**** BEGIN LOGGING AT Wed May 02 02:59:56 2007 May 02 03:30:51 how do I make sure I cleanly removed a package from the build May 02 03:31:13 bitbake -c clean only remove the stuff in tmp/work May 02 03:32:03 does it? May 02 04:11:29 morning :) May 02 05:59:59 03lenehan 07org.oe.dev * rf91b0576... 10/ (1 packages/net-snmp/net-snmp_5.1.2.bb): May 02 05:59:59 net-snmp 5.1.2: Fix for do_configure hanging when using bitbake May 02 05:59:59 1.8.2. We need to pass "--with-defaults" to configure tell it we want May 02 05:59:59 the defaults so it doesn't warn that the defaults are going to be May 02 05:59:59 used and ask them to confirm each of the defaults. May 02 06:20:18 Laibsch : Bug 2200 looks like its not caused by squashfs-lzma-tools-native it self but from the required libs May 02 06:21:44 Laibsch : but in any case thanks should go to Likewise :) He is the one "responsible" for the the package :) May 02 06:58:33 good morning all May 02 06:59:02 zecke: I just downloaded the ptxdist crosstool, and they are shipping OE patches, but they renamed zecke-sane-readelf to sane-readelf May 02 07:01:26 <_law_> there is a bluez (hciattach-ti-bts.patch) patch available in OE is the hciattach bts-script also available somewhere in OE ? May 02 07:09:56 koen: good May 02 07:10:24 stefan_schmidt: http://cgi.ebay.de/SUPER-KNULLER-FERNSEHEN-HANDY-MOTOROLA-DVB-H_W0QQitemZ320109684176QQihZ011QQcategoryZ9359QQssPageNameZWDVWQQrdZ1QQcmdZViewItem May 02 07:10:31 stefan_schmidt: you might want buy another EZX device May 02 07:11:28 Einfach hammer! May 02 07:12:05 koen: Einfach der Hammer? May 02 07:12:31 just quoting from the article May 02 07:13:14 zecke: http://www.rpsys.net/openzaurus/temp/oe-install-change.patch May 02 07:13:55 I won't say it looks sane :} May 02 07:18:32 zecke: it should fix the weirdness people after they restart a failed build May 02 07:21:21 koen: I assumed that May 02 07:21:48 zecke: http://butterfeet.org/?p=22 May 02 07:22:06 Do you know of any other events beside: ELCE,OLS,LinuxTag,T-Dose? May 02 07:22:25 fosdem, bossa May 02 07:24:22 Morning May 02 07:25:20 koen: could you please pastebin include/config/auto.conf and include/linux/autoconf.h from your last poodle kernel build? I am still trying to trace down the CONFIG_UNAME_RELEASE nonsense I am seeing May 02 07:25:39 koen: fosdem2008, bossa2008? May 02 07:25:46 yes May 02 07:26:00 hvontres|home: I don't have a poodle build May 02 07:26:45 koen: who did the test image in unstable then? May 02 07:26:58 zecke: Thanks. Added to my wishlist. :) May 02 07:26:58 ever heard of 'rm'? May 02 07:27:31 koen: ok. nevermind... May 02 07:28:04 ~lart kernel build system for seemingly pulling obscure config variables out of its ass May 02 07:28:04 * ibot puts kernel build system into a headlock and administers a mighty noogie, rubbing half of kernel build system's hair of for seemingly pulling obscure config variables out of its ass May 02 07:30:55 good morning everyone May 02 07:31:33 For the record, latest oe mipsel compile try to compile wrong platform xptcall source May 02 07:31:40 user@ubuntu:/stuff/tmp/work/mipsel-angstrom-linux/minimo-0.016+cvs20060720-r4/mozilla/xpcom/reflect/xptcall/src/md/unix$ ls *.o May 02 07:31:41 xptcinvoke_unsupported.o xptcstubs_unsupported.o May 02 07:32:00 time to go to bed May 02 07:32:34 It should have to compile below source, but it didn't May 02 07:32:37 user@ubuntu:/stuff/tmp/work/mipsel-angstrom-linux/minimo-0.016+cvs20060720-r4/mozilla/xpcom/reflect/xptcall/src/md/unix$ ls *mips* May 02 07:32:37 xptcinvoke_asm_mips.s xptcinvoke_mips.cpp xptcstubs_asm_mips.s.m4 xptcstubs_mips.cpp May 02 07:32:56 s/compile// May 02 07:33:43 morning all May 02 07:33:47 hi May 02 07:33:54 afternooon May 02 07:34:00 Morning RP May 02 07:35:07 zecke: Any thoughts on http://www.rpsys.net/openzaurus/temp/oe-install-change.patch ? May 02 07:35:42 RP: looks sane, the wished result can happen. I'm too scared to say things about the side-effects :) May 02 07:35:50 polyonymous: morning :) May 02 07:35:56 [09:13] zecke: I won't say it looks sane May 02 07:36:15 :) May 02 07:36:22 zecke: There are problems if do_package fails as do_install has to run again... May 02 07:36:50 Ah, I should read scrollback more carefully :) May 02 07:36:50 right, your patch is likely to be the only solution to that problem May 02 08:10:00 If COMBINED_FEATURES does not have bluetooth in there, why does task-base want to be able to build bluez then? May 02 08:10:33 for task-bluetooth I guess May 02 08:10:51 yes May 02 08:11:02 03koen 07org.oe.dev * r4775c587... 10/ (4 files in 3 dirs): glib-2.0: add 2.12.12, a bugfix release May 02 08:11:58 likewise: btw, cairo 1.4.6 doesn't have the workqueue patch May 02 08:13:04 Sort of chicken egg problem: bluez depends on glib, which needs i18n (USE_NLS). If I keep USE_NLS = "no", bluez cannot build, so I cannot build my image, even if I do not need bluez/bluetooth at all. May 02 08:13:28 koen: workqueue patch is now solved differently, or not at all? May 02 08:14:30 koen: I froze on an older OE revision for our products, too much problems with current OE unfortunately. May 02 08:14:53 likewise: bluez can use it's own internal glib May 02 08:15:10 likewise: the workqueue is the re-implementation of the surface-cache May 02 08:15:12 koen: yes, but that's not the real issue. May 02 08:16:29 koen: For any configuration using task-* stuff, it seems all the packages must be buildable for OE in order to build anything at all. May 02 08:16:45 for debian.bbclass May 02 08:16:52 even if most of those packages are not wanted. May 02 08:17:26 that's why I split up the sdk tasks May 02 08:18:32 But the task-* is supposed to be distro/machine/package handling independent, right? May 02 08:26:24 I've just added bugs 2204, 2205, 2206 and 2207 for new gstreamer stable releases and a bug in gst-plugins-good May 02 08:36:21 morning May 02 08:36:26 ~change 60 gbp to pln May 02 08:36:30 hrw: good morning! May 02 08:37:02 i got some error trying that May 02 08:37:14 hi hrw .) May 02 08:37:34 ~change 60 gbp to pln May 02 08:37:41 i got some error trying that May 02 08:37:45 ibot: ok May 02 08:37:50 xe.com is probably a bit sick May 02 08:38:01 fine May 02 08:44:05 all RPATH needs to be reported in bugtracker? May 02 08:44:33 madplay has this problem May 02 08:44:35 hrw: yes May 02 08:44:48 and make it depend on the rpath metabug May 02 08:44:54 ok May 02 08:45:07 RP: feel free to push, it can't get more unstable :} May 02 08:45:53 morning all May 02 08:45:55 zecke: Someone needs to test it first ;-) May 02 08:46:08 !oebug 2208 May 02 08:46:09 * * Bug 2208, Status: NEW, Created: 2007-05-02 01:45 May 02 08:46:10 * * openembedded(AT)hrw.one.pl: RPATH in madplay May 02 08:46:11 * * http://bugs.openembedded.org/show_bug.cgi?id=2208 May 02 08:46:13 hm gersternhi rp May 02 08:46:15 hi ade|desk May 02 08:46:38 * zecke 's FreeSoftware contribution is currently mostly limited to chewing mails May 02 08:47:49 ups May 02 08:47:51 hi rp May 02 08:47:52 03hrw 07org.oe.dev * r4c02388e... 10/ (1 packages/settings-daemon/settings-daemon_svn.bb): settings-daemon: merge changes from Poky (proper PV, dependencies) May 02 08:56:25 hi greentux May 02 08:57:04 hi hrw May 02 08:57:51 morning all May 02 09:00:25 hi Dirk May 02 09:01:54 hi Marcin May 02 09:02:57 hi do13 May 02 09:03:25 hi woglinde May 02 09:11:20 ~ping koen May 02 09:11:35 pong koen May 02 09:17:09 is there a task to create the ipkg index,copy all the package to an image, and configure ipkg.conf? May 02 09:17:45 morning mickeyl May 02 09:18:02 Morning btw May 02 09:18:29 hi May 02 09:19:08 hi mickeyl May 02 09:19:52 mickeyl: How's openmoko progressing? May 02 09:22:20 do13: not bad. hardware issues should be sorted out by now, IIRC the first production run is in the factory May 02 09:22:43 sounds good May 02 09:23:18 yeah, we had tons of mispredictions from HW sourcers May 02 09:23:39 bought tons useless stuff while we are missing useful components May 02 09:23:39 heh May 02 09:32:50 03dirk 07org.oe.dev * r23dd62a5... 10/ (1 packages/valgrind/valgrind_3.2.3.bb): valgrind: 3.2.3 added May 02 09:32:56 03dirk 07org.oe.dev * rf82dd6e7... 10/ (3 files in 2 dirs): syslinux: 3.36 added, also install extlinux May 02 09:33:03 03dirk 07org.oe.dev * r54c79c33... 10/ (1 packages/ethtool/ethtool_5.bb): ethtool: version 5 added May 02 09:40:13 morning May 02 09:40:20 hi xora May 02 09:43:04 Hi XorA May 02 09:44:34 I wonder if oprofile 0.9.2 in oe has all the recent arm improvements and the remote mode, I've seen people at o-hand has added. May 02 09:45:01 sirfred: RP or hrw probably the ones to ask that May 02 09:45:30 XorA: Waiting for their irc clients to beep. May 02 09:47:19 good morning May 02 09:49:45 03lenehan 07org.oe.documentation * r76b9cce3... 10/ (1 usermanual/chapters/usage.xml usermanual/usermanual.xml): May 02 09:49:45 usermanual: Clean up the using bitbake/oe chapter to the point where it's May 02 09:49:45 almost useable. Still not included in user manual yet, probably needs a bit May 02 09:49:45 more work before that hppens. May 02 09:50:51 sirfred: do not know May 02 09:51:25 hi florian May 02 09:53:10 hrw: Is there a proper recipe in the o-hand repository I could use? May 02 09:54:22 hrw: Or perhaps the oprofile_cvs.bb recipe? May 02 09:55:12 sirfred: oprofile 0.9.2 is same May 02 09:55:42 but oprofile cvs Poky is extended May 02 09:57:43 btw.. May 02 09:58:05 speaking about r/w access for polyonymous I would also like to give r/w for sirfred May 02 09:58:32 hrw: I'm disturbing with too many patches. ;-) ? May 02 09:58:55 ;D May 02 09:58:59 hrw: +1 for me, sirfred does good work May 02 09:59:15 Thanks, guys. May 02 09:59:21 mickeyl, koen, RP, rwhitby: what about you? May 02 09:59:33 +1 May 02 10:03:03 hrw: I found sirfred's work very good, too. I thought he already had RW access. May 02 10:03:14 Laibsch: he does not yet May 02 10:03:43 The more quality people who have RW, the less bug work for me ;-) May 02 10:03:49 ;D May 02 10:03:55 It's a surprise for me people being aware of my work. May 02 10:04:06 sirfred: you are the pr0n player provider :-D May 02 10:04:10 I thought it was something more anonymous. May 02 10:04:10 sirfred: send me a monotone key with sirfred@openembedded.org as key address and cc' koen. May 02 10:04:28 03Manuel 07org.oe.dev * r1bb95c93... 