**** BEGIN LOGGING AT Sat Jun 25 23:59:56 2005 Jun 26 01:06:42 morning Jun 26 01:36:56 i gave a bitbake bootstrap-image and left and when i came back linux hanged; nothing was working keyboard. so i rebooted the system. how do i know has the image built or how much has been built. Jun 26 01:40:05 hello... Jun 26 01:45:41 hello how do i know if everything is completely built and where do i find error logs if i found any during building.. Jun 26 01:47:10 morning all Jun 26 01:47:52 sriram: Any image would be in the tmp/deploy directory Jun 26 01:49:17 i have a directory by name ipk and so many files inside it. Jun 26 01:49:31 what is the name of the image file. Jun 26 01:49:50 sriram: That's normal. There would also be an images directory if it had got up to building one Jun 26 01:50:12 It will resume from where it got up to if you restart it Jun 26 01:50:37 so. i started all over again and it is right now building the busybox now... Jun 26 01:50:43 how many more to go ? Jun 26 01:50:47 any ideas ? Jun 26 01:51:31 It won't be starting over again - it will continue where it left off... Jun 26 01:51:50 and no idea how many to go.... Jun 26 01:52:05 ok where will be the error log ?. the error when the last build stopped ? Jun 26 01:52:42 In the tmp/work/xxx/temp directory where xxx is the package it was up to... Jun 26 01:53:02 there is no single log file ? Jun 26 01:53:20 No Jun 26 01:53:59 i have stored my oetmp in /tmp directory. so will the system slow down if i stored in /tmp directory ? Jun 26 01:54:24 i read some where if you store large amts of data in /tmp dir then ur system will be slow ? Jun 26 01:55:27 I doubt it would make that much of a difference... Jun 26 01:55:35 the build process is very slow... so i suspect Jun 26 01:55:52 Its just slow I'm afraid Jun 26 01:55:54 it is over 12 hours and still it is building ? Jun 26 01:56:03 What speed machine? Jun 26 01:56:09 pentium 4 Jun 26 01:56:22 1.2 GHZ or more Jun 26 01:56:50 That's probably about right for a 1.2... Jun 26 01:57:04 It takes my 2.8Ghz twin Xeon about four hours... Jun 26 01:57:25 what does bootstrap-image do ? Jun 26 01:57:54 Gives you an image that should give you a shell prompt on your machine of choice Jun 26 01:58:45 if i give gpe-image and bootstrap-image do you know how many more packages it need to download. Jun 26 01:59:00 i guess it would need only gpe + gtk and xserver Jun 26 01:59:04 am i right ? Jun 26 01:59:20 kind of Jun 26 02:00:08 if i dont want bitbake to download all the packages all over again. what directories should i save. Jun 26 02:00:16 should i save TMPDIR Jun 26 02:00:34 Just kill the current build and run bitbake gpe-image instead Jun 26 02:01:11 suppose i reboot the system. i want to store the downloaded directories somewhere right. Jun 26 02:01:18 i dont want to start all over gain. Jun 26 02:01:24 so what do i need to do ? Jun 26 02:01:29 what all should i save ? Jun 26 02:01:38 Your /tmp directory is wiped over a reboot? Jun 26 02:01:50 Just leave all the files alone Jun 26 02:02:13 if i give gpe-image. will it start all over gain ? Jun 26 02:02:14 RP: Morning. Jun 26 02:02:30 sriram: No, it shouldn't Jun 26 02:02:33 hi SirFred Jun 26 02:02:55 fine. get back later. i will give gpe-image and see Jun 26 02:02:57 bye Jun 26 02:03:00 take care Jun 26 02:03:49 LEAVE Jun 26 02:04:42 Ciao all Jun 26 02:06:20 hi Pigi Jun 26 02:07:03 hi RP Jun 26 02:08:55 I need to switch-user for an bb install-phase. How can this be done in oe ? Jun 26 02:09:48 Anybody knows how could I make gdb tell me in which object a symbol is defined? Jun 26 02:10:01 Pigi: I thought bb was supposed to run as one user only? Jun 26 02:10:13 Well, I've found it using a dirty solution. Just a breakpoint. Jun 26 02:10:36 yes, RP, but in install phase it could be possible to su to root ( for example ) Jun 26 02:11:08 and I'm sure that there should be something.... as every kernel module or the kernel itself are installed as root. Jun 26 02:11:37 Pigi: the conversion from $USER to root is done at image creation time using fakeroot IIRC Jun 26 02:11:56 that's what I supposed. :) Jun 26 02:12:04 What an strange thing in qte. Jun 26 02:12:07 I've always wondered how that was handled :) Jun 26 02:14:06 There's a method called QLinuxFbScreen::uncache() to free a chunk of offscreen memory. That method calls in sequence: sync() and deleteEntry(). The comment says something like "This function will first sync the graphics card to ensure the memory isn't still being used by a command in the graphics card fifo queue". But there is no sync() method in QLinuxFbScreen neither in its superclasses. Jun 26 02:14:50 So, the sync() that's called is the one from to commit buffers to disk. Jun 26 02:15:03 I think it's a bug. Jun 26 02:15:04 Ouch Jun 26 02:15:13 Yes, that sounds like a nasty bug Jun 26 02:15:16 A very bad bug, because the machine is syncing everytime. Jun 26 02:15:21 That will destroy flash :-( Jun 26 02:15:32 RP: I have put a breakpoint and: Jun 26 02:15:44 #0 0x40639e70 in sync () from /lib/libc.so.6 Jun 26 02:15:45 #1 0x40351ed8 in QLinuxFbScreen::uncache () from /mnt/card/w100/libqte.so.2 Jun 26 02:15:45 #2 0x40357948 in QW100Screen::uncache () from /mnt/card/w100/libqte.so.2 Jun 26 02:16:00 RP: uncache is called thousand of times all the time. Jun 26 02:16:23 SirFred: Does it run a bit faster if you remove the sync? Jun 26 02:16:30 RP: I've just found it. Jun 26 02:16:45 RP: I found it because sometimes I had screen corruption on rotations. Jun 26 02:16:57 RP: And some SIGSEGV in that uncache. Jun 26 02:17:27 RP: Probably, some time ago, there should be a sync() method in QScreen. Jun 26 02:17:48 Because the comment of the method also says: Jun 26 02:17:50 That sounds likely Jun 26 02:17:57 \sa deleteEntry() sync() Jun 26 02:18:22 RP: I don't know if current openzaurus is using offscreen memory to cache pixmaps. Jun 26 02:18:31 RP: If that's used, it's dangerous. Jun 26 02:18:51 SirFred: I guess we need a mickeyl to comment on that :) Jun 26 02:18:54 RP: Yes. Jun 26 02:19:08 * SirFred invokes mickeyl Jun 26 02:19:29 Nothing. Jun 26 02:19:46 Wild party yesterday, perhaps. Jun 26 02:19:54 Hopefully he'll scroll back when he wakes up :) Jun 26 02:20:19 Well, I'm going to remove that sync and make some tests. I expect it to fly now. Jun 26 02:20:20 He's been idle for 17 hours Jun 26 02:20:26 Fingers crossed :) Jun 26 02:25:05 RP: Are you working in kdrive? Jun 26 02:25:32 SirFred: Kind of Jun 26 02:25:45 RP: I was thinking that perhaps it is the way to go. Jun 26 02:26:07 SirFred: Once I finish with atilib, mallum will probably write an accel driver for it Jun 26 02:26:09 RP: As time passes I dislike more this qte qws stuff. Jun 26 02:26:20 I've never looked at it much... Jun 26 02:26:33 RP: And having in mind that my first motivation was have an accelerated sdl implementation... Jun 26 02:26:51 If kdrive is light enough could be the solution. Jun 26 02:27:52 Everything does seem to be heading in the xserver direction... Jun 26 02:27:59 Always supposing that sdl over x is using correctly the x apis. Jun 26 02:28:14 RP: Because sdl-qte uses all the time direct framebuffer access. Jun 26 02:28:29 RP: So, having an accelerated qte wont give us an accelerated sdl-qte. Jun 26 02:28:59 SirFred: Does it use the fb accell calls? Jun 26 02:29:13 RP: No. It uses the QDirectPainter qte classes. Jun 26 02:29:25 RP: And manages the rotations itself. It's funny. Jun 26 02:29:39 I keep wondering if its worth adding accell to the w100fb driver... Jun 26 02:29:56 RP: You can have a situation where QTE is rotating the framebuffer and SDL is rotating again it to put it in the original orientation. Jun 26 02:30:15 RP: Well, I've never look at the FB accel apis. Jun 26 02:30:20 That sounds plain broken Jun 26 02:30:29 RP: It a big waste