**** BEGIN LOGGING AT Sat Jul 09 23:59:56 2005 Jul 10 01:32:17 Hi Jul 10 02:45:21 how can i uninstall (remove the ipks of packages) if i remove them manually. next time i make an image it says it can edit /usr/lib/ipkg file: permission denied. is there any mechanism which allows me to uninstall ipkgs from bitbake Jul 10 02:53:39 the problem is i already had compiled my image with 2.6.12-omap1-rc2 and if i compile my image with 2.6.12-omap1 it is always taking image with omap1-rc2. some one suggested here that ipkg thinks omap1-rc2 is latest when compared to omap1. so it takes it always. i tried deleting files omap1-rc2 manually. but, when i build i get an error saying permission denied (as ipkg is trying to write to /usr/lib) Jul 10 02:55:30 Hello Jul 10 02:56:32 I'am trying to setup a VPN with racoon (ipsec-tools) on a Zaurus with OZ 3.5.3 Jul 10 02:58:58 as far as I found out, there are modules for ESP, AH and so on available for 2.4.18-rmk7-pxa3-embedix. Jul 10 02:59:54 Is there anybody who can give me a link in the right direction? Jul 10 03:02:19 <[cc]smart> pb_: ping Jul 10 03:04:58 Zecke:ping Jul 10 03:23:40 [cc]smart: hello Jul 10 03:25:40 pb_: i having an ipk issue. if i delete them; during the next image building i get an permission error. but i have build oe environment as a user and not root. i dont understand why it wants to access /usr/lib directory ? Jul 10 03:26:00 eJumbo: that's harmless. Jul 10 03:26:32 do u know what my issue is ?. if u want to to explain. Jul 10 03:26:52 no, I don't know what your issue is. Jul 10 03:26:55 the problem is i already had compiled my image with 2.6.12-omap1-rc2 and if i compile my image with 2.6.12-omap1 it is always taking image with omap1-rc2. some one suggested here that ipkg thinks omap1-rc2 is latest when compared to omap1. so it takes it always. i tried deleting files omap1-rc2 manually. but, when i build i get an error saying permission denied (as ipkg is trying to write to /usr/lib) Jul 10 03:27:55 i removed kernel-image for rc2 then i get an build error saying it needs kernel-image rc2 even though i have done a clean. Jul 10 03:28:19 i mean kernel image ipk Jul 10 03:30:36 no body knows anything about ipsec? Jul 10 03:31:42 pb: how do i resolve my issue. does bitbake offer any method to remove my ipks. Jul 10 03:31:56 eJumbo: I have no idea. Jul 10 03:32:53 the problem is if i remove any of the ipks manually. the next build fails. it does not build the ipks again. Jul 10 03:34:22 pb_: did u understand my problem ? Jul 10 03:36:47 Ok! I have a outher Question: does anybody know a CF2PCMCIA-Adapter which makes it possible to us a D-Link Air in a Notebook. Jul 10 04:11:27 mickey: i am having a problem and need ur advice Jul 10 04:11:49 The Issue is I have compiled a image using Linux Kernel 2.6.12omap1-rc2. now latest kernel has been released which is Linux 2.6.12. If i use it to build my image. i always get the old one (2.6.12-omap1-rc2). this is an ipkg problem. i delete rc2 ipk files from my ipk directory. if i build an image now i get an error saying it requires the deleted ipks. i dont know why ? Jul 10 04:12:30 that is because of caching issues... Jul 10 04:12:43 deleting the ipks is not enough Jul 10 04:13:12 you could check the preferred_versions stuff in the .bb files, too Jul 10 04:13:28 u mean to say i must delete the cache. Jul 10 04:13:54 well, the best way obviously is to delete the complete tmp dir Jul 10 04:14:00 i have give DEFAULT_PREFERENCE="-1" for omap1-rc2 so it should not be built. i confirmed it. Jul 10 04:14:02 at least that's what mickey told me Jul 10 04:14:24 yeah but you should check for the kernel version in the machine conf file, too Jul 10 04:14:34 maybe distro config file, too Jul 10 04:14:38 i have done that now and a build is going on. Jul 10 04:15:00 problem with deleting tmp dir is, you'll have to build the who bootstrap again Jul 10 04:15:11 can take a while on a slow box, particulary glibc Jul 10 04:15:20 but mickey told me there's no real good way around it Jul 10 04:16:14 btw, you won't believe what i saw on the tele yesterday Jul 10 04:16:59 the problem is it takes time. is there no short cut. i am struck with this problem since yesterday and dont know how to solve it. Jul 10 04:17:32 well, you could try an delete selected files in the cache and stamps directory Jul 10 04:18:01 just want to delete the ipks related to kernel 2.6.12-rc2 not all. Jul 10 04:18:13 the ipks are not important Jul 10 04:18:31 the problem is that the cache still has your old .bb file Jul 10 04:18:52 and if you don't delete the stamp files, too there may be weird problems Jul 10 04:19:00 ..i guess Jul 10 04:21:32 i did a old-kernel clean new kernel clean Jul 10 04:21:49 that won't remove the files in the cache Jul 10 04:21:52 i think clean only removes files from work directory alone Jul 10 04:22:05 as for building the bootstrap, buying a faster comp is the only thing that helps *ggg* Jul 10 04:22:28 03mickeyl 07org.openembedded.dev * re80a27c7... 10/packages/dropbear/ (5 files in 2 dirs): Add dropbear 0.46. Patch courtesy Erik Hovland. Jul 10 04:22:48 deleting TMPDIR is only solution u say ? Jul 10 04:22:59 for details best ask mickey Jul 10 04:23:16 i said you could try and fidlle around with tmp/cache tmp/stamps Jul 10 04:23:21 but dunnow how... Jul 10 04:24:58 mickey prefers more serious questions. actually i know this is not the forum to ask questions like i do. but since there is no place to ask i need to hang on here pester,request,command. Jul 10 04:25:27 well, i asked him similiar questions a couple of days ago, and he was very helpful indeed Jul 10 04:25:41 but yes, hee Jul 10 04:25:41 i think time is right for a seperate forum for users of oe. so that this forum can be used for pure development questions. Jul 10 04:25:42 hehe Jul 10 04:25:56 there is an oe forum at oesf, isn't there? Jul 10 04:26:07 but it is a mailing list. Jul 10 04:26:26 need a chat room irc. so that we dont disturb the "core" guys. Jul 10 04:26:57 no: http://www.oesf.org/forums/index.php?showforum=136 Jul 10 04:27:06 i always feel bad to ask when they are discussing something important regarding oe enhancement. but forced to ask. Jul 10 04:27:18 that is not a mailing list btw Jul 10 04:27:28 and anyways, devs usually prefer mailing list Jul 10 04:28:04 darn, my kvire doesn't do sound alarm anymore :-( Jul 10 04:29:15 brb Jul 10 04:30:20 back Jul 10 05:07:58 03tmbinc 07org.openembedded.dreambox * r5ae89864... 