10/ (3 files in 3 dirs): gpe-conf: make it build with wireless includes changes - close #2095 May 02 10:04:47 sirfred: and welcome on board. Now you can work on fast fixing some bugs ;D May 02 10:05:00 hrw: Thanks. :-) May 02 10:05:33 sirfred congratulations! May 02 10:06:29 mickeyl: I'm afraid I'm not very confident with monotone (other than R/O features) May 02 10:06:43 mickeyl: How can I generate a key? May 02 10:07:01 sirfred man mtn May 02 10:07:03 sirfred: that is in monotone manual May 02 10:07:03 sirfred: you should read enough of the manual to answer that question before tinkering with write access :D May 02 10:07:12 *g* May 02 10:07:13 03lenehan 07org.oe.documentation * r207bcfd2... 10/ (7 files in 3 dirs): May 02 10:07:13 usermanual: The image_ipkg class is now just the image class. Update May 02 10:07:13 all references. Modify description to include details about how it May 02 10:07:13 can potentially support other packaging types now. Probably needs a May 02 10:07:13 more detailed check to see if anything else changed. May 02 10:07:43 * sirfred sees how the doors of the rw heaven closes in front of him May 02 10:07:58 mickeyl: Don't worry, I won't dare to do anything without being sure. May 02 10:08:18 mickeyl: It was just to send you the key as requested, as fast as possible. May 02 10:08:23 sirfred: the good thing is if you balls it up we can always mtn dissaprove May 02 10:09:20 sirfred: mtn genkey sirfred@openembedded.org May 02 10:09:30 mickeyl: Yes, I've found it. :) May 02 10:10:19 mickeyl: The email address where you want it sent ? May 02 10:10:24 sirfred: congrats :) May 02 10:10:31 mickeyl@linuxtogo.org, cc: koen@linuxtogo.org May 02 10:10:32 * sirfred expects that not to be on the man page. May 02 10:10:44 or rather May 02 10:10:53 substitute linuxtogo.org w/ openembedded.org May 02 10:10:57 zecke, keesj : Thanks. May 02 10:10:59 sirfred: few commands to read about: add/drop/rename and commit/merge/push May 02 10:11:20 sirfred: and change 'pull update' to 'pull merge update' May 02 10:11:30 hrw: I will read the whole manual before trying to do anything more than I used to do. May 02 10:11:50 sirfred: sure - I am giving most used ones May 02 10:11:58 hrw: Thanks May 02 10:12:32 sirfred gets my vote too FWIW May 02 10:12:53 sirfred: poky does have decent oprofile packages, you need the cvs version and the correct cvs date May 02 10:13:07 * RP needs to sort out getting them upstream... May 02 10:13:55 RP: OK, I will try to generate it from oe. May 02 10:16:43 Any other thing I should read, related with commit policy or so? May 02 10:16:56 Other than the monotone manual May 02 10:16:59 first line of commit is summary May 02 10:16:59 sirfred: we have some policy pages on the wiki, please be sure to read them all May 02 10:17:13 mickey|bbl: Will do. May 02 10:17:30 http://www.openembedded.org/policies May 02 10:18:25 hrw: thx May 02 10:19:00 sirfred: and your first task is to clear out your own bugs :-D May 02 10:19:08 Aaaag May 02 10:19:22 XorA: :) May 02 10:24:36 http://www.openembedded.org/commitpolicy - feel free to extend May 02 10:30:04 http://worsethanfailure.com/Articles/The-Loony-BIN.aspx May 02 10:30:50 +1 for both May 02 10:30:52 hrw: The link from http://www.openembedded.org/policies does not work. Or does this just need a little time? May 02 10:32:13 mickey|bbl: private key alert May 02 10:33:10 did I told that I do not like drupal? May 02 10:33:35 Laibsch: fixed May 02 10:34:51 mickey|bbl: Humm, I'm thinking that perhaps the only thing I had to send is the public key of the generated pair, true? May 02 10:35:53 sirfred: remove that key, generate new one, send public part May 02 10:36:12 hrw: ok. May 02 10:52:24 huhu May 02 10:57:01 good morning boys and girls May 02 10:57:36 Laibsch: stop telling people to do ASSUME_PROVIDED += "qemu-native" May 02 10:57:48 hi polyonymous May 02 11:00:11 ASSUME_PROVIDED += "qemu-native" did also not work for me because my version did not containt options like "-r" May 02 11:00:25 it's faq right? May 02 11:00:52 it's only faq because people keep saying to do ASSUME_PROVIDED += "qemu-native" May 02 11:01:06 which is wrong, since you need a patched qemu May 02 11:01:41 but the faq also containt ASSUME_PROVIDED qemu-native http://www.openembedded.org/wiki/OeFaq May 02 11:01:47 ah god I wish I never mentioned I had patched up fedora qemu to work May 02 11:01:50 aaaaaargh! May 02 11:03:12 * koen corrects the faq May 02 11:03:26 the right thing is to do the ENABLE_BINARY_LOCALE_GENERATION right? May 02 11:03:50 no, the correct thing is to build qemu-native May 02 11:04:01 the only option were that doesn't work is in x86_64 May 02 11:04:53 since doing ENABLE_BINARY_LOCALE_GENERATION=0 will most likely give you a broken image May 02 11:04:56 does not work on my machine either (gentoo x86 32) May 02 11:05:11 keesj: emerge gcc 3.4? May 02 11:05:20 Laibsch, yes, those two bugs are dupes (qte) but they contain different solutions and both should work:) May 02 11:06:54 koen broken in which way? May 02 11:07:09 woglinde: guess May 02 11:07:17 LOCALES? May 02 11:07:22 yes May 02 11:07:29 if dont want locales May 02 11:07:32 *g* May 02 11:07:44 then you should disable locales May 02 11:07:48 e.g. using USE_NLS May 02 11:08:43 koen is there a chance that armv4 will be supported for eabi May 02 11:08:56 after reading the debian wiki site May 02 11:08:59 I guess no May 02 11:09:17 woglinde: it needs some patches to glibc May 02 11:09:34 but I don't think it's worth the effort May 02 11:09:55 since you will hit a slowpath on each function return May 02 11:12:11 RP: I'm starting a build from scratch using your install-patch May 02 11:26:35 koen: You sound like I was giving out this advice on an hourly basis despite being told this was wrong. Neither of which was right. May 02 11:26:56 koen: I had been told to do this previously because of another problem I had May 02 11:27:14 And at some point in time it seems this was the "accepted" approach May 02 11:27:54 Laibsch: for the past year with angstrom, it never has been the accepted approach May 02 11:38:36 03hrw 07org.oe.dev * r53c6a446... 10/ (3 files in 2 dirs): libglade: set '-lz' in LDFLAGS instead of CFLAGS - close #1774 May 02 11:38:40 03hrw 07org.oe.dev * r09dbe920... 10/ (1 classes/image.bbclass): image.bbclass: remove *all* IPKG list files May 02 11:40:34 koen: Well, I believe I made that change in 2006. So it depends on your definition of "past year". ;-) Though I am almost certain I made that change less than 12 months ago. Have reverted it now and will stop giving out that advice. May 02 11:41:06 have you been building angstrom for the past year? May 02 11:43:36 RP: The opreport-xml-output-fixes.patch does not apply against current oprofile cvs. Could you give me a safe SRCDATE to use, please? May 02 11:43:58 RP: 12 of 14 hunks failed, fyi May 02 11:44:09 koen: On and off, yes. But started flashing only recently due to lack of time. May 02 11:44:29 sirfred: moment May 02 11:44:50 SRCDATE_oprofile ?= "20070106" May 02 11:45:04 hrw: Thanks a lot. May 02 11:45:08 Laibsch: I doubt you built glibc with an assumed qemu-native, since no provided qemu has the nptl patch May 02 11:48:26 Great nptl, a pair of weeks ago, top showed a lot of gst-launch processes (threads, actually) eating all my CPU May 02 11:48:40 Now, only a single gst-launch is eating it all May 02 11:49:23 Koen: How was celf ? May 02 11:49:37 gerwinin: lots of nice cookies May 02 11:49:57 Koen: Do we got any pictures ? Received new donations ? May 02 11:50:15 I expect oprofile to find some room for improvement for gstreamer hackers. May 02 11:51:04 <---------- received a usb development kit for making some usb devices May 02 11:51:09 gerwinin: http://www.flickr.com/photos/koenkooi/tags/elc2007/ May 02 11:52:54 * koen reads http://www.math.utah.edu/~palais/pi.html May 02 11:53:19 Koen: I will update the conference schedule this week and I will issue some press releases soon as everything is confirmed May 02 11:53:38 Confirmed for sure is t-dose (nl) May 02 11:54:22 koen: Looking for glibc in my angstrom dir I find for example glibc-2.5-r5. May 02 11:54:35 --------> I will do it at the end of the week still got a shit load of hardware to make May 02 11:55:18 Laibsch: and you generated the locales with your host qemu? May 02 11:55:52 gerwinin: hi May 02 11:55:58 gerwinin: which hardware to make May 02 11:56:00 ? May 02 11:57:18 hrw: I am working on several project now : usb webcam (waiting for pcb to come in), usb heartbeat monitor (just did the design), usb stepcounter (finished samples out now) May 02 11:59:20 hrw: very small projects :) May 02 11:59:44 gerwinin : Usb 2.0 or 1.1 ? May 02 12:00:08 hrw: usb 2.0 mostly with a tiusb host chip , sometimes with a nifty fti chip May 02 12:00:41 USB step counter? Hope it doesn't need to be connected to work.. ;) May 02 12:00:50 koen lol "what worries me is that the first thing we broadcast into the cosmos..." May 02 12:01:03 CM: no it doesn't May 02 12:01:48 hrw: Although I must admit that I am getting some help with the layout , it is not so difficult to make usb devices as I thought May 02 12:06:46 the drivers/firmware is where most of the development is... May 02 12:07:13 especially if your device does not fall on a "standard" profile May 02 12:16:16 lfaistos true , but I am only going to develop linux drivers , the windows they need to do themselves , me and windows are incompatible for one of other reason May 02 12:44:30 so far no problems with the install-patch May 02 13:05:59 RP: any plans for a linux-rp 2.6.21? May 02 13:06:44 koen: feel free to prepare such one May 02 13:15:45 03koen 07org.oe.dev * rc3acd01a... 10/ (1 packages/gpe-dm/gpe-dm_0.51.bb): gpe-dm: add 0.51, doesn't need procps anymore May 02 13:27:29 hrw: Any hint about how the remote/client oprofile mode is used? I may be blind, but the options of oprofiled and/or opcontrol are not helping me. May 02 13:28:03 I was expecting a way to tell it the port, to disable the session directory,... May 02 13:32:28 never used