**** BEGIN LOGGING AT Tue Aug 26 02:59:57 2008 Aug 26 03:26:06 * * OE Bug 4522 has been created by xydarcher(AT)163.com Aug 26 03:26:08 * * pango.module file generation problems. Aug 26 03:26:10 * * http://bugs.openembedded.net/show_bug.cgi?id=4522 Aug 26 05:49:06 * * OE Bug 4523 has been created by  Aug 26 05:49:08 * * poodle support for linux-kexecboot Aug 26 05:49:10 * * http://bugs.openembedded.net/show_bug.cgi?id=4523 Aug 26 06:54:00 good morning Aug 26 06:58:39 moring Aug 26 07:17:47 hi Aug 26 07:32:21 morning all Aug 26 07:32:33 thesing: wb Aug 26 07:34:06 thesing: during your absence (holiday hopefully), me and hvontres did extend kexecboot (oebug 4520, 4523) Aug 26 07:34:22 now I see an isue: the defconfigs.... Aug 26 07:34:51 most of them are lacking +CONFIG_BLK_DEV_INITRD=y +CONFIG_INITRAMFS_SOURCE="initramfs.cpio.gz" Aug 26 07:35:03 ant_work: yes I was on holiday. Aug 26 07:35:08 :-) Aug 26 07:37:51 ant_work: I'm working trough my mail now. Will look at the bugs later. Aug 26 07:44:16 polyonymous: moin Aug 26 07:57:27 bonjour Aug 26 08:02:12 hi / bonjour Aug 26 08:12:36 wb hrw Aug 26 08:12:45 re Aug 26 08:28:14 hi Aug 26 08:47:42 hi woglinde, by by woglinde_ Aug 26 08:48:01 he ant Aug 26 09:01:07 hi woglinde Aug 26 09:01:30 hi boris Aug 26 09:11:59 florian: good morning Aug 26 09:14:30 good morning Aug 26 09:15:03 morning Aug 26 09:15:28 currently i experienced no problems with directfb-1.2.3 on angstrom , woglinde :) Aug 26 09:16:07 unless a seg fault but i think it's not due to directfb Aug 26 09:16:10 i hope ^^ Aug 26 09:18:32 boris_OmegA: while you're there, please feel free to update with your findings these bugs: http://bugs.openembedded.net/buglist.cgi?quicksearch=directfb Aug 26 09:18:34 boris hm Aug 26 09:18:58 boris is other stuff segfaulting too? Aug 26 09:21:35 ant ah I can close one bug Aug 26 09:21:37 its fixed Aug 26 09:21:50 one less Aug 26 09:21:55 good Aug 26 09:22:18 ah its stable Aug 26 09:22:19 args Aug 26 09:22:25 naah Aug 26 09:23:30 hm Aug 26 09:23:38 has some stable checkedout? Aug 26 09:23:58 nope, too old for my uses Aug 26 09:24:14 or someone knows if there is a webfrontend anymore Aug 26 09:24:39 hi all! Aug 26 09:24:45 hi xlele Aug 26 09:25:14 I'm using stable, I changed nothing but local.conf Aug 26 09:25:19 i have this error: Aug 26 09:25:20 CROSS COMPILE Badness: /usr/include in INCLUDEPATH: /usr/include/ Aug 26 09:25:32 during uclibc compilation (angstrom distro) Aug 26 09:25:35 xlele which package? Aug 26 09:25:57 I run into this in the past, but I can't remember how i solved it... Aug 26 09:26:16 xlele could you please look at packages/diretctfb/files if there is a file dont-use-linux-config.patch ? Aug 26 09:26:20 woglinde: I think .stable is only on mtn and not imported in git Aug 26 09:26:26 ant yes Aug 26 09:27:01 ah hehe Aug 26 09:27:02 works Aug 26 09:27:04 http://amethyst.openembedded.net/oe/viewmtn/viewmtn.py/ Aug 26 09:27:30 ah, ok, 153 days ago Aug 26 09:27:38 hm Aug 26 09:27:38 :=) Aug 26 09:27:52 could be that know one checke stuff in Aug 26 09:28:08 woglinde: no, no such file exists in this dir Aug 26 09:28:16 xlele okay Aug 26 09:28:24 than I can close the bug Aug 26 09:28:26 cannt Aug 26 09:28:38 joking, 153 days ...that was the phoenix 'angstrom-stable' Aug 26 09:28:38 I am using stable, downloaded from mtn Aug 26 09:29:10 is the stable repository moving to git too? Aug 26 09:29:20 * ant_work think probably this link should be on OE wiki Aug 26 09:29:28 no woglinde, but i currently try to debug, but without gdbserver it's not easy :s Aug 26 09:29:36 hm Aug 26 09:29:41 than build gdbserver Aug 26 09:29:45 bitbake Aug 26 09:29:48 gdb Aug 26 09:30:10 bitbake gdb builds gdbserver too ? Aug 26 09:30:21 yes Aug 26 09:30:24 i have gdb but not gdbserver currelty on my arch ....streange Aug 26 09:30:30 or gdb-native Aug 26 09:30:37 letme see Aug 26 09:30:45 great news :) Aug 26 09:31:27 do_install () Aug 26 09:31:34 install -m 0755 gdb/gdbserver/gdbserver ${D}${bindir} Aug 26 09:31:50 otherwise I would be schocked a little bit Aug 26 09:31:51 found :) Aug 26 09:35:26 sorry to bother you guys Aug 26 09:36:15 can you please help me with a clue for "cross compile badness" error during uclibc package compilation? Aug 26 09:36:42 I tried to search into the wiki and the ML, but without success :( Aug 26 09:37:38 I can't figure out why /usr/include is in my INCLUDEPATH Aug 26 09:38:01 I did not change anything inside my .bb files! Aug 26 09:38:14 xlele could you try dev branch? Aug 26 09:38:19 and test again Aug 26 09:38:30 I fear the uclibc in stable is in bad shape Aug 26 09:38:59 i had terrible experiences with .dev in the past! Aug 26 09:39:11 ???? Aug 26 09:39:57 I'm not a cross-compiling geek, unfortunately Aug 26 09:40:19 I just need to make an embedded system work for a project I'm working in Aug 26 09:40:35 and I'd like to go the simplest way possible Aug 26 09:40:38 xlele and stable dont has the toolchain improvement from khem Aug 26 09:41:10 03  07org.openembedded.dev * r855bf06575 10OE.dev/packages/images/fso-image.bb: fso-image: remove bogus compatible machine Aug 26 09:41:19 03  07master * r855bf06575 10OE.dev/packages/images/fso-image.bb: fso-image: remove bogus compatible machine Aug 26 09:41:38 woglinde: in other words, you think that with .dev i'd have less pain in making it work? Aug 26 09:42:10 xlele I think so Aug 26 09:42:23 can you tell us about the cpu or architecture? Aug 26 09:42:47 rschuster is jsut checked in my latest classpath changes Aug 26 09:42:47 I'm trying to cook an image for vortex86 platform Aug 26 09:42:56 ah right x86 Aug 26 09:42:58 args Aug 26 09:43:00 basically it is a x86 system Aug 26 09:43:12 to be honest x86 is tested very littl Aug 26 09:43:14 e Aug 26 09:43:24 indeed :( Aug 26 09:43:54 but why the stable branch is not stable anymore? Aug 26 09:44:03 hm it is stable Aug 26 09:44:07 but has bugs Aug 26 09:44:09 *g* Aug 26 09:44:13 :D Aug 26 09:44:38 woglinde, i don't manage to debug with gdb ... gdbserver running on arm9, gdb on x86 Aug 26 09:44:41 got it... maybe the fastest way is to solve the problem with uclibc Aug 26 09:44:45 isn't it possible ? Aug 26 09:44:46 in stable Aug 26 10:20:52 boris it is Aug 26 10:21:13 but you have to point to the right dir where the crosscompiled libs reside Aug 26 10:21:40 search google for solib-absolute-prefix and solib-search-path Aug 26 10:22:01 boris I thought you already debuged some stuff with it Aug 26 10:47:48 *poof* Aug 26 10:47:56 one bug shquashed Aug 26 10:49:06 * * OE Bug 1179 has been RESOLVED (FIXED) by heinold(AT)inf.fu-berlin.de Aug 26 10:49:08 * * linux-uclibc gpe-image arm ep93xx build failed due to patch failure of uclibc Aug 26 10:49:11 * * http://bugs.openembedded.net/show_bug.cgi?id=1179 Aug 26 11:07:55 woglinde: he he 2006-07-16 Aug 26 11:20:56 ant_work: some bugs just fall off the radar Aug 26 11:20:59 that is