**** BEGIN LOGGING AT Sat Feb 10 02:59:57 2007 Feb 10 04:29:30 03rwhitby * r685 10kernel/trunk/patches/2.6.20/98-sata_via-pata-support-upstream.patch: Refreshed against 2.6.20 Feb 10 04:30:24 03rwhitby * r686 10kernel/trunk/patches/2.6.20/ (00-fix-redboot-bad-code.patch series): Fix the redboot.c breakage in 2.6.20 - upstream got it wrong. Feb 10 05:31:45 SlugOS/LE should now boot if built with the latest 2.6.20-r2.686 kernel Feb 10 05:33:08 Now that SlugOS works again, I can get back to arm-kernel-shim and MokoSlug. Feb 10 05:59:50 Hmm - the push is taking some time ... Feb 10 06:01:25 there, now it's done. Feb 10 10:32:41 03marceln * r5453 10optware/trunk/ (Makefile make/memtester.mk): optware: Added memtester Feb 10 16:00:40 03bzhou * r5454 10optware/trunk/make/asterisk14.mk: asterisk14: avoid picking up amd k6 specific optimization, upstream should properly fix Feb 10 16:13:53 03bzhou * r5455 10optware/trunk/make/ossp-js.mk: ossp-js: 1.6.20060820 -> 1.6.20070208 Feb 10 16:14:27 03bzhou * r5456 10optware/trunk/make/fish.mk: fish: 1.22.2 -> 1.22.3 Feb 10 16:19:48 eno: ping Feb 10 16:28:41 03bzhou * r5457 10optware/trunk/make/postgresql.mk: postgresql: 8.2.2 -> 8.2.3 Feb 10 16:28:47 osas: pong Feb 10 16:30:07 Hi eno Feb 10 16:30:18 what was the issue with asterisk14? Feb 10 16:30:25 which arch? Feb 10 16:31:59 only when host=amd (k6) and target=powerpc for ds101g Feb 10 16:32:33 see http://logs.nslu2-linux.org/buildlogs/autobuild-nudi-last.txt Feb 10 16:33:33 and our official build machine happens to be amd (k6) Feb 10 16:33:51 I see Feb 10 16:35:09 codecs/gsm/Makefile line 212 needs fix Feb 10 16:46:36 eno: basically we need to add powerpc along with arm, armv5b and armeb to the list of $(PROC). right? Feb 10 16:51:47 osas: think so Feb 10 16:52:10 I'm confused ... from the logs I see the host being x86 Feb 10 16:52:25 bfin mipsel )) Feb 10 16:52:25 GSM_SOURCES+= $(SRC)/k6opt Feb 10 16:52:30 Ooops ... Feb 10 16:52:35 checking build system type... i386-pc-linux-gnu Feb 10 16:52:41 checking host system type... powerpc-603e-linux-gnu Feb 10 16:53:41 i actually haven't looked too closely, but in any case, on my build machine (p4) ds101g builds ok Feb 10 16:54:02 and on nudi (amd) ds101g it got problem Feb 10 16:54:22 can you run uname -m on your machine? Feb 10 16:54:24 and we don't want GSM_SOURCES k6opt.s for any of our target Feb 10 16:54:50 x86_64 Feb 10 16:55:05 I agree with that .... I just what to understand the issue if I'm going to hunt asterisk with this :-) Feb 10 16:55:19 yup Feb 10 16:55:25 that explains Feb 10 16:55:45 on my PIII build machine I have i686 Feb 10 16:56:03 On your machine is skiping this on the previous line Feb 10 16:56:23 ok Feb 10 16:56:32 on the build macine we have i386 Feb 10 16:56:41 and it is not caught by the first test Feb 10 16:56:43 and on nudi, $ uname -m Feb 10 16:56:44 i686 Feb 10 16:56:54 so it goes to the second one Feb 10 16:57:00 I got it now Feb 10 16:57:01 thx Feb 10 16:57:51 thank you Feb 10 17:10:41 Can I move /var/log to a flashdisk w/ a debianslug? Feb 10 18:51:23 03bzhou * r5458 10optware/trunk/make/asterisk14.mk: asterisk14: need to change two places to disable k6opt Feb 10 20:28:17 the mk file for faad2 doesn't work on unslung-dev Feb 10 20:28:40 scant: what's the error? Feb 10 20:28:53 another aclocal error like with esound Feb 10 20:29:06 esound.mk also doesn't work for unslung-dev Feb 10 20:29:12 but i patched that one Feb 10 20:34:57 rwhitby: u there Feb 10 20:35:00 ? Feb 10 20:35:41 yep Feb 10 20:36:35 Can't exec "aclocal": No such file or directory at /opt/bin/autoreconf line 182. Feb 10 20:36:35 Use of uninitialized value in pattern match (m//) at /opt/bin/autoreconf line 182. Feb 10 20:36:35 Can't exec "automake": No such file or directory at /opt/bin/autoreconf line 183. Feb 