oprofile May 02 13:32:35 likewise: cairo's workqueue patch should be landing very soon now, (just after cairo 1.4.6). May 02 13:33:10 Is it just me or is glib-2.0 failing? I complaines about some iconv.h in and libiconv stuff May 02 13:35:05 hrw: ok, I will ask RP, I think he made some of the changes. May 02 13:35:11 RP: Any hint, please? May 02 13:35:19 pastebin.co.uk/13879 May 02 13:36:03 cworth: thanks for the update push :-) May 02 13:36:50 goxobx hm http://www.gtk.org/api/2.6/glib/glib-building.html May 02 13:41:33 still no problems with that install patch May 02 13:42:27 koen: can you update snapshots of OE? May 02 13:44:46 koen is it in oe or onlz local at zour side? May 02 13:45:14 damnit english keyborad May 02 13:45:37 crappy german imitations of keyboard May 02 13:45:50 * hrw is fan of US layout May 02 13:45:50 no no May 02 13:46:04 hehe you have no umlauts May 02 13:46:21 * woglinde shakes head May 02 13:46:37 qtopia4 compiles takes now 1 gig space May 02 13:46:48 woglinde: I have ąćęłńóśźż instead May 02 13:47:02 hm actual my irssi cant display utf8 May 02 13:47:24 and can enter öäëïüä without problem May 02 13:47:36 ðđŋħ? May 02 13:47:49 let me see if euro works May 02 13:47:55 hm no May 02 13:48:28 I'd like to patch BitchX for UTF8 May 02 13:48:36 imho still the best irc client around :> May 02 13:48:48 in your opinion May 02 13:48:54 * hrw started with ircII May 02 13:49:06 then epic3, epic4/2000, irssi May 02 13:49:25 Jin^eLD: It's not as good as Irssi ;) May 02 13:49:38 I switched from BX to irssi. May 02 13:49:38 tried bitchx, xchat, amirc, blackirc, telnet, mirc, grapevine and few others May 02 13:49:42 even macirc May 02 13:50:03 There's no better irc-client than screen+irssi ;) May 02 13:50:26 hrw: update in what way? May 02 13:50:28 NAiL: I tried switching, I even installed the BX theme for irssi, but its not the same :) I guess I am too conservative when it comes to changing look/behaviour of stuff that I got used to May 02 13:51:02 Jin^eLD: It took a while getting used to irssi, but when I got used to it, there was no going back. May 02 13:51:09 koen: 0.29/0.30/0.31 are quite old ones May 02 13:51:18 and rightfully so May 02 13:51:42 I'm not going to waste time on updating those May 02 13:51:48 NAiL: well thats my problem now - got used to BX too much :) I am sure I can somehow configure irssi to behave same as BX but I don't have the patience :) May 02 13:51:49 you're welcome to do so May 02 13:51:49 I prefer UK keyboards, US ones have silly little LF May 02 13:52:01 LF? May 02 13:52:25 hrw: he means the tiny enter key :> May 02 13:52:41 hm right May 02 13:52:54 enter, return LF , get me a new line key ... all of those May 02 13:53:22 don't forget "execute" May 02 13:53:33 But that is only a matter of the keyboard model , isn't it? I've seen spanish keyboards with the little and horrible, one-row high enter key, and models with the two-rows high key. May 02 13:53:35 ade|desk: thats depend on keyboard itself not on layout May 02 13:53:35 or 'hit and hope' May 02 13:54:02 That's what I wanted to mean. :) May 02 13:54:43 * XorA has a # where the rest of enter should be :-D May 02 13:54:50 ok then i like my IBM UK layout keyboard, it works. May 02 13:55:21 I am used to one line ENTER key - my microsoft natural has it that way May 02 13:55:32 i wonder what happened to the rest of the IBM that perhaps went with it May 02 13:55:47 The best keyboard I had is one of those heavy IBM PS/2 ones May 02 13:55:58 RP: still no troubles with that patch (it's at libx11 now, from scratch) May 02 13:56:06 the IBM XT one was nice May 02 13:56:07 The only matter is that your neighbours will hear you typing in the night. May 02 13:56:08 * woglinde likes his original german DEC keyboard May 02 13:56:38 koen: oe/snapshots are oe:www-data only and I am not oe and not in www-data.. May 02 13:56:41 * ade|desk still fears the AS/400 terminals May 02 13:56:58 ade he F22 May 02 13:57:24 morning May 02 13:57:26 hrw: I offered multiple times to add your key, but you always refused May 02 13:57:34 hi chouimat May 02 13:58:08 I know May 02 13:58:31 lack of time etc May 02 14:02:30 sirfred: Look in conf/distro/poky.conf in poky for the srcdate that works May 02 14:02:50 koen: I will do a 2.6.21 linux-rp at some point when I get a moment :) May 02 14:02:55 RP: hrw kindly got me a proper SRCDATE, thanks. May 02 14:03:00 koen: and its good to know that install patch is working :) May 02 14:03:23 RP: Now, I have an oe generated oprofile, but I don't know how to use it in remote mode. May 02 14:03:44 RP: You will see my question if you keep scrolling down . :) May 02 14:03:51 sirfred: Have you build oprofileui? May 02 14:04:20 RP: I've downloaded a prepackaged one for my ubuntu. I assume I only need it in my desktop. May 02 14:04:38 sirfred: right. You also need a modified oprofile on your desktop May 02 14:05:04 RP: ok, I need to do that, so. And then, what? May 02 14:05:31 sirfred: Ah, you also need oprofileui's server on the client May 02 14:05:49 RP: ok. May 02 14:06:02 sirfred: The reciepe from poky should help with that May 02 14:06:18 RP: Yes, I saw it before. Thanks. May 02 14:06:45 Hello. Could somebody with a recent linux-rp-2.6.20 build pastebin their include/conf/auto.conf and include/linux/autoconf.h files please? May 02 14:07:06 hvontres|poodle, don't you have it in stagin? May 02 14:08:17 polyonymous: I want to see what somebody elses looks like. CONFIG_UNAME_RELEASE keeps getting set to uname -r from my build machine. I wanted to see what it looks like for a build that works May 02 14:08:35 http://rafb.net/p/QJDJC471.html May 02 14:08:41 polyonymous: thanks :) May 02 14:08:54 hvontres|poodle, the same. May 02 14:09:08 brb May 02 14:09:12 #define CONFIG_UNAME_RELEASE "2.6.18-4-686" May 02 14:09:40 it's either "safely ignore" or "by design" :) May 02 14:10:06 polyonymous: do your sound modules load? May 02 14:10:20 hvontres|poodle, haven't tried yet :) May 02 14:10:30 well, according to lsmod, yes. May 02 14:10:50 * hvontres|poodle scratches head..... May 02 14:11:16 well, I guess I need to start looking someplace else now... May 02 14:11:25 Thanks for your help guys :) May 02 14:17:34 16:17 hrw@home:build$ grep CONFIG_UNAME_RELEASE tmp/staging/akita-poky-linux-gnueabi/kernel/include/linux/autoconf.h May 02 14:17:37 #define CONFIG_UNAME_RELEASE "2.6.21" May 02 14:18:04 so it is wrong May 02 14:18:53 #define CONFIG_UNAME_RELEASE "2.6.20-15-generic" May 02 14:24:29 RP: Why is oprofile in the viewer side needed. I thought it was something like: [embedded device: oprofile - oprofileui-server] --- [oprofileui-viewer: Desktop] May 02 14:24:57 hmmm, so I wonder where this gets used: #define CONFIG_DEFCONFIG_LIST "/lib/modules/$UNAME_RELEASE/.config" May 02 14:25:33 sirfred: We use some of oprofile to do the analysis May 02 14:26:04 RP: Are there prepackaged versions somewhere? May 02 14:26:15 sirfred: debian.o-hand.com May 02 14:26:32 sirfred: The README should mention this May 02 14:26:35 RP: Oh, I didn't see it before. May 02 14:27:04 hvontres|poodle: The recompressed data for wikipedia in normal squashfs came out to be even 2G May 02 14:27:12 Uncompressed is 13G May 02 14:27:40 RP: Sorry, I didn't see the README before. May 02 14:28:33 Laibsch: with squashfs-lzma? May 02 14:28:59 * XorA could fit that on his Z with space to spare May 02 14:29:56 koen: Still going May 02 14:30:19 I used my faster laptop for this but it has been about 16 hours now. May 02 14:30:33 data is being pulled in over NFS, though May 02 14:30:36 RP: linux-ezx fails in do_deploy, dunno if that's caused by the patch May 02 14:30:41 * koen investigates May 02 14:30:58 XorA: You have more than 13G on your Z? May 02 14:31:07 Laibsch: I mean the 2G May 02 14:31:17 XorA: that is no feat ;-) May 02 14:31:46 I could plug my 160G usb drive into Z for the 13G May 02 14:31:55 XorA: I would buy a new card so I could. But I guess I will get back to around the 700M this was when I pulled it off static.wikipedia.org May 02 14:32:15 RP: do_install[dirs] = "${D} ${S} ${B}" May 02 14:32:28 RP: I suspect that's missing in package.bbclass May 02 14:32:39 XorA: I could use wireless and $whatwasthenameofthisworldwideFSthing May 02 14:32:42 ;-) May 02 14:32:58 XorA: Now, top THAT! May 02 14:33:49 Laibsch: GFS May 02 14:34:21 if SDHC cards worked in 2.6.20 then 13G would be no issue May 02 14:34:48 they should work in .21 May 02 14:36:03 XorA: What is the cost? May 02 14:36:14 Laibsch: no idea May 02 14:36:25 oh, that expensive May 02 14:36:28 Thought so May 02 14:36:29 ;-) May 02 14:36:38 Not yet for the average joe May 02 14:36:42 But time will come May 02 14:36:44 4G SD + 4G CF not unaffordable May 02 14:36:52 Absolutely May 02 14:37:01 that gives me 14G May 02 14:37:10 8GB Sd is €100 May 02 14:37:11 I have 1G SD and 256M CF and have no issue May 02 14:37:35 But I am probably going to up that if my wikipedia stuff gets going May 02 14:37:52 I am very eager to see the results May 02 14:38:14 Laibsch: don't forget to add squashfs-lzma support to the kernel May 02 14:38:29 so 200 EUR and I could have 22G May 02 14:38:38 thats more than my PC used to have May 02 14:39:35 koen: I'll get back on that once my compilation of the rest and the recompression progresses May 02 14:41:30 that reminds me that I have to resize ext3 partition with /home/ May 02 14:42:36 294GB -> 402GB May 02 14:43:37 * sirfred thinks he should update his hardware May 02 14:49:10 koen: The do_install[dirs] should be fine where it is. What was the failure? May 02 14:49:37 +addtask deploy before do_populate_staging after do_package May 02 14:49:38 -addtask deploy before do_package after do_install May 02 14:49:43 Hi! May 02 14:50:10 took me a while to pair bluetooth on angstrom's opie with my machine :) May 02 14:50:11 Laibsch: what's that wikipedia stuff you mentioned? is there's an app to use wikipedia dump on a pda? May 02 14:50:19 RP: so the error was to be expected May 02 14:50:29 I the his app of choice is web browser. May 02 14:50:33 I think May 02 14:50:36 psokolovsky__: yes, it's called 'webbrowser' :) May 02 14:50:43 that sucks ;-) May 02 14:50:59 koen: Hmm, I'm not sure that will work for the kernel