natural Aug 26 11:26:29 I wonder where would be the best place to define a sane default for XSERVER for all machines (unless they want to deviate from it in their own conf/machine/$machine.conf)? Aug 26 11:28:40 XorA, Crofton: you have an idea? Aug 26 11:29:03 no idea Aug 26 11:29:05 I don't use displays :) Aug 26 11:29:14 serial console for me (or ssh) Aug 26 11:29:21 you just whistles in PSK to control your boards :-) Aug 26 11:29:28 FSK I meant Aug 26 11:29:36 PSK also Aug 26 11:29:43 Phase shift keying Aug 26 11:29:49 harder for human ear to detect PSK though Aug 26 11:29:53 "better" the FSK Aug 26 11:29:57 yeah Aug 26 11:30:02 harder to whistle also Aug 26 11:30:18 03  07org.openembedded.dev * rdba8ba7f2e 10OE.dev/packages/images/fso-image.bb: disapproval of revision '9cdbd45b5ff507c6e00fc91c08ec5e17929302f4' Aug 26 11:30:22 heh, we turn this channel really geek with the radio HAM speak Aug 26 11:30:25 03  07master * rdba8ba7f2e 10OE.dev/packages/images/fso-image.bb: disapproval of revision '9cdbd45b5ff507c6e00fc91c08ec5e17929302f4' Aug 26 11:32:11 uf.. Aug 26 11:32:16 3000 mails passed Aug 26 11:33:45 No idea where the best place for setting global variables would be? Aug 26 11:33:54 jokes aside, I mean Aug 26 11:35:13 Laibsch: about $XSERVER there is anarchy (http://bugs.openembedded.net/show_bug.cgi?id=3428#c5) Aug 26 11:35:14 Laibsch: you are aware that not having XSERVER= defined wont actually break the fso-image Aug 26 11:38:36 why does is break the image? Aug 26 11:38:40 what is an fso btw Aug 26 11:39:19 hrw: 3000 spam filtered e-mails? Aug 26 11:39:35 behind on his email I suspect :) Aug 26 11:39:37 FreeSmartphone.org Aug 26 11:39:43 ah Aug 26 11:39:44 == FSO Aug 26 11:39:46 ant_work: not quite Aug 26 11:55:29 XorA: Try "MACHINE=spitz bitbake fso-image" Aug 26 11:55:58 Crofton: the image definition relies on $XSERVER being defined which is not always the case Aug 26 11:56:03 which is what I am trying to fix Aug 26 11:56:35 what error does it actually fail with? Aug 26 11:56:44 unbuildable dependency Aug 26 11:58:49 can you test if the var is defined and print an error if it is not defined? Aug 26 11:59:13 this might be a more general solution Aug 26 11:59:34 * XorA is not sure what is causing the error and doesnt have time to look right now, but this all sounds fishy Aug 26 11:59:52 actually, at the risk of sounding like koen, I think that the reference to ${XSERVER} there is probably... well, bogus. You would probably be better off using virtual/xserver as the dependency at that level, and allowing individual machines which have a strong preference to call in the right one via IMAGE_INSTALL_prepend in their own conf files. Aug 26 11:59:59 as if XSERVER is not defined all that ends up is that it is missing from IMAGE_INSTALL Aug 26 12:02:32 XorA: no, undefined variables don't expand to the empty string. Aug 26 12:03:07 pb___: oh, I always thought they did Aug 26 12:03:10 if XSERVER isn't set at all (as opposed to being set with an empty value) then it will stay as a literal "${XSERVER}" and you will indeed find it blowing up with an unresolveable dependency Aug 26 12:04:19 XorA: I'm pretty sure they don't. I can see why you might get the idea that they do, though, because bitbake.conf and the classes do pre-initialise quite a lot of things to "". Aug 26 12:04:31 pb: You are right and but this is not what koen said. Aug 26 12:04:36 pb___: probably just my mistake, never looked into it closely Aug 26 12:04:52 The right solution is to define XSERVER generally Aug 26 12:05:01 And deviate only for a few machines where necessary Aug 26 12:05:12 Currently, XSERVER is not defined at all for some machines Aug 26 12:05:21 probably better to check if its set in the fso-image and default if its not Aug 26 12:05:27 The fbdev server should be the default Aug 26 12:05:43 or XSERVER ?= "fbdev" in fso-image :-) Aug 26 12:05:45 Laibsch: if you wanted to go ahead and set XSERVER ?= "virtual/xserver" in some suitable place, I'm sure that would be fine. Aug 26 12:06:04 I'm all set Aug 26 12:06:18 My initial question was where that suitable place would be Aug 26 12:06:25 That question is still open Aug 26 12:06:28 I'm not sure that making fbdev be the default is necessarily the right thing. You could make a fairly cogent argument that xorg would be a better default. Aug 26 12:06:36 bitbake.conf is the obvious place Aug 26 12:06:42 OK Aug 26 12:06:48 Wasn't obvious to me Aug 26 12:06:51 or, if it's only used in fso-image.bb, put it there. Aug 26 12:07:08 No, this has been inconsistently spread all over the place Aug 26 12:07:21 It'll be some work to consolidate Aug 26 12:07:26 maybe in image.bbclass Aug 26 12:07:37 no Aug 26 12:07:43 as all the blah-image.bb files already have XSERVER ?= "xserver-kdrive-fbdev" Aug 26 12:07:43 I thought image.bbclass as well Aug 26 12:07:48 and already made the change Aug 26 12:08:07 But that won't help you with "bitbake $XSERVER" Aug 26 12:08:16 no, I don't think image.bbclass would be the right thing. You might want to use it in a task package for example. Aug 26 12:08:28 the image definition is the wrong place to define XSERVER Aug 26 12:08:37 this is more a machine attribute Aug 26 12:09:07 task-gmae.bb: ${XSERVER} \ Aug 26 12:09:07 task-openmoko-bare.bb: ${XSERVER} \ Aug 26 12:09:07 task-openmoko-ui.bb:XSERVER ?= "xserver-kdrive-fbdev" Aug 26 12:09:07 task-openmoko-ui.bb: ${XSERVER} \ Aug 26 12:09:13 so yeah its used in tasks as well Aug 26 12:09:13 I know Aug 26 12:09:28 I will drop all the definitions in images Aug 26 12:09:44 definitions of XSERVER Aug 26 12:09:55 anyway Aug 26 12:09:59 I'll try to come up with something later and put it up for review Aug 26 12:12:01 ok, very good Aug 26 12:14:07 Laibsch, do post to the list, it help more people learn this stuff Aug 26 12:14:27 yes Aug 26 12:15:11 I wonder what the default should be. I talked with mickeyl about this and we both agreed it should be xserver-kdrive-fbdev Aug 26 12:15:37 Looking at the machines, it also seems to be the most common Aug 26 12:16:03 We would need to define the least number of exceptions with this choice of default Aug 26 12:18:19 I still kind of feel that virtual/xserver is the best default. Most platforms are only going to have a single xserver available anyway. Aug 26 12:19:24 but of course I will defer to überhacker mickeyl if he thinks differently Aug 26 12:20:48 do our xservers still all provide virtual/xserver? Aug 26 12:20:51 hi, btw. Aug 26 12:21:08 if so, i wonder who invented XSERVER in the first place Aug 26 12:21:15 Laibsch: c7x0 will the typical case of excpetion (xserver-kdrive-imageon will differ both from standard XSERVER and from Zaurus family PREFERRED_PROVIDER_virtual/xserver Aug 26 12:21:28 hi mickeyl :-) Aug 26 12:22:09 well, I guess you do need ${XSERVER} set to something different if you are on a platform where multiple xservers are