**** BEGIN LOGGING AT Tue Nov 06 02:59:56 2007 Nov 06 04:32:18 Is it just me, or should bitbake re-do a step if it gets a ^C in the middle? Nov 06 04:32:25 Just like make does. Nov 06 08:27:18 03koen 07org.oe.dev * rd6978786... 10/ (1 conf/distro/angstrom-2007.1.conf): angstrom-2007: stop mickeyl's baseless complaints about 'downgrading' from kdrive 1.4, angstrom 2007 has never used it Nov 06 08:30:47 koen|away: he complained where? Nov 06 08:32:43 http://hentges.net/tmp/logs/irc/livelogs/%23oe.livelog?lines=500#[20071106%2000:13:34] Nov 06 08:34:43 i am getting '..... arm-angstrom-linux-gnueabi/bin/ld: cannot find -lqte' when i run 'bitbake -b kdepimpi_2.2.7' could anyone tell me why this is? other than the obviuos reason as to the script cannot find it, there is no makeable package calle lqte or maybe libqte i have sucsesfully built qte qte-mt libqpe-opie Nov 06 08:35:26 koen|away: wierdo, does he not realise 1.4 is shafted so we couldnt use it :-D Nov 06 08:35:33 qte is not supported in angstrom Nov 06 08:35:51 oh right Nov 06 08:35:54 ok Nov 06 08:36:10 because it probably should use g++ as LD, not $(LD) Nov 06 08:36:31 ok Nov 06 08:37:24 XorA: he probably misunderstood http://lists.linuxtogo.org/pipermail/angstrom-distro-devel/2007-November/000849.html and how ?= works in bitbake Nov 06 08:42:13 morning Nov 06 08:42:33 Phoix: you really need to read wiki page about building opie in angstrom Nov 06 08:45:09 koen|away: BTW did you see the insanity on the last comment on http://bugs.openembedded.org/show_bug.cgi?id=3007 Nov 06 08:45:20 re Nov 06 08:45:28 Im not sure how to answer that, how the hell did the machine boot if he flashed a .ipk file Nov 06 08:47:44 yeah Nov 06 09:08:35 cbrake_away: around? when using GPIO pins.. have you tried the little tool provided by compulab to read gpio levels? Nov 06 09:13:56 how can i find out what branches are in the OE.mtn file? Nov 06 09:17:31 Phoix: http://monotone.ca/monotone.html Nov 06 09:17:40 thanks Nov 06 09:27:43 which is the most stable branch out of .dev .dreambox .packaged-staging Nov 06 09:30:07 hi all ! Nov 06 09:30:47 There seems to be a problem with gcc-packaging.inc when used on a powerpc with softfloat that puts things in /nof/ dir... Nov 06 09:31:43 steliosk: again? Nov 06 09:31:52 I thought that had been addressed? Nov 06 09:32:15 before it was at gcc-cross Nov 06 09:32:42 ah, my bad Nov 06 09:34:05 hm..which package contains the C std headers? Nov 06 09:36:55 ah.. linux-libc-headers i suppose Nov 06 09:37:15 steliosk: regarding the MPC8323E-RDB kernel, any chance of activating automatic module loading? Nov 06 09:37:38 no problem here Nov 06 09:40:53 basically: Nov 06 09:41:04 -# CONFIG_KMOD is not set Nov 06 09:41:05 +CONFIG_KMOD=y Nov 06 09:44:33 cbrake: let me know when you are around please.. seems like the 2.6.23 kernel could use a linux headers package.. Nov 06 09:45:41 gilligan_: like the one I added 3 weeks ago? Nov 06 09:46:46 koen: sometimes I hate my work ... firewall --> no mtn updates at work Nov 06 09:47:29 gilligan_: fetch snapshots then? Nov 06 09:47:41 gilligan_: or use git mirror? Nov 06 09:47:53 steliosk: I have a (although ugly) workaround that fixes gcc for powerpc. Nov 06 09:48:12 I just have to spend some more hours to find the real solution. Nov 06 09:48:27 But this is for hardfloat, which also breaks. Nov 06 09:48:40 hrw: yes right, i can fetch a snapshot.. still this is quite annoying to say the least Nov 06 09:48:56 Something is getting mixed up with the LIBGCC2_LONG_DOUBLE_TYPE_SIZE Nov 06 09:49:15 it is being set to 64 when long double is actually 128 bits. Nov 06 09:49:40 could be something similar for softfloat... Nov 06 09:50:26 hrw: I take it http://www.openembedded.org/snapshots/OE.mtn.bz2 is always the latest one Nov 06 09:51:41 refreshed every day via cron Nov 06 09:51:45 gilligan_: snapshots are generated daily Nov 06 09:52:20 okay nice Nov 06 09:52:27 EsbenH : Well the problem is different