module stuff linux-rp does :/ May 02 14:51:09 I mean carrying WP ona handheld for offline use ;-) May 02 14:51:14 why sucks? wp dump as html May 02 14:51:20 yes, he's using html dump May 02 14:51:21 RP: Great, I have oprofile client/server working. Thanks a lot! May 02 14:51:29 polyonymous: you couldn't pair OE.dev opie with anything at all May 02 14:51:37 RP: D'oh! May 02 14:51:43 psokolovsky__, Huh? I did. May 02 14:51:58 It's NOT easy :) May 02 14:52:03 but it's possible. May 02 14:52:12 polyonymous: no, you used some tricks, for eaxmple didn't do that using *opie* May 02 14:52:21 RP: About symbols, I suppose the only way to get them is to have unstripped libs in the target device, true? May 02 14:52:24 sirfred: Great. Feel free to mail us with any feedback (we've already made some comments on the oprofile list) May 02 14:52:28 ah, yes, I meant running opie image, not using opie. May 02 14:52:38 that is - without gpe dbus helper, etc. May 02 14:52:39 polyonymous: gotcha May 02 14:52:44 sirfred: Install -dbg packages is the way to do it atm May 02 14:52:58 RP: 4.0K modules-2.6.21-ezxdev-a780.tgz May 02 14:53:01 sirfred: In due course, we hope to support pointing it at OE's staging area too May 02 14:53:03 psokolovsky__, have you seen qte breakage yet? May 02 14:53:12 RP: That should be nice. May 02 14:53:17 koen: Nearly empty - I guessed that would happen :/ May 02 14:53:23 RP: I only looked at the file, not the size :( May 02 14:53:34 anyone tried this? www.ext4cow.com May 02 14:53:39 polyonymous: if you have plans to patch bluez3 to heal its dbus OCD, let me know ;-) May 02 14:53:41 err ext3com.com May 02 14:53:49 polyonymous: no breakage here with yesterday's build May 02 14:53:59 koen: The kernels are going to have to inject a package function May 02 14:54:02 err ext3cow.com ... look like I can't type today May 02 14:54:13 psokolovsky__, I may, but not right now that I've paired :) May 02 14:54:19 RP: looks like it May 02 14:54:24 psokolovsky__, hmm... with recent kernel headers? May 02 14:54:31 koen: I guess this is a chance to srtop duplicating all this stuff in each kernel .bb May 02 14:54:47 :) May 02 14:54:48 koen: i.e make a standard kernel do_deploy task May 02 14:54:53 polyonymous: I rebuilt yesterday opie-image with cvs20070428, committing now May 02 14:54:57 psokolovsky__, it's in bugs 2201 2128 May 02 14:55:19 psokolovsky__, in OE from scratch? May 02 14:55:21 koen: Anyhow, at least we know what the bugs are with this approach and they're solvable May 02 14:55:31 for angstrom? May 02 14:56:44 polyonymous: how qte-static crawls into opie? opie uses normal qte May 02 14:57:05 psokolovsky__, it's just in the name of the bug, it applies to both. my bug (2201) is for normal qte. May 02 14:57:31 because if this qte-static I filed another bug, thought it's inapplicable :) May 02 14:58:09 psokolovsky__, what's in your staging asm/page.h? (not kernel one) May 02 14:58:29 polyonymous: then it appears Z problem, it was reported only for tehm May 02 14:58:55 ah... Well, maybe. You use different kernel headers, I take it? May 02 14:59:07 polyonymous: no %) May 02 14:59:20 but I don't have any such issue, and I build for h4000/hx4700 May 02 14:59:25 hmm... May 02 14:59:31 weird. May 02 15:00:11 psokolovsky__, do you have it like this: http://rafb.net/p/6zfDaS70.html May 02 15:00:26 psokolovsky__: I am working on a solution May 02 15:00:34 I'll let you know May 02 15:00:47 This just likely going to be just data + browser May 02 15:02:06 Laibsch: that's cool, but some compression/lookup would be helpful. like CHM ;-) May 02 15:02:23 polyonymous: yes, that's what I have, but I also have h4000-angstrom-linux-gnueabi/kernel/include/asm-arm/page.h May 02 15:02:26 psokolovsky__, I think he went for compressed filesystem. May 02 15:02:53 psokolovsky__: Compression is squashfs May 02 15:02:54 polyonymous: can you paste ls -l of your staging? May 02 15:02:54 psokolovsky__, But It doesn't -I${STAGING_KERNEL_DIR}... I had to add it to bb... May 02 15:02:57 lookup is tracker May 02 15:03:19 like this - http://rafb.net/p/GDDuGS25.html ? May 02 15:03:27 Laibsch: already used tracker on pda? May 02 15:03:34 no May 02 15:03:41 koen committed only yesterda May 02 15:03:43 day May 02 15:03:51 cool! May 02 15:04:11 polyonymous: so, please check for page.h in spitz-angstrom-linux-gnueabi, and is it being used May 02 15:04:34 If not tracker then some other indexing solution. No indexing solution is also one small step forward if I only get wikipedia for my Z May 02 15:05:07 psokolovsky__, it's not. it IS being used after I add -I${STAGING_KERNEL_DIR} to .bb. Moreover, I had to patch #include to #define __KERNEL__ before including. Otherwise my kernel headers do not define it. May 02 15:05:32 Laibsch: yep, good plan May 02 15:05:43 psokolovsky__, basically, there are two solutions in these two bugs - one does this stance around kernel includes, the other defined PAGE* literally. May 02 15:05:55 I think we should pick one up and see if it breaks other archs. May 02 15:06:19 polyonymous: no, let's pick one which for sure won't ;-) May 02 15:06:45 psokolovsky__: I'll keep you posted. You'll see what is coming up. Thanks to koen who got me looking in the right direction. May 02 15:06:50 polyonymous: and again, ipaqs work, so you could try to compare that with Z May 02 15:06:53 psokolovsky__, that would be the one that #defines stuff literally. May 02 15:07:05 but I'm not sure if it doesn't, though... May 02 15:07:25 well, again, Z - the way it is in the tree doesn't :) May 02 15:07:35 polyonymous: I guess, safe solution is which uses PAGE_* from kernel May 02 15:08:22 psokolovsky__, but see the patch: http://bugs.openembedded.org/attachment.cgi?id=1395 May 02 15:09:33 polyonymous: well, that's good, I guess. Or maybe, the whole thing is linux-headers package issue? why it stages empty file? May 02 15:09:46 http://static.wikipedia.org/wikipedia/ja/%E8%AA%BF/%E3%81%B9/%E3%82%82/Wikipedia%7E%E8%AA%BF%E3%81%B9%E3%82%82%E3%81%AE%E6%A1%88%E5%86%85.html May 02 15:09:46 I fear that the navigation bar on the left is going to suck on a PDA, though. But I guess that might get solved talking to the person in charge at wikipedia. May 02 15:10:21 You also cannot readily redistribute recompressed static wikipedia (-> copyright since author info is left out) May 02 15:10:27 psokolovsky__, that is also a good question. I don't know the answer, though ;-) it looks like page.h stripped out of #ifdef __KERNEL__ stuff, maybe. May 02 15:10:41 Thus everybody has to redo it themselves :-( May 02 15:11:24 Laibsch: iirc andrew flegg has a custom css for that May 02 15:11:26 psokolovsky__, the bad part of the patch is #define __KERNEL__ before including, but if we test it on other archs, well, at least your ipaq, that should do... May 02 15:12:05 Laibsch: see http://bleb.org/software/770/ (scroll down to wikipedia) May 02 15:12:48 koen: Thank you VERY much for that. I had not inspected it yet but had hoped this could be solved by CSS May 02 15:13:11 it is a bit tricky with a static snapshot, but nothing that couldn't be solved May 02 15:13:22 I agree May 02 15:13:38 And in any case "one step after the other" May 02 15:13:44 polyonymous: what's kernel ver is for spitz anyway? can you paste its page.h too? May 02 15:13:53 koen: We have made a big step with your suggestion May 02 15:14:00 koen: for squashfs May 02 15:14:59 psokolovsky__, http://rafb.net/p/10ijrq71.html May 02 15:15:36 Laibsch: you'd have to thank likewise for getting squashfs support in OE, otherwise I wouldn't have know of its existance May 02 15:15:37 polyonymous: and what's that version? May 02 15:15:47 I'm waiting for localedef -f UTF-8 on Z to finish :) May 02 15:15:56 but basically .20* May 02 15:16:04 * Laibsch thanks likewise May 02 15:16:34 * koen waves to mallum May 02 15:17:28 polyonymous: well, in hh.org kernel, page stuff is out of __KERNEL__, and it's also .20. who's right? May 02 15:17:38 psokolovsky__ : i think the current version of linux-rp is 2.6.20.6 May 02 15:17:39 hey koen May 02 15:17:48 psokolovsky__, some other .20 then :) May 02 15:18:01 Laibsch: you're welcome, but I needed it myself anyway :-) May 02 15:18:58 psokolovsky__, linux-rp-2.6.20-r15 May 02 15:19:47 2.6.20.6-g50ac3650 to be precise :) May 02 15:21:18 polyonymous: well, what about alternative locations for PAGE_* stuff? shouldn't glibc define it? May 02 15:21:39 psokolovsky__, grep didn't give any satisfactory results :) May 02 15:21:56 that was of course the first thing I thought of :) May 02 15:22:27 "Sanitise the ARM headers exported to userspace. Signed-off-by: David Woodhouse". how funny. May 02 15:22:41 heh May 02 15:22:59 I'm just looking at gitweb :) May 02 15:23:15 polyonymous: well, now last q: can you please try once again that only #define __KERNEL__ is not enough? May 02 15:23:32 psokolovsky__, you mean, without -I ? May 02 15:23:38 polyonymous: yep May 02 15:23:47 polyonymous : The bitbake patch worked fine. thanks May 02 15:23:48 Well, I'm pretty sure, but lemme try. May 02 15:23:54 Ifaistos, you're welcome :) May 02 15:24:42 03ifaistos 07org.oe.dev * r27d1edd8... 10/ (1 packages/gcc/gcc_4.1.1.bb): May 02 15:24:42 packages/gcc/gcc_4.1.1.bb: Do not apply ppc soft-float patches when building with uclibc May 02 15:24:42 This is a dirty hack to get gcc 4.1.1 to compile for glibc AND uclibc on ppc May 02 15:24:42 the patch that is need it to get gcc support soft-floats with glibc, makes gcc fail with uclibc May 02 15:24:49 03ifaistos 07org.oe.dev * r645107cc... 10/ (1 conf/bitbake.conf): conf/bitbake.conf:make default values for jffs2, squashfs and squashfs-lzma generated images architecture neutral May 02 15:24:58 03ifaistos 07org.oe.dev * r67086aca... 10/ (1 conf/machine/magicbox.conf): conf/machine/magicbox.conf: Make sure squashfs-lzma image builds for big endian arch May 02 15:24:58 likewise: Oh, come one. I was under the illusion you did it just for me :-) May 02 15:25:07 03ifaistos 07org.oe.dev * re9a5ebf7... 10/ (5 files in 3 dirs): packages/initscripts/initscripts_1.0.bb:Add mountall.sh script for oplinux-uclibc May 02 15:25:11 03ifaistos 07org.oe.dev * radf21433... 