available. Aug 26 12:22:44 right, but still, i would think that the machine knows best about the best xserver for it Aug 26 12:22:56 yes, indeed Aug 26 12:23:00 so, why invent an additional logic on top of PREFERRED_PROVIDER Aug 26 12:23:55 well, PREFERRED_PROVIDER is only a build time thing; it doesn't necessarily help you with image construction if you have multiple xservers in your ipk repository. Aug 26 12:24:16 or, at least, I think that was the situation at the time ${XSERVER} was added. it might be a bit different now with bitbake's eleet reverse dependency stuff, I'm not entirely sure. Aug 26 12:24:27 hmm, right Aug 26 12:25:11 then again, i don't think people would cry if they had to change PREFERRED_PROVIDER in the rare case that two (or more) are available for your device Aug 26 12:25:23 ah well Aug 26 12:25:28 the task package, of course Aug 26 12:26:40 i agree with virtual/xserver in any case Aug 26 12:26:53 as for the runtime settings, we need to discuss that Aug 26 12:28:15 ok, great Aug 26 12:28:45 to answer your original question, it looks like the answer is yes: all the xservers do PROVIDE virtual/xserver. Aug 26 12:29:00 if there are any that don't then it should be easy enough to fix them :-} Aug 26 12:29:19 I guess they should also RPROVIDE virtual-xserver to solve the runtime issue in the same way, but they don't seem to do that at present. Aug 26 12:29:41 * pb___ lunchtime now Aug 26 12:29:42 bbiab Aug 26 12:30:32 ~bon appetit Aug 26 12:30:33 [bon appetit] smacznego. Guten Appetit. Eet Smakelijk. God Appetitt. Buon Appetito. Buen apetito Bom Apetite. buen apetito Smaklig måltid!. Hyvää ruokahalua. Bo Proveito Aug 26 12:35:37 re Aug 26 12:36:44 hi all! Aug 26 12:37:12 in stable, compiling for x86 the package uclibc, i get the following error: Aug 26 12:37:14 CROSS COMPILE Badness: /usr/include in INCLUDEPATH: /usr/include/ Aug 26 12:37:22 re Aug 26 12:37:28 hi pb Aug 26 12:37:58 I've tried to look where the include path is set, but I didn't find anything... Aug 26 12:38:20 could anyone of you please address me in the right direction? Aug 26 12:41:11 thesing: so, for kexecboot the plan is to remove the init and the shell? Aug 26 12:42:34 ant_work: yes. But not near future. I won't touch any code thats not related to my diploma thesis in near future. Aug 26 12:43:19 thesing *g* Aug 26 12:43:27 do you think it makes sense to parse other args beside 'mtdparts=' ? Aug 26 12:43:45 because I'd import some c classes I found on the net Aug 26 12:44:00 but really you make me reinvent 'split' Aug 26 12:44:04 :=) Aug 26 12:45:30 btw is 'strtok' implemented in klibc? Aug 26 12:46:03 ant_work: I don't know if we need more parsing. ATM mtdparts seems to be enough. Aug 26 12:46:10 woglinde: similar problem? Aug 26 12:46:38 thesing: I'd suppose woglinde is on the other part of the desk Aug 26 12:46:45 referring exams Aug 26 12:47:12 ant_work: klibc seems to have klibc Aug 26 12:47:18 thesing in near fututre yes Aug 26 12:56:05 pb: RPROVIDE = "virtual/$something" is incorrect Aug 26 12:56:08 * Laibsch points to Aug 26 12:56:13 !oebug 2412 Aug 26 12:56:13 * * Bug 2412, Status: NEW, Created: 2007-06-04 07:03 Aug 26 12:56:14 * * rodrigo.vivi(AT)gmail.com: dpkg-deb failing when package name contains character '/' Aug 26 12:56:15 * * http://bugs.openembedded.net/show_bug.cgi?id=2412 Aug 26 13:00:09 thesing: INITRAMFS_IMAGE = "initramfs-kexec-image" Aug 26 13:00:14 not necessary anymore? Aug 26 13:00:27 I'm kind of lost as to the latest state of affairs Aug 26 13:01:10 How do I specify the size for the Angstrom rootfs? It is defaulting to 64MB but I have a 256MB card... Aug 26 13:01:30 Laibsch: you don't need it in local.conf anymore. Aug 26 13:01:33 local.conf Aug 26 13:01:51 thesing: I had it in sonkei.conf, distro config Aug 26 13:02:05 Laibsch: you don't need it there either. Aug 26 13:02:11 OK, thanks Aug 26 13:03:57 I'm happy to see that qt-embedded is working again :) Aug 26 13:04:06 piroko? Aug 26 13:04:21 woglinde: qt-embedded didn't compile for me for the longest time. Now it does. Aug 26 13:04:24 qte2? Aug 26 13:04:27 Laibsch: btw. in Talk:Monotone_to_Git_ID_mapping you're giving me "thomas@openedhand.com" which is not one of my addresses. Aug 26 13:04:33 or qt4? Aug 26 13:04:34 woglinde: qte4.4.1 Aug 26 13:04:54 thesing: If it was me who made the change I copied that from the non-talk page Aug 26 13:05:02 speaking of which, is the final conversion scheduled to any time? :) Aug 26 13:05:16 thesing: it's a wiki, update it to what you want, rather Aug 26 13:05:53 Man, this rootfs size thing is killing me... Trying to write 71MB to a 64MB partition isn't a good idea O_o Aug 26 13:06:07 xd Aug 26 13:06:19 But I can't figure out how to change it! :( Aug 26 13:06:35 Laibsch: which non-talk page? Aug 26 13:06:48 priroko which cpu? Aug 26 13:06:54 woglinde: ARM9 Aug 26 13:07:02 machine? Aug 26 13:07:03 sorry Aug 26 13:07:17 IMAGE_ROOTFS_SIZE? Aug 26 13:07:25 thesing: you pointed to a talk page. There is a normal page associated with it. Just strip "Talk:" Aug 26 13:07:38 woglinde: at91sam9263 Aug 26 13:07:45 ah this page. Aug 26 13:09:00 Laibsch: the error was already in RP's version of the page. Aug 26 13:09:08 polyonymous-work: Set that in my local.conf? Aug 26 13:09:08 ok Aug 26 13:09:15 go ahead and fix it Aug 26 13:09:25 if you like Aug 26 13:09:27 try to. Aug 26 13:09:39 piroko, basically, grep conf/ for SIZE for hints :) Aug 26 13:09:57 piroko: I use IMAGE_ROOTFS_SIZE_ext2 = "25000" in my local.conf Aug 26 13:10:02 prioko try ROOT_FLASH_SIZE in local.conf Aug 26 13:10:22 piroko: so, i guess, you have to specify the fs type Aug 26 13:10:50 xlele: Ah! Aug 26 13:13:02 Laibsch: indeed it is, that's why I said "virtual-xserver" and not "virtual/xserver" Aug 26 13:13:33 oh, sorry Aug 26 13:13:40 Details, details Aug 26 13:13:50 heh Aug 26 13:13:55 much more involved this stuff than I ever thought Aug 26 13:15:39 well, I never understood fully how many subst are done in background... ('_', '/') Aug 26 13:15:59 I remember some talk about separators years ago... Aug 26 13:17:32 ant_work, that's because you can't use _ in dns! :) Aug 26 13:19:25 good point Aug 26 13:29:57 ~hail debian installer Aug 26 13:29:58 * ibot bows down to debian installer and chants, "I'M NOT WORTHY!!" Aug 26 13:30:13 hrw: You're disgusting Aug 26 13:30:23 ? Aug 26 13:30:25 Making the poor bot bow before things Aug 26 13:30:35 You should be ashamed Aug 26 13:30:51 ah Aug 26 13:30:58 ;) Aug 26 13:31:07 `,) Aug 26 13:31:30 anyway I am now installing Debian on laptop which has winxp on ntfs and all what I need is in installer Aug 26 13:31:42 resizing ntfs, creating crypto lvm... Aug 26 13:36:27 hrw: The crypto lvm is pretty sweet I must admit Aug 26 13:43:19 xlele: I tried ROOT_FLASH_SIZE_jffs2 = "256000" in my build/conf/local.conf and it still only made a 64MB partition... :'( Aug 26 13:43:59 piroko: mmmm... let me check on my build env... Aug 26 13:44:05 prioko try ROOT_FLASH_SIZE = "256" in local.conf Aug 26 13:45:33 woglinde: 256 or 256000? Aug 26 13:45:39 256 Aug 26 13:46:12 Wait, no. I did IMAGE_ROOTFS_SIZE Aug 26 13:47:24 hm strange Aug 26 13:47:30 bitbake.conf:ROOT_FLASH_SIZE ?