**** BEGIN LOGGING AT Thu Aug 10 02:59:56 2006 Aug 10 03:37:25 poushag: it's not something likely to be added to monotone itself, so sort of out of my hands... is viewmtn not working for you? Aug 10 06:50:14 njs, viewmtn works but it doesnt appear to allow me to get whole history of an individual package Aug 10 06:51:48 does package mean... a file? a directory? Aug 10 06:51:57 (I don't know how OE arranges things) Aug 10 06:52:24 njs: iirc, it's multiple files with optional subdirectories Aug 10 06:53:38 it means a package - all the stuff that is sourced in single buildfile i think - does that sound right nabyss? Aug 10 06:55:15 poushag: Yeah.. mtn doesn't store any of the metadata re links between the .bb file and anything that's sourced.. so you're up the creek unless it's all in one folder Aug 10 06:55:27 morning Aug 10 06:55:31 well, you could write some tool that did that, I guess :-) Aug 10 06:55:59 (knew about links between .bb file and things it depended on, and also knew how to trace history) Aug 10 06:56:26 njs, i wish i could - just not part of my skillset unfortunately Aug 10 06:56:32 http://sourceware.org/bugzilla/show_bug.cgi?id=3004 Aug 10 06:56:34 hehe Aug 10 06:57:16 hey zecke Aug 10 06:57:43 zecke, why do you hate intel architecture? Aug 10 06:57:48 hehe Aug 10 06:58:01 poushag: it is full of legacy crap and not optimal hacks Aug 10 06:58:08 at least apple uses EFI Aug 10 06:58:31 but ia64 will fix all that! Aug 10 06:59:09 ...indeed, at the moment, it looks like the next version or so of monotone may have ia64-optimized sha1, but not x86-optimized... Aug 10 06:59:19 njs: ys, IA64 was promising. It looks like it is killed Aug 10 07:00:16 It seems AMD is considering about opening ATI drivers. Aug 10 07:01:36 a Aug 10 07:02:37 zecke, the way you describe intel products sounds much like those of a certain other monopolistic company.... Aug 10 07:03:21 poushag: ?? I like for intel opening so much stuff Aug 10 07:03:32 poushag: specially I just read the announcement by KeithP Aug 10 07:03:52 poushag: just the ia-32 architecture is old, CISC... Aug 10 07:04:27 well i think they actually have serious competition in amd so that is helping Aug 10 07:05:15 the new interrupt structure for virtualization will be nice Aug 10 07:14:52 good morning all Aug 10 07:16:01 koen: hi Aug 10 07:16:10 "So everything's great, unless you're one of those fringe weirdos that Aug 10 07:16:11 uses an x86." Aug 10 07:16:16 * koen likes njs Aug 10 07:18:58 zecke: which announcement by keithp? Aug 10 07:20:10 morning all Aug 10 07:27:02 | checking sysdep dirs... configure: error: The arm is not supported. Aug 10 07:27:02 | FATAL: oe_runconf failed Aug 10 07:27:02 NOTE: Task failed: /home/kergoth/code/oe/build/tmp/work/arm-linux/glibc-2.4-r6/temp/log.do_configure.5869 Aug 10 07:27:10 * kergoth is confused Aug 10 07:28:43 morning Aug 10 07:28:56 kergoth: I think I cracked the fan problem on shuttle Aug 10 07:29:20 morning XorA Aug 10 07:29:34 hey mikearthur early start this mornng? Aug 10 07:30:06 XorA: ah, hows that? Aug 10 07:30:39 hmn Aug 10 07:30:49 XorA: yeh, have to get some medication on the way home, pharmacy shuts at 5 :s Aug 10 07:30:56 kergoth: got a TLS enabled glibc-initial? Aug 10 07:31:11 kergoth: lm-sensors contains a script called fancontrol, as it stand the script assume linear mapping of number 0-255 onto fan speed, fix this assumption and it works quite well Aug 10 07:31:38 kergoth: and will adjust the fan speed higher as temp rises Aug 10 07:32:23 When I try to make a bootstrap-image do_rootfs ends with an error. I says he can not find the package task-bootstrap. The package exists and stands in the 'Packages' file. The name of the package is task-bootstrap_1.0-r25_coolfox-II.ipk Aug 10 07:32:47 morning Aug 10 07:32:51 hey hrw|work Aug 10 07:33:49 Cobelius: tried manually building that package? Aug 10 07:33:54 Cobelius: arch looks weird Aug 10 07:36:20 what could be a solution for this problem? Aug 10 07:37:47 Cobelius: s/coolfox-II/collfoxII/ tried? I do not know does ipkg handle '-' in arch names Aug 10 07:38:44 well it worked before I updated bitbake and OpenEmbedded Aug 10 07:40:11 Cobelius: check etc/ipkg/arch.conf then Aug 10 07:40:25 Cobelius: IPKG_ARCHS handling was changed a bit Aug 10 07:40:40 okay Aug 10 07:41:46 arch.conf in rootfs ofcourse Aug 10 07:42:45 the file is not existing Aug 10 07:44:35 meh Aug 10 07:44:48 bulding from scratch show a lot of missing DEPENDS Aug 10 07:45:05 hmn I'll rename the Machine and try it again Aug 10 07:45:43 koen: when you commit packaged staging to mainlne someone can setup a buld cruncher that tests every .bb for from scratch :-) Aug 10 07:45:53 :) Aug 10 07:46:02 I'm writing *$%(#*$(@# docs for that right now Aug 10 07:46:56 * rwhitby just received a Freecom FSG3 freebie :-) Should be a good SlugOS target ... Aug 10 07:48:52 v8jlene: don't forget to add yourself to the copyright section in the docs Aug 10 07:48:54 zecke: http://www.openembedded.org/user-manual&dpage=ch07 Aug 10 07:50:12 Cobelius: can you show machine conf? Aug 10 07:52:26 yes Aug 10 07:53:18 http://pastebin.de/10554 Aug 10 07:56:13 Cobelius: IPKG_EXTRA_ARCHS = "x86" instead of IPKG_ARCHS Aug 10 07:56:36 Cobelius: otherwise you do not get coolfox-II in ipkg archs Aug 10 07:57:09 ah okay, so IPKG_ARCHS get filled automatlically and I don't have to manipulate it? Aug 10 07:57:46 yes Aug 10 07:57:53 ah okay Aug 10 07:57:55 thank you Aug 10 07:58:06 IPKG_ARCHS='all none any TARGET_ARCH MACHINE IPKG_EXTRA_ARCHS' iirc Aug 10 07:58:19 do I need to rebuild everything from scratch now? Aug 10 08:00:18 Cobelius: no Aug 10 08:00:29 okay Aug 10 08:04:35 03koen 07org.oe.dev * r346469f1... 10/ (1 packages/xorg-xserver/xserver-kdrive_git.bb): xserver-kdrive-git: fix DEPENDS Aug 10 08:10:27 new net connection at work ;) Aug 10 08:12:14 zecke: eh, this is an arm build, no glibc-initial is built at all Aug 10 08:12:18 meh Aug 10 08:12:36 oe hates me lately, flat out fails to get even a toolchain built Aug 10 08:12:57 kergoth: glibc 2.4 uses nptl now, so you need glibc-i Aug 10 08:13:12 kergoth: or hack linuxthreads back in Aug 10 08:13:12 glibc-i? Aug 10 08:13:25 intermediate or initial Aug 10 08:13:26 i presume you meant intermediate Aug 10 08:13:33 I always confuse the two Aug 10 08:13:42 right, it built intermediate, then exploded in glibc saying it doesnt support arm Aug 10 08:13:50 ah Aug 10 08:14:01 zecke had asked about initial :P Aug 10 08:14:05 :) Aug 10 08:14:22 I suspect that's because only the eabi crowd is using 2.4 on arm Aug 10 08:15:13 heh Aug 10 08:15:22 libpng got with the program and uses pkgconfig now Aug 10 08:15:31 I did however forget to inherit it Aug 10 08:15:47 symptoms: -L@libdir@ everywhere in makefiles Aug 10 08:17:12 heh Aug 10 08:18:14 hmmm Aug 10 08:18:17 that's not it Aug 10 08:20:35 JustinP: ping Aug 10 08:22:20 JustinP: I found why emacs-x11 does not build in .oz354x - full x11 cannot be choosen because pref_provider for x11 is diet-x11 Aug 10 08:22:27 JustinP: will build it by hand Aug 10 08:27:57 hrw|work: ehm, that will mess up the feeds Aug 10 08:28:11 hrw|work: since diet-x11 and libx11 produce the same ipkg Aug 10 08:28:44 hrw|work: it will also mess up future builds Aug 10 08:31:12 ok then no emacs-x11 for OZ Aug 10 08:32:13 koen: http://sourceware.org/bugzilla/show_bug.cgi?id=3004 Aug 10 08:32:36 zecke: yes, I read that Aug 10 08:32:59 hrw|work: you could put emacs + libx11 in a seperate feed Aug 10 08:33:11 hrw|work: people advanced enough to use emacs..... Aug 10 08:33:41 koen: time for contrib/DEVELOPERNAME/ feeds then Aug 10 08:34:51 hrw|work: that's one option Aug 10 08:35:03 the other option is to statically link in libx11 Aug 10 08:36:09 a sed a day keeps the doctor away Aug 10 08:36:31 I love the ego trip in that bug Aug 10 08:37:29 XorA: One needs to love Drepper Aug 10 08:37:49 zecke: I assume they are bitter about the lack of GNU/OSX Aug 10 08:38:21 XorA: gnu/osx is missing because glibc needs readelf to compile ;) Aug 10 08:38:34 iirc there's a gnu/darwin project Aug 10 08:38:36 03koen 07org.oe.dev * r53586020... 