:34 it's very consistent. Not a single recent build went under about 2mn Jun 28 12:42:39 pkgadd: I'm fine to merge PR#1126 to master, but I don't feel comfortable to backport the ath10 leds stuff to 18.06 now that we have release candidate Jun 28 12:42:43 which seems extremely slow Jun 28 12:43:04 I can imagine that the new git clean tree stuff causes this Jun 28 12:43:09 feeds are now recloned on every build Jun 28 12:43:17 which means all pacakge metadata is rescanned on every build Jun 28 12:43:27 but idk, can't investigate that right now Jun 28 12:43:29 build 781 has the full clean (with -x) Jun 28 12:43:32 and defconfig took 12s Jun 28 12:43:35 pkgadd: I'm the opinion, ony bugfixes should go into release branches as soon they are declared as rc1 Jun 28 12:43:58 i asked if defconfig does network things because three commits have changed download.mk Jun 28 12:44:03 it does not Jun 28 12:44:08 mkresin: totally agreed, I was just concerned about the the other two patches (which would have enabled the white LEDs in addition to the amber ones) going into the openwrt-18.06 branch late/ alone Jun 28 12:44:15 super weird Jun 28 12:44:47 mkresin: now that they'll won't be cherry-picked, everything is fine Jun 28 12:45:51 hmm Jun 28 12:46:17 i realize that the 'culprit' (if any) isn't necessarily in the source git, but can come from the feeds Jun 28 14:32:23 how clever do you need to be to port ar71xx devices? Jun 28 14:33:00 fairly Jun 28 14:33:07 :( Jun 28 14:33:38 nevermind then :D Jun 28 14:34:53 given all the a tivity on ath79, people seem to really like atheros devices Jun 28 14:35:03 heh Jun 28 14:35:05 theres lots of them Jun 28 14:35:25 just booted up snapshot on this rb2011, not sure if its working properly Jun 28 14:35:42 seems to detect both switches but only see swconfig vlans for one Jun 28 14:35:49 or is that just config? Jun 28 14:36:03 needs more DSA :D Jun 28 14:37:12 dsa won't come until the qca8k driver gets fixed Jun 28 14:37:24 oh? Jun 28 14:37:39 who broke it Jun 28 14:37:57 it never worked Jun 28 14:38:00 ah Jun 28 14:38:37 thats a shame Jun 28 14:39:27 seems to work ok on a71xx and swconfig though, so that will do for now Jun 28 14:54:39 [ 33%] Linking C executable rtl_power Jun 28 14:54:39 CMakeFiles/rtl_power.dir/rtl_power.c.o: In function `fix_fft': Jun 28 14:54:39 rtl_power.c:(.text+0xc16): undefined reference to `FIX_MPY' Jun 28 14:54:44 dammit Jun 28 14:54:48 that's a compiler bug, I think Jun 28 15:02:33 or just cmake cross compiler failure Jun 28 15:06:49 no Jun 28 15:06:54 inline int16_t FIX_MPY(int16_t a, int16_t b) Jun 28 15:07:01 with -Os it just doesn't exist Jun 28 15:07:04 it's not "extern inline" Jun 28 15:07:18 surely the compiler should emit it, if it decides not to inline it? Jun 28 15:12:33 no, it's a code bug Jun 28 15:12:46 just 'inline' still lets the compiler assume there's an extern version it can call Jun 28 15:12:48 and there isn't Jun 28 15:16:22 those openssl vulns, will there be updated 17.01 packs or will we have to wait for 1806? Jun 28 15:16:58 inline is just a hint anyway. Jun 28 15:21:08 hm Jun 28 15:21:35 wonder why I'm getting no carrier on the second switch Jun 28 15:21:56 blogic: do you remember if the gige switch in the rb2011 was ever fixed? Jun 28 15:22:11 or maybe its the fast ethernet one actually Jun 28 15:22:41 what number do openwrt patches start with, for patches submitted upstream? Jun 28 15:22:51 1? Jun 28 15:29:41 hm Jun 28 15:29:55 link is up, maybe it only shows carrier if at least one switch port is up? Jun 28 16:01:51 karlp: yeah, I don't blame GCC any more Jun 28 16:01:55 it's purely a rtl-sdr bug Jun 28 16:02:15 although nobody in their right mind would compile rtl-sdr with -Os so nobody's noticed before :) Jun 28 16:04:31 dwmw2_gone: it used to be 0xx for upsteam Jun 28 16:04:39 dwmw2_gone: for kernel now you have a separated directory Jun 28 16:04:41 called backports Jun 28 16:04:52 0xx for already upstream, 1xx for submitted? Jun 28 16:04:56 dwmw2_gone: target/linux/generic/PATCHES Jun 28 16:05:03 yes Jun 28 16:05:04 this is rtl-sdr Jun 28 16:05:12 https://github.com/openwrt/packages/pull/6377#issuecomment-401084671 Jun 28 16:05:50 i don't think many *packages* follow that schema Jun 28 16:05:57 but it sounds like a nice idea probably Jun 28 16:06:06 I was heckled when I didn't :) Jun 28 16:07:26 ok, train home... Jun 28 16:08:42 * karlp waves Jun 28 16:19:53 hm Jun 28 16:20:12 not seeing any traffic on the built in ar9344 switch on this routerboard :( Jun 28 16:20:59 is it supposed to be forced up at a gig? (its an fe switch afaik?) Jun 28 16:32:16 oh hm, works if I disable vlan Jun 28 16:36:14 odd Jun 