**** BEGIN LOGGING AT Mon Mar 28 02:59:57 2011 Mar 28 03:35:42 build #82 of etrax is complete: Failure [failed compile_3] Build details are at http://tksite.gotdns.org:8010/builders/etrax/builds/82 Mar 28 04:57:02 philipp64|laptop: why not ask your question, maybe someone knows the answer ... can't tell without asking Mar 28 06:37:50 russell--: I had done a commit to my local git cloning, but wanted to revert it to the upstream version. Mar 28 06:39:33 git reset --hard Mar 28 06:40:10 generally, you probably want to create a branch for your local changes Mar 28 06:40:13 commit to them Mar 28 06:40:37 git checkout -b Mar 28 06:41:31 commit any changes to your branch, then checkout master when you want to pull from upstream again Mar 28 06:43:10 then, if you want to build with the new upstream and your patches, do something like this: git checkout master ; git checkout -b mybuild ; git merge mychanges-branch Mar 28 07:05:01 whaat? Mar 28 07:05:04 we're using git now? Mar 28 07:23:55 no... I'm trying to get my device driver fixes on the fast-track upstream, and linux (netdev) uses git. Mar 28 07:33:25 morning Mar 28 07:52:37 <[florian]> KanjiMonster: allright, thanks! Mar 28 09:23:56 build #99 of ramips is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/ramips/builds/99 Mar 28 10:03:10 philipp64|laptop: does that make sense? Mar 28 12:04:18 acoul * r26331 /trunk/target/linux/atheros/patches-2.6.37/240-ar231x_netpoll.patch: linux/atheros: add missing netpoll option. (closes #9130) Mar 28 14:05:16 build #82 of iop32x is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/iop32x/builds/82 Mar 28 14:57:41 build #78 of mpc52xx is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/mpc52xx/builds/78 Mar 28 15:48:04 mb * r26332 /trunk/ (8 files in 6 dirs): Update 2.6.38 to 2.6.38.2 Mar 28 15:50:17 build #77 of octeon is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/octeon/builds/77 Mar 28 15:57:30 mb * r26333 /trunk/target/linux/omap24xx/config-2.6.38: Update omap24xx kconfig Mar 28 17:15:20 build #88 of ar71xx is complete: Failure [failed shell compile_6] Build details are at http://tksite.gotdns.org:8010/builders/ar71xx/builds/88 Mar 28 17:18:05 build #35 of lantiq is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/lantiq/builds/35 Mar 28 17:22:27 build #113 of at91 is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/at91/builds/113 Mar 28 18:06:42 ping KanjiMonster Mar 28 18:08:03 KanjiMonster: I think I asked before, but is 2.6.37 currently up-to-date and working for brcm63xx? I think we need to update it soon, but I don't have much time to work on it at the moment, so I'm hoping you already know the answer. Mar 28 18:16:09 build #97 of pxcab is complete: Failure [failed compile_3] Build details are at http://tksite.gotdns.org:8010/builders/pxcab/builds/97 Mar 28 18:20:15 cshore: the patches are from me, and it's working on my devices (haven't tested my 6348 though, only 6358) Mar 28 18:20:48 KanjiMonster: ok, I'll try to make time to test on 6348 then Mar 28 18:21:59 cshore: I think I successfully booted one of the .37-rc's on my 6348, but haven't tested any newer kernels Mar 28 18:22:19 Can an Openwrt dev take a look at #8601? It should be fixed with the attached patches Mar 28 18:22:25 https://dev.openwrt.org/ticket/8601 Mar 28 18:23:01 KanjiMonster: did you send the patches to the mailing list (I have to hunt them down again)? Mar 28 18:23:17 cshore: they are in the svn already ;) Mar 28 18:23:31 KanjiMonster: oh, ok, just needs testing Mar 28 18:23:40 KanjiMonster: then made default Mar 28 18:23:44 right Mar 28 18:24:06 KanjiMonster: I will try to find time this week to do that Mar 28 18:25:59 cshore: these patches also work for 2.6.38, except that for 2.6.38 squashfs isn't currently working since mksquashfs doesn't create xz images yet Mar 28 18:40:39 Originally not my intention but as I sent to the mailing list Mar 28 18:40:53 if somebody know about a job please let me know ;) Mar 28 19:08:56 solca: just make your work more