**** BEGIN LOGGING AT Sun Feb 18 02:59:57 2007 Feb 18 06:00:28 Any advice? I'm trying to build LyX according to the recipe I get from the oz354x repository from monotone. It chugs for a while, builds all deps, then dies with a bunch of errors about libraries being in the wrong format. I notice on the path to gcc it's trying to link against things in /usr/lib. Any idea how to fix? Everything it needs to link against exists in ARM format in my staging dir. (Output: http://pastebin.ca/345807) Feb 18 06:48:32 any idea why gettext depends on gettext-dev ? Feb 18 06:48:54 lol no Feb 18 07:03:30 could some one explain for me what a "hense" is on a cell phone? Feb 18 07:04:02 It says: "The phone works fine but has some marks on it hense" Feb 18 07:04:07 What does he mean? Feb 18 07:08:27 goxboxlive: "Some marks on it hense" would be an unusual way to put it.. it's probably just generically warning that there's a few marks on it Feb 18 07:09:11 right, so it might be the display. Thanks a lot. Feb 18 07:54:05 good morning all! Feb 18 07:54:25 howdy Feb 18 07:54:53 koen morning Feb 18 08:00:52 koen : when we specify to a machine config as PACKAGE_EXTRA_ARCHS ="all mpla mpla mpla" does the all mean packages that are not target specific ? Feb 18 08:01:47 koen : i am having troubles with multimachine builds Feb 18 08:01:56 apart from the fact that OE adds 'all' by default, yes, packages with TARGET_ARCH ="all" aren't target specific Feb 18 08:02:47 koen, not sure if I already asked you: I'm trying to build LyX according to the recipe I get from the oz354x repository from monotone. It chugs for a while, builds all deps, then dies with a bunch of errors about libraries being in the wrong format. I notice on the path to gcc it's trying to link against things in /usr/lib. Any idea how to fix? Everything it needs to link against exists in ARM format in my staging dir. (Output: h Feb 18 08:03:29 calmofthestorm7: no idea how to fix that, we fixed a lot of those bugs in .dev Feb 18 08:03:43 oh neat Feb 18 08:03:54 kind of counterintuitive that dev would be more likely to build than stable Feb 18 08:04:04 then I shall download dev and build the deps, while I sleep tonight. Thanks;) Feb 18 08:04:13 *crosses fingers for magic fix* Feb 18 08:04:24 it's a matter of manpower Feb 18 08:04:36 .oz354x is maintained by hrw, .dev by all of us Feb 18 08:04:41 aah Feb 18 08:04:54 so you get more breakage in .dev, but also more fixes Feb 18 08:05:30 interesting Feb 18 08:08:42 how hard would it be for me to mix packages from dev with stable? Feb 18 08:08:56 eg, if I manage to bake lyx from dev would it run on a oz with a stable image? Feb 18 08:08:57 or is it all or nothing? Feb 18 08:12:05 hi psokolovsky_ Feb 18 08:12:14 Hi! Feb 18 08:12:40 * gremlin[it] is becoming crazy trying to schedule what he want to attend at FOSSDEM !!! Feb 18 08:13:40 03koen 07org.oe.dev * rd6b3b1b5... 10/ (3 files in 2 dirs): openmoko-panel-plugins: add gsm and mainmenu Feb 18 08:13:45 03koen 07org.oe.dev * rd6f1925c... 10/ (1 packages/tasks/task-openmoko.bb): task-openmoko: convert to matchbox-panel-2 and clean up a bit Feb 18 08:15:16 * gremlin[it] would be a read gremlin as film ... with water them multiplicate !!! Feb 18 08:15:26 would be a reaL Feb 18 08:18:30 gremlin[it]: Just don't end up like the guy in Multiplicity with a Copy of a copy :) Feb 18 08:21:40 hahahah yes yes ... and heve to admit ... the original is already far from "normal" :) :P Feb 18 08:27:37 anyone knows what "BINFMT_FLAT: reloc outside program" means? Feb 18 08:41:56 03pfalcon 07org.oe.dev * r8cde5e50... 10/ (1 packages/linux/linux-handhelds-2.6_2.6.19-hh11.bb): (log message trimmed) Feb 18 08:41:56 linux-handhelds-2.6 2.6.19-hh11: Add new release. Feb 18 08:41:56 * Changes since -hh10 include: Feb 18 08:41:56 1. aximx5 port revived - boots with LCD working. UDC known not to work. Feb 18 08:41:56 2. aximx50 buttons driver. Feb 18 08:41:57 3. Macine-independ touchscreen driver for ADS7846/TSC2046, first version so Feb 18 08:42:00 far, used for h4000 and works for hx4700 too. Still needs tweaking to Feb 18 08:50:31 Laibsch, do you think that "To achieve what we are trying to accomplish" is good wording? ;-) Feb 18 08:54:26 has anyone noticed header files to be included in a uclibc build ? Feb 18 08:54:49 for some reason gettext seems to depend on gettext-dev which depends on uclibc-dev Feb 18 08:55:13 as a result i get all kind of header files included in the rootfs Feb 18 08:59:20 Ifaistos: INHERIT += "insane" in local.conf and rebuild Feb 18 08:59:56 that should error out on packaging errors Feb 18 09:00:08 ok Feb 18 09:03:08 03pfalcon 07org.oe.dev * r1fc4ea96... 10/ (1 packages/linux/linux-handhelds-2.6_2.6.19-hh12.bb): (log message trimmed) Feb 18 09:03:08 linux-handhelds-2.6 2.6.19-hh12: Add new release. Feb 18 09:03:08 * Changes since -hh11 include: Feb 18 09:03:08 1. rx1950 port added. Feb 18 09:03:08 2. cpugreq fixes to support PCMCIA working over freq changes. Feb 18 09:03:10 3. cpufreq support for htcmagician. Feb 18 09:03:12 4. ads7846_adc_ssp fixes. Feb 18 09:30:52 koen : insane did not report any error but the header files are there :( Feb 18 09:34:00 does this have to do anything ? -> PREFERRED_PROVIDER_virtual/libintl = "libintl" Feb 18 09:34:16 the provider of libint is gettext Feb 18 09:36:43 psokolovsky_: yes, although I hesitated about that particular one. If you have a better idea, feel free to change. At least it is a sentence now which was not the case before. Feb 18 09:41:31 Laibsch, well, it's sound like tautology. if in real world there's idioms like "if we want to do what we want to do", then using more funky than "do" verbs, and synonym-interleave them, makes it a bit funny. ymmv ;-) Feb 18 09:41:43 anyway, I'm not native English speaker ;-) Feb 18 11:00:03 03koen 07org.oe.dev * re63a2eef... 10/ (4 files in 3 dirs): openmoko distro: add openmoko distro, slightly modified from openmoko svn Feb 18 11:10:57 03koen 07org.oe.dev * r04a1a2e0... 