10 20:36:35 Use of uninitialized value in pattern match (m//) at /opt/bin/autoreconf line 183. Feb 10 20:36:35 autoreconf: Entering directory `.' Feb 10 20:36:35 autoreconf: configure.in: not using Gettext Feb 10 20:36:37 autoreconf: running: aclocal --output=aclocal.m4t Feb 10 20:36:39 Can't exec "aclocal": No such file or directory at /opt/share/autoconf/Autom4te/FileUtils.pm line 290. Feb 10 20:36:41 autoreconf: failed to run aclocal: No such file or directory Feb 10 20:36:43 ./configure: No such file or directory Feb 10 20:36:45 that's the error Feb 10 20:36:57 i think i need to put in this line Feb 10 20:36:58 ACLOCAL="aclocal-1.10 -I $(STAGING_DIR)/opt/share/aclocal" AUTOMAKE=automake-1.10 \ Feb 10 20:37:37 sounds reasonable, but I'll defer to eno for comment. Feb 10 20:38:01 but when i do, i get this error, make/faad2.mk:118: *** missing separator. Stop. Feb 10 20:38:15 find $(FAAD2_BUILD_DIR) -name \*.am -or -name \*.in | xargs sed -i -e 's/ Feb 10 20:38:15 $$//' Feb 10 20:38:15 (cd $(FAAD2_BUILD_DIR); \ Feb 10 20:38:15 ACLOCAL="aclocal-1.10 -I $(STAGING_DIR)/opt/share/aclocal" AUTOMAKE=automake-1.10; \ Feb 10 20:38:15 autoreconf -vif; \ Feb 10 20:38:16 $(TARGET_CONFIGURE_OPTS) \ Feb 10 20:38:23 what's wrong with this snippet of the mk file? Feb 10 20:39:37 missing separator usually means spaces instead of a tab Feb 10 20:40:44 i have to use tabs? Feb 10 20:41:27 makefile command lines must start with a tab. Feb 10 20:42:27 i have all tabsl Feb 10 20:45:28 rwhitby: it starts with a tab Feb 10 20:48:13 rwhitby: it's working now, getting warnings like with esound, let's hope it compiles Feb 10 20:48:20 thanks for your help rwhitby Feb 10 20:48:47 warnings, with faad2 and esound with aclocal 1.10 deal with underquoted strings Feb 10 20:54:44 is apex as second stage bootloader not build in mastermakefile debianslug-image any more? Feb 10 21:04:23 Sinclair73de: that's correct. Feb 10 21:04:50 I sent email about that to the lists last week or so Feb 10 21:05:12 how it works now? Feb 10 21:05:31 it's going to use arm-kernel-shim soon (just debugging LE) Feb 10 21:05:40 currently it uses in-kernel cmdline hacks. Feb 10 21:06:09 and it will be slugosle-image real soon too Feb 10 21:06:31 (debianslug is now deprecated as a name, since there is real Debian/NSLU2 now) Feb 10 21:07:13 hmmm... i'm interested because of setting then cmdline parameters for my 128MB Fatslug Feb 10 21:07:56 yep, you'll be able to set them in arm-kernel-shim which gets prepended to the kernel Feb 10 21:09:00 until apex has network and host usb support, it wasn't worth adding it for the standard build Feb 10 21:09:24 (see the mailing list message for more details) Feb 10 21:09:49 hm... Feb 10 21:09:49 ok, thanks for these information Feb 10 21:10:02 rwhitby, getting another error with faad2 Feb 10 21:10:17 configure: no xmms build configured Feb 10 21:10:17 configure: error: conditional "HAVE_BMP" was never defined. Feb 10 21:11:20 scant: doesn't look like something I can help you with ... Feb 10 21:12:49 =( Feb 10 21:12:56 is faad2 an integer decoder? Feb 10 21:14:03 faad2 is available in integer mode too Feb 10 21:14:34 NAiL: do u use faad2? Feb 10 21:14:46 nope Feb 10 21:14:51 I just googled a bit Feb 10 21:15:09 i wonder if the one in the optware sources is configured for integer or float Feb 10 21:19:42 does anyone here use FAAD2? Feb 10 21:25:52 Hi. From where do I get the optware source for apache2 (used in Brian's build for SlubOS/BE) ? Feb 10 21:31:48 chacko: www.apache.org? Feb 10 21:32:01 read the optware makefile Feb 10 21:32:32 Nail: Brian must be using a specific stream src? Feb 10 21:33:26 I am seeing a problem with apache2 in SlugOS/BE, but not on unSlung. Feb 10 21:34:43 Variables that provide the remote ip# or REMOTE_HOST are coming out as 0.0.0.0 Feb 