**** BEGIN LOGGING AT Sun Feb 27 02:59:57 2011 Feb 27 04:51:19 aha! Feb 27 04:51:24 so it was a luci bug Feb 27 04:51:31 (oh noes) Feb 27 05:05:20 https://dev.openwrt.org/ticket/7925 <-- it applies here using the latest trunk :\ Feb 27 05:05:29 not even backfire Feb 27 05:15:36 but well... the issue is different here :\ Feb 27 06:00:52 build #75 of x86 is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/x86/builds/75 Feb 27 06:39:28 build #93 of atheros is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/atheros/builds/93 Feb 27 11:08:22 xMff: you dead? Feb 27 11:22:44 nbd * r25748 /trunk/target/linux/generic/config-2.6.38: kernel 2.6.38: add a few missing config symbols Feb 27 11:22:49 nbd * r25749 /trunk/package/mac80211/ (27 files in 2 dirs): mac80211: update to 2011-02-25 Feb 27 11:22:55 nbd * r25750 /trunk/include/netfilter.mk: netfilter.mk: put ipv6 conntrack in the right package Feb 27 11:26:16 nbd: you use luci? Feb 27 11:26:31 no Feb 27 11:47:35 xMff: seems iwinfo is mandatory again (can't compile it without wifi stuff: iwinfo_nl80211.h:29:29: fatal error: netlink/netlink.h: No such file or directory) Feb 27 12:10:24 updated openwrt/upstream, https://home.comcast.net/~sdwalker/uscan/index.html Feb 27 12:18:27 [florian] around? Feb 27 12:24:14 Kaloz_: ping Feb 27 12:25:23 ahh :\ Feb 27 12:26:44 Wipster? Feb 27 12:29:41 loswillios: no its not Feb 27 12:30:54 some commiter around? xD Feb 27 12:37:29 hi Feb 27 12:37:41 crap Feb 27 12:38:05 does dnsmasq read /etc/dnsmasq.conf at all anymore? Feb 27 12:41:56 did the parallel build option (-j X ) disappear from the .config file? (it used to be in advanced (for developers) -> build options ? Feb 27 12:44:13 Weedy_laptop: yes Feb 27 12:44:53 Weedy_laptop: uci stuff only builds commadnline args, dnsmasq still tries to read /etc/dnsmasq.conf if it exists Feb 27 12:46:13 aroscha: yes Feb 27 12:46:17 aroscha: you can use make -j ... now Feb 27 12:46:34 nbd: ok, thx... so no special stuff needed anymore :) Feb 27 12:46:56 hmm... having another buildbreak with -j 4 Feb 27 12:47:19 checking for utime... yes Feb 27 12:47:19 configure: creating ./config.status Feb 27 12:47:19 config.status: error: cannot find input file: `Makefile.in' Feb 27 12:47:19 make[3]: *** [/media/SeagateSWRaid0V/mangroveOS/seed/trunk/backfire.trunk/build_dir/host/opkg-576/.configured] Error 1 Feb 27 12:47:19 make[3]: Leaving directory `/media/SeagateSWRaid0V/mangroveOS/seed/trunk/backfire.trunk/package/opkg' Feb 27 12:47:19 make[2]: *** [package/opkg/host/compile] Error 2 Feb 27 12:47:30 was this tree fresh or updated? Feb 27 12:47:38 updated Feb 27 12:47:44 svn up Feb 27 12:47:57 should I fetch a fresh one? Feb 27 12:47:58 after the large number of autofools related changes, a make dirclean is in order Feb 27 12:48:00 https://dev.openwrt.org/ticket/8418#comment:1 Feb 27 12:48:05 ok thx Feb 27 12:48:33 <[florian]> PsyMan: yes Feb 27 12:53:37 jow * r25751 /packages/net/chillispot/patches/150-chillispot-rmtctrl.patch: [packages] chillispot: use memset instead of bzero Feb 27 12:54:15 xMff: does our dnsmasq support txt records? Feb 27 12:54:22 i think i'm doing it wrong Feb 27 12:54:27 Weedy_laptop: yes Feb 27 12:54:35 Weedy_laptop: we even ship examples for it Feb 27 12:54:56 https://dev.openwrt.org/browser/trunk/package/dnsmasq/files/dnsmasq.conf Feb 27 12:56:05 nice Feb 27 12:56:19 so i've got wireless working Feb 27 12:56:24 anyone here able to confirm the GPIO values of the QSS and RESET buttons on the WR1043ND? Feb 27 12:57:01 it was luci making the router freeze Feb 27 12:57:19 * Olipro_ pings Kaloz_ Feb 27 12:57:49 PsyMan: it is just