10/ (5 files in 3 dirs): (log message trimmed) Feb 18 11:10:57 tslib: merge patch from openmoko: Feb 18 11:10:57 This patch adds support for "EVIOCGRAB" on the input device, which Feb 18 11:10:57 tells the kernel _not_ to deliver events of the touchscreen to Feb 18 11:10:57 /dev/input/mice. Feb 18 11:10:58 This is probably what most people want, since unprocessed raw touchscreen Feb 18 11:11:02 events should not be converted to emulated PS/2 mouse movements. Feb 18 11:16:20 mithro: hey, how are you? Feb 18 11:16:32 okay Feb 18 11:16:44 mithro: did you finish your hardware? Feb 18 11:18:14 zecke: not due to middle of this year Feb 18 11:30:28 been up to anything intresting? Feb 18 11:30:54 can't comment :) Feb 18 11:31:07 mithro: the most interesting now is being abused by a cold/flu Feb 18 11:32:45 i put my back out and was in bed on pain killers for a few days Feb 18 11:33:30 otherwise I would be waiting on the first complete prototype of the device Feb 18 11:34:51 koen: have a look at this -> http://www.pastebin.ca/361894 is this normal ? Feb 18 11:39:04 welp off to bed Feb 18 11:40:07 koen|away: well, emulators are not overrated. They are needed for systems where you don not emulate 90% out of the box. Feb 18 11:44:46 Ifaistos: no, it means that uclibc packaging sucks Feb 18 11:53:03 i wonder if other uclibc distros have the same issue and no one noticed or if this is a 2.9.28-svn issue Feb 18 12:02:02 ljp: Will Trolltech release SXE now as OLPC develops Bitfrost? Feb 18 12:08:55 aeh, what is wrong with our website? Feb 18 12:18:29 zecke: apache doesn't access the framebuffer correctly, so image loading is slow Feb 18 12:19:10 s/correctly/directly/ Feb 18 12:20:40 hehe, I should really cache images in SRAM Feb 18 12:22:10 koen: do you know who removes the $WORKDIR/install/package-name directory and when? Feb 18 12:22:31 koen: package_qa fails as the dir is empty for the next run (only a do_install helps here) Feb 18 12:23:16 koen : And this is for gettext -> http://www.pastebin.ca/361952 Feb 18 12:23:18 BTW: I hate this world, klibc has EABI support (again and again every time I check) Feb 18 12:27:47 koen : these are the header files that get included in the rootfs -> http://www.pastebin.ca/361956 Feb 18 12:29:06 hmm this dir is not empty Feb 18 12:31:27 zecke : are you refering to what i posted ? Feb 18 12:33:15 Ifaistos: no :) Feb 18 12:35:45 koen: okay the dir gets emptied on the second run of do_package Feb 18 13:08:11 zecke: do you have some time for me? Feb 18 13:11:08 some Feb 18 13:11:25 I hope I remember some of the stuff you will ask Feb 18 13:13:00 zecke: do you know how the simpad wakes up? in simpad.c is a line /* enable wake-up on GPIO0 (Assabet...) */ PWER = GFER = GRER = 1; Feb 18 13:13:41 zecke: is it neccessary to have a driver for gpio0 or should it wake up by just pressing the power button Feb 18 13:14:23 mr_nice: this is a stock StrongARM thing. We define GPIO0 which is the Power Button (IIRC) as wakeup source Feb 18 13:14:43 mr_nice: this should make a reset, the cpu should jump to the reset vector and the device should resume Feb 18 13:15:27 mr_nice: the tricky part, which Chicken (Alexander Lange), struggled with when porting bootldr was, you had to toggle something to keep the machine on Feb 18 13:16:26 zecke: So there is no irq handler for gpio0 needed? Feb 18 13:16:33 It is amusing that every map of Brussels I find, has differnet names for the same streets Feb 18 13:17:41 Crofton|laptop: it's dead center in the language divide Feb 18 13:17:49 yeah Feb 18 13:17:50 Crofton|laptop: like quebec in canada Feb 18 13:17:50 zecke: I ask because the driver for the power button isn't written atm. And I need to know if I have to write it first to resume the simpad. Feb 18 13:17:53 It shows :) Feb 18 13:18:16 oh, nothing could be as bad as quebec :) Feb 18 13:20:11 Crofton|laptop: seeing asd your now the expert, you'll have to do the OE map ->astrid -> beer Feb 18 13:20:16 mr_nice: no, just make sure the power button (would be a really simple driver) is defined as wakeup source Feb 18 13:20:47 mr_nice: e.g. pb_ created a generic GPIO based btn 'framework' Feb 18 13:21:11 I only think I know where the Astrid is :) Feb 18 13:21:37 zecke: is it in mainline or on hh.org? Feb 18 13:21:37 Crofton|laptop: I am landing at Charlaroi (or however you spell it) and info on that airport is kinda vague Feb 18 13:21:42 but by Friday, I will know how to get from Astrid to beer Feb 18 13:21:59 zecke: where do I have to devine the wakeup source? Feb 18 13:22:07 zecke: in simpad.c? Feb 18 13:22:12 I come in th einternational airport and can get a trin to crentral Feb 18 13:22:39 I saw something on the train website about Charleoi Feb 18 13:22:46 the advantage I suppose of coming in some wierd Airport is the airfar is 10.18 GBP Feb 18 13:23:00 that is crazy Feb 18 13:23:13 do you have to hand crack the engines? Feb 18 13:23:29 Crofton|laptop: bugger knows how Ryanair make money Feb 18 13:23:51 Crofton|laptop: maybe they ship drugs, thats why your luggage allowance is so low Feb 18 13:24:32 heh Feb 18 13:24:33 the air stewardesses do tend to be rather high quality though Feb 18 13:25:59 mr_nice: the machine init function would make a good candidate Feb 18 13:26:15 mr_nice: but it should be already setup properly (I assume this) Feb 18 13:26:54 zecke: Ah, yes PWER=PWER_GPIO0 | PWER_RTC; defines it as wakeup source in simpad.c? Feb 18 13:27:20 mr_nice: yes Feb 18 13:27:30 zecke: sorry for asking such stupid but I want to figure out why it won't wake up ;) Feb 18 13:28:15 mr_nice: I wondered why you asked. How do you define not waking up? is the serial/screen dead? Feb 18 13:28:28 zecke: yes Feb 18 13:28:46 zecke: If I do a apm -s. I can't get it back Feb 18 13:29:26 mr_nice: try to find out if it tries to wakeup (it should) Feb 18 13:29:37 zecke: How to do this? Feb 18 13:29:43 mr_nice: e.g. go somewhere in the resume