of resources. Jun 26 02:30:33 I have but only very loosly... Jun 26 02:31:20 RP: But the sdl-qte driver is a hack in my opinion. It just setup an qteapp, creates a QImage filling all the screen and then direct access to that Image. Jun 26 02:31:32 That's all, more or less. Everything else is made in software. Jun 26 02:31:52 I expect the X version to be more polished. Jun 26 02:32:03 That sounds likely Jun 26 02:34:58 hi chaps Jun 26 02:35:25 hi pb_ Jun 26 02:39:43 good morning pb_ Jun 26 02:42:40 Pigi: did you make any progress with the alsa drivers? Jun 26 02:43:30 it seems, yes. Jun 26 02:43:39 I do managed to compile the drivers. Jun 26 02:43:51 Now I'm a bit stucked on the package and install phase. Jun 26 02:44:02 Do you have some minutes to share some ideas ? Jun 26 02:45:39 sure Jun 26 02:45:56 got an error : please see http://pastebin.com/304160 for details Jun 26 02:46:01 well, right now I'm going to eat breakfast, but in 15 minutes or so I will be at your disposal Jun 26 02:46:18 good breakfast then Jun 26 02:46:19 got an error ; defconfig not found. Jun 26 02:47:36 can anyone tell me what is the problem and how do i solve it ? Jun 26 02:48:02 hello Jun 26 02:49:31 hmm wasn't "defconfig" the machine specific kernel .config? Jun 26 02:49:44 omap processor Jun 26 02:50:10 what's your MACHINE set to? Jun 26 02:50:22 *yawn* Jun 26 02:50:24 morning Jun 26 02:50:27 hi mickeyl Jun 26 02:50:28 NAiL: pong Jun 26 02:50:43 morning mickeyl Jun 26 02:51:08 omap5912osk Jun 26 02:51:16 MACHINE=omap5912osk Jun 26 02:52:37 and your kernel is "./linux-omap-2.6-2.6.9-omap1" or "./linux-omap1-2.6.12-rc2"? Because there's only a .config for these two for your machine Jun 26 02:53:11 hey Jun 26 02:53:17 hi zecke Jun 26 02:53:50 how to find out ? Jun 26 02:53:59 mickeyl: would it be ok for you to have checkversions only check for YYYYMM, not YYYYMMMDD for snapshot kernels? Checking for DD is a PITA :\ Jun 26 02:54:40 sriram: from your pastebin: linux-omap1-2.6.12-rc2 Jun 26 02:54:49 CoreDump|home: i could do that. i rather disable checkversion completely for DEBUG builds though Jun 26 02:55:02 DISTRO_VERSION != "release" that is Jun 26 02:55:07 yeah Jun 26 02:55:09 would that work for you as well? Jun 26 02:55:14 will do it later today Jun 26 02:55:23 ok Jun 26 02:55:39 so what do i do now ? Jun 26 02:56:06 sriram: good question Jun 26 02:56:54 SirFred: now that's an interesting finding. good work! Jun 26 02:56:59 ~lart TT Jun 26 02:56:59 * ibot pours hot grits down the front of TT's pants Jun 26 02:57:33 seriously i dont know what to do and am waiting for a response. Jun 26 02:58:23 mickeyl: Morning. Jun 26 02:58:46 mickeyl: What do you think it's a better solution? Jun 26 02:58:57 sriram: from the looks of it your "MACHINE" is either wrong, or contains a typo. OE should pick up /linux-omap1-2.6.12-rc2/omap5912osk/defconfig using overrides. Jun 26 02:59:12 mickeyl: I think that putting a virtual sync() in QScreen is the cleaner one. But there is that BC issue. Jun 26 02:59:36 mickeyl: The other one is just comment out that sync() call from QLinuxFbScreen::uncache. Jun 26 02:59:48 SirFred: sorry I missed 80% of the talk but what 'sync()' do you need? Jun 26 03:00:04 SirFred: could you abuse beginDisplayReconfigure (probably beginReconfigure...) Jun 26 03:00:14 anyways, bbl Jun 26 03:00:32 zecke: SirFred found out that qte call libc's sync instead of what they really wanted to do is qscreen::sync() Jun 26 03:00:33 zecke: One for QScreen, to assure that all graphic operations are finished. Jun 26 03:00:34 which is quite funny Jun 26 03:01:14 SirFred: given that many other accellerated drivers implement sync, i'm actually for adding that method. If it's not needed, just leave it empty. Jun 26 03:01:19 zecke: The thing is that one QScreen::sync() seems to have existed, at least in some developer mind. Jun 26 03:01:33 mickeyl: Well, they don't implement sync. Jun 26 03:01:37 i am pasting my local.conf please see http://pastebin.com/304163 Jun 26 03:01:39 mickeyl: They implement sync but for the Gfx. Jun 26 03:01:44 mickeyl: Not for the QScreen. Jun 26 03:01:52 SirFred: oh right, that's different entities Jun 26 03:01:58 Pigi: right Jun 26 03:01:59 ~lart TT again Jun 26 03:01:59 * ibot chops TT again in half with a free AOL CD Jun 26 03:02:03 mickeyl: good morning Jun 26 03:02:06 :) Jun 26 03:02:09 SirFred: break BC in Qt/E Jun 26 03:02:11 good morning pb_ Jun 26 03:02:21 zecke: good morning Jun 26 03:02:29 pb_: hell good morning here ;) Jun 26 03:02:33 well, afternoon for you I suppose Jun 26 03:02:59 OK. So, I'm going to put a virtual void sync() {} in the QScreen class. Jun 26 03:03:03 yep Jun 26 03:03:10 SirFred: or look at QStyle Jun 26 03:03:11 And then, we can reimplement it in the accellerated drivers. Jun 26 03:03:18 i wonder how we could have overlooked that Jun 26 03:03:21 it's nasty Jun 26 03:03:24 SirFred: they forgot a method as well and created a global function pointer... Jun 26 03:03:26 12 in the morning actually :) i always forget which 12:00 is AM and which is PM Jun 26 03:03:36 zecke: Humm. Jun 26 03:03:50 SirFred: a) do it properly and notify TT Jun 26 03:03:57 SirFred: b) look if we can do it in a bc way Jun 26 03:04:08 zecke: Are they still maintaining that version? Jun 26 03:04:28 CoreDump|afk: i have pasted my local.conf file Jun 26 03:04:32 SirFred: some how yes, but I think main focus is getting Qt4 usable Jun 26 03:05:21 so pb_ To be able to do something with those nasty alsa driver I took the following direction: Jun 26 03:05:21 SirFred: you know how to profile things? I'd be interested to know how often that sync() is getting called Jun 26 03:05:21 sorry gotta really go now. later Jun 26 03:05:40 mickeyl: Well, it's called a lot of times, at least using offscreen memory. Jun 26 03:05:47 SirFred: hell Jun 26 03:05:51 bbl, breakfast time. Jun 26 03:05:53 mickeyl: I don't know if current openzaurus is using offscreen memory for anything. Jun 26 03:06:04 I did what jamey advice in README.familiar, so I override the configure phase with the config.ipaq script Jun 26 03:06:17 coredump: Jun 26 03:06:18 and now I did almost the same to create the ipk. Jun 26 03:06:20 are you there ? Jun 26 03:06:28 I think I've read that it's used to cache pixmaps. But I'm not sure. With the accellerated driver is used a lot, of course. Jun 26 03:06:34 can you tell me what i should do to resolve my problem ? Jun 26 03:06:53 jamey do prepared a Makefile that create the ipk's Jun 26 03:07:10 so in do_install() i invoked it. Jun 26 03:07:32 now I have an alsa-whatsoever-xx.ipk in a dir, and don't know : Jun 26 03:07:40 ah, right Jun 26 03:07:44 1) if is it right what I did until now Jun 26 03:07:53 oe likes to generate the ipks for itself, rather than have you do that by hand Jun 26 03:07:55 2) how to move this ipk to deploy area Jun 26 03:08:27 can you paste a bit of the makefile here, so I can see what it looks like? Jun 26 03:08:41 I thought oe likes to generate the ipks for itself, but I'm not able to say him that Jun 26 03:08:43 sure Jun 26 03:08:57 You probably need to replace that with a do_install() method that just copies all the files into ${D}, and then appropriate settings of some other variables to have the packages generated Jun 26 03:09:35 please find the problem at http://pastebin.com/304160 and local.conf file at http://pastebin.com/304163 Jun 26 03:09:53 pb_ that's the "familiar/Makefile" : http://pastebin.com/304164 Jun 26 03:10:44 okay Jun 26 03:10:47 pigi can you have a look at my build issue Jun 26 03:11:20 so, if I'm reading that right, it installs all the modules, deletes the stuff in ${extra_modules}, and packages the rest. Jun 26 03:11:35 good afternoon all Jun 26 03:11:41 pb_ yes Jun 26 03:12:08 koen: some build problems http://pastebin.com/304163 and