calling /sbin/wifi and /etc/config/network restart - if this is freezing your router there's either an oom condition or a bug in the scripts Feb 27 12:58:11 * /etc/init.d/network Feb 27 12:58:13 if i access the router using the cmd line edit config and reboot Feb 27 12:58:16 it works Feb 27 12:58:28 yeah, reboot is different Feb 27 12:59:31 if i edit using luci it freezes the router Feb 27 13:00:04 :\ Feb 27 13:00:14 (13:57:40) xMff: PsyMan: it is just calling /sbin/wifi and /etc/config/network restart - if this is freezing your router there's either an oom condition or a bug in the scripts Feb 27 13:00:51 ehm Feb 27 13:01:17 network? :p Feb 27 13:02:15 you mean /etc/config/wireless Feb 27 13:03:43 it was freezing when enabling the controller Feb 27 13:09:23 so restarting the network... does not exactly fit..? :\ Feb 27 13:10:01 /etc/init.d/network restart Feb 27 13:10:11 that is redoing everything, including switch and wifi Feb 27 13:10:24 hmm Feb 27 13:10:28 and that is what luci is calling Feb 27 13:10:34 so maybe try it on the cli Feb 27 13:10:41 the problem is Feb 27 13:11:04 even after rebooting, wireless stays off, but only in ap mode Feb 27 13:12:16 if i have some other mode selected, enable the device and apply the changes, it still freezes the router, but the settings are applied after rebooting the router Feb 27 13:12:47 the exact same thing in ap mode would keep the device disabled Feb 27 13:13:11 right, so you should debug it on the cli Feb 27 13:13:23 maybe some process is hanging Feb 27 13:13:27 during network restart Feb 27 13:13:35 or wpad cannot be brought up properly Feb 27 13:14:11 nbd, sorry for bothering :). The problem was so simpler than what I said about DMA failed to stop. Actually each time that I run wifi at client, it makes the connection unstable. Even I can't ping the AP properly. I see strange ping times. You can see that here: http://pastebin.com/uWkizTcS . I have to run wifi at AP after this to fix this problem. I repeated this test a lot and every time I see that result Feb 27 13:16:13 hmm Feb 27 13:32:46 nbd, and I get 5 or 6 ping responses at once. the strange ping times are because of that Feb 27 13:34:12 what encryption do you use? Feb 27 13:41:33 xMff: do i query the txt of the router ip? Feb 27 13:41:46 +how ? Feb 27 13:42:03 I suppose with dig Feb 27 13:42:09 but I do not know Feb 27 13:42:51 empty :/ Feb 27 13:44:04 http://openwrt.pastebin.com/6FmmTVvk Feb 27 13:44:10 i r confuse Feb 27 13:49:11 xMff: ah ok. do you know what's the issue? http://pastebin.com/RSVu1r89 Feb 27 13:49:32 loswillios: some heisenbug in dependency handling Feb 27 13:49:42 eeks Feb 27 13:49:49 it is supposed to depend on libnl-tiny if kmod-mac80211 is enabled Feb 27 13:50:05 it does work in menuconfig but it seems it is not properly propagated as build dependency Feb 27 13:50:24 its --- in menuconfig Feb 27 13:50:31 libnl-tiny ? Feb 27 13:50:42 iwinfo Feb 27 13:50:55 yeah because the "luci" meta package depends on it Feb 27 13:51:03 ah, now I get it Feb 27 13:51:04 you can just deselect that Feb 27 13:51:14 and choose the components individually Feb 27 13:51:50 okies Feb 27 14:05:09 tornado: do you have those ping time problems with the lastest backfire builds? I have them as well... Feb 27 14:05:13 <[florian]> Wipster: ping? Feb 27 14:05:58 Erich_123, I use the latest trunk Feb 27 14:06:37 Erich_123, do you use ath9k? Feb 27 14:06:42 tornado: yes Feb 27 14:06:59 torndo: dir-825 and wndr3700 Feb 27 14:07:59 Erich_123, Router Station Pro here at both sides Feb 27 14:08:23 tornado: latest trunks as of yesterday works fine tough. something changed