here. It compiles but OE fails to install as its does not expect /nof dir Nov 06 09:52:38 ok. Nov 06 09:52:52 do you only use softfloat? Nov 06 09:52:59 EsbenH : but what you say about the long double is interesting Nov 06 09:53:13 EsbenH : no efika builds with hard float Nov 06 09:53:41 try to build meta-toolchain for efika. Nov 06 09:53:58 I did, and it breaks on this long double bug. Nov 06 09:54:07 in general i prefer soft float for the cpu's without fp as its much much faster tahn the kernel emulation Nov 06 09:55:22 EsbenH : will do. Also post your patch to the OE bugzilla for reference Nov 06 09:55:53 Hmm.... I would prefer to post something a tad more clean. Nov 06 09:55:55 I think we will see a lot of *interesting* stuff with the SDK.... Nov 06 09:56:08 what are you thinking of? Nov 06 09:57:35 all these packaging/building issues comming up, means there is something warped Nov 06 09:58:03 or gcc shouldn't build in the first place at all Nov 06 10:01:29 That 64/128 bit issue looks like its pulling some host stuff during config and gets things mixed up Nov 06 10:02:17 steliosk: should I file a bug for the CONFIG_KMOD / MPC8323E-RDB change? Nov 06 10:02:53 no just pastbin the diff and i'll push it Nov 06 10:03:56 steliosk: http://www.pastebin.ca/763279 Nov 06 10:04:26 there's another change I think would be usefull : adding "pci" and "uboot" features to the MPC8323E-RDB machine definition Nov 06 10:04:57 usecase for "pci" : this feature is used as a switch for some madwifi-related features Nov 06 10:05:43 http://www.pastebin.ca/763281 Nov 06 10:13:38 steliosk: I've just switched back to u-boot 1.1.6, and it's considerably faster than 1.3.0-rc3 on initialisation, I wonder why.. Nov 06 10:19:57 gm Nov 06 10:22:16 hey likewise Nov 06 10:22:23 hey Jin^eLD Nov 06 10:33:08 hm.. seems like the compulab won't run at 512mhz even when set in armmon.. cbrake?! ;] Nov 06 10:35:23 jeremy_laine : not sure why, haven't tried the 1.3.0 Nov 06 10:35:50 jeremy_laine : trying to get the SDK in a working condition Nov 06 10:39:38 steliosk: where you present on ELCE/RTLW? Nov 06 10:42:30 what's the MACHINE_CLASS variable that is used in the distro confs? Nov 06 10:48:30 is there a nice way to override the BUILD_SYS generated by conf/bitbake.conf ? Nov 06 10:52:06 hey florian Nov 06 10:52:13 the metatoolchain thing almost worked out Nov 06 10:52:59 I got a bunch of those: http://pastebin.org/7206 Nov 06 10:54:50 hi all Nov 06 10:55:36 * florian looks Nov 06 10:59:16 BUILD_OS is set to a value not supported by config.sub on my system. Nov 06 10:59:23 it is set to cygwin_nt-5.1 Nov 06 10:59:34 the '-' confuses config.sub Nov 06 11:00:03 config.guess works correctly though.... Nov 06 11:08:18 Jin^eLD: looks like binutils-cross-sdk packaging is bad... Nov 06 11:08:50 aha.. so the problem is actually with the binutils-cross-sdk .bb file? Nov 06 11:11:00 likewise : nop. i am grounded till baby comes :) Nov 06 11:14:13 ~fish lrg Nov 06 11:14:16 * ibot slaps lrg around with a large trout Nov 06 11:21:43 broonie: is lrg in work today? Nov 06 11:23:02 XorA: Working from home. Nov 06 11:23:27 broonie: bah Nov 06 11:23:44 XorA: But that was because he was off to get some jabs this morning so he may not be around right now. Nov 06 11:23:57 broonie: ah, he off abroad again Nov 06 11:24:31 XorA, broonie: hey Nov 06 11:24:44 hey lrg, msged yah Nov 06 11:30:07 lrg: Morning. Nov 06 11:35:04 * * OE Bug has been RESOLVED by jeremy.laine(AT)bolloretelecom.eu Nov 06 11:35:06 * *  Enable u-boot autoboot on MPC8323E-RDB Nov 06 11:35:08 * * http://bugs.openembedded.org/show_bug.cgi?id= Nov 06 11:40:20 lego catalogue arrives Nov 06 11:41:12 Jin^eLD: yes right, looks like the file mask for the debug package is wrong or missing. Nov 06 11:41:24 * florian is slow today... Nov 06 12:02:34 re Nov 06 12:02:47 florian: oh ok, I think I may be able to figure that out Nov 06 12:03:08 allthough if you do it, you could commit the