**** BEGIN LOGGING AT Thu Apr 04 02:59:58 2013 Apr 04 05:23:28 build #198 of au1000 is complete: Failure [failed compile_9] Build details are at http://buildbot.openwrt.org:8010/builders/au1000/builds/198 Apr 04 07:00:22 cyrusff: netifd w/ ip6hint works great for my use case. Thanks a lot! Apr 04 07:00:42 sungami: you're welcome. glad it works as expected Apr 04 08:51:48 build #182 of mcs814x is complete: Failure [failed compile_9] Build details are at http://buildbot.openwrt.org:8010/builders/mcs814x/builds/182 Apr 04 11:34:52 build #211 of avr32 is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/avr32/builds/211 Apr 04 11:55:08 hi folks. is there a way to force 'rm' (unlink syscall) to actually delete the file without waiting for the close of the last file descriptor open? Apr 04 11:55:25 <_rmk_> no. Apr 04 11:55:29 I'm in trouble with space in disk... Apr 04 11:55:46 <_rmk_> you could truncate it if you haven't already deleted it Apr 04 11:58:17 my issue is with opkg. when it is upgrading a package, it first removes the old version to then copy the new one. the problem is that the old file is not really deleted and then the copy of the new file fails due to space in disk Apr 04 12:03:06 hmmm maybe patch opkg to truncate the file before unlink it... Apr 04 12:08:12 jwendell: are you using squashfs by chance? Apr 04 12:08:34 jffs2 Apr 04 12:09:25 ok, but well you cannot really do anything about that matter, i guess opkg itself doesn't hold the old files open Apr 04 12:09:46 it calls unlink to remove old files Apr 04 12:10:04 it's just a matter of call truncate() before unlink() Apr 04 12:10:11 *I guess* :) Apr 04 12:10:47 well the problem is this can lead to corruption if the files are in use Apr 04 12:11:21 e.g. an executable / library that is currently being used or a file used by an executable and so on Apr 04 12:12:13 the data is most likely kept for a reason Apr 04 12:12:42 indeed, that's what I'm afraid Apr 04 12:13:02 but, in theory, all libraries in use are in memory, right? Apr 04 12:14:07 my use case is related to libopenssl. it's in use by openvpn binary. I need to do some tests.... Apr 04 12:15:02 well libs are usually copy-on-write memory-mapped Apr 04 12:15:09 from the files Apr 04 12:16:13 mmaps and truncates together can get nasty Apr 04 12:57:42 uhm, can anyone help me out with this Makefile? http://paste.debian.net/247226/ basically i copied the net/bird , added -devel and "tried" to make it get the source from git.. it doesnt work Apr 04 12:58:03 of course it sits in net/bird-devel directory Apr 04 13:01:40 and this is the error i get: http://paste.debian.net/247235/ Apr 04 13:18:29 dape8708, it's missing something like PKG_SOURCE:=$(PKG_NAME)-$(PKG_VERSION).tar.bz2 Apr 04 13:36:56 when source is git ? Apr 04 13:45:56 dape8708, yep, this is use internally by the build system. it checks out the git and make a tarball upon it Apr 04 13:46:19 the it extracts that tarball and builds it Apr 04 13:46:27 *then Apr 04 13:50:31 ah.. duhh.. thanks for enlighting me Apr 04 13:55:41 jwendell, if i may: if the previous bird was using a patches/0000-blabla and this one doesn't need one.. how can i use svn diff to mark that it doesn't need a patch? Apr 04 13:56:04 if i run it i get the Makefile diff but it complains about "svn: Directory 'feeds/packages/net/bird/patches' is missing" .. Apr 04 14:01:59 meh, got it. subversion stuff :D Apr 04 14:08:33 there, on the openwrt-devel list, the patch format is okay Apr 04 14:09:30 your mailer broke long lines Apr 04 14:11:33 thats the punishment for using the crap here instead the mutt at home, let me check some stuff anyway Apr 04 14:16:35 dape8708: git send-email ? Apr 04 14:16:47 works everytime Apr 04 14:21:23 looks awesome, let me check git help send-email, thanks for hint Apr 04 14:27:22 