in backfire between Feb 11th and Feb 24th... Feb 27 14:09:52 is 0x18040004 the correct base address for reading the button GPIO inputs on WR1043ND? Feb 27 14:10:36 tornado: running WPA2 with wds enabled and bridging with stp Feb 27 14:11:32 Erich_123, I don't use any encryption. wds enabled Feb 27 14:12:56 Erich_123, nbd is expert at this. I think we should wait for him Feb 27 14:15:27 Erich_123, at your situation, running wifi at AP solves the problem? Feb 27 14:17:04 tornado: to be honest I didn't try, 2-3x reboot solved it for me and then I switched back to my previous builds Feb 27 14:19:01 Erich_123, I think it's started after wireless-compat-2011-02-07. I didn't check ath9k after that till some days ago Feb 27 14:19:01 tornado: the problem was not there right from the start, it came after 20-60 minutes or so... Feb 27 14:20:20 tornado: that is my guess too, but I didn't look at the change timeline close enough Feb 27 14:40:47 so, I just built a custom uBoot for the WR1043ND Feb 27 14:40:52 anyone feel like flashing it? :) Feb 27 14:41:41 heh Feb 27 14:42:20 the changes I've made won't do any damage if you have serial because they'll only execute if you don't interrupt the boot Feb 27 14:42:57 hm, I should probably say what I changed Feb 27 14:44:09 basically, if you hold in the QSS button, the bootloader will load an image from TFTP into RAM and boot it Feb 27 14:44:39 if you hold in the reset button the device will load an image from TFTP into RAM and write to flash Feb 27 14:44:47 Feb 27 14:44:49 although I might change that to require you to press both Feb 27 14:57:02 ehm Feb 27 14:57:05 question Feb 27 14:57:28 the mini versions of wpad, hostapd, wpa sup Feb 27 14:57:40 they support only wap1 Feb 27 14:57:44 standard Feb 27 14:58:07 apart from that, do they support only tkip or aes too? Feb 27 14:58:47 the mini variants support wpa2 with aes and tkip Feb 27 14:58:54 but only personal mode Feb 27 14:58:59 what? Feb 27 14:59:02 since when? Feb 27 14:59:08 no wpa enterprise (radius and stuff), just psk Feb 27 14:59:14 i thought they were only wpa1 capable Feb 27 14:59:16 since they where introduced? Feb 27 14:59:19 :o Feb 27 14:59:23 many thanks Feb 27 14:59:42 11n mandates aes support Feb 27 15:00:00 wopuld be kind of pointless to ship authenticators without support for it Feb 27 15:00:18 i see Feb 27 15:00:26 another thing Feb 27 15:00:43 adls info is available throught luci? or lua only? Feb 27 15:01:19 ehh Feb 27 15:01:23 adsl info, like snr / line status? Feb 27 15:01:23 cli i mean Feb 27 15:01:27 yeah Feb 27 15:01:32 only cli Feb 27 15:01:34 ahh Feb 27 15:01:41 there is no standard interface I could support Feb 27 15:01:49 like /proc/dsl_status Feb 27 15:02:02 oh, they are many? Feb 27 15:02:10 had no clue about it Feb 27 15:02:25 well there is ueagle, solos and the ar7 thing Feb 27 15:02:34 and then trendchip Feb 27 15:02:43 and lantiq/ifxmips Feb 27 15:03:14 looks liek we need some libatminfo too ;) Feb 27 15:03:15 ouch Feb 27 15:03:31 but I have no access to any such hw so I cannot implement it Feb 27 15:03:41 I do not even have access to a dsl line :D Feb 27 15:04:09 hmm Feb 27 15:04:33 avalanche/avsar wasn't it? Feb 27 15:04:44 yes Feb 27 15:04:54 ah, there it is Feb 27 15:05:10 stats didn't change since the f/w change :\ Feb 27 15:06:13 not that they are bad Feb 27 15:06:33 I suppose I could hunt down some examples from the net and attempt to build a wrapper for dsl info Feb 27 15:06:39 according to that it can go up to 1.1ul/18dl Feb 27 15:06:43 ar7 has some proc file afaik Feb 27 15:07:00 that would take up too much space Feb 27 15:07:12 