10/ (3 files in 3 dirs): packages/uclibc/uclibc-cvs/magicbox: Add uclibc config for Magicbox May 02 15:25:42 psokolovsky__, building it now... Just a thought - maybe 2.3.12 worth looking at? May 02 15:25:58 On the other hand, it's not likely that this recent problem dealt with there... May 02 15:26:45 polyonymous: 2.3.12 qt/e? May 02 15:26:54 polyonymous: yep. plus, I personally still perplexed at what the heck .12 is. on d/l site where .11 lies, there's no .12. though hrw tells it exists somewhere... May 02 15:27:13 ahh... Well, here he is ;-) May 02 15:27:17 hrw, yes. May 02 15:27:20 ask ljp which qt/e is current :D May 02 15:27:31 qtopia-core 2.2 contain qt/e 2.3.12 May 02 15:27:40 psokolovsky__, | kernel/qpixmapcache.cpp:541: error: 'PAGE_SIZE' was not declared in this scope May 02 15:28:00 anyway I would rather stick with 2.3.10 as known as working May 02 15:28:15 well, I actually meant looking at, not switching to :) May 02 15:28:37 psokolovsky__, rebuilding with -I now. May 02 15:28:45 polyonymous: Ah, here's real fix: getpagesize() should be used, not compile-time constant May 02 15:28:57 Laibsch: Open Illusion May 02 15:29:01 psokolovsky__, ah... Okay, I'll look into it. May 02 15:29:02 polyonymous: ok, we did the testing then, let's go with that patch ;-) May 02 15:29:12 Laibsch: The Dream Build System and Metadata May 02 15:29:24 psokolovsky__, unless we want to go for the real fix? May 02 15:29:25 polyonymous: nope, I don't mean *we* should patch qte to do thaat now ;-) May 02 15:29:31 Ah okay :) May 02 15:29:36 polyonymous: that's just the asnwer why it's ni longer exported May 02 15:30:02 I see. We could do that as well, but let's see if the patch breaks things and if it's not we postpone it indefinitely :) May 02 15:30:22 ok May 02 15:34:04 psokolovsky__, don't feel like trying to build with the patch in your ipaq setup? Just for the sake of complete testing :) May 02 15:41:38 polyonymous: sure, I will do. results will be in few hours, I guess - gotta run now May 02 15:41:48 no hurry ;) May 02 15:43:03 polyonymous: Do you have RW access now? We promoted sirfred just this morning. If you have not done so, read the first paragraph on http://www.openembedded.org/wiki/MonotonePhraseBook May 02 15:43:40 yeah and I should read up more on monotone too. And no, I do not have rw access yet. May 02 15:45:06 BTW, I think your problem yesterday might possibly have been solved by "mtn merge". But you should have really just asked in #monotone instead of reverting. May 02 15:45:59 Laibsch, nope, merge didn't solve the problem, I think there are other ways than reverting, but I think in this particular case revert was the easiest approach. May 02 15:46:45 I've had locally added files which were getting in the way. And I should've gone to #monotone just to let them know it would be nice to give full path instead of just the name :) May 02 15:47:02 03pfalcon 07org.oe.dev * rba1cabe5... 10/ (1 conf/distro/include/preferred-opie-versions-1.2.3-pre.inc): preferred-opie-versions-1.2.3-pre: Bump CVS version to 20070428. May 02 15:48:58 cu guys May 02 15:49:24 hrw|gone: Later! May 02 15:52:23 After I sort out the russian multikey keymap problem I'm going to switch to angstrom's opie image as my primary image on handheld :) May 02 15:59:50 polyonymous: You were able to compile a full opie image? May 02 15:59:59 We need to get your patches in fast May 02 16:00:07 * Laibsch oogles psokolovsky__ May 02 16:00:17 Laibsch, yes. May 02 16:00:26 lftp currently does not install any files. I'm sure something simple is missing. If I add "autotools_do_install" in do_install the image/ dir gets populated, but no ipk is created May 02 16:00:36 Not only to compile, but to run and see if bluetooth works :) May 02 16:00:41 (sorry for the noob-type question) May 02 16:01:14 heh, actually, I only made statements. Anyone know of the solution? (there's the question) May 02 16:01:18 :) May 02 16:01:31 Laibsch, I'm rebuilding with the lastest update of opie-versions from psokolovsky, so, maybe more patches are to follow, I don't know :) May 02 16:02:27 tnb: Did it ever work in the past? May 02 16:02:42 Laibsch: don't know May 02 16:02:52 first time I've tried to build it May 02 16:03:08 tnb: checked the bug tracker? May 02 16:04:32 Laibsch: I'd think it should work given that lftp is referenced in the bitbake user manual :) May 02 16:09:50 tnb: checked the bug tracker? May 02 16:10:09 It might be that others have already the problem (and possibly a fix) May 02 16:10:45 Laibsch: none found May 02 16:11:03 well, then please report it May 02 16:11:18 I don't have time now but I might try and build it later May 02 16:12:00 Laibsch: I'll figure it out, then commit a fix May 02 16:12:31 that is even better May 02 16:16:20 the lzma compression really makes a difference May 02 16:16:26 -rw-r--r-- 1 stelios stelios 1916992 May 2 19:03 bootstrap-image-magicbox-20070502160041.rootfs.squashfs May 02 16:16:27 -rw-r--r-- 1 stelios stelios 1568832 May 2 19:03 bootstrap-image-magicbox-20070502160041.rootfs.squashfs-lzma May 02 16:18:39 Ifaistos: Indeed. I'm creating a tiny-tiny kernel+rootfs for linkstation/kurobox. I wouldn't be able to without lzma. May 02 16:19:12 sound is damn low... May 02 16:20:21 NAiL : the only "problem" is that you have to mount etc also in ram May 02 16:21:08 * Ifaistos wishes jffs2 had lzma compression May 02 16:21:11 well, my rootfs is strictly readonl ;) May 02 16:22:37 for devices that are set to do a task, that's much better and safer. For "generic" usage i.e when you need to add packages change things it creates a small problem May 02 16:23:13 yeah May 02 16:23:24 this rootfs is intended as a rescue-fs in flash. May 02 16:23:26 Ifaistos: I think RP added some patches for linux-rp to support jffs2 with lzma.... May 02 16:23:36 so it *should* be "untouchable" ;) May 02 16:23:56 hvontres|poodle: lzo May 02 16:24:36 RP : Any numbers regarding compression ? May 02 16:24:45 10-15% ? May 02 16:25:08 Ifaistos: Nearly as good as zlib May 02 16:25:49 RP: oops.. wrong lz May 02 17:08:56 svn is much nicer wrt to mv/cp stuff. And in svn/git combo it's even equally distributed :) May 02 17:10:23 * polyonymous is struck by RAS syndrome :) May 02 17:10:31 RAS? May 02 17:11:10 would it be possible to "pacakge" qemu-native , so I can use it on different machines? May 02 17:11:34 chouimat, redundant acronym syndrome, I think. May 02 17:11:43 ok May 02 17:11:57 http://en.wikipedia.org/wiki/RAS_syndrome May 02 17:12:29 keesj, iirc it dislikes gcc version used as gcc-native. May 02 17:14:41 polyonymous: yes , my problem with building is that I am compiling with 4.1 . I have installed 3.4.6 now but I don't want to do that all the time on all machines :p May 02 17:15:17 keesj, ah, by packaging you mean binary package? May 02 17:15:31 yes May 02 17:15:40 keesj, that is what you better do on the system level and assume it's provided. May 02 17:15:57 Because OE, you know, works on different host archs. May 02 17:16:38 I just disabled binary locales generation, but I can do that on Z with swap :) May 02 17:16:39 * JustinP has Gentoo and just uses gcc-config to switch compilers for qemu-native May 02 17:16:58 Yup, that's also the possibility. May 02 17:17:12 JustinP: I will try the same. May 02 17:17:31 I also have gentoo, but thought it's easier to generate locales on Z :) May 02 17:18:18 I don't really care about locales right now (specialy not for the tools I am building ) May 02 17:18:45 cripts would not work with french error messages May 02 17:36:03 Ola May 02 17:40:33 well, newer opie image is built okay. May 02 17:41:05 hmm... that's about it. I can go for switching roots, woohoo :) May 02 17:41:55 Laibsch: it works when FILES_${PN} are added May 02 17:42:07 thought it has to do with FILES_ May 02 17:42:39 (s/it/lftp/) May 02 17:42:54 sorry for not telling you, didn't feel experienced - never written .bb, just fixing them :) May 02 17:43:51 tnb: now is the time to write a bug report and document the fix May 02 17:45:13 ah, forgot the russian multikey thing... Dunno if it should go into OE tree, but it's sitting on opie bugtracker for ages now.. May 02 17:47:17 polyonymous: it should go to Opie, ask psokolovsky__ or ljp to commit, or ask for commit rights right-a-way May 02 17:47:52 hello?? May 02 17:48:11 zecke, I already talked to psokolovsky__ about it and he said he'd do his best to get it through, but as I understand he's just opie-in-oe maintainer. May 02 17:48:28 polyonymous: as I suggested, you should drop an email to opie-devel, with request to include easy patches by you into 1.2.3 May 02 17:49:00 is there any one here who can give me some advice on how to set up OE? May 02 17:49:02 polyonymous: but his kernel stuff and coordination makes him a highly trusted pal May 02 17:49:19 paul_uk_lime: generally, just ask :) May 02 17:49:25 psokolovsky__, yes, I told you don't feel like subscribing and I don't feel like bombarding list from the outside even if allowed... May 02 17:49:45 polyonymous: and otherwise, I didn't mention that, because I saw your changes as good thing to do anyway, but multijey ri layout works ok for me with angstrom (and it should be using utf-8 by default) May 02 17:49:49 zecke, well, I'm only reiterating his words. I highly trust him :) May 02 17:50:31 psokolovsky__, no, it doesn't use utf8 locale. And without turning on utf8 locale you don't get utf8 filenames. May 02 17:50:54 polyonymous: how comes it doesn't? May 02 17:51:14 psokolovsky__, it may well use utf8, for its own purpose, but not locale, not for filenames. May 02 17:51:27 well, it's never set to en_US.utf8 May 02 17:51:47 polyonymous: you mean LC_ALL not set or what? May 02 17:52:03 psokolovsky__, well, LC_CTYPE in particular. ALL would also do, of course. May 02 17:52:34 and this is exactly what breaks multikey loading and allows for utf8 filenames. May 02 17:52:47 polyonymous: well, angstrom should be utf-8 by default, you should post about that to angstrom-devel May 02 17:53:12 psokolovsky__, before doing that I should check "supported" image :) May 02 17:53:21 thankyou! im trying to to set up OE and i have followed the instructions on openembedded and have got to the point where i need to update