10/packages/linux/linux-dm7020.bb: fix PN for dm7020 linux kernel Jul 10 05:14:05 is there anyway we can find out what packages have been used to build an image ? Jul 10 05:14:14 any log file ? Jul 10 05:17:07 besides reading tmp//temp/log.do_rootfs., there's also tmp/rootfs/usr/lib/ipkg/status, in case you are doing ipkg based images. Jul 10 05:20:41 is there any way to find out what happends when I boot with the bootstrap image and init doesn't seem to do nothing? Jul 10 05:21:51 "nothing"? What's the last thing displayed? Jul 10 05:21:54 several ways Jul 10 05:22:12 set console=ttyS0 and read output on serial Jul 10 05:22:29 or boot manually with init=/bin/busybox sh Jul 10 05:22:48 I have the console on ttyS0.. I see the kernel booting.. than it says freeing init memory 88k and nothing more.. Jul 10 05:23:24 ah, ok I try booting with busybox sh as init.. Jul 10 05:23:29 one reason could be that you don't have hardware floating point nor floating point emulator, but a hard-float userspace. not sure which architecture you have, but on ppc, this will cause exactly that problem. Jul 10 05:23:52 but i assume there are 1000 other ways to do something wrong with exactly the same result Jul 10 05:23:56 I have a pxa270 Jul 10 05:31:23 hum, with busybox sh as init it's the same.. nothing happens.. Jul 10 05:31:30 can I enable softfloat in oe? Jul 10 05:38:47 i had a problem: i had built a image using linux2.6.12rc2 and now Linux2.6.12 has been released. when i build an image with the latest kernel. always my file system contains the old rc2 image. why ?. even i have done clean. but nothing seems to work. is deleting the TMPDIR the only solution. Jul 10 05:39:31 mickeyl: can u clarify on this. please Jul 10 05:41:20 that can have lots of reasons. In that case removing TMPDIR and building again is probably the most sane solution. Jul 10 05:41:43 Hi Zecke Jul 10 05:41:59 hey Jul 10 05:42:04 mickeyl: pybison is crappy Jul 10 05:44:04 mickey: when i delete a linux kernel using -c clean why it should not delete ipks releated to it and delete what ever data it has created during the creation of that ipk. when u clean u assume that what all bitbake has done to produce that package.ipk is undone Jul 10 05:44:32 now bitbake only removed /work/package directory only i assume. Jul 10 05:45:20 eJumbo: right Jul 10 05:45:23 eJumbo: and the stamps Jul 10 05:47:23 but why not the ipks and its associated files. according to me. it should remove what all it has done to create that package. correct me if i am wrong. Jul 10 05:47:35 maybe it should do that Jul 10 05:47:37 it's hard to do Jul 10 05:47:54 cleaning stage and removing ipks would be nice Jul 10 05:48:00 patches appreciated :D Jul 10 05:48:35 is there any documentation about how to setup oe for a specific hardware I can read? Jul 10 05:48:51 Delta_: mickeyls FOSDEM slides? Jul 10 05:49:17 I already have those.. Jul 10 05:49:58 i deleted a TMPDIR and started all over again and i get the following error: http://pastebin.com/310506 Jul 10 05:50:13 Delta: what hardware u got ? Jul 10 05:50:39 guess that error is pretty straightfoward Jul 10 05:50:56 you really don't want me to explain what "ERROR: Cannot satisfy the following dependencies for task-basic: Jul 10 05:50:58 " means, or do you ? Jul 10 05:52:31 a pxa 270 Jul 10 05:53:02 it means that task-basic needs these listed package to be made. am i right ? Jul 10 05:53:17 it cant find them or knows how to make them ? Jul 10 05:54:18 it can't find them Jul 10 05:54:28 obviously they haven't been built Jul 10 05:55:44 but why do i get this error. these come as a part of the kernel sources right. it worked for me the last time why not now ? Jul 10 05:56:40 kernel sources are there. it has compiled it. no issues. as a user what am i supposed to do ? Jul 10 05:56:45 Delta_: take a look at another pxa270 machine and copy that Jul 10 05:57:37 how should we have an idea about what packages you put into your RDEPENDS and what not? Jul 10 05:57:48 you need to stop blindly trying out things Jul 10 05:57:57 you need to start to understand what the hell you are doing Jul 10 05:58:08 you really have used 200% of your support credits here Jul 10 06:00:46 but we dont specify the kernel in RDEPENDS or DEPENDS. it takes which ever kernel it finds in packages/linux where DEFAULT_PREFERENCE != -1 probaly the latest Jul 10 06:01:10 correct me if i am wrong. Jul 10 06:02:16 eJumbo: do you really think it decided it needs a kernel? Jul 10 06:03:02 I tried building the image for the ipaq-270.. it's the same cpu.. but it didn't do neither... Jul 10 06:03:12 but ok, I will read some more.. Jul 10 06:03:18 i mean these modules come part of the kernel sources. this is what i think. Jul 10 06:03:36 eJumbo: start improving your debugging skills Jul 10 06:03:51 eJumbo: you've a MACHINE = "foo" entry in your local.conf Jul 10 06:04:10 yes i have it. Jul 10 06:04:16 Delta_: I need to understand what you want to achieve Jul 10 06:04:32 eJumbo: ask yourself why it is there and what it does Jul 10 06:04:45 Delta_: do you've custom hardware and want to build a distribution? Jul 10 06:05:05 what i mean was it is building the modules which i marked 'm' in the defconfig (.config) which is used to build the kernel. Jul 10 06:05:57 MACINE indicates for what machine ur software will actually run ie target. Jul 10 06:06:16 so it will be built for that machine Jul 10 06:06:24 correct me if i am wrong. Jul 10 06:09:04 zecke I bought the toradex pxa270 board, it's a custom board, containing everything you need to get started, now I try to build my own initrd that loads the cardmgr and switches to a rootfs on compact flash, on that rootfs I want to run tools like dropbear, tinyx, and eventually sablevm.. Jul 10 06:10:51 Delta_: you need a file in conf/machine/yourmachine.conf Jul 10 06:11:26 Delta_: where you can specify your prefered kernel (if any), optimisations (xscale tuning), preferred x server, architecture... Jul 10 06:12:05 Delta_: a bbfile describing how to build your kernel (optional) and a bbfile with dependencies to say what to install into your image Jul 10 06:12:17 I have that file already.. Jul 10 06:12:27 Delta_: I consider it unfriendly to query me Jul 10 06:12:37 Delta_: if you've a concrete idea use pastebin to paset data Jul 10 06:12:41 ~pastebin Jul 10 06:12:41 methinks pastebin is a place to paste your stuff without flooding the channel - try http://pastebin.ca Jul 10 06:12:46 oh, I am sorry.. Jul 10 06:13:17 Delta_: it was your first time Jul 10 06:13:44 so I will use pastebin the next time.. Jul 10 06:14:44 zecke: i am confused what i should do ?. Jul 10 06:15:31 eJumbo: find the answer yourself Jul 10 06:15:36 eJumbo: are you a student? Jul 10 06:15:40 yes Jul 10 06:16:18 u know what my understanding is. am i right or wrong ?. Jul 10 06:17:16 eJumbo: No I just wanted to know if your debugging/analysis skills should be already developed Jul 10 06:17:30 eJumbo: I gave you hint MACHINE Jul 10 06:17:38 why cant it find the modules. these are modules that are needed to support some devices. these come along with kernel sources. Jul 10 06:17:43 eJumbo: what did you tried to do after getting this hint Jul 10 06:18:10 eJumbo: try to understand what you're doing Jul 10 06:20:02 is there any documentation about the variables I can use in the machine.conf file? Jul 10 06:20:10 i checked my local.conf file. MACHINE suggested to me that it is being built for my target machine. and then checked defconfig file i have given some as modules. so i thought these either come with kernel sources (mostly). before when i gave module it worked. i did not give external driver source to bitabake at all. so just wondering why it is not able to find sources of those modules or built modules themselves. Jul 10 06:21:54 Delta_: sadly documentation is still not too good Jul 10 06:22:15 eJumbo: MACHINE -> kernel -> defconfig = you do not understand what you're doing Jul 10 06:22:28 eJumbo: -> is your think chain/logic chain Jul 10 06:22:37 ~lart pybison Jul 10 06:22:38 * ibot runs at pybison with an origami Swiss Army knife, and inflicts a nasty paper cut Jul 10 06:22:47 zecke: i know what i am doing. just that i can understand the error. but since i have just started linux image building and filesystem. i dont know how to debug. if i know something easy to debug. if u know nothing about it. its hard and then i need to turn to people here. thats the problem. Jul 10 06:24:14 eJumbo: did you even look at the content of your machine configuration file? Jul 10 06:24:25 MACHINE is machine (target) mine is omap5912. my kernel sources are cross compiled to give me an image that runs on omap5912 and defconfig is a file which says these are the configuration of the kernel Jul 10 06:24:29 yes Jul 10 06:24:30 eJumbo: and why did you rename to eJumbo frim sriram Jul 10 06:25:11 eJumbo: so your defconfig is embedded in omap5912.conf? I think that is not true Jul 10 06:25:25 no. Jul 10 06:25:28 eJumbo: ask yourself how you get from a omap5912.conf to building a kernel Jul 10 06:27:39 pb_: hey Jul 10 06:27:50 pb_: pybison is crap Jul 10 06:28:18 zecke_: yo Jul 10 06:28:21 zecke_: oh dear Jul 10 06:29:32 pb_: I've one bison problem though Jul 10 06:29:46 pb_: My lexer is reentrant and bison tries to use yylineno Jul 10 06:30:05 pb_: but yylineno is just a define, I need to make bison aware of it Jul 10 06:30:11 %pure_parser does not help though Jul 10 06:31:07 zecke: omap5912.conf says these features are there and u need these modules to use these features. so u need to install them. --BOOTSTRAP_EXTRA_RDEPENDS. but this RDPENDS is already there in my .bb file. Jul 10 06:31:27 i mean these features are supported in osk. Jul 10 06:34:15 eJumbo: sorry if you do not listen to hints I won't waste my time repointing to them Jul 10 06:37:47 i am trying to. but could not get what u wanted to convey. i told my understanding. Jul 10 06:43:21 hm, I should have stayed in amsterdam, today it's even better weather than yesterday.. Jul 10 06:57:01 TARGET_FPU="soft" will do the trick.. Jul 10 06:57:46 oh sorry, could have told you, but missed your question Jul 10 06:58:12 on ARM, you probably have the luck that multilib isn't used and won't screw gcc packaging ;) Jul 10 07:00:38 I'll see what he does :) it takes a long long time here to bitbake an image.. Jul 10 07:01:22 glibc is hell, yes. Jul 10 07:01:57 could not really get it. Jul 10 07:09:52 zecke: i am afraid to ask u any further. but honestly i could not find out what is the problem. Jul 10 07:11:08 Delta_: is there no kernel fpu emulation? Jul 10 07:11:17 you could try that first to make sure that this is the problem Jul 10 07:11:52 it is ipkg problem. and these modules are needed thats all i know. seriously. Jul 10 07:14:20 the kernel makefile doesn't say anything about fpu emu.. Jul 10 07:25:28 strange thing is, when I boot the target with the provided image, and run the init program I made with oe in it, it runs.. I copied the init.sysvinit from the bootstrap image over and ran it succesfully.. Jul 10 07:30:52 eJumbo: seriously giving you more hints is wrong Jul 10 07:30:59 eJumbo: you need to improve your skills Jul 10 07:31:07 eJumbo: I would check if my kernel got built Jul 10 07:31:27 eJumbo: afterwards I would check who needs/RDEPENDS on these modules Jul 10 07:31:41 eJumbo: then I would check if I've built the required modules at all Jul 10 07:41:54 a5I don't understand this.. the binaries I build with oe work on the target, but it doesn't boot with the image, nor can I run busybox sh as init with it.. Jul 10 07:42:52 is there anything special about init I don't know about? Jul 10 07:47:42 Delta_: no Jul 10 07:47:55 Delta_: as long as your libc.so and ld.so is compatible Jul 10 07:51:34 compatible to what? Jul 10 07:53:49 ABI Jul 10 08:00:43 abi? why abi? it's both linux, the only difference is that I compiled the binaries with oe, and the other roofs was build with buildroot.. Jul 10 08:04:27 Delta_: You already heard Soft Float vs Hard Float Jul 10 08:04:36 Delta_: then you could compile your apps against version 2.3.5 of glibc Jul 10 08:04:47 Delta_: and they might not run with earlier versions of glibc Jul 10 08:06:44 I understand that, but I assume oe builds the right glibc with the apps