probably true Feb 27 15:07:31 and the modem driver is too big already Feb 27 15:08:15 hmm Feb 27 15:08:26 4mb is too few space Feb 27 15:08:38 last time i got stuff out it wouldn't boot :| Feb 27 15:09:55 jow * r25752 /branches/backfire/package/cyassl/Makefile: [backfire] merge r24346 (#8947) Feb 27 15:22:17 must be careful not to delete anything needed Feb 27 15:39:54 jow * r25753 /trunk/ (include/autotools.mk scripts/config.rpath): [buildroot] add config.rpath and link it in autotools.mk if missing, some packages might need it, especially users of lib-link.m4 or iconv.m4 Feb 27 15:41:40 jow * r25754 /branches/backfire/ (include/autotools.mk scripts/config.rpath): [backfire] merge r25753 Feb 27 15:42:28 jow-laptop: Cool and thanks! Cheers. Feb 27 15:43:03 mazilo: ? Feb 27 15:43:27 xMff: I meant the r25753 and r25754. Feb 27 15:43:45 you noticed this error in some packages? Feb 27 15:44:56 xMff: No. But, some new packages I tried to compile seemed to have problem with autotools. Let's see if these two revision updates will fix them. Feb 27 15:45:31 florian * r25755 /trunk/target/linux/ar7/Makefile: [ar7] use kmod-acx-mac80211 by default Feb 27 15:45:45 florian * r25756 /trunk/package/acx-mac80211/Makefile: [package] acx-mac8011: bump to 20110123 Feb 27 15:46:54 oh look Feb 27 15:47:20 that version is said to be very stable Feb 27 16:28:16 ffffffffffffffffffffffffff Feb 27 16:28:28 luci is broke man :/ Feb 27 17:12:32 indeed Feb 27 17:23:51 mb * r25757 /trunk/target/linux/omap24xx/patches-2.6.38/ (505-cbus-retu-write-lock.patch 505-retu-write-lock.patch): rename patch Feb 27 17:24:25 mb * r25758 /trunk/target/linux/omap24xx/patches-2.6.38/811-mmc-null-ptr-fix.patch: omap24xx: Fix possible MMC null ptr deref Feb 27 17:24:53 mb * r25759 /trunk/target/linux/omap24xx/patches-2.6.38/501-cbus-retu-irq-nr-fix.patch: cbus-retu: Fix nested IRQ handling Feb 27 18:30:29 build #87 of orion is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/orion/builds/87 Feb 27 18:44:09 build #93 of at91 is complete: Failure [failed shell compile_6] Build details are at http://tksite.gotdns.org:8010/builders/at91/builds/93 Feb 27 19:04:22 florian * r25760 /trunk/target/linux/generic/patches-2.6.37/ (20 files): [kernel] refresh 2.6.37 patches Feb 27 19:04:30 florian * r25761 /trunk/target/linux/au1000/patches-2.6.37/005-au1000_phy_addr.patch: [au1000] make au1OOO_eth probes all PHY addresses for MTX-1 Feb 27 19:52:44 nbd * r25762 /trunk/target/linux/generic/ (2 files in 2 dirs): kernel: add a patch for isolating individual bridge ports Feb 27 19:52:47 nbd * r25763 /trunk/target/linux/generic/ (8 files): kernel: disable CONFIG_WIRELESS_EXT_SYSFS by default Feb 27 19:52:54 nbd * r25764 /trunk/target/linux/generic/ (8 files): kernel: disable CONFIG_LEDS_TRIGGER_HEARTBEAT by default, nothing uses it Feb 27 19:52:55 nbd * r25765 /trunk/target/linux/generic/ (8 files): kernel: disable CONFIG_SYSCTL_SYSCALL_CHECK by default Feb 27 19:53:00 nbd * r25766 /trunk/target/linux/generic/ (8 files): kernel: disable CONFIG_CRYPTO_ALGAPI* by default, it will be built into the kmod-crypto-core package Feb 27 19:53:06 nbd * r25767 /trunk/target/linux/generic/ (2 files in 2 dirs): kernel: de-bloat the lzma library used by jffs2 Feb 27 19:53:10 nbd * r25768 /trunk/target/linux/generic/ (2 files in 2 dirs): kernel: when disabling common quirks, disable them in the usb subsystem as well - bios handover is useless for most platforms Feb 27 19:53:17 nbd * r25769 /trunk/Config.in: kernel: make AIO support configurable (disabled by default) Feb 27 19:53:19 nbd * r25770 /trunk/target/linux/generic/ (8 files): kernel: disable /proc/kcore Feb 27 19:53:23 nbd * r25771 /trunk/ (9 files in 2 dirs): kernel: disable scheduler actions by default, they will be enabled only when kmod-sched is selected Feb 27 19:56:07 nbd: Do we actually need sysctl syscall? The sysctl utility should use the proc interface. Feb 27 19:57:56 mb__: good idea Feb 27 19:59:22 nbd: And the help text to config KERNEL_AIO seems slightly wrong :D Feb 27 20:00:20 oops Feb 27 20:00:22 copy&paste error Feb 27 20:06:34 nbd * r25772 /trunk/Config.in: fix wrong copy&paste in the helptext for kernel AIO support Feb 27 20:06:40 nbd * r25773 /trunk/target/linux/generic/ (8 files): kernel: disable sysctl syscall support - nothing uses it Feb 27 20:07:36 kernel lzma size went down from ~950k to 910k with all those changes Feb 27 20:07:42 cool Feb 27 20:16:08 I'm wondering why my omapfb based LCD doesn't work on 2.6.38... Feb 27 20:16:39 And the MMC doesn't work properly, too Feb 27 20:17:21 mb__: when I ported Android to the Nokia Internet Tablets I had to patch omapfb Feb 27 20:17:34 then mainline break it Feb 27 20:17:40 .37 works fine for me Feb 27 20:18:07 which board are you using? Feb 27 20:18:14 rx44 Feb 27 20:18:22 N810? Feb 27 20:18:24 yeah Feb 27 20:18:46 last I recall was unusable after 2.6.30 Feb 27 20:19:03 good to know it works, tahvo and everything works too? Feb 27 20:19:42 There are no differences between .37 and .38 in omapfb_main and MIPID. I suspect a bug way deeper in the architecture code. Because MMC behaves strange, too Feb 27 20:20:15 i'm considering looking into omap4 soon Feb 27 20:20:41 * solca should port android to openwrt base, that would be very cool! Feb 27 20:20:56 on .37 basically everything except battery charging works. But I'm working on that, too. If anybody is experienced in reverse engineering soft double precision floating point ARM code that would help :D Feb 27 20:21:32 solca: you can continue with my android openwrt port Feb 27 20:21:34 ;) Feb 27 20:21:40 mb__: do you ever use my NITdroid port? http://guug.org/nit/nitdroid/ Feb 27 20:21:52 nbd: didn't know about it... Feb 27 20:22:00 solca: no Feb 27 20:22:02 solca: when i worked on it, i got to the point where most of the android 1.6 core worked Feb 27 20:22:37 nbd: p54 throws a NULL deref after the mac80211 update. Feb 27 20:22:46 mb__: details please Feb 27 20:22:51 yep, second Feb 27 20:23:31 nbd: that's indeed very cool! I was thinking in dropping all bionic and android shitty stuff and using openwrt to fill it Feb 27 20:23:54 nbd: http://pastebin.com/EabjLj0y Feb 27 20:24:46 solca: i added templating code for compiling android components and split the codebase up into several separate packages Feb 27 20:25:20 nbd: but you were porting bionic to openwrt? Feb 27 20:25:58 solca: no Feb 27 20:26:01 i used eglibc Feb 27 20:26:07 and eventually wanted to move to uclibc Feb 27 20:26:12 back then uclibc nptl didn't work yet ;) Feb 27 20:26:35 nbd: I'm interested in continue your work, my Android port to NITs worked fine but I hate Nokia closed daemons Feb 27 20:26:53 eventually OpenWRT could be used in phones!!! Feb 27 20:27:23 http://nbd.name/gitweb.cgi?p=borg.git;a=summary git://nbd.name/borg.git Feb 27 20:29:05 great! will take a look Feb 27 20:29:53 hmm, it seems mb__ put many love for ARM in OpenWRT Feb 27 20:30:32 mb__: hm, that p54 code is weird Feb 27 20:31:23 mb__: does your device have rssical data? Feb 27 20:31:52 my guess is that data is missing and because of that the rssi lookup table is NULL and there's a missing null pointer check there Feb 27 20:33:25 The eeprom is generated by feeds/packages/utils/calvaria/files/src/cal2p54.c