= "256" Aug 26 13:47:32 hm Aug 26 13:47:33 hm Aug 26 13:47:41 http://www.nabble.com/ROOT_FLASH_SIZE-td156075.html Aug 26 13:47:56 That says I should modify the machine.conf Aug 26 13:48:02 But not what to Aug 26 13:48:54 prioko I looked already there Aug 26 13:49:02 for your machine it is not set Aug 26 13:49:28 piroko: I use x86 as machine Aug 26 13:50:02 piroko: in my .conf I don't specify any fs size Aug 26 13:50:12 but in my local.conf I have the following: Aug 26 13:50:13 IMAGE_FSTYPES = "ext2 squashfs" Aug 26 13:50:14 IMAGE_ROOTFS_SIZE_ext2 = "25000" Aug 26 13:50:26 Did that for jffs2, did nothing Aug 26 13:50:30 Maybe jffs2 is special? Aug 26 13:50:34 piroko: indeed Aug 26 13:50:52 piroko: maybe jffs is managed in a different way... Aug 26 13:51:07 xlele: I believe it is, as the kernel manages its boundaries iirc Aug 26 13:51:15 Only problem is I have no idea how to change that parameter Aug 26 13:51:21 Where rather Aug 26 13:52:49 piroko: ROOT_FLASH_SIZE IMAGE_ROOTFS_SIZE are not used Aug 26 13:53:14 piroko: you build jffs2 image? Aug 26 13:53:17 hrw: Yep Aug 26 13:53:27 your device require padded image? Aug 26 13:54:11 hrw: I don't think so. My image is 71MB, I have a 256MB flash card, but only a 64MB partition is created Aug 26 13:54:45 'flash card' is nand? nor? onenand? sd? Aug 26 13:54:56 nand Aug 26 13:55:02 whats onenand? Aug 26 13:56:15 woglinde: onenand is nand chip with some extras Aug 26 13:56:33 piroko: so you need to strip something from image or resize flash partitions Aug 26 13:56:54 piroko: OE does not use any variable to limit jffs2 image size Aug 26 13:56:57 hrw I think he wants the later one Aug 26 13:57:18 hrw: I need to resize the flash partitions I guess. Aug 26 13:57:38 Just not sure how to go about that in OE, or without OE for that matter Aug 26 13:58:35 piroko: consult your board manual I think Aug 26 13:58:54 hrw???? Aug 26 13:59:06 piroko: what's the bootloader? Aug 26 13:59:16 woglinde: ? Aug 26 13:59:18 Hi there! I'm trying to bitbake with a new branch but I don't know what I have to change from the "Getting Started"-example Aug 26 13:59:43 hm Aug 26 13:59:50 does oe create the partition Aug 26 13:59:52 This was my try but it didn't work: mtn --db=/media/daten/OE/oe.db pull monotone.openembedded.org org.openmoko.asu.stable Aug 26 13:59:58 or is they already theire Aug 26 14:00:42 Proton23 openmoko is not hosted at oe Aug 26 14:00:57 proton it is hosted at git.openmoko.org Aug 26 14:01:24 try to understand the git manual(which is not this complicated) and check it out form there Aug 26 14:02:59 you don't need to understand manual, there's a walkthrough at http://wiki.openembedded.net/index.php/Git Aug 26 14:03:35 Thanks a lot you two Aug 26 14:03:58 before I sent this to the ml, what do you think about http://oss.leggewie.org/wip/XSERVER.txt? mickeyl, XorA|lunch, pb__, Crofton? Aug 26 14:04:12 The patch is big enough as it is, I left the RPROVIDES stuff out for now Aug 26 14:04:21 That should be easy to add afterwards Aug 26 14:04:40 It's also somewhat only loosely related to this. Aug 26 14:04:57 ant_work: uboot. I modified bootargs to have mem=256M but I don't think that's right... Aug 26 14:05:01 piroko: if it's u-boot, look at the changes to u-boot_git.bb about resizing using ROOT_FLASH_SIZE as placeholder Aug 26 14:06:07 Nope. That wasn't it Aug 26 14:06:27 piroko mem is for memory Aug 26 14:06:31 woglinde: I figured :( Aug 26 14:08:33 ant_work: Well they set ROOT_FLASH_SIZE in there. Should I try putting that in my machine.conf instead of my local.conf maybe? Aug 26 14:08:51 Laibsch: I would suggest that you leave the existing XSERVER assignments in the machine.conf files alone, let the individual maintainers remove them as they see appropriate. They are harmless, and your new weak assignment in bitbake.conf is not 100% equivalent to the previously-existing strong assignments. Aug 26 14:09:01 Laibsch: also, there is no need to increment PR if you aren't changing any output packages. Aug 26 14:09:07 piroko: yes, it is a machine setting. ROOT_FLASH_SIZE = partition limits, max rootfs partition size as defined in mtdparts Aug 26 14:09:50 but remember this pertition can contain a much bigger *compressed* image Aug 26 14:10:01 * hrw agrees with Phil Aug 26 14:10:06 Laibsch: also also, and at the risk of sounding like a broken record this time, I would still suggest virtual/xserver :-} Aug 26 14:10:57 I have ROOT_FLASH_SIZE = 256 in my machine.conf Aug 26 14:11:17 pb__: Oh, right Aug 26 14:11:36 virtual/xserver Aug 26 14:11:43 piroko: where is u-boot? other chip? Aug 26 14:11:53 ant_work: Yeah. ON a dataflash Aug 26 14:11:55 *On Aug 26 14:11:59 PR is harmless, and I thought, I'd rather bump it than think too hard on whether it was necessary ;-) Aug 26 14:12:20 piroko: so 256 should be fine Aug 26 14:12:40 Actually, I had 252 set Aug 26 14:12:42 touching package/images or not is something, I'm not sure about yet. Aug 26 14:12:55 piroko: but where is the kernel? Aug 26 14:12:55 And it wasn't doing anything. Still 64MB partition Aug 26 14:13:01 ant_work: On the dataflash Aug 26 14:13:08 pb__: I like lean stuff Aug 26 14:13:15 pb__: thanks for taking a look Aug 26 14:13:34 rootfs built by OE is on nandflash, custom kernel is copied manually to dataflash Aug 26 14:15:27 did tou clean/rebuild u-boot after patching bootargs? Aug 26 14:16:49 grep -rn -e pci_alloc_consistent /usr/src/linux-headers-2.6.20-15-generic | wc -l gives me 17? Aug 26 14:17:51 never mind that question Aug 26 14:18:00 ant_work: What do you mean clean/rebuild? I just modified the environment on the board Aug 26 14:18:04 hi kgilmer Aug 26 14:18:09 ant_work: Using printenv/setenv Aug 26 14:18:14 What file should I include to use pci_alloc_consistent()? Aug 26 14:18:24 OE does nothing about u-boot for me Aug 26 14:18:29 hey woglinde :) Aug 26 14:18:29 ah, ok, good for you your u-boot is sane enough :-) Aug 26 14:18:57 ant_work: Is there some environment variable I should change for u-boot to make the partition size bigger? Aug 26 14:19:02 thesing: morining Aug 26 14:19:12 Or to make linux think the partition is bigger? Aug 26 14:19:33 piroko: dunno your board and flash specs, probably it defaults to only one partition Aug 26 14:20:15 kernel takes mtdparts from cmdline and/or atags Aug 26 14:20:41 if these are not specified I suppose it defaults to physmap or so Aug 26 14:23:26 Argh. mtdparts looks confusing Aug 26 14:31:50 Laibsch: it's true that bumping PR unnecessarily will not actually break anything, but it is annoying for other people to end up with useless rebuilds and even more