visible, then maybe the jobs come to you ;) Mar 28 19:09:01 that's the nice thing about open source Mar 28 19:13:28 nbd: yeah, thanks, I would love to work all day on embedded stuff ;) Mar 28 19:13:44 nbd: OTOH what would you recommend to make my work more visible? Mar 28 19:15:14 maybe write about what you're working on, especially if it's something that could be interesting for many people Mar 28 19:15:25 or do enough stuff publicly that people can't help but notice your work Mar 28 19:15:31 if they're working on a related field Mar 28 19:16:44 worked for me with ath9k Mar 28 19:17:15 though i didn't really work on it with the intention of attracting attention Mar 28 19:17:22 i just wanted stable 11n wifi ;) Mar 28 19:18:04 picky picky Mar 28 19:21:13 nbd: ok thank you for the advice, I'm just to finish a long work and then I'll have time Mar 28 19:21:26 for example I want to do something with the borg stuff Mar 28 19:21:48 I know a full GPL android will be very successfull Mar 28 19:21:49 great. a few days ago i got in contact with another guy here in berlin that is interested in this as well Mar 28 19:21:56 i'm meeting him again the day after tomorrow Mar 28 19:22:21 i'm really interested in resuming my work on this as well Mar 28 19:22:40 just haven't done so because right now i don't have time to go deep on this yet Mar 28 19:22:59 cool, it's a lot of work if one wants all cases covered Mar 28 19:23:08 yeah, definitely Mar 28 19:23:27 there are so many things in android that really need some serious fixing Mar 28 19:23:50 and some areas touch parts where google has financial reasons to have a different opinion ;) Mar 28 19:23:56 yes that's for sure! Mar 28 19:24:10 and now that android is not developing it in the open it sucks Mar 28 19:24:22 android was never developed in the open Mar 28 19:24:38 they've just been a bit more honest about holding it back this time Mar 28 19:24:42 before 1.6 was in the open Mar 28 19:24:43 which is shocking people Mar 28 19:24:47 I pull daily Mar 28 19:24:59 there's some limited merging of community patches Mar 28 19:25:14 well that's true Mar 28 19:25:16 that's not open development, when google makes huge code drops for new releases Mar 28 19:25:30 with stuff that they've been developing in private over longer periods of time Mar 28 19:26:01 not that this is in any way surprising Mar 28 19:26:08 they just merge simple patches but from time to time they make some big code drops Mar 28 19:26:17 since google needs to keep tight control over android so that it continues to do what they need it to do Mar 28 19:26:36 but after 1.6 they just release code after a new version comes out Mar 28 19:27:13 in fact they used internally something different than git (per...) Mar 28 19:27:17 i hope that we can make our stuff interesting enough so that we can eventually assimilate stuff like cyanogenmod Mar 28 19:27:34 quite the ambition Mar 28 19:27:34 not just the code, but the people as well Mar 28 19:27:46 they are doing some good work Mar 28 19:28:13 but my goal is to make it possible to easily deviate more from what google does without having to deal with a huge fork maintenance overhead Mar 28 19:28:27 and i think that's something the cyanogenmod guys would appreciate as well Mar 28 19:28:35 yea, when cyanogenmod bases it stuff on borg which based on openwrt that will be a major win for all! Mar 28 19:30:53 got_milk: working without ambition is boring ;) Mar 28 19:32:37 true enough Mar 28 19:45:03 nbd: so let me know what are your plans with the other guy re borg Mar 28 19:45:26 * solca knows germans are extremelly good hackers Mar 28 19:45:38 :) Mar 28 19:46:10 sure, i'll let him know that you're interested in this project as well, so maybe at some point we should set up a mailing list for that Mar 28 19:46:44 cool! Mar 28 21:26:18 florian * r26334 /packages/multimedia/ffmpeg/Makefile: Mar 28 21:26:18 [package] update ffmpeg to 0.5.4 Mar 28 21:26:18 Version bump ffmpeg from 0.5.2 to 0.5.4. 