changes... Nov 06 12:05:47 Jin^eLD: I don't think i'll manage to do this today - but as soon as I need it I will run into that and fix it if noone else did it so far. Nov 06 12:07:27 ok, let's see, if I am the first then I will open a bug and add a patch Nov 06 12:07:29 :) Nov 06 12:12:05 someone has working SRCREV for openocd? Nov 06 12:12:37 ops - openmoko insane it Nov 06 12:46:33 silicon is getting dirty cheap -> http://www.wirelesslan.gr/product_info.php?products_id=973&osCsid=e9ad2aa0fa2f4d1adce85ad20433c97e Nov 06 12:46:56 steliosk: req login.. Nov 06 12:47:50 hrw : http://www.wirelesslan.gr/index.php?cPath=127 Nov 06 12:48:05 fck. go to www.wirelesslan.gr Nov 06 12:48:17 and on the front page there are 2 ixp based wifi routers Nov 06 12:48:44 one is 89E the other 139E Nov 06 12:48:52 nice Nov 06 12:49:37 steliosk: which minipci card (chipset/driver) works stable as AP? Nov 06 12:50:01 steliosk: I am tinking about making router/wifiAP from alix1c Nov 06 12:50:33 If you are looking for 802.11b 2.4Ghz/10mbit nothing beats old prism cards Nov 06 12:50:44 for the rest anything with Atheros Nov 06 12:50:59 atheros/madwifi Nov 06 12:51:05 I would prefer b/g something Nov 06 12:51:28 use atheros then Nov 06 12:51:34 ok Nov 06 12:52:15 if possible switch to 802.11a 5Ghz, provided that you don't have too thick walls you can get up to 30-35mbits transfer rates Nov 06 12:52:31 libaio and coreutils need updating in org.openembedded.dreambox ... i have taken the versions from .dev branch and placed them into my .dreambox branch directory .. as i am trying to bitbake world the org.openembedded.dreambox branch using distro Angstrom Nov 06 12:52:54 i.e advertized wifi speed <> actual throughput Nov 06 12:53:21 Phoix : Could luck :) Nov 06 12:53:27 :) Nov 06 12:54:14 steliosk: my router do not do 11a currently Nov 06 12:55:00 hrw : you avoid gloiwing too much at night then ;) Nov 06 12:56:00 * steliosk wonders what the cumulative effects of all the wifi/bluetooth/gsm/dect will have on him after 20 years Nov 06 12:56:45 probably will be glowing at night with all that HF radiation we are absorving Nov 06 12:57:51 Phoix : Well, we (as in i am ) are supposed to get dreambox merged with Angstrom Nov 06 12:58:59 steliosk: I have a friend working on driving small embedded things only using the antenna to catch energy from all the radio waves :) Nov 06 13:00:42 CM : heh if he comes close to our office probably he wan't even need an antenna :D Nov 06 13:00:53 steliosk: Hehe Nov 06 13:01:17 steliosk: cool Nov 06 13:01:55 CM: really .. thats cool Nov 06 13:02:54 Phoix: Clocks and such is no problem, even a small radio or so is ok when using headphones Nov 06 13:04:16 its something like the technology that the aliens have in independance day where the mothership supplies all the power through the air to the other ships.. hehe Nov 06 13:04:47 Something like that ;) Nov 06 13:07:35 have any of the branches ever succsesfully been built as world? Nov 06 13:08:46 * Phoix is wondering how long building world on a core 2 quad 2.2ghz will take Nov 06 13:09:04 I doubt it will run through at all Nov 06 13:09:06 :) Nov 06 13:09:11 Phoix : World is a big thing... Nov 06 13:09:13 :) Nov 06 13:09:33 actually it would be cool to setup buildbot to do that once in a while Nov 06 13:09:51 Poky autobuilder builds world incrementally Nov 06 13:09:57 starts each saturday from scratch Nov 06 13:10:04 Phoix : Try building angstrom for x86.... at least you will get an image you can use on a pc :) Nov 06 13:10:24 yeah Nov 06 13:10:40 i may try that out Nov 06 13:10:49 didn't think of that Nov 06 13:11:28 Phoix : For other powerpc targets that work try dht-walnut,efika mpc8323e-rdb Nov 06 13:11:54 ok Nov 06 13:12:22 time to go. later all Nov 06 13:12:34 later Nov 06 13:24:23 hi cbrake Nov 06 13:30:41 gilligan_: hello Nov 06 13:31:10 gilligan_: I have never tried the compulab gpio utility -- I wrote my own which just maps the GPIO regs