it puts into a bootstrap image.. Jul 10 08:08:23 I would assume that as well Jul 10 08:08:34 * zecke_ rereads what Delta_ wrote Jul 10 08:09:13 Delta_: indeed this sounds weird Jul 10 08:11:43 does the kernel have any logging output of what he does when he tries to run init? Jul 10 08:16:44 Delta_: not that I know of Jul 10 08:17:09 Delta_: do you've floating point emulation compiled as module or such? Jul 10 08:29:58 there is nothing about floating point in the kernel config.. Jul 10 08:36:09 doesn't every app that needs floating point arithmetic do it itself in software when I build it with softfloat enabled, so it's independent from the fact if there is fpu support in hardware or in the kernel? Jul 10 08:36:53 yes Jul 10 08:37:04 when you build soft-float, you don't need any kernel stuff Jul 10 08:40:17 when I make a change in the machine.conf , can I just run bitbake again, or do I have to delete any temp. files to make sure it builds everything again? Jul 10 08:40:37 no need. Jul 10 08:40:53 ok.. Jul 10 08:41:20 u can run it. please clean ur image files first using bitbake -c clean just to ensure they dont use the old copy. Jul 10 08:42:14 zecke: ran out of ideas. really cant make out how to figure out whats the problem Jul 10 08:44:17 eJumbo: It might be best to send me your OSK Jul 10 08:44:43 eJumbo: what of the stuff I told you do check did you check? Jul 10 08:44:48 yes Jul 10 08:44:56 kernel is built Jul 10 08:45:30 the modules listed are from BOOTSTRAP-RDEPENDS which are in my machine.conf file Jul 10 08:46:28 Delta: erm not sure, but if you change stuff like TARGET_FPU, i'm not sure that everything required (which will be about everything) is being rebuilt. Jul 10 08:46:39 i have given all modules to be within kernel. all drivers built into kernel. Jul 10 08:48:47 checked for each package to determine where it is built could not find it in packages. perhaps they are drivers. Jul 10 08:49:05 built in the kernel source code. Jul 10 08:50:20 zecke: beyond this i cannot really understand. i assure you i tried my best. i know i am not good enough to be doing what i am doing. but i will try to improve my skills. Jul 10 08:52:20 "module to be within kernel" Jul 10 08:52:30 that is a contradiction Jul 10 08:53:05 eJumbo: see I try to help you to understand the problem yourself Jul 10 08:53:15 what i mean is the drivers are build with kernel. in the defconfig if u specify =y instead of =m then they are built within the kernel. Jul 10 08:54:22 ok Jul 10 08:54:27 eJumbo: so you won't build modules -> no modules packages are built Jul 10 08:54:32 eJumbo: no packages to install? Jul 10 08:54:46 eJumbo: solution if you do not build modules, do not RDEPEND on them? Jul 10 08:54:47 yes Jul 10 08:55:20 so i removed BOOTSTRAP_RDEPENDS from my .bb file. but still getting the same error Jul 10 08:55:40 eJumbo: touch conf/local.conf? Jul 10 08:58:18 ok Jul 10 09:01:48 tmbinc: so you say I should remove target files by hand before building? just to be on the safe side? Jul 10 09:11:16 zecke: still get the same error Jul 10 09:13:24 eJumbo: man grep Jul 10 09:13:52 eJumbo: I've learned C with grepping the kernel for certain strings when I was 12 Jul 10 09:14:27 grep what. i grep those modules in packages. but could not find any. Jul 10 09:14:29 eJumbo: grep for the module names? (skip the kernel-module-) Jul 10 09:15:03 eJumbo: if it wants to install the the module Jul 10 09:15:09 eJumbo: someone RDEPENDS on them Jul 10 09:15:25 <[cc]smart> is there sbdy actually maintaining cyrus, or can i feel free to fiddle with it as to my likings ? Jul 10 09:15:58 [cc]smart: fiddle with it create a patch put into our bugzilla Jul 10 09:16:04 [cc]smart: wait a day, get maintainer Jul 10 09:16:16 <[cc]smart> i'm on nslu monotone Jul 10 09:16:28 <[cc]smart> but i was recommended to check back on more intrusive changes Jul 10 09:16:29 <[cc]smart> so i do Jul 10 09:20:17 Zecke: these modules are not found in packages. only found in defconfig file of the KERNERL Jul 10 09:22:23 Delta_: remove whole TMPDIR, yes. you need to rebuild mostly everything anyway. Jul 10 09:22:48 I am at 54% rebuilding :) Jul 10 09:24:33 03zecke123 * r280 10bitbake/bin/bitbake: Jul 10 09:24:33 trunk/bitbake/bin/bitbake: Jul 10 09:24:33 -Fix findBestProvider to work with BB collections (hopefully) Jul 10 09:24:33 without adding new regressions. Jul 10 09:24:33 This fixes bug #137 the patch is courtsey Seb (proti) Jul 10 09:26:40 03zecke123 * r281 10bitbake/contrib/vim/ftplugin/: Remove Empty directory Jul 10 09:29:32 zecke_: very good Jul 10 09:30:30 eJumbo: sorry you wasted my time too much today Jul 10 09:30:56 sorry zecke Jul 10 09:30:56 eJumbo: A little quiz: Jul 10 09:31:04 eJumbo: What is a bbfile? What is it for? Jul 10 09:31:20 eJumbo: what does INHERIT += "package_ipk" do? Jul 10 09:31:24 eJumbo: What is a task? Jul 10 09:31:32 eJumbo: What is a ipk Package Jul 10 09:31:41 eJumbo: How does one create an image? Jul 10 09:31:44 c'mon zec Jul 10 09:31:46 eJumbo: what is DEPENDS, RDEPEND? Jul 10 09:32:08 eJumbo: if it wants to install the module, someone RDEPENDS on it Jul 10 09:32:26 eJumbo: grep is an awesome tool to recursively find certain things Jul 10 09:32:42 .bb file is a bitbake (bb) conf file where u specify what u want; dependpencies(DEPENDS) and what u want to install (RDEPENDS) Jul 10 09:32:46 eJumbo: During the day you showed 0% of knowledge 0% of improving 0% of actually trying to improve Jul 10 09:33:18 eJumbo: specially the last point is disappointing Jul 10 09:33:35 ipk is a package built out of an ipkg-build utility Jul 10 09:34:06 and ipkg-build utilities contain goedel numbers for packages? Jul 10 09:34:18 out of an/using Jul 10 09:34:20 creating an image ; inherit image-ipk Jul 10 09:34:29 no Jul 10 09:34:52 if u specify that it will result in creating an image. Jul 10 09:35:16 INHERIT more_magic Jul 10 09:35:35 eJumbo: what is IPKG_INSTALL? Jul 10 09:35:43 sorry using an ipkg-build utility Jul 10 09:36:45 dont know zecke Jul 10 09:37:20 zecke: i tried to improve apply myself. Jul 10 09:37:45 thanks zecke for ur help. Jul 10 09:38:05 eJumbo: what do you study? what semester? Jul 10 09:38:14 dont say 0% improvement. actually i tried to improve myself Jul 10 09:38:15 