0.5.3 and 0.5.4 are maintenance only releases for security and minor bug fixes. Mar 28 21:26:18 Signed-off-by: Ian Leonard Mar 28 21:26:20 florian * r26335 /packages/net/gpsd/Makefile: (log message trimmed) Mar 28 21:26:21 [package] gpsd: use autogen.sh Mar 28 21:26:21 attach is the patch, fix the compile error: Mar 28 21:26:21 libtool: Version mismatch error. This is libtool 2.2.6 Debian-2.2.6a-4, but the Mar 28 21:26:21 libtool: definition of this LT_INIT comes from libtool 2.2.6b. Mar 28 21:26:21 libtool: You should recreate aclocal.m4 with macros from libtool 2.2.6 Debian-2.2.6a-4 Mar 28 21:26:22 libtool: and run autoconf again. Mar 28 21:26:23 florian * r26336 /packages/admin/pmacct/ (Makefile patches/ patches/001-honor_cflags.patch): Mar 28 21:26:23 [package] update pmacct to 0.12.5 Mar 28 21:26:24 This patch update pmacct from 0.10.3 (21-06-2006) to 0.12.5 (28-12-2010) Mar 28 21:26:24 Thanks in advance. Mar 28 21:26:25 Signed-off-by: Etienne CHAMPETIER Mar 28 21:26:27 florian * r26337 /packages/utils/usbutils/Makefile: Mar 28 21:26:27 [package] update usbutils to 002 Mar 28 21:26:27 Update to version 002, also adds dependencies on librt and libpthread Mar 28 21:26:27 Signed-of-by: Maarten Bezemer Mar 28 21:26:32 florian * r26338 /packages/libs/avahi/ (Makefile files/avahi-daemon.init): (log message trimmed) Mar 28 21:26:33 [package] update avahi to included D-Bus support, libavahi-client library and avahi-utils tools (#8929) Mar 28 21:26:33 This updates the avahi package as follows: Mar 28 21:26:33 1. The libavahi-client library can be selected. Mar 28 21:26:33 2. The following avahi utilities can be selected: Mar 28 21:26:44 The build fails since the php module is statically disabled. This patch Mar 28 21:26:44 adds the required magic. Mar 28 21:26:45 This should also finally close #7758. Mar 28 21:26:45 Signed-off-by: Michael Heimpold Mar 28 21:26:46 florian * r26342 /packages/net/freeradius2/ (Makefile patches/009-sql_sqlite_c.patch): Mar 28 21:26:46 [package] freeradius2: add rlm_sql_sqlite and rlm_passwd Mar 28 21:26:47 This patch adds rlm_sql_sqlite, rlm_passwd modules for freeradius2. Mar 28 21:26:47 It patches the rlm_sql_sqlite in order to work (authorization is working, accounting not yet tested). Mar 28 21:26:48 It adds radwho to the utils. Mar 28 21:26:48 Thanks in advance. Mar 28 21:26:49 Signed-off-by: Etienne CHAMPETIER Mar 28 21:34:34 florian * r26343 /packages/utils/flashrom/Makefile: Mar 28 21:34:34 [package] flashrom: update to 0.9.3 and fix dependency Mar 28 21:34:34 Bump flashrom to current version, and reflect that it uses dmidecode as a helper. Mar 28 21:34:34 Signed-off-by: Philip Prindeville Mar 28 21:35:06 florian * r26344 /trunk/target/linux/x86/geos/target.mk: Mar 28 21:35:06 [x86] geos: support cpu MSR register and flashrom for coreboot upgrades Mar 28 21:35:06 The early versions of the platform shipped with coreboot 3.0, which had some issues rebooting correctly the first time. Mar 28 21:35:06 Allow the user to upgrade his BIOS from Linux. Mar 28 21:35:06 Signed-of-by: Philip Prindeville Mar 28 22:07:30 florian * r26345 /packages/libs/libmodbus/ (. Makefile): Mar 28 22:07:30 [package] add libmodbus Mar 28 22:07:30 Apparently people are already using libmodbus on OpenWrt, but there is no package. Mar 28 22:07:30 (see http://zagrodzki.net/~sebek/openwrt/packages/libmodbus.so ) Mar 28 22:07:30 So I made a Makefile for it, I'd be happy to be the package maintainer for that. Mar 28 22:07:32 florian * r26346 /packages/utils/collectd/ (Makefile patches/300-fix-modbus.patch): Mar 28 22:07:32 [package] collected: add modbus plugin. Mar 28 22:07:32 This enables the modbus plugin of collectd and patches it so it works with the current version of libmodbus Mar 28 22:07:32 Signed-off-by: Daniel Golle Mar 28 22:07:33 florian * r26347 /packages/utils/mc/ (Makefile patches/100-fix-missingsyntax.patch): Mar 28 22:07:34 [package] upgrade mc to 4.7.5.1 Mar 28 22:07:34 Signed-off-by: Luka Perkov < openwrt ->-to->- lukaperkov.net > Mar 28 22:36:07 xMff: I don't think