10/ (1 packages/libpng/libpng_1.2.12.bb): libpng 1.2.12: fix libpng.pc Aug 10 08:39:15 kergoth: I hope you don't mind me putting a ton of sed in one the .bbs you MAINTAINER Aug 10 08:39:22 * XorA reads the horrors of win32 driver development Aug 10 08:39:56 feel free, i havent actively maintained in a while, though thats gonna be changing as soon as i can get a damn successful build on a regular basis :P Aug 10 08:56:05 lol -> http://go-team-ari.livejournal.com/14323.html Aug 10 08:56:15 morning kergoth Aug 10 08:56:21 he pb Aug 10 08:56:42 hi woglinde Aug 10 08:57:10 hey pb__ Aug 10 08:57:15 hi kergoth Aug 10 08:57:38 morning all Aug 10 08:57:49 hi pb RP Aug 10 08:57:51 hi rp Aug 10 08:58:33 03koen 07org.oe.dev * r103510e2... 10/ (1 packages/pango/pango_1.14.0.bb): pango: update to 1.14 Aug 10 08:58:44 hey RP Aug 10 08:59:16 do13: ping Aug 10 09:08:07 koen: binutils 2.17 just compiled on OSX ( a patch was needed) Aug 10 09:08:19 he zecke Aug 10 09:08:50 zecke: cool Aug 10 09:10:36 morning Aug 10 09:11:07 hi CoreDump|home Aug 10 09:11:58 ah Aug 10 09:12:10 I finally found the unresolved symbols bug Aug 10 09:12:16 hey CoreDump|home Aug 10 09:12:17 koen hm? Aug 10 09:12:40 woglinde: libpng: unresolved symbols to Aug 10 09:12:55 hu? Aug 10 09:13:15 hope It have nothing to do with my patches Aug 10 09:13:26 hi CoreDump|home Aug 10 09:13:31 hi hrw Aug 10 09:13:32 woglinde: no, libpng has a *really* broken .pc Aug 10 09:13:39 woglinde: I added a -lz Aug 10 09:13:44 hi hrw|work Aug 10 09:13:55 koen hm -lpng needs -lz? Aug 10 09:13:57 CoreDump|home: any plans to fix poodle keyboard in gpe? Aug 10 09:14:00 that suckz Aug 10 09:14:01 koen: that sounds like a broken .so, not a broken .pc file Aug 10 09:14:07 woglinde: yes Aug 10 09:14:27 than compiling isnt rigth Aug 10 09:14:31 rpath-link Aug 10 09:14:37 hrw|work: of course, eventually Aug 10 09:15:23 * CoreDump|home just ordered the parts for his new PC =) Aug 10 09:15:32 coredump amd64? Aug 10 09:15:46 nope, Core 2 Duo Aug 10 09:16:09 pfff Aug 10 09:16:13 32bit is dead Aug 10 09:16:15 *g* Aug 10 09:16:21 heh Aug 10 09:16:42 but it sooo blows opteron and x2 out of the water performence wise Aug 10 09:16:52 and ms-office dont profit from dual-core Aug 10 09:16:57 *duck and hide* Aug 10 09:17:04 ~lartv woglinde Aug 10 09:17:23 woglinde: it probably means the stupid papers clip only brings one cpu to a standstll Aug 10 09:17:50 xora I alwasy switch to mother earth Aug 10 09:20:58 03koen 07org.oe.dev * r913ac3f0... 10/ (1 packages/libpng/libpng_1.2.12.bb): libpng: add -lz to the .pc to fix cairo->pango->gtk+. I'll leave it up to the knowledgeable people to fix up the .so Aug 10 09:23:15 hm Aug 10 09:30:31 jesus Aug 10 09:30:42 libpng is the next candidate for autotools Aug 10 09:34:20 yay Aug 10 09:34:33 ipkg .163 solved a segfault in .159 Aug 10 09:35:12 03koen 07org.oe.dev * r8cbe2a36... 10/ (1 packages/ipkg/ipkg_0.99.163.bb): Aug 10 09:35:12 ipkg: add .163 Aug 10 09:35:12 * no new bugs noticed after >1 week of on-device testing Aug 10 09:36:34 hi florian Aug 10 09:36:34 hi florian_kc Aug 10 09:39:05 good morning Aug 10 09:44:15 koen: re sylpheed plugin, is that ${PV} valid? Aug 10 09:44:39 XorA: I copied the maildir plugin Aug 10 09:44:45 03koen 07org.oe.dev * red4ac7ec... 10/ (1 packages/angstrom/task-angstrom.bb): task-angstrom: remove dbus-glib hack Aug 10 09:45:09 koen: then you typoed it Aug 10 09:45:30 koen: mailmbox_1.12.1_2.3.0.bb ?? Aug 10 09:45:52 XorA: doesn't it need the sylpheed-c version as well? Aug 10 09:46:04 XorA: please correct if it's wrong Aug 10 09:46:09 koen: I though _ was illegal in ${PV} Aug 10 09:46:10 * koen <- sylpheed n00b Aug 10 09:46:33 koen: all the ones I did are -1.12.1_${PV}.bb Aug 10 09:46:55 where ${PV} comes from the extra-plugins release version Aug 10 09:49:17 ah Aug 10 09:49:57 mainly because I couldnt think of a way to make upgrades work when there are two different ${PV}s Aug 10 09:53:34 03koen 07org.oe.dev * r7152ad3b... 10/ (3 files in 2 dirs): sylpheed-claws-plugin mailmbox: de-typo PV Aug 10 09:56:12 koen: damn I was hoping you had a genious idea for versioning there :-) Aug 10 09:56:26 XorA: not after midnight Aug 10 10:00:34 | ./bits/endian.h:13:2: #error Machine byte order unknown. Aug 10 10:00:35 argh! Aug 10 10:00:38 glibc hates me. Aug 10 10:00:54 glibc hates everyone Aug 10 10:02:04 it hates me more Aug 10 10:02:06 but its mutual Aug 10 10:15:45 kergoth: I seem to remember that error from somewhere. Can't think where though :-/ Aug 10 10:18:24 good morning Aug 10 10:18:38 hey Dirk Aug 10 10:19:07 hi dirk Aug 10 10:19:16 hi Kerwood ade|desk Aug 10 10:19:25 morning do13_ Aug 10 10:19:34 hi Koen, Marcin, mikearthur Aug 10 10:19:36 who is Chris Humbert? Aug 10 10:20:17 mikearthur: http://oz.drigon.com/ Aug 10 10:23:10 do13: whats this? Aug 10 10:25:01 Chris Humberts page Aug 10 10:25:04 mikearthur: webpage of Chris work on tosa usb Aug 10 10:25:16 ok, cool Aug 10 10:26:36 hrw|work: morning Aug 10 10:27:05 do13_: is the LCD/touch interefe Aug 10 10:27:07 woops Aug 10 10:27:27 do13_: is the LCD/touch interference problem on 2.6/Tosa just a Tosa problem, or does it affect other Sharp hardware? Aug 10 10:31:25 mikearthur: tosa only Aug 10 10:31:34 mikearthur: It affects the others too. Aug 10 10:31:49 There is a fix in the ad78xx ts driver. Aug 10 10:31:57 ops Aug 10 10:32:27 hi Aug 10 10:32:52 mikearthur: the ts driver waits a specified time after a sync signal and then scans the touch Aug 10 10:33:38 can anybody give me a small tip? i still have problems with bitbake task-bootstrap (asked some weeks ago)... Maybe i have to change something in my local.conf to use another glibc version? Aug 10 10:33:38 the tosa 2.4 driver does this in the same way. The idea is to use the mask input of the wm9712 instead of waiting in a loop Aug 10 10:34:00 yeh, I was looking at the 2.4 code for it Aug 10 10:34:13 hm. I have a provider, gst-plugins-base (see http://gstreamer.freedesktop.org/documentation/splitup.html for their reasons to have multiple plugin pacakges), which emits, for example, gst-plugin-audioconvert. Now i have an application which depends on this plugin. I thought DEPEND=gst-plugins-base and RDEPEND=gst-plugin-audioconvert would be the solution - but i was wrong. I'm getting "ERROR: Nothing provides runtime dependency gst-plugin-audioconvert Aug 10 10:34:42 so basically, what there is to do, is port the relevent fix from the ad78xx to the wm97xx code Aug 10 10:34:43 This once worked, with half-a-year (or so) old version of openembeded, and i remember that "something changed regarding RDEPENDs" Aug 10 10:35:24 03koen 07org.oe.dev * rd0c77a4c... 10/ (1 packages/libpng/libpng_1.2.12.bb): libpng: *sigh* more .pc munging Aug 10 10:35:27 tmbinc: you need to set PACKAGE_DYNAMIC = "gst-plugin*" in gst-plugin-base iirc Aug 10 10:35:36 tmbinc: but better ask RP Aug 10 10:35:39 hi tmbinc :) Aug 10 10:35:52 mikearthur: basically yes. better use the mask input. this would do the sync inside the wm9712 Aug 10 10:35:56 koen: but gst-plugins-notbase also give gst-plugin* :( Aug 10 10:36:05 hrw|work: exactly Aug 10 10:36:15 should i rename the packages which are emitted by gst-plugins-base to "gst-plugin-base-audioconvert" ? Aug 10 10:36:17 but that's ugly. Aug 10 10:36:20 hrw|work: that's why you need the DEPENDS Aug 10 10:36:46 tmbinc: no renames Aug 10 10:36:49 mikearthur: with the mask input we do not waste time waiting for the sync signal Aug 10 10:37:07 dyanmic package naming must *must* must be unique. If its not, it needs to be changed so it is Aug 10 10:37:13 do13_:yeh, makes sense. Does the ads78xx have a similar mask bit? Aug 10 10:37:26 mikearthur: I don't think so Aug 10 10:37:33 ok, cheers Aug 10 10:37:54 RP: unique so you can make out the provider name out of a package name? Aug 10 10:38:25 mikearthur: If the mask input won't work (for whatever reason) we have the software workaround Aug 10 10:39:10 do13_: aye. Is this ads78xx fix in mainline, or in the patched openzaurus kernel or in another patch (that I currently don't have) Aug 10 10:39:16 tmbinc: If you have dynamically generated packages gst-plugin-*, you can't have static packages that match that expression (say gst-plugin-base if its not generated dynamically) Aug 10 10:39:35 mikearthur: mainline, in the corgi-ts driver Aug 10 10:39:43 koen: you are writing Re: Help a newb? mails :) Aug 10 10:40:08 RP: it's called gst-plugins-base Aug 10 10:40:36 tmbinc: That should work ok then Aug 10 10:40:36 however both gst-plugins-base and gst-plugins-ugly (and some others) are emitting gst-plugin-..s Aug 10 10:40:56 and i don't know in advance which one is emitting what. That would be a big list i want to avoid. Aug 10 10:42:28 tmbinc: Ah, I've just looked at the .bb files. You can't have multiple and different providers with the same PACKAGES_DYNAMIC Aug 10 10:42:43 ok :/ Aug 10 10:42:50 any hints how to go on? Aug 10 10:43:13 tmbinc: Make each one emit some unique name string - gst-plugin-ugly-* Aug 10 10:43:24 :/ ok Aug 10 10:43:34 You can still use the .inc with some variable magic Aug 10 10:44:02 yep it's more that the end-user, when installing packages, shouldn't need to know about the plugin splitup Aug 10 10:44:06 but i guess that's impossible now Aug 10 10:44:39 tmbinc: Even with the different names, why would the user need to know about the splitup? Aug 10 10:45:20 because the package would be named gst-plugin-base-audioconvert, not gst-plugin-audioconvert Aug 10 10:45:27 but it's a minor issue, i agree. Aug 10 10:46:12 tmbinc: If they know to install gst-plugin-*, I'm not sure I see the problem Aug 10 10:47:12 tmbinc: If the names really bother you, you could have some PROVIDES magic although I think it will just get too complicated Aug 10 10:52:01 tmbinc: Having all these files PROVIDES gst-plugins also looks wrong btw Aug 10 10:52:19 tmbinc: Simply because they don't all provide the same thing Aug 10 10:53:40 what about creating a .bb that builds them all in one run? Aug 10 10:53:46 like the collie-kernels did Aug 10 10:54:40 that's probably a good idea Aug 10 10:55:22 is there special magic available to do this? Aug 10 10:56:07 tmbinc: Have a look in packages/linux for the collie kernel example Aug 10 10:56:53 kergoth: that "machine byte order unknown" error is usually symptomatic of a mixed-up cvs checkout Aug 10 10:57:12 florian_kc: good morning Aug 10 10:57:27 pb__: hi Aug 10 10:57:33 hail zecke Aug 10 10:58:35 hmm cool, i managed to make glib segfault in a function the documentation says it would do nothing in unix environments Aug 10 10:58:44 pb__: ah, will check that. thanks Aug 10 10:59:12 kergoth: check the bits/ mangling in the .bb Aug 10 10:59:15 florian_kc: awesome Aug 10 10:59:34 kergoth: I think v8jlene corrected some flaws in it Aug 10 10:59:41 RP: i don't get the idea completely there - does not just collie-kernels-2.4-embedix.bb DEPENDS on the different other kernels? so, you mean all i need is creating a gst-plugins.bb which DEPENDS on gst-plugins-good, -base, -ugly etc.? Aug 10 10:59:43 my knowledge of hte toolchain stuff isnt what it used to be :) /me needs to fix that Aug 10 11:00:25 tmbinc: exactly Aug 10 11:00:33 tmbinc: Basically yes, which solves some of your problems but perhaps not all... Aug 10 11:00:53 kergoth: the trouble is that the 2.4 .bb is based on the cvs one, and we haven't got rid of all the cvs-isms Aug 10 11:06:02 03florian 07org.oe.dev * r61158ddc... 10/ (1 conf/machine/htcuniversal.conf): htcuniversal.conf: Update to latest Goxboxlive version. Aug 10 11:06:07 03florian 07org.oe.dev * r4d81e47b... 10/ (1 packages/bluez/bluez-utils-nodbus_2.25.bb): bluez-utils-nodbus: Add ti patch used in bluez-utils-dbus. Aug 10 11:06:12 03florian 07org.oe.dev * r12e9019b... 10/ (3 files in 3 dirs): irda-utils: Add HTC Universal support to init script. Aug 10 11:06:18 03florian 07org.oe.dev * r283f041a... 10/ (1 packages/linux/handhelds-pxa-2.6_cvs.bb): Aug 10 11:06:18 handhelds-pxa-2.6_cvs: Add Universal to compatibility list, don't ship an image in Aug 10 11:06:18 its filesystem. Aug 10 11:06:23 03florian 07org.oe.dev * r991900e1... 10/ (1 packages/opie-taskbar/opie-taskbar.inc): opie-taskbar.inc: Add Universal bits by Goxboxlive. Aug 10 11:06:28 03florian 07org.oe.dev * r5c865738... 10/ (3 files in 3 dirs): sysvinit: Fix inittab for Universal. Aug 10 11:20:01 03koen 07org.oe.dev * r79101310... 10/ (1 packages/eel/eel_2.12.2.bb): eel 2.12.2: fix DEPENDS Aug 10 11:20:06 03koen 07org.oe.dev * rd91baaa5... 10/ (1 conf/machine/htcuniversal.conf): htcuniversal.conf: unbreak it Aug 10 11:25:25 koen: weird, glibc-intermediate failed once Aug 10 11:37:55 zecke, I have been thinking about my u-boot problems Aug 10 11:38:34 and am thinking about trying to unset vars like CFLAGS etc so u-boot can set the flags itself Aug 10 11:39:42 my thinking is that since you configure u-boot for the specific machine Aug 10 11:40:05 it will do a better job of setting compile options than OE, for the case of building the bootloader Aug 10 11:42:21 Crofton: understandable. theres a reason we dont override the flags in kernel builds either :) Aug 10 11:43:08 hmm, you have become nocturnal? Aug 10 11:43:39 How do you avoid overriding the flags in the kernel builds? Aug 10 11:44:39 working for a company thats 8 hours ahead of you tends to make one nearly nocturnal :) Aug 10 11:44:44 yeah Aug 10 11:45:00 and living in the Phoenix area over the summer might do the same :) Aug 10 11:45:32 see the MAKEFLAGS variable. by default it has -e, which lets env override vars set from makefiles. a lot of packages will set MAKEFLAGS = "'somespecificflagvar=-O'" or whatever, thereby overriding the -e and passing in only what you want passed in Aug 10 11:45:37 hehe Aug 10 11:45:50 ok Aug 10 11:46:52 meeting... Aug 10 11:47:44 Crofton: er, s/MAKEFLAGS/EXTRA_OEMAKE/g Aug 10 11:48:02 ok Aug 10 11:48:06 silly memory, pulling up gmake variable names instead of oe variable names Aug 10 11:50:41 copying the EXTRA_OEMAKE from u-boot 1.1.2 ... Aug 10 11:51:02 I was cutting out settings I didn't undertsnad ... Aug 10 11:56:33 kergoth, koen: I've built glibc 2.4 for epia and sh4 without too many issues. Needed both glibc-initial and glibc-intermediate and needed make 3.81 for epia. No idea about arm. Aug 10 11:56:47 v8jlene: ok, thanks for the info Aug 10 11:56:49 see http://www.mail-archive.com/oe@handhelds.org/msg02118.html for details on epia build, sh4 is pretty much the same Aug 10 11:58:00 hm ship make with oe? Aug 10 12:01:17 now I get: Aug 10 12:01:18 arm-linux-ld: unrecognized option '-Wl,-rpath-link,/home/balister/oe_home/tmp/staging/arm-linux/lib' Aug 10 12:01:28 linking u-boot Aug 10 12:02:11 Crofton: unset if good Aug 10 12:02:18 Crofton: -Wl are gcc options Aug 10 12:02:23 Crofton: we mostly use gcc for linking Aug 10 12:02:33 Crofton: unset LDFLAGS