28 16:48:42 for iptables if I do not put -p tcp|udp will it default to both ? or will it make its own decisions(e.g. port 22 will always be -p tcp if no -p used)? Jun 28 16:48:59 trying to reduce iptable rules. Jun 28 16:49:31 specifically, do i have to redirect 53 to 5353 by two lines with -p tcp and -p udp each Jun 28 16:49:59 if i don't put -p i don't want it enforce udp only, as 53 could be done via tcp too though not default Jun 28 16:56:26 find the answer, must specify -p Jun 28 17:16:17 hi, looking at https://openwrt.org/toh/start?dataflt%5BBrand*~%5D=tp-link&dataflt%5BModel*~%5D=1043 it shows ath10k driver for all versions. shouldn't that be ath9k? Jun 28 17:19:15 your link doesn't show what you think it might... Jun 28 17:20:13 where does it say ath10k? Jun 28 17:33:35 the link will show you every version of 1043. open some and you'll find section driver -> ath10k Jun 28 17:34:37 *section Wlan driver Jun 28 17:34:55 so.. next thing in my tree, http://ix.io/1e54 Jun 28 17:34:59 is that just horrible? Jun 28 17:35:31 ignore broken indents Jun 28 17:38:21 aversario: on what? device page? device techdata? what? Jun 28 17:39:06 your description was wayyyy to broad, I presume you're complaining that for isntance this page: https://openwrt.org/toh/hwdata/tp-link/tp-link_tl-wr1043nd_v2 says ath10k ? Jun 28 17:48:22 karlp: yes, you linked the perfect example and when you do a search for ath10k on that page, it gives you only 1 hit Jun 28 17:49:13 I only want to know if that is correct, otherwise I will change it to ath9k Jun 28 17:50:23 I haven't really heard of 802.11g/n chipset using ath10k Jun 28 17:57:55 * ldir suggests no one expect any sense from this direction for at least the next 2 hours. Jun 28 18:01:22 aversario: QCA99xx series used ath10k for b/g/n Jun 28 18:01:24 *use Jun 28 18:13:46 hi Jun 28 18:14:17 hi yall im comiling openwrt from scratch for my device and id like to add a device driver, i have this https://gist.github.com/Ins1ghtLabs/c346f7ed4f257d1b4a2d Jun 28 18:21:11 mangix: ok, nice to know that, but QCA9563 most certainly uses ath9k Jun 28 18:21:23 yes Jun 28 18:23:00 https://gist.github.com/Ins1ghtLabs/c346f7ed4f257d1b4a2d Jun 28 18:23:04 thats the file Jun 28 18:24:18 ldir: not happening anyway heh Jun 28 18:28:29 stintel: wasn't aware that the previous set had collateral damage... thanks for the heads up! looks good. Jun 28 18:29:00 so im trying to add rtl8812au suport to my version of openwrt Jun 28 18:29:05 i dont see it in the menuconfig Jun 28 18:29:22 just one more and then I'll quit: is here channel #openwrt missing? -> https://openwrt.org/contact#irc_channels Jun 28 18:30:11 hi Jun 28 18:30:50 hi Brother_Lal Jun 28 18:30:52 stintel: for what it's worth, though, I wouldn't be surprised if the VMware emulation itself isn't broken. when I was at Brocade (Vyatta) we used VMware a lot for testing and had a lot of false positives with VMware... they were always the last to pick up patches for their drivers... KVM was surprisingly quick at patching things, and that's not a commercial product. Go figure. Jun 28 18:30:53 Borromini, Jun 28 18:31:45 hi Jun 28 18:32:36 "Some tracing shows that after e1000e_trigger_lsc() is called, ICR reads out as 0x0 in e1000_msix_other() on emulated e1000e devices. In comparison, on real e1000e 82574 hardware, icr=0x80000004 (_INT_ASSERTED | _LSC) in the same situation." Jun 28 18:32:49 Yup, it's a bug in VMware. Jun 28 18:35:32 "The 82574 specification update errata 12 states that [...]" wow! No wonder that so many people hate this chipset. Jun 28 18:37:49 rofl Jun 28 18:39:54 stintel: i have a support question Jun 28 18:40:25 philipp64: The 82574 is literally the worst LAN chip that Intel ever made. Some people claim that they are solid, but I don't believe them. :P Jun 28 18:40:40 philipp64: Although 12 errata takes the cake. I only remember counting 6 and lately, I have only been able to find 3 or 4. Jun 28 18:43:59 but only realtek makes crap and anything intel produces is A-OK. Jun 28 18:58:30 philipp64: emulating real devices in VMs is crazy :D Jun 28 18:58:33 virtio or vmxnet Jun 28 18:58:33 Borromini: lol Jun 28 18:59:06 football! Jun 28 18:59:11 barely Jun 28 18:59:12 the stuff intel makes is question begging Jun 28 18:59:35 that new 28-core cpu they showed off was epic fail Jun 28 18:59:52 what happened? Jun 28 19:00:13 they used a 1 kilowatt chiller to raise the clock speed to 5ghz Jun 28 19:00:16 Borromini: realtek make some good stuff now :D Jun 28 19:00:16 and told nobody Jun 28 19:01:05 also, nobody knows the termperature of the CPU. current guesses are -20C Jun 28 19:01:06 mangix: yeah. Jun 28 19:01:15 i