eJumbo: use grep -rni modulename * Jul 10 09:38:34 did that grep -Ri Jul 10 09:39:31 modulename != kernel-module-modulename Jul 10 09:40:14 zecke; i will paste my errors in paste.bin and am leaving as it is too late now. 10PM Jul 10 09:40:20 got to go home Jul 10 09:40:26 cya Jul 10 09:41:55 zecke_ what did you study? Jul 10 09:42:41 Delta_: I'm an undergraduate Jul 10 09:43:30 http://pastebin.com/310661 Jul 10 09:43:40 please help if u get any idea Jul 10 09:43:43 thanks for ur time. Jul 10 09:43:59 i will improve my skills and thanks once again. Jul 10 09:44:51 bye Jul 10 09:45:11 see u tommorrow. Jul 10 09:45:33 please reply i will check in the chat logs file early morning. Jul 10 09:49:07 ah, ok, going for a bachelor? Jul 10 09:54:58 Delta_: no the German Diplom Jul 10 09:55:48 proti: thanks for the patch Jul 10 09:56:42 I am doing that too.. but the dipl.ing.fh not uni.. Jul 10 10:31:48 Delta_: do you still have the problem? Jul 10 10:35:46 I am still compiling.. when it has finished I can tell if the problem is still there.. Jul 10 11:03:26 03zecke123 * r282 10bitbake/ (8 files in 3 dirs): Jul 10 11:03:26 Prepare for the new tokenizer C-based parser as optional parser. Jul 10 11:03:26 For now import bb.parse will function the way it used to work Jul 10 11:19:31 hell, tailgate just fell off my car. Jul 10 11:22:09 :( Jul 10 11:23:15 yeah, that was my reaction as well. Jul 10 11:24:00 very annoying Jul 10 11:24:23 not to mention somewhat embarrassing. Jul 10 11:25:24 Were you going at any speed? Jul 10 11:26:21 no, only about 5mph. Jul 10 11:26:33 quaility american product? Jul 10 11:29:39 03zecke123 * r283 10bitbake/lib/bb/parse/parse_c/ (. bitbakeparser.l bitbakeparser.y): (log message trimmed) Jul 10 11:29:39 Start with the excellent lexer and lemon Grammar from Jul 10 11:29:39 Marc Singer and turn it into a C module (after pybison Jul 10 11:29:39 was a total failure). Jul 10 11:29:39 Lexer Changes: Jul 10 11:29:41 Do not allow unquoted R_VALUE Jul 10 11:29:43 Grammar Changes: Jul 10 11:34:10 treke|home: heh. au contraire, quality british product. Jul 10 11:36:11 pb, what what your answer for br #140 all about? Jul 10 11:36:23 incinerator: er, the bug. Jul 10 11:36:24 ^2nd what|was Jul 10 11:36:30 zroadmap Jul 10 11:37:53 treke|home: seems that some joker had unfastened the pins that hold the tailgate shut, and I had neglected to check that it was secure before driving off. normally, that doesn't cause any great drama: it just drops open, and I have to get out and close it up again. But, this time, it seems that the sudden impact when it fell open was enough to tear off both the hinges, so it dropped off completely. Jul 10 11:38:51 which is a bit weird: the metal wasn't visibly corroded. Must just have been fatigue, I suppose. Jul 10 11:39:49 incinerator: I don't understand what you're asking. Jul 10 11:41:48 Are there any known issues with compiling opie-pcmciaapplet? I'm having a bitbake opie-image fail on it; monotone tree updated this evening (was doing it before that though), building for Zaurus Spitz. Jul 10 11:42:44 well, your comment is alright, but if i don't change anything i cannot link anything Jul 10 11:43:10 i need that patch or else the package won't build, otherwise I wouldn't have posted the br Jul 10 11:43:34 Noodles: I had the same issue a while ago Jul 10 11:43:40 pb_: ouch Jul 10 11:43:47 was a bug in the code, wrote a patch and it worked Jul 10 11:44:09 however, that must have been fixed in cvs a while ago Jul 10 11:44:23 i did a clean build yesterday and it compiled straight through Jul 10 11:45:16 Noodles: has CVSDATE changed when you did the monotone update? Jul 10 11:45:40 Oh, possibly not. I certainly didn't do anything to change it. Jul 10 11:45:55 RP: ping Jul 10 11:46:02 20050627, so no. Jul 10 11:46:05 k, then it probably still fetches the wrong code from upstream cvs Jul 10 11:46:44 i compile oz-3.5.4 for c7x0, that has the fix in it already, so it seems Jul 10 11:47:23 if you look at the compile errors you could probably fix it yourself though Jul 10 11:47:52 wasn't difficult, the code called a function whose name had changed in the library it imported.... Jul 10 11:48:03 renamed the function call => worked Jul 10 11:48:26 oh, my CVSDATE is 20050704 byraway Jul 10 11:49:19 unfortunately, I don't remember the exact details anymore, and I have delete the patch :-( Jul 10 11:50:25 pb: anyways, I know that my patch won't really fix the bug. but my box can't be arsed to do the automatic lookup you suppose it should Jul 10 11:50:49 maybe it's just that my computer doesn't like you and wants me to make your life miserable with bogus BRs *ggg* Jul 10 11:54:02 arg, does anybody know how to tell bugzilla to display local zone time instead of some 'murrican zone time I cannot figure out? Jul 10 11:54:08 or maybe UTC Jul 10 11:54:26 what is responsible for auto-mounting MMC cards on 2.6? hotplug, udevd or both? Jul 10 11:54:49 udev only creates the device link, everything else has to be done by other software Jul 10 11:54:49 mickeyl: hotplug Jul 10 11:55:06 mickeyl: this is what mmc.agent normally does Jul 10 11:55:07 ok, then i'll try to find out why it's not working here Jul 10 11:55:18 though, theoretically you _could_ use RUN="whatever" to execute a script that mounts it *ggg* Jul 10 11:56:36 shouldn't udev create the device node when i insert the mmc ? Jul 10 11:56:53 mickey: just caught another bug Jul 10 11:57:06 yes Jul 10 11:57:11 yes, it should Jul 10 11:57:16 depending on its ruleset Jul 10 11:58:39 mickey: something's bonky with c7x0 Jul 10 11:58:44 indeed Jul 10 11:58:55 irda is broken on 2.6.12, SD no longer works Jul 10 11:59:03 bluetooth OOPSes like hell when you remove the modules Jul 10 11:59:05 tsss Jul 10 11:59:06 I just did an opie-image build with a hacked ROOT_FLASH_SIZE=23 Jul 10 11:59:18 image had the same size as for ROOT_FLASH_SIZE=32 Jul 10 11:59:29 though judging distro configs etc. it should not Jul 10 11:59:36 should be bigger Jul 10 11:59:44 did you rebuilt everything? Jul 10 11:59:49 i.e. rm -rf $TMPDIR ? Jul 10 11:59:54 ca. 18MB mebbe, but now it is only 14MB Jul 10 11:59:57 yes I did Jul 10 12:00:02 rm -rf tmp Jul 10 12:00:13 newest bitbake revision 281 etc. Jul 10 12:00:24 even