as well Aug 10 12:02:43 yeah Aug 10 12:02:57 I don't think u-boot should try and link against stging :) Aug 10 12:03:15 try and should are different type of things :) Aug 10 12:03:20 heh Aug 10 12:03:25 woglinde: Is that in reference to my needed make 3.81 for epia? I just built make-native 3.81 (after adding it, just a cp of the 3.80 files) and then rebuilt.. that picked up the make build by oe just fine. I have a fix for make 3.80 but I suspect that'll have side effects ;( Aug 10 12:21:26 RP: it would be neat if OE could do this: http://www.0d.be/2006/08/10/manu-cornet-is-a-gtk-torturer/ Aug 10 12:22:17 ok I can build u-boot from git Aug 10 12:22:31 had to force unset of LDFLAGS, CFLAGS, and CPPFLAGS Aug 10 12:22:51 Crofton: congrats Aug 10 12:22:54 koen: First step would be to add it to OE :) Aug 10 12:23:21 now I need to get up the courage to fash it into my OSK :) Aug 10 12:27:26 koen: I have an EABI toolchain on OSX Aug 10 12:27:58 zecke_: yay! Aug 10 12:28:56 zecke_: now I have to convince my parents to contribute to the "Koen wants a macbook pro" fund :) Aug 10 12:29:03 oops Aug 10 12:30:56 why would you want a MacBook pro? Aug 10 12:31:10 one can create professional sofware on macbook's as well Aug 10 12:31:36 zecke_: http://ewi546.ewi.utwente.nl/tinderbox/showlog.pl?machine_id=79&logfile=20060810134519.log <- it's 'e-image', not 'e17-image' Aug 10 12:31:42 I know Aug 10 12:31:48 I do not want to halt my script Aug 10 12:31:51 zecke_: &*$#&*# coating Aug 10 12:37:09 koen: private flags = 4000002: [Version4 EABI] [has entry point] Aug 10 12:37:12 is that any good? Aug 10 12:37:22 yes Aug 10 12:55:51 koen: glibc gets cured by magic built Aug 10 12:57:03 Does anyone remember the purpose of IMAGE_FSTYPE (not IMAGE_FSTYPES) ? Aug 10 12:58:11 RP: its old var before FSTYPES was born iirc Aug 10 12:58:34 lrg: hey Aug 10 12:58:39 morning all Aug 10 12:58:41 hi Liam Aug 10 12:59:00 hrw|work: So it doesn't have a real purpose anymore? Aug 10 12:59:04 hey Liam Aug 10 12:59:30 RP: iirc not have Aug 10 12:59:36 * RP adds to the cleanup list... Aug 10 12:59:52 RP: is that list in the wiki? Aug 10 13:00:01 koen: It used to be Aug 10 13:00:12 * RP has no idea if it was migrated Aug 10 13:00:41 I meant my personal list of things to do sometime but it would be better on a global one if we have it... Aug 10 13:01:22 'if it is not in email then it does not exist' Aug 10 13:01:46 hrw|work: The one on the wiki was dicussed via email :) Aug 10 13:02:39 ;D Aug 10 13:03:07 email is scary, lets add a forum to oe.org Aug 10 13:03:11 * koen runs Aug 10 13:10:29 hrw|work: yes, I know....but you dropped both emacs in ox354x Aug 10 13:11:03 who does the openzaurus vector art thats on the splashscreens? Aug 10 13:11:30 mikearthur: that is the result of a contest Aug 10 13:11:32 JustinP: I see normal emacs Aug 10 13:11:34 Personally, I can't stand it :) Aug 10 13:12:20 RP: then remove it from own images Aug 10 13:12:36 hrw|work: I already do ;-) Aug 10 13:12:43 the oz bootsplash looks funny on my netbook Aug 10 13:12:47 and h5550 Aug 10 13:13:31 hrw|work: ah...nm, then Aug 10 13:13:45 koen: it should give info that it lack image Aug 10 13:14:11 hrw|work: as I said, I can give you ipks for both. While emacs-x11 will only work if people switch to real x11, it would still be good to have in contrib I think Aug 10 13:14:22 hrw|work: emacs takes forever to build... Aug 10 13:14:34 JustinP: give me it and I will create feed/contrib/ with it Aug 10 13:15:30 JustinP: does diet-x11 in .oz have the all the patches from ade|desk? Aug 10 13:16:33 I don't know what you mean Aug 10 13:16:55 * JustinP uses full x11 in his oz354x images... Aug 10 13:17:13 ok, time for me to wake up....back in a half hour or so Aug 10 13:17:16 those patches stub out a lot, so it could be emacs builds against diet-x11 with those patches Aug 10 13:20:59 time to write 'oz-make-release-archives.sh' script Aug 10 13:21:38 damnit Aug 10 13:21:46 I want heading home and starts raining Aug 10 13:24:36 bye till later Aug 10 13:28:11 hmm.. collie opie-image for 3.5.4.2-rc1 is a bit too big ;( Aug 10 13:53:43 koen: there are still more stubs i need to look into for vnc stuff but i haven't had time to look as yet, maybe when i am on holiday i can get a better idea Aug 10 14:02:00 ade|desk: I was getting at patches getting lost for no obvious reason Aug 10 14:04:52 ade|desk: you made patches for stubs for the diet stuff? Where can i find them? Aug 10 14:07:13 JustinP: in .dev Aug 10 14:09:07 ah Aug 10 14:09:21 do they help with vnc then? Aug 10 14:09:27 * JustinP will try synergy Aug 10 14:22:11 ade|desk nope.... XkbUseCoreKbd, XkbLockGroup Aug 10 14:24:15 * koen kicks CIAbot Aug 10 14:24:31 03mickeyl 07org.oe.dev * rd0d679b8... 10/ (1 packages/opie-networksettings/opie-networksettings.inc): opie-networksettings: remove me as MAINTAINER and set it to Team Opie Aug 10 14:24:36 03koen 07org.oe.dev * r963bef78... 10/ (1 packages/linux/handhelds-pxa-2.6_cvs.bb): handhelds-pxa-2.6: set maxdate for iwmmxt patch Aug 10 14:24:41 03koen 07org.oe.dev * r08e70776... 10/ (3 files in 3 dirs): handhelds-pxa-2.6: add defconfig for h5xxx Aug 10 14:24:47 03koen 07org.oe.dev * rebfa533f... 10/ (1 conf/machine/htcuniversal.conf): Aug 10 14:24:47 htcuniversal.conf: "Since TARGET_ARCH is in IPKG_ARCHS by default you can lose the "arm" in Aug 10 14:24:47 there." - RP Aug 10 14:24:52 03koen 07org.oe.dev * rdae6d011... 10/ (1 packages/linux/handhelds-pxa-2.6/h2200/defconfig): handhelds-pxa-2.6: update h2200 defconfig Aug 10 14:24:58 03nail 07org.oe.dev * rf51ee4ac... 10/ (1 packages/meta/slugos-native.bb): slugos-native: Fix libc6-dev deps Aug 10 14:25:03 03koen 07org.oe.dev * r1415f8fe... 10/ (1 packages/linux/handhelds-pxa-2.6/ipaq-pxa270/defconfig): handhelds-pxa-2.6: update hx4700 defconfig Aug 10 14:25:08 03nail 07org.oe.dev * re159af24... 10/ (1 packages/db/db_4.3.29.bb): db: Remove --enable-smallbuild and replace with separate --disable-* options, removing --disable-hash Aug 10 14:27:03 shit.. what to remove from collie... Aug 10 14:27:25 hrw|work: bluetooth, browser? Aug 10 14:27:33 games? Aug 10 14:27:46 no such things already Aug 10 14:27:52 hrw|work: strip the kernel modules? Aug 10 14:27:55 bt modules are in image anyway Aug 10 14:28:31 2.6M of modules... big Aug 10 14:28:47 wpa-supplicant will go out (use openssl) Aug 10 14:31:38 grrr.... Aug 10 14:31:55 ade|desk: nm.....will re-test.... Aug 10 14:33:55 stripping modules gave 1M... Aug 10 14:34:19 2640372 contra 1599208 Aug 10 14:34:30 but does they work? Aug 10 14:35:38 they will not Aug 10 14:40:41 http://lists.arm.linux.org.uk/pipermail/linux-arm-toolchain/2006-August/000502.html Aug 10 14:40:56 I just realised that the original mail is from the ptxdist author Aug 10 14:57:58 * hrw|work want -= operator Aug 10 15:07:00 hrw|work: in bb? Aug 10 15:07:11 yes Aug 10 15:07:18 collie suxx with its 16M flash Aug 10 15:08:11 oh yes Aug 10 15:09:07 hrw|work: would a bit of python/sed in -image.bb work? Aug 10 15:10:00 koen: thats an option. but what to remove.... Aug 10 15:10:56 hrw|work: everything except pacakage manager? Aug 10 15:11:11 which suxx Aug 10 15:11:16 I wanted to suggest altboot, but that kind of sucks without SD Aug 10 15:11:51 I know - will do opie-minimal image for it Aug 10 15:11:54 does someone have all accumulated all the collie-sd patches? Aug 10 15:11:54 no PIM at all Aug 10 15:12:15 if only the connect+ would work as expected Aug 10 15:12:21 that'd rule for collie Aug 10 15:12:45 koen: we should have had a COnnect+ SoC Aug 10 15:14:45 XorA: I've been thinking about a fundraiser to do our own SoC sometime Aug 10 15:15:25 koen: the connect+ forum once had tempting signs that someone had cracked the problem, but no code ever appeared :-( Aug 10 15:15:30 koen: OE Foundation need to be created first? Aug 10 15:15:39 hrw|work: yes Aug 10 15:15:48 which we can do at OEdem if we want Aug 10 15:16:17 regardless of that, we could use a clearly visible TODO list Aug 10 15:18:59 collie without PIM fits Aug 10 15:24:36 koen, what irc client do you use on mac? Aug 10 15:25:12 im using chatzilla but its a pain that i cant start/stop ff separately from chatzilla Aug 10 15:26:47 hey xora i noticed you worked on libetpan recently - what changes did you ci? Aug 10 15:27:49 poushag: inherit bnconfig Aug 10 15:28:19 is it something noticable to the user such that i would want to build and install it? (i dont know what bnconfig is) Aug 10 15:28:59 poushag: nope, its to allow other programs to -letpan Aug 10 15:30:13 hehe - ok you went from speaking greek to chinese ;) or maybe burmese since letpan is a place in burma... Aug 10 15:30:36 what does -letpan option do for other progs? Aug 10 15:31:05 poushag: if you dont know, you really dont care :-) Aug 10 15:32:04 well i know theres some kind of bug that keeps opie-mail from successfully using pw auth on smtp sending - just trying to confirm if this is unrelated Aug 10 15:32:44 poushag: well sylpheed-claws can, and that uses libetpan, so I guess a opie-mail bug Aug 10 15:33:56 i filed the bug on opie-mail but apparently its libetpan known bug (according to harlekin) Aug 10 15:34:15 poushag: remind me to retest sometime then Aug 10 15:34:40 do you have an opie image around? Aug 10 15:34:52 poushag: no, I mean test sylpheed-claws Aug 10 15:35:16 k Aug 10 15:35:36 if you have gmail addr its easy to test Aug 10 15:35:48 03koen 07org.oe.dev * r6be389aa... 10/ (1 packages/fakeroot/fakeroot_1.2.13.bb): fakeroot: add staging Aug 10 15:35:51 gmail has smtp server set up for pw auth Aug 10 15:35:51 poushag: I have smtp auth on my emal serrver Aug 10 15:36:09 oh ok Aug 10 15:36:19 anyway time for me to catch train home Aug 10 15:36:56 hmm..well, I just obsserved OE bug#1290 on my work coomputer but not on my home system Aug 10 15:36:56 both identical distros and freshly updated, as was OE.mtn Aug 10 15:36:56 II'm tempted to reopen it Aug 10 15:37:32 oh man......this ahtlon64 is *humming* Aug 10 15:37:40 I can't believe the speed difference Aug 10 15:37:49 * JustinP hugs his new build computer Aug 10 15:40:16 im kinda new at this but im trying to get 2 files from the kdepimpi cvs repos so i can try to patch stuff in oe - any guru here want to help me get the cvs command straight? Aug 10 15:40:18 hmmm Aug 10 15:40:35 I am building busy box and the c compiler is eating memeory .... Aug 10 15:41:53 hi wookey Aug 10 15:45:08 brb Aug 10 15:45:23 Crofton: we set the 'build all at once' param in OE for busybox Aug 10 15:45:43 you mean the take forever to build parameter? Aug 10 15:45:46 according to the busybox docs that will generate a more l33t version Aug 10 15:46:01 ok Aug 10 15:46:05 It is going again Aug 10 15:46:20 that took a while Aug 10 15:46:29 03koen 07org.oe.dev * r3b5717e8... 10/ (1 packages/scratchbox packages/scratchbox/sbrsh_7.1.bb): shrsh: add 7.1 Aug 10 15:47:26 I only have about 512K of phys ram in build bx Aug 10 15:47:40 that's not much Aug 10 15:47:47 half a meg Aug 10 15:47:54 My first computer had 128 bytes Aug 10 15:48:03 er 512M Aug 10 15:48:06 :) Aug 10 15:48:14 mine only has 512 too - took about 2 or 3 hrs to build busybox on mine Aug 10 15:48:49 http://en.wikipedia.org/wiki/MEK6800D2 Aug 10 15:49:57 http://www.computermuseumgroningen.nl/motorola/mek6800d2.html Aug 10 15:50:40 hi wookey Aug 10 15:51:00 Crofton: http://www.computermuseumgroningen.nl/motorola/mek6800d2dammit Aug 10 15:51:08 dammit :( Aug 10 15:51:11 oops - i thought you said 512mb - yeah i remember when a pc with no hdd and 16mb of ram was awesome and cost about $10k in 2006 $ Aug 10 15:51:50 (price adjusted for inflation) Aug 10 15:51:52 Crofton: aye, but we were happy then... :) Aug 10 15:51:57 heh Aug 10 15:51:58 yeah Aug 10 15:53:46 i agree - win-3.1 made lots of folks happy and every upgrade since then was a mix of some good stuff with a good deal of bad Aug 10 15:54:19 but now i switched to mac osx and have regained my digital happiness Aug 10 15:55:07 my roommate had a mac back then (cs major) but us engineers were only allowed to have ibms Aug 10 15:55:23 03koen 07org.oe.dev * re383b4da... 10/ (1 packages/scratchbox/sbrsh_7.1.bb): sbrsh: RRECOMMEND nfs module Aug 10 15:55:36 i had to beg to get him to let me play the games on there Aug 10 16:04:03 hrw, kergoth - im trying to see if i can find the code zautrix used to patch his kdepimpi mail prog - i have the source - but i dont knwo where to start looking - i do know that he made his own libs Aug 10 16:04:25 libs: /opt/QtPalmtop/lib/libmicrolibetpan.so.2.2.5 and /opt/QtPalmtop/lib/libmicromailwrapper.so.2.2.5 Aug 10 16:05:21 so im trying to find the code he wrote that changed these from whatever they were before - im pretty lost - any help? Aug 10 16:06:34 i found a tmake file that indicates he used arm-linux-g++ crosscompiler but not sure where to go from there Aug 10 16:06:50 inherit pkgconfig binconfig pkgconfig Aug 10 16:07:08 ?? was that for me woglinde? Aug 10 16:07:18 poushag nope Aug 10 16:07:26 thats from libpng.bb Aug 10 16:07:27 cr Aug 10 16:07:32 maintainer kergoth *g* Aug 10 16:07:49 hrw|work: we are getting closer to a new gpe-conf release... i found another issue with some bets which are too specific to Familiar. Aug 10 16:07:50 re Aug 10 16:07:56 florian_kc: ok Aug 10 16:08:07 florian_kc: oz 3.5.4.2-rc1 released with 0.2.1 Aug 10 16:08:18 poushag: 1. kergoth has nothing common with kdepimpi Aug 10 16:08:35 but i managed to fix the keys&buttons problem Aug 10 16:08:39 i know that - but hes listed as an opie-developer Aug 10 16:08:50 gpe-screenshot needs to be fixed in oe too Aug 10 16:08:52 thought he might have interest in fixing opie-mail Aug 10 16:09:06 poushag: 2. libetpan<>libmicroetpan changes can be interesting, 3. libmailwrapper (OPIE) <> libmicromailwrapper probably are important Aug 10 16:09:25 poushag: kergoth is also OZ, OE, Bitbake, tslib and many other projects developer Aug 10 16:09:40 poushag: then ompi<>opie-mail things Aug 10 16:10:08 poushag: first check WHEN ompi was created and on which opie-mail version it was based (same for libs) Aug 10 16:10:11 poushag: lot of work Aug 10 16:10:38 how much a lot? days or weeks? Aug 10 16:11:18 hours, days Aug 10 16:11:24 sok Aug 10 16:11:38 JustinP: sorry was afk, will read back through the log Aug 10 16:12:03 i dont know how to isolate the source used for building those two libs though Aug 10 16:12:17 but i think im geting closer Aug 10 16:13:40 JustinP: XUnsetICFocus Aug 10 16:13:40 XGetIMValues Aug 10 16:13:40 XSetICValues Aug 10 16:13:40 XDisplayOfIM Aug 10 16:13:40 Xutf8TextPropertyToTextList Aug 10 16:13:40 XwcLookupString Aug 10 16:13:42 XSetICFocus Aug 10 16:13:44 XVaCreateNestedList Aug 10 16:13:46 are on my list Aug 10 16:13:58 grr that was not meant to happen Aug 10 16:14:04 sorry folks Aug 10 16:14:20 ade no prob Aug 10 16:17:15 looks like the source i need is in two subdirs: ./kdepim/kmicromail/libmailwrapper and ./kdepim/libetpan Aug 10 16:21:00 hrw, i found opiemail.h in src of zautrix - // CHANGED 2004-09-31 Lutz Rogowski - thats the first line Aug 10 16:22:08 and opiemail.cpp - // CHANGED 2004-08-06 Lutz Rogowski Aug 10 16:22:20 ~emulate zautrix Aug 10 16:22:24 this is exactly what the GPL is about. grab the stuff and run away Aug 10 16:22:44 mr_nice_slacker: bummer with the flash, ill reflash it tonight maybe Aug 10 16:23:32 so how do i figure out the ver of opie-mail upon which this is based? Aug 10 16:25:01 im thinking to start with diff of libmailwrapper/smtpwrapper.h from opie-mail to ompi Aug 10 16:27:58 What package makes 'strip'? Should it be present in .../tmp/staging/i686-linux/bin? Aug 10 16:28:42 binutils(-cross) Aug 10 16:28:47 Kerwood: you hitting the qemu thing? Aug 10 16:29:34 I just remove the -s from the Makefiles.... Aug 10 16:29:37 looYes, because I have two systems at my fingertips, one with the bug, one without Aug 10 16:29:51 s/loo// Aug 10 16:30:04 always happens for me... Aug 10 16:32:01 I think OE expects it to be present on your host for -native packages Aug 10 16:33:25 03cbrake 07org.oe.dev * r113954a1... 