path and set one of the LEDs to on Feb 18 13:30:08 zecke: thx I will try this :) Feb 18 13:31:16 mr_nice: or enable low level debugging and print something to the serial Feb 18 13:31:58 zecke: How to enable low lever debugging? Feb 18 13:35:35 mr_nice: check for CONFIG_DEBUG_LL you need to enable it with make menuconfig... Feb 18 13:35:49 zecke: thx Feb 18 13:37:34 zecke: If waking up,powerbutton driver, battery driver and the bug in the mq200 driver is solved I would say the 2.6 driver is as competed as the 2.4 driver is :) Feb 18 13:38:33 battery is simple, powerbutton is trivial, and with the datasheet mq200 is possible as well Feb 18 13:39:04 zecke: yes :) Feb 18 13:45:19 I am talking with a guy running an image I buitl from OE on the OSK Feb 18 13:45:32 he istrying to use the serial, after stopping getty Feb 18 13:45:40 he can write chars to it, but not read them Feb 18 13:45:47 ring a bell with anyone? Feb 18 13:53:37 koen: I think you have to search a new topic. QtopiaCore can also work with IO based Graphic Controlers Feb 18 13:56:36 is it possible to connect a laptop to say, another laptop directly via ethernet w/o crossover cable? Feb 18 13:56:59 sure via cross ethernet cable Feb 18 13:57:42 no ... i think about no one ethernet on laptop have auto MDX capability ... Feb 18 13:58:42 thought maybe there was a software trick =) To bad Feb 18 13:59:16 CoreDump|home: if you have a mac Feb 18 13:59:27 zecke: nahh, I'm cheap Feb 18 13:59:37 CoreDump|home: mac's are cheap too Feb 18 14:00:30 zecke: they are not exactly expensive if you consider the built-in hardware. However, they are a bit over my budget =) Feb 18 14:00:50 hehe Feb 18 14:01:14 morning joshin ;) Feb 18 14:01:16 now where did I put this ancient 10Mbit hub Feb 18 14:01:34 Morning NAiL! Feb 18 14:06:42 re Feb 18 14:19:03 mickey|slides: mithro : how do I read a byte in python? how do I swap endianes of a byte? Feb 18 14:22:54 use struct module Feb 18 14:25:08 mickey|slides: hmm that is bad api :( Feb 18 14:25:18 well Feb 18 14:25:20 you get what you ask for Feb 18 14:25:32 low level access in python is wrong tool for the job as well :) Feb 18 14:25:38 mickey|slides: which module allows me to use inline assembly? Feb 18 14:25:50 dunno offhand, i don't do such crazy things Feb 18 14:26:12 mickey|slides: I think doing inline assembly would increase the readbility above struct.unpack Feb 18 14:26:32 * zecke tries to read the elf header from within python Feb 18 14:27:05 * Crofton|laptop thinks zecke is nuts Feb 18 14:29:07 mickey|slides: back to your slides! (which day is the presentation on?). I have read about the different filesystems Feb 18 14:29:23 mickey|slides: and I find initramfs quite funny, specially the merging it is supposed to do Feb 18 14:30:59 zecke: Thursday, 22th. 14:00 presentation 14:15 query session 16:30 life or die Feb 18 14:31:36 mickey|slides: and after 16:30 is overtime with sudden death? Feb 18 14:31:42 ah this is defense? Feb 18 14:31:42 mickey|slides: or cookies and biscuits? Feb 18 14:31:54 zecke: cookies and biscuits in any case Feb 18 14:32:04 Crofton: yep Feb 18 14:32:08 Hmmm, I suppose it is in German also? Feb 18 14:32:15 yep Feb 18 14:32:18 heh Feb 18 14:32:26 which is ok, since it's my mother tongue Feb 18 14:32:29 heh Feb 18 14:32:33 MickeyL: Can you arrange the flyers and posters for the booth ? Feb 18 14:32:34 I arrive in Brussels at 7:30 AM Thrusday Feb 18 14:32:36 yeah Feb 18 14:32:54 gerwinin: sorry, no way. i have my presentation on thursday and i'm flying on friday Feb 18 14:33:20 I can print/plot anything you throw at me Feb 18 14:33:35 ya, problem is the content Feb 18 14:33:37 MickeyL: Zecke before when do I need to deliver the materials Feb 18 14:33:49 MickeyL : I will deliver the content Feb 18 14:34:00 gerwinin i send u an email ... if other help is need ask :) Feb 18 14:34:12 Gremlin : cool :) Feb 18 14:34:14 Thursday 16h so I can go and plot the poster in a shop Feb 18 14:34:40 gerwinin: printing flyers until friday morning Feb 18 14:34:53 Zecke: Okay and can you get a receipt so that we can pay you from the donations we get Feb 18 14:35:29 MickeyL: can you send me your slides so we can publish them on the website Feb 18 14:35:56 which slides are you referring to? Feb 18 14:36:11 MickeyL: the slides you are making for your presentation Feb 18 14:36:21 ah, for the OE one? Feb 18 14:36:34 yes, I can send them to you. i hope i will at least finish them in time for FOSDEM. Feb 18 14:37:02 MickeyL: okay Feb 18 14:37:29 MickeyL: I will send the pressrelease to linuxdevices tomorrow Feb 18 14:37:34 cool Feb 18 14:37:36 thanks Feb 18 14:37:55 MickeyL: I will announce your lecture there as well Feb 18 14:38:55 I made some appointment with somebody from kde who raised a foundation for kde , he will explain us how it is done Feb 18 14:40:19 MickeyL: I am thinking how we can present the project gpe/opie/gpephone/openmoko Feb 18 14:43:53 gerwinin: and the name is? Feb 18 14:44:18 Zecke: A german dude that speaks dutch very well Feb 18 14:44:25 Zecke: let me check Feb 18 14:44:34 gerwinin: Sebastian? Feb 18 14:45:54 Zecke: yes Feb 18 14:45:56 gerwinin: Don't announce it yet, I only have a backup talk Feb 18 14:46:08 which means, in case one is cancelling Feb 18 14:46:23 gerwinin: hehe, we have our KDE e.V. for years Feb 18 14:46:29 MickeyL: I hate it that people from trolltech and stuff get time to talk and we don't Feb 18 14:46:34 hehe Feb 18 14:46:37 well we have the booth Feb 18 14:46:40 and we have been there twice in a row Feb 18 14:46:48 mickey|slides: or someone falls down the stairs and breaks a leg... :} Feb 18 14:46:53 hah Feb 18 14:46:57 mickey|slides: Get me drunken on sunday afternoon and you get my slot. :) Feb 18 14:47:01 and we will be mentioned in both OpenMoko and OpenEZX presentation Feb 18 14:47:06 Zecke: We should ask them how they did it Feb 18 14:47:09 stefan_schmidt: nah, OpenEZX is important Feb 18 14:47:27 gerwinin: I'm a member