10 21:34:51 APACHE_SITE=http://www.apache.org/dist/httpd Feb 10 21:34:51 APACHE_VERSION=2.2.4 Feb 10 21:34:51 APACHE_SOURCE=httpd-$(APACHE_VERSION).tar.bz2 Feb 10 21:35:08 There's a few patches Feb 10 21:35:33 rwhitby, i'm supposed to use this line to fix the error, but i don't know where to put it in the mk file Feb 10 21:35:45 sed -i '/E_B/d' configure.in Feb 10 21:35:55 can you suggest a place to put it Feb 10 21:36:17 I know very little about builds. Do you mean there are some patches that have not got into Brian's build that we can ipkg get? Feb 10 21:37:54 scant: I expect it would go in the .configured target, just after the 'find' line. Feb 10 21:38:01 let me see Feb 10 21:38:04 (before the call to autoreconf) Feb 10 21:38:46 sed -i '/E_B/d' $(FAAD2_BUILD_DIR)/configure.in Feb 10 21:38:57 something like that Feb 10 21:39:02 i have this Feb 10 21:39:03 find $(FAAD2_BUILD_DIR) -name \*.am -or -name \*.in | xargs sed -i -e 's/ Feb 10 21:39:03 $$//' Feb 10 21:39:03 (cd $(FAAD2_BUILD_DIR); \ Feb 10 21:39:03 sed -i '/E_B/d' configure.in; \ Feb 10 21:39:10 is that wrong? Feb 10 21:39:40 that should work too Feb 10 21:39:50 i'll tell you if it does Feb 10 21:40:16 i still can't believe that faad2 and esound on optware for unslung don't compile correctly Feb 10 21:40:23 aren't these things tested? Feb 10 21:41:54 rwhitby: a different error now Feb 10 21:41:57 config.status: error: cannot find input file: plugins/Makefile.in Feb 10 21:42:01 scant: if they are in the feeds, then by definition they build on the autobuild machine Feb 10 21:42:08 Nail: I just noticed that I am using the last build of a few weeks ago...apache - 2.0.59-3 Feb 10 21:42:39 Nail: and there is new version in the optware pkg feed! Feb 10 21:46:01 rwhitby: can you help me with this new error? Feb 10 21:50:13 can anyone here help me build faad2 on an unslug dev? Feb 10 22:18:43 03rwhitby * r687 10kernel/trunk/Makefile: Update arm-kernel-shim to 1.3, should fix LE problems Feb 10 22:24:58 scant: the official build machine are cross compilation Feb 10 22:25:22 it's not absolutely required to have a package also working on native build Feb 10 22:26:20 if a package is already cross, making it also build natively is just nice-to-have Feb 10 22:47:43 pushing major slugos build changes now Feb 10 23:09:35 eno: I just noticed that you have the 2.2 version of apache2 and php-apache in the stream now !!! Feb 10 23:13:49 chacko: r5078 | oleo | 2007-01-08 01:33:20 -0800 (Mon, 08 Jan 2007) | 1 line Feb 10 23:13:52 apache: 2.0.59->2.2.3 Feb 10 23:14:23 not sure if that solve your problem or not Feb 10 23:15:07 eno: Since mine is a production machine, I need to get a reading of the changes before trying it, it might solve my problem! Feb 10 23:15:28 eno: Do you know anybody has been using it with php? Feb 10 23:15:51 yeah you should try it on a staging slug server Feb 10 23:16:26 I am actually on transition to this one, so I do not have a spare one at the moment :( Feb 10 23:17:07 you can also try a different web server (like cherokee, lighttpd, or nginx) with PHP Feb 10 23:17:25 just so you can know whether the problem is on apache side, or php side Feb 10 23:18:12 eno: I think I know today it is on the apache side, because the apache logs do not show the remote_ip either Feb 10 23:18:27 php is not involved in the apache log generation Feb 10 23:18:36 k Feb 10 23:18:57 I have another short q. you have a min more? Feb 10 23:19:06 apache always feel a bit too heavy for the slug Feb 10 23:19:07 hi eno Feb 10 23:19:27 go ahead Feb 10 23:19:34 eno do you use bind (named) ? Feb 10 23:19:57 It is the dns server Feb 10 23:19:59 used it for a little while when i had my debian slug up Feb 10 23:20:27 eno: can you help me when you're done with chacko? Feb 10 23:20:38 I am curious how it gets