**** BEGIN LOGGING AT Fri Sep 22 02:59:56 2006 Sep 22 03:16:24 hey . folks. Sep 22 03:16:28 hey psokolovsky Sep 22 03:18:05 hmm, how ENABLE_BINARY_LOCALE_GENERATION works ? Sep 22 03:36:44 it enables qemu to build the locales Sep 22 03:36:52 or thats all it used to do Sep 22 03:39:12 emte, so . which glibc version do you preferred ? Sep 22 03:40:03 emte, so. if ENABLE_BINARY_LOCALE_GENERATION=0 ? how it generate binary locale ? natively ? Sep 22 03:40:39 and what is the difference between glibc-intermediate and glibc ? Sep 22 03:42:24 using qemu means its emulated durring compile Sep 22 03:42:40 not using it means you need an intermediary compile Sep 22 03:46:25 03tmbinc 07org.oe.dreambox * r71b18e03... 10/ (3 files in 3 dirs): busybox-1.2.1: re-add hdparm -M patch Sep 22 03:46:30 03tmbinc 07org.oe.dreambox * rddee9f13... 10/ (1 packages/busybox/busybox-1.2.1/opendreambox/defconfig): busybox-1.2.1/opendreambox: fix configuration Sep 22 03:46:35 03tmbinc 07org.oe.dreambox * r71b2c9c6... 10/ (5 files in 5 dirs): initscripts/opendreambox: fix hdparm calls, fix /var/tmp Sep 22 03:46:45 03tmbinc 07org.oe.dreambox * raa39ac7b... 10/ (5 files in 2 dirs): twisted: update to 2.4.0 Sep 22 03:46:48 03tmbinc 07org.oe.dreambox * rcf1f48ec... 10/ (1 packages/twisted/twisted-web2_0.2.0.bb): twisted-web2: update to 0.2.0 Sep 22 03:47:18 emte, hmm, I found my opie somewhere. maybe the binary locale is something wrong . Sep 22 03:47:39 your comment makes no sense Sep 22 03:47:50 emte, and my gdb is either crash .. :( Sep 22 03:48:19 emte, so. do you familiar with opie ? Sep 22 03:50:13 * emte looks for a sentance sorter for leoncamel so he makes sense Sep 22 03:57:03 hi Sep 22 03:57:25 emte, http://pastebin.ca/179353 Sep 22 03:57:34 emte, so, any suggestion ? Sep 22 03:57:49 somebody could give a help Sep 22 03:58:09 i hsving problems with boot on omap Sep 22 03:58:25 it boots but do not open the console Sep 22 03:59:34 lvcargnini, what do you mean "do not open the console" exactly ? Sep 22 04:00:01 leoncamel, this OMAP GPIO switch handler initializing Sep 22 04:00:01 VFS: Mounted root (jffs2 filesystem). Sep 22 04:00:01 Freeing init memory: 116K Sep 22 04:00:16 it's the last lines of my boot Sep 22 04:00:35 next step should be appear the login: Sep 22 04:02:09 lvcargnini, whic kernel do you use ? 2.4 series ? or 2.6 ? Sep 22 04:03:09 2.6 Sep 22 04:03:14 leoncamel, http://pastebin.ca/179355 Sep 22 04:03:21 the full boot process Sep 22 04:04:26 2.6.18-rc7 is the latest git tree official from kernel.org, the main omap git treee Sep 22 04:05:06 lvcargnini, OK. I think the kernel is OK. Sep 22 04:05:17 lvcargnini, how do you build the rootfs ? Sep 22 04:05:36 i used the openembedded bitbak bootstrap-image Sep 22 04:05:48 lvcargnini, and which distro ? Sep 22 04:05:52 it created almost everything Sep 22 04:06:13 distro ???? of bitbake or openembedded ? Sep 22 04:06:20 bitbake Sep 22 04:06:28 leoncamel, 1.6 Sep 22 04:06:32 lvcargnini, the distro in you local.conf Sep 22 04:06:45 lvcargnini, I mean the DISTRO option in your local.conf Sep 22 04:07:10 leoncamel, openomap Sep 22 04:07:42 leoncamel, MACHINE = omap5912osk Sep 22 04:08:11 lvcargnini, I never use that distro before. Sep 22 04:08:32 lvcargnini, so, you should mount your rootfs and read through the boot script. Sep 22 04:09:18 leoncamel, are you still not using a stable distro? Sep 22 04:09:37 leoncamel, the boot script ??? Sep 22 04:09:45 leoncamel, inittab ? Sep 22 04:09:49 lvcargnini, sure. Sep 22 04:10:30 emte, so. add ENABLE_BINARY_LOCALE_GENERATION=1 into my distro/xxx.conf Sep 22 04:11:40 leoncamel, you still have not stated how that relates to anything Sep 22 04:11:51 leoncamel, my inittab http://pastebin.ca/179360 Sep 22 04:12:58 emte, yes. I am also don't know what happened there. So, could you give me some advice on debugging about it ? Sep 22 04:14:44 http://pastebin.ca/179364 Sep 22 04:15:56 emte, what ? Sep 22 04:20:02 emte, so, "using default" would cause this crash ? Sep 22 04:23:30 leoncamel, do you think if my inittab is correct ? Sep 22 04:24:04 lvcargnini, the inittab is OK, IIRC. do you got /dev/console ? Sep 22 04:25:27 no this is the case it no open the console, how could i know if i got the /dev/console ?? Sep 22 04:25:37 leoncamel, i'm using udev Sep 22 04:25:47 leoncamel, until why i expect Sep 22 04:25:59 leoncamel, excuse Sep 22 04:26:11 until i know I'm using udev Sep 22 04:26:18 lvcargnini, hmm, I have no idea about it . Sep 22 04:26:57 leoncamel, thi is the case it boot until that step and it's all Sep 22 04:27:20 do not show "login:" Sep 22 04:27:53 my boot args is bootargs=console=ttyS0,115200n8 noinitrd rw rootfstype=jffs2 root=/dev/mtdblock3 mem=32M udev Sep 22 04:43:42 in bitbake, what is a "providee" ?? Sep 22 04:45:34 I'm trying to modify the source in tmp/work/ to make some changes so I can generate a patch file against that source. Is it possible to do that with bitbake, or, do I need to copy the source out to someplace else and use my other ARM devtools to do this? Sep 22 04:47:52 nm, run.do_compile Sep 22 07:09:25 bonjour Sep 22 07:29:05 morning Sep 22 07:32:16 morning all Sep 22 07:50:34 03blaster8 07org.oe.dev * r1a44d9f2... 