annoying for end users if "ipkg upgrade" downloads and installs a pile of stuff that isn't actually any different to what they had before. Aug 26 14:31:58 pb__: Does http://oss.leggewie.org/wip/XSERVER.txt look better now? Aug 26 14:32:18 pb__: This is images Aug 26 14:32:46 plus task-openmoko-ui Aug 26 14:32:52 I kind of like the idea to be able to tell whether something was built before or after this change, just by looking at the PR of built stuff Aug 26 14:33:20 angstrom-x11-base-depends doesn't look like an image either Aug 26 14:33:44 yes, but it has nobusiness defining XSERVER, AFAICT Aug 26 14:34:04 I don't think these two actually build anything Aug 26 14:34:05 well, no, but my point was that it is a real package and bumping its PR that you don't want to do capriciously. Aug 26 14:34:14 I'd agree with you if this was glibc or something Aug 26 14:35:18 Laibsch: the patch to htckaiser certainly doesn't look right Aug 26 14:35:35 PREFERRED_PROVIDER_virtual/xserver = "xserver-kdrive" Aug 26 14:35:36 -XSERVER = "xserver-kdrive-fbdev" Aug 26 14:35:36 +PREFERRED_PROVIDER_virtual/xserver = "xserver-kdrive-fbdev" Aug 26 14:35:40 Oops, correct Aug 26 14:35:44 thanks for spotting Aug 26 14:35:53 wonders of sed at work ;-) Aug 26 14:36:17 ... and, that makes me wonder whether there is actually an xserver-kdrive-fbdev.bb at all Aug 26 14:36:35 if there isn't, then you want to be saying "xserver-kdrive" not "xserver-kdrive-fbdev" everywhere that you mention the former Aug 26 14:36:44 er, the latter that is Aug 26 14:36:54 OK, enough of that for today Aug 26 14:37:02 Looks like this will take more looking into Aug 26 14:37:10 Need to get some real work done now Aug 26 14:37:22 righto Aug 26 14:56:23 hvontres|work: please update *all* defconfigs for initramfs support in linux-kexecboot 2.6.24 Aug 26 15:00:30 hvontres|work: BTW these defconfigs are producing too big kernels...(over 1,2mb) Aug 26 15:07:06 hi rschuster Aug 26 15:07:14 hi ken Aug 26 15:08:07 woglinde, do you know a channel for directfb ? i guess #directfb is not the only one ... Aug 26 15:08:49 ant_work: Ok, I'll try, but I only have a poodle to test on so far. Aug 26 15:09:45 most of them are lacking +CONFIG_BLK_DEV_INITRD=y +CONFIG_INITRAMFS_SOURCE="initramfs.cpio.gz" Aug 26 15:10:33 feel free to add it for 2.6.24. I'll bother thesing for 2.6.26 Aug 26 15:11:23 ant_work: ok. But I will only guarantee a working kernel for poodle :) Aug 26 15:11:56 np, nobody else apart our hackers Stanislav and Dmitry is touching these defconfigs Aug 26 15:32:18 Do we still have the old webiste somewhere? I'm missing http://oe.linuxtogo.org/successtories Aug 26 15:32:36 All in all it looks like we lost quite a lot of data... Aug 26 15:33:51 eeks Aug 26 15:34:07 mickeyl, kergoth`work: Are you able to fix this: http://oe.linuxtogo.org/successtories ? Aug 26 15:34:11 eh Aug 26 15:34:22 This one I mean: http://openembedded.sourceforge.net/ Aug 26 15:34:34 * florian needs another coffee Aug 26 15:35:33 uuh Aug 26 15:35:52 no idea, this is so old Aug 26 15:36:03 i don't even know whether i have my login for sf.net Aug 26 15:36:05 i'll check Aug 26 15:37:01 * florian needs OE success stories to show off ;) Aug 26 15:40:50 is there a dedicated IRC channel for FreeSmartphone.Org stuff? Aug 26 15:43:04 florian: bug kgilmer then Aug 26 15:44:35 kgilmer: ping Aug 26 15:44:51 hello hello hrw and florian Aug 26 15:45:02 i was away from my desk putting stickers on my bicycle Aug 26 15:45:31 kgilmer: hi... are you able to fix http://openembedded.sourceforge.net/ ? Aug 26 15:46:47 florian did you see this? http://bugblogger.com/author/ken-gilmer/ Aug 26 15:46:49 florian: bug him about success story Aug 26 15:46:56 second one down Aug 26 15:47:18 * kgilmer can't seem to directly link to the post Aug 26 15:47:45 florian not sure what you're referring to IRT HTTP://openembedded.sourceforge.net Aug 26 15:48:11 florian you may also want to bug the neuros guys Aug 26 15:48:15 (#neuros) Aug 26 15:50:15 forian, we'd also be happy to talk more specifically about OE and how it helps us, etc. Aug 26 15:50:24 we are big fans at bug labs :) Aug 26 15:50:50 ah :) Aug 26 15:52:04 well, I do not need to know that many details... I'd like to list a few companies and projects mostly. Aug 26 15:52:20 florian: bugged beagleboard team? Aug 26 15:52:38 but someone seems to have decided to throw away the list we had Aug 26 15:55:03 well the list is bigger now florian :) Aug 26 15:55:17 That's why I recommended *not* to start the wiki from scratch. The information we have is even worse than before... Aug 26 15:55:53 kgilmer: True... but we do not have place for it any more. Aug 26 15:56:34 Someone threw it away and someone else will have to rewrite it in order to add some bits... that's not good practice. Aug 26 15:56:34 ic...well i am just getting started with the oe wiki Aug 26 15:56:41 i'm working on a java page currently Aug 26 15:57:03 I'll check of Google has a backup Aug 26 16:03:07 bye all Aug 26 16:03:37 thx for your help hrw|gone Aug 26 16:18:19 thesing: ping Aug 26 16:19:25 kgilmer: Maybe you want to add aline or even a quote? :-) http://wiki.openembedded.net/index.php/SuccessStories Aug 26 16:20:27 ant_work: pong Aug 26 16:20:31 thesing: so we need smthg like FBANGLE_akita='270', FBANGLE_spitz='270', FBANGLE_collie='270' and '0' as default? Aug 26 16:20:42 hvontres|work: poodle? Aug 26 16:21:15 thesing: what was your original idea? rotation like service menus or landscape? Aug 26 16:22:07 on clamshells (c7x0) the cursor are fine if I let FBANGLE="0" Aug 26 16:22:18 if we have 3 machines using 270 and on 0 we set 270 as default and 0 for the machine using an override Aug 26 16:22:39 akita and spitz are somehow rotated...dunno all others... Aug 26 16:22:48 c7x0 is fine with 0 Aug 26 16:22:49 I think we should use portrait for collie /poodle and landscape for all clamshells Aug 26 16:24:38 but apart Zaurus? default 0 seems sound to me... Aug 26 16:26:15 we can change it if more machines use it. Aug 26 16:26:28 atm only zaurus use it. Aug 26 16:30:34 hvontres|work: you don't have commit rights? You are in the Monotone_to_Git_ID_mapping list. Aug 26 16:33:34 anyone building ffmpeg for x86 successfully? Aug 26 16:34:50 thesing: I think I was listed as auhor on a couple of patches pushed by hrw :) Aug 26 16:37:18 hvontres|work: could you rename devicetable-poodle to devicetable-kexecboot, delete devicetable-collie and change the machines that used it to use devicetable-kexecboot? Aug 26 16:38:28 hvontres|work: btw. does it work with this devicetable with multiple mmc partitions? Aug 26 16:39:07 thesing: yes, multiple partitions work. I'll update stuff when I get home tonight Aug 26 16:39:17 bbl Aug 26 16:39:33 hvontres|work: thanks. Aug 26 16:44:06 Jay7: hello Aug 26 16:44:32 ant_work: hi :) Aug 26 16:44:41 Jay7: some progress here... Aug 26 16:44:41 ant_work: what