into user space, etc. Nov 06 13:32:10 what is world, as in once world is built what does it become? Nov 06 13:40:12 cbrake: pxaregs could not be used? Nov 06 13:40:38 hrw: I must not know about that one :-) Nov 06 13:40:41 * cbrake googles ... Nov 06 13:40:49 cbrake: "bitbake pxaregs" Nov 06 13:41:22 hrw: guess I re-invented the wheel on that one Nov 06 13:41:32 happens Nov 06 13:43:59 koen, I am at http://www.sdrforum.org/pages/sdr07/forumMeeting_sdr07_main.asp Nov 06 13:47:34 yo Nov 06 13:49:35 cbrake: I see.. what do you mean by mapping gpio registers into user space though? Nov 06 13:49:53 gilligan_: see the pxaregs app hrw mentioned -- it is a good example Nov 06 13:50:41 gilligan_: allows you see/change gpios (or any pxa reg) from a user space app Nov 06 13:50:59 cbrake: oh, nice Nov 06 13:52:11 cbrake: I would just have customized the stuff written by the compulab guys (also looking at the gpio stuff in the kernel helps) .. but no need for me to ALSO reinvent the wheel i suppose :) Nov 06 14:15:17 cbrake: ah okay.. that pxaregs is quite useful (and looking at the source code it is also pretty much straight forward) Nov 06 14:16:11 gilligan_: nod Nov 06 14:27:38 gilligan_ which card of compulab do you use ? Nov 06 14:32:52 Genesis: the cmx270 (the w model) Nov 06 14:33:23 i look Nov 06 14:33:27 i work on CM-iGLX Nov 06 14:34:05 weird card compulab i think Nov 06 14:35:00 has somebody ever used the usbvision or bttv drivers on ARM ? Nov 06 14:35:24 compulab says they wont do any further development on the cm-x270 board Nov 06 14:35:30 for so-called performance reasons Nov 06 14:35:55 I guess the biggest part of the performance issues they cite come from the FPA-enabled distro they ship Nov 06 14:36:36 on our card , nand flash are connected trough ... usb . Nov 06 15:13:58 morning Nov 06 15:45:33 hmm that's odd Nov 06 15:46:02 I did a pull && update rebuild my rootfs (with angstrom) and now I get nfs server timeouts when booting the rootfs via nfs Nov 06 15:46:14 works fine with 1.5-2 weeks old build/checkout Nov 06 15:46:25 I wonder what could be responsible.... Nov 06 15:46:45 no kernel modules involved, same kernel, so its really something with the rootfs Nov 06 15:50:50 mysql does not build for angstrom Nov 06 15:51:28 there is a problem with the gen_lex_hash.patch Nov 06 15:51:45 and after generating the lex_hash.h fil;e it still wont build Nov 06 15:53:05 * * OE Bug 868 has been RESOLVED (FIXED) by Nov 06 15:53:07 * *  initial support for vmware virtual machines Nov 06 15:53:09 * * http://bugs.openembedded.org/show_bug.cgi?id=868 Nov 06 15:55:15 mhm, I guess I just screwed up, as always.. Nov 06 15:55:24 Jin^eLD: what are you running Angstrom on? Nov 06 15:55:28 heh Nov 06 15:55:45 Phoix: a dev board that is "under construction" so to speak Nov 06 15:56:01 basically using the minimal image for a start Nov 06 15:56:32 what arch is it? Nov 06 15:56:38 arm Nov 06 15:56:44 cool Nov 06 15:57:05 :) Nov 06 15:57:14 what are you up to? Nov 06 15:58:12 im kinda new to this, but am attempting to compile in slackware 12 packages/Angstrom distro for my SL-C3200 Nov 06 15:58:21 am running into lots of problems Nov 06 15:58:39 03florian 07org.oe.dev * rc5d7d939... 10/ (4 files in 4 dirs): vmware.conf: New machine definition and defconfig for vmware disk images. Fixes #868 Nov 06 15:58:52 aha.. zaurus Nov 06 15:59:01 and because im not too good at programming im not too good at solving the problems Nov 06 15:59:04 yeah :) Nov 06 15:59:19 but i enjoy tinkering around with this sort of stuff Nov 06 15:59:53 re Nov 06 16:00:25 hey zecke Nov 06 16:01:05 re? Nov 06 16:01:26 zecke: assuming I want to start of angstrom and customize, can I add a require angstrom-2007.1.conf in my custom-distro.conf and just override a couple of values? Nov 06 16:01:46 I did not want to just copy or touch the original angstrom conf file Nov 06 16:02:09 but my require approach did not quite work out so I wonder