local.conf file at http://pastebin.com/304163 Jun 26 03:12:09 sriram, I'm not very good at build. Just studing now... Jun 26 03:12:13 zecke: Do you think posting the problem to qt-embedded-interest would be appropiate? Jun 26 03:13:00 in that case, you want something like: do_install() { make -k NODEPMOD=yes DESTDIR=${D} install; for i in ${extra_modules}; do rm -f ${D}/lib/modules/*/misc/$$i; done } Jun 26 03:13:21 sriram: could you remove all the spaces around '=' operators in local.conf? Jun 26 03:13:24 SirFred: appropriate yes, but if it results in anythin is doubtful Jun 26 03:13:32 zecke: :) Jun 26 03:13:35 zecke: What do you suggest? Jun 26 03:13:59 do you think that is the problem. it is not finding the correct patch Jun 26 03:14:34 sriram: I've had problems myself with things like that Jun 26 03:14:43 if i remove it. will bitbake start all over again ?. till now it has taken me 10 hours flat Jun 26 03:15:15 it shouldn't start over again Jun 26 03:17:34 so shall i give bitbake or do you think there is some other problem Jun 26 03:18:41 is machine anem correct ? Jun 26 03:19:03 sorry machine name is it correct or not ?. getting doubts ? Jun 26 03:19:10 pb_ oe create the ipk's in the do_install phase usually ? Jun 26 03:19:12 SirFred: use the list you mentioned, and qt-bugs@trolltech.com Jun 26 03:19:24 sriram: MACHINE=omap5192osk should work Jun 26 03:19:25 zecke: Ok. Jun 26 03:19:42 Pigi: it creates them in do_package_ipk, which runs just after do_install. Jun 26 03:19:45 use parenthesis Jun 26 03:19:49 but it is omap5912 and not omap5192 ? Jun 26 03:19:54 ok. Jun 26 03:20:17 sriram: ehm, oops, yes omap5912osk Jun 26 03:22:28 SirFred: void foo(void) == void foo() in C++ Jun 26 03:22:29 sriram: I'm testing a patch Jun 26 03:22:42 guys, should go to eat to my parents-in-law :( Jun 26 03:22:45 see you later Jun 26 03:22:57 cu Jun 26 03:23:29 zecke: Why do you say that? Jun 26 03:23:58 SirFred: I some how remembered you pasted a method with (void) as it's parameters Jun 26 03:24:14 SirFred: it might be ~20 hours too late and completely out of context Jun 26 03:24:27 zecke: Yes, I always do it that way, but when coding for qte I try to follow its style. Jun 26 03:24:39 koen: what should i do now ?. Jun 26 03:24:53 zecke: I always liked to put the (void) as args. That gives me some trouble in java. Jun 26 03:24:55 koen: we now have the org.openembedded.nslu2-linux branch created. But please do not pull from our server into your server. We will always do the push for openembedded transfers. (if you pull, you will get all the nslu2-linux non-openembedded stuff ending up in your repo, and you don't want that). We are serving org.* (both org.openembedded.nslu2-linux and org,nslu2-linux.*). Jun 26 03:26:11 rwhitby-away: ok Jun 26 03:29:28 sriram: change linux-omap1_2.6.12_rc2.bb to look like this: http://pastebin.ca/15882 Jun 26 03:29:31 03koen 07 * r1.3622.1.1 10openembedded/packages/linux/linux-omap1_2.6.12_rc2.bb: Jun 26 03:29:31 linux-omap1_2.6.12_rc2.bb: Jun 26 03:29:31 revert my misguided change to src_uri Jun 26 03:29:56 sriram: so remove the SRC_URI_append and change the SRC_URI= Jun 26 03:30:12 koen: once we have a tested build, we will propagate across to org.openembedded.dev and push back to your server. Jun 26 03:30:31 sriram: see the patch at http://oe-devel.bkbits.net:8080/openembedded/diffs/packages/linux/linux-omap1_2.6.12_rc2.bb@1.4?nav=index.html|ChangeSet@-7d|cset@1.3622.1.1 Jun 26 03:30:46 1 sec Jun 26 03:32:00 koen: in which dir ? Jun 26 03:32:08 packages/linux/ Jun 26 03:34:18 just look at all this beautiful branching: http://ewi546.ewi.utwente.nl/tmp/viewmtn/revision.psp?id=259042bb5d1b8bf09aad8d974f9bb40744c90066&ancestry_limit=0 Jun 26 03:37:30 a marriage of two families :-) Jun 26 03:37:37 :) Jun 26 03:37:40 ok i appended file=defconfig" shall i run bitbake ? Jun 26 03:37:48 now Jun 26 03:38:01 file://defconfig Jun 26 03:38:23 yes yes Jun 26 03:38:36 it should be good now Jun 26 03:38:38 shall i run bitbake now or do i need to make some more changes ? Jun 26 03:38:43 as in, the kernel builds over here Jun 26 03:39:49 koen: do i need to make anymore changes ? Jun 26 03:41:17 not that I know off Jun 26 03:42:33 koen: one more thing. i have give bitbake bootstrap-image and if i want to make it gpe-image will it start all over again ?. Jun 26 03:42:46 i am afraid becuase it takes lot of time. Jun 26 03:42:49 no, it will use the stuff it build before Jun 26 03:43:30 if i want to copy to some other machine. should i copy TMPDIR also ? Jun 26 03:43:58 if you don't want to start again, yes Jun 26 03:44:17 i dont want to start all over again. Jun 26 03:45:15 koen i got build error after the chages: Jun 26 03:45:54 http://pastebin.ca/15886 Jun 26 03:46:17 ah yes Jun 26 03:46:55 'bitbake -c clean -b packages/linux/linux-omap1_2.6.12_rc2.bb' Jun 26 03:51:02 sriram: you might want to add 'include conf/distro/preferred-gpe-versions.inc' to conf/distro/generic. Jun 26 03:51:05 conf Jun 26 03:52:23 i have started bitbake. should i add include conf/distro now ?. Jun 26 03:52:30 will it make a difference > Jun 26 03:55:10 it shouldn't make a difference, but it's to make sure there aren't any surprises with X Jun 26 03:57:49 should i add PREFERRED_PROVIDERS and PREFERRED_X11 as in zuarus.conf Jun 26 03:59:10 got an warning: see http://pastebin.ca/15887 Jun 26 03:59:16 can i ignore this warning. Jun 26 03:59:56 if it continues, it should be ok Jun 26 04:00:57 and adding the X11 stuff from another conf/machine/*.conf would help Jun 26 04:01:28 you'll probably need the ones with diet-x11 and xserver-fbdev-dkrive Jun 26 04:01:33 kdrive* Jun 26 04:19:57 i got NOTE: package bootstrap-image-1.0-r0: task do_build: completed Jun 26 04:19:57 NOTE: package bootstrap-image-1.0: completed Jun 26 04:19:57 NOTE: build 200506261621: completed Jun 26 04:19:57 Build statistics: Jun 26 04:20:07 is the build complete. Jun 26 04:20:14 where can i find it ? Jun 26 04:21:47 tmp/images Jun 26 04:22:06 tmp/deploy/images Jun 26 04:22:14 fine Jun 26 04:22:16 3.0M bootstrap-image-omap5912osk-20050626104045.rootfs.ext2.gz Jun 26 04:22:27 does it contain all the libs ? Jun 26 04:22:42 is it a jffs2 image or just a file system . Jun 26 04:23:00 file system in .gz format Jun 26 04:24:00 it appears to be an ext2 loop-file Jun 26 04:24:49 what is that ? Jun 26 04:25:41 mount -o loop ./bootstrap-blahh /mnt/somewhere Jun 26 04:25:46 it can't be flashed Jun 26 04:26:11 i want 1) a single kernel file and file system in .tar format Jun 26 04:26:23 2) a jffs file system Jun 26 04:26:28 what should i do ? Jun 26 04:26:29 never used the tar FSTYPE Jun 26 04:27:15 ok. i want a kernel image and a directory which can be mounted as root so that i can mount it via NFS Jun 26 04:27:22 how do i do that ? Jun 26 04:29:00 IMAGE_FSTYPE = "tar" in local.conf Jun 26 04:29:28 well, you could use the ext2 file, too. mount it somewhere and export it via NFS Jun 26 04:30:37 if i change it right now in local.conf file. will it start all over again. Jun 26 04:31:12 I doubt you'd have to start over Jun 26 04:31:22 most of the files will be ccached, most packages built Jun 26 04:31:46 just the image build would probably have to go again Jun 26 04:32:40 fine Jun 26 04:32:51 sriram: you can stag my local.conf + build scripts here: http://www.northern.ca/projects/openembedded/oe.tar.gz Jun 26 04:33:10 it's set to build OZ-latest Jun 26 04:33:29 i did not get you ? Jun 26 04:33:59 ok. what does the file system contain ? Jun 26 04:34:19 only busybox ? Jun 26 04:34:57 suppose i want to make some changes to the kernel configurations and busybox conf files what should i do ? Jun 26 04:35:11 file system contains the same thing as the ext2, but it's will be in .jffs2 Jun 26 04:35:34 format Jun 26 04:36:07 i am talking of the utilities /bin