news here? :) Aug 26 16:44:53 basically we have all...:-) Aug 26 16:45:06 but with old u-boot yet Aug 26 16:45:10 good :) Aug 26 16:45:26 and how is InSearchOf's work on new u-boot? Aug 26 16:46:18 http://insearchof.blogsite.org/ Aug 26 16:47:32 Jay7: http://bugs.openembedded.net/show_bug.cgi?id=4520 Aug 26 16:47:40 if you already have some code... Aug 26 16:48:11 I'm finishing the new emergency-image in th spare time... Aug 26 16:48:57 I started with minimal-image-with-mtd-utils (+nandlogical) but eventually moved to a initramfs one Aug 26 16:49:24 ant_work: my primary question - is kexecboot + u-boot now working with resized rootfs? :) Aug 26 16:49:29 yes Aug 26 16:49:34 nice Aug 26 16:49:56 but thesing gave me 'hausaufgabe' to write the parsing code in C Aug 26 16:49:56 I was too busy this time Aug 26 16:50:11 now it seems like I have some time for experiments Aug 26 16:50:44 ant_work: to parse u-boot's cmdline? Aug 26 16:50:50 yes Aug 26 16:51:03 I was looking at strtok Aug 26 16:51:30 seems in klibc stings.h Aug 26 16:52:33 ant_work: from memory: maybe strstr is more what you need. Aug 26 16:52:52 yes, but only for one token... Aug 26 16:53:03 ant_work: strtok is ok :) Aug 26 16:53:10 I'm thinking about custom 'console=' redirections... Aug 26 16:53:26 I remember some other like strspn, strbrk, strsep Aug 26 16:53:46 Jay7: lately I only use php with strings... Aug 26 16:53:54 we only need mtdparts atm. So find mtdparts and the space after and you have what you want. Aug 26 16:54:03 ant_work: I'm too.. php/perl/shell :) Aug 26 16:54:21 ant_work: I can write it I think Aug 26 16:55:08 in 2002-03 I have spent some time on writing text processing application :) Aug 26 16:55:15 I'll be at home in a few hours..please post your progress here on ml :-) Aug 26 16:55:33 otherways I'll start from scratch Aug 26 16:55:42 thanks in advance, btw Aug 26 16:55:42 It was kinda of autoclassificator based on neural network :) Aug 26 16:56:10 ant_work: It simpier to start from scratch Aug 26 16:56:22 here is not so much work Aug 26 16:56:41 thesing: one more question Aug 26 16:57:06 is it really convenient to split the patch in two? Aug 26 16:57:19 I mean, the very same line is patched twice... Aug 26 16:57:23 most hard is to parse something like spaces within quotes.. Aug 26 16:58:48 Jay7: http://www.java2s.com/Code/C/String/Truncatestringbydelimiterhowtousestrtok.htm Aug 26 16:59:23 and nest a check for 'mtdparts=' in it Aug 26 16:59:25 http://www.freebsd.org/cgi/man.cgi?query=strtok&apropos=0&sektion=0&manpath=FreeBSD+7.0-RELEASE&format=html Aug 26 16:59:34 or just for '=' Aug 26 16:59:42 look at EXAMPLE section Aug 26 17:00:41 he he, same idea Aug 26 17:01:09 but from right to left :) Aug 26 17:01:50 see also to SEE ALSO ;) Aug 26 17:02:27 btw, I have using strsep Aug 26 17:02:32 http://www.freebsd.org/cgi/man.cgi?query=strsep&sektion=3&apropos=0&manpath=FreeBSD+7.0-RELEASE Aug 26 17:02:38 see EXAMPLE also Aug 26 17:02:47 when it is in klibc's strings.h Aug 26 17:03:39 if there is, all will be few lines Aug 26 17:04:06 Jay7: look at the sources of klibc... Aug 26 17:04:14 I don't have acheckout here Aug 26 17:04:17 but it looks like BSD-specific.. Aug 26 17:04:46 google has some results Aug 26 17:05:52 hehe.. gentoo man page have it but recommends to avoid using strsep and strtok because it modifies original string.. Aug 26 17:06:09 ok, I'm going to look into klibc :) Aug 26 17:06:21 yes..http://lkml.org/lkml/2006/6/25/200 Aug 26 17:08:44 bbl Aug 26 17:21:16 florian: will do :) Aug 26 17:23:34 For the record, I fixed my 64MB partition problem by updating to the latest u-boot : Aug 26 17:23:35 :) Aug 26 17:33:51 Is there a recommended way to make qt apps in OE? I have a qt app that compiles fine outside of OE, but I'm not sure the proper way to get it compiled under OE... Aug 26 17:36:25 what buildsystem? Aug 26 17:38:26 piroko: qmake is easy but we should also stage the pkg-config files of qt Aug 26 17:49:31 zecke_: How does one go about it? Aug 26 17:49:38 I have inherit qmake2 so far Aug 26 17:49:54 But my package fails trying to find a bunch of shared library providers Aug 26 17:50:20 piroko: hmm, takr Aug 26 17:50:30 piroko: take a look at qsvn.. this used to be my test case Aug 26 17:55:06 re Aug 26 17:55:15 So I should inherit qt4x11? Aug 26 17:56:02 piroko: if you want Qt4/X11 yes :) Aug 26 17:56:12 bbl Aug 26 18:02:44 re Aug 26 18:19:50 hm.. strtok in klibc code uses strsep Aug 26 18:20:03 strsep uses strpbrk Aug 26 18:20:37 and strpbrk uses __strxspn Aug 26 18:26:09 morning kergoth Aug 26 18:26:19 re pb Aug 26 18:26:20 hm Aug 26 18:26:28 monotone server slow again Aug 26 18:26:34 hi woglinde Aug 26 18:28:48 Has anyone in here successfully compiled a custom qt app in OE? Aug 26 18:29:11 okay wolfson again Aug 26 18:31:28 with qte2 yes Aug 26 18:31:41 woglinde: Nothing with qt4? Aug 26 18:31:45 no Aug 26 18:31:50 :( Aug 26 18:31:56 qt4 last 5 hours to build Aug 26 18:31:58 or so Aug 26 18:34:46 woglinde: Well I already have qt-embedded built through OE Aug 26 18:34:59 woglinde: I'm just trying to get this simple custom app I wrote to compile under OE Aug 26 18:35:04 Using qt-embedded Aug 26 18:35:09 where is the problem? Aug 26 18:36:17 I don't think I made the bb file correctly. I think it ends up compiling an x86 binary instead of an arm binary. It also complains saying NOTE: Couldn't find shared library provider for libQtGui.so.4 and many other libraries when it tries to do do_package Aug 26 18:37:41 hm seems there are only classes for qt4x11 and qtopia4 Aug 26 18:37:53 have you pastebin somewhere? Aug 26 18:38:06 where is the qt stuff installed? Aug 26 18:38:39 woglinde: What do you mean? I'm trying to use OE's built version of qt-embedded Aug 26 18:39:06 http://rafb.net/p/VGW0qA65.html Aug 26 18:39:31 prioko yes Aug 26 18:39:39 I just looked at the recipe Aug 26 18:39:45 florian: done Aug 26 18:39:58 prioko this is not the recipe Aug 26 18:40:11 Oh you want the recipe, sorry Aug 26 18:40:52 http://rafb.net/p/vlZhFL39.html Aug 26 18:40:56 prioko and NOTE: multiple providers are available for qtopia-core (qtopia-core, qt-embedded); Aug 26 18:41:00 NOTE: consider defining PREFERRED_PROVIDER_qtopia-core Aug 26 18:41:07 I don't know how to fix that Aug 26 18:41:22 you should set PREFERRED_PROVIDER_qtopia-core = "qt-embedded" inc local.conf Aug 26 18:41:34 easier then the app says how to fix something it cant bee Aug 26 18:42:10 Alright, done Aug 26 18:42:19 okay Aug 26 18:42:22 I'll tell you if it comes back with the same error Aug 26 18:42:22 then to the recipe Aug 26 18:42:33 because qtopia-core is virtual Aug 26 18:42:55 you should let depend it on qtopia-core instead of qt-embedded directly Aug 26 18:43:40 Alright Aug 26 18:44:02 I still don't know the difference between qtopia-core and qt-embedded Aug 26 18:44:23 names and versions :) Aug 26 18:44:45 yes and some