removed all patches from my tree except the one for hotplug Jul 10 12:00:29 yeah. in that case, it looks like something has been broken Jul 10 12:00:47 pb: I'll post the compile log etc. for bug 140# rsn Jul 10 12:01:05 strangely enough, it worked a couple of days ago Jul 10 12:01:32 you know what that means... we need more automated regression tests... asap. Jul 10 12:02:53 well, I'd just be happy if i could groke this whole bitbake stuff Jul 10 12:02:59 ^groke|grok Jul 10 12:04:12 I'll restart a build and have a look Jul 10 12:04:15 bbl Jul 10 12:06:21 oh Jul 10 12:07:33 pb: attachment is filed at bugzilla Jul 10 12:10:38 pb you still here? Jul 10 12:44:52 mickey|24: ping? Jul 10 12:47:00 mickey|24: oh.. 24... aha. When you're back: I keep running into the same problems with packages requiring users. But adding them to base-passwd isn't really a good solution, since bumping the rev won't add the users to the already-installed passwd base... Jul 10 12:48:50 mickey|24: So I'm thinking.. What if there is a list of reserved UID/GIDs in base-passwd, that maintainers can check, find a free UID/GID, and thus make it safe to echo > /etc/passwd during ipkg install? Jul 10 12:50:33 There can even be an example of how to do that in do_install in that README. That checks passwd to see if it's there, then just echos if it isn't already in there.. Jul 10 12:53:30 depending on the developer, not all use more than one user Jul 10 12:54:11 mickey|24: are you only developing wellenweiter for oz now, or are there your updates avail for the sharp rom too? Jul 10 12:54:24 * france is back (gone 22:19:49) Jul 10 12:56:56 RP: ping Jul 10 12:57:38 emte: what do you mean? Jul 10 12:58:28 on a pda there really is no need for multiple users Jul 10 12:58:54 emte: well, do you really need mysql and cyrus-imapd on a pda? Jul 10 12:59:09 root, zaurus, maybe more if you want to run apache+mysql+php Jul 10 12:59:15 depending on application yes Jul 10 12:59:54 but you can set those to a specific user Jul 10 13:00:29 it works! Jul 10 13:00:39 it really works.. Jul 10 13:00:56 why bother changing the user to match your system? just use whatever it wants, less work Jul 10 13:01:13 if they ever have updates, then you'd have to recall what you did earlier Jul 10 13:01:23 emte: Any good suggestions then? openntpd requires a user, defined with configure. Jul 10 13:01:41 try root Jul 10 13:01:45 lol Jul 10 13:01:51 i think subdue is ahead of you Jul 10 13:02:01 you know root will exist Jul 10 13:02:05 it needs to update things that root has access for Jul 10 13:02:29 Kinda makes the whole point of privilege separation useless though :P Jul 10 13:02:52 so sudo when you run it Jul 10 13:03:35 subdue: uh.. openntpd switches to an unprivileged user. If that user is UID 0, the point is kinda... lost Jul 10 13:03:52 NAiL, aside from bluetooth, what need do you have for that on a pda? Jul 10 13:04:23 unless you like to lose your device Jul 10 13:04:29 shrug Jul 10 13:04:32 whichever Jul 10 13:05:13 emte: I'm not the one running a pda, but I can see use for openntpd if I care to run cyrus-imapd Jul 10 13:05:25 on a desktop/workstation/server the issue IS relevent because you are always connected to a network from a static(or semi-static) location Jul 10 13:05:55 hence my BT exception Jul 10 13:06:48 yes, but do you really care if there's a few lines added to /etc/passwd? Jul 10 13:07:15 if the lines overwrite ones already there, yes Jul 10 13:07:31 emte: that's the whole point. It's not supposed to Jul 10 13:07:49 That's what I'm trying to achieve Jul 10 13:08:04 Making *sure* that nothing is overwritten or duplicated Jul 10 13:08:16 then you need to modify ipkg or the install meta to check Jul 10 13:09:06 It's kind of an ardous task to undertake though... Jul 10 13:09:13 why? Jul 10 13:09:38 if you do it through the meta you would just diff and do soemthing based on output Jul 10 13:09:58 or you could grep Jul 10 13:10:06 looking for the user Jul 10 13:10:22 if it exists do nothing, if it doesnt add Jul 10 13:10:23 I can see a solution using grep, but finding a free uid with grep? Jul 10 13:10:44 you dont need to find the uid Jul 10 13:10:49 just the username Jul 10 13:10:54 uh? Jul 10 13:11:09 it will autoassign the next uid if you add a user Jul 10 13:11:30 emte: You are aware that busybox adduser is interactive? Jul 10 13:11:35 and use the existing uid if the user exists Jul 10 13:11:44 you can override that Jul 10 13:11:51 oh? How? Jul 10 13:11:57 use the flags luke Jul 10 13:12:05 which flags? :) Jul 10 13:12:28 you ever run a server without a gui? Jul 10 13:12:43 adduser/useradd have flags associated with them Jul 10 13:12:47 yes Jul 10 13:12:53 do "adduser --help" Jul 10 13:12:56 specifiing every feild Jul 10 13:13:01 or man Jul 10 13:13:04 eitehr works Jul 10 13:13:28 adduser [options] [--home DIR] [--shell SHELL] [--no-create-home] [--uid ID] Jul 10 13:13:28 [--firstuid ID] [--lastuid ID] [--ingroup GROUP | --gid ID] [--disabled-pass- Jul 10 13:13:28 word] [--disabled-login] [--gecos GECOS] user Jul 10 13:13:58 hmm Jul 10 13:14:03 why don't I get that... Jul 10 13:14:06 man Jul 10 13:14:16 adduser: unrecognized option `--disabled-login' Jul 10 13:14:21 its part of info i belive Jul 10 13:16:01 prefer useradd myself ... Jul 10 13:16:29 small diiferences tho Jul 10 13:17:34 * NAiL realises the problem Jul 10 13:17:45 adduser comes from tinylogin Jul 10 13:17:50 on openslug Jul 10 13:18:43 its just a script Jul 10 13:20:37 useradd? Jul 10 13:21:14 er wait ... adduser is the script Jul 10 13:21:19 useradd is a binary Jul 10 13:21:24 bad me Jul 10 13:22:06 oh well... with tinylogin there's only adduser (binary). And that is a problem ;) Jul 10 13:23:28 you should be able to do: adduser --disabled-password test Jul 10 13:23:38 and not get an interactive Jul 10 13:24:26 neat Jul 10 13:24:45 that good enough for you? Jul 10 13:25:20 almost ;) Jul 10 13:25:24 lol Jul 10 13:25:51 It doesn't let me specify homedir. But that's fixable in other ways Jul 10 13:26:07 adduser [--home DIR] Jul 10 13:26:09 ie, --home doesn't work Jul 10 13:26:11 sure it does Jul 10 13:26:26 root@gunslinger:/mnt/var# adduser --disabled-password --home /ttt test2 Jul 10 13:26:26 Tinylogin v1.4 (2005.07.01-01:45+0000) multi-call