if it was supposed to.. Nov 06 16:03:00 Jin^eLD: yeah, see openmoko as example Nov 06 16:03:28 I did.. Nov 06 16:03:45 for starters the only thing I did override was the distro name, just to check if it works; and the build itself did Nov 06 16:03:45 What is this IBM RS6000 thingy? Nov 06 16:03:46 but Nov 06 16:03:52 my rootfs was bad Nov 06 16:04:03 am I really supposed to be including rs6000.h in my ppc603e gcc builds? Nov 06 16:04:04 EsbenH: in oe?! Nov 06 16:04:14 well, in gcc. Nov 06 16:04:23 i.e. when booting from nfs I got timeouts and such.. uuh... because probably other packages had some _angstrom things in them? Nov 06 16:04:56 EsbenH: That's what IBM P-Series is now... PPC based machines - nice hardware some time ago. Nov 06 16:05:28 yes, but is GCC supposed to be including rs6000.h when building a xcompiler for ppc603e? Nov 06 16:05:42 Jin^eLD: if i was to go buy some books to get some better knowledge of whats being done here, what books do you suggest i should start reading first, or :) is the answer anything and everything to do with architechtures and linux? Nov 06 16:06:10 Phoix: well, I guess the question is - what do you want to do? Nov 06 16:06:20 I do not think there is an OE book around Nov 06 16:06:42 actually all my OE knowledge comes from being in this channel and asking people for help:) Nov 06 16:07:20 i guess i want to be able to tinker with OE but know what i am doing, Nov 06 16:07:26 oh right, cool Nov 06 16:07:26 i just fckd my uboot enviroment vars..... Nov 06 16:07:43 Phoix: I still did not reach that point ;) each time I think I know what I am doing I find something new ;) Nov 06 16:08:09 EsbenH, the RS6000 is the grandfather of all powerpc :) Nov 06 16:08:09 Jin^eLD: hmm, I don't know if we have the DISTRO_NAME in OVERRIDES but we probably have Nov 06 16:08:20 zecke : hi ! Nov 06 16:08:39 zecke: what does OVERRIDES do.. what exactly does it override and where? Nov 06 16:08:43 steliosk: so my gcc/tm.h which includes rs6000.h is not broken? Nov 06 16:09:01 EsbenH : nop. its normal Nov 06 16:09:15 Jin^eLD: see the bitbake manual. This allows us to apply a patch only when XYZ (machine is that, distro...) Nov 06 16:09:32 ok, then I am one step closer. the wrong LIBGCC2_LONG_DOUBLE_TYPE_SIZE is set by rs6000.h Nov 06 16:09:34 aha Nov 06 16:09:38 Jin^eLD: openmoko.conf cheats and puts OVERRIDES = "local:${MACHINE}:${DISTRO}:angstrom:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}" into the OVERRIDE to get all angstrom patches applied Nov 06 16:09:53 ook, now I know what I missed Nov 06 16:10:00 well. should I cheat the same way? Nov 06 16:10:09 Jin^eLD: e.g. SRC_URI_arm = "file://patch-is-only-for-arm;patch=1" means when TARGET_ARCH is ARM apply the patch Nov 06 16:10:24 this openmoko cheat looks a little hack'ish Nov 06 16:10:39 yeah, that I learned already, the _something thing Nov 06 16:10:56 but I missed the OVERRIDES in openmoko.conf, or - I did not understand what it does Nov 06 16:10:59 so now it makes sense of course Nov 06 16:11:58 well.. so should I do the same or is there a nicer way? Nov 06 16:20:08 steliosk: hmm..... Nov 06 16:20:37 where did the ppc-gcc-41-20060515.patch from gcc-4.1.1 go in gcc-4.1.2 ? Nov 06 16:20:48 was it actually merged upstream, or are we missing it? Nov 06 16:22:38 it does not seem to be in gcc-4.1.2 Nov 06 16:23:51 Jin^eLD: you don't have much of a choice Nov 06 16:24:07 what are those fail-fast things Nov 06 16:24:18 there are lots of things there Nov 06 16:24:43 is it ok to copy the openmoko OVERRIDES as it is to pretend to be Angstrom? or do I need less stuff in there? Nov 06 16:36:58 bye Nov 06 16:39:35 Esben : well its partially merged and that's why the ppc's without fp still use gcc 4.1.1 Nov 06 16:56:45 Why doesn't bitbake recognize when I've edited configure.ac and re-configure a package when I tell it to build? Nov 06 17:01:17 nelson: bitbake has stamps Nov 06 17:01:33 nelson: it has a stamp that