of this KDE e.V. (not since founding but still) Feb 18 14:47:42 mickey|slides: just do it like mjg, and hijack one session Feb 18 14:47:48 Zecke: We should have a look how they do things Feb 18 14:47:59 mickey|slides: heh, ok. That means I need to get my slides ready. :) Feb 18 14:47:59 gerwinin: ev.kde.org Feb 18 14:48:36 Zecke: cool Feb 18 14:48:49 Zecke: ev.openembedded.org :) Feb 18 14:49:26 gerwinin: in Berlin we slightly talked about the possibilities we have: e.V. in germany, using SPI and Ltd. in England Feb 18 14:50:05 gerwinin: at least we pretty much agreed on the purpose and DO NOTS of such an association :} Feb 18 14:50:14 for the rest I failed to investigate (~lart time) Feb 18 14:50:36 Zecke: okay , so we should make the next step ? Feb 18 14:51:17 gerwinin: draw a matrix with the possible forms of the association as rows (e.V., SPI, Ltd.) Feb 18 14:51:24 Zecke: ok Feb 18 14:51:32 gerwinin: and check stuff like impeachment, liability, membership, needed work, etc. pp. Feb 18 14:51:41 Zecke: okay Feb 18 14:51:44 gerwinin: summarize them and put them up for discussion Feb 18 14:52:13 Zecke: Let me prepare this Feb 18 14:53:04 gerwinin: e.g. I'm the president of the assocation and I get less or more insane/nuts (whatever is worse) and will spend even more money on crack, how do you get rid of me. How do you prevent me spending money on bogus things :) Feb 18 14:53:36 Disclaimer: Dependant on your judgement crack might not be classified at bogus, in this case I sincerley apologize Feb 18 14:56:07 zecke: I will put some things in the paper Feb 18 14:57:13 gerwinin: but Sebastian is good to talk with. He is _the_ KDE marketing guy, also part of the head of KDE e.V. Feb 18 14:57:27 (even if I have met him only once) Feb 18 15:01:00 Zecke: Yes Feb 18 15:01:00 good mornin all Feb 18 15:02:59 I'm having some difficulty dealing with paths in a .bb recipie Feb 18 15:03:40 hillct: ask your questions, program outputs more than two lines go to pastebin Feb 18 15:03:42 ~pastebin Feb 18 15:03:46 well, pastebin is a place to paste your stuff without flooding the channel - try http://pastebin.ca, or http://channels.debian.net/paste, or http://rafb.net/paste/ Feb 18 15:04:10 the build is simple. The makefile is basic and does not change - it's fully formed in the unpacked source. Feb 18 15:04:35 I'm doing an oe_runmake in do_compile() Feb 18 15:04:50 I get the error, unknown target Feb 18 15:05:13 hillct: make sure your 'S' dir is the right one?! Feb 18 15:05:53 I added 'oenote ${PWD} above the oe_runmake and it appears to be the top level directory, not the source Feb 18 15:06:25 this build does not use automake or any of the other snazzier elements Feb 18 15:06:49 hillct: okay, enough monologue, answer the above question Feb 18 15:07:21 ok. I'll review all the paths, but what's the preferred method for changing the path in do_compile? simple cd or some other mathod? Feb 18 15:07:38 16:38 < zecke> hillct: make sure your 'S' dir is the right one?! Feb 18 15:08:01 will know in a moment Feb 18 15:08:13 hillct: do_compile is called from the dir inside the 'S' variable (see base.bbclass and bitbake.conf for the default value of S) Feb 18 15:08:39 K. soungd good Feb 18 15:08:39 checking Feb 18 15:12:29 Path is /home/slug/releases/slugos-3.10-beta/openslug-nslu2.tmp/work/ntpclient-2003_194-r0/ntpclient-2003_194 Feb 18 15:12:39 but Feb 18 15:12:47 that directory remains empty Feb 18 15:13:11 source is located in /home/slug/releases/slugos-3.10-beta/openslug-nslu2.tmp/work/ntpclient-2003_194-r0/ntpclient Feb 18 15:13:39 hillct: check bitbake.conf for the default value of 'S' Feb 18 15:14:03 hillct: which is package-name + -version. so you are free to set 'S' to something else inside your bbfile Feb 18 15:14:15 hillct: e.g. S = "${WORKDIR}/ntpclient" Feb 18 15:15:06 the bitbake config seems valid Feb 18 15:15:08 S = "${WORKDIR}/${P}" Feb 18 15:15:31 P = "${PN}-${PV}" Feb 18 15:15:33 hillct: check what 'P' is Feb 18 15:15:37 which seems fine Feb 18 15:15:44 hillct: hehe Feb 18 15:15:45 and works for other packages Feb 18 15:15:59 so I'll manually set it Feb 18 15:16:01 but Feb 18 15:16:12 yes? Feb 18 15:16:29 ntpclient vs. "ntpclient-2003_194" Feb 18 15:16:38 is there something 'wrong' that would cause the source not to be placed in the directory which includes the package version? Feb 18 15:16:42 yes Feb 18 15:16:44 the one who packaged the ntpclient did not put the version number into the base dir Feb 18 15:16:54 d'oh Feb 18 15:16:59 it's that simple Feb 18 15:17:01 ok Feb 18 15:17:11 hillct: Actually we would prefer if maintainers put the version number into the dir name, we can't enforce that though Feb 18 15:17:21 sure Feb 18 15:17:30 hillct: this is why you can set 'S' to point BitBake/OE to the real dir Feb 18 15:17:31 I was looking for some deper issue Feb 18 15:17:37 I should know better Feb 18 15:17:47 hillct: deeper than human behaviour? Feb 18 15:17:49 thanks Feb 18 15:18:06 deeper than author packaging siliness Feb 18 15:19:51 OT: Is the kernel stripped before it is zImaged? Feb 18 16:16:45 okay, after a week or two of tinkering i have OE making bootable images, kernels and our patched uboot for HG linkstations and kuroboxes Feb 18 16:17:07 * timtim wiggles Feb 18 16:17:16 it just booted first time!!!! Feb 18 16:17:20 congrats Feb 18 16:17:51 i guess i should read a manual somewhere before submitting stuff upstream Feb 18 16:18:17 or submit, read manual, resubmit Feb 18 16:24:14 i'll go with reading first. :) Feb 18 16:26:42 all i read was getting started hehe Feb 18 16:28:24 so any luck porting the sharp_mmcsd driver to 2.6? Feb 18 16:29:01 I have http://www.dmatech.org/alt/sharp_mmcsd_m.asm but that's all I have so far. Feb 18 16:30:59 dma`: is that like http://josejx.net/collie/ ? Feb 18 16:31:13 yeah Feb 18 16:31:43 As you might gather, I'm not very familiar with the arm instruction set. Feb 18 16:32:20 although IDA did some pretty nice work on