Feb 27 20:33:44 There's a PDR_RSSI_LINEAR_APPROXIMATION_CUSTOM thing Feb 27 20:35:25 please try http://nbd.name/710-p54_rssi_crash_fix.patch Feb 27 20:40:42 mb * r25774 /trunk/target/linux/omap24xx/patches-2.6.38/900-n810-battery-management.patch: n810bm: Compile fix for kernel with lock debugging Feb 27 20:42:54 nbd: what omap4 device do you have? Feb 27 20:42:59 pandaboard Feb 27 20:43:19 cool Feb 27 20:43:28 TI sent me one to check out their wireless stuff Feb 27 20:43:34 wow Feb 27 20:44:11 some nokia engineers seem to be at TI now. At least their email adresses changed :D Feb 27 20:47:18 :) Feb 27 20:47:40 they probably really appreciate not having to work for microsoft now ;) Feb 27 20:47:47 :) Feb 27 20:49:04 nbd: That p54 fix works Feb 27 20:49:52 ok, i'll commit & submit Feb 27 20:50:08 thanks Feb 27 20:51:22 nbd * r25775 /trunk/package/mac80211/patches/710-p54_rssi_crash_fix.patch: p54: fix a null pointer dereference bug Feb 27 20:51:31 But something is still broken. It succeeds to connect to the WPA AP, but when pinging there are delays of 20 seconds or so between successful packets. I guess this is a PS bug. Feb 27 20:52:43 what kind of AP is this? Feb 27 20:53:18 ath5k on vanilla 2.6.37 Feb 27 20:53:58 hostapd is probably a few months old, though. But it used to work. Feb 27 21:37:49 nbd * r25776 /trunk/package/mac80211/Makefile: mac80211: set CONFIG_RTLWIFI Feb 27 21:38:43 mb * r25777 /trunk/target/linux/omap24xx/patches-2.6.38/301-nokia-board-additional.patch: n810: Fix tsc2005 busnum Feb 27 21:45:50 nbd: I like the 'borg' project name Feb 27 21:45:55 last commit 20 months ago Feb 27 21:46:07 * solca cloning... Feb 27 21:46:35 nbd * r25778 /trunk/package/mac80211/Makefile: mac80211: override CONFIG_IWLWIFI_LEGACY Feb 27 21:46:42 nbd * r25779 /trunk/package/mac80211/patches/ (2 files): mac80211: improve ath9k AP A-MPDU PS buffering Feb 27 21:48:34 solca: when you start working on it, i'll give you commit access to the git tree if you want Feb 27 21:49:39 nbd: ok let me try to build something useful first Feb 27 21:49:46 sure Feb 27 21:49:58 this android-* folders are openwrt packages right? Feb 27 21:50:05 it's all just openwrt packages Feb 27 21:50:15 excellent! Feb 27 21:50:16 i tested this stuff in the 'goldfish' target Feb 27 21:50:25 openwrt can also build the emulator for it Feb 27 21:50:25 goldfish is an emultaor AFAIR Feb 27 21:50:28 yes Feb 27 21:51:32 great, it seems there is the android core Feb 27 21:52:14 yes, i had the GUI running, and the browser worked Feb 27 21:53:41 eh, that's a lot Feb 27 21:53:44 everything built by the openwrt build system Feb 27 21:53:46 no external binaries Feb 27 21:54:19 of course all of this still need to be rebased to a recent version Feb 27 21:54:23 but i got quite far with it Feb 27 21:54:32 very far Feb 27 21:54:51 will try it first on goldfish and my old G1 is waiting... Feb 27 21:54:54 too my N810 Feb 27 21:55:04 and a recent android Feb 27 21:55:23 nbd: why you stop? Feb 27 21:55:38 because i didn't have time for it anymore Feb 27 21:55:41 too many other projects Feb 27 21:56:02 but i am still interested in putting some more work into this Feb 27 21:56:04 once i find the time for it Feb 27 21:56:55 and why you say it's not finished? running the GUI and the browser is all about Feb 27 21:57:14 that's the only thing i tested Feb 27 21:57:19 lots of components probably still have bugs Feb 27 21:57:30 and it was still emulator-only Feb 27 21:57:46 my long term goal is to have a decent full-featured phone running openwrt Feb 27 21:58:16 and then i plan on fixing some of the core problems