was, obviously, kidding. Jun 28 19:01:25 or trolling, although i do a weak troll unfortunately :( Jun 28 19:01:39 i think intel was trolling Jun 28 19:01:50 yeah in the end big corporate trolls us all. Jun 28 19:02:01 laughing all the way to the bank, those fucks. Jun 28 19:02:26 'oh so you want *four* cores in your laptop? haha. wait till AMD *maybe* puts something on the market that can make it worth your while Jun 28 19:02:34 until then, ka-ching Jun 28 19:03:30 Borromini: The 82574 is an exception to Intel making good things. Jun 28 19:08:39 philipp64: Only 12? That's puny. Jun 28 19:12:25 well, it's had a *lot* longer production run than a lot of other chips... that's part of the problem. when it started to be clear how many issues there were with it, someone should have EOP'd it. Jun 28 19:12:41 Too many designs relying on it Jun 28 19:12:56 there are compatible replacements... Jun 28 19:13:09 Pin, strap, and firmware compatible? Jun 28 19:14:02 which firmware? the BIOS driver level? yes. the configuration block that BIOS pushes to the EEPROM? maybe not, but that's trivial to change. Jun 28 19:14:41 Maybe, but customers get grumpy :) Jun 28 19:15:35 the smarter ones, maybe. I was dealing with Lanner and they hadn't updated BIOS or the EEPROM block in 5 years despite being notified of production issues several times by Intel... Jun 28 19:15:45 heh Jun 28 19:15:51 But yeah, are there pin compatible replacements? Jun 28 19:15:56 we had to threaten to cancel a contract to get them to update. Jun 28 19:16:04 Most are very reluctant to respin boards Jun 28 19:16:48 yeah, because the massive hours of simulation to validate the new board... but usually the Intel reference designs are pre-validated. Jun 28 19:17:20 stick to those, remove only the components that they say you can safely remove, and you're okay. start removing components without termination, and you can run into trouble... Jun 28 19:18:02 Ah, yes, but the one tenth cent per board must be saved. Jun 28 19:18:43 Ethernet's slow and robust enough it's hard to properly break it Jun 28 19:18:56 (okay, gigabit is - 10GigE's touchier, let's not speak of faster) Jun 28 19:21:58 Anyway, people love to hate on that 82574, but it still works adequately most of the time Jun 28 19:22:05 And it's beginning to go away. Jun 28 19:22:06 Holy crap, they've pushed the Hartwell discontinuance back 7 quarters to Q1 2020. That's antisocial. Jun 28 19:23:06 Good grief, it's already a decade old Jun 28 19:23:08 Wait until Q4 2019 when they push it out to Q4 2022... Jun 28 19:23:11 I know it's popular.. Jun 28 19:24:16 My desktop has a "Intel Corporation 82574L Gigabit Network Connection". It never seems to have caused a problem for me. Jun 28 19:24:43 The i350 and the 82576 are better chips... Jun 28 19:25:31 even the i210 is better. Jun 28 19:25:53 My router has 2x i210 and an i354. Jun 28 19:26:31 The number of MSI interrupts in /proc/interrupts from those cards is nothing short of crazy. Jun 28 19:26:37 yeah... my build server has an i350, 2x i210, and 1x X550. Jun 28 19:28:24 to answer your question, I think the i82583 is highly compatible with the 82574... but it's PCI 1.0 and not 1.1. Jun 28 19:31:20 I have always found it weird which of the Intel controllers are e1000e and which are igb. For example, the i210 is igb but the i219 is e1000e. Jun 28 19:33:04 philipp64: Well, yeah, they're newer chips, I'd imagine they're better Jun 28 19:34:01 Although I figured the 82574 and the 82576 were related, but I guess they work in mysterious ways Jun 28 19:48:42 ldir: ping Jun 28 19:54:24 yay we won <3 Jun 28 19:54:32 ldir: sorry mate. better luck next time :^) Jun 28 19:55:48 I wondered why it was so quiet. Jun 28 19:56:35 hehe Jun 28 19:56:48 hi id like to build a kernel module for an external rtl device into the firmware im compiling. How do i go about adding it, as its not in the menuconfig. I can compile the module seperately but how do i chose a platform Jun 28 19:57:28 Well the football was crap! Jun 28 19:58:08 Borromini well dun, but we did not make it hard did we! :-D Jun 28 19:58:26 Tapper: no offense but we didn't field our a team either :P Jun 28 19:58:33 Isn't that the definition of football? Jun 28 19:59:27 True Jun 28 20:00:40 Joking aside our second eleven are an average bunch. Perhaps this will serve as a reality check to those that were getting way to carried away after the Panama win. Jun 28 20:01:01 There is no reality checking football fans Jun 28 20:01:01 below average Jun 28 20:01:06 :D Jun 28 20:01:09 it seems the platform is a mips24k but i dont see that in config_platform Jun 28 20:02:25 Tapper: england did score more easily