10/ (3 files in 2 dirs): Sep 22 07:50:34 ixp4xx-npe: quick refresh Sep 22 07:50:34 * update install location of npe microcode Sep 22 07:50:34 * remove obsolete version Sep 22 07:50:46 03blaster8 07org.oe.dev * r1ef7e13b... 10/ (1 packages/linux/ixp4xx-kernel_2.6.18.bb): Sep 22 07:50:46 ixp4xx-kernel_2.6.18: General Update Sep 22 07:50:46 * Patches updated to reflect NSLU2-Linux SVN repository Sep 22 08:05:56 http://www.hrw.one.pl/2006/09/21/one-step-closer-to-oedem/ Sep 22 08:14:08 morning Sep 22 08:14:13 will buy tickets today as well Sep 22 08:14:27 Tomas and myself have flights and also have a hotel - not the Berliner Baer though :-/ Sep 22 08:14:36 ah bummer. was it full? Sep 22 08:14:54 no twin rooms and singles were too expensive Sep 22 08:15:00 T0mW: a providee is an entity that provides something. Sep 22 08:15:08 T0mW: to manage patches I recommend quilt Sep 22 08:15:12 *nod* Where are you now? Sep 22 08:15:29 mickeyl: as in a task-image ? Sep 22 08:15:46 T0mW: yes. example: Sep 22 08:15:49 mickeyl: or a .bb file would be a providee? Sep 22 08:16:00 foo_1.2.3.bb Sep 22 08:16:04 ah Sep 22 08:16:05 automagically provides: Sep 22 08:16:05 foo Sep 22 08:16:07 foo-1.2.3 Sep 22 08:16:14 if foo would have something like Sep 22 08:16:19 PROVIDES += virtual/superfoo Sep 22 08:16:23 then it would provide this one as well Sep 22 08:16:37 bitbake help screen mixes the term providee vs. bb file a lot, it is not very clean Sep 22 08:16:39 so a bb file usually is a providee for two or more things Sep 22 08:16:45 s/clean/clear Sep 22 08:17:12 i get confused about the terms all the time as well Sep 22 08:17:13 mickeyl: thanks, just putting the touches on the project, heh, the splash screen Sep 22 08:17:27 hi mickeyl Sep 22 08:17:29 perhaps it helps to think in terms of caller and callee Sep 22 08:17:33 mickeyl: "Berlin Aparthotel Residenz 2000" was the closest thing we could fine with spaces Sep 22 08:17:47 let me check how far this is Sep 22 08:18:03 mickeyl: Its ~6km Sep 22 08:18:11 (I think) Sep 22 08:18:47 sounds not too bad Sep 22 08:19:18 It should be ok :) Sep 22 08:20:19 ~lart citymaps Sep 22 08:20:20 * ibot makes a balloon animal out of citymaps Sep 22 08:20:59 here I can choose 3 Berlin citymaps - each end on Tempelhof Sep 22 08:21:48 heh Sep 22 08:22:55 probably will have to search for one in Berlin Sep 22 08:23:47 I collect citymaps as mementos from cities which I visited Sep 22 08:24:06 * mickeyl needs to pack all the demo and give-away PDAs for odem Sep 22 08:24:13 oedem, even Sep 22 08:24:13 hrw|work: I used the map on http://www.booking.com/city/de/berlin.html Sep 22 08:24:39 * RP wonders what air port security is going to be like with the laptop/handhelds... Sep 22 08:25:24 RP: german security is used to it. you just need to unpack everything into seperate of their containers Sep 22 08:25:33 RP: http://maps.google.com/maps?daddr=Im+Rosengrund,+12347+Berlin,+Germany&saddr=attilastr,%2Bberlin,%2Bgermany&f=li&hl=en&ie=UTF8&om=0&z=13&ll=52.476821,13.425636&spn=0.097341,0.199127 Sep 22 08:27:06 mickeyl: For a while, you weren't allowed hang luggage here which would have meant no laptop as computer equipment can't go in the hold. Thankfully they've relaxed a bit but things are still pretty tight :-/ Sep 22 08:27:37 oh right, of course Sep 22 08:27:39 you're coming from UK Sep 22 08:27:45 * mickeyl understands Sep 22 08:29:11 hrw|work: One location is the meeting, is the other your accomodation? Sep 22 08:30:03 Basically, our hotel is about the same distance in the opposite direction Sep 22 08:32:40 morning Sep 22 08:32:59 RP: other is Zecke's place (also my accomodation) Sep 22 08:33:01 hi XorA Sep 22 08:34:25 morning xora Sep 22 08:35:02 speaking about packing... I need to buy bag first (or will travel with big backpack) Sep 22 08:35:05 hrw|work: ah, ok. Its handy to know where that is :) Sep 22 08:35:12 morning XorA Sep 22 08:35:23 hrw|work: *nod* a good bag rocks Sep 22 08:35:41 two years ago I bought the largest one that is still allowed to go as handluggage via pane Sep 22 08:35:43 plane, even Sep 22 08:35:50 RP: http://maps.google.com/maps?f=q&hl=en&q=poznan,+poland&ie=UTF8&z=14&ll=52.375704,16.964607&spn=0.03584,0.10231&om=1 Sep 22 08:35:56 I like big backpacks. airlines don't though... Sep 22 08:36:00 *nod* Sep 22 08:36:12 RP: thats my place Sep 22 08:38:09 so OEDEM'07 in Frankfurt? Sep 22 08:38:17 or UK? :) Sep 22 08:38:32 mickeyl: Ania suggest in PoznaƄ, Poland Sep 22 08:38:38 hehe, i can imagine Sep 22 08:38:41 i guess it will be a good place Sep 22 08:38:44 i never was there Sep 22 08:38:48 * RP votes for Italy Sep 22 08:39:25 hrw|work: Its interesting look at maps and trying to imagine what it looks like. Invariably, you're totally wrong :) Sep 22 08:39:48 * mickeyl votes for Ibiza Sep 22 08:39:49 morning all Sep 22 08:39:52 but then we wouldn't be able to work Sep 22 08:39:53 I'm sure something could be arranged in the UK :) Sep 22 08:39:53 *cough* Sep 22 08:39:56 morning dirk Sep 22 08:40:00 hi Dirk Sep 22 08:40:00 hi Dirk Sep 22 08:40:15 hey mickeyl, Richard Sep 22 08:42:20 mickeyl: I do keep