**** BEGIN LOGGING AT Mon Aug 24 02:59:59 2015 Aug 24 04:56:10 build #81 of rb532 is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/rb532/builds/81 Aug 24 05:16:45 after 4 years finally a new version of ncurses "officially" available Aug 24 05:17:09 most distros using the "dev" version atm probably Aug 24 08:29:04 cyrus r46716 trunk/package/network/ (5 files in 3 dirs) * omcproxy: add new igmpv3 & mldv2 multicast proxy Aug 24 08:54:09 cyrus r46717 trunk/package/network/services/omcproxy/Makefile * omcproxy: fix last commit Aug 24 09:34:13 swalker: btw, you scan script seems to not pick up some mainteiners; e.g. packages net/bind does have a PKG_MAINTANER Aug 24 10:40:53 KanjiMonster: MAINTAINER="$(grep -E -m 1 '(^|[[:space:]]|PKG_)'MAINTAINER:= $MAKEFILE)"; bind had to be different and add too much spacing Aug 24 10:42:59 swalker: there's also one example of "PKG_MAINTAINER=" (i.e. no ":") Aug 24 10:48:43 KanjiMonster: fixed the 8 offending packages Aug 24 10:50:05 swalker: something along "sed -n -e 's/.*MAINTAINER.*=[ ]*\(.*\)/\1/p' " is what I used when I started my "create a lookup table for maintainer => github handle" project Aug 24 10:55:20 KanjiMonster: libs/alsa-lib/Makefile is special as well ;) Aug 24 11:02:13 swalker: yeah, we started getting a few packages with more than one maintainer ;) Aug 24 11:02:18 if a package doesn't build with format-security type problems, is the best approach to try hacking on the old branch of code for that package to fix it, or can you just turn off format-security checking for a single package? Aug 24 11:02:25 (net-snmp in this case) Aug 24 11:02:43 I'm not sure trying to hack on the old 5.4.2.1 branch code to fix them is worthwhile.... Aug 24 11:03:14 ahh, PKG_CHECK_FORMAT_SECURITY:=0 in the package might be useful... Aug 24 11:03:42 ah, you already found it yourself ;) Aug 24 11:05:25 eck, I have some local net-snmp patches too :| fun times. Aug 24 11:07:22 doesn't everyone have a few packages they use themselves, but don't want to take the burden of maintaining them? ;) Aug 24 11:08:06 I was actually tidying up my own things locally, then someone else offered to take it before I got there :) Aug 24 11:08:56 well, with net-snmp, I want agentx support, and tcp support, and I'm not sure if those are really appropriate in teh base build. Aug 24 11:09:14 so it's sort of "have to have my own anyway" Aug 24 12:04:32 blah, everytime I try and touch trunk the kernel versions changed and abandonded files in an patch directory Aug 24 12:55:11 hmmm, this is nasty. when I try to ssh to my ERL, I get this: http://pastebin.com/nK63d6K8 Aug 24 13:07:54 sometimes when I use the scripts/remote-gdb, I get "Process xxx created; pid = yyy" but then nothing else. it just hangs. SOmetimes it works and the process actually starts, anyone seen that or have any ideas? Aug 24 14:21:36 build #81 of cobalt is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/cobalt/builds/81 Aug 24 14:24:48 build #81 of cns21xx is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/cns21xx/builds/81 Aug 24 14:27:51 build #81 of orion is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/orion/builds/81 Aug 24 14:37:01 build #76 of ar71xx.mikrotik is complete: Failure [failed compile_8] Build details are at http://buildbot.openwrt.org:8010/builders/ar71xx.mikrotik/builds/76 Aug 24 14:56:02 build #76 of cns3xxx is complete: Failure [failed shell_14] Build details are at http://buildbot.openwrt.org:8010/builders/cns3xxx/builds/76 Aug 24 15:02:02 wtf, I can no longer ssh to hosts in different network segments that are routed by my ERL :/ Aug 24 15:02:23 seems to give the same as when trying ssh to the ERL itself, http://pastebin.com/nK63d6K8 Aug 24 15:05:16 hmm. after a reboot I can ssh to the ERL again but not to host on different segment. that's now giving this: Port 0 receive error code 2, packet dropped Aug 24 15:07:54 stintel: trunk? Aug 24 15:08:29 diizzy: yes, r45712 Aug 24 15:09:04 but it's weird, seems like the octeon ethernet driver is dropping the packets, while that worked before Aug 24 15:09:26 I guess I should update octeon to 4.1 kernel, there were some modifications to the driver after 3.18 Aug 24 15:09:44 stintel: works fine here, r46386 Aug 24 15:09:53 but it does crash on boot sometimes Aug 24 15:10:06 this image is compiled with gcc 4.9 though Aug 24 15:10:31 I have been testing different combinations to see if it would fix the tcpdump segfault Aug 24 15:10:32 stintel: 4.8 here Aug 24 15:11:01 compiling current trunk with default gcc (4.6) now, to see what gives Aug 24 15:11:21 stintel: I do get sometimes irq 146 error or such Aug 24 15:11:47 seems to work fine now at least (some image) Aug 24 15:11:50 same* Aug 24 15:12:02 oh my bad, 4.6 here Aug 24 15:12:21 Linux version 3.18.18 (diizzy@buildbox-v2) (gcc version 4.6.4 (OpenWrt/Linaro GCC 4.6-2013.05 r46386) Aug 24 15:12:38 stintel: when does tcpdump segfault? Aug 24 15:13:17 bummer that musl doesn't have mips64 support.. this is going to be an annoying target now that trunk uses musl for everything else :() Aug 24 15:13:29 diizzy: always Aug 24 15:13:37 right at start Aug 24 15:14:24 build #79 of netlogic is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/netlogic/builds/79 Aug 24 15:14:29 it doesn't with glibc Aug 24 15:15:19 stintel: gimme a sec Aug 24 15:15:56 works (glibc) Aug 24 15:16:14 I don't see why it should use uclibc(ng) on mips64 Aug 24 15:16:26 space isn't an issue Aug 24 15:22:00 hmm, even with gcc 4.6 I still get that Port 0 receive error code 2, packet dropped Aug 24 15:26:36 stintel: making glibc standard and call it a day isn't an option? Aug 24 15:27:13 I doubt it is related Aug 24 15:27:16 but I will try it Aug 24 15:27:27 interesting, 4.1 kernel seems to have octeon optimized crypto modules Aug 24 15:41:22 stintel: dunno if sha1,256,512 md5 is of much use though Aug 24 15:42:50 yeah, aes would be better :) Aug 24 15:49:47 aren't those "Cavium Liquidio" LAN cards for octeon platform too ? Aug 24 15:50:18 or at least from the same vendor that sells octeon stuff Aug 24 15:56:23 heh, luci.fs no longer exists. that'll teach me... Aug 24 15:59:11 jow_laptop: documentation/JsonRpcHowTo.md still talks about exporting rpc/fs/* as being the complete luci.fs library, and of course, http://luci.subsignal.org/api/luci/modules/luci.fs.html still exists, even though you dropped it with no futher explanation in e91b603acc0920f998ae57704a9d97b83ee284e4 Aug 24 16:09:57 bleh, luci.__version__ gone too :) w Aug 24 16:10:26 seems "lv = require("luci.version"); lv.luciversion" orks on both BB and "new" luci Aug 24 16:15:02 KanjiMonster: i updated the mac80211 patches with the changes you and hauke suggested Aug 24 17:42:14 plntyk: thanks Aug 24 19:08:06 build #74 of octeon is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/octeon/builds/74 Aug 24 20:04:29 is there a standard way to support more than just wifi LED for new configurations? Aug 24 20:20:45 zinx: this question needs more context Aug 24 20:21:42 KanjiMonster: i ported OpenWrt to the BR-6475nD, it has LEDs that need setup for wifi, 5ghz wifi, and power, on a fresh install. openwrt has a standard way of setting it up for wifi. i want to know if there's a standard way for other LEDs. Aug 24 20:22:32 i did the port a while back; just figured i'd add stuff. Aug 24 20:24:12 zinx: not really well documented, but there are quite a few led types supported -> http://git.openwrt.org/?p=openwrt.git;a=blob;f=package/base-files/files/lib/functions/uci-defaults.sh;h=8ba95a22ff71893288b14e8d5338fd2795ee354d;hb=HEAD Aug 24 20:24:47 thanks Aug 24 20:26:59 hmm looking at the board.d file seems they are used now too >_< Aug 24 20:30:56 build #77 of avr32 is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/avr32/builds/77 Aug 24 21:25:54 I'm working to a @GITHUB scheme for download, I can download directly a tar.gz from github and download in 1/3 of the time Aug 24 21:30:43 https://github.com/openwrt/packages/issues/1704#issuecomment-134369612 Aug 24 22:29:57 matteo: how are you going to handle some of the useless tarball names you'll end up fetching? Aug 24 22:31:04 you can choose the archive name in the makefile Aug 24 22:31:35 PKG_SOURCE:=$(PKG_NAME)-$(PKG_VERSION)-$(PKG_SOURCE_VERSION).tar. Aug 24 22:31:55 if you don't want the git hash, then just use the version and release Aug 24 22:33:29 gnite **** ENDING LOGGING AT Tue Aug 25 02:59:58 2015