do_configure on the .bb was executed Nov 06 17:01:50 nelson: imagine bitbake would have to track everyfile in $WORKDIR on every parse... people would complain badly Nov 06 17:05:33 I'd rather have it work correctly than be quick but wrong. Nov 06 17:05:59 anyway ... I can rm the stamp. Where is it? Nov 06 17:08:52 nelson: lol Nov 06 17:09:00 nelson: bitbake -h and search for 'force' Nov 06 17:09:43 that's too hard. Is there some place I could point and click? Nov 06 17:10:05 nelson: no idea ;) Nov 06 17:10:33 nelson: if you change a file used by configuration of a package (it is not always configure.ac....) you can just call bitbake foo -cconfigure -f afterwards... Nov 06 17:10:59 nelson: or if you are in the ${S} just use the ../temp/run.do_configure* script to reconfigure Nov 06 17:22:05 * * OE Bug 3264 has been created by eha(AT)doredevelopment.dk Nov 06 17:22:07 * * linux-libc-headers depends on gettext-native Nov 06 17:22:09 * * http://bugs.openembedded.org/show_bug.cgi?id=3264 Nov 06 17:25:36 zecke: is it somehow possible to achieve the following behaviour... as you know, I can define my own site directory with an own site/conf/* etc. and tell bitbake to look for it there and that works fine; I can also have my collections and it will look for packages there, works fine too; now: Nov 06 17:25:46 there are certain packages that have subdirectories with some custom stuff Nov 06 17:25:52 custom meaning - depending on the distro Nov 06 17:26:11 can I somehow add my custom stuff separately? Nov 06 17:26:27 so far it only works if I cp -a the whole directory for that package to my collection Nov 06 17:26:53 because then I am overriding the "default" package completely... but actually I only need one small addition, without even changing the .bb file Nov 06 17:26:55 do you know what I mean? Nov 06 17:30:20 Jin^eLD: you have to copy... Nov 06 17:31:24 ok.. thx Nov 06 17:32:40 Jin^eLD: It isn't possible to have custom versions from collections pulled in atm and its shortcomming of collections Nov 06 17:33:11 Jin^eLD: I'd love to see that fixed and I have a few ideas but they are going to slow bitbake down :( Nov 06 17:33:26 :( Nov 06 17:33:53 well, no big deal Nov 06 17:34:11 just a little more difficult to sync, i.e. not to forget to copy over from OE trunk to collection Nov 06 17:34:29 but apart from that I worked out a very nice setup to keep local changed stuff separate from the OE tree Nov 06 17:41:38 RP: could you verify the git tree? I need to get some space and would remove my copy here **** BEGIN LOGGING AT Tue Nov 06 17:51:34 2007 Nov 06 18:00:48 zecke: ok, will have a look now Nov 06 18:04:28 hi all, I want to change the default kernel for simpad from 2.4 to 2.6. I wonder if I should just change it in machine/simpad.conf to 2.6 and not mention at local.conf.sample. or if I need to hint that if you want a simpad 2.4 kernel you need to set 2.4 for MACHINE_KERNEL_VERSION Nov 06 18:05:33 i ask because i want to solve bug 2973 Nov 06 18:08:44 mr_nice: Is the 2.6 kernel able to replace 2.4 completely on Simpads? Nov 06 18:09:39 If yes 2.4 is obsolete for this platforms and does not need to be mentioned any more... Nov 06 18:11:17 florian: 2.4 is no longer maintained. 2.6 should be able to replace 2.4 completely. 2.4 is better tested. Nov 06 18:11:52 florian: so I will not mention it at local.conf.sample but leave the ability to use 2.4 as value Nov 06 18:11:58 ? Nov 06 18:12:13 mr_nice: I'd make 2.6 default and just let the 2.4 bb live Nov 06 18:12:17 yeah Nov 06 18:12:30 florian: ok, thx Nov 06 18:21:38 RP: looks right, these are the stubs Nov 06 18:21:48 RP: and I still plan... but I'm so lost ATM Nov 06 18:22:21 zecke: Can you mail mallum about this so I/he knows what to do with it :) Nov 06 18:22:35 RP: on my list :) Nov 06 18:22:40 zecke: ok Nov 06 18:23:03 zecke: Anyhow, I've confirmed I have it which is what you mainly wanted to know. Nov 06 18:23:33 thanks Nov 06 18:26:13 * steliosk