thinking about writing an OE book myself btw, I'd quite like to. It'd just be a question of time :-/ Sep 22 08:42:39 RP: that sounds like we would perfectly match as coauthors Sep 22 08:44:50 mickeyl: We should discuss it at oedem :) Sep 22 08:44:57 RP: absolutely Sep 22 08:45:13 I know mallum would like documentation for poky and I just hope there wouldn't be a conflict of interest :-/ Sep 22 08:46:14 RP: well, that should be ok. considering that my book is not an OE book Sep 22 08:46:23 it's a book about developing and building software for embedded linux Sep 22 08:46:32 OE and poky could well coexist in that book Sep 22 08:46:45 with the differences and similarities Sep 22 08:47:43 mickeyl: Interesting. Roughly what percentage of the material would you expect to be OE related and what else were you planning to cover? Sep 22 08:48:11 RP: give me a second, I'll put the current TOC for you online Sep 22 08:51:10 http://vanille.de/temp/toc.txt Sep 22 08:51:19 please skip all the german text which is the proposal for a german publisher Sep 22 08:52:04 so chapter 5 should be at least 35% of the book Sep 22 08:52:25 could just make a chapter about poky as well then Sep 22 08:53:29 A chapter about branching and creating your own customistions with poky as an example perhaps Sep 22 08:53:39 ah yeah! Sep 22 08:53:57 I must admit I had been thinking of more secific OE content but I realise a publisher would be unlikely to take that :) Sep 22 08:54:19 The TOC looks good - I think it will work Sep 22 08:54:38 cool Sep 22 08:55:24 as you say, i went for a broader scope because it should attract decision makers and folks starting to consider embedded linux as a foundation for a project as well Sep 22 08:57:04 *sigh* i get more and more spam daily Sep 22 08:57:08 ~lart spammers Sep 22 08:57:08 * ibot executes killall -TERM spammers Sep 22 08:58:24 hrw|work: your OZ kernel update, is that including the newer ASOC? Sep 22 08:58:40 XorA: did not planed it Sep 22 08:59:07 hrw|work: It compiles, but as I am having preblems getting my kernel to see my init I am having troubles testing Sep 22 08:59:51 * XorA apologises for bad typing, but for first time in months I am not on ergonomic keyboard Sep 22 09:04:18 Hi, all! Is it possible to force bitbake log build process output somewhere? Sep 22 09:05:03 slapin: look in WORKDIR/temp/ Sep 22 09:05:25 XorA: its pain to use 'normal' keyboard after long usage of ergo ones Sep 22 09:06:01 hrw|work: Thanks a lot! Sep 22 09:11:15 morning all Sep 22 09:13:40 hey liam Sep 22 09:13:42 morning koen Sep 22 09:13:59 hey mickeyl Sep 22 09:14:01 * koen is en route to a wedding Sep 22 09:14:26 morning Sep 22 09:16:03 bonne route ;) Sep 22 09:16:13 NOTE: package hackndev-2.6trunk-2.6.16-r0: task do_populate_staging: completed Sep 22 09:16:13 NOTE: package hackndev-2.6trunk-2.6.16-r0: task do_package: started Sep 22 09:16:13 NOTE: Task failed: Failed to extract module dependencies. Sep 22 09:16:13 NOTE: package hackndev-2.6trunk-2.6.16-r0: task do_package: failed Sep 22 09:16:13 ERROR: TaskFailed event exception, aborting Sep 22 09:16:14 NOTE: package hackndev-2.6trunk-2.6.16: failed Sep 22 09:16:16 ERROR: Build of gpe-image failed Sep 22 09:16:40 eeew Sep 22 09:16:55 someone needs to get hit with the package naming and versioning cluebat Sep 22 09:17:01 Where could I find logs about packaging errors? Sep 22 09:17:47 $tmp/work// hackndev/tmp Sep 22 09:17:55 something like that Sep 22 09:18:48 Thanks! (that's probably me, who have to be hit, anyway) Sep 22 09:24:25 lrg: fix your poodle problem PREFERRED_PROVIDER_dbus-glib = "dbus-glib" Sep 22 09:25:55 i've some pb with dbus too :/ Sep 22 09:27:27 lrg: I think the locale problem stems from the fact that we're now using pregenerated locales, and your build is older than the changeover Sep 22 09:27:35 lrg: but I'm not very sure about that Sep 22 09:28:04 later all Sep 22 09:41:55 lets see if this image/kernel combo boots Sep 22 09:53:23 nope, kernel cant find init, is there a change in jffs2 that means 2.6.16 kernel cant read images produced by .dev? Sep 22 09:54:40 03lrg 07org.oe.dev * r114e1056... 10/ (1 packages/linux/mainstone-kernel_2.6.18.bb): mainstone_kernel_2.6.18| virtual/kernel : added mainstone kernel 2.6.18 to support Angstrom and EABI Sep 22 10:53:08 * XorA now wonders what a .dev image is including as init as kernel sure doesnt like it Sep 22 11:04:49 interesting, gpe-image in .dev doesnt build busybox, or put it into rootfs Sep 22 11:05:46 at least that explains why init is missing Sep 22 11:06:06 mickeyl: did you used qt4? Sep 22 11:06:30 not on a Z yet Sep 22 11:06:59 mickeyl: I did not asked about Z :) Sep 22 11:07:05 aha :) Sep 22 11:07:18 yes, I had a quick glance at qt4. working with it would be exaggarated Sep 22 11:07:25 ~spell exaggarated Sep 22 11:07:26 I wonder about using qt4 + regexp + sqlite Sep 22 11:07:27 possible spellings for exaggarated: exaggerated exaggerates exaggerate exaggeratedly execrated exaggerator Sep 22 11:07:42 hrw|work: sounds like a kick-ass combo Sep 22 11:08:34 morning Sep 22 11:09:01 2.6.18 kernel UTS_RELEASE is in utsrelease.h instead of version.h - which breaks the definition of KERNEL_VERSION in kernel.bb ... Sep 22 11:09:22 See http://kerneltrap.org/node/7144/199809 (end of page) Sep 22 11:09:38 "(Issue is that UTS_RELEASE define has moved from include/linux/version.h to the (new) include/linux/utsrelease.h)" Sep 22 11:09:47 * lrg fingers are crossed for RP's patch 3847/1 Sep 22 11:10:34 Does anyone know enough python to be able to check for utsrelease.h and use that if it exists, else use version.h ? Sep 22 11:10:48 KERNEL_VERSION = "${@get_kernelversion('${S}/include/linux/version.h')}" Sep 22 11:11:08 lrg: Anything can happen with those patches. I worry about 3848/1 :) Sep 22 11:11:42 RP how is the openzaurus 2.6.18 kernel working after this 2.6.18 change? Sep 22 11:11:50 rwhitby: if os.path.isfile(somefile) Sep 22 11:12:20 rwhitby: It probably isn't - we just added it as a work in progress - I've not tested anything other than compiles with it Sep 22 11:12:46 well, I've saved you some debugging time then :-) Sep 22 11:13:07 rwhitby: Yes, thanks for the pointer - we'd no doubt hit the same problem :) Sep 22 11:13:19 (thank blaster8 - he found it) Sep 22 11:13:49 http://bugs.openembedded.org/show_bug.cgi?id=1431 Sep 22 11:13:49 blaster8 is the newest member of the nslu2-linux core team, working on integrating the new open source ixp4xx ethernet driver into the kernel Sep 22 11:14:04 probably make the get_kernelversion accept a list of possible sources for version.h Sep 22 11:14:11 should something in gpe-image/opie-image chain not depend on task-bootstrap somwhere? Sep 22 11:16:03 * rwhitby goes to bed, and dreams of a fix by RP or mickeyl ;-) Sep 22 11:16:05 XorA: task-base should be in there instead for Zs Sep 22 11:16:13 rwhitby: :) sleep well Sep 22 11:16:40 mickeyl: thx - night all Sep 22 11:16:57 mickeyl: I suspect we should just pass ${S} to get_kernel_version and teach it knowledge internally Sep 22 11:16:59 'night rwhitby Sep 22 11:17:11 RP: ack Sep 22 11:17:43 XorA: I suddenly get the feeling we might be missing some core packages in task-base... Sep 22 11:17:53 RP: something is wrong, Im getting no busybox Sep 22 11:18:18 RP: :-D Sep 22 11:18:29 that's expected Sep 22 11:19:12 i have the following here: Sep 22 11:19:12 RDEPENDS_task-xxxxxxxx-core = "\ Sep 22 11:19:12 task-base \ Sep 22 11:19:12 base-files \ Sep 22 11:19:12 base-passwd \ Sep 22 11:19:13 busybox \ Sep 22 11:19:15 initscripts \ Sep 22 11:19:17 sysvinit \ Sep 22 11:19:19 tinylogin \ Sep 22 11:19:21 dropbear \ Sep 22 11:19:23 " Sep 22 11:19:41 Yes, this was a slight oversight in task-base :) Sep 22 11:20:10 Basically, we need to create a tasks package containing them and have it elected on a per distro basis - let me check something in ;-) Sep 22 11:20:19 cool Sep 22 11:21:56 where did mickey|bbl pull that from? Sep 22 11:22:12 XorA: I think he has it locally Sep 22 11:22:22 RP: ah, I thought I was going insane Sep 22 11:22:26 I have other things which meant I didn't realise this was missing Sep 22 11:22:47 It proves nobody's tested .dev for a while - I thought the lack of bugs was worrying Sep 22 11:22:48 RP: I only noticed on a new /tmp Sep 22 11:23:41 RP: well I am testing oz-unstable right now Sep 22 11:23:53 RP: just yesterday I was being an idiot and not seeing the obvious Sep 22 11:28:12 afternoon Sep 22 11:30:33 hi reenoo Sep 22 11:30:58 hey RP Sep 22 11:31:55 03rpurdie 07org.oe.dev * r637a94bb... 10/ (1 packages/tasks/task-base.bb): task-base: Add a distro selectable base set of depends so we can mimic task-bootstrap correctly Sep 22 11:34:32 XorA: Applying that to oz354x should give you what you need Sep 22 11:35:01 RP: Im building oz-unstable Sep 22 11:35:42 XorA: Sorry, I just can't read :) Sep 22 11:36:41 RP: typo at line fuser, no \ Sep 22 11:38:59 XorA: pushed, thanks Sep 22 11:39:51 03rpurdie 07org.oe.dev * r3313872a... 10/ (1 packages/tasks/task-base.bb): task-base: Fix typos (thanks XorA) Sep 22 11:40:29 RP: bitbake told me ~:-) Sep 22 11:44:04 ~curse trolltech for lack of search funtion in docs Sep 22 11:44:07 May the fleas of a thousand camels infest your most sensitive regions, trolltech for lack of search funtion in docs ! Sep 22 11:45:41 camels in norway seem unlikely Sep 22 12:04:57 bother, I think I need to commit a fstab for the OSK :) Sep 22 12:08:26 hmmm Sep 22 12:08:48 I am wondering if task-base stuff is biting the guy that just posted to the list about his OSK board ... Sep 22 12:14:50 koen|away: thanks for removing the broken gpe-plucker stuff :) Sep 22 12:17:54 does anyone know if boostrap-image behavior has changed lately? Sep 22 12:20:02 hmmm Sep 22 12:20:22 wtf is ${MACHINE_TASK_PROVIDER} Sep 22 12:20:25 uf. my feed-browser query util in qt4 works Sep 22 12:21:04 hrw|work, got a clue about my question? Sep 22 12:21:11 Crofton: nope Sep 22 12:21:16 * florian_kc wants a gtk one ;) Sep 22 12:21:16 I suspect I need to add something to the OSK machine file Sep 22 12:21:51 stress kills Sep 22 12:21:57 * Crofton implodes Sep 22 12:30:14 exactly what does {$S} mean? Sep 22 12:30:24 Sourcedir Sep 22 12:31:09 i.e. the directory unpacked from a tarball within the work directory Sep 22 12:31:25 yes Sep 