against panama than we did, at least Jun 28 20:02:50 i'm a bit worried about our defense. scoring seems to work, but locking things down is something else Jun 28 20:23:12 muhaha: pong Jun 28 20:26:57 im adding an entirely new kernel module to openwrt, some help would be nice Jun 28 20:32:39 ldir: Did you look at the gpio-nct5104d patch, please? Thanks Jun 28 20:33:49 CONFIG_PLATFORM_MIPS_AR9132 does that cover the atheros 9331 Jun 28 20:36:47 rpifan: you should check other modules to see how they are picked up Jun 28 20:37:07 some of them have some exclusion statements to make sure it doesn't show up on unsupported architectures e.g. Jun 28 20:37:43 ldir: congrats on the promotion Jun 28 20:37:52 now that you're a dev i'll be nice to you ;) Jun 28 20:40:31 thanks Borromini its been hell of a time i just want to add rtl8812au to the glinet1500 Jun 28 20:42:14 you could use this as a template e.g. https://git.openwrt.org/?p=openwrt/openwrt.git;a=blob;f=package/kernel/kmod-sched-cake/Makefile;h=ad49bdaa19dca1c9c62a72aea262172a22c0e292;hb=HEAD Jun 28 20:45:08 good luck, i gotta go. Jun 28 20:51:40 muhaha: Not yet - I've had some work come in so Sunday probably earliest I can look. Jun 28 20:52:35 muhaha: I was sure you had pointed at a previous build/patch/PR/something that contained the original tweak. Jun 28 20:59:16 ldir: this https://forum.lede-project.org/t/pc-engines-apu2-leds-not-working-in-lede-17-01/12722/19 ? Jun 28 21:02:21 Yes I think so. Will look at that as a reference. Jun 28 21:05:28 mm, is there a simple example of ar71xx to ath79 i can look at to try and get an idea about what needs to be done? Jun 28 21:06:06 helpful if its ar9334 Jun 28 21:06:27 9344 even Jun 28 21:30:09 jwh: the only currently supported ar9344 device is the tl-wdr3600/ tl-wdr4300, which is working very nicely Jun 28 21:30:30 well, at least as far as ath79 support goes Jun 28 21:31:26 https://forum.lede-project.org/t/adding-rtl8812au-support-to-glinet150m/16273 Jun 28 21:31:33 i made a forum post to make it simpler i guess Jun 28 21:31:43 pkgadd: ta Jun 28 21:34:24 now I just need to trawl commit log for the relevant ones :D Jun 28 21:35:05 not really, you just need to look at the DTS Jun 28 21:36:00 target/linux/ath79/dts/ar9344.dtsi target/linux/ath79/dts/ar9344_tl-wdr4300.dtsi target/linux/ath79/dts/ar9344_tl-wdr4300.dts Jun 28 21:36:03 is that all it needs? Jun 28 21:36:19 and a board definition I guess? Jun 28 21:36:27 that's more or less the ar9344 specifics Jun 28 21:36:32 yes Jun 28 21:37:08 guess I can give that a go Jun 28 21:37:39 if you need further driver changes is another question., but there the git log wouldn't help you that much either Jun 28 21:37:47 yeah Jun 28 21:38:30 it *looks* on the surface to be a generic 9344, except theres an ar8327 on one of the ports (probably need leds and other gpio stuff too) Jun 28 21:38:50 its an rb2011 Jun 28 21:39:01 quite a lot, if not most, of the driver support is already there, you 'just' need to select and configure it via the DTS (and that's not trivial) Jun 28 21:39:08 yeah Jun 28 21:39:21 I get the idea behind dts but I don't really understand anything about hardware heh Jun 28 21:39:25 not how buses work etc Jun 28 21:40:07 unfortunately you can't extract all the necessary information from the existing mach files Jun 28 21:40:30 :( Jun 28 21:41:07 some you can, some you need to determine from the running system, some you need to check the data sheets, some you need to just try (serial console and initramfs images pretty much necessary) Jun 28 21:41:07 well, I have a spare board so I can at least have a fiddle, see how broken I can make it Jun 28 21:41:37 kinda want 4.14 heh Jun 28 21:41:51 the good part, I can confirm that ar9344 SOC support is fully working on my tl-wdr3600 and tl-wdr4300, http://paste.debian.net/hidden/a49524c9/ Jun 28 21:42:22 offload, DSA would be nice at some point (when the driver works I guess) Jun 28 21:42:37 the cpu is pretty weak, so offload would be a huge help Jun 28 21:43:24 that and version parity across all devices heh Jun 28 21:44:38 actually I can go down next week and take the spare board to the DC Jun 28 21:44:52 for me personally, the major reason was being able to forget about kernel 4.9 ;) whenever I build an image, I do bump it to the most current -stable release (if master is lagging a bit) - that's easier to do if you can concentrate on a single kernel, rather than two (and yes, that's also the reason why I'm using the ar71xx/ 4.14 bump on my two remaining ar71xx devices, tl-wr941nd v2 and Jun 28 21:44:58 tl-wr1043nd v1) Jun 28 21:45:06 yeah Jun 28 21:45:57 well especially since these are all