wrapper stuff Aug 26 18:44:46 just wait.. it'll get better... Aug 26 18:44:59 so you link against qtopialib Aug 26 18:45:02 and so on Aug 26 18:45:19 hm is it wrapper bridge Aug 26 18:45:28 lorn might knew it better Aug 26 18:45:49 prioko has nxlogger to do with nx stuff? Aug 26 18:46:02 No. It's just a name I came up with Aug 26 18:48:02 Now I have a dependency loop Aug 26 18:48:10 hm Aug 26 18:48:33 args hm Aug 26 18:48:36 yes okay Aug 26 18:48:45 I would suggest Aug 26 18:50:25 where is the loop? Aug 26 18:50:30 in qt-embedded Aug 26 18:50:47 could ypu show the error? Aug 26 18:51:10 http://rafb.net/p/yeI4Wr95.html Aug 26 18:53:20 hms Aug 26 18:53:37 seems the deve for this stuff didnt think or test enough Aug 26 18:53:45 Go figure :-/ Aug 26 18:53:59 okay delete the PREFFERED think Aug 26 18:54:01 for now Aug 26 18:54:46 Alright Aug 26 19:01:00 I wonder why OE installs the qt-embedded libraries with an "E" at the end of their names... Aug 26 19:03:13 hm so you could do qt4x11 too Aug 26 19:03:19 hm but the name suckz Aug 26 19:03:29 qt4-embedded would more clear Aug 26 19:03:42 woglinde: Plus I don't want all the X11 stuff that comes with it Aug 26 19:04:15 hm Aug 26 19:04:19 this suckz too Aug 26 19:04:25 qmake is used form 4.4.0 Aug 26 19:04:35 so both 4.4.0 and 4.4.1 is downloaded Aug 26 19:04:36 args Aug 26 19:04:37 While the newest qt is 4.4.1 :( Aug 26 19:04:41 Yeah Aug 26 19:04:49 Qt and OE aren't very good quite yet :) Aug 26 19:06:10 okay uicmoc4 I meant Aug 26 19:11:37 *g* so just copy 4.4.0 uicmoc4 to 4.4.1 Aug 26 19:11:46 now I have only one download for it Aug 26 19:11:55 dont know why this was forgetten Aug 26 19:13:57 re Aug 26 19:14:03 re flo Aug 26 19:15:50 woglinde: Alright, I did that. What should I inherit to compile my package with qt-embedded? Aug 26 19:17:21 piroko hm I fear you have to write a class like qtopia4core.bbclass Aug 26 19:17:27 for it Aug 26 19:17:37 Crap :( Aug 26 19:17:50 shouldnt be to hard Aug 26 19:17:55 woglinde: Oh, and moving uicmoc4 fails because of the patches Aug 26 19:18:14 piroko yes I will look at it soon Aug 26 19:18:19 if it fails here to Aug 26 19:18:23 is at unpacking Aug 26 19:19:06 qkay Aug 26 19:21:26 ah seems qapp patch is applied Aug 26 19:21:31 upstream Aug 26 19:23:46 hmm, opkg asks questions during image building thats bad (tm) Aug 26 19:27:55 XorA: heh Aug 26 19:30:08 it just sits taking 100% cpu waiting for you to type y/n Aug 26 19:30:22 I am testing add -force-defaults to commandline Aug 26 19:59:34 he bluelightning Aug 26 20:00:15 bluelightning I pushed some fixes for opie-console with gcc-4.3 maybee you can check it in the opie-cvs Aug 26 20:01:01 hi woglinde Aug 26 20:02:52 woglinde: sure thing, could you post them to opie-devel@handhelds.org? Aug 26 20:02:52 thanks Aug 26 20:03:26 will do Aug 26 20:03:37 tommorrow Aug 26 20:04:31 cool Aug 26 20:05:02 or later Aug 26 20:05:08 when my mouse is working again Aug 26 20:05:19 hm cordless laser dont last a day of acion Aug 26 20:08:59 hmmm Aug 26 20:10:10 he kergoth Aug 26 20:10:16 hey Aug 26 20:10:24 whats up? Aug 26 20:10:39 hi...is there anybody compiling on x86_64 for machine type "native"? Aug 26 20:10:56 there's almost nothing about that on google Aug 26 20:11:08 ???? Aug 26 20:11:21 you mean image for x86_64? Aug 26 20:11:35 hmm, that reminds me Aug 26 20:11:56 my machine is x86_64, i want to build for x86_64...so after some googling i thought of using machine type "native" Aug 26 20:12:00 is that wrong? Aug 26 20:12:07 machine type x86_64 doesnt work either Aug 26 20:12:22 hm? Aug 26 20:12:49 * kergoth wonders if oe ever got multilib support for its distros Aug 26 20:12:49 why you want ran an embedded os on x86_64? Aug 26 20:13:11 OE is a wonderful way to generate a just-enough OS Aug 26 20:13:17 for use in virtual machines Aug 26 20:13:26 it's better than anything else existing Aug 26 20:13:27 hi mr_nice Aug 26 20:13:36 woglinde: hi Aug 26 20:13:56 mrt2k7 I can slim down a debian system enough Aug 26 20:14:09 woglinde: funnily enough I was thinking about using oe to generate livecd images for x86 and x86-64 the other day. it's certainly not a completely stupid thing to want to do. Aug 26 20:14:13 what would be responsible for a shell recipe to create a symlink from /bin/bash to /bin/sh? the bash package doesn't do this, yet busybox does Aug 26 20:14:24 woglinde, can you shrink it so far that a ZIP of the rootfs is of 2,7mbyte? ;) Aug 26 20:15:01 tharvey: update-alternatives is probably the right mechanism for that Aug 26 20:15:09 mrt2k7 hm so you need it only for some numbercracking like action Aug 26 20:15:10 not that I have any idea whether busybox actually uses u-a, but still Aug 26 20:15:30 mrt2k7 hm nobdy consider this now I think Aug 26 20:15:34 only pb Aug 26 20:15:42 comes out with it now Aug 26 20:15:44 imo bash should really never be /bin/sh, too many people blindly use bashisms in #!/bin/sh scripts and then whine when it breaks :P Aug 26 20:15:51 woglinde: some kind of, i call it "task virtualization", a VM for each task someone has Aug 26 20:15:55 ~fishslap kergoth Aug 26 20:15:56 * ibot slaps kergoth up side the head with a wet fish. Aug 26 20:16:03 like f.e., download a file, encode a video Aug 26 20:16:12 it works pretty well. its my diploma thesis Aug 26 20:16:33 you guys are of course mentioned in my preface with thanks for your work :) Aug 26 20:17:27 the goal is to migrate VMs by a P2P overlay so that less energy is needed Aug 26 20:17:52 kergoth: by that logic, more or less any backwards-compatible extension would be a bad idea. I know what you mean about creeping bashisms but I'm not sure that requiring everybody to install a separate shell just to police them is the right answer. Aug 26 20:18:11 pb__: have you just thought of it or did you accomplish anything? just wondering... Aug 26 20:18:16 I guess the ideal thing would be for bash to notice when it was being invoked as /bin/sh and turn off some of its more outlandish features Aug 26 20:18:22 pb__, know of an example that uses update-alternatives for this? I assume there is some way you call update_alternatives during native build to setup conf files? Aug 26 20:18:52 mrt2k7: thus far I haven't accomplished anything to speak of, I keep getting distracted by other projects. but we do have a fairly pressing need to replace our old (DSL-based) livecd system so I guess I will have to do something with it soon. Aug 26 20:19:19 mrt2k7: I guess you could start with 32-bit x86 (which is already pretty well supported in oe) for prototyping, and then switch to x86-64 later. Aug 26 20:19:36 true. that's a good point. thatd be nice, if bash was smarter about it Aug 26 20:20:04 pb_: yes, x86-32 works well, i use x86-prescott in .stable - crosscompiling from amd64 worked nice AFAIK Aug 26 20:20:57 tharvey: not offhand, but 'grep -r "inherit update-alternatives" packages' should reveal