that. Feb 18 16:39:30 is there a consistant method for testing whether an RRECOMMEND has been met, from within a postinit script block? Feb 18 16:42:06 hillct: dunno, the only thing I can think of is 'test' Feb 18 16:42:58 tough word on which to search the docs Feb 18 16:43:21 hillct: the shell command test Feb 18 16:43:31 oh Feb 18 16:43:42 ok Feb 18 16:43:44 hillct: you should ask yourself if RRECOMMEND is the right thing in this case though Feb 18 16:44:02 yeah Feb 18 16:44:28 I don't want to make cron a dependency. it's an alternative method for deploying ntpclient Feb 18 16:44:39 calling as a cron job offers additional functionality Feb 18 16:45:05 I may just remove the dependency completely, and ditch the recommendation Feb 18 16:45:34 hillct: test for /usr/bin/cron being present? Feb 18 16:47:35 yeah. that's one approach - probably the way to go. I was just thinking about testing for the installation of the package itself Feb 18 16:47:41 rather than a particular binary Feb 18 16:47:49 seems more flexible Feb 18 16:48:11 hillct: not if you have /usr/bin/cron hardcoded Feb 18 16:48:23 03freyther 07org.oe.dev * rc416a736... 10/ (1 classes/insane.bbclass): Feb 18 16:48:23 classes/insane.bbclass: Print the actual error, write a log file Feb 18 16:48:23 -Print all errors of the built packages and if we have printed Feb 18 16:48:23 any errors fail Feb 18 16:48:23 -Log the errors to a log file, so we can see them afterwards Feb 18 16:48:24 -Print the path (- the TMPDIR) of the package in question Feb 18 16:48:24 test -e `which cron` Feb 18 16:48:27 03freyther 07org.oe.dev * r7ebd0697... 10/ (1 classes/insane.bbclass): (log message trimmed) Feb 18 16:48:30 classes/insane.bbclass: Allow errors to be ignored Feb 18 16:48:32 Add a method to implement koen's skipping of Feb 18 16:48:34 errors of type .so in non dev package. This Feb 18 16:48:36 will turn into a list of known and acked issues which Feb 18 16:48:38 we decided to ignore. I do not yet know how to maintain Feb 18 16:48:40 such a list properly though Feb 18 16:48:58 which gives useful return codes? Feb 18 16:49:03 guess that'd work Feb 18 16:49:27 dunno, I#m not good at shell stuff Feb 18 16:49:48 hillct: but this is racy, someone could install cron after installing ntpclient ;) Feb 18 16:50:28 true Feb 18 16:50:50 in which case, it won't have used the cron-specific portion of the setup script Feb 18 16:51:08 but would execute the cron-specific portion of the prerm script Feb 18 16:51:12 yeah. it would be bad Feb 18 16:51:18 + bb.error("QA issue, koen give us a better msg!!!") Feb 18 16:51:19 decision made. Feb 18 16:51:21 Thanks Feb 18 16:51:22 * koen ponders Feb 18 16:51:43 koen: this was always there Feb 18 16:51:51 I know Feb 18 16:52:01 Better message: *** Out Of Cheese Error *** Feb 18 16:52:01 just got reminded of it Feb 18 16:52:33 ) Feb 18 16:54:14 hehehe Feb 18 16:55:07 zecke: can we make insane.bbclass behave like bug-buddy? Feb 18 16:55:26 should i add an enrty tot the bugtracker regarding uclibc shipping header files to the rootfs? Feb 18 16:55:45 zecke: e.g. report the issues to our QA server, which will do some dupe checking and file a bug? Feb 18 16:55:52 ifaistos_ppc: yes Feb 18 16:56:12 koen: BTW: miguel's blog entry (or the reference from the M$ engineer is funny) Feb 18 16:56:29 koen: like bug buddy? Crash and recursively exec bug-buddy? Feb 18 17:00:34 how can I start a framebuffer programm over a serial connection? Feb 18 17:01:02 hostname $ app Feb 18 17:01:10 just like any other app :) Feb 18 17:01:21 koen: thx Feb 18 17:01:29 mr_nice: cat /dev/urandom > /dev/fb0 Feb 18 17:02:21 that'll make an interesting display, methinks Feb 18 17:02:36 ) Feb 18 17:02:50 koen: actually this on my todo. I will continue on the checks for now Feb 18 17:02:52 matrix style Feb 18 17:03:03 zecke: I will try :) Feb 18 17:04:47 koen: did you try the angstrom image on the efika Feb 18 17:04:59 ifaistos_ppc: not yet Feb 18 17:06:01 changes to the network tab do not get applied Feb 18 17:09:00 sound also does not work Feb 18 17:18:14 Laibsch: PING Feb 18 17:18:20 hi guys Feb 18 17:21:46 hi Feb 18 17:22:00 hey hrw Feb 18 17:22:33 hey hrw Feb 18 17:22:49 hi hrw Feb 18 17:23:27 which device give to Laibsch? tosa or spitz... I need to send 'fosdem hardware insurance' mail and do not want to take 2 zaurii with me Feb 18 17:23:29 03pfalcon 07org.oe.dev * rdb6d8e18... 10/ (1 conf/distro/angstrom-2007.1.conf): (log message trimmed) Feb 18 17:23:29 angstrom-2007.1.conf: Ship just en-gb locale by default. Feb 18 17:23:29 * Ship just basic locale by default. Locales are big (~1Mb uncompr.), so Feb 18 17:23:29 shipping some adhoc subset will be still useless and size burden for Feb 18 17:23:29 users of all other languages/countries. Instead, worth to make it easy Feb 18 17:23:29 to install additional languages: installer/wizard + metapackages which Feb 18 17:23:36 will RRECOMMEND as much as possible content for a given language Feb 18 17:23:37 03pfalcon 07org.oe.dev * r98a9f60d... 10/ (1 conf/distro/angstrom-2007.1.conf): angstrom-2007.1.conf: Allow to set DISTRO_TYPE in local.conf. Feb 18 17:27:17 psokolovsky_: how big is an h3600 image after those changes? Feb 18 17:28:49 top - 19:02:19 up 347 days, 5:29, 10 users, load average: 12.87, 7.86, 4.18 Feb 18 17:29:12 koen, 16Mb straight (release) ;-) which is still too big, of course Feb 18 17:34:10 koen, task-base: Feb 18 17:34:10 ${@base_contains("COMBINED_FEATURES", "bluetooth", "${task-base-bluetooth-rdepends}", "",d)} \ Feb 18 17:34:10 ${@base_contains("DISTRO_FEATURES", "wifi", "${task-distro-wifi-rdepends}", "",d)} \ Feb 18 17:34:29 koen: hostname $ app don't work :(. I am getting a "Not running in a graphics capable console, and unable to find one". cat /dev/urandom > /dev/fb works fine. Feb 18 17:34:36 koen, what about making wifi COMBINED_FEATURE too? Feb 18 17:34:56 psokolovsky_: consult with RP or hrw Feb 18 17:35:05 ok Feb 18 17:35:18 hrw, ^^^ any thoughts? Feb 18 17:37:41 psokolovsky_: go ahead Feb 18 17:37:52 hrw, ok, thanks Feb 18 17:42:45 I have total redefinition of task-base in queue Feb 18 17:43:03 but will add any changes which will be created by you Feb 18 17:44:00 ok Feb 18 17:44:53 cu Feb 18 18:11:46 03pfalcon 07org.oe.dev * r342f7925... 