my copy of the OE database. Except it tells me that i need to 'migrate' my copy. When i try to migrate my OE.mtn i get told ...'mtn: calculating necessary migration steps May 02 17:53:21 mtn: error: database schema 48fd5d84f1e5a949ca093e87e5ac558da6e5956d is unknown; cannot perform migration' i have followed the instructions to the letter, any suggestions? May 02 17:54:10 paul_uk_lime: what version of monotone do you have? May 02 17:54:26 0.31 May 02 17:54:28 paul_uk_lime: you have downloaded a too new db file for a old version of mtn :) May 02 17:54:34 paul_uk_lime: which file did you download? May 02 17:54:48 hm May 02 17:54:56 ERROR: No providers of build target angstrom-minimal-image.bb (for []) May 02 17:54:59 paul_uk_lime: http://www.openembedded.org/snapshots/OE-this-is-for-mtn-0.31.mtn.bz2 is probably the one you want May 02 17:55:05 woglinde: set BBPATH :) May 02 17:55:27 ok i'l give it a go, thankyou!:) May 02 17:56:04 zecke it is set May 02 17:56:08 paul_uk_lime: after wget and bunzip2 you need to pull and this might take some time May 02 17:56:13 woglinde: how many bbfiles were parsed? May 02 17:56:14 args May 02 17:56:19 shound add .bb May 02 17:56:20 heheh May 02 17:56:31 shpuldnt May 02 17:56:33 args May 02 17:59:48 ok, im getting the file just now... May 02 18:04:29 woglinde: You sent med a link here, but shouldnt glib-2.0 have been build without any local methods? May 02 18:05:09 goxbox hm as I understand it there is something mixed up with iconv in your envoriemnet May 02 18:07:53 ahh right May 02 18:08:36 hmm.. btw sound in my opie works... May 02 18:08:53 I was told it shouldn't :) May 02 18:09:07 system sounds? May 02 18:09:15 goxboxlive, yes, screen taps. May 02 18:09:24 do you use angstrom? May 02 18:09:38 yes, angstrom and today's cvs bump of opie. May 02 18:09:50 I haven't tried it yesterday with older cvs. May 02 18:10:29 but it's veeery low. This is a module issue, though. May 02 18:11:23 ok, good. I'll try my self as soon as i have solved the glib-2.0 issue May 02 18:11:47 I was surprised :) May 02 18:12:50 okay, I've sent it upstream and I can provide interested parties with a patch against oe as well. I just didn't fancy the idea of taking care of the file on each reinstall. May 02 18:13:03 polyonymous: so, tell us again, what machine do you use? May 02 18:13:09 borzoi May 02 18:13:15 borzoi? May 02 18:13:15 that is spitz. May 02 18:13:30 yup. 3100 May 02 18:14:22 woglinde: Is there something i can rebuild to do something about the mixed up iconv in my envirement? May 02 18:17:00 goxbox sorry dont know May 02 18:17:29 ok May 02 18:17:51 lastime I build glib-2.0 it works May 02 18:17:57 polyonymous: see privmsg May 02 18:18:02 but there are serval changes since then oin the .bb May 02 18:18:22 psokolovsky, I haven't got any. have you identified yourself with nickserv? :) May 02 18:21:10 oehm May 02 18:21:23 why angstroem minimal needs gtk-doc-native-1.0-r1 May 02 18:21:40 woglinde, after all it's native :) May 02 18:22:44 hm python native and glib2-native too May 02 18:22:51 * woglinde shakes head May 02 18:26:25 okay, now I'm rebooting into my old image, backup the whole thing and switch roots :) May 02 18:28:23 ljp: ping May 02 18:29:43 hello all May 02 18:32:30 mreimer_: Hi! May 02 18:32:59 hi psokolovsky! May 02 18:33:04 polyonymous: feel free to add your sound experience to http://bugs.openembedded.org/show_bug.cgi?id=2211 May 02 18:33:29 okay. May 02 18:41:30 * polyonymous is backing up the whole microdrive. May 02 18:45:16 Laibsch: we need more bugs for opie ;-) see current dep tree: http://bugs.openembedded.org/showdependencytree.cgi?id=2186 May 02 18:46:00 wtf, where's CIA? May 02 18:50:04 hey zecke May 02 18:50:32 this 2181 issue isn't clear to me. May 02 18:50:34 ljp: did you get my mail? May 02 18:50:44 Anyone recommend any commercial JTAG programmer/debuggers that have good support from OSS? May 02 18:51:20 it isn't clear what the original issue was - the one "reproduced" by Laibsch with the log attached is fixed but it's not necessarily the one reported. May 02 18:51:45 vervain: Commercial? As in pre-built, or do you want support for it? ;) May 02 18:51:55 NAiL, pre-build programmer?/? May 02 18:52:04 (pardon my spelink) May 02 18:52:08 home-grown May 02 18:52:20 ah, that's me :) May 02 18:52:21 polyonymous: pre-assembled? !home-made? May 02 18:52:30 vervain: http://cgi.ebay.com/JTAG10-Parallel-Port-JTAG-Programmer-for-Blackfin-ARM_W0QQitemZ230124683519QQihZ013QQcategoryZ50915QQssPageNameZWDVWQQrdZ1QQcmdZViewItem May 02 18:52:41 I've got a couple of those. I'm using them for ARM and PPC atm. May 02 18:52:53 NAiL, never tried to disassemble programmers before. I'm sure all commercial programmers are disgusting inside. May 02 18:53:09 They're cheap, wiggler-compatible (ie, dog slow), but they work. May 02 18:53:12 ;-) May 02 18:54:21 zecke: yes. May 02 18:54:34 ljp: comments :) May 02 18:55:03 stioll thinking on how best to do this May 02 18:56:00 ljp: a) you could build a toolchain and bare-bone immage and ask your folk to integrate Qtopia to it May 02 18:56:16 ljp: this would help you to learn :) May 02 18:56:23 ljp: then you could do the same with uclibc May 02 18:56:30 ljp: and then you can move on to ppc and mips May 02 18:56:34 and redo :) May 02 18:56:55 psokolovsky, nice tree. I'm going to try irda with palm. It didn't work quite well with bt phone. May 02 18:57:02 morning May 02 18:57:20 psokolovsky, not bt, of course (even though the phone is bt:)) May 02 18:57:30 psokolovsky and polyonymous: Thanks for the great work. I will check it out and report more bugs as requested ;-) May 02 18:57:36 bitbake opie-image? May 02 18:57:45 or what is it called these days? May 02 18:57:48 Laibsch, furthermore, opie-kdepim-image :) May 02 18:57:59 Laibsch, opie-image is the one, yes. May 02 18:58:02 oh, we have kdepim's back? May 02 18:58:08 Laibsch, ah... May 02 18:58:09 AWESOME! May 02 18:58:13 Laibsch, no, not in the tree. May 02 18:58:17 It's in bugzilla, though. May 02 18:58:28 polyonymous: We will have it soon, though, I am sure. May 02 18:58:39 polyonymous: where? May 02 18:58:41 Sure, it is working fine here. May 02 18:58:45 hang on May 02 18:58:50 Laibsch: opie-image is basic image, let's take it as groundline, and first of all make it perfect May 02 18:59:01 psokolovsky: Sounds reasonable May 02 18:59:07 psokolovsky: regarding the sound problems you had, did you have snd_mixer_oss and friends loaded? May 02 18:59:11 Ifaistos: lzma give good compression but is slow - too slow for using it on device May 02 18:59:11 But compilation will likely take me a while. May 02 18:59:19 Laibsch, http://bugs.openembedded.org/show_bug.cgi?id=1076 May 02 18:59:37 hrw: Too slow for ondevice *de*compression as well? May 02 18:59:40 hvontres|poodle: yes, that was static kernel with all the stuff. the sound worked in gpe and didn't - in opie May 02 18:59:50 Laibsch, took me almost 5 hours, but including patching on the way :) May 02 18:59:59 Laibsch: decompression is fast enough May 02 19:00:06 polyonymous: You are my hero, REALLY! May 02 19:00:15 superhero, I hope? :) May 02 19:00:20 psokolovsky: I think gpe is using alsa directly, while opie goes through the oss compatibility layer. May 02 19:00:30 hrw: lzma is about compressing it once and not caring about time May 02 19:00:33 hvontres|poodle, yes, qss May 02 19:00:38 hrw: Only size is what counts. May 02 19:00:51 hrw: And I would always compress on the desktop May 02 19:00:55 Laibsch: I know May 02 19:00:58 psokolovsky, have you installed qss? It's not installed by default on opie-image, I believe. May 02 19:01:04 hvontres|poodle: no ui framework stuff in our realm uses alsa directly ;-) May 02 19:01:14 Laibsch: I mentioned lzma to few people here as good companion for squashfs May 02 19:01:29 polyonymous: best hero of all (with sugar and cream on top) May 02 19:01:40 polyonymous: likely, no May 02 19:01:41 * Laibsch pads polyonymous for a job very well done May 02 19:01:44 psokolovsky: I also stumbled across an issue with the early alsa implementations not having a master volume channel May 02 19:01:44 Yeah, I consume a lot of sugar :) May 02 19:01:56 * Laibsch pads psokolovsky, too, for his continued service May 02 19:02:23 hvontres|poodle: sound worked for me in familiar 0.8.4 opie May 02 19:02:41 Laibsch, I take it pads as in pats? :) May 02 19:02:54 Laibsch: thank you too for starting this new wave, I'm sure we'll get opie up to speed soon! May 02 19:03:51 btw, this beast we have in opie - it's an OZ beast? :) May 02 19:04:14 polyonymous: I need to check the meaning of pats. "to pad" is a verb. "auf die schulter klopfen" May 02 19:04:37 ah ok May 02 19:04:38 polyonymous: nahh, I think that one is an opie creation :) May 02 19:04:49 psokolovsky: I humbly do my best (which unfortunately cannot match the skills of most around here) May 02 19:04:56 hvontres|poodle, ok. Thought it's the opened zaurus :) May 02 19:05:34 polyonymous: nahh, that one actually came out looking pretty cool... May 02 19:05:46 Laibsch, strange, I couldn't find this meaning. Nevermind, though. May 02 19:06:14 polyonymous: http://opie.handhelds.org/cgi-bin/moin.cgi/Opie_20General_20FAQ?action=highlight&value=mascot May 02 19:07:00 psokolovsky, thank you! Now at least I can call him by name :) May 02 19:07:09 hmm.. bin-utils-cross is failing to build here May 02 19:07:18 ljp: amd64 or x86? May 02 19:07:28 Laibsch, "to pat so. on the back jmdm. auf die Schulter klopfen" (said dict.leo.org) May 02 19:07:33 ljp: what distro/machine? May 02 19:07:39 polyonymous: You are right. It is indeed "to pat" not "to pad" May 02 19:07:44 Laibsch, ah ok. May 02 19:07:51 polyonymous: Just checked leo.org, too May 02 19:07:54 :))) May 02 19:07:54 angstrom/tosa May 02 19:08:01 host is x86 May 02 19:08:11 ljp: pastebin bug May 02 19:10:47 hmm pastebin doesn't like it - Query failure: Can't open file: 'pastebin.MYI'. (errno: 145) May 02 19:11:02 ~pastebin May 02 19:11:06 rumour has it, pastebin is a place to paste your stuff without flooding the channel - try http://pastebin.ca, or http://channels.debian.net/paste, or http://rafb.net/paste/, or http://pastebin.com is usually painfully too