and /sbin /usr/sbin what do they contain ? Jun 26 04:37:08 I'd think... (I'm new at this), you do a patch against the source code and then add it to the patch data of the .bb file Jun 26 04:37:12 for busybox Jun 26 04:37:13 sriram: well, a complete bootstrap image Jun 26 04:37:15 or kernel Jun 26 04:37:37 then build the package again Jun 26 04:37:58 then build the image again Jun 26 04:38:49 do bitbake -f -c compile -b Jun 26 04:39:08 if i want to modify kernel settings (say i want only usb gadget support - ethernet gadget) and say i want to turn on/off a specific busybox utility what should i do ? Jun 26 04:40:28 sriram: edit defconfig (heh) Jun 26 04:41:25 i changed the FSTYPE to tar. but still i am getting the XXX.ext2.gz Jun 26 04:41:26 CoreDump: is defconf same as the .bb in /stuff/openembedded/packages/linux? Jun 26 04:41:33 for the kernel that is. Busybox has its own config packages/busybox Jun 26 04:42:13 i have to make modifications to defconfig and do a clean on that package and give bitbake bootstarp-image thats all Jun 26 04:42:22 am i right ? Jun 26 04:42:36 jumpkick: it's not a .bb, just a file containing a kernel .config. It's in linux/$your-kernel/$your-machine Jun 26 04:42:46 yes Jun 26 04:42:51 fine Jun 26 04:43:07 ah... I see Jun 26 04:43:56 is there a way to clean up old stuff? Jun 26 04:44:10 sir, if i want a kernel image say a uImage and a directory (which contains busybox and others) what should i do ? and that directory should be able to mount as a root file system ? Jun 26 04:44:15 what should i do ? Jun 26 04:44:21 bitbake -c clean -b path/to/.bb Jun 26 04:45:18 Sirfred|beer: you already have a kernel. It was build during bootstrap-image. The ext2 image can be mounted as a loop device and will do the trick Jun 26 04:46:12 the advantage of a ext2 file is that all permissions inside the bootstrap image are set correctly Jun 26 04:47:00 s/Sirfred|beer/sriram Jun 26 04:47:44 can i the jffs image onto a flash ? Jun 26 04:48:28 yep, jffs images can be flashed. But I dunno your machine/device so I'd ask someone else if flashing your device is supported Jun 26 04:49:12 i have a omap 5912 eval board. the board has an inbuilt uboot 1.1 Jun 26 04:49:26 so i want to flash the jffs2 file system and kernel ? Jun 26 04:55:45 CoreDump: bitbake -c clean doesn't seem to actually free disk space for me Jun 26 04:55:59 is it just deleting timestamps? Jun 26 04:56:42 jumpkick: oh, diskspace. No it just cleans up all timestamps and sources for a .bb to allow a clean rebuild Jun 26 04:57:01 jumpkick: you can safely delete tmp/ipk/morgue Jun 26 04:57:07 but that's about it Jun 26 04:57:15 ok... well that's a start Jun 26 04:57:19 thanks Jun 26 04:57:31 np Jun 26 04:58:48 you mean tmp/deploy/ipk? Jun 26 04:58:54 that won't do much Jun 26 04:59:17 oh Jun 26 04:59:26 there actually a dir morgue Jun 26 04:59:28 got ya Jun 26 05:00:08 still not much disk to be had there though Jun 26 05:00:22 400k Jun 26 05:02:08 the best way to "clean up" is to trash tmp and sources and build from scratch :\ Jun 26 05:03:15 ugh Jun 26 05:04:08 I think I'll parse out stuff in the deploy/ipk/morgue and then delete from /tmp/work/old-package Jun 26 05:04:28 I'll cook up a script to do it Jun 26 05:05:24 03CoreDump 07 * r1.3623.1.1 10openembedded/packages/initscripts/ (initscripts_1.0.bb initscripts-1.0/openzaurus/checkversion): checkversion: Do not check kernel version for snapshots Jun 26 05:10:16 i got the ext2 file system and mounted it via -o loop option. Jun 26 05:10:23 but where is the kernel image Jun 26 05:12:53 sriram: that depends on your image Jun 26 05:13:02 either it is in /boot/ Jun 26 05:13:23 it is zImage. Jun 26 05:13:40 can i convert it into a uImage and download using uboot Jun 26 05:13:55 and can i mount the ext2 file system via NFS Jun 26 05:14:00 can i do that ? Jun 26 05:15:09 usually whet we do is get the kernel source compile it and give it mount via NFS arguments in defconfig and get an uImage and build a directory which contains all the bustbox and which can be mounted as a root file system. Jun 26 05:15:14 thats what we do here. Jun 26 05:15:25 i dont how it has to be done with OE Jun 26 05:15:49 and i am not a usually kernel guy. just a stupid c programmer Jun 26 05:17:24 Koen: please throw some light on this it will be very useful for idiots like me who have very little exposure of file systems,mount Jun 26 05:19:57 your starting to get pretty off topic Jun 26 05:20:08 OE deals with how to build software, not how to install it to your device Jun 26 05:20:15 you better ask that on a dedicated channel Jun 26 05:20:51 our omap guys are all asleep anyway Jun 26 05:28:33 fine mickey Jun 26 05:32:34 Got the Board to Bootup i think Jun 26 05:34:19 but the file system size is 8 MB which is too high for me. i am looking at 4 MB or less than that. Jun 26 05:34:53 in that case you want to remove things from task-bootstrap.bb Jun 26 05:35:03 or make your own task-minimal-bootstrap Jun 26 05:35:18 mickeyl: do you mind if I remove your Buttons SIMpad patch? Jun 26 05:35:37 from qte ? Jun 26 05:35:43 is there any step by step instructions ... Jun 26 05:36:00 sriram: unfortunately not. only the Getting Started, the bitbake manual, and the FOSDEM2005 pdf. Jun 26 05:36:21 zecke: go ahead, but wait a minute, i'm committing something to qte-2.3.10 atm. Jun 26 05:37:10 only removing them in bootstrap.bb is enough Jun 26 05:37:25 is it some thing equivalent to main make file ? Jun 26 05:37:27 zecke: you going the input layer route for 2.6 right? Jun 26 05:37:28 03mickeyl 07 * r1.3624.1.1 10openembedded/packages/qte/ (qte_2.3.10.bb qte-2.3.10/fix-linuxfb-setmode.patch): qte-2.3.10: add and apply a patch that moves setting the linuxfb screen mode to a more sane location. Patch courtesy Manuel Teira Jun 26 05:37:41 sriram: no. bitbake works pretty different from Make. Jun 26 05:38:01 sriram: i recommend reading the three articles I mentioned before continuing. Jun 26 05:38:21 so if i want say only busybox in my file system, and only one run level. what files need i to change Jun 26 05:38:46 ok fine mickey. will read it and get back to you on this matter. Jun 26 05:38:52 sriram: you need to find out which packages contain the things you want to tweak Jun 26 05:39:10 bootstrap-image = task-bootstrap Jun 26 05:39:25 runlevel = base-files or initscripts Jun 26 05:40:03 Sirfred|beer: all your patches are in. The qte patch made it to OE, the others were applied upstream, so make sure you set the CVSDATE to tomorrow. Jun 26 05:42:57 Sirfred|beer: small note for further patches: Please comply with the formatting already present in the files, i.e. tabs vs. spaces and so on Jun 26 05:44:10 mickeyl: Sorry for the space/tab mistake. Jun 26 05:44:34 mickeyl: I didn't setup my emacs to use spaces. Never remember to do so. Jun 26 05:48:03 03mickeyl 07 * r1.3623.2.1 10openembedded/conf/distro/openzaurus-3.5.4.conf: oz-3.5.4: bump CVSDATE Jun 26 06:08:54 03mickeyl 07 * r1.3628 10openembedded/ (5 files in 2 dirs): add gpsd 2.26 and kill off a couple of older ones Jun 26 06:58:48 koen: just propagated org.openembedded.nslu2-linux to org.openembedded.dev, but for some reason it created a new head. Any idea why? And can you merge them please? Jun 26 06:59:33 wow Jun 26 06:59:44 * koen looks at a 3 way split merge window Jun 26 07:00:05 not automatic? Jun 26 07:00:54 nope Jun 26 07:02:09 koen: we did a pull from ewi, checked there was only one head, then did a propagate, checked again there was only one head, then did a push back to ewi. still only one head that we can see Jun 26 07:02:33 monotone: misuse: merge of 'packages/meta/openslug-packages.bb' : '6c99609c7a05d4ec58b2086c263a396da9af067e' -> '53d30a56bd0b86b02cb60ab9e3e4c207d927c0b6' vs '8e327852d945fd291a7ac9e3aa70c34ef6faadba' failed Jun 26 07:02:42 