that i see in android Feb 27 21:58:29 e.g. that a lot of stuff on there tries hard to violate your privacy Feb 27 21:59:31 hmmmm. I'd like to run this on the freerunner Feb 27 21:59:33 i want the core API to provide only fake data for address book, gps, serial number, etc. except for apps where i explicitly (and only temporarily) give permission to read the real stuff Feb 27 22:00:37 right, the data partition and sql files have clear text passwords and tons of private data not very good protected Feb 27 22:00:45 another thing that i want to do is move a few abstraction layers for phone and system stuff out of android Feb 27 22:00:54 have them run as small C services hooked up to ubus Feb 27 22:01:05 and have android connect to that instead of using its crappy java stuff Feb 27 22:01:27 that way it's easier to have a common abstraction layer for multiple different software stacks Feb 27 22:01:33 yea, I played with that and it sucks, I'm pretty sure it can be done faster at the system level Feb 27 22:02:20 I had your same ideas but I was looking for speed, my old G1 is decent hw for openwrt but now it's slow for a recent android Feb 27 22:02:24 and connecting ubus to services written in C, java or any other language is trivial Feb 27 22:03:41 and finally borg could be the base for many open firmwares, as everything would be linked to a LGPL libc it would become and GPL Android Feb 27 22:04:01 it seems Google wants the Apache license to give freedom to phone makers and carriers Feb 27 22:04:08 yeah Feb 27 22:04:37 certainly Feb 27 22:04:41 also, android is such a huge messy codebase and 'repo' makes maintaining forks with lots of changes unnecessarily hard Feb 27 22:05:00 i think having a patch structure makes it easier to maintain and rebase changes Feb 27 22:06:41 yes, it's much easier for me to work on OpenWRT than on Android AOSP Feb 27 22:07:25 and we can replace google-forked libraries with openwrt packages Feb 27 22:07:33 i already did that to a few libraries that android uses ;) Feb 27 22:08:18 yea, that's was what I was talking about, openwrt already has many components needed in android Feb 27 22:20:49 it's a pain mixing svn and git Feb 27 22:21:11 nbd: backfire.git is your openwrt trunk? Feb 27 22:21:26 huh? Feb 27 22:21:30 backfire.git is the backfire branch Feb 27 22:21:34 openwrt.git is the trunk branch Feb 27 22:21:45 ohh I miss it Feb 27 22:21:55 those trees are automatically synced via cron job from my server Feb 27 22:22:17 and your goldfish.git you just link it in target/ ? Feb 27 22:22:52 no, that one's obsolete Feb 27 22:22:59 there's a goldfish target in trunk Feb 27 22:23:21 i think i'll delete goldfish.git Feb 27 22:23:37 ok Feb 27 23:01:19 build #17 of lantiq is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/lantiq/builds/17 Feb 27 23:08:40 this is not good Feb 27 23:08:56 luci pages take ages to load Feb 27 23:10:01 sometimes they don't even come up Feb 27 23:13:02 nbd: feeds.conf.default supports something like 'src-git packages git://nbd.name/packages.git' ? Feb 27 23:15:36 yes Feb 27 23:16:27 yea, was reading scripts/feeds and yes it's supported Feb 27 23:16:31 but I get: Ignoring feed 'packages' - index missing Feb 27 23:16:58 did you run ./scripts/feeds update? Feb 27 23:17:25 ./scripts/feeds list Feb 27 23:17:33 * solca retrying... Feb 27 23:19:18 cool, it works for borg! Feb 27 23:19:20 no more svn Feb 27 23:19:45 btw. for trees that i work on, i usually use src-link Feb 27 23:19:51 to point at external directories Feb 27 23:20:12 that way stuff doesn't get lost if i decide to nuke the openwrt checkout Feb 27 23:20:40 what's the