out in the field, would be nice to not have to remember that some devices are running older kernels Jun 28 21:46:08 or more importantly, code for that case heh Jun 28 21:48:33 rb2011 brings me up to 5 active targets Jun 28 21:50:19 at least theres no wifi, that makes it abit simpler heh Jun 28 21:54:16 the one I really care about are ipq806x, lantiq and ath79/ ar71xx, brcm47xx for some really legacy devices that aren't in daily use anymore Jun 28 21:55:35 yeah Jun 28 21:56:05 the ones I have in active use are.. x86, octeon, ramips, lantiq and ar71xx (hopefully ath79 eventually) Jun 28 21:56:29 the only problem one in that list is ar71xx Jun 28 21:56:59 mt7621/ mt7622 is something I'd like to add to my zoo, seems to be rather nice Jun 28 21:57:12 (and mvebu) Jun 28 21:57:27 yeah, mvebu is one I'm also interested in Jun 28 21:57:42 but haven't convinced the finance people to buy e one yet Jun 28 21:57:44 me* Jun 28 21:58:40 hehe, same here (even if I just need to talk to my own wallet), it doesn't help if I don't really /need/ yet another device, but merely would like to look into it Jun 28 21:59:01 :D Jun 28 21:59:17 I'm well out of my depth here, sigh Jun 28 22:00:53 amusingly its actually the board/target definitions that I'm stuck on :D Jun 28 22:28:31 yeah I suck, who do I send money/booze/women to? Jun 28 22:45:08 philipp64: thanks Jun 28 22:48:51 bugs.openwrt.org is unbelievably slow for me right now... Jun 28 22:49:15 hu. 404 when trying to add a new report Jun 28 22:50:47 26368 pts/0 S+ 0:00 | \_ make V=s defconfig Jun 28 22:50:47 26464 pts/0 S+ 0:00 | \_ /bin/sh -c export MAKEFLAGS= ;make V=ss -j1 -r -s -f include/scan.mk SCAN_TARGET="packageinfo" SCAN_DIR="package" SCAN_NAME="package" SCAN_DEPS="/home/varenet/source/include/pac Jun 28 22:50:47 26465 pts/0 S+ 0:00 | \_ make V=ss -j1 -r -s -f include/scan.mk SCAN_TARGET=packageinfo SCAN_DIR=package SCAN_NAME=package SCAN_DEPS=/home/varenet/source/include/package*.mk SCAN_DEPTH=5 SCAN_EXTRA= Jun 28 22:50:48 26806 pts/0 S+ 0:00 | \_ /bin/sh -c { ?printf "\033[M\rCollecting package info: package/network/utils/arptables" >&2; echo Source-Makefile: package/network/utils/arptables/Makefile; true; make V Jun 28 22:50:52 26807 pts/0 S+ 0:00 | \_ make V=ss --no-print-dir -r DUMP=1 FEED= -C package/network/utils/arptables Jun 28 22:50:55 26809 pts/0 R+ 0:00 | \_ python /home/varenet/source/scripts/download.py dl_method --url https://git.netfilter.org/arptables --proto=git Jun 28 22:50:58 was trying to point out that 75ab064d2b38e70746af1718ed7cdbafb906249e is the cause of super slow 'make defconfig'. Jun 28 22:51:01 oops Jun 28 22:51:03 sorry Jun 28 23:33:18 https://forum.lede-project.org/t/adding-rtl8812au-support-to-glinet150m/16273/2 Jun 28 23:33:20 can anyone help Jun 28 23:35:36 does it build now? Jun 28 23:38:11 rpifan, any reason you picked that driver? Jun 28 23:39:27 that specific repo Jun 28 23:42:14 not really, i just need to build rtl8812au for that specific device and i coulldnt find a better source Jun 28 23:48:01 I'm biased but you should give a try to https://github.com/aircrack-ng/rtl8812au/ Jun 28 23:48:24 Mister_X, but how do i add it / make it for the my gli_net device Jun 28 23:49:00 you'll have to follow the compiling instructions for openwrt Jun 28 23:49:22 I'm not familiar with it Jun 28 23:49:57 you dont do openwrt dev work? Jun 28 23:50:49 no Jun 28 23:51:34 oh Jun 28 23:51:51 well i guess the bigger issue is how do i know what kind of platform i need ot compile for Jun 28 23:51:53 my guess is that you just have to cross compile the driver Jun 28 23:52:18 well, you have to target one and you need the compiler to compile and libs to link to Jun 28 23:52:39 you need the specific compiler Jun 28 23:52:52 what im saying Jun 28 23:52:53 is Jun 28 23:52:58 i dont what if it is mips Jun 28 23:53:01 or arm Jun 28 23:53:03 or x86 Jun 28 23:53:06 and which kind of mips etc Jun 28 23:53:30 you need to have the compiler and libs specific to that arch Jun 28 23:53:48 to the arch you want to target Jun 28 23:54:56 well i dont know what arch i want Jun 28 23:54:57 you are targetting the glinet150m, I assume it is the AR-150m which is atheros Jun 28 23:54:58 that sthe issue Jun 28 23:55:09 you can find that info on the openwrt website Jun 28 23:55:20 nop Jun 28 23:56:36 https://wiki.openwrt.org/toh/gl-inet/gl-ar150 Jun 28 23:56:51 Atheros AR9331 Jun 28 23:56:58 ar71xx Jun 28 23:57:49 you need to build a firmware for it and then use that toolchain to build the driver as well Jun 28 23:58:40 have you looked at openwrt documentation on how to build a firmware? Jun 28 23:58:49 or get the toolchain Jun 29 00:00:07 and someone is