probably the dropped key problem Jun 26 07:03:10 hmmm Jun 26 07:03:14 ok, just take the latest one of those. Jun 26 07:03:23 how the hell do I do that? Jun 26 07:03:44 or just do a propagate from org.openembedded.nslu2-linux to org.openembedded.dev there Jun 26 07:04:03 koen@bitbake:/data/build/oe$ monotone --db=OE-koen.db propagate org.openembedded.nslu2-linux org.openembedded.dev Jun 26 07:04:07 monotone: misuse: branch 'org.openembedded.dev' is not merged Jun 26 07:04:14 Funny... I've started memorizing sha-1 hashes Jun 26 07:05:02 koen: I guess now is the time where we both look up the manual on how to resolve merge conflicts ... :-) Jun 26 07:06:17 "If there had been a conflict, monotone would have invoked an external merging tool to help resolve it." Jun 26 07:06:58 koen: can you put the server back on line - I want to see if I can get the two heads here and do the merge. Jun 26 07:07:17 ok, done Jun 26 07:07:54 nope, it successfully syncs, but says it's merged Jun 26 07:08:02 monotone: branch 'org.openembedded.dev' is currently merged: Jun 26 07:08:02 891a0add04844a80f9eef545fe76ab9b3684db4f nslu2-linux@nslu2-linux.org 2005-06-26T13:52:37 Jun 26 07:09:04 I suspect it is because of the cross-collection merges we did previously. Jun 26 07:09:53 * koen puts a slug on the "things to buy after vacation" Jun 26 07:11:35 koen: try this: monotone disapprove of the following three revisions: Jun 26 07:12:00 * NAiL wait for three of his hashes to appear :P Jun 26 07:13:05 051fc6f9e28e16d274e2cd95db273eb514d8fb33 Jun 26 07:13:14 0e46f5262b548693d0460cf2c1b103cc455e4aae Jun 26 07:13:25 e939e179c1552200dcb9eb26b181ac051f11e4be Jun 26 07:14:20 that's the three blue ones on this: http://ewi546.ewi.utwente.nl/tmp/viewmtn/revision.psp?id=e939e179c1552200dcb9eb26b181ac051f11e4be Jun 26 07:15:14 * Titeuf is away: Je suis occup ! Jun 26 07:15:31 * Titeuf is back (gone 00:00:03) Jun 26 07:15:37 koen: the other way to solve this is to pull from ewi to another database, and then replace ewi's database with that new one. that will remove all the errant org.nslu2-linux.* stuff Jun 26 07:22:14 monotone: misuse: revision e939e179c1552200dcb9eb26b181ac051f11e4be has 2 changesets, cannot invert Jun 26 07:23:04 rwhitby: we can't remove branches in monotone 0.19 Jun 26 07:23:08 koen: hmm - will have to try the second approach to filter out the org.nslu2-linux.* cruft, as I only get one head here. Jun 26 07:23:45 koen: should be able to remove the org.nslu2-linux.* branches by pulling from a server that does not serve them up (like ewi) Jun 26 07:23:48 the monotone people have committed a patch for it Jun 26 07:57:30 becuase of space constraints i want to delete TMPDIR. will all the tars, cvs, all the downloads will be done again or will it take from the sources directory. Jun 26 07:58:18 i dont find any cvs directories there only tar files. Jun 26 07:58:24 morning Jun 26 07:58:26 that is in sources Jun 26 07:58:39 sriram: rm -rf tmp/work/ Jun 26 07:59:08 and stamps Jun 26 07:59:40 sriram: keep DL_DIR outside of TMPDIr Jun 26 07:59:43 ok. what directories i should take backup if i down want future downnloads ? Jun 26 07:59:47 hi folks Jun 26 07:59:50 sriram: DL_DIR Jun 26 08:00:00 hey chris144|home Jun 26 08:00:05 what about TMPDIR Jun 26 08:00:06 hi koen Jun 26 08:00:06 koen: could you use one-commit for one-change? Jun 26 08:00:38 hi hrw Jun 26 08:00:51 hi RP Jun 26 08:01:17 can i delete it. what happens ? Jun 26 08:01:28 if i delete TMPDIR what happends Jun 26 08:01:31 hrw: I usually try to do that Jun 26 08:01:38 sriram: tmpdir is for all build stuff - so you will need to rebuild whole Jun 26 08:02:06 building will use the sources in source directory right and it wont download Jun 26 08:02:25 download takes time and i have a slow link. Jun 26 08:02:26 sriram: yes Jun 26 08:02:55 so backup means taking sources, openembbeded Jun 26 08:03:23 sriram: keep openembedded, sources outside of builddir (TMPDIR) and it will work Jun 26 08:03:35 where do cvs downloads go ... Jun 26 08:03:45 i have kept it like that only. Jun 26 08:03:49 into CVSDIR iirc which is DL_DIR/cvs/ Jun 26 08:03:54 folks, I just did a fresh build and binutils-cross complaind about "target system type". Conf looks ok to me. See http://pastebin.ca/15905 Jun 26 08:05:08 chris144|home: do you have a GA-586ATE.conf? Jun 26 08:05:24 koen: yes Jun 26 08:05:39 and does that set TARGET and such? Jun 26 08:05:39 * RP can't move today. Too painful :-/ Jun 26 08:05:46 koen: yes Jun 26 08:05:51 RP: what did you do? Jun 26 08:06:14 chris144|home: "INVALID-linux" is OE's default, so it seem,s it doesn't get set Jun 26 08:06:31 koen: ok... Jun 26 08:06:36 koen: Something a little insane, even by my standards... Jun 26 08:06:39 koen: http://www.rpsys.net/mx1.jpg Jun 26 08:07:15 looks cool Jun 26 08:07:17 RP: nice Jun 26 08:07:20 koen: ok, let me check that the machine symlink is REALLY there... Jun 26 08:08:39 hrw: Was good fun but I'm paying for it now. Only fell off once... Jun 26 08:12:19 good morning kergoth Jun 26 08:12:26 ood morning chouimat Jun 26 08:12:28 +g Jun 26 08:12:29 hey Jun 26 08:12:33 hi kergoth Jun 26 08:12:45 www.dilbert.com and http://www.userfriendly.org/ Jun 26 08:13:56 got this error while building gpe-image please see http://pastebin.com/304252 Jun 26 08:14:35 koen: it is giving segmentation fault by compiler ? Jun 26 08:14:48 what should i do ? Jun 26 08:15:31 sriram: see conf/machine/h3900.conf and copy the bits with 'xserver' to conf/machine/omap5912osk.conf Jun 26 08:17:57 did not get you ?. Jun 26 08:18:12 there is only one line with xserver which is Jun 26 08:18:13 PREFERRED_PROVIDER_xserver = "xserver-kdrive" Jun 26 08:18:16 re Jun 26 08:18:28 this line is already in omap5912.conf Jun 26 08:19:26 pb_ ping Jun 26 08:19:37 * koen wonders why it's trying to build xorg instead Jun 26 08:20:32 do you want to see my h3900 file ? Jun 26 08:20:41 koen ? Jun 26 08:21:39 no, your omap5912osk.conf file Jun 26 08:21:46 ok Jun 26 08:22:03 Pigi: hello Jun 26 08:22:30 hi pb_. The first phase of install is ok now, but oe don't do the ipk .... Jun 26 08:22:32 please see http://pastebin.com/304256 Jun 26 08:22:37 I suspect I need another bit of help Jun 26 08:23:08 Pigi: okay. what does bitbake say? Jun 26 08:24:14 Let me see if I'm doing something wrong. Jun 26 08:24:42 If I do bitbake -c install alsawhatsoever, I'm execting he does also the ipk or it only install ? Jun 26 08:24:55 only instal Jun 26 08:24:55 l Jun 26 08:24:57 sriram: try adding the line PREFERRED_PROVIDER_virtual/xserver = "xserver-kdrive" Jun 26 08:25:03 good. Jun 26 08:25:19 what happens if you use "-c package_ipk"? Jun 26 08:25:33 so maybe it's my brain trouble , and not oe trouble Jun 26 08:26:11 ugh this axim is driving me nuts.. and the pocketpc community overall is less than helpful Jun 26 08:26:26 pb_ Jun 26 08:26:27 NOTE: package alsa-driver-0.9.6-hh4c: started Jun 26 08:26:27 ERROR: InvalidTask event exception, aborting Jun 26 08:26:27 NOTE: package alsa-driver-0.9.6-hh4c: failed Jun 26 08:26:27 ERROR: Build of 'alsa-driver' failed Jun 26 08:26:32 I joined a pocketpc channel and now I'm helping people rather than getting helped.. had the unit what 3 days now? =-s Jun 26 08:26:49 Pigi: hm, that's not good. what's your DISTRO? Jun 26 08:27:08 DISTRO = "familiar-0.8.3" Jun 26 08:27:21 koen: after the change shall i run bitbake ? Jun 26 08:27:37 yes Jun 26 08:27:42 ah, maybe you just can't do -c package_ipk. Jun 26 08:27:47 try without any -c option Jun 26 08:27:59 that's what I was thinking. Jun 26 08:28:04 what was the error; i did not understand ? Jun 26 08:28:36 sriram: it was trying to build xserver-xorg, which was feeding x86 assembler to your compiler Jun 26 08:30:38 i have just inserted the line you gave instead of the existing one. have delete