the "" is the problem after all. I do use config_get_bool, and when I disable extroot (comment out the look for rootfs in the extroot scripts), I get things mounted on /tmp/overlay-disabled as it should be. I don't see why trumee was getting a double mount when rootfs mount failed. Mar 28 22:36:40 cshore: users manage to come up with weird unreproducable errors all the time... Mar 28 22:37:22 often its unclean trees, not quite correct config, incoherent states, outdated versions Mar 28 22:37:56 btw, did you look int othe overlay enable action? Mar 28 22:38:06 xMff: ah, you missed my post Mar 28 22:38:14 xMff: I did it Mar 28 22:38:19 ah great Mar 28 22:38:29 because users begun to develop myths already Mar 28 22:38:41 like format drive, or copy magic md5sums etc. Mar 28 22:39:04 /etc/init.d/overlay_enable for overlay-type and /etc/init.d/whole_root_enable for the non-overlay type Mar 28 22:39:11 xMff: interesing Mar 28 22:39:40 uhm, so you did a whoƶe new init script? Mar 28 22:39:45 *whole Mar 28 22:39:58 xMff: nope, just added EXTRA_COMMANDS Mar 28 22:40:02 ah Mar 28 22:40:15 I suggest to rename "whole_root_enable" to "root_enable" Mar 28 22:40:19 xMff: you told me about that Mar 28 22:40:25 yes, I did Mar 28 22:40:36 but you just wrote "/etc/init.d/overlay_enable" Mar 28 22:41:00 xMff: oh, sorry /etc/init.d/fstab overlay_enable Mar 28 22:41:09 great Mar 28 22:41:14 I assume its in the wiki too? Mar 28 22:41:30 let me check if I actually wrote that or not Mar 28 22:42:06 wow who injected florian with caffeine Mar 28 22:42:14 http://wiki.openwrt.org/doc/howto/extroot#system.upgrade Mar 28 22:47:05 xMff: could you edit that sometime, or add/change to your thoughts....you interact with users more than I do. Mar 28 22:47:20 yes but might take a while Mar 28 22:47:35 I'm a bit out of the loop atm Mar 28 22:48:34 xMff: been too busy to follow the commits? Mar 28 22:49:00 well, kind of. You loose track very fast if you don't do it daily :) Mar 28 22:49:09 xMff: so I've noticed Mar 28 22:53:52 hell and damnation... we version bumped to a broken version of flashrom. Mar 28 22:54:13 it builds fine, but has a regression when erasing flash hanging off a CS5536 controller. Mar 28 22:54:36 yay Mar 28 22:54:43 I'll investigate tomorrow as to whether 0.9.2 is also broken. Mar 28 22:54:43 just revert it? Mar 28 22:54:53 yeah, for now. sorry for the trouble. Mar 28 22:55:08 its not like its used by thousands of people right now anyway Mar 28 22:55:24 actually, the dependence on dmidecode was correct, though. as was the correction for the typo. :-) Mar 28 22:55:56 so dmidecode dep ok, but version downgrade? Mar 28 22:56:16 I'll submit another patch tonight, resetting the version, but fixing the typo, the dependency, and fixing the path to /dev/msr0 and not /dev/cpu/0/msr ... Mar 28 22:56:31 yeah, in short, yes. Mar 28 22:56:50 all that's missing (other than winding back the verison) is the MSR device path patch. Mar 28 22:57:04 ok, hitting the road... Mar 28 22:59:14 xMff: oh yeah, I've also merged block-mount, block-hotplug, and block-extroot into a single block-mount package Mar 28 22:59:23 nice Mar 28 22:59:32 I think everybody just installed all of them anyway :) Mar 28 22:59:39 flyn * r26348 /packages/ (libs/libdmapsharing/Makefile net/dmapd/Makefile): Update dmapd and libdmapsharing to new upstream versions Mar 28 22:59:56 xMff: or forgot one and wished they had Mar 28 23:00:11 which reminds me on the discussions on the list back then about 2MB devices Mar 28 23:00:24 I said the packaging overhead is complicating stuff Mar 28 23:01:35 xMff: yeah, you were right. Too busy trying to keep crazy people (2MB with a 2.6 kernel?) happy Mar 28 23:01:56 at least in the present state of the kernel anyway Mar 28 23:02:10 yeah, and in the end they just use custom solutions anyway Mar 28 23:02:23 yeah Mar 28 23:02:45 Though I did add the ability to do non-overly rootfs Mar 28 23:02:56 it was almost no work Mar 28 23:04:18 which is, together with the preinit splicing, a