22 12:32:29 good :), thanks Sep 22 12:35:09 and any shortcuts for $STAGING_DIR/i686-linux/bin ? Sep 22 12:35:30 $STAGING_BIN_DIR maybe ... Sep 22 12:35:43 grep other bb files and see if it is real Sep 22 12:36:12 ${STAGING_BINDIR} Sep 22 12:37:32 yeah Sep 22 12:37:37 something like that :) Sep 22 12:53:47 morning Sep 22 13:10:29 hi chouimat Sep 22 13:17:06 30s contra 19s... php script rewritten to c++/qt4 Sep 22 13:25:17 Is it possible to make kernel provider package, that will checkout from SVN 'trunk' without any -r settings to SVN (without rewriting too many rules)? Sep 22 13:28:22 slapin, use SRCDATE="now" Sep 22 13:30:23 psokolovsky: any file I could use as template? Sep 22 13:31:01 slapin, try grepping the tree ;-) Sep 22 13:31:33 03tmbinc 07org.oe.dreambox * r7c628933... 10/ (1 packages/twisted/twisted_2.4.0.bb): twisted-2.4.0: work around broken installscript Sep 22 13:34:18 slapin, Making .bb for hackndev kernel? Sep 22 13:34:27 psokolovsky: yeah. Sep 22 13:34:42 slapin, you should have told that ;-) Sep 22 13:36:36 slapin, keep in mind that BitBake 1.6 and older *re-checks out* repo for such .bb on each build! Sep 22 13:37:30 slapin, This was fixed in BB trunk, but it maybe broken to use it. So, I suggst backporting that change to 1.6. There's also OE bug ticket on that, can get patch from that. Sep 22 13:37:57 03tmbinc 07org.oe.dreambox * rf69d6376... 10/ (1 conf/machine/dm600pvr.conf conf/machine/dm7020.conf): machine/dm7020.conf,dm600pvr.conf: use a known-good glibc-initial Sep 22 13:38:49 psokolovsky: do you remember ticket name? Sep 22 13:39:42 slapin, http://bugs.openembedded.org/show_bug.cgi?id=1367 Sep 22 13:44:53 RP: well it boots, now just to work out how to install ipkg without having ipkg Sep 22 13:46:50 XorA: ar -x ipkg.ipk;tar zxf data.tar.gz Sep 22 13:58:17 i would like to compile sdl on diet-x11 Sep 22 13:58:22 i've this error : Sep 22 13:58:36 dependency virtual/libx11 (for libsdl-x11) not satisfied Sep 22 13:58:54 is it possible to use it with diet-x11 ? Sep 22 13:59:06 Genesis: yes Sep 22 14:01:07 cool :] Sep 22 14:01:26 XorA, : how to do ? Sep 22 14:01:35 Genesis: it just works here Sep 22 14:01:51 Genesis: though its been a few weeks since I tried Sep 22 14:02:14 perharps modify the bb file Sep 22 14:02:16 DEPENDS = "alsa-lib virtual/libx11 libxext" Sep 22 14:02:31 remplacing libx11 by diet-x11 Sep 22 14:02:51 i think that there is a better solution ... Sep 22 14:04:57 RP: so missing to make booting system is ipkg, procps, util-linux-mount Sep 22 14:06:21 virtual/libx11 is really bigger than diet-x11 ? Sep 22 14:06:39 XorA: heh, your PC isn't here yet, but don't worry - we have a doz PC for you until it arrives ;) Sep 22 14:06:57 RP: but two of them are due to regressions in busybox so I dont want to force them on people Sep 22 14:07:16 lrg: heh, Wolfson do a bring your own PC policy :-) Sep 22 14:07:31 XorA: do you have laptop ? Sep 22 14:08:11 lrg: no, only a shuttle and a Z Sep 22 14:21:13 03tmbinc 07org.oe.dreambox * rbe8f8350... 10/ (1 packages/twisted/twisted_2.4.0.bb): twisted: experimental package split Sep 22 14:24:05 time for me to format the machine, so I switch to gprs Sep 22 14:29:13 Anybody remembers how to fix plenty of NOTE: bb.build.FuncFailed:("can't read from Sep 22 14:29:13 file '%s' (%s)", Sep 22 14:29:13 ('/oe/build.tosa/tmp/staging/tosa-linux/kernel/kernel-abiversion', Sep 22 14:29:13 )) while evaluating: Sep 22 14:29:13 ${@base_read_file('/oe/build.tosa/tmp/staging/tosa-linux/kernel/kernel-abiversion')} Sep 22 14:29:14 NOTE: bb.build.FuncFailed:("can't read from file '%s' (%s)", Sep 22 14:29:16 ('/oe/build.tosa/tmp/staging/tosa-linux/kernel/kernel-abiversion', Sep 22 14:29:18 )) while evaluating: Sep 22 14:29:20 ${KERNEL_VERSION} Sep 22 14:30:21 (paths are not mine there - actually I have the same with familiar tree, upgraded bitbake to 1.6) Sep 22 14:35:02 slightly delayd there Sep 22 14:50:43 time to watch the worst movie ever Sep 22 14:51:08 hrw|work: whats that Sep 22 14:51:16 apart from titanic Sep 22 14:51:17 plan 9 from outer space Sep 22 14:51:49 :) thats a cult over here Sep 22 14:56:11 hrw|work: it's endearingly dreadful! http://i8.photobucket.com/albums/a8/KerwoodDerby/VampiraAndTor.jpg Sep 22 15:12:56 XorA|gprs: The missing ipkg is a bug - I guess it gets added to some RDEPENDS we now bypass... Sep 22 15:14:11 slapin: Thats a bug in familar which they refused to check in a fix for. Not our problem. Sep 22 15:14:17 RP: I guess thats a distro option as well Sep 22 15:14:41 XorA|gprs: It needs to be set by the rootfs generation class really... Sep 22 15:14:51 RP: package_ipk.bbclass adds ipkg stuff Sep 22 15:15:04 I added that years ago.. Sep 22 15:15:12 hrw|work: It shouldn't - it should get added, only if the image is generated with ipks Sep 22 15:15:27 RP: or was it rootfs_ipk.. Sep 22 15:15:40 hrw|work: generated images are now missing ipkg Sep 22 15:15:44 package_ipk.bbclass:BOOTSTRAP_EXTRA_DEPENDS += "ipkg-collateral ipkg ipkg-link" Sep 22 15:15:47 .oz354x Sep 22 15:15:57 package_ipk.bbclass:BOOTSTRAP_EXTRA_RDEPENDS += "ipkg-collateral ipkg ipkg-link" Sep 22 15:16:01 .dev and .oz354x Sep 