just telling you the same thing as here Jun 29 00:00:53 there is the 300M but 150M doesn't exist Jun 29 00:05:21 well on the github it says to change the config_platform Jun 29 00:05:25 to the correct platform Jun 29 00:05:26 right? Jun 29 00:09:19 Mister_X, cant i just build the driver then load it as a module Jun 29 00:12:07 you need to build it with openwrt toolchain against the kernel, can't just build it externally Jun 29 00:14:19 jwh, oh why is that? one the github it says CONFIG_PLATFORM_I386_PC = n Jun 29 00:14:19 ... Jun 29 00:14:19 CONFIG_PLATFORM_ARM_RPI = y Jun 29 00:14:33 makes it seems all you do is hoose the platform and compile Jun 29 00:16:14 umm Jun 29 00:16:51 i have compiled it before for my amd64 of course Jun 29 00:17:06 that doesn't look like its even used Jun 29 00:17:20 but even so, that is supposed to be built as a package Jun 29 00:17:30 one sec Jun 29 00:17:41 https://github.com/gnab/rtl8812au Jun 29 00:17:52 that builds it as a module i think Jun 29 00:19:45 it still needs to be built with the correct toolchain against the kernel in the openwrt image Jun 29 00:20:13 i see, is there a wiki to show me how to do that. https://www.securityartwork.es/2018/02/21/laboratorio-hacking-i-wifi-gl-ar150-pineapple-nano-2-0-2-enhanced/ Jun 29 00:20:16 i followed that guide Jun 29 00:20:23 and got to the make menuconfig part Jun 29 00:20:37 so tahts all set up, but i stopped cause i didnt see the firmware that i needed Jun 29 00:30:24 how do i set up the toolchain and such then jwh Jun 29 00:34:51 they both build as a module Jun 29 00:35:02 probably you should read https://wiki.openwrt.org/doc/howto/build Jun 29 00:35:13 I refered to the other github because I know the guy does a significant amount of work on it Jun 29 00:37:17 clone repo, drop openwrt bits for that driver in packages, configure and build for your target, should be on the wiki Jun 29 00:39:26 hw do i know what openwrt bits i drop into the git ? Jun 29 00:41:18 the gist in your post has most of it, may be better repos though Jun 29 00:42:38 well one of those gists has a github link that no longer works Jun 29 00:42:45 in fact, the most basic of search provides this Jun 29 00:42:51 https://gist.github.com/jow-/920dfe708e8997e23f71ca453df54ed9 Jun 29 00:43:00 probably enough Jun 29 00:43:21 modify to suit, ofc Jun 29 00:43:39 the pkg source is no longer any good, so where do i stikc this gist into tho Jun 29 00:44:07 into a package, theres a page on the wiki somewhere Jun 29 00:44:53 i cant just compile this directly? Jun 29 00:47:22 "directly" ? Jun 29 00:47:59 you can use the sdk I'd imagine Jun 29 00:48:05 but thats as direct as you're going to get Jun 29 00:49:33 im reading the wiki Jun 29 00:49:47 and it says basically to use menuconfig to select the packages you want but thats part of the issue this package doesnt exist Jun 29 00:50:06 https://openwrt.org/docs/guide-developer/obtain.firmware.sdk Jun 29 00:50:23 create package, most of the work has already been done Jun 29 00:52:05 sigh im confused Jun 29 00:52:54 for one i already have all the ddwrt stuff ready thats why i was able to run make menuconfig Jun 29 00:53:08 so there is no point in redownloading all of that but how do u add that information into menuconfig Jun 29 00:55:50 hmm well in the packages directory there is a file called linux firmware and inside a make file for r8169 firmware Jun 29 00:56:24 you need to download at least the sdk Jun 29 00:56:42 but the sdk will save you from having to build the entie thing Jun 29 00:56:43 entie Jun 29 00:56:48 ugh, entire Jun 29 00:57:19 its not that difficult to figure out from the wiki, otherwise you'll have to see if someone will do it Jun 29 00:57:20 well either way i need to build the entire image as the current .bin i have wasnt working right for my device anyway Jun 29 00:57:47 is it actually a supported device? Jun 29 00:57:50 yes Jun 29 00:58:05 spent one day with dnsmasq, to no avail. unless I run dnsmasq on port 53, I can never get a DNS server assigned to dhcp client Jun 29 00:58:10 so in the packages/firmware there are a few folders, do i just pick a random name for it? Jun 29 00:58:38 if i changed default 53, to say 5353, i can get IP, route all correct, but no DNS server is assigned to my dhcp client, what for Jun 29 00:59:01 no, you'll need to create a new package Jun 29 00:59:13 see the gist, use that Jun 29 00:59:14 change port=5353 in /etc/config/dhcp, restart it, i can now ping 8.8.8.8, but no dns, i.e. no point to any domain name Jun 29 00:59:46 dhcp doesn't contain ports, so that seems like the expected result Jun 29 00:59:55 or are you still doing the redirect? Jun 29 