10/ (1 packages/tasks/task-base.bb): (log message trimmed) Feb 18 18:11:46 task-base: Make "wifi" belong to COMBINED_FEATURES. Feb 18 18:11:46 * This is consistent with how "bluetooth" is handled already, and, assuming Feb 18 18:11:46 that distro supports wifi, essentially allows machine to decide if wifi packages Feb 18 18:11:46 go into rootfs or would be installed from feed. This is helpful for machines Feb 18 18:11:46 which don't have wifi and devoid of any means of external card extension, or Feb 18 18:11:46 for machines with very small ROM (like, for example, h3600 with its 16Mb, Feb 18 18:14:29 ok, with that change, h3600 jffs2 is 15872K. should be pretty good, hope gremlit[it] will try it now Feb 18 18:15:02 still a bit tight for jffs2, but at least it's flashable now Feb 18 18:17:53 psokolovsky_: I still think 'fuser' and 'setserial' shouldn't be in task base Feb 18 18:17:57 yes, I hope ;-). too bad I chicken to flash yet, would need working g_ether to try over nfs ;-) Feb 18 18:19:39 koen, hm, don't even know. I guess, fuser indeed better be pulled by deps only if needed. as for setserial, dunno about its usecases, still too green with using serial stuff on pdas ;-) Feb 18 18:20:06 it's as usefull for a user as gdb Feb 18 18:21:22 03pfalcon 07org.oe.dev * rfc57f4ef... 10/ (1 packages/angstrom/task-angstrom-x11.bb): Feb 18 18:21:22 task-angstrom-x11: Remove dependency on detect-stylus. Feb 18 18:21:22 * detect-stylus is no longer used in OE, which adopted convention of Feb 18 18:21:22 having the touchscreen device at /dev/input/touchscreen0 (for all 2.6 Feb 18 18:21:22 devices being created by udev rule). Feb 18 18:21:27 03pfalcon 07org.oe.dev * rd94ee355... 10/ (1 packages/angstrom/task-angstrom-x11.bb): task-angstrom-x11: Bump PR for latest change. Feb 18 18:23:14 just hope that openmoko isn't stuck with detect-stylus, and won't need to pull it back ;-I Feb 18 18:25:54 (I don't say openmoko uses it btw - just wonder if it does) Feb 18 18:26:05 no, they use the udev link Feb 18 18:41:27 XorA: Hello, could I ask you some asoc related questions? Feb 18 18:41:46 (if you have time) Feb 18 18:46:13 mmp: ask, so in the event one of the experts gets back to their computer and sees it, they can respond Feb 18 18:48:28 hillct: the problem is that there is no official way to contact persons behind ASoC and the only one I know and is here is XorA:). But good point, I almost forgot to not ask to ask. Feb 18 18:50:46 My problem is dapm_power_widgets function in soc-dapm.c, which should power up all neccessary widgets. The power-up sequence doesn't contain widgets of type 'switch', which are used here and there in ak4535 driver, so they don't get activated. Feb 18 18:50:58 Hi Feb 18 18:54:17 hillct: and moreover, I did not read your message precisely enough and partially misunderstood it. :) Feb 18 18:55:41 heh Feb 18 18:56:10 mmp: we have Xora and lrg here Feb 18 18:56:25 mmp: both get their food/pay if you ask them Feb 18 18:59:16 zecke: yes; I asked just XorA because lrg told me he will be not available these days. Feb 18 18:59:24 hehe Feb 18 18:59:57 * mmp should stop use semicolons in normal sentences/code more in python Feb 18 19:00:07 hehe Feb 18 19:00:13 I used to do that Feb 18 19:00:19 and printf statements Feb 18 19:00:22 scary Feb 18 19:00:31 :) Feb 18 19:01:00 There was one year when I ws dreaming in C on a regular basis Feb 18 19:04:05 hillct: computer sciences student? :) Feb 18 19:05:20 My degree is in finance actually Feb 18 19:05:24 * mmp is one of them Feb 18 19:05:32 this was 20 years ago though Feb 18 19:05:42 I wasn't getting laid much back then Feb 18 19:05:51 ahh, sorry :) Feb 18 19:05:59 so I didn't have much else to dream about Feb 18 19:06:07 don't be sorry. I was 11 years old Feb 18 19:06:10 hillct: so you switched to computer science :) Feb 18 19:06:29 no, I was just a geek as a kid Feb 18 19:36:58 anyone with access to armeb here? could you send me any small binary? Feb 18 19:42:06 what do you need? Feb 18 19:42:45 any binary or someting specific? Feb 18 19:45:05 oh. 'any'. K just a sec Feb 18 19:49:30 zecke: download an nslu2 ipkg if you need an arm(v5t)eb binary Feb 18 19:52:44 zecke: http://vmaster.scoophost.com/logrotate_armeb Feb 18 20:00:38 i think i have hit a multimachine problem Feb 18 20:01:39 it build for one machine ok, but when i try to build for a second it fails when creating rootfs Feb 18 20:01:48 http://www.pastebin.ca/362409 Feb 18 20:17:26 hillct: thanks Feb 18 20:17:37 np Feb 18 20:21:44 Ifaistos, one thing you can try is to kill Packages* files and retry Feb 18 20:23:01 g'day zecke Feb 18 20:23:08 hey master Feb 18 20:24:17 psokolovsky_ : Any idea where the problem is ? The multimachine.bbclass or in rootfs.bbclass ? Feb 18 20:25:06 Ifaistos, try what i say, and let's see if it helps Feb 18 20:25:51 Ifaistos: e.g check if you have these packages, and check the architectures of them Feb 18 20:26:02 Ifaistos: then check the used ipkg.conf and the defined set of archs Feb 18 20:26:45 psokolovsky_ : cleaned tmp/ rebuilding from scratch Feb 18 20:29:18 zecke : ok Feb 18 20:34:46 mmp: can you email linux@wolfsonmicro.com and Ill see if I can find answers at work tomorrow Feb 18 20:35:16 mmp: Im not really here just now Feb 18 20:36:48 XorA :) Feb 18 20:37:42 XorA: I actually solved few issues and found new ones; maybe I'll solve it by myself. Feb 18 20:38:29 thanks:) Feb 18 20:42:39 koen: do you feel like extracting an EABI image for me? Feb 18 20:42:48 koen: and offering a direct download? Feb 18 20:43:52 zecke: will we be showing openmoko hardware ? Feb 18 20:44:24 gerwinin: I think this is safe to say. Even