PREFERRED_PROVIDER line and added ur line Jun 26 08:30:43 is it ok ?. Jun 26 08:31:23 no, Add it, not replace it Jun 26 08:33:09 pb_ http://pastebin.com/304261 Jun 26 08:34:05 Pigi: ah, you need to adjust your FILES. Jun 26 08:34:26 I'm not sure why it's generating an error, though. That's a bit weird. Jun 26 08:40:44 is there a variable for "/etc/modutils/" ? Jun 26 08:41:03 /etc is ${sysconfdir}. Jun 26 08:41:09 ok Jun 26 08:41:11 I don't think there's a variable for /etc/modutils as such. Jun 26 08:42:17 not a problem for this.... Jun 26 08:44:57 brb Jun 26 08:45:01 * Pigi is away: I'm busy Jun 26 08:45:43 koen: as always my problem a few minutes ago was a stupid mistake: Jun 26 08:46:14 I started a build with the links to machine and distro not set, got an error. Jun 26 08:46:26 things got cached... Jun 26 08:46:46 then I set the proper links to machine and distro... Jun 26 08:47:27 and built agin. now with the mentioned errors. deleting tmp and restarting fixed it. Jun 26 08:47:53 ~lart myself Jun 26 08:47:53 * ibot slams myself against a large cement Tux Jun 26 08:48:10 ~lart chris144|home Jun 26 08:48:10 * ibot pours hot grits down the front of chris144|home's pants Jun 26 08:48:22 chris144|home : you are not alone Jun 26 08:48:24 :-) Jun 26 08:49:25 sriram: what is your problem, I have a large portion of newbie recipies... Jun 26 08:49:58 koen: it builds like hell now... Jun 26 08:50:11 i faced many problems koen and others helped me and will continue to face problems and i am sure they will help me. Jun 26 08:50:15 chris144|home: that's good to hear Jun 26 08:50:30 this irc room is god sent for people like me Jun 26 08:50:47 thanks koen. Life would be hell without people like you. Jun 26 08:51:34 koen why is it building glibc after i changed omap5912.conf ? Jun 26 08:51:44 this is going to take time. Jun 26 08:51:50 no idea, sorry Jun 26 08:51:57 sriram: yes, this is a very good irc room. Many good experiences which a lot of folks. thumbs up to the oe core guys Jun 26 08:52:30 glib-2.6.2.tar.bz2 ? Jun 26 08:53:31 sorry it is glib and not glibc. Jun 26 08:53:46 so has x server properly compiled ?. how to check ? Jun 26 08:55:16 tmp/deploy/ipk should contain the xserver binaries Jun 26 08:55:32 but glib doesn't need X ~ Jun 26 08:55:51 does glib build before X or after X Jun 26 08:57:15 I assumes before X, but I'm not sure Jun 26 08:58:00 fine Jun 26 08:58:04 time to go. Jun 26 08:58:09 and meet you tommorrow. Jun 26 08:58:15 cu Jun 26 08:58:20 take crae Jun 26 08:58:25 sorry take care Jun 26 08:58:36 * Pigi is back (gone 00:13:34) Jun 26 08:59:05 pb_ sorry.... Jun 26 09:06:10 I added : FILES_alsa-driver = "${D}/lib/modules/${KERNEL_VERSION}/misc/snd* ${sysconfdir}/init.d/ " Jun 26 09:06:26 Pigi: that ${D} is spurious Jun 26 09:06:50 I added it because I can't get the lib in package Jun 26 09:07:10 on the first try it was without ${D} Jun 26 09:07:36 Pigi: you probably need to be setting FILES_${PN}, not FILES_alsa-driver Jun 26 09:07:58 let's try ( i had copied the idea from hostap :) Jun 26 09:14:18 also this try didn't worked. Do you want to give a look at the bb file ? Jun 26 09:14:55 sure Jun 26 09:15:04 http://pastebin.com/304282 Jun 26 09:15:31 you've still got that ${D} there Jun 26 09:15:33 you need to take that out Jun 26 09:15:50 Yes, I' ve removed now ad trying again Jun 26 09:16:12 but my machine is too slow for this try&error mechanism :) Jun 26 09:17:01 Pigi: you do know the -b trick, right? Jun 26 09:17:37 yes, I know, but it works also without the -c option ? Jun 26 09:17:57 I thought it was warking only with the -c option. Jun 26 09:18:18 the -c option is... optional. Jun 26 09:18:23 it will work fine without that Jun 26 09:18:35 but you can do "-c build", if you like. Jun 26 09:18:52 then I have found a good way to workaround the "slow" problem Jun 26 09:19:11 very good Jun 26 09:28:50 It didn't work :( Jun 26 09:28:55 re Jun 26 09:33:39 what happened? Jun 26 09:34:13 nothing, always the same "NOTE: file /lib/whatsoever installed but not shipped" Jun 26 09:34:19 or something like this. Jun 26 09:34:30 okay Jun 26 09:34:39 check with oeread to see what your FILES is ending up set to Jun 26 09:35:15 can you point me to the manual ( so I will avoid to ask stupid question ) ? Jun 26 09:35:37 heh Jun 26 09:35:44 oeread is a bit retro Jun 26 09:35:48 I don't know the location offhand. It's on the berlios site somewhere. Jun 26 09:35:56 RTFMing is always a good thing :) Jun 26 09:36:16 http://bitbake.berlios.de/ Jun 26 09:38:59 google don't know nothing about oeread :( Jun 26 09:39:33 it's bbread now Jun 26 09:39:35 Pigi: s/oeread/bbread Jun 26 09:39:46 ah ok.. Jun 26 09:40:12 oh, er, right. Jun 26 09:40:19 I've never really got to grips with these new names. Jun 26 09:40:22 18:35 < koen> oeread is a bit retro Jun 26 09:40:45 ehe Jun 26 09:41:25 bbread path_to_bb/alsa.bb : FILES_alsa-driver="/lib/modules/${KERNEL_VERSION}/misc/snd* /etc/init.d/" Jun 26 09:41:35 ah, that's no good. Jun 26 09:41:43 looks like you aren't inheriting module Jun 26 09:42:05 (or module-base) Jun 26 09:42:10 that's the strange thing: If I do inherit modules, I get a compilation error Jun 26 09:42:22 okay Jun 26 09:42:23 use module-base, then Jun 26 09:42:33 in the sense that he try to compile with arm-linux-arm-linux-gcc Jun 26 09:43:02 that does seem odd Jun 26 09:43:40 | /bin/sh: arm-linux-arm-linux-gcc: command not found Jun 26 09:43:56 actually, module-base won't work, since then you won't get any CC setting at all. Jun 26 09:44:07 I guess you'll have to figure out why module.bbclass is going wrong. Jun 26 09:44:28 I think it depends from the configure script. Jun 26 10:06:55 is it possible to inherit modules after the make phase ? Jun 26 10:10:40 no Jun 26 10:16:53 maybe I found a workaround. Jun 26 10:19:49 if I add qpf fonts into upgrades feed then should I include source of qpf fonts or ttf fonts? Jun 26 10:20:28 depends on the license Jun 26 10:20:40 will put both Jun 26 10:24:08 I managed to have the ipk for the pxa arch. Jun 26 10:28:20 time to eat now. Jun 26 10:28:23 see you later Jun 26 10:28:26 * Pigi is away: I'm busy Jun 26 10:29:08 * chouimat is away: clothes shopping Jun 26 10:44:10 hrw: hunkyfonts only contain one rotation Jun 26 10:46:24 *cough* use ttf *cough* Jun 26 10:46:59 mickeyl: hunkyfonts or hunkysans/hunkyserif 0.3.0? first was done by gints, second by me Jun 26 10:47:03 koen: agreed Jun 26 10:47:09 * hrw hacked opie pim again Jun 26 10:47:13 no,no! use PCF! Jun 26 10:47:23 Twiun: Intellifonts! Jun 26 10:47:24 ah Jun 26 10:47:40 hrw: hunkysans Jun 26 10:47:51 at least the version on oz 3.5.3 feed Jun 26 10:47:55 didn't test a newer one Jun 26 10:48:07 mickeyl: place a bug in bugzilla then Jun 26 10:48:15 and assign it to me Jun 26 10:48:19 will do Jun 26 10:48:57 ~lart users of XML pim backend in opie Jun 26 10:48:57 * ibot takes a rusty axe and swings it violently, taking users of XML pim backend in opie's head off Jun 26 10:50:12 xml backend does not have sorting at all Jun 26 10:51:21 hrw: what's your bugtracker email? Jun 26 10:51:28 openembedded@hrw.one.pl Jun 26 10:52:33 #119 Jun 26 10:53:03 ok Jun 26 10:53:17 do you've ttf-verdana handy ? Jun 26 10:53:40 I use dejavu now Jun 26 10:53:59 ah Jun 26 10:56:01 mickeyl: hi, FYI - I don't know if I will be able to finish the OResource conversion for 1.2.1...my Linux box died...:( Jun 26 10:56:11 drw: again ;( Jun 26 10:56:51 hi hrw Jun 26 10:57:14 luckily, there is only launcher, mail, opietooth and ns2 left to do Jun 26 10:57:36 launcher and mail being the important ones Jun 26 10:58:55 drw: don't worry about that. what's broken? Jun 26 11:00:17 I'm hoping it's just the power supply, it had locked up hard (which is unusual), but I turned it off and then back on a minute or 2 later and nothing Jun 26 