**** BEGIN LOGGING AT Sun Dec 26 02:59:56 2010 Dec 26 03:17:18 xMff: Here is the print out of bootstrap.sh about its usage of autoconf, automake, aclocal, libtool, and libtoolize: Dec 26 03:17:22 Bootstrapping using: Dec 26 03:17:24 autoconf : /opt/tmp/openwrt-svn-trunk/staging_dir/host/bin/autoconf Dec 26 03:17:25 automake : /opt/tmp/openwrt-svn-trunk/staging_dir/host/bin/automake Dec 26 03:17:27 aclocal : /opt/tmp/openwrt-svn-trunk/staging_dir/host/bin/aclocal Dec 26 03:17:29 libtool : /opt/tmp/openwrt-svn-trunk/staging_dir/host/bin/libtool (2.4.) Dec 26 03:17:30 libtoolize: /opt/tmp/openwrt-svn-trunk/staging_dir/host/bin/libtoolize Dec 26 03:54:55 jow * r24831 /packages/Xorg/lib/gtk2/Makefile: [packages] gtk2: depend on glib2, remove libintl, libiconv stuff, that is glib2's business Dec 26 04:08:42 jow * r24832 /packages/lang/pygtk/patches/ (. 001-automake-compat.patch): [packages] pygtk: fix autoreconf Dec 26 04:20:14 cshore * r24833 /trunk/package/mtd/src/ (mtd.c mtd.h trx.c): Dec 26 04:20:14 mtd: Changed mtd_fixtrx and related functions to be weak references. The weak references only lead to actual functions for brcm47xx Dec 26 04:20:14 Signed-off-by: Daniel Dickinson Dec 26 04:20:19 cshore * r24834 /trunk/target/linux/brcm63xx/patches-2.6.35/040-bcm963xx_flashmap.patch: Dec 26 04:20:19 mtd flashmap: Switched to calculating the rootfs size rather than depending on Dec 26 04:20:19 being given the size in the imagetag. This is because solving the problem of Dec 26 04:20:19 second boot CRC errors requires changeing the rootfs size in the image to zero. Dec 26 04:20:20 Signed-off-by: Daniel Dickinson Dec 26 04:20:23 cshore * r24835 /trunk/package/mtd/src/ (Makefile bcm_tag.h imagetag.c mtd.c): Dec 26 04:20:23 mtd: Added fixtrx for brcm63xx imagetag. This allows brcm63xx boards which experience a CRC on second boot due to the JFFS2 creation / DEADC0DE deletion to work correctly when mtd fixtrx is run on first boot Dec 26 04:20:23 Signed-off-by: Daniel Dickinson Dec 26 04:20:26 cshore * r24836 /trunk/target/linux/brcm63xx/base-files/etc/uci-defaults/ (. brcm63xx_fixcrc.sh): Dec 26 04:20:27 base-files: Added uci-defaults script to run mtd fixtrx on firstboot for devices that need it. Dec 26 04:20:27 Signed-off-by: Daniel Dickinson Dec 26 04:20:30 cshore * r24837 /trunk/package/mtd/src/ (imagetag.c jffs2.c): Dec 26 04:20:30 mtd: Added trx_fixup for brcm63xx imagetag, and made references to fix_trx use the weak reference rather than the brcm47xx ifdef. This fixes a bug in which sysupgrade failed due to changing bad CRC on reboot. Dec 26 04:20:30 Signed-off-by: Daniel Dickinson Dec 26 04:20:35 cshore * r24838 /trunk/ (4 files in 4 dirs): Dec 26 04:20:35 Added OpenWRT-specific field to imagetag so that we can record the real root length, so that when the CRC fixup is applied and the root length is recorded as zero we can still calculate the rootfs mtd partition size. Dec 26 04:20:35 Signed-off-by: Daniel Dickinson Dec 26 10:53:34 gmorning Dec 26 12:57:28 build #43 of adm5120 is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/adm5120/builds/43 Dec 26 12:58:03 build #49 of ramips is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/ramips/builds/49 Dec 26 13:02:03 build #49 of rb532 is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/rb532/builds/49 Dec 26 13:58:07 updated openwrt/upstream, https://home.comcast.net/~sdwalker/uscan/uscan.shtml Dec 26 15:05:58 build #42 of rdc is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/rdc/builds/42 Dec 26 16:06:25 build #47 of sibyte is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/sibyte/builds/47 Dec 26 16:08:44 xMff: R U available? Dec 26 19:01:11 nedos: eh, just found your email in my gmail spam foder Dec 26 19:09:48 thepeople: ping Dec 26 19:21:26 Just want to report back. In compiling libs/flac, if I don't manually first compiled libs/libogg and libs/libvorbis, I get a configuration error as shown in http://pastebin.com/WxUZazGB Dec 26 19:58:02 dumb question... which kernel.org git do netdev patches show up in? Dec 26 20:00:25 philipp64|laptop: netdev patches should be in linux/kernel/git/davem/net-next-2.6.git at first Dec 26 20:02:52 I'm showing that hasn't changes in a while... 