01:00:37 with 53 the default, everything of course works, change 53 to anything else, no DNS server will be assigned(on the router I redirect 53 requests to 5353,e.g) Jun 29 01:00:41 there is an easier fix for giving a seperate machine a different dns server, just alias an address on lo and bind it to that, then hand it out as the server Jun 29 01:00:49 ah Jun 29 01:00:55 maybe it checks for sane config Jun 29 01:00:56 heh Jun 29 01:00:58 jwh: this is for a group of people, so it has to be automatic Jun 29 01:01:48 nah it seems this all fruitless Jun 29 01:01:52 other ppl ahve tried and failed Jun 29 01:01:52 if you have the mac address (or I guess you could match on client hostname), you could just tag them and then override dns server Jun 29 01:01:55 now to experiment, i run one dnsmasq on 53, another on 5353 at the same router openwrt, and redirect my pc's 53 request to 5353, PC got a correct DNS server then Jun 29 01:01:57 https://forum.lede-project.org/t/rtl8812au-and-or-rtl8814au-drivers/67/52 Jun 29 01:01:58 none of its pretty elegant though Jun 29 01:02:44 so i must have one dns running at port 53 with the dhcp server even though my pc client does not use that 53 port Jun 29 01:03:29 otherwise pc does not have a DNS server assigned, very strange, compared both dnsmasq configs, even added no-dhcp-interface to the dnsmasq5353.conf for the redirect instance Jun 29 01:04:31 tbf it sounds like misconfig or misunderstanding of whats actually happening Jun 29 01:04:43 how are you redirecting? mac address? Jun 29 01:05:22 dnsmasq(dns53+dhcp), works; dnsmasq(dns5353+dhcp+redirect), no dns server; dnsmasq(dhs53+dhcp)+dnsmasq(dns5353), works; Jun 29 01:06:00 in short i must have one dns at 53 port even though my MAC addr will redirect my 53 request to 5353 Jun 29 01:07:10 jwh: iptables -t nat -A PREROUTING -p tcp -dport 53 -m mac --mac-source xxxx -j REDIRECT --to-port 5353 Jun 29 01:07:39 this redirect works fine as long as there is another dnsmasq running 53 idle Jun 29 01:08:06 I wouldn't even do that, I'd just tag those mac addresses in dnsmasq, and supply a different address for dns Jun 29 01:08:18 maybe it's time to seperate dhcp and dns... Jun 29 01:08:20 coz thats pretty horrible Jun 29 01:08:32 so like Jun 29 01:08:55 you could do dhcp-host=set:altdns,XX:XX:XX:XX:XX:XX Jun 29 01:08:59 or similar, can't remember Jun 29 01:09:08 just have another server bound to aliased address Jun 29 01:09:20 jwh: the point is to enforce dns to avoid using random external DNS servers, so dhcp-tag does not help Jun 29 01:10:07 well you can just deny them access to !$ip Jun 29 01:10:25 any request on 53 port must abide whatever openwrt does on dns, no override on dns is permitted Jun 29 01:10:50 forbid change of dns vs deny-ip-access are a bit different... Jun 29 01:11:15 jwh, im gonna shift gears on this a bit. The different with the M seems to be the external antenna. How do i ensure the device is able to use usb to receive internet. I want to plug it into my comptuer and share the connection to it. What modules do i need to set up when using menuconfig to make that work Jun 29 01:12:45 ausjke: I'm not sure -j REDIRECT is the right target either, but not done that much with it Jun 29 01:13:21 should be, as it works as long as there is an extra dns53 port openning idle Jun 29 01:13:36 exactly Jun 29 01:13:53 i can see my MAC's 53 is directed to 5353 Jun 29 01:13:57 it shouldn't need to listen Jun 29 01:14:16 however without that idle 53-dnsmasq i can never get a DNS server assigned,which is the core issue Jun 29 01:14:48 what does the broken config look like? Jun 29 01:14:50 generated, not uci Jun 29 01:15:25 i cut that to minimum so the same as openwrt.org, only changed port=5353 and added an iptable-redirect... Jun 29 01:15:52 on 5353 dnsmasq i removed dhcp-authorative and anything dhcp related Jun 29 01:17:00 how about a similar take on what I suggested previously, but keeping the redirect... except send to different aliased address, rather than changing the port? Jun 29 01:17:04 hmm let me turn off dhcp in dnsmasq53 and move them to 5353 Jun 29 01:17:27 did not quite get the alias idea, let me re-read Jun 29 01:18:08 obviously it should still work regardless of port, but something isn't right Jun 29 01:18:12 can't tell without configs :D Jun 29 01:21:01 https://pastebin.com/KXUGDKwv for 53 port without dhcp Jun 29 01:22:16 https://pastebin.com/txpduvzn for the default dhcp config but changed to 5353 Jun 29 01:22:59 if i swap 5353 and 53 port in both configs it worked fine, but i want to put 5353 in default dhcp Jun 29 01:23:26 it seems the dhcp likes to stay with port53 no matter what Jun 29 