is trying to debrick a board uploading 16MB over a 33600kbit serial line with kermit protocol Nov 06 18:26:46 * steliosk takes a mental note never to erase uboot before testing that the tftpboot server is *really* working Nov 06 18:50:32 moin Nov 06 19:30:37 http://linuxdevices.com/news/NS3871478989.html looks cute Nov 06 19:31:52 kergoth: true. Eclipse based editor on the bottom. nice. Nov 06 19:32:24 yay lego Nov 06 19:32:45 that could be fun to play with, though i don't want to know how much it'll cost Nov 06 19:35:00 i just installed bitbake, forgetting about the recommendation not to install it and now whenever i run it i get just errors the main recurrence of 'local variable 'bb' referenced before assignment' anyone know how to resolve this issue? Nov 06 19:35:27 kergoth: heh - yeah, but a pretty clueful idea Nov 06 19:36:12 Phoix: I doubt that has to do with installed vs. not-installed, did you change your config files or BBPATH? Nov 06 19:36:54 no Nov 06 19:37:08 i may have changed bbpath actually Nov 06 19:37:19 does that have an impact? Nov 06 19:40:00 kergoth: It does look like fun :) Nov 06 19:44:34 khem: Good day Nov 06 19:45:24 hey likewise Nov 06 19:45:26 whats up Nov 06 19:45:34 * Khem is sneezing Nov 06 19:46:47 cheers Nov 06 19:47:25 khem: everything fine here, you? Nov 06 19:47:43 just a bit sick else all ok :) Nov 06 19:48:10 I have installed gutsy so will try a fresh OE build Nov 06 19:48:19 today Nov 06 19:48:30 khem: I *just* finished my gutsy install yesterday - am still installing everything I am missing Nov 06 19:48:44 went x86 instead of x86_64 this time round. Nov 06 19:48:56 ah ok I have to setup the OE rest is completed :) Nov 06 19:49:07 so far mtn is working Nov 06 19:49:07 There is this URL you can add the Synaptic. Nov 06 19:49:23 It installs openembedded-essential: everything needed Nov 06 19:49:34 ah I remember that Nov 06 19:49:35 cool Nov 06 19:49:41 its somewhere in wiki Nov 06 19:50:01 khem: deb http://www.openembedded.org/dl/ packages/ Nov 06 19:50:25 http://www.openembedded.org/wiki/OEandYourDistro, look for Debian. It still works for every Ubuntu I have seen. Nov 06 19:50:35 monotone is now 0.35, good enough. Nov 06 19:51:15 The wiki said 0.36 was required... Nov 06 19:51:25 hmmm Nov 06 19:52:32 no, but 0.36 is faster Nov 06 19:52:43 http://www.openembedded.org/snapshots/ Nov 06 19:53:52 likewise: is it considerably fast ? Nov 06 19:54:14 likewise: So that I can look for a .deb for 0.36 Nov 06 19:54:40 http://www.monotone.ca/downloads/0.36/ Nov 06 19:54:44 but I dont do so much mtn ops so prolly once a while suckage is acceptable Nov 06 19:54:44 0.37 is already out, I now see. Nov 06 19:55:03 oh they have debs too cool Nov 06 19:55:08 03Jeremy_Laine 07org.oe.dev * r043837f9... 10/ (1 packages/linux/linux-2.6.23/mpc8323e-rdb/defconfig): packages/linux/linux-2.6.23/mpc8323e-rdb/defconfig : Enable auto module loading Nov 06 19:55:16 03ifaistos 07org.oe.dev * rfa812c3e... 10/ (1 packages/linux/linux_2.6.23.bb): packages/linux/linux_2.6.23.bb : Bump PR Nov 06 19:55:19 will 0.37 work with oe Nov 06 19:55:21 03Jereme_Laine 07org.oe.dev * r0dfbd216... 10/ (1 conf/machine/mpc8323e-rdb.conf): conf/machine/mpc8323e-rdb.conf : Add pci and uboot to MACHINE_FEATURES Nov 06 19:56:44 Khem: usually it keeps working, as the network sync interface is often identical or compatible. Database formats might change, but are forwardable through the monotone executbale. Nov 06 19:57:03 no debs for 0.37 Nov 06 19:57:07 yet Nov 06 19:57:15 there are statically linked binaries to test. I will try soon. No, packages take time to appear. User contributed. Nov 06 19:57:40 * RP just built monotone 0.37 Nov 06 19:58:17 RP: cool any significant differences you find let us know too Nov 06 19:58:23 Khem: do you know if the eglibc project keeps the same license if glibc goes GPL v3? This question was raised during the ELCE/RTLW event in Euope Nov 06 19:58:51 Khem: I'm upgrading