10/ (1 packages/python/python-pysqlite2_2.2.2.bb): python-pysqlite2 2.2.2: add python-datetime to RDEPENDS Aug 10 16:36:25 koen: but it is Aug 10 16:36:43 /usr/bin/strip Aug 10 16:40:03 ok, so if 'install' is present in the tmp/staging/i686-linux/bin, the staging of qemu-native fails on 'cannot run strip" Aug 10 16:40:03 but if it's missing, staging uses /usr/bin/install, which finds runs strip with no problem Aug 10 16:40:03 so user should clean and rebuild binutils to fix? Aug 10 16:41:44 hmm...it was a recent build which put the 'install' there Aug 10 16:46:00 and deleting (or renaming) tmp/staging/i686-linux/bin/install allows the build to complete Aug 10 16:54:07 good to know Aug 10 16:55:42 so staging install is broken somehow.... Aug 10 16:56:04 the staged install is 'special' Aug 10 16:56:57 special how? Aug 10 16:59:38 it has to be special, since it's causing a shitload of problems that nobody is caring to fix Aug 10 16:59:49 ah Aug 10 17:02:05 I see... well, 'bitbake binutils-cross -c clean' had no effect on the tmp/staging/i686-linux/bin directory Aug 10 17:02:05 so I guess stomping on the rogue 'install' executable will have to do for now Aug 10 17:09:05 time to finally go Aug 10 17:09:42 bbiab Aug 10 17:12:51 kerwood: clean only affects tmp/work Aug 10 17:13:01 kerwood: clean only affects tmp/work Aug 10 17:13:09 unless you're using packaged-staging Aug 10 17:13:38 Kerwood: can you paste your run.do_install script somewhere? Aug 10 17:14:51 koen: right, that came to me after my brain engaged... :P Aug 10 17:15:37 pb__: I can in a few minutes Aug 10 17:17:07 koen: and how is packaged-staging these days? Aug 10 17:17:28 JustinP: nearly finished, I'm writing the docs right now Aug 10 17:19:11 koen: very cool. :-) Aug 10 17:21:07 hi andersee Aug 10 17:21:32 woglinde: morning Aug 10 17:21:37 hehe Aug 10 17:21:42 here its evening Aug 10 17:22:08 woglinde: still morning here for another 15 min Aug 10 17:23:02 * koen heats up the laminator Aug 10 17:23:22 hu? Aug 10 17:23:24 ~ugt Aug 10 17:23:25 well, ugt is Universial Greeting Time. Created in #mipslinux, it is a rule that states that whenever somebody enters an IRC channel it is always morning, and it is always late when the person leaves. The local time of any other people in the channel, including the greeter, is irrelevant. Aug 10 17:23:32 new badget? Aug 10 17:23:33 +s Aug 10 17:23:39 ibot: ugt also is http://www.total-knowledge.com/~ilya/mips/ugt.html Aug 10 17:23:42 koen: what are you talking about? Aug 10 17:23:53 ibot: ugt is also http://www.total-knowledge.com/~ilya/mips/ugt.html Aug 10 17:23:55 koen: okay Aug 10 17:23:58 ~ugt Aug 10 17:24:00 i heard ugt is Universial Greeting Time. Created in #mipslinux, it is a rule that states that whenever somebody enters an IRC channel it is always morning, and it is always late when the person leaves. The local time of any other people in the channel, including the greeter, is irrelevant. http://www.total-knowledge.com/~ilya/mips/ugt.html Aug 10 17:25:29 brb Aug 10 17:30:35 can some please paste the contents of his $staging/$arch/share/pkgconfig/libpng.pc and staging/$arch/lib/libpng.la? Aug 10 17:30:44 pb__: Well, I guess I can't show you the log.do_install because it's empty Aug 10 17:30:44 pb__: and the log.do_stage which had the record of the 'install' failure is gone because I did a 'bb -c clean' and a rebuild to make sure the workaround "worked" :( Aug 10 17:32:55 woglinde: comin' right up Aug 10 17:32:56 run.do_install, not log.do_install Aug 10 17:33:19 or run.do_stage, if that's where the error showed up Aug 10 17:33:21 woglinde: http://ewi546.ewi.utwente.nl/tmp/libpng.pc Aug 10 17:33:57 woglinde: and I don't have an .la Aug 10 17:37:21 pb__: shall I first let do_configure qemu-native with the offensive 'install'? Aug 10 17:37:26 s/let// Aug 10 17:37:26 yes, please Aug 10 17:37:50 bbl Aug 10 17:41:51 mr_nice_slacker: ping Aug 10 17:42:52 koen hm that cant be Aug 10 17:43:08 mr_nice_slacker: could you pull an image off your 16mb flash via jtag?takes a shitload of time, so maybe over night?wonna try that Aug 10 17:43:25 pb__: btw, flashing the hh.org bootldr directly worked just fine, but it wont boot Aug 10 17:43:29 actually wont even turn on Aug 10 17:43:41 x29a: oh dear Aug 10 17:44:02 he nice inherit autotools does the trick for libpng Aug 10 17:45:18 pb__: any other hints/tipps? Aug 10 17:49:28 sorry, nothing I can think of Aug 10 17:49:39 * pb__ going home now Aug 10 17:50:32 pb__: http://rafb.net/paste/results/p4dVam63.html Aug 10 17:51:24 pb__: sorry for my slow computer 9_9 Aug 10 17:55:07 heh Aug 10 17:55:10 well, that all looks fine Aug 10 17:55:28 maybe you could try running the offending "install" command under strace to see what it's doing. Aug 10 17:55:45 if you email the output to the list, or file a bug or something, I'll look at it in the morning Aug 10 17:55:50 going home now, really Aug 10 17:55:51 later all Aug 10 17:56:10 ciao pb Aug 10 18:08:27 stupid ipkg Aug 10 18:09:03 it tried to upgrade libgcc, removed the old version, but then of course the next program does not run Aug 10 18:09:07 ~lart ipkg Aug 10 18:09:08 * ibot takes a large goose feather pillow and swings it wildly in ipkg's direction, hitting ipkg and sending ipkg flying into the closet Aug 10 18:09:13 again! Aug 10 18:09:15 again! Aug 10 18:09:18 ~lart ipkg Aug 10 18:09:18 * ibot takes large quantities of Krispy Kream donuts and stuffs them one after another down ipkg's throat until ipkg puts on 150lbs Aug 10 18:09:31 ah, that'll teach him Aug 10 18:10:45 Sigh. ipkg is already quite big enough... Aug 10 18:16:46 mwester: it could be a lot smaller if someone taught it to link to libbb instead of building it's own copy Aug 10 18:17:48 I'll check to see if the Unslung rootfs has that lib -- if so, that could be a relatively easy way to reclaim some flash space. Thanks! Aug 10 18:18:49 mwester: busybox started to support a shared libbb starting from 1.2.0 Aug 10 18:19:05 for <1.2.0 there was a patch to patch ipkg into busybox Aug 10 18:20:03 Well, there it is then -- Unslung's slingbox version of busybox as well as ipkg should be built against a common libbb. Aug 10 18:20:16 * mwester goes to write down another to-do for Unslung 6.9... Aug 10 18:26:55 koen could you please test http://page.inf.fu-berlin.de/~heinold/libpng.diff Aug 10 18:27:04 and look if it works on a device too Aug 10 18:27:47 woglinde: right now all my devices I have buildtrees for are off Aug 10 18:28:17 someone else? Aug 10 18:29:34 woglinde: I wouldn't know what to test... Aug 10 18:30:46 justinp apply the patch rebuild libpng and pug the ipk on the device Aug 10 18:31:15 woglinde: and wee if what breaks? Aug 10 18:31:16 see Aug 10 18:31:35 justino see that it breaks nothing Aug 10 18:31:43 thats what I hope Aug 10 18:32:01 as in breaks what....? Aug 10 18:32:12 hm Aug 10 18:32:14 right Aug 10 18:32:16 mom Aug 