11:00:41 no fan noise, lights...no smell of smoke at least Jun 26 11:01:18 it's still under warranty, so I need to call and see about sending it in Jun 26 11:01:24 so now all that fun with checking piece by piece what work.. Jun 26 11:01:31 uff. warranty makes it easier Jun 26 11:01:58 yeah, that's what I thought...send it to them and let them figure it out Jun 26 11:02:11 I don't have any extra PSU's laying around Jun 26 11:02:21 otherwise would have tried that Jun 26 11:02:48 so probably be stuck with just my powerbook for awhile Jun 26 11:02:58 'stuck' Jun 26 11:03:46 no qvfb for osx? Jun 26 11:04:00 drw: dont forget to get hdd from machine before sending Jun 26 11:04:22 koen: yes, but I was never able to get things (opie) compiled, running on os x Jun 26 11:04:27 hrw: right Jun 26 11:05:38 should have installed YDL on my laptop, but not going to try and mess around with it now as it is my only way to do my 'real' work Jun 26 11:06:43 sure Jun 26 11:07:26 you could always try a ubuntu live cd Jun 26 11:07:56 koen: yeah maybe I'll give that a shot Jun 26 11:08:49 I always get the shivers when using rpm Jun 26 11:09:14 heh, no worse than ipkg ;) Jun 26 11:09:31 ipkg actually supports remote repos Jun 26 11:09:31 hrw: is there anything that would speak against making ttf-dejavusans the default on the hires devices? Jun 26 11:09:48 mickeyl: I can only say one thing Jun 26 11:09:53 Go for it! Jun 26 11:09:54 :) Jun 26 11:09:58 hehe Jun 26 11:10:08 ok, I will think about that for 3.5.4 Jun 26 11:10:10 on collie I use ttf too Jun 26 11:10:28 still not feeling it being slower ? Jun 26 11:11:00 mickeyl: months passed since last time I used qpf for >5minutes Jun 26 11:11:06 heh Jun 26 11:11:25 hmm, guess I don't need more arguments 'pro' then Jun 26 11:11:26 mickeyl: in gpe we turn on AA inspite it being 2 - 5 times slower Jun 26 11:11:28 and there is no tool to benchmark it Jun 26 11:11:43 the only thing we heard: "It looks more professional" Jun 26 11:11:59 yeah, opie-sysinfo is quite odd regarding text performance. i need to look at that some time Jun 26 11:12:01 mickeyl: btw - hunkysans 12 works on my c7x0 in two rotations without problem Jun 26 11:12:12 koen: absolutely. Jun 26 11:12:22 qpf don't look good with their static AA Jun 26 11:12:41 mickeyl: qt/e 2.3.10 glyph canching Jun 26 11:12:49 ehm Jun 26 11:12:49 caching* Jun 26 11:13:09 koen: hmm Jun 26 11:13:11 memory hog then? Jun 26 11:13:37 you are triggering my irony alert, be carefull Jun 26 11:13:43 oops :) Jun 26 11:14:29 gints|home: you will have one package less to maintain in OE Jun 26 11:14:39 now before i attempt to convert us to using TTF i need to look into where the odd fontsizes are. i want 13, 15, and 17 Jun 26 11:14:40 heh Jun 26 11:15:09 qpf-dejavu* contain 13 Jun 26 11:15:18 i mean when using TTF Jun 26 11:15:30 opie doesn't show odd sizes in the range of 10-20 Jun 26 11:15:33 ofontrequester magic Jun 26 11:15:39 possible Jun 26 11:15:56 ok. back to editing mail to france Jun 26 11:21:34 hmm Jun 26 11:21:39 ofontrequester is unguilty Jun 26 11:22:21 aah ha! Jun 26 11:22:29 did we ~lart TT lately ? Jun 26 11:22:32 no ? Jun 26 11:22:34 ~lart TT Jun 26 11:22:34 * ibot strangles TT with a 9-pole serial cable Jun 26 11:22:35 ~curse TT Jun 26 11:22:35 May the fleas of a thousand camels infest your most sensitive regions, TT ! Jun 26 11:22:35 not enough Jun 26 11:22:42 static int s[]={ 8, 9, 10, 11, 12, 14, 16, 18, 20, 22, 24, 26, 28, Jun 26 11:22:43 36, 48, 72, 0 }; Jun 26 11:22:48 heh Jun 26 11:23:08 * mickeyl patches that to get more sizes Jun 26 11:23:14 welcome in fscked world of TT Jun 26 11:23:38 * koen shouldn't gloat Jun 26 11:23:58 hehe Jun 26 11:27:15 * mickeyl turns on the Linux Workstation again and patches that out Jun 26 11:29:07 * hrw finish email to get r/w on opie cvs Jun 26 11:29:51 urghs Jun 26 11:30:01 bad carma Jun 26 11:30:06 ? Jun 26 11:30:10 my workstation is stuck on a reset loop Jun 26 11:30:13 can't get it on :/ Jun 26 11:30:19 * mickeyl curses Jun 26 11:30:49 * koen starts thinking about a PR blurb for Familiar <-> ngstrm Jun 26 11:30:52 hmm Jun 26 11:31:05 guess i'm the 2nd person with a broken computer now Jun 26 11:31:10 koen: excellent Jun 26 11:31:59 "ngstrm, not angstROM" Jun 26 11:32:38 mickeyl: just one of those weekends... Jun 26 11:33:50 koen: you dont use utf8? Jun 26 11:34:00 I do on osX Jun 26 11:34:08 not to sure on the linux server where irssi runs on Jun 26 11:34:34 ah.. I use irssi under screen on old FreeBSD Jun 26 11:34:42 screen -RDU and all works Jun 26 11:35:51 * mickeyl tries clearing the CMOS Jun 26 11:36:18 󶼿 for example (Polish 9 national characters) but "ngstrm" fails - no A nordic... Jun 26 11:37:00 ok, can someone do some funky utf8 tricks? Jun 26 11:37:06 󶼿 for example (Polish 9 national characters) but "ngstrm" fails - no A nordic... Jun 26 11:38:36 깾ʩ Jun 26 11:38:44 hmmm, still no luck Jun 26 11:39:06 Jun 26 11:39:15 La cigea vol. mprobo esfuerzo. Jun 26 11:39:56 koen: メインページ Jun 26 11:40:40 still no luck Jun 26 11:40:43 äöü§ περιμένω עברי 메인페이지 ĄČĘĖĮŠŲŪŽ Заглавная страница ویکی Jun 26 11:40:45 I guess irssi is too stupid Jun 26 11:40:52 I have irssi here Jun 26 11:40:55 and here Jun 26 11:41:41 I see some of those Jun 26 11:41:55 * koen /set term_charset = UTF-8 Jun 26 11:41:57 all the ones before Hrww logged on Jun 26 11:43:14 here I see only part of them but from home (Hrww) I saw all Jun 26 11:43:36 texts taken from http://www.pmichaud.com/wiki/UTF8/UTF8 Jun 26 11:43:47 hmmm Jun 26 11:44:41 my irss doesn't have term_charset Jun 26 11:44:46 err, irssi Jun 26 11:44:48 my too Jun 26 11:45:09 mine neither Jun 26 11:45:16 stinking gentoo docs! Jun 26 11:45:25 koen: you use gentoo? Jun 26 11:45:51 I'm on Debian Jun 26 11:46:04 hrw: no Jun 26 11:46:14 but google returned a gentoo doc Jun 26 11:46:27 * koen not 1337 enough for g3nt00 Jun 26 11:46:31 :) Jun 26 11:47:20 aargh Jun 26 11:47:25 I have irssi 0.8.9 Jun 26 11:47:31 0.8.10 adds utf8 Jun 26 11:47:41 Irssi 0.8.9 (20031210) - http://irssi.org/ Jun 26 11:47:48 lol Jun 26 11:47:55 * Twiun checks apt Jun 26 11:47:56 FreeBSD blue.stone.pl 4.9-RELEASE-p8 FreeBSD 4.9-RELEASE-p8 #4: Fri May 21 12:02:09 CEST 2004 black@blue.stone.pl:/usr/obj/usr/src/sys/BLUE i386 Jun 26 11:48:44 nope, 0.8.9 is the latest for me Jun 26 11:49:09 ubuntu has 0.8.10rc Jun 26 11:49:10 5 Jun 26 11:50:10 * koen /set term_type utf-8 Jun 26 11:54:21 mickey|PCdiag: font fixed Jun 26 11:57:59 I have only 0.8.4-3 Jun 26 11:58:14 how big is the newer one? Jun 26 11:58:49 the only one is about 373K Jun 26 11:58:53 s/only/old Jun 26 12:00:00 probably isn't all that big Jun 26 12:06:18 re Jun 26 12:06:33 中文 - 中文化-說明文字 - 中e文 - 简体中文 - 繁體中文 - CJK 没有 - Jun 26 12:06:47 heh.. I can input chinese but it is wrong displayed Jun 26 12:06:49 I get the 'CJK' part Jun 26 12:07:09 urdu text: ویکیپیڈیا میں خوش Jun 26 12:07:15 pb_ there ? Jun 26 12:08:11 gmame is cool Jun 26 12:08:14 http://thread.gmane.org/gmane.comp.handhelds.openembedded/4853 Jun 26 12:08:33 it includes other mailinglists in the threaded view Jun 26 12:09:54 Pigi: hello Jun 26 12:10:24 hi pb_ I have an ipk for pxa ( and should work also on sa ) for alsa drivers.... Jun 26 12:10:46 Now..... I think we could have something to discuss about alsalib and alsautils.... Jun 26 12:11:25 Pigi: you're my alsa-hero! Jun 26 12:11:31 Pigi: ah, very good Jun 26 12:11:34 ahahah Jun 26 12:11:58 koen, I suspect I won't be no more your alsa hero when you see the .bb file ;) Jun 26 12:12:35 well, now two question: 1) kernel.class ( or whatever ) need to be improved. Jun 26 12:12:50 2) what should we do with alsa-lib and alsa-utils ? Jun 26 12:13:03 I don't understand your first question. Jun 26 12:13:13 I think (2) is easy: nothing. Jun 26 12:13:41 pb_ for compiling the drivers I had the need of copy