01:23:39 otherwise it won't give me DNS server Jun 29 01:23:43 to my pc client Jun 29 01:24:19 probably makes sense, no point handing out dns server if its listening on 5353 Jun 29 01:24:44 in 5353 though, if you give it another address (like 8.8.8.8) or something, does it hand that out? Jun 29 01:27:07 no Jun 29 01:27:20 option server '8.8.8.8' does not help Jun 29 01:27:25 I'd avoid using uci/init.d to start 5353 instance anyway Jun 29 01:27:32 I'm sure I said this the other day as well Jun 29 01:27:36 so maybe dnsmasq does not give out dns server if non-standard port is configured Jun 29 01:27:56 why, i can run it manually though Jun 29 01:27:58 let me try Jun 29 01:28:14 because it generates config you haven't otherwise defined Jun 29 01:29:00 the fine /tmp/etc/dnsmasq.conf is sane tough Jun 29 01:29:03 s/fine/final/ Jun 29 01:30:16 WARNING: your configuration is out of sync. Please run make menuconfig, oldconfig or defconfig! Jun 29 01:30:29 i dont get it i ran make menuconfig saved it then pressed esc esc Jun 29 01:30:32 and it still says that Jun 29 01:30:52 does it have to be called .config then Jun 29 01:31:47 rpifan: just run make defconfig Jun 29 01:32:18 better, ./scripts/diffconfig.sh > /tmp/1 && make defconfig && ./scripts/diffconfig.sh > /tmp/2 && diff /tmp/1 /tmp/2 Jun 29 01:32:24 and see the diff on what config changed Jun 29 01:32:57 hmm, if i can at least get the usb interent working Jun 29 01:35:52 then maybe i can get someone to make that rtl into a package for me, maybe in exchange for a pizza anyone? Jun 29 01:36:04 its not just for me its for the community Jun 29 01:42:17 /home/phantonhawx/Projects/gl-ar150-pineapple-firmware-builder/openwrt-cc/include/toplevel.mk:181: recipe for target 'world' failed Jun 29 01:44:23 CC is 3 years old Jun 29 01:44:32 and that sin't real openwrt, so prob won't get any help with that Jun 29 01:44:42 jwh: i moved dhcp portion to 53 out of /etc/config/dhcp and now it work! Jun 29 01:44:48 heh Jun 29 01:44:56 made my day, thinks for the hints Jun 29 01:45:05 what is real openwrt then? Jun 29 01:45:41 jwh: the key is whatever you try, dnsmasq needs a 53 port to dispense dns server to its dhcp clients Jun 29 01:45:42 also that has nothing to do with the wifi card in question? Jun 29 01:46:01 yeah Jun 29 01:46:15 thats what I suspect initially, sensible sanity check I guess Jun 29 01:46:18 suspected Jun 29 01:48:16 well the device is supported by lede Jun 29 01:48:28 but i fillowed this guide and it uses CC https://www.securityartwork.es/2018/02/21/laboratorio-hacking-i-wifi-gl-ar150-pineapple-nano-2-0-2-enhanced/ Jun 29 01:48:36 how can i update it to lede Jun 29 01:48:55 why are you using that if its supported? Jun 29 01:49:11 just flash it with the appropriate image Jun 29 01:49:16 https://github.com/KhasMek/gl-ar150-pineapple-firmware-builder Jun 29 01:49:30 thats why, because it isnt just plain lede is a hack of a the wifi pineapple Jun 29 01:49:44 so basically you take the wifi pineapple bin extract put taht into lede then compile Jun 29 01:50:21 theres probably a better way of doing that Jun 29 01:50:25 what wifi is it for? Jun 29 01:50:39 wifi? Jun 29 01:51:37 https://openwrt.org/toh/gl-inet/gl-ar150 Jun 29 01:53:37 yea thats the clean but the idea is to merge the wifi pineapple stuff to that, but the instructions all pull from chaos calmder, i guess if i pull lede it should just work too Jun 29 01:53:53 why do you need the "wifi pinapple stuff"|? Jun 29 01:53:56 what does it do? Jun 29 01:55:03 the wifi pineapple is a 100 dollar pent test tool Jun 29 01:55:09 that happens to hae the sam eboard as this cheap thing Jun 29 01:55:40 right, ok Jun 29 01:55:49 so it isn't supported by openwrt Jun 29 01:55:55 at least, not officially Jun 29 01:56:18 its built on open but not official no Jun 29 01:59:09 still there should be no difference in building Jun 29 02:00:43 I don't see anything of value in that pineapple repo Jun 29 02:00:58 so why not just flash it with openwrt and see if it works Jun 29 02:01:51 https://www.wifipineapple.com/ Jun 29 02:02:52 you can probably port the apps Jun 29 02:03:09 thats probably harder to do Jun 29 02:03:34 yup Jun 29 02:03:38 though it would be good for the community Jun 29 02:03:50 obviously they provide firmware though Jun 29 02:03:55 so just ask them to do stuff? Jun 29 02:03:56 still i have an important decision to make Jun 29 02:04:02 do i eat an entire pizza Jun 29 02:04:04 or not Jun 29 02:05:51 guess i am Jun 29 02:09:42 im gnna update that github thing and make a newer version lets see if it works then i can finally post something on a github **** ENDING LOGGING AT Fri Jun 29 03:00:00 2018