slow and unresponsive to use, use one of the other pastebin sites, or dpaste.com is a very nice pastebin as well May 02 19:11:09 use other site May 02 19:11:22 pastebin.osuosl.org is very nice, and fast. May 02 19:12:31 http://pastebin.ca/467839 May 02 19:12:59 I see May 02 19:15:06 ljp: up-to-date OE? May 02 19:15:40 ljp: this bug is workarounded in binutils.inc May 02 19:16:43 should be up to date, started fresh on new machine at work May 02 19:16:59 polyonymous: if you get a chance, could you roll me a poodle opie image for testing? I am still trying to sort out my kernel build failures May 02 19:17:19 hvontres|poodle, sure, I'll try. May 02 19:17:33 ljp: http://pastebin.ca/467854 in binutils.inc take care of it May 02 19:17:45 ljp: btw - which version of qt/e 2.3.x is last one? May 02 19:17:50 NAiL: I mean commercial as in commercial product not necessariy support. I'm currently looking at the Atmel JTAGICE but I need to make sure we have gdb support. May 02 19:17:51 polyonymous: thanks. drop me a line at hvontres(AT)gmail(DOT)com May 02 19:18:08 hvontres|poodle, ok. May 02 19:18:13 hmm.. 2.3.11? May 02 19:18:17 hvontres|poodle, but I'm more likely to drop you a line here :) May 02 19:18:19 would have to check May 02 19:18:26 We're investigating next gen system which will be an arm9 and I'm looking at the JTAG implementation for programming/testing. May 02 19:18:46 ljp: qtopia 2.2 contain 2.3.12/13 iirc May 02 19:19:28 polyonymous: ok, but I may not be on later :) May 02 19:19:30 hvontres|poodle, but I can't generate you binary locales :) May 02 19:19:39 * hvontres|poodle is on PDT May 02 19:19:52 hvontres|poodle, I'm not sure when the image is complete, anyway :) May 02 19:20:00 psokolovsky, polyonymous: I just thought it would be a nice idea to have a dual-GUI GPE/Opie image as well to get more people to try out GPE stuff. I will open a bug report and try working on it time permitting. Feel free to comment and work on it. May 02 19:20:15 seem to recall later qt/e versions only in qtopia 2.2 ... May 02 19:20:47 Laibsch, I also thought about it, but I'd postpone it till we have both working :) May 02 19:20:53 Laibsch: such image will be big probably May 02 19:21:49 hvontres|poodle, http://rafb.net/p/UfG3d095.html is that ok? May 02 19:22:02 hrw, just fine for dogs with microdrive :) May 02 19:22:08 hrw: thats alread in binutils.inc May 02 19:23:12 ljp: no idea then. I built binutils-cross-sdk so many time during last weeks... on misc x86, amd64 boxes May 02 19:23:14 polyonymous: looks good to me :) May 02 19:23:25 hvontres|poodle, okay, opie-image or opie-kdepim-image ? May 02 19:23:52 polyonymous: lets start with plain opie-image first :) May 02 19:24:02 ok :) May 02 19:24:08 polyonymous: I still have my OZ setup for daily work :) May 02 19:24:28 parsing :) May 02 19:25:47 Laibsch: nice idea, in my deep todo is creating LiveCards (cf. LiveCD) for people to try stuff ;-) May 02 19:26:10 psokolovsky, if only all devices could boot cards.. May 02 19:26:38 polyonymous: yep, that would be major trouble to get it really cross-device ;-) May 02 19:26:57 polyonymous: well, if we can get kexec to twiddle the cmdline....:) May 02 19:27:49 hvontres|poodle, even if we can, it is likely to require some knowledge level having which user wouldn't really need "livecard" to try things out :) May 02 19:30:01 yep, assuming target audience are folks with orginal Zs, it would take to find a security hole in embeddix kernel and reboot it into new kernel ;-) May 02 19:30:13 fortunately, it's easier with wince ;-E May 02 19:30:26 ;-) May 02 19:30:29 polyonymous: actually, we could probably come up with some sort of statup script the user can run from the card that could take care of some of the more advanced things...we would probably need to custumize the card image for different devices May 02 19:31:11 psokolovsky: ahh, drat, I forgot about the crapix kernels...I've been on 2.6 for almost a year now...:) May 02 19:31:19 hvontres|poodle, startup script? For what I remember original sharp rom wouldn't let you kexec a thing. May 02 19:31:29 exactly. May 02 19:32:22 Now I need to find out how can I swap the contents of / and /chroot/ live.... :) May 02 19:35:37 heh, I found the way, it seems. May 02 19:38:12 cu May 02 19:40:54 hrw|gone: I know. But there should be solutions for the size. This will not be a flash installation for collie for example. May 02 19:46:42 time to flash the new kernel May 02 19:48:47 May 02 19:49:13 koen: not sure if it matters, but we are missing the tail section of this patch: http://gcc.gnu.org/bugzilla/attachment.cgi?id=12705 May 02 19:54:11 woohoo, I'm in. May 02 20:21:01 where? May 02 20:21:48 where in?:) May 02 20:22:15 * polyonymous has just swapped his old primary OZ root and brand new angstrom opie root. May 02 20:29:53 polyonymous: well, for me, sound doesn't work eevn with opie-qss and a ipaq-specific workaround which wasn't applied previously ;-( May 02 20:30:31 psokolovsky, then the bug confirmed for your machine. May 02 20:38:57 hi, I'm trying to bitbake zroadmap... and it fails on building qte-mt-2.3.10 with `clipbounds' undeclared... can anyone advice on this? please? May 02 20:40:06 hmm... qte-mt should build fine... May 02 20:40:27 re May 02 20:40:52 polyonymous: do I really have to build all these things just for zroadmap? can't I just build this package? May 02 20:41:11 Yes, I'm afraid, you do. May 02 20:41:30 zecke: ping May 02 20:41:41 polyonymous: so any idea where to start looking what i have done wrong? May 02 20:42:05 do you have recent tree? May 02 20:42:15 that would be the first thing to check. May 02 20:42:24 ljp: pong May 02 20:42:48 polyonymous: well... it's a couple of weeks old... I'll update it and try again May 02 20:43:33 hvontres|poodle, "your" build failed :) May 02 20:44:28 did you get the email about an address for the greenphone? May 02 20:51:31 I know this has been asked before but I cannot find the solution for it now; I removed "qemu-native" from ASSUME-PROVIDED and now it complains about gcc4. gcc3 is also installed. I read something about sdl binaries and dev packages but wonder if that is relevant here and what packages it would be on ubuntu. Any help? May 02 20:51:56 I do have libsdl-dev installed (well, the package it points to, it is a virtual package on ubuntu) May 02 21:02:19 polyonymous: what did it fail on? May 02 21:02:53 hvontres|poodle, bad RPATH in gcc-cross. I don't understand it - I have 'return False' commented out in my insane class... May 02 21:02:58 lookinginto it now. May 02 21:04:11 weirdness. May 02 21:11:21 hvontres|poodle, I was building from the wrong tree :)) May 02 21:13:24 polyonymous: ooops...:) May 02 21:14:21 hvontres|poodle, yup. Building afresh now. May 02 21:17:22 ljp: yes, I got two mails. We had a holiday yesterday, and I wrote a rather lengthy reply to Knut May 02 21:19:16 ljp: ah I got both mails from you now May 02 21:27:53 zecke: Sorry to bother you. But can you tell me what the solution was again for non-compiling qemu-native? See my question from 22:51 May 02 21:28:46 laibsch ENABLE_BINARY_LOCALE_GENERATION = "0" May 02 21:29:06 Laibsch: you don't bother. who complains? sane.bbclass or do you get a failure on do_configure? May 02 21:30:19 woglinde: Thank you. I had that enabled previously, but koen said it will likely create broken images May 02 21:30:26 Laibsch: I do ;-) May 02 21:30:34 laibsch test it May 02 21:30:37 zecke: Let me post the output May 02 21:30:41 my gpe image was functional May 02 21:30:54 woglinde: you are a latin1 moron ;) May 02 21:31:02 um yes May 02 21:31:54 Laibsch, does machine have enough swap/mem to generate locales? May 02 21:34:08 zecke: http://rafb.net/p/uGCpIT54.html May 02 21:34:39 polyonymous: yes, I think so. It has ~700MB of RAM, headless machine, LAN server May 02 21:34:51 Laibsch: could you run ./configure --help in the qemu dir for me? May 02 21:34:52 Laibsch, should do the trick, then. May 02 21:35:07 psokolovsky: What should do the trick? May 02 21:35:12 zecke: hang on May 02 21:35:29 polyonymous, you can do without binary locales, you can generate them on device. May 02 21:35:34 ouch May 02 21:35:42 Laibsch, that was for you, I'm talking to myself now :) May 02 21:36:03 Laibsch: there was patch posted reported to allow build qemu with gcc4 May 02 21:36:55 sorry, p* guys. Mixed 'ya up ;-) May 02 21:37:24 at least you didn't confuse yourself with one of us :) May 02 21:37:52 zecke: http://rafb.net/p/Nc5IAD34.html May 02 21:38:26 Laibsch: Do you have a gcc-3.3 on your system? May 02 21:38:56 check if there is EXTRA_OEFLAGS="--cc=gcc-3.3" in the .bb file May 02 21:39:29 zecke: yes May 02 21:39:36 gcc 4 and 3 both installed May 02 21:40:06 cool klee im fernsehen May 02 21:40:08 oops sorry May 02 21:40:37 altboot doesn't build for spitz! May 02 21:40:50 zecke: zecke, who the f... is klee? May 02 21:40:58 music :) May 02 21:41:04 polyonymous: It built for me. But it did not work. May 02 21:41:20 zecke: I thought you were talking about Kleeberg. You know the guy? May 02 21:41:20 I hope you know it. you might recognize it from commercials May 02 21:41:25 Laibsch: well, but did you try patch at #2157 after all? May 02 21:41:26 no May 02 21:42:14 zecke: I hardly get commercials, not having a TV and haunting everyone with BDSG that sends me junk mail via snail mail May 02 21:42:41 zecke: I was wondering since he is a celebrity chef cook from Berlin. May 02 21:42:55 !oebug 2157 May 02 21:42:56 * * Bug 2157, Status: NEW, Created: 2007-04-29 03:49 May 02 21:42:57 * * bero(AT)arklinux.org: qemu fails to compile with gcc 4.x May 02 21:42:58 * * http://bugs.openembedded.org/show_bug.cgi?id=2157 May 02 21:43:08 Ah, there seems to be some info May 02 21:43:53 Laibsch, for spitz? May 02 21:44:19 polyonymous: Well, it might have been for collie. May 02 21:44:30 Actually I think it *was* for collie May 02 21:44:31 Sorry May 02 21:45:33 zecke: It looks like there is some sophisticated if structure in qemu-native.inc that should set EXTRA_OEFLAGS="--cc=gcc-3.3" May 02 21:45:40 But I am not sure it really does May 02 21:45:53 Laibsch: do you have a gcc-3.3? May 02 21:46:41 ah... | install: cannot stat `/home/hacker/OE/angstrom/tmp/work/spitz-angstrom-linux-gnueabi/altboot-1.1.1+wip-SVNR30-r1/trunk//spitz/altboot-menu/*-*': No such file or directory May 02 21:46:53 zecke: Will http://rafb.net/p/9ATRzK31.html do what I think it is there to do? May 02 21:47:40 zecke: package name is gcc-3.3-base, version is 3.3.6-13ubuntu2 May 02 21:48:04 I also have 4.1 and 4.0 May 02 21:48:06 Laibsch: well, for the third and last time. Do you have an executable called gcc-3.3 May 02 21:48:22 zecke: Must have missed the first two ;-) May 02 21:48:49 gcc-3.3-base isnt qnough May 02 21:48:53 gcc-3.3-base isnt enough May 02 21:48:56 Laibsch, you could catch up at the fourth and the last later on :) May 02 21:48:57 which `gcc-3.3` shows nothing May 02 21:49:48 cu May 02 21:49:59 see you May 02 21:50:00 Laibsch: apt-get install gcc-3.4 May 02 21:50:01 laibsch you nee gcc-3.3 paclage May 02 21:50:09 or 3.4 May 02 21:50:14 Thanks, guys May 02 21:50:25 hrw: Maybe add to openembedded-essential? May 02 21:51:05 Laibsch: don't be lazy, please try that patch, if you have this issue! ;-) May 02 21:52:11 psokolovsky: I am not lazy. Don't worry I will look at the bug. May 02 21:52:20 But unlike my computer I am single-task May 02 21:52:41 Laibsch, upgrade :) May 02 21:52:43 although I can breathe and type at the same time ;-) May 02 21:52:58 Laibsch: it's just too many people have been having this issue. if qemu 0.9 indeed buildable with gcc4, let's just make it such, and forget about... May 02 21:53:06 that sounds like RT-11 BG/FG monitor :) May 02 21:54:22 Strange, "dpkg -l" did not list gcc-3.3 as a package, only gcc-3.2 May 02 21:56:09 than you have gcc-3.2 May 02 21:56:15 as command May 02 21:57:04 * Laibsch finally took a look at bug 2157 May 02 21:57:07 psokolovsky: Have you tried the patch? May 02 21:57:39 Laibsch: nope, I don't have such prob ;-) May 02 21:57:54 woglinde: No, it was shown as uninstalled. With "dpkg -l" you should get installed and installable packages. At least that is what I think. May 02 21:58:01 psokolovsky: Why not? May 02 21:58:11 Laibsch: in queue ;-) May 02 21:58:28 I'm not sure I have gcc4 installed ;-) May 02 21:59:00 Why EXTRA_OECONF="--disable-gcc-check" if this should really solve the issue? May 02 21:59:05 Seems a bit weird May 02 21:59:23 Laibsch, probably patch doesn't take care of gcc detection part? May 02 22:06:27 polyonymous: I removed the gcc3.x packages to test the patch May 02 22:07:11 psokolovsky: I removed the gcc3.x packages to test the patch May 02 22:07:12 Laibsch, I meant the --disable and fix. The patch from busybox guys doesn't touch autoconf. May 02 22:07:32 nice, thanks! May 02 22:10:05 OK, fails without the patch May 02 22:10:33 hmm.. ipkg: (null): Bad address -- seen that? May 02 22:12:36 polyonymous: yes. The times I saw it was harmless and I ignored it May 02 22:13:11 Laibsch, yes, seems to be harmless here, just wondered what the hell is that. May 02 22:15:50 psokolovsky: configure ran through this time May 02 22:16:15 polyonymous: ipkg needs lots of love, too ;-) May 02 22:16:20 Laibsch: with patch? May 02 22:16:40 psokolovsky, uhh... Dunno if I'm ready to love it :) May 02 22:18:26 actually, I'm rather bored with ipkg's slowness, probably, I'll try to look into it on a calm day (but where such come?..) May 02 22:18:32 psokolovsky: Yes, second run with patch. compile still going strong. May 02 22:19:52 psokolovsky: that would be great :) May 02 22:20:07 I also had such idea... May 02 22:20:19 yeah, it's much slower on angstrom for some reason... May 02 22:20:19 * mmp is also waiting for one calm day May 02 22:20:28 Hmm.. after suspend/resume my g-ether doesn't work... May 02 22:20:55 polyonymous: at least something. that's compensation for working sound ;-P May 02 22:21:03 psokolovsky, :)))) May 02 22:22:51 it used to spit a few some usb reset messages after loading, but now it doesn't... May 02 22:35:06 03philippe 07org.oe.dev * r8aeb8aa9... 10/ (1 packages/gpe-mininet/gpe-mininet_svn.bb): May 02 22:35:06 gpe-mininet: added new svn build recipe May 02 22:35:06 packages/gpe-mininet/gpe-mininet_svn.bb May 02 22:35:06 * new svn build recipe May 02 22:38:45 polyonymous: I'll be taking off in about 1/2 hour. Thanks for working on an image for me :) May 02 22:39:07 I'll pass your thanks to the cpus. Both cores :) May 02 22:39:31 It's at glibc, so I think you'll get the image tomorrow, May 02 22:39:38 polyonymous: what kind of box are you running on ? May 02 22:40:03 I'll go to sleep in about half an hour too. amd64 x2 2.2GHz May 02 22:40:20 (that is 4200+) May 02 22:40:43 polyonymous: hehe...my last attempt from scratch only took about 20 hrs on my poor old dual P-II 333's :) May 02 22:41:05 hvontres|poodle, much faster here. May 02 22:41:24 and it builds way more than I would expect it to build :) May 02 22:53:04 night May 02 22:53:29 night all May 02 23:13:45 * koen finds out that the job tomorrow has been cancelled May 02 23:14:09 koen: Lucky you ;-) May 02 23:15:02 he koen May 02 23:15:27 yeah, means that I will get more than 6 hours of sleep :) May 02 23:15:29 hey woglinde May 02 23:16:10 Laibsch: a dual gpe/opie image is a bad idea, since libxine and libsdl are incompatible with libsdl-qpe and libxine-qpe May 02 23:16:31 koen hm minimal-angstroem shipps bluez-utils and bluezprobe which need perl-native pyhton-native cups and serval more native packages May 02 23:16:50 that longs the buildtime May 02 23:17:26 woglinde: no, minimal-image depends on task-base, which contains all that and is built because of debian.bbclass May 02 23:17:39 yes May 02 23:17:55 but I dont need bluez-cups in a minimal image May 02 23:19:29 it isn't in there May 02 23:19:58 args yes but the build time May 02 23:20:10 I am littel stubborn May 02 23:20:12 *g* May 02 23:20:25 [01:17] koen: woglinde: no, minimal-image depends on task-base, which contains all that and is built because of debian.bbclass May 02 23:20:59 so fixing debian.bbclass? May 02 23:21:18 no, put all task-* in their own recipes May 02 23:25:23 koen: It was possible on OZ so I am sure there is a way to make this work on angstrom. May 02 23:25:41 Let's see May 02 23:25:51 It would certainly be nice. May 02 23:26:08 Laibsch: you can get two basic GUIs running at the same time May 02 23:26:22 but as soon as you install libsdl or libxine, you loose May 02 23:26:52 unless you use 2 chroots, and at that point the advantage of a dual-image is gone May 02 23:27:24 koen different RPATHS? May 02 23:27:32 or is this too much overhead May 02 23:28:16 Laibsch: something like altboot and booting an image from sd would be better May 02 23:36:11 nite May 02 23:48:15 psokolovsky__, the qte-mt patch breaks fresh build, it seems. Well, I'd say it doesn't fix it. We need to depend on staged kernel. May 02 23:48:51 polyonymous: then, consider hacking in getpagesize() ;-) May 02 23:49:00 psokolovsky__, or depending on the kernel May 02 23:49:09 how do we depend on the kernel? May 02 23:49:40 polyonymous: no-how. generic userspace should not depend on kernel May 02 23:50:02 psokolovsky__, what do you mean "generic"? It's machine-specific. May 02 23:50:47 after all it's built in work/spitz*/ :) May 02 23:50:51 polyonymous: only "kernel userspace" (modutils, blah-blah) may depend on kernel. May 02 23:51:15 polyonymous: plus, it is no machine specific, it being so is a software defect, being fixed gradually. May 02 23:51:25 Hmm.. May 02 23:51:54 matchbox-panel depends on virtual/kernel.. May 02 23:52:21 polyonymous: that sux May 02 23:52:29 but indeed it can be avoided here with getpagesize() thing... May 02 23:52:43 polyonymous: plus, kernel really should be built before userspace May 02 23:52:57 polyonymous: should be avoided, I'd say ;-) May 02 23:53:29 psokolovsky__, if it was built before userspace then we wouldn't even need to depend on it, we would just rely on it :) May 02 23:54:16 polyonymous: that's how it works, I'd think. for ipaqs, kernel is built much before UI stuff. May 02 23:54:44 well, I just started a fresh poodle build for hvontres and it failed. That means kernel wasn't built before UI.. May 02 23:55:14 then adding such a dependency shouldn't hurt, I think. May 02 23:55:53 polyonymous: just not on opie May 02 23:56:06 what do you mean not on opie? May 02 23:56:32 polyonymous: no-how. generic userspace should not depend on kernel May 02 23:57:12 well, you've just said kernel should be built before userspace. We do not make kernel depend on qte-mt, it's the other way around. May 02 23:57:28 In general, I share this sentiment about generic userspace... May 02 23:58:03 Well, okay, nevermind, I'll hack in a getpagesize() solution and that will be it. May 02 23:58:09 polyonymous: please find proper fix for this dep problem, or just fix qte crap May 02 23:58:16 yep, thanks May 02 23:58:38 I think the dep on virtual/kernel should be fine and should do the trick, but since I'm at it, I'll see into getpagesize(). May 02 23:58:53 The question here is if it would be slower or not. May 02 23:59:41 heh May 02 23:59:47 I tried to build angstrom-e-image... May 02 23:59:54 | evas_fb_main.c:533: error: 'PAGE_MASK' undeclared (first use in this function) May 02 23:59:54 :) May 03 00:00:22 I'll got for a getpagesize(). May 03 00:00:39 I'm afraid there will be too many packages depending on virtual/kernel otherwise :) May 03 00:28:12 psokolovsky__, http://bugs.openembedded.org/show_bug.cgi?id=2201 May 03 02:31:01 (09:11:03 PM) xjqian: the README in the opie-1.2.3 reads OPIE II, but I though the opie-1.2.3 is just a minorly improved (release frozen) version of opie-1.2.2 (a.k.a opie I) May 03 02:31:01 (09:11:10 PM) xjqian: now I'm confused May 03 02:31:01 (09:14:51 PM) xjqian: the qtopia/qt files in the opie-1.2.3 do look like newly released version from Trolltech. hence it does indicate OPIE II from the perspective it uses new qtopia libraries May 03 02:31:01 (09:16:26 PM) xjqian: so does opie-1.2.3 mean a mixture of new qtopia libraries with old qtopia applications, which implies that most of the application will be broken and need to be rewritten? May 03 02:31:47 sorry, but nobody in the opie channel responded. so I'm trying my luck here **** ENDING LOGGING AT Thu May 03 02:59:56 2007