binary Jul 10 13:26:43 adduser --home=/ttt --disabled-password test Jul 10 13:27:08 it would try to read /ttt as a user i'd guess Jul 10 13:27:11 man, you're the MAN Jul 10 13:27:12 ;) Jul 10 13:28:24 and --ingroup works Jul 10 13:28:38 * NAiL hugs emte Jul 10 13:29:42 hmm where did esmart_conatainer.h go ... Jul 10 13:45:32 é Jul 10 13:53:09 mickey|24, or any other python*.bb maintainer: is there a reason why libpython isn't built? Jul 10 13:53:17 (at least not as a shared library) Jul 10 13:55:58 tmbinc: it works now, my system boots and everything looks fine.. Jul 10 13:56:22 Delta_: ah fine. what was the problem? Jul 10 13:56:44 the softfloat.. and me not deleting the tmp directory the first time.. Jul 10 13:56:58 ah ok Jul 10 13:57:06 btw what board do you use exactly? do you have any link? Jul 10 13:58:06 I use the toradex colibri: Jul 10 13:58:09 http://www.toradex.com/e/products.html Jul 10 13:59:38 ah Jul 10 14:00:27 it's cheap, it's 99 euros.. and the devkit is only 300 euros.. Jul 10 14:01:01 but I am off to bed now.. I spent enough time in front of the computer today.. Jul 10 14:01:04 goodnight.. Jul 10 14:06:23 tmbinc: builds fine here. Jul 10 14:09:22 subdue: sorry, no updates for the SharpROM nor for SharpROM clones. I have lost interest in everything compiled with gcc 2.95.3 Jul 10 14:10:06 ok Jul 10 14:10:22 thanks for the reply. wasn't sure what to put up on elsix in regards to that Jul 10 14:10:43 I was hoping the only issue would be the wrapper Jul 10 14:11:28 well, the whole thing started out with Sharp not having wireless extensions in the kernel and with insisting on a brain dead wireless driver Jul 10 14:11:49 ah Jul 10 14:12:03 I can see how that's be a hassle Jul 10 14:12:05 then i've become dragged into OpenZaurus and I have no time to think about the old things. from an open source perspective the sharprom and all its clones are a dead end Jul 10 14:12:34 i wanted to make a closed source Wellenreiter/SharpROM to sell Jul 10 14:12:42 ah Jul 10 14:12:49 but then sharp pulled the trigger in everything but japan Jul 10 14:12:52 and the market vanished Jul 10 14:13:02 mickeyl: i have looked at this issue with the image size Jul 10 14:13:04 with it the least bit of my motivation to do anything for sharprom Jul 10 14:13:15 i think it must have something to do with the opie-collections.inc Jul 10 14:13:29 i guess that lookup to get the value of ROOT_FLASH_SIZE goes wrong Jul 10 14:13:31 yeah, that was bothersome for us. on the other side though, the 3k+ are all sharp roms, so if you created something for the sharp rom side, it'd work on those Jul 10 14:14:08 because default return value is 16, so i guess that bb.getvar doesn't see the envvar Jul 10 14:14:28 pb_:ping Jul 10 14:16:12 subdue: true, however i'm already doing so many open source things for the modern systems, there is no way i can do even more. to the contrary, i will have to give up a lot of things soon, because my contract in university ends and then I have orders of magnitudes less time for open source. I may have do most of things commercially soon. Jul 10 14:16:33 understood Jul 10 14:17:00 I like eongame's stuff - if you created stuff like that, I'm sure lots would like it :) Jul 10 14:17:18 unfortunately, when sharp pulled things, he stopped having interest too Jul 10 14:20:53 03zecke123 * r284 10bitbake/lib/bb/parse/parse_c/ (5 files): Jul 10 14:20:53 More things done on the Bitbake C parser: Jul 10 14:20:53 -Implement token_t and switch the grammar actions Jul 10 14:20:53 to the new class Jul 10 14:20:53 -Start with the lex_t as well Jul 10 14:20:54 -Start with implementing the high level python Jul 10 14:20:56 side. Jul 10 14:22:29 subdue: yeah, it makes no sense to run a business when there's not really a market. The plethora of images and distributions makes it even worse Jul 10 14:22:43 the irony is that i'm also responsible for that.. but hey... that's life :) Jul 10 14:22:49 LOL that's true Jul 10 14:23:03 i have faith in two things Jul 10 14:23:19 a) the mobile phone/pc from road-gmbh.de Jul 10 14:23:30 b) Nokia Jul 10 14:23:38 yeah the 770 Jul 10 14:23:38 so lets see how that works out Jul 10 14:23:54 I dont think the work here will be wasted with that coming out Jul 10 14:23:58 they already have a dev kit Jul 10 14:25:26 yeah, that phone from road-gmbh is cool Jul 10 14:25:45 mickeyl: python - yes, it builds, but i have another application which wants to link against libpython. does libpython*.so build for you? Jul 10 14:25:48 i can see me working on a new Wellenreiter edition for that Jul 10 14:26:04 tmbinc: yes, that's what I wanted to say with 'it builds' Jul 10 14:26:10 2.3 or 2.4? Jul 10 14:26:10 it = the library. some weeks ago i changed that Jul 10 14:26:13 2.4.1 Jul 10 14:26:16 i didn't change it for 2.3 Jul 10 14:26:19 ah ok Jul 10 14:26:35 i'll just switch to 2.4 then Jul 10 14:26:42 Hmm... is it possible to set up so that an ipkg executes commands after install? Jul 10 14:27:09 NAiL: sure. we do that in a lot of packages Jul 10 14:27:11 postinstall scripts Jul 10 14:27:31 pkg_postrm() { Jul 10 14:27:35 your commands here Jul 10 14:27:36 } Jul 10 14:27:41 pkg_postinstall() { Jul 10 14:27:43 foo Jul 10 14:27:43 } Jul 10 14:27:56 some of our classes do that automatically, i.e. have a look at update-rc.d Jul 10 14:28:07 ok, thanks a bunch :) Jul 10 14:33:23 does anyone know how those network connection monitors work in principle? i.e. on mandriva i have an icon in the taskbar that automatically changes its state when I 'ifup' and then some bytes go through. How does it determine if I'm really "connected". I mean the interface could go up but without any bytes coming in. Jul 10 14:34:12 there's some info on the sharp side I saw a while back (re kismet) if you were interested in that Jul 10 14:34:16 it will probably check if an ip address is set Jul 10 14:34:29 hmm... but an ip address doesn't mean anything Jul 10 14:34:33 true Jul 10 14:34:50 even a default route is meaningless until a bit comes in (as opposed to go out) Jul 10 14:35:00 i don't think they're using libpcap or so to investigate traffic Jul 10 14:35:29 usbctl on 1 Jul 10 14:35:29 sleep 1 Jul 10 14:35:29 wlanctl-ng wlan0 lnxreq_ifstate ifstate=enable Jul 10 