difference with src-git? Feb 27 23:20:52 with src-git the feeds script managed the checkout Feb 27 23:20:55 and it uses a shallow history Feb 27 23:20:58 and i think i'm missing packets :\ Feb 27 23:21:03 and the checkout is in feeds/ Feb 27 23:21:12 with src-link i can keep the checkout of the package feed anywhere i like Feb 27 23:21:40 yeah, that's what I want Feb 27 23:21:51 * solca redoing it's feeds Feb 27 23:23:07 jow * r25780 /packages/net/dante/ (6 files in 2 dirs): [packages] add dante - provides libsocks, sockd and socksify Feb 27 23:24:12 it seems doesn't work with relative paths Feb 27 23:24:30 it does but its one level down Feb 27 23:24:30 yeah, you have to use absolute paths Feb 27 23:24:32 ah Feb 27 23:24:36 ../ Feb 27 23:24:51 xMff: I tried one level ( ../ ) Feb 27 23:24:58 then it was two :) Feb 27 23:25:22 * solca retrying... Feb 27 23:25:23 or just use an absolute path ;) Feb 27 23:25:27 or that Feb 27 23:27:01 jow * r25781 /packages/net/transocks/ (. Makefile): [packages] add transocks - a transparent SOCKS relay proxy Feb 27 23:27:22 it works relative with two levels :) Feb 27 23:27:41 it's 'make package/symlinks' necessary? Feb 27 23:28:03 no Feb 27 23:28:10 that's old stuff Feb 27 23:28:15 we just left it in for compatibility Feb 27 23:28:20 you can use ./scripts/feeds install Feb 27 23:29:25 install when a package is right? Feb 27 23:29:37 no Feb 27 23:29:38 it it's <*> should be in the image Feb 27 23:29:53 generates a package Feb 27 23:29:57 <*> puts it into the image Feb 27 23:30:15 ./scripts/feeds install makes a package visible in menuconfig Feb 27 23:31:56 ./scripts/feeds install -a -p packages ? Feb 27 23:33:03 I already see all packages but I did a 'make package/symlinks' Feb 27 23:33:15 thats equivalent to install -a -p packages Feb 27 23:33:29 a wrapper so to say Feb 27 23:34:12 xMff: so how do I reset to a clean state the packages symlinks? Feb 27 23:34:20 to start over? Feb 27 23:34:25 ./scripts/feeds uninstall -a Feb 27 23:36:09 alright! finally I have this packages feeds under control! thx Feb 27 23:40:34 './scripts/feeds install -a' seems to do the whole work after 'make distclean' Feb 28 00:18:15 jow * r25782 /packages/net/transocks/Makefile: [packages] transocks: add missing dependency on libsocks Feb 28 00:44:16 ping nbd Feb 28 00:47:18 nbd: this changeset https://dev.openwrt.org/changeset/25768 breaks the build for me - im already using 2.6.37.2 btw - when i remove the patch the kernel compiles fine and works ... http://nopaste.info/092fe3c7de.html Feb 28 00:47:56 ok, i'll fix Feb 28 00:52:39 nbd * r25783 /packages/net/samba3/ (3 files in 2 dirs): samba3: remove rpc support, fix various compile errors Feb 28 01:17:49 nbd * r25784 /packages/net/mtr/patches/502-fix-res_ninit.patch: mtr: remove some dead code Feb 28 01:21:17 nbd * r25785 /trunk/target/linux/generic/ (2 files in 2 dirs): kernel: the usb pci quirks removal patch was a bit overzealous, leave in some functions referenced from other modules Feb 28 01:21:18 tripolar: fixed Feb 28 01:26:56 thanks Feb 28 01:27:09 ill go now have a nice night :) Feb 28 01:46:59 cshore * r25786 /trunk/package/block-mount/files/mount.sh: [package] block-mount,block-extroot: Modify mount.sh so that block-extroot mounts target /overlay as rootfs (use if is_rootfs is now deprecated) when block-extroot is installed, and as /tmp/overlay-disabled otherwise Feb 28 01:47:14 cshore * r25787 /trunk/package/block-mount/files/mount.sh: [package] block-mount: Fixed shell error on empty string due to -eq to number for the optional find_root variable **** ENDING LOGGING AT Mon Feb 28 02:59:57 2011