from 0.2x in the case of one machine and 0.33 on the other :) Nov 06 19:58:56 LOL Nov 06 19:59:05 Giant Leaps Nov 06 19:59:23 likewise: yes generally eglibc follows FSF copyrights but I am not a lawyer so dont sue me if I am wrong. Nov 06 19:59:36 I can bring this up in consortium meeting Nov 06 19:59:40 khem: trust me, I won't. Nov 06 19:59:44 They'll have to if they want to keep pulling stuff from glibc. Nov 06 20:00:10 Khem, broonie: yes, but there was concern in the embedded market (read: Sony) about his. Nov 06 20:00:17 this. Nov 06 20:00:28 have to go now. Thanks for all the fish. Nov 06 20:00:31 likewise: about LGPL v3 ? Nov 06 20:00:59 Khem: not sure, the question was a bit vague. Nov 06 20:01:03 RP: Do I have to get a new snapshot to start with if I upgrade to mtn 0.36? Nov 06 20:01:16 I don't know if a v3 of LGPL even exists, I am not a lawyer myself. Nov 06 20:01:24 And I won't sue you :-) Nov 06 20:01:31 Khem: I think the last db format changes were around 0.33 so probably not Nov 06 20:01:35 likewise: hehe :) Nov 06 20:01:48 * Khem trusts RP Nov 06 20:02:27 RP: I have a strange issue which sounds a arm kernel bug do u have any armv6 hardware Nov 06 20:02:29 yes, even in real person RP seems trustworthy :-) Nov 06 20:03:01 Khem: I do have an armv6, yes Nov 06 20:03:03 khem: read the gcc bug about conditional instructions? Nov 06 20:03:24 likewise, what is bug # Nov 06 20:03:26 likewise: :) Nov 06 20:03:41 sorry have to go now, it was on LKML. Nov 06 20:03:46 just yesterday Nov 06 20:03:48 cya Nov 06 20:04:01 likewise: OK Nov 06 20:04:42 RP: I can give you a testcase and see if it fails for you. What kernel are u using on this Nov 06 20:05:49 * Khem 's lunch is sponsored by boss today Nov 06 20:05:51 Khem: Its an n800 Nov 06 20:06:12 RP: ah I think its close to my ref board I have a 2430sdp Nov 06 20:06:21 n800 is 2420 based Nov 06 20:09:56 RP: you familiar with the memory mangerment stuff in kernel Nov 06 20:10:08 Khem: More than average Nov 06 20:10:18 RP: Cool so I can bug you then Nov 06 20:13:42 what is latest production bitbake? 1.8.9 ? Nov 06 20:14:36 how do i clear bitbakes cache becase i think my error is coming from the fact that my cache is invalid then it tries to rebuild it and then just spits out tunns of compile errors Nov 06 20:15:09 Phoix: unlikeley Nov 06 20:15:20 Phoix: rm ${TMPDIR}/cache? Nov 06 20:15:32 cool, thanks Nov 06 20:15:39 hopefully that works Nov 06 20:15:55 Phoix: your issue is somewhere else, most likely in your local.conf where you do a recurring assignment... Nov 06 20:16:31 i see.. ok Nov 06 22:07:47 re Nov 06 22:08:10 re Nov 06 22:11:48 hi all ... Nov 06 22:12:40 someone have dettailed and verified spec about nokia N810 ? Nov 06 23:21:34 03koen 07org.oe.dev * r85dec212... 10/ (1 classes/rm_work.bbclass): rm_work: only remove ${S}, leave temp and install alone Nov 07 00:01:34 Zecke: i forgot to say sorry :) Nov 07 00:01:59 gerwinin: sorry for what? Nov 07 00:02:25 Zecke: for the accomodation thing Nov 07 00:02:32 gerwinin: I'm over it Nov 07 00:02:48 Zecke: sent today the gerbers for the cheap arm board Nov 07 00:02:58 gerwinin: Eindhoven sucks, the train station closes at 3 o'clock Nov 07 00:03:09 Zecke: true it does Nov 07 00:03:22 Zecke: I am also thinking to leave eindhoven Nov 07 00:03:44 Zecke: but for one of other reason it seems that getting here is more easy than leaving again Nov 07 00:04:15 zecke: it was really cool to see you and knut :) Nov 07 00:04:15 hehe, it is not too bad... Nov 07 00:04:30 gerwinin: sorry, I'm soooo tired :) Nov 07 00:04:52 zecke: good night :) need to work some with the sdcc compiler :( Nov 07 00:05:33 gn Nov 07 01:40:38 I'm getting fetch failures where the fetch fails while the file exists and is in my sources directory. every week it seems to be a different problem file, quite odd. Nov 07 02:50:42 hmmm openembedded-essential does not include help2man **** ENDING LOGGING AT Wed Nov 07 02:59:57 2007