if our phones might not arrive in time, mickeyl can bring something Feb 18 20:44:32 gerwinin: oh wait, I better not talk for mickeyl ;) Feb 18 20:44:40 koen: alternatively run this script http://rafb.net/p/p2oNb630.html Feb 18 20:44:45 koen: and paste the results Feb 18 20:45:09 zecke: I am writing now the pressrelease , also the nokia webpad will be shown ? Feb 18 20:45:26 this is not using OE :( Feb 18 20:45:38 zecke: okay Feb 18 20:46:05 gerwinin: we can show iPAQs, nslu2, zaurus, dev board Feb 18 20:47:01 dev board = ? Feb 18 20:47:39 zecke: EABI image: http://www.openembedded.org/~koen/a780/ Feb 18 20:47:54 oh wait I could bring a/the dreambox, and/or pxa270 development board Feb 18 20:48:24 efika runs angstrom (with some minor glitches) Feb 18 20:48:28 koen: hehe, you got me. I wanted an elf executable ;) Feb 18 20:49:42 and you can use an ATI vga + usb mouse + usb keyboard and connect it to an lcd for a "bigger" picture Feb 18 20:49:51 Zecke: how about a oe certification programm ? Feb 18 20:50:16 gerwinin: hehe, not yet. What would you certify? Enough insanity to use OE? Feb 18 20:50:33 zecke: http://www.angstrom-distribution.org/unstable/feed/armv5te/base/elfutils_0.108-r0_armv5te.ipk Feb 18 20:50:37 i think most us (here at least) of are "certifiable".... ;) Feb 18 20:50:46 koen: I'm too lazy for ar and tar Feb 18 20:51:34 dpkg -x /tmp ? Feb 18 20:52:45 koen: that would work, if I would have dpkg installed Feb 18 20:53:01 koen: I will simply not test EABI and assume it will work :) Feb 18 20:59:11 zecke: do you have a high resolution pic of oe logo for me ? Feb 18 20:59:25 gerwinin: no, just the svg ;) Feb 18 20:59:41 gerwinin: which I will have to search Feb 18 21:00:18 zecke: making the flyers now Feb 18 21:00:28 cruiseoveride: I have never done things like that, but, from what I know, there were some ipaqs equipped with Consumer IR port, so they could be used as remote controls. Feb 18 21:00:32 uhh, sorry:) Feb 18 21:00:40 bad window Feb 18 21:00:44 zecke: pressrelease sent for review Feb 18 21:02:28 which address did you send it to? Feb 18 21:04:19 gerwinin: what about changing "OpenEmbedded" to "The OpenEmbedded Project"? Feb 18 21:08:40 koen|away: ping? Feb 18 21:09:05 Koen okay Feb 18 21:10:45 zecke: http://users.dls.net/~mwester/unslung/busybox Feb 18 21:11:19 thanks Feb 18 21:11:24 IIRC, that should be a staticly-linked program. Feb 18 21:11:45 but not EABI? Feb 18 21:13:52 okay, my mini selfcontained ELF header parser http://rafb.net/p/AFDjgH20.html Feb 18 21:15:06 Don't think its EABI... Feb 18 21:15:21 good, that is what my parser told me :) Feb 18 21:15:37 OSVERSION 97, ABIVERSION 0, or the otherway around Feb 18 21:16:05 Ifaistos: can you test the parser for efika binaries? Feb 18 21:16:15 There's a useful utility called "wackelf" in the madwifi-ng source, that must be able to do some parsing as well - FYI. Feb 18 21:16:42 mwester: readelf can do it too, I just wanted to handle bytes in something like python Feb 18 21:16:53 zecke : which parser ? Feb 18 21:17:08 Ifaistos: the pastebin from above Feb 18 21:17:19 Ifaistos: http://rafb.net/p/AFDjgH20.txt python download binary Feb 18 21:17:49 zecke : ok Feb 18 21:21:09 Ifaistos: haha the last lines are missing ;) Feb 18 21:21:44 http://rafb.net/p/NwzloD80.html Feb 18 21:26:26 koen|away: where can I put my elf hdr parser? Feb 18 21:27:27 zecke : you want any file in particular ? Feb 18 21:27:37 Ifaistos: ppc file Feb 18 21:27:47 or an executable and a library Feb 18 21:28:06 ok Feb 18 21:30:42 gerwinin: did you find the logo? Feb 18 21:30:43 zecke : http://www.pastebin.ca/362540 Feb 18 21:30:47 gerwinin: and where did you mail it to? Feb 18 21:31:43 zecke: no Feb 18 21:31:56 Zecke: mailt it only to you, mickeyl and koen Feb 18 21:33:05 gerwinin: http://www.openembedded.org/repo/org.openembedded.dev/contrib/oe.svg Feb 18 21:33:15 gerwinin: which email addess did you mail it to? Feb 18 21:33:21 zecke : ok ? need anymore cause i am leaving in a couple of minuts Feb 18 21:34:01 zecke : to your gmx address Feb 18 21:34:22 Ifaistos: I hate pastebin.ca :) Feb 18 21:34:47 Ifaistos: thanks, have a good night Feb 18 21:35:07 zecke : you want them in rafb ? Feb 18 21:35:32 Ifaistos: on third reload it actually worked :) Feb 18 21:35:49 Ifaistos: at least machine and endianes was detected properly Feb 18 21:36:10 when a recipie includes 'inherit update-rc.d', do I then need to call update-rc.d from postinit and prerm or is it handled automatically? Feb 18 21:36:43 zecke : do you want me to send you some files to have localy for the tests ? Feb 18 21:37:01 Ifaistos: no, I'm fine for now :) Feb 18 21:37:34 hillct: that is handled automatically, you can/should specify the number and the runlevels Feb 18 21:37:53 hillct: check the update-rc bbclass Feb 18 21:38:06 K Feb 18 21:38:41 hillct: it has some weird syntax I can never remember ;) Feb 18 21:38:45 I was working from the example of the cron package which calls update-rc.d in preinit and postrm, but it seems that generates warnings about duplication Feb 18 21:39:04 * hillct reads up on it Feb 18 21:39:19 is the automation of that behavior a recent change? Feb 18 21:39:49 hillct: we have always been lazy, so it is around for almost forever Feb 18 21:40:34 night everyone Feb 18 21:42:52 laziness is a virtue Feb 18 21:47:08 zecke: yeah, it appears to be fully automated. Guess it's just the cron recipie that's broken Feb 18 22:00:57 gerwinin: the mail did not arrive, but I hope you found/can use the svg Feb 18 22:34:37 dma`: Don't bother. I've already done the whole driver. Someone else is currently writing a driver from the specs. Feb 18 22:35:22 Ah good. I take it you just did a clean room operation. Feb 18 22:35:27 dma`: Yep Feb 18 22:35:34 dma`: I have not written any code Feb 18 22:35:40 dma`: But I wrote the specs :) Feb 18 22:35:48 Chinese wall :p Feb 18 22:36:00 The guy writing the driver is making good progress. Feb 18 22:36:01 Yeah I