2 years or so. but romieu has. Dec 26 20:03:26 philipp64|laptop: no the last change was 16h ago Dec 26 20:03:38 philipp64|laptop: http://git.kernel.org/?p=linux/kernel/git/davem/net-next-2.6.git;a=summary Dec 26 20:22:55 I just did an "svn update" and things are broken... http://pastebin.com/agfcEBH4 known issue? do I need to back up to an earlier rev? Dec 26 20:24:56 Hauke: I was looking at linux/kernel/git/jgarzik/netdev-2.6.git Dec 26 20:25:39 * philipp64|laptop muffles a scream as he realizes how out of date the solos-pci driver is on SF's openadsl project. Dec 26 20:28:33 -fhonour-copts ? Is that what's breaking my build? Dec 26 20:29:22 philipp64|laptop: What is your current revision? Dec 26 20:29:39 r24838 Dec 26 20:30:51 philipp64|laptop: Did you do a 'make distclean' to clean everything and start the compilation from scratch? Dec 26 20:31:52 philipp64|laptop: I have also 'svn up' this afternoon and now the compilation already passes building freeswitch. Dec 26 20:53:44 when did "make distclean" start removing .config.* ? Dec 26 20:55:45 that's a bit brutal. I'd rather have it enumerate the files to zap... Dec 26 20:56:04 especially since I tend to save .config files as .config.geos-uclibc for later use. Dec 26 20:58:49 philipp64|laptop: I told you that 'make distclean' will clean everything. Dec 26 20:59:15 well, there's varying degrees of 'everything'. :-) Dec 26 21:00:00 philipp64|laptop: True. You can do a 'make clean', make 'dirclean', etc. Dec 26 21:00:32 looks like it left dl/ intact. Dec 26 21:01:09 philipp64|laptop: AFAIK, 'make distclean' cleans everything, even the dl directory, IIRC. Dec 26 21:05:28 The freeswitch_git breaks the compilation with some undefined reference to pthread_ functions as shown in http://pastebin.com/1CGnehGu Dec 26 21:06:10 I have included the libpthread package dependency in the OpenWRT Makefile for FS git file. Dec 26 21:11:08 <_trine> mazilo, I PM'ed you Dec 26 21:11:37 _trine: PM'ed to where? Dec 26 21:12:02 <_trine> can you not see it Dec 26 21:12:34 _trine: Got it on the other chat? Dec 26 21:32:51 xMff: Have you gotten some times to fix this http://pastebin.com/1CGnehGu ? Dec 26 21:33:22 The FS git compilation couldn't seem to find the libpthread. Dec 26 22:05:18 mazilo: did you do freeswitch_git yourself (v.s the 1.0.6 that's in the packages feed)? Dec 26 22:06:25 cshore_: Yes. I followed http://wiki.freeswitch.org/wiki/Download_FreeSWITCH to fetch a git version Dec 26 22:07:43 mazilo: ah: the instructions on the freeswitch wiki for OpenWRT are incorrect Dec 26 22:08:12 cshore_: And, in my OpenWRT Maefile for FS git, I included a '(cd $(PKG_BUILD_DIR); git pull)' in the 'define Build/Prepare' section to update every time I freshly compile the git version. Dec 26 22:08:59 mazilo: how much of the existing freeswitch git did you copy? Dec 26 22:09:07 I mean freeswitch Makefile Dec 26 22:09:36 cshore_: I started out from a fresh empty directory and did a 'git clone git://git.freeswitch.org/freeswitch.git .' and then 'git pull'. Dec 26 22:10:16 mazilo: freeswith is a bear to get to install correctly for cross-compilation Dec 26 22:10:24 *compile Dec 26 22:10:25 cshore_: I used existing freeswitch-1.0.6 OpenWRT Makefile and made the necessary changes. Dec 26 22:11:12 mazilo: fs probably needs to be patched to build on OpenWRT Dec 26 22:12:19 cshore_: Yes. Dec 26 22:12:22 mazilo: are you overriding the default OpenWRT download and compile behaviour Dec 26 22:12:34 cshore_: Here is my OpenWRT Makefile for FS git version: http://pastebin.com/4tmfrHma Dec 26 22:13:28 mazilo: I think the problem is you're not doing what you think you're doing with that Makefile Dec 26 22:13:48 mazilo: You using some weird mix of the tarball and git that way Dec 26 22:14:23 cshore_: not the FS tar ball. Dec 26 