22 15:16:30 We can add a new verison for task-base, in the right place this time... Sep 22 15:16:37 XorA|gprs: inherit package_ipk SHOULD then do DISTRO_EXTRA_RDEPENDS += same stuff Sep 22 15:16:49 RP: what is actual fix for it? So I'd fix that in my tree? Sep 22 15:17:19 Also a question - is it possible to disable packages generation? Sep 22 15:17:25 hrw|work: it doesnt, and at moment Im on a mobile so can fix Sep 22 15:17:33 slapin: some change to kernel.bbclass. Sep 22 15:18:45 RP: there is no fix for it that could be done outside bitbake. there's only a nasty workaround for it in base.bbclass that will cause an error condition to be silently ignored. Sep 22 15:18:51 s/can/cant/ Sep 22 15:19:31 RP: I'll be glad to check in an actual fix if there is one Sep 22 15:21:20 reenoo: We've had this dicussion before. We can't have variables that error when bitbake reads them. It is as simple as that Sep 22 15:21:21 yo incinerator Sep 22 15:21:23 reenoo: iirc there checkings are in sharp-binary/sdmmc-support recipes Sep 22 15:21:34 hi Sep 22 15:22:32 reenoo: I have no issue if it returns nothing as the version, "INVALIDVERSION" or anything else it would like to return but raising an exception when the variable is read is never going to work Sep 22 15:22:59 and we do need to read RDEPENDS Sep 22 15:24:06 not in our tree Sep 22 15:24:28 right. Please ban people from using bitbake > 1.3 with it then Sep 22 15:24:37 or whenever the problems started Sep 22 15:24:55 "ban people"? wtf? Sep 22 15:26:27 there's no easy way to do that anyway since bitbake doesn't check which version of python modules it runs against Sep 22 15:27:13 reenoo: See sanity.bbclass in OE - it checks which version of bitbake is running Sep 22 15:28:24 reenoo: Above you should s/not in our tree/not with an old version of bitbake/ Sep 22 15:28:39 bitbake will happily pick up python modules from earlier or later versions Sep 22 15:29:03 yay for subtle breakage all over the place Sep 22 15:29:09 That is a bug in bitbake we should address and it a separate issue Sep 22 15:29:47 reenoo: Our metadata shouldn't make unreaonable demands on bitbake... Sep 22 15:30:08 reenoo: As I've explained before, it is totally unavoidable to have to do this to develop the bitbake codebase any further Sep 22 15:39:52 time for me to head pubwards, bye Sep 22 15:42:59 RP, do you know if building bootstrap-image will wtill work as it did a few weeks back? Sep 22 15:43:19 Crofton: For which machine? Sep 22 15:43:26 omap5912osk Sep 22 15:43:50 There was an email to the list from a guy I have been working with Sep 22 15:43:57 and he has some problem booting Sep 22 15:44:13 cu Sep 22 15:44:22 unfortunately I won't be able to rebuild on my machine for several days Sep 22 15:44:32 Crofton: It should still work as it did although recent versions of busybox sound broken Sep 22 15:44:59 I did make some invasive changes but they'd only affect zaurus users mainly Sep 22 15:45:36 so boostrap-image still makes task-bootstrap? Sep 22 15:45:50 I don't need to tell the machine file to call that task? Sep 22 15:46:14 Crofton: It defaults to the old behaviour Sep 22 15:46:22 ok Sep 22 15:46:56 I am runnning busybox 1.2.1 in my osk Sep 22 15:47:19 you mean recent changes to that bb file broke busybox? Sep 22 15:48:08 RP: it all depends on your (impersonal) agenda. if you want to market/sell bitbake as a tool of its own you may have to do it that way especially when deadlines are involved. if you want to get work done in the sense of putting out a distribution your choices may be different. playing catchup games because bitbake keeps redefining concepts is certainly not an option. Sep 22 15:50:02 RP: this extends to OE as well. for several packages in the base system you'd get better results by going back to what we had in mid-2004. an no, unfortunately I'm not kidding. Sep 22 15:50:14 03tmbinc 07org.oe.dreambox * raf5d7454... 10/ (1 conf/bitbake.conf): bitbake.conf: remove space after -I Sep 22 15:51:09 a+ Sep 22 15:51:28 reenoo: Keep in mind I maintain a stable distro myself... Sep 22 15:52:13 reenoo: and bitbake didn't redefine a concept. The metadata tried to take advantage of something it shouldn't which happened to work for a while. Sep 22 15:53:37 reenoo: hmm... which packages in the base system do you have in mind? Sep 22 15:54:08 RP, is the busybox breakage you mention linked with the boot process not finding init? Sep 22 15:56:05 RP: I've been around long enough to know that RDEPENDS was originally never known or intended to be known by bitbake Sep 22 15:58:33 * florian_kc really dislikes this kind of discussions Sep 22 16:00:27 * Crofton gets annoyed when things stop working for no apparent reason Sep 22 16:00:37 Crofton: That sounds like task-base problems which were fixed in .dev today but your board shouldn't be using that? Sep 22 16:00:42 especially when he doesn't have time to fix them Sep 22 16:00:53 I do not think it is Sep 22 16:01:02 it should be using task-bootstrap Sep 22 16:01:12 is busybox in the image? Sep 22 16:01:16 yeah Sep 22 16:01:25 where is init usually Sep 22 16:01:25 it should be ok... Sep 22 16:01:35 I f have