14:35:29 ifconfig wlan0 1.1.1.1 up Jul 10 14:35:37 that's what they do on the sharp rom Jul 10 14:35:41 subdue: yeah i know about these things. like i said, I _could_ make Wellenreiter work on SharpROM clones, but I'm just not interested Jul 10 14:35:46 ok Jul 10 14:36:13 wasn't trying to convince you, just wasnt sure if that was new info or not Jul 10 14:36:46 Wellenreiter already contains a backend for wlan-ng. For the sharprom you would just need to remove the WE check and hard code it to wlan0 then it would work out of the box Jul 10 14:37:13 heh Jul 10 14:41:15 all in all I won't touch the current code base any more. Wellenreiter was my first full-fledged application based on Qt/Embedded 2.x and I have learned a lot since these days. I'd rather rewrite it from scratch now that I have some more clues about how to do it right :) Besides, it could benefit a lot from Qt4. Jul 10 14:46:14 mickeyl: about that network stuff, you don't mean "link state", like ifplugd? (http://www.stud.uni-hamburg.de/users/lennart/projects/ifplugd/, probably boils down to polling with SIOCETHTOOL). Jul 10 14:47:10 otherwise you could poll ifconfig stats (not sure which ioctls) Jul 10 14:48:24 tmbinc: ah righto. that's probably how they do it. I guess there's nothing like a push way-of-things here Jul 10 14:48:48 perhaps I could convince the netlink layer to call me when a packet arrives Jul 10 14:49:47 mickeyl: yes, you can ask netlink to deliver an event on interface up/down, address add/delete, routing changes, all that kind of thing. Jul 10 14:50:03 pb_: excellent. I need to have a closer look at RT_NETLINK Jul 10 14:50:56 i'm still a bit paranoid about all the polling that goes on in our software Jul 10 14:51:04 like... battery-wise... Jul 10 14:51:44 yeah, polling is always bad news Jul 10 14:52:21 of course, but it's probably once every few seconds. but sure, polling is bad. Jul 10 14:52:26 unfortunately it's the simplest solution and not only beginners get trapped into using it to make things work asap Jul 10 14:52:52 * mickeyl won't count the places where Qtopia polls Jul 10 14:53:01 wow, really? Jul 10 14:53:19 those crazy trolls Jul 10 14:54:04 almost all applets poll individually, libqpe polls apm, interfaces, mountpoints, qte seems to poll as well instead of using a select loop in the innards Jul 10 14:54:52 freaky Jul 10 14:55:04 well, admittedly, for apm you have no choice but to poll Jul 10 14:55:05 I'm pretty sure the whole thing could be snappier and more durable batterywise when all those APIs would be converted to be event based Jul 10 14:55:21 pb_: isn't keventd working on a solution for that? Jul 10 14:55:22 mickeyl: well, the screenshots do look pretty :) Jul 10 14:55:24 I'm not sure if you can get events for mountpoint things. Jul 10 14:55:26 subdue :) Jul 10 14:55:54 if you do remove the wrapper, I'll be glad to see if it works on my 6k with the sharp rom Jul 10 14:56:20 I've been pretty good about feedback on stuff Jul 10 14:56:22 pb_: there's quite some movement going on to finalize a generic userland kernelland interface and even apm would benefit from that. Jul 10 14:56:29 yeah Jul 10 14:56:39 i just hope the big guys can eventually agree Jul 10 14:56:50 * pb_ afk again for a while Jul 10 14:56:55 subdue: who knows. I'll ring when there's anything in place Jul 10 14:57:05 * mickeyl going to bed Jul 10 14:57:09 g'night folks Jul 10 14:57:14 sure. nite :) Jul 10 14:58:17 pb_: I think one can get kevents for mounts nowadays Jul 10 14:58:28 pb_: or hotplug could send something over d-bus or similiar Jul 10 14:59:30 03zecke123 * r285 10bitbake/lib/bb/parse/parse_c/ (bitbakeparser.l bitbakeparser.py bitbakeparser.y lexer.h): Getting closer... Jul 10 16:19:36 keturn: ping? Jul 10 19:21:52 hey raster Jul 10 19:22:07 i have some light reading for ya in the portal Jul 10 20:14:03 moo Jul 10 20:14:11 emmipoos Jul 10 20:14:16 there you are Jul 10 20:14:24 i deny that allegation! Jul 10 20:14:25 lies Jul 10 20:14:27 lol Jul 10 20:14:31 spread by my competitors Jul 10 20:14:42 ecore_event_handler_add(ECORE_EVENT_SIGNAL_EXIT, sig_exit_cb, NULL); Jul 10 20:14:48 that look wrong to you? Jul 10 20:15:00 this is out of the cookbook btw Jul 10 20:15:17 sig_exit_cb is giving me some issues Jul 10 20:17:20 sig_exit_cb is a function ... but adding the brackets gives ... hmm i need to typecast as a pointer Jul 10 20:20:08 loosk fine Jul 10 20:20:34 that is correct Jul 10 20:20:37 what u wrote Jul 10 20:21:00 i assume your sig_exit_cb function is prototyped correctly and defined before u call that fn call Jul 10 20:21:04 at least the prototype is Jul 10 20:21:07 the fn can be anywhere Jul 10 20:21:08 yeah Jul 10 20:21:21 but it at least needs to KNOW its going to be a fn to look up later Jul 10 20:21:30 error: `sig_exit_cb' undeclared (first use in this function) Jul 10 20:21:55 if i add the brackets that goes away but i get a intger to pointer warning Jul 10 20:24:21 no Jul 10 20:24:24 dont add brackets Jul 10 20:24:30 its because u doidnt DECLARE the fucntion Jul 10 20:24:37 see how header files DECLARE functions Jul 10 20:24:43 like void blah(void); Jul 10 20:24:50 yeah Jul 10 20:24:52 its there Jul 10 20:24:58 that tells c "there is a funciton called blah - it takes no arguments and returns nothing" Jul 10 20:25:01 it doesnt say where it is Jul 10 20:25:03 just that it exists Jul 10 20:25:10 its liek declaring variables before u use them Jul 10 20:25:13 declarer functiosn before u use them Jul 10 20:25:23 is it BEFORE the callback add Jul 10 20:25:27 thats the thing Jul 10 20:25:30 bet u it isnt Jul 10 20:25:38 er wait no its not ... me beats self with the stupid stick Jul 10 20:25:53 i swear i had checked Jul 10 20:26:42 * raster beats emte witht the stupid stick too Jul 10 20:28:15 i must have deleted them when i removed those stupid gotos Jul 10 20:33:10 yeah that was it Jul 10 20:38:31 hmm edc parse error ... Jul 10 20:39:23 aha format change Jul 10 21:04:29 hmm Jul 10 21:04:40 frkn lib problem .... Jul 10 21:13:50 well trhats good it all behaves Jul 10 21:24:09 now to get the client server stuff to work with it **** BEGIN LOGGING AT Sun Jul 10 23:36:06 2005 **** BEGIN LOGGING AT Sun Jul 10 23:49:56 2005 **** ENDING LOGGING AT Sun Jul 10 23:59:57 2005