great thing since it consolidates all the previous pivot root hacks Mar 28 23:04:35 xMff: yep Mar 28 23:04:50 xMff: now I just have to get the overlay fixing stuff worked out Mar 28 23:05:08 you mean the duplicate mounts? Mar 28 23:05:31 xMff: not the the script for making it safe to enable again after flashing Mar 28 23:05:35 *no Mar 28 23:05:42 ah Mar 28 23:06:06 I'm not worrying about the duplicate mounts unless I get more reports because I can't replicate the problem Mar 28 23:06:25 I did before, but I what I did should have fixed it Mar 28 23:06:42 so it's probably some inconsistency in build or user state Mar 28 23:07:10 xMff: do the buildbots do a fresh checkout? Mar 28 23:07:29 yes, distclean, svn up, echo ... > .config; make defconfig; make Mar 28 23:08:26 xMff: ok, so probably not a build state issue then Mar 28 23:08:31 echo ... > .config <-- wait you can echo the packages you want and defconfig fills in the blank Mar 28 23:08:37 ? Mar 28 23:08:38 yes Mar 28 23:08:49 FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF Mar 28 23:08:58 HOW DID I NOT KNOW THIS Mar 28 23:09:29 http://nbd.name/release-build.sh Mar 28 23:09:34 see prepare_config() Mar 28 23:09:41 buildbots do something very similar Mar 28 23:09:49 without luci obviously Mar 28 23:12:44 xMff: wait, will defconfig figure out deps? or do i have to echo everything thats not a stock package? Mar 28 23:14:50 xMff: apparently some had bad overlay creation instructions on the wiki Mar 28 23:14:54 *someone Mar 28 23:15:05 Weedy: it will figure out deps of the packages you marked as =y or =m Mar 28 23:15:35 Not only were they intending to copy all of /, they got it wrong and were copying nothing Mar 28 23:21:07 yeah, glad you reviewd it now Mar 28 23:22:14 me too - I don't like chasing ghosts Mar 28 23:23:05 I also got rid of configuration notes in the build instructions and pointed that page at the appropriate extroot page section Mar 28 23:23:22 Weedy: and ./scripts/diffconfig.sh generates compact config data that can be used as input for make defconfig Mar 28 23:23:40 also makes it easy to see what was changed compared to the default config Mar 28 23:23:55 nbd: yes it is a cool script. Thanks for it btw Mar 28 23:26:46 i'm currently extending it so that it handles CONFIG_ALL, CONFIG_DEVEL and CONFIG_TOOLCHAINOPTS properly as well Mar 28 23:26:59 that will reduce the output for config files with those options set Mar 28 23:28:09 ah, nice. all i need to do is move these options to .diffconfig.head Mar 28 23:28:14 then the rest is handled properly automatically Mar 28 23:29:08 fffffffffffffffffffffffffffffffff Mar 28 23:29:18 so much time wasted all these years Mar 28 23:29:21 nbd * r26349 /trunk/scripts/diffconfig.sh: scripts/diffconfig.sh: handle CONFIG_ALL, CONFIG_DEVEL and CONFIG_TOOLCHAINOPTS properly Mar 28 23:29:30 wait how old is openwrt? Mar 28 23:29:41 i was still in highschool... Mar 28 23:30:14 nbd * r26350 /branches/backfire/scripts/diffconfig.sh: scripts/diffconfig.sh: handle CONFIG_ALL, CONFIG_DEVEL and CONFIG_TOOLCHAINOPTS properly (backport of r26349) Mar 28 23:33:38 i should really use my patchteam access Mar 28 23:34:24 * nbd is currently writing documentation about protocol handlers in netifd Mar 28 23:35:01 didn't we have a webapp that tags patches from the mailing list Mar 28 23:35:15 * cshore has been distracted from work by openwrt for the afternoon Mar 28 23:35:32 oh Mar 28 23:35:34 found it Mar 28 23:46:12 http://nbd.name/gitweb.cgi?p=luci2/netifd.git;a=blob;f=DESIGN;hb=HEAD - updated for protocol handlers Mar 28 23:48:18 i think with that design, netifd should stay responsive even if protocol handlers need a lot of time for setup/teardown Mar 28 23:48:50 and i intend to add some helper functions to uloop to deal with child process completion Mar 28 23:49:03 that should make task management straightforward enough Mar 28 23:50:17 and of course i will still support an easy way of writing a protocol handler in shell **** ENDING LOGGING AT Tue Mar 29 02:59:57 2011