the truth Aug 26 20:21:11 pb__, thx Aug 26 20:22:43 mrt2k7: there is already a site file for x86-64 (and even for x86-64 with uclibc) so I guess someone must have worked on this already Aug 26 20:24:14 mrt2k7: it should be easy enough to create a conf/machine/x86-64-generic.conf based on the i686-generic one Aug 26 20:25:18 pb_, there is conf/machine/x86-64.conf - but doesn't work :( Aug 26 20:25:54 AFAIR I got some errors related to http://projects.linuxtogo.org/pipermail/openembedded-devel/2006-November/000748.html Aug 26 20:26:22 mrt2k7: oh, odd. I don't seem to have x86-64.conf in my .dev tree Aug 26 20:26:25 are you using .stable? Aug 26 20:26:37 yes Aug 26 20:26:47 too much noise in .dev for me ;) Aug 26 20:27:08 so i started with stable, was hard enough at the beginning anyways Aug 26 20:28:21 pb__: huh, bash already does change its behavior somewhat when invoked as /bin/sh, enables "posix mode", but that seems to only be for behaviors where it differs from posix, not the features above and beyond it Aug 26 20:28:40 re Aug 26 20:29:05 * kergoth ponders bitbake/oe changes and the possibility of a new similar tool Aug 26 20:29:05 who is to lart for the broken ppp_2.4.3.bb commit? Aug 26 20:29:27 ant_: dunno, what does mtn say? Aug 26 20:29:39 unparsed line: ' file:// Aug 26 20:29:51 and a lot of garbage... Aug 26 20:29:52 kergoth: oh, right, hm. Aug 26 20:30:11 oh well Aug 26 20:30:13 should just be a typo Aug 26 20:30:17 I guess that's the same as you get with POSIXLY_CORRECT Aug 26 20:30:58 pb__: http://rafb.net/p/QGZnWv82.html Aug 26 20:31:10 kergoth: what sort of new tool are you thinking of? Aug 26 20:31:52 ant_: right, I was thinking more in terms of the output from "mtn log" Aug 26 20:32:06 just mulling over some oe weaknesses i've had problems with over the years. no hardware resource management for the scheduling, no ability to offload package builds to other machines, i never wanted BBFILES to be a permanent mechanism, etc Aug 26 20:32:20 some problems can be fixed as is, others would have compatibility issues Aug 26 20:32:26 ant stable? Aug 26 20:32:27 do I need to depend on sed-native If I use sed in a .bb file ? Aug 26 20:32:36 in my ppp.bb is no line wwith Ifile Aug 26 20:36:02 pb__: is trivial: missing '\' http://rafb.net/p/fQCBTX93.html Aug 26 20:36:55 ant_: I'm writing cmdline parser right now :) Aug 26 20:36:58 mr_nice most distros has sed anyway Aug 26 20:37:07 ant_: ~1 hour I think :) Aug 26 20:37:17 mr_nice but for quality stuff I would add it Aug 26 20:37:37 woglinde: ok Aug 26 20:37:54 ant_: what parameters exactly we needed? mtdparts, what else? Aug 26 20:39:03 ok, so I understand how update-alternatives.bbclass works - very intuitive. Why wouldn't bash configure itself as an alternative for /bin/sh? Aug 26 20:39:05 Jay7: thesing said only this...the rest should be handles in /boot/kernel-cmdline Aug 26 20:39:06 re polyonymous Aug 26 20:39:20 ant_: what about 'console'? Aug 26 20:39:29 tharvey because the dev was lazy? Aug 26 20:39:30 Jay7: you read un my mind... Aug 26 20:39:34 oops..my code Aug 26 20:39:45 ant_: yes, I've :) Aug 26 20:39:47 Jay7: great, I'm investigating uImage (btw the std kernel size is too big, so you'll have to add -C gzip to mkimage args) Aug 26 20:40:01 woglinde, ok... I'll believe that - just wanted to make sure I am not missing anything regarding bash not being a possible alternative Aug 26 20:40:13 gzip can compress the kernel decompressor...2-3 kb are precious ! Aug 26 20:40:33 tharvey I see nothing which speaks against Aug 26 20:41:09 Jay7: let open the possibility to add other tokens Aug 26 20:41:25 I think it is wise Aug 26 20:42:13 ant_: ok, then I will write parser, not just substring searching :) Aug 26 20:42:32 if thesing removes the shell, space will not be a problem Aug 26 20:42:37 :=) Aug 26 20:43:07 a pareser could take a couple of kb... Aug 26 20:43:33 s/a pareser could/a parser should/ Aug 26 20:44:26 it's about ~1kb in binary I think :) Aug 26 20:45:07 feel free to add fancy pr0n boot-videos :-) Aug 26 20:45:26 ha ha Aug 26 20:45:45 somebody asked for music during boot IIRC Aug 26 20:45:52 no comment Aug 26 20:45:56 yeah, we already have framebuffer.. may be add mplayer into? :) Aug 26 20:51:58 Jay7: stupi question: can one recycle the parser code in /kernel/drivers/mtd/cmdline.c ? Aug 26 20:52:23 ant_: good point.. I've not think about this.. Aug 26 20:54:36 oh...question again...anyone of you having troubles with local overlays too? i noticed that f.e. things in conf/distro are not taken from my local overlay, although the file in there are newer an the priority is set higher Aug 26 20:56:44 does packaged staging handle everything, native and cross as well as regular packages? Aug 26 20:57:38 kergoth hm what you mean with regular packages? Aug 26 20:58:04 that which is not -native and -cross. packages built for the target. Aug 26 20:58:43 kergoth I would say yes Aug 26 21:01:42 ant_: parser code from cmdline.c is only for mtdparts parsing Aug 26 21:06:49 rwhitby: wb Aug 26 21:25:06 * * OE Bug 4524 has been created by  Aug 26 21:25:08 * * add aufs (patch included) Aug 26 21:25:10 * * http://bugs.openembedded.net/show_bug.cgi?id=4524 Aug 26 21:31:13 hi pigi_ Aug 26 21:31:30 hi pg__ Aug 26 21:31:38 hi pb__ either Aug 26 21:31:55 how are you doing ? Aug 26 21:33:27 pretty good. right now I am trying to debug some crazy ipsec problem, but other than that all is well :-} Aug 26 21:39:18 pb__: fine. I'm just fighting with tuxonice Aug 26 21:41:46 after building a recipe is there a task that can be called to package it aside from building an image that uses it? Aug 26 21:45:25 later all Aug 26 21:45:46 tharvey, if I got you right, building a recipe includes packaging it. Aug 26 21:54:38 polyonymous, ah yes your correct - sorry I thought otherwise Aug 26 21:55:32 ant_: I'm going to bed now.. Aug 26 21:55:34 ah, I know what I was missing - building a recipe includes packing it but it doesn't update the Package manifest Aug 26 21:56:42 ant_: I have some skeleton and write it tomorrow Aug 26 21:56:51 polyonymous, is there a way to update the package manifests? Aug 26 21:57:08 tharvey, you mean bitbake packge-index ? Aug 26 21:57:10 * Jay7 -> sleep() Aug 26 21:57:35 package-index, even Aug 26 21:58:55 polyonymous, ya... thats it... didn't realize that was an explicit package - makes sense that it is. It was hiding in meta Aug 26 21:59:12 it *is* meta :) Aug 26 21:59:32 anyway, I'm gone. Have fun :) Aug 27 00:23:17 hi all, how do I get osb-jscore to build from dev? Aug 27 00:25:41 make that: osb-jscore-0.5.2+svnr117-r1 Aug 27 00:37:55 error is: Aug 27 00:37:58 | collector.cpp: In static member function 'static void KJS::Collector::markCurrentThreadConservatively()': Aug 27 00:37:58 | collector.cpp:307: error: 'pthread_getattr_np' was not declared in this scope **** ENDING LOGGING AT Wed Aug 27 02:59:56 2008