22:14:52 mazilo: I'm not sure OpenWRT is able to do a compile from a git repository Dec 26 22:15:06 it is Dec 26 22:15:16 loswilios; how do you do it ? Dec 26 22:15:36 see the acx-mac80211 package for an example Dec 26 22:15:40 cshore_: After I did the git, I tarred ball the git to a file called freeswitch_git-20101225.tar.bz2 and use it from there. Dec 26 22:16:17 I think nbd even introduced branches support Dec 26 22:16:53 mazilo: is the MD5SUM of the tarball from files.blah or from the git tarball in the Makefile you posted Dec 26 22:17:03 cshore_: On my Makefile (line 12), it shows the PKG_NAME with $(NAME)_git. Dec 26 22:17:30 mazilo: but you still have PKG_SOURCE_URL:=http://files.$(NAME).org/ Dec 26 22:17:38 I think that means openwrt will try to download that Dec 26 22:17:39 cshore_: Yes and that one was generated from 'md5sum freeswitch_git-20101225.tar.bz2' Dec 26 22:18:06 cshore_ I forgot to comment that line. Dec 26 22:18:45 cshore_: I guess it won't hurt the compilation because OpenWRT sees a copy under the dl directory. Dec 26 22:26:09 cshore_: When I did a clean compilation of FS git, here is the dump that shows the extraction of the FS git tarball file: http://pastebin.com/4V4HWEfi Dec 26 22:28:17 mazilo: it looks like the problem is libtool fixups Dec 26 22:29:10 mazilo: it's not going to be easy to fix at all Dec 26 22:29:17 cshore_: Is it fixable? Dec 26 22:29:35 cshore_: OK and I understood. Dec 26 23:31:05 mazilo: those messages are uncritical Dec 26 23:37:34 hmmmm......apparently I much to learn about autotools Dec 26 23:39:48 build #44 of ixp4xx is complete: Exception [exception failed slave lost shell_13 compile_12] Build details are at http://tksite.gotdns.org:8010/builders/ixp4xx/builds/44 Dec 26 23:39:49 build #43 of au1000 is complete: Exception [exception failed slave lost shell_13 compile_12] Build details are at http://tksite.gotdns.org:8010/builders/au1000/builds/43 Dec 26 23:39:52 build #47 of ar71xx is complete: Exception [exception failed slave lost shell_13 compile_12] Build details are at http://tksite.gotdns.org:8010/builders/ar71xx/builds/47 Dec 26 23:48:05 ping xMff Dec 27 01:31:28 xMff: IC your point. I guess the problem is something else. Dec 27 01:47:36 xMff: When I started to peruse further at the dump (http://pastebin.com/1CGnehGu), I noticed line #27 "make[5]: *** [tone2wav] Error 1" indicates the actual error. If you take a look at somewhere in the middle of line #5, it shows '-o .libs/tone2wav tone2wav-tone2wav.o' which is to compile the object file (tone2wav-tone2wav.o) into an executable file (.libs/tone2wav). The compiler did generate the tone2wav-tone2wav.o file (not . Dec 27 01:47:37 libs/tone2wav file due to the error message). Also, if you take a look at the Makefile file (http://pastebin.com/iatsEegC), the line #260 indicates where the object file is. OTOH, on line #300 of the Makefile.am file (http://pastebin.com/3V1jtxn4), indicates where the tone2wav is defined. Notice automake translates tone2wav from the Makefile.am to tone2wav-tone2wav in Makefile. Is automake supposed to do that? FYI, here is the Dec 27 01:47:39 Makefile.in (http://pastebin.com/cgt8UUPv). Dec 27 01:57:25 xMff: BTW, if I manually compile the tone2wav using the line #5 of http://pastebin.com/1CGnehGu and added -lpthread at the end of line, it compiles and produces the .libs/tone2wav file. Dec 27 02:05:37 mazilo: perhaps try adding -lpthread to the approprate configure variable in the freeswitch OpenWRT Makefile Dec 27 02:06:22 e.g. LDFLAGS or TARGET_LDFLAGS, or whatevever Dec 27 02:25:40 cshore_: I alredy did and create a patch for the Makefile.am file now it compiles and produces the .libs/tone2wav file. Apparently, some other related to -lpthread crops up and needed the same patch on the same Makefile.am file. :) Dec 27 02:29:11 cshore_: I will post a file on FS mailing list later on so that developers there will be aware of the issue to fix this Makefile.am file. **** ENDING LOGGING AT Mon Dec 27 02:59:57 2010