started on bootup... I see the S09named file in the boot script directory Feb 10 23:21:26 on slugosbe? Feb 10 23:21:27 eno, S09named requires a parameter start ... Do not see how that parameter gets passed on to the screen Feb 10 23:21:52 Same on both slugbosbe and unslung Feb 10 23:22:14 It is not really a problem for me, but a curiosity... It does start ok Feb 10 23:23:06 if you execute S09named from the command line, it will ask for a parameter, yet it works ok at boot ... Feb 10 23:23:31 eno, about I menat 'script' not 'screen' Feb 10 23:23:59 i just read optware/sources/bind/S09named Feb 10 23:24:58 eno: i'm getting this error with faad2 on unslug dev, can you help me with it? config.status: error: cannot find input file: plugins/Makefile.in Feb 10 23:25:35 i think on unslung you should read /etc/rc.d/rc.unslung-start Feb 10 23:26:06 scant: i assume it's native compilation error, not runtime error Feb 10 23:26:56 eno: yes, it's a native compilation error Feb 10 23:28:20 eno, bye. talk to you later Feb 10 23:28:28 eno: any ideas? Feb 10 23:29:17 scant: one second Feb 10 23:33:31 i'm launching a distcc native build of faad2 right now Feb 10 23:37:20 eno: cool Feb 10 23:37:38 eno: can you help me while you are doing that? Feb 10 23:38:19 first i got a problem of not finding aclocal, i try to remember why i have the autoreconf there Feb 10 23:39:18 ah, actually the faad2 package does not come with configure Feb 10 23:39:19 eno: i fixed that issue Feb 10 23:39:35 by using aclocal-1.10 Feb 10 23:39:39 yes Feb 10 23:41:22 we have a problem that the autotool versions of cross and native build do not match exactly Feb 10 23:42:34 i've noticed Feb 10 23:43:42 do you have this error message? Feb 10 23:43:45 Can't exec "automake": No such file or directory Feb 10 23:44:26 ACLOCAL="aclocal-1.10 -I $(STAGING_DIR)/opt/share/aclocal" AUTOMAKE=automake-1.10 \ Feb 10 23:46:05 i got this error: "configure: error: conditional "HAVE_BMP" was never defined." Feb 10 23:46:15 it's possible faad2 is not compatible with the new autotool Feb 10 23:46:29 and used this line to get rid of the above error: sed -i '/E_B/d' configure.in; \ Feb 10 23:47:05 then i get this error: config.status: error: cannot find input file: plugins/Makefile.in Feb 10 23:47:15 and i don't know how to fix this one Feb 10 23:48:28 if faad2 is ONLY compatible with the old autotool, then this path requires lot of work Feb 10 23:49:17 it may even be easier if you replace autotools with old versions Feb 10 23:49:22 someone said to just create the Makefile.in file and it will go Feb 10 23:49:34 i'm trying that now Feb 10 23:49:48 it didn't work Feb 10 23:50:12 how do i replace? Feb 10 23:50:33 and may i ask why do you need to build faad2 in the first place? any problem with the binary in the feed? Feb 10 23:51:40 it's like "ipkg remove automake"; manually download automake-1.9...ipk from feed; "ipkg install automake_1.9...ipk" Feb 10 23:53:53 with all the effort, you might as well setup a cross build environment and tackle any runtime problem Feb 10 23:54:12 i was trying to compile myself Feb 10 23:54:42 i don't have a distro to use, can you reccomend one that works well in a vm? Feb 10 23:55:12 i'm running ubuntu 6.10 under vmware server Feb 10 23:55:25 works pretty well Feb 10 23:55:53 i just installed the server version, which is quite small Feb 10 23:56:21 cool Feb 11 01:42:03 i don't think the mpd ipk on the feed works for unslung and alsa Feb 11 02:08:01 scant: http://tech.groups.yahoo.com/group/nslu2-linux/message/17478 Feb 11 02:09:03 SlugOS/BE 4.3-beta nslu2 ttyS0 Feb 11 02:09:22 First boot of SlugOS/BE (formerly known as OpenSlug) Feb 11 02:28:01 one poster said it worked Feb 11 02:28:11 it can't find my audio device, but my own builds do **** ENDING LOGGING AT Sun Feb 11 02:59:57 2007