a couple pof files in kernel staging area. Jun 26 12:13:45 s/pof/of Jun 26 12:13:55 ah, right Jun 26 12:14:04 can you make a patch for kernel.bbclass? Jun 26 12:14:28 * _chronic has been hit with a stupid stick. trying to port and include icecast-1.3.12 to arm/familiar & bitbake variable expansion is doing my head in... it keeps either expanding $PV/$PN with -r0 on the end or putting version 1.0 through it. name of bb file is icecast-1.3.12.bb but can't seem to get it to name & fetch *properly*. help anyone? Jun 26 12:14:32 pxa? Jun 26 12:14:36 I can try, but I'm not sure it would be easy for me if there is someone that knows it better Jun 26 12:14:47 btw I can try. Jun 26 12:15:05 and also I need to discover which files are needed for sa arch Jun 26 12:15:36 for the 2nd question, alsa need at least alsactl to load the driver ( or similar ) Jun 26 12:15:51 _chronic: ${PN}_${PV}.bb Jun 26 12:16:04 _chronic: note the underscore Jun 26 12:16:12 _chronic: and the r0 is intended Jun 26 12:16:35 <_chronic> koen: ok. cheers big ears, will try that .... Jun 26 12:16:43 and also I need to hack /etc/init.d/alsasound to finish the whole job Jun 26 12:16:57 <_chronic> koen: re r0, sure. Jun 26 12:17:25 r0 is set by PR Jun 26 12:17:36 <_chronic> koen: yup. read that. Jun 26 12:17:51 so if you change the .bb in a way that effects the outcome, you need to bump up the PR Jun 26 12:22:36 Pigi: hm, I didn't think alsactl was required Jun 26 12:22:43 did we have that in 0.7? Jun 26 12:23:01 that's what I'm trying to understand Jun 26 12:23:15 the /etc/init.d/alsasound use it . Jun 26 12:23:24 I'm not sure we need alsassound btw Jun 26 12:25:05 bye Jun 26 12:25:59 I suspect we don't. Jun 26 12:26:06 <_chronic> koen: thanks, werked like a charm.... perhaps got dependency issues now, as compilation borks halfway through with Jun 26 12:26:10 I don't remember ever using that in the past. Jun 26 12:26:11 <_chronic> | commands.o: In function `com_dump': Jun 26 12:26:11 <_chronic> | commands.c:(.text+0x4f6c): undefined reference to `errno' Jun 26 12:26:11 <_chronic> | collect2: ld returned 1 exit status Jun 26 12:26:27 <_chronic> koen: but thanks. one step closer .... Jun 26 12:26:32 Me too, but in the jamey makefile is referenced. Jun 26 12:28:59 pb_ well, in 0.7,2 there were the alsasound script, but it wasn't working, as it needs bash that wasn't in the busybox applets Jun 26 12:30:02 heh Jun 26 12:30:16 I guess that means it isn't necessary. Jun 26 12:30:22 hey Jun 26 12:30:26 hail zecke Jun 26 12:30:35 I will remove it from the ipk then :) Jun 26 12:31:15 I have only one problem: "Warning: loading snd-pxa-uda1380 will taint the kernel: no license" Jun 26 12:31:58 ah, right Jun 26 12:32:10 stick MODULE_LICENSE("GPL") somewhere in the source file Jun 26 12:32:17 assuming the licence is actually gpl, of course Jun 26 12:32:25 I will. ( it should ) Jun 26 12:33:18 very good Jun 26 12:34:17 hrw|gone: oh I thought you had an account already Jun 26 12:34:27 which is the script that creates the devices at first boot ? Jun 26 12:35:52 on 2.4, it's handled automatically by the kernel Jun 26 12:35:54 (via devfs) Jun 26 12:36:35 but maybe alsa 0.9.6 don't handle devfs Jun 26 12:36:53 I guess that's possible. Jun 26 12:37:00 are you having trouble with /dev/sound/dsp not showing up? Jun 26 12:37:12 there is nothing in /dev/sound Jun 26 12:37:21 you remembered to load snd-pcm-oss, right? Jun 26 12:37:55 you rule :) Jun 26 12:38:11 heh Jun 26 12:38:18 you'll also need snd-mixer-oss to get /dev/sound/mixer Jun 26 12:38:35 snd_pcm_oss load it Jun 26 12:38:42 aha Jun 26 12:41:43 I had sound for some seconds, and now he is mute again :) Jun 26 12:45:29 with gxine I have some troubles : " alignement trap " while with vlc-gpe it seems to work. Jun 26 12:45:37 Anyone interessed in test it ? Jun 26 12:49:20 <_chronic> Pigi: vlc-gpe works for me out of the fam box, but only just .... it seems cpu bound on a h38000 .... Jun 26 12:49:30 * chouimat is back. Jun 26 12:50:00 <_chronic> s/h38000/h3800 Jun 26 12:50:23 _chronic, I was talking about alsa-drivers, not vlc-gpe :) Jun 26 12:51:28 <_chronic> Pigi:yup. ok. Jun 26 12:51:53 OT: can gstreamer also record sound? Jun 26 12:52:09 zecke: sure Jun 26 12:52:46 gstreamer itself is just a device for moving bits from one place to another. if you tell it to move bits from a file to the audio device, it plays sound; if you tell it to move bits in the other direction, it records. Jun 26 12:52:51 bbiab Jun 26 12:52:53 * pb_ -> Jun 26 12:53:16 Everything is a pipeline ;) Jun 26 12:53:22 hey bluelightning Jun 26 12:53:35 gstreamer will be definately part of my platform Jun 26 12:53:52 <_chronic> anyone tried libxml2 post version 2.6.10? | ./.libs/libxml2.so: undefined reference to `inithtmlDefaultSAXHandler' Jun 26 12:53:52 <_chronic> | collect2: ld returned 1 exit status Jun 26 13:00:46 pb_ I should hack kernel.bbclass in kernel_do_stage for the files I need , isn't it ? Jun 26 13:01:41 hi Jun 26 13:03:13 hey Bernardo Jun 26 13:03:18 adsl all set up? Jun 26 13:23:10 koen, ping Jun 26 13:24:00 pong Jun 26 13:24:16 do you have multiple oe ambient, isn't it ? Jun 26 13:24:26 yes Jun 26 13:24:47 well, actually only one where I can build muitlipole distros and images Jun 26 13:25:08 if I do file a bug in oe with the files for alsa, you are able to compile for sa and gives me back the errors ? Jun 26 13:25:19 sure Jun 26 13:25:23 but not tonight Jun 26 13:25:50 no, not tonight. I will try to compile it by myself,, but in case I will ask your help Jun 26 13:28:03 How can I test a new kernel.bbclass ? I should recompile everithing ? Jun 26 13:53:46 lol Jun 26 13:53:53 F7 is firefox's key of death Jun 26 13:58:47 zecke: for me not Jun 26 13:59:16 gints|home: hey, what do you refer to? ;) Jun 26 13:59:39 * chouimat is away: food and stuff Jun 26 13:59:50 hey firefox v1.0.4 Jun 26 14:00:38 Pigi: yeah Jun 26 14:00:50 gints|home: ah, F7 enters this mode Jun 26 14:00:58 gints|home: now press f7 for a long time Jun 26 14:01:29 gints|home: you get infinite number of message boxes (looks like F7 up KeyEvent is swallowed and this dialog triggers on KeyDown...) Jun 26 14:02:07 zecke: odd, that doesn't happen for me Jun 26 14:02:19 I can hold the key down as long as I like, and I only get one dialog Jun 26 14:03:11 zecke: no for me no... Jun 26 14:04:03 zecke: have you hacked firefox on some way? :) Jun 26 14:05:21 That is the Qt version of firefox (sadly not) Jun 26 14:28:58 'night all Jun 26 14:29:05 ciao koen Jun 26 14:40:40 pb_ in bug.oe.org (#120 ) there is the patch for kernel.bbclass Jun 26 14:40:54 Pigi: awesome Jun 26 14:41:45 I have tried alsa on my 3900 and it work. Should only see if this compiles for h3600,but not tonight :) Jun 26 14:47:40 what's alsa? Jun 26 14:48:06 oh. the audio library? Jun 26 14:48:19 Pigi: righto :-) Jun 26 14:49:01 good night to all. Jun 26 15:00:42 evening Jun 26 15:23:01 * chouimat is back. Jun 26 15:35:48 hi Jun 26 17:20:33 http://www.openzaurus.org/oz_website/faq/faq?id=84 Jun 26 17:20:42 Warning: main(PEAR.php): failed to open stream: No such file or directory in /home/groups/o/op/openzaurus/htdocs/oz_website/include/db.inc on line 6 Jun 26 17:49:19 ...why are alsa-drivers needed? Jun 26 17:49:22 ALSA is part of Linux now Jun 26 17:50:41 <[g2]> kergoth ping Jun 26 17:53:16 mickey|zzZZzz: you totally missed that guy's question on oz-devel... =p Jun 26 19:35:03 Luke-Jr, not all ALSA drivers are open or availible Jun 26 19:57:50 and there is also the fact user-land tools are not part of the kernel either Jun 26 20:01:05 my zaurus has an IPv6 address! :) Jun 26 20:01:37 thats too cool Jun 26 20:10:39 * BigAl grins. Jun 26 20:19:09 emte: user-land tools are not part of alsa-drivers, either... Jun 26 20:20:07 nope **** ENDING LOGGING AT Sun Jun 26 23:59:56 2005