noticed http://josejx.net/collie/index.php?wiki=FrontPage Feb 18 22:36:17 He's got it initializing and detecting cards\ Feb 18 22:36:27 The IRQs work properly, etc. Feb 18 22:36:35 Just a bit more for TX/RX and it should be working Feb 18 22:36:50 I've been prodding him to release the code, but he doesn't want to yet :p Feb 18 22:37:55 Well who needs TX/RX anyway... Feb 18 22:38:01 :p Feb 18 22:38:31 looks like there's code for CRC and detection of the presence of a card too Feb 18 22:38:50 I just threw it into IDA for fun. Feb 18 22:39:11 Yeah. Like I said, I have the whole thing translated already Feb 18 22:39:23 I hand translated to C Feb 18 22:39:52 The hardware specific stuff is in the wiki Feb 18 22:40:05 Everything else should already be handled by the mmc stack in the kernel Feb 18 22:40:16 Yeah it's far more complex than I expected. Feb 18 22:40:29 It's really not Feb 18 22:40:45 The driver itself is pretty bad though Feb 18 22:40:49 Lots of globals Feb 18 22:40:58 Lots of long sleeps :p Feb 18 22:42:05 Well the sharp one is pretty complex. Feb 18 22:42:52 Are you talking about the montavista one? Feb 18 22:43:15 No, the sharp one is what I'm talking about Feb 18 22:43:21 mickey|dinner: thanks for the struct hint, writing the parser with it took only a couple of minutes :) Feb 18 22:43:22 I didn't think it was that complex to be honest. Feb 18 22:43:50 dma`: any reason you prefered ida over objdump? Feb 18 22:43:50 Well it actually doesn't look complex at all. Just quite a few lines of code. Feb 18 22:44:01 zecke: I just had it lying around. Feb 18 22:44:13 03freyther 07org.oe.dev * r1d0feb92... 10/ (1 classes/insane.bbclass): Feb 18 22:44:13 classes/insane.bbclass: Add a small elf reader and check ABI,ARCH and Endian Feb 18 22:44:13 Add a small ELF parser and read OSABI, ABIVERSION, ENDIAN and MACHINE. This Feb 18 22:44:13 code compares TARGET_OS and TARGET_ARCH to a list of known versions. Currently Feb 18 22:44:13 we have values for linux and linux-uclibc for arm, armeb and powerpc. mips, Feb 18 22:44:13 i386, amd64, ia64 are missing. ABI wise ARM EABI is missing as well, but koen Feb 18 22:44:17 is likely to enter the values into the documented dict Feb 18 22:44:20 03freyther 07org.oe.dev * rd92fd0f1... 10/ (1 classes/insane.bbclass): classes/insane.bbclass: Be less verbose Feb 18 22:44:58 * NAbyss_ shudders at the mention of ELF parsers Feb 18 22:45:18 NAiL: written in python ;) Feb 18 22:45:28 Uh isn't there a library for that? Feb 18 22:45:38 ELF is a lovely format, it seems, but I only parse the well defined headers Feb 18 22:46:08 good night, I have to charge my macbook now Feb 18 22:46:10 zecke: It's hard to find the authoritative source of ABIs etc.. had to write an ELF/DWARF parser for one of my recent contracts.. slightly painful Feb 18 22:46:15 Seeyas Feb 18 22:46:50 looks like sco and binutils are authoritive Feb 18 22:47:17 later Feb 18 22:48:53 hmm never even noticed linux/drivers/mmc/mmc.c actually. Feb 18 22:52:03 heh Feb 18 22:54:28 so it's just a matter of writing a few functions and pointing to a few of them in mmc_host_ops and platform_driver structures. Feb 18 22:58:31 good night:) Feb 18 22:58:52 dma`: Yeah, pretty much. Feb 18 22:59:11 dma`: Like I said, he's got it mostly working already, just a bit more to do Feb 18 22:59:46 with oe, is it possible to depend on a specific version of a package? Feb 18 23:28:02 hi all. I've got a pretty basic question about _append functions in bitbake recipes under OE Feb 18 23:28:35 hola alguien escribe en español? Feb 18 23:29:03 Iberiana, yo. Feb 18 23:29:19 aaahhh por fin alguien Feb 18 23:29:23 I don't know if its ok to do it. But I speak Spanish. Feb 18 23:29:33 (To do it here, I mean) Feb 18 23:29:39 no se de donde es esta sala Feb 18 23:30:07 If I use the bitbake recipe chunk at http://rafb.net/p/QKSH2747.html Feb 18 23:30:08 De desarrollo, Openembedded, GNU/Linux. Es muy técnica. Tal vez no sea lo que buscas. Feb 18 23:30:22 I get 4:SyntaxError:' oenote "Performing ${DESCRIPTION} unpack append here "' Feb 18 23:30:27 bueno, si ute lo dice, debe ser Feb 18 23:47:52 no-one has come across that? Feb 18 23:56:18 jselect: Might be best to wait for one of the more serious bitbake hackers to wander in.. most of them are in Europe (UTC/CET), so they'll probably be around in approx. 8-9 hours Feb 18 23:56:46 I'd hazard a guess that 'oenote' isn't defined though Feb 19 00:02:47 Thanks NAbyss_, I've had a look and the odd thing is that oenote works if I use it in do_unpack but not if I use it in do_unpack_append Feb 19 00:03:06 all I can think of is that the base bbclass isn't being imported somehow Feb 19 00:04:56 Yeah, sounds possible.. I'm really not sure about namespaces etc. in bitbake Feb 19 00:17:53 ahhh it's not oenote at all Feb 19 00:18:30 it's the def for do_unpack not prepending bb.build.exec_func('base_do_unpack', d) properly Feb 19 00:18:38 fabulous, the errors get worse :) Feb 19 01:28:49 anyone have a .bb for r8187? (wireless card driver) Feb 19 01:52:22 silly question: I just built gpe-image. In the images directory, there's a bunch of files, which ones do what? I have a .rootfs.jffs2, .rootfs.summary.jffs2 rootfs.tar.bz2 modules.tgz (updater and zImage, but I know what thsoe are) Feb 19 02:28:16 weee. I just updated bitbake and the org.openembedded.dev monotone repository and now I can't compile glib Feb 19 02:28:31 fails with checking size of pthread_t... configure: error: cannot run test program while cross compiling Feb 19 02:51:36 I'm trying to add python to the oe/slugos-image by only changing the local.conf file. My trick that worked over in the NSlug build doesn't work for the current OE tips. Any ideas of what I should try? Feb 19 02:55:24 what did you use previously? Feb 19 02:56:03 from what I can see as a relatively new OE user is that you should set your preferred provider for virtual python (if there is one) **** ENDING LOGGING AT Mon Feb 19 03:00:00 2007