done some partial rebuilds Sep 22 16:01:37 /sbin/init Sep 22 16:01:53 yeah Sep 22 16:01:55 ok Sep 22 16:02:14 hmmm Sep 22 16:02:29 I think I need to nuke tmp an rebuild Sep 22 16:02:56 but just no time to investigate and the guy with the problem doesn't speak such good english and brand new to oe Sep 22 16:03:54 it is also possible the problem may be due to updates sent to git omap kernel Sep 22 16:04:16 * Crofton hates grad school Sep 22 16:04:26 too many demands from too many projects Sep 22 16:04:35 * Crofton talks to himself Sep 22 16:20:06 03blaster8 07org.oe.dev * r0b92f5db... 10/ (3 files in 2 dirs): Sep 22 16:20:06 ixp4xx-kernel-2.6.18: pull patches automatically from SVN Sep 22 16:20:06 * Improved previous system so patch tree is pulled automagically Sep 22 16:20:06 * Specific SVN revision can be specified Sep 22 16:20:06 * Patches applied using quilt Sep 22 16:27:01 blaster8: http://www.rpsys.net/openzaurus/temp/oe-kernelversionfix.patch might be a fix for a problem you were seeing - I've not tested yet though Sep 22 16:27:16 florian_kc: "this kind of discussion" could have easily been avoided by accurately describing the problem along the lines of "bitbake has developed in a non-backwards compatible direction that is particularly incompatible with how OE handles kernel<->external module deps. use the version of bitbake that comes with the familiar tree or grab the workaround for base.bbclass from org.oe.dev". it was RP's choice to continue the bashing, not mine. Sep 22 16:27:51 hm? Sep 22 16:27:55 whats wrong now Sep 22 16:31:07 i don't really understand waht reenoo wanted to say... complaining about packages and going back to a status of 2004 without one word of explaining. Sep 22 16:32:03 RP: OK, I will test Sep 22 16:34:11 florian_kc: Someone asked why famailar won't work with bitbake 1.6 - I told them. reenoo felt he should dicuss the issue with me. Both of our positions remain unchanged :-( Sep 22 16:35:11 What makes it so annoying is the underlying change is trivial and has never caused a problem in .dev and never should cause a problem. I've agreed its far from ideal but it is a metadata issue, not a bitbake one Sep 22 16:36:58 RP: yeah... i really wonder why this is necessary :-/ Sep 22 16:37:38 RP: but don't worry, he complained about my gpe-screenshot fix in a similar way. "gtk is broken..." Sep 22 16:37:58 florian_kc: I also wonder why :-( Sep 22 16:40:25 strange Sep 22 16:47:48 RP: no depmod error in 2.6.18 build with your patch Sep 22 16:48:04 should I test backward compatibility (2.6.17 build)? Sep 22 16:49:15 bbl Sep 22 16:50:01 blaster8: It should be ok but if you have a build you can test with... Sep 22 16:50:15 * RP back later Sep 22 16:50:30 OK, I'll report 2.6.17 build results Sep 22 16:52:50 what does this kernel patch do? Sep 22 16:53:20 it allows the kernel bbclass to read from the new linux version file in 2.6.18 Sep 22 16:53:23 hmm multiple heads from RP and blaster? Sep 22 16:53:35 hmmm Sep 22 16:53:53 I wonder if I need this for the omap kernel bb file? Sep 22 16:58:08 03crofton 07org.oe.dev * re1ceeb76... 10/ (3 files in 3 dirs): * base-files: add machine specific fstab for omap5912osk machine Sep 22 17:00:50 RP: 2.6.17 also reports no error Sep 22 17:01:03 Time to commit your patch I think :) Sep 22 18:08:50 ~seen minipanda Sep 22 18:09:04 minipanda was last seen on IRC in channel #oe, 8d 6h 29m 24s ago, saying: 'crap, in $OESTUFF/build/tmp/rootfs/ have /dev, but in bootstrap-image.tar.gz, no!'. Sep 22 18:18:52 ~seen pgfeller Sep 22 18:18:55 pgfeller was last seen on IRC in channel #opie, 3d 23h 35m 45s ago, saying: 'yang: sorry; have to go ... - cu.'. Sep 22 18:22:10 I have a bluetooth network set up using pand, everything connects and works but if I disconnect the slave, I can't get it to reconnect to the master unless I restart the master PAND server. Sep 22 18:29:18 well after more testing it is not pand that needs re-run but the network connection restarted on the server "ifconfig bnep0 192.168.2.1" Sep 22 18:40:13 Lol, missed hot discussion ;-) Sep 22 18:41:26 Anyway, Familiar tree *does* work with bitbake 1.6, and I even dare to recommend it to all people who has to work with Familiar. Sep 22 18:44:53 which tree? Sep 22 18:46:28 emte: There's single Familiar tree, isn't it? ;-) the git one Sep 22 18:46:35 no there are two Sep 22 18:46:44 three actually Sep 22 18:46:59 emte: lol. actually, three. no, four, count mine too ;-) Sep 22 18:47:00 and which devices are you refering to? Sep 22 18:47:24 emte: ? Sep 22 18:47:28 i suspe3ct any of the 2.4 kernel devices wont build with the 1.6 OE bitbake Sep 22 18:47:31 suspect* Sep 22 18:47:49 emte: yep, 2.4 is too old on its own Sep 22 18:48:02 then you should avoid reccomending it Sep 22 18:48:05 emte: afaik, no one does develop 2.4 kernels nowadays Sep 22 18:48:10 sure they do Sep 22 18:48:34 since there are not enough devs to quickly bring up the 2.6 stuff on older devices Sep 22 18:48:37 emte: nope, because it *does* work, and that's good path for migration to newer versions. Sep 22 18:49:27 i still dont expect the h36 2,6 stuff to work for another 2 years Sep 22 