build #237 of orion is complete: Failure [failed compile_8] Build details are at http://buildbot.openwrt.org:8010/builders/orion/builds/237 Apr 04 14:58:27 blogic r36194 trunk/package/mac80211/patches/617-rt2x00-load-eeprom-on-SoC-from-a-mtd-device-defines-.patch Apr 04 14:58:27 [mac80211] add a helper that allows loading the eeprom from a MTD partition pointed to by a phandle Apr 04 15:51:56 gosh i hope this time the patch for bird looks okay Apr 04 16:32:29 build #238 of orion is complete: Failure [failed compile_8] Build details are at http://buildbot.openwrt.org:8010/builders/orion/builds/238 Apr 04 17:42:00 build #181 of iop32x is complete: Failure [failed shell_12] Build details are at http://buildbot.openwrt.org:8010/builders/iop32x/builds/181 Apr 04 18:43:40 the PCI and the USB don't work in the arv4518pw r01A Apr 04 18:44:43 maybe an issue in the EBU or a gpio misconfigured in some Lantiq boards Apr 04 18:45:07 better open a ticket? Apr 04 18:52:24 maybe more debug Apr 04 20:05:38 nbd r36195 trunk/package/ubus/Makefile * ubus: update to latest version, fixes a crash on reconnect Apr 04 20:08:50 Pteridium: same thing on 7518pw Apr 04 20:13:14 build #184 of ep93xx is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/ep93xx/builds/184 Apr 04 20:14:32 Delboy: sure? Apr 04 20:15:06 i go to test with 3.7 kernel Apr 04 20:17:30 build #164 of octeon is complete: Failure [failed shell_12] Build details are at http://buildbot.openwrt.org:8010/builders/octeon/builds/164 Apr 04 20:19:13 Delboy: i go to change in trunk/target/linux/lantiq/Makefile the kernel version 3.7.9 instead 3.8.10 Apr 04 20:21:16 blogic made a lot of changes and improvements for 3.8 kernel so is very easy to forgot something Apr 04 20:32:21 Pteridium: i didn't realize they had released 3.8.10! ;-) Apr 04 20:36:32 russell--: no, the actual kernel for the lantiq target is 3.8.3 Apr 04 20:36:50 at least for the danube SoC Apr 04 20:39:46 confirmed, 3.8.3; but with the kernel change he made a lot of improvements and rewrote almos all the patches. Apr 04 20:41:47 was a very big job and a lot of changes, and logically some issues happen; it's normal Apr 04 20:46:16 * russell-- was really just commenting on the fact that kernel.org says 3.8.5 is most recent, not 3.8.10 ... i got excited for a minute ;-) Apr 04 20:46:36 XD Apr 04 20:48:00 and if you can test 3.9-rc1 Apr 04 20:48:16 rc5! Apr 04 20:49:18 Wow! Apr 04 20:50:40 maybe in some days will be released 3.9.1 Apr 04 20:58:39 What the hell is happening in the wiki? Apr 04 20:58:41 http://wiki.openwrt.org/profile_nancyfxtnkiuthw Apr 04 21:01:28 it is called spam Apr 04 21:02:06 and it makes no sense to write "SPAM" into each page, that only makes it harder for me to look up the original author Apr 04 21:02:22 ok Apr 04 21:02:50 then only notification? Apr 04 21:02:50 <_trine> yes and she might be offering a deal you cant refuse :P Apr 04 21:03:07 Pteridium: would be better yes, if you like simply pm me the links of spammy pages Apr 04 21:03:37 ok, i'll do as you say, thanks Apr 04 22:51:56 Delboy: changing in (trunk_dir)/target/linux/lantiq/Makefile the kernel version 3.8.3 by 3.7.9 logically usb and wifi works Apr 04 22:55:29 but still couldn't find why usb and pci don't work. Maybe EBU, a couple of stolen gpios, a lost irq or any other simple thing. Apr 04 22:55:46 good night Apr 05 00:31:28 build #199 of au1000 is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/au1000/builds/199 Apr 05 01:26:26 build #183 of mcs814x is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/mcs814x/builds/183 Apr 05 02:32:45 build #177 of gemini is complete: Failure [failed compile_8] Build details are at http://buildbot.openwrt.org:8010/builders/gemini/builds/177 Apr 05 02:53:32 anyone here? **** ENDING LOGGING AT Fri Apr 05 02:59:58 2013