10 18:32:21 some package thats use libpng Aug 10 18:32:31 is this for .dev or oz354x? Aug 10 18:32:38 for dev Aug 10 18:32:49 should work in .oz as well Aug 10 18:33:07 hm cairo need libpng Aug 10 18:33:17 qte too Aug 10 18:35:50 .oz354x doesn't have 1.2.12 Aug 10 18:37:47 mr_nice_slacker: flashed simpadSL.rom back, booted, have a serialsignal again, but now a lot of errors Aug 10 18:40:55 djeez, I suck at configing postgress Aug 10 18:41:35 koen what? Aug 10 18:41:49 hmmm Aug 10 18:42:00 it could be that perl hates the password I entered Aug 10 18:45:05 ah yes, it does Aug 10 18:52:50 how come that my libs also produce a .dbg file? this ends up in the rootfs as well Aug 10 18:53:34 eFfeM: perhaps you're not including the default -dbg package? Aug 10 18:55:39 JustinP: can you be more specific? I didn't get those .dbg files before and I don't know about -dbg Aug 10 18:56:37 eFfeM: I responded to your email to nslu2-developers; I had to "hack" the debug files out of the rootfs image for Unslung manually. Aug 10 18:56:50 eFfeM: it's something new, not sure what it is myself. But I noticed it with my new emacs builds Aug 10 18:57:04 eFfeM: add ${PN}-dbg to your PACKAGES Aug 10 18:57:30 they seem to normally go into -dbg packages for most packages for me... Aug 10 18:57:41 mwester, didn't check nslu2-developers the last hr or so Aug 10 18:57:47 :) Aug 10 18:57:51 JustingP tnx Aug 10 18:57:53 eFfeM: you work at philips? Aug 10 18:57:56 yes Aug 10 18:58:02 which department? Aug 10 18:58:03 how did you guess Aug 10 18:58:08 CE Aug 10 18:58:08 zecke told me Aug 10 18:58:10 ah Aug 10 18:58:11 okay Aug 10 18:58:14 i'm at research Aug 10 18:58:20 connected planet (if that rings a bell) Aug 10 18:58:25 ofcourse it does :) Aug 10 18:58:28 conf/bitbake.conf:PACKAGES = "${PN}-dbg ${PN} ${PN}-doc ${PN}-dev ${PN}-locale" Aug 10 18:58:29 conf/bitbake.conf:FILES_${PN}-dbg = "${bindir}/.debug ${sbindir}/.debug ${libexecdir}/.debug ${libdir}/.debug \ Aug 10 18:58:35 research ehv ? Aug 10 18:58:38 ye Aug 10 18:58:39 s Aug 10 18:58:44 ok Aug 10 18:58:53 you do the streamium stuff? Aug 10 18:58:56 chat with you later, heading back to the movie i was watching Aug 10 18:59:00 okay Aug 10 18:59:01 streamium yes Aug 10 18:59:05 cool Aug 10 18:59:07 ttyl Aug 10 18:59:15 later (or drop me an email) Aug 10 18:59:33 k Aug 10 19:01:59 hey mallum Aug 10 19:02:16 hey koen Aug 10 19:03:14 hey zecke Aug 10 19:03:23 hi Aug 10 19:03:29 hi Aug 10 19:03:33 koen: when will you be offline? Aug 10 19:03:34 re hrw Aug 10 19:03:36 he zecke Aug 10 19:03:45 zecke: 15-8 till 21-8 Aug 10 19:03:55 I just found that collie doe snot have tslib config for 2.6 and I shipped 3.5.4.2-rc1 images... Aug 10 19:04:02 :D Aug 10 19:04:19 hrw: touchscreens are fluff Aug 10 19:04:27 koen: progress on writing things up? Aug 10 19:04:35 zecke: yes Aug 10 19:08:52 aaah Aug 10 19:08:55 hey zecke Aug 10 19:09:01 postgress only runs on a socket Aug 10 19:09:15 koen: or port Aug 10 19:09:32 koen: 5432 iirc Aug 10 19:09:38 hrw: yes, but it isn't allowing connections right now Aug 10 19:10:04 osx leopard is cool - looks like i wont need quicksilver anywmore - and i wont have to keep my cobbled together DAV/ics calendar server Aug 10 19:11:08 and the app integration of time machine looks really great - no more retrospect for me Aug 10 19:12:56 man the mac pro is sweet too - will be about 10x faster than my current dual g4 - of course i actually need to earn some real $ so i can buy the hw (and then use the sw) Aug 10 19:13:25 poushag: really will be faster or marketoids told that? Aug 10 19:13:41 it's intel Aug 10 19:13:43 03hrw 07org.oe.oz354x * rbd53b6ad... 10/ (1 conf/distro/openzaurus-3.5.4.2.conf): oz 3.5.4.2: collie will get 2.6 image too - it will suxx but will show how 2.6 works on this machine Aug 10 19:13:48 03hrw 07org.oe.oz354x * rc44b5430... 10/ (1 packages/meta/opie-image.bb): opie-image: do not install extra stuff on tosa & poodle as this nearly fills rootfs Aug 10 19:13:49 so you can run specFP real fast Aug 10 19:13:52 03hrw 07org.oe.oz354x * rd7068be0... 10/ (6 files in 3 dirs): tslib: copied 2.4/2.6 switch from poodle for collie - have to test on device Aug 10 19:14:19 koen: thats just benchmark... Aug 10 19:14:29 that's my point Aug 10 19:14:35 :) Aug 10 19:14:52 really faster - measured in apps such as Xcode, final cut pro, etc Aug 10 19:14:56 anyway my plans for changing machines is: amd64 for builds (headless) + cheap laptop Aug 10 19:15:02 and building oe! Aug 10 19:16:08 hrw: I have a similar setup Aug 10 19:16:29 koen: I do not have - only plans Aug 10 19:16:41 cu Aug 10 19:16:53 k later hrw Aug 10 19:17:27 zecke can you talk about opie-mail vs ompi with me now? Aug 10 19:18:37 well ping me when you can pls Aug 10 19:20:07 if I want to do a fresh build of all the packages built by a specific package is there an easy way to do a clean with bitbake that cleans the package and its dependancies? Aug 10 19:20:38 and the dependancies dependancies.... Aug 10 19:21:43 hmmm Aug 10 19:21:55 koen, et al., FWIW, the troublesome 'install' executable was put there during staging of coreutils-native-5.1.3-r7 Aug 10 19:21:57 I think I'm running into the firewall Aug 10 19:22:09 Kerwood: I know Aug 10 19:22:13 ok Aug 10 19:30:39 can I go ahead and check in the bb file I did that builds u-boot from a git tag? Aug 10 19:30:51 this builds for the omap5912osk machine Aug 10 19:31:05 Crofton: sure Aug 10 19:31:28 I need to get up the nerve to actually flash it into my OSK Aug 10 19:31:37 guess I hould remember how to do that first :) Aug 10 19:32:04 it also builds with gcc-4.1.1 Aug 10 19:36:00 step one, clean off desk to find OSK Aug 10 19:44:23 when oe meeting is berlin is planned? Aug 10 20:05:56 good evening/afternoon ! Aug 10 20:08:15 darkschneider from 5-8.10 Aug 10 20:08:23 reminds me to write the email Aug 10 20:08:27 to the dekan Aug 10 20:08:30 JustinP: finally came to add ${PN}-dbg to PACKAGES, should I also add ${PN}-doc ${PN}-dev ${PN}-locale ? Aug 10 20:08:30 woglinde: thanks Aug 10 20:08:47 woglinde: mmm is it open to users? Aug 10 20:09:03 darkschneider hm Aug 10 20:09:23 woglinde: look like a no :) Aug 10 20:09:23 dont know, but personal I didnt have problems with that Aug 10 20:09:54 if you hapen to hear any opinion about that, let me know, considering i will be in bruxelles byt hat time, berlin is not far Aug 10 20:25:35 what is an easy yet fast way to regenerate the rootfs (without nuking all of tmp or so) Aug 10 20:26:08 bitbake -b foo-image.bb Aug 10 20:28:00 tnx Aug 10 20:28:41 still trying to get rid of the bb files, not too sure if adding ${PN}-dbg to PACKAGES actually fixes this Aug 10 20:29:31 hm, doesn't seem to work, digging further Aug 10 20:29:54 it is the installing of the libs in the rootfs that I need (and that seems to be the bad guy) Aug 10 20:29:58 it should get added automagically Aug 10 20:30:13 only apps that frob PACKAGES or FILES are prone to errors Aug 10 20:30:19 eFfeM: perhaps you're missing the FILES setting for it as well... Aug 10 20:30:40 JustinP: there is no FILES setting if I recall correctly Aug 10 20:30:41 * eFfeM checks Aug 10 20:31:32 there is no FILES in slugos-packages.bb Aug 10 20:31:44 PACKAGES now reads PACKAGES = "${PN}-dbg ${PN}" Aug 10 20:32:00 I get dbg ipk's already before making this change Aug 10 20:32:14 issue is that apparently they end up in the rootfs which I do not want Aug 10 20:32:38 eFfeM: that's not the issue Aug 10 20:32:48 is FILES inherited from bitbake.conf if not installed Aug 10 20:32:49 eFfeM: some package frob PACKAGES and FILES Aug 10 20:33:00 and unsurpisingly need to get adjusted Aug 10 20:33:01 frob ? Aug 10 20:33:07 touch/poke/etc Aug 10 20:33:42 eFfeM: you don't need that setting in slugos-packages.bb I think Aug 10 20:33:46 tweak Aug 10 20:33:58 eFfeM: likely some actual packages need to have their PACKAGES lines changed Aug 10 20:34:06 ah, ok Aug 10 20:34:21 which lib is installing debud files? Aug 10 20:34:24 debug? Aug 10 20:34:44 eFfeM: did you get the mail I sent you tonight? Aug 10 20:34:53 or did gmane screw up? Aug 10 20:35:22 to my yahoo account? no Aug 10 20:35:31 just checked again, no Aug 10 20:36:12 "http://bugs.openembedded.org/show_bug.cgi?id=1211 Aug 10 20:36:12 But RP changed it to use a .debug/ subdir a while ago, so most of .dbg Aug 10 20:36:12 file should be gone." Aug 10 20:36:20 that's the important part Aug 10 20:36:29 JustinP: every lib file in rootfs/lib seems to have the .dbg file Aug 10 20:37:30 eFfeM: try rebuilding one of the packages and rebuilding the rootfs and see if it goes away for that lib Aug 10 20:38:23 JustinP: just to be sure: with or without adding ${PN}-dbg to PACKAGES in slugos-packages.bb ? Aug 10 20:38:56 eFfeM: remove it from your slugos thing Aug 10 20:39:11 the complete packages line Aug 10 20:39:15 ok will do Aug 10 20:44:37 building, meanwhile searched for dbg in the rootfs, it is there for /bin/devio, /usr/lib/libipkg and most libs in /lib (which all seem to come from glibc, eg resolv) Aug 10 20:45:09 eFfeM: .dbg files have ceased to exists since a few days Aug 10 20:45:26 eFfeM: you should look for .debug/ dirs Aug 10 20:46:31 koen, just rebuild ipkg and then the rootfs and indeed the .dbg file for it is gone and there is a .debug dir Aug 10 20:46:36 eFfeM: http://www.openembedded.org/viewmtn/revision.psp?id=d7cc6ec7ff432b7592af0d14fbe1777a3bebd543 Aug 10 20:46:54 I've removed the PACKAGES line from slugos-packages.bb Aug 10 20:48:11 koen, I feel I now have only moved the problem, this .debug dir is in my rootfs, and I expect that it will also show up in my rootfs.jffs2 right? Aug 10 20:50:13 yes Aug 10 20:50:36 please add bus for those packages to #1211 Aug 10 20:50:40 (or fix them :)) Aug 10 20:50:49 and actually what I wanted is to get rid of these to save memory Aug 10 20:51:58 i'd rather fix, if I know how Aug 10 20:52:46 you need to fix the FILES_foo of the affected .bb files Aug 10 20:55:04 add a FILES_foo_dbg or so ? Aug 10 20:55:43 if needed Aug 10 20:56:08 but I think most recipes need a less broad FILES_${PN} Aug 10 20:56:21 the problem is that they include too much Aug 10 20:56:36 like FILES = "/bin/" instead of "/bin/*" Aug 10 20:56:42 the bad guys I see are ipkg, devio and glibc Aug 10 20:57:24 well, as long as FILES_foo-dbg picks all the debug files it shouldn't matter since foo-dbg comes before foo in PACKAGES (by default) Aug 10 20:57:40 but yeah, perhaps /bin/* would be better Aug 10 20:57:52 there is no FILES in packages/devio/* Aug 10 20:58:08 ipkg says: Aug 10 20:58:08 ipkg.inc:FILES_libipkg-dev = "${libdir}/*.a ${libdir}/*.la ${libdir}/*.so" Aug 10 20:58:08 ipkg.inc:FILES_libipkg = "${libdir}" Aug 10 20:58:22 the last line is the bad guy I assume Aug 10 20:58:29 looks like it Aug 10 20:58:45 that should be ${libdir}/*.so.* Aug 10 20:59:38 changed it Aug 10 21:00:05 yuk, some of the ipkg bb files do not even have a ${PR} Aug 10 21:01:15 eFfeM: PR is optional Aug 10 21:01:30 if it's missing it defaults to r0 Aug 10 21:01:41 yeah I know, but I think it is good practice to have it Aug 10 21:02:01 wanted to bump PR for all ipkg bb files, would that be ok (the change I made is in an .inc file) Aug 10 21:02:12 that should be ok Aug 10 21:02:36 zecke and I usally bump PR on the toolchains, which is a bit more annoying :) Aug 10 21:02:45 anyway, night all Aug 10 21:08:40 ok, fixed and pushed ipkg, for the others it is in glibc but I think these are ok after the move from *.dbg to .debug, need a rebuild to verify that Aug 10 21:08:45 tomorrow or so ... Aug 10 21:08:51 nite Aug 10 21:28:20 I'm new to the list and I use OE for our boards at work. Is this where I could ask questions ? Aug 10 21:28:32 yep Aug 10 21:29:06 Ok, our processor is s3c2410 ie. like h1940 Aug 10 21:30:17 It seems when I try to use gcc-4.1.1 with the .dev branch (generic-unstable.conf or generic.conf) I hang in Freeing init memory Aug 10 21:30:55 thats strange Aug 10 21:31:48 Wondering if anyone has actually built images from scratch for h1940 Aug 10 21:32:13 nite Aug 10 21:32:33 I not using the same kernel version as h1940, but I could try to confirm Aug 10 21:33:19 If I use the .oz354x branch, no problems, for a full gpe-image Aug 10 21:47:05 w00t, my git based u-boot boots my OSK Aug 10 22:08:45 re Aug 10 22:13:07 Hi! Aug 10 22:13:33 I've got a couple of questions about OZ on the spitz Aug 10 22:13:58 does it support 4GB SD cards? Aug 10 22:15:27 that depends more on the hardware than OZ Aug 10 22:15:32 and if I create a FAT32 partition on my microdrive, can I get it mounted as a USB drive via a cable? Aug 10 22:15:33 I wouldn't count on it Aug 10 22:15:42 but some work? Aug 10 22:16:01 my 4GB SD card works under cacko fine Aug 10 22:16:01 I have never heard of anyone getting more than a 2GB working Aug 10 22:16:08 then it should work fine in OZ Aug 10 22:16:18 good! Aug 10 22:16:19 but we use a completely different kernel Aug 10 22:16:24 if it doesn't work file a bug report Aug 10 22:16:27 thats why i ask Aug 10 22:17:07 I just commited a new u-boot bb file to my repo Aug 10 22:17:44 and there is a remark that this revision creates divergence, you may (or may not) wish to run mtn merge Aug 10 22:17:48 advice? Aug 10 22:17:49 justinp: what about the usb link? Aug 10 22:18:48 pdaxrom and cacko have a little gui utility to set the usb link setting- how about OZ/GPE? Aug 10 22:19:08 I've never got USB networking to work Aug 10 22:19:25 under any of them Aug 10 22:19:43 with windows or linux Aug 10 22:19:59 usb networking worked with 2.4 for me but sometimes required rebooting Aug 10 22:20:09 usb networking works fine with a linux box with a newer kernel Aug 10 22:20:14 usb networking does not work with windows Aug 10 22:20:22 I haven't tried the usb storage thing in a while Aug 10 22:20:36 Crofton: did you not update before comitting? Aug 10 22:20:43 not bothered about that though, i want the usb drive cable link to work easily like under sharp rom Aug 10 22:20:47 Crofton: try mtn merge Aug 10 22:21:08 I pulled, forgot to update Aug 10 22:21:18 I merged and am down to one head :) Aug 10 22:21:33 my commit was the addtion of a file Aug 10 22:22:03 dmd: well, I don't know if it's working Aug 10 22:22:32 anybody here tried usb storage usb link under oz/gpe 3.5.4.1? Aug 10 22:23:01 -1 usb there ;) Aug 10 22:26:37 03crofton 07org.oe.dev * r0fee5c6e... 10/ (1 packages/uboot/u-boot_LABEL.2006.06.30.2020.bb): Aug 10 22:26:37 Add u-boot built from git repository. Source pulled from tag LABEL.2006.06.30.2020. Aug 10 22:26:37 This version compiles with gcc4 and has some compilation fixes needed Aug 10 22:26:37 to build u-boot for the omap5912osk board. Aug 11 01:41:15 03eFfeM 07org.oe.dev * rd18d8a8b... 10/ (1 packages/glibc/glibc-package.bbclass): glibc-package.bbclass: added libc6 to PACKAGES_DYNAMIC Aug 11 01:41:20 03eFfeM 07org.oe.dev * r51b8951a... 10/ (1 packages/meta/slugos-packages.bb): Aug 11 01:41:20 slugos-packages.bb: removed PACKAGES line, Aug 11 01:41:20 should not be needed according to the wizards in #oe Aug 11 01:41:25 03eFfeM 07org.oe.dev * ra4130542... 10/ (14 files in 2 dirs): Aug 11 01:41:25 ipkg: refined FILES spec to exclude debug info in package Aug 11 01:41:25 bumped PR for all bb files as the change is in a .inc file **** ENDING LOGGING AT Fri Aug 11 02:59:58 2006