18:50:19 emte: Them I miss things with 2.4, I can admit it. Still, hackndev people, with whose questions the discussion started, use 2.6 kernel. Sep 22 18:50:44 emte: That's bad. Because in 2 years, 3600 maybe itself too old develop anything at all for it ;-( Sep 22 18:51:29 psokolovsky, its been in progress since the 2.6 kernel came out Sep 22 18:51:54 if that gives you any indication on the speed of porting devices Sep 22 18:53:24 emte: 2.6 version for 3600? - I know. Someone should just try and finish it, voila. And I really prefer to believe that such person arrive faster than in 2 years. Again, it's matter of faith ;-) Sep 22 18:54:12 psokolovsky, just try and finish it? ... most everything needs to be rewritten Sep 22 18:54:28 gremlin[it], is one of the most active devs on it Sep 22 18:54:42 comeone talk about me ?!?!?! Sep 22 18:54:49 emte: was, he said he's out of it for last time ;-( Sep 22 18:54:51 yup i did :) Sep 22 18:55:23 psokolovsky surelly fast than a year ... surelly slower than a couple of month :( Sep 22 18:56:47 emte: anyway, we started with bitbake 1.6 and Familiar. My point is that even maintainer can't predisct usages Familiar tree will have. So, people use bb 1.6 on it, and it works. And there's nothing to do about it. Even Microsoft can't predict ways Windows is used. People install cygwin on it, and use as GNU/Windows system. ;-) Sep 22 18:57:30 gremlin[it]: That sounds much more cheerful! Sep 22 18:57:40 ohh thanks :) Sep 22 18:58:29 gremlin[it], thank you, and I hope h3600 will progress! Sep 22 18:58:30 psokolovsky, encouraging people to use a tool that is not supported for development purposes seems like you may be intentionally causing dev problems Sep 22 18:58:30 i would like to do a xmas gift to "all" :) Sep 22 18:59:05 psokolovsky sincerelly if i wasn't busy every night with a work on ep93xx i would on h3600/3700/3800 Sep 22 18:59:49 but that is due to behaviour and changes in the tools themselves Sep 22 19:00:50 emte: I'm not the maintainer, just dev who shares his experiences with another devs. No need to suspect me in conspiracy ;-). But yep, I try to communicate with Palm people for example and just tell them what issues I had with Familiar/OE/Bitbake myself, and how I solved them. Sep 22 19:00:58 its not any different that trying to compile apps for the Zaurus OS using gcc-4.x Sep 22 19:01:01 than* Sep 22 19:01:28 gremlin[it], Nice to now you leave room for it in your plans! Sep 22 19:02:35 aha ... lets see if i finally rebound my wireless card ... Sep 22 19:02:44 anyone got a Dapper Drake monotone 0.26 package available for am64? Sep 22 19:03:36 emte: Btw, most frequent complaint of people trying to develop own distro/build systems, port Gentoo and stuff, etc. - is OE/bitbake being too slow and uncomfortable to use. But that's old news, and neither OE.dev not bitbake has those issues any more. Sep 22 19:05:03 psokolovsky i have an h3600 and i bought on ebay a h3800 ... i want to support all ipaq with sa1110 Sep 22 19:05:04 * emte witholds comment Sep 22 19:05:14 psokolovsky: is that so? I find it to work perfectly *iff* its works, but if not, you have to be quite familiar with the details in order to fix stuff. Sep 22 19:05:46 likewise, what part of discussion you refer to? Sep 22 19:06:16 psokolovsky: OE/BB slow and unconfortable to use. Sep 22 19:07:53 likewise, yes, and how familiarity needed to fix relates to that? ;-) . BB 1.6 starts up fast after it parsed .bb once, and that works out of the box, no need to "fix" for that ;-) Sep 22 19:08:42 psokolovsky: True, but I was more referring to the 'uncomfortable' part :-) Sep 22 19:10:50 likewise, well, that was mostly referring to startup speed, too ;-). But yes, complexity of OE and BB is the second complaint in the list. Alas, noone yet did cross-build system supporting so many architectures. So, nobody told it will be easy ;-) Sep 22 19:11:32 psokolovsky: True, and in fact yes it works great if it works. :-) Sep 22 21:40:56 hi Sep 22 21:57:08 blaster8: thanks for testing - I'll commit it Sep 22 21:57:43 psokolovsky: Does bitbake 1.6 work with familar without a patch to kernel.bbclass? Sep 22 21:58:25 RP: Yes, it is. those file read message appear, sure, but they are harmless and go away once kernel is built. Sep 22 21:58:47 psokolovsky: Fair enough :) Sep 22 21:59:05 RP: I baked a Fam0.8.4 release for h4000 for users to play in the meantime, and didn't have any visible issues, and images work nice ;-) Sep 22 21:59:40 Which kind of proves we didn't break backwards compatibility that much... Sep 22 21:59:45 * RP sighs Sep 22 21:59:52 yep, I second that Sep 22 22:05:08 03rpurdie 07org.oe.dev * r6f348542... 10/ (1 classes/kernel.bbclass): kernel.bbclass: Update get_kernelversion for changes in kernels >= 2.6.18 Sep 22 22:31:22 * hvontres|poodle gets ready to fill in time shee and call it a week :) Sep 23 01:19:48 anyone can help with a virgin oe user ? Sep 23 02:01:37 dos000: may koen have mercy upon your soul... Sep 23 02:02:18 hah Sep 23 02:02:30 kerwood: are you da man ? Sep 23 02:03:16 dos000: I am so NOT the man, sorry **** ENDING LOGGING AT Sat Sep 23 02:59:56 2006