**** BEGIN LOGGING AT Tue Oct 12 02:59:57 2010 Oct 12 03:16:48 Crofton: no not much Oct 12 03:17:08 Crofton: but I am building gnuradio Oct 12 03:17:12 as I say Oct 12 03:17:18 my machine is a bit slow Oct 12 05:56:01 gm Oct 12 05:58:15 khem, i'm not really sniffing down your neck, but as I completed cleaning the orphaned files, I decided to do a sweep to see if I by accident overlooked something and at that point noticed your openjade patch and some libxine orphaned files and removed them as part of the finising off Oct 12 05:58:35 i'm definitely not planning to do this daily or so Oct 12 06:07:23 03Andreas Mueller  07org.openembedded.dev * r703feb98b8 10openembedded.git/recipes/gnome/gedit_2.30.0.bb: (log message trimmed) Oct 12 06:07:23 gedit: add `enchant` to `DEPENDS` Oct 12 06:07:23 checking for ENCHANT... no Oct 12 06:07:23 configure: error: Enchant library not found or too old. Use --disable-spell to build without spell plugin. Oct 12 06:07:23 ERROR: Function do_configure failed Oct 12 06:07:23 I build tested this with `minimal-eglibc` for `MACHINE = "beagleboard"` and it fixes task `do_configure`. Oct 12 06:07:24 Signed-off-by: Andreas Mueller Oct 12 06:07:42 03Andreas Mueller  07org.openembedded.dev * r88354ba421 10openembedded.git/recipes/gnome/gedit_2.30.0.bb: (log message trimmed) Oct 12 06:07:42 gedit: add `gnome-doc-utils-native` to `DEPENDS` Oct 12 06:07:42 Otherwise task `do_compile` fails with the following error. `xml2po` is provided by `gnome-doc-utils-native`. Oct 12 06:07:42 | xsltproc -o gedit-C.omf --stringparam db2omf.basename gedit --stringparam Oct 12 06:07:42 db2omf.format 'docbook' --stringparam db2omf.dtd "-//OASIS//DTD DocBook XML Oct 12 06:07:42 V4.1.2//EN" --stringparam db2omf.lang C --stringparam db2omf.omf_dir Oct 12 06:07:43 "/usr/share/omf" --stringparam db2omf.help_dir "/usr/share/gnome/help" Oct 12 06:07:43 03Paul Menzel  07org.openembedded.dev * rec4a0bf8e4 10openembedded.git/recipes/gnome/gedit_2.30.0.bb: Oct 12 06:07:44 gedit: sort `DEPENDS` alphabetically Oct 12 06:07:45 Signed-off-by: Paul Menzel Oct 12 06:07:46 Signed-off-by: Frans Meulenbroeks Oct 12 06:07:46 03Paul Menzel  07org.openembedded.dev * r860f2947b5 10openembedded.git/recipes/gnome/gedit_2.30.0.bb: (log message trimmed) Oct 12 06:07:47 gedit: GPLv2+ is the correct the `LICENSE`. Oct 12 06:07:47 `COPYING` [1] and the file headers, for example [2], say GPLv2+. Only `README` [3] just mentions GPLv2. Oct 12 06:07:48 [1] http://git.gnome.org/browse/gedit/tree/COPYING?id=dada4388aa0c33486a6d48ee1c8f2adac67abb01 Oct 12 06:07:48 [2] http://git.gnome.org/browse/gedit/tree/gedit/gedit-view-frame.c?id=dada4388aa0c33486a6d48ee1c8f2adac67abb01 Oct 12 06:07:49 [3] http://git.gnome.org/browse/gedit/tree/README?id=3fdf01ec9e0ac5f48d22e7e5d369c81fa357ea49 Oct 12 06:07:49 Signed-off-by: Paul Menzel Oct 12 06:40:28 I'm running "bitbake virtual/kernel -c menuconfig", but I just see a windows flashing a short time, then it says that menuconfig succeeded. Oct 12 06:40:49 tassleho1f: have you installed ncurses on your host (or ncurses-dev) Oct 12 06:42:32 eFfeM_work: ncurses5-dev it was :). reinstalled the OS since my last menuconfig and forgot. thanks :) Oct 12 06:42:42 yw Oct 12 06:43:17 if i do a reinstall or move to a new system this is one of the things that often bites me too Oct 12 06:44:22 and its not in build-essential meta package either Oct 12 06:56:37 good morning Oct 12 07:06:18 After a fetch of upstream I get "RDEPENDS/RRECOMMENDS or otherwise requires the runtime entity 'libqtcoree4' but it wasn't found in any PACKAGE or RPROVIDES variables" Oct 12 07:09:26 I get the same for libusb (I have "IMAGE_INSTALL += libusb1") Oct 12 07:14:55 no idea on libqt but i would expect libusb1 to come from e.g. libusb1_1.0.8.bb Oct 12 07:20:01 when I add libusb1 to IMAGE_INSTALL, it needs to find a libusb-ipk in deploy, right? Oct 12 07:21:02 guess so (or maybe a libusb1 ipk) Oct 12 07:21:16 i did never bother to peek at the internals there Oct 12 07:21:39 wish I didn't have to either :) Oct 12 07:30:14 Good morning Oct 12 07:32:21 hmmz, can anyone give me pointers on how to find out the physical memory map on a custom pxa270 board? Oct 12 07:34:45 Sorry for off topic question: Why read to 0008000 while actual RAM is addr is 50008000, pastebin: http://pastebin.com/7wsuFSyZ Oct 12 07:35:56 ehlo # Oct 12 07:36:49 Sorry for off topic question: Why read to C0008000 while actual RAM is addr is 50008000, pastebin: http://pastebin.com/7wsuFSyZ Oct 12 07:44:22 uboot? Oct 12 07:44:55 tasslehoff: about libusb read http://lists.linuxtogo.org/pipermail/angstrom-distro-users/2010-October/003401.html Oct 12 07:50:11 hrw: Yes, about u-boot. Oct 12 07:51:01 I was told to do this but not know the underlying reason. Oct 12 07:54:57 mfhk, how much memory has your board? Oct 12 07:55:13 it has 128MB Oct 12 07:55:15 mfhk: some systems use the high bit to inhibit things like caching Oct 12 07:56:45 you might want to peek in this: Oct 12 07:56:46 http://tldp.org/LDP/khg/HyperNews/get/memory/linuxmm.html Oct 12 07:56:59 c0000000 is a logical address where the kernel lives Oct 12 07:57:56 who dictates the addr to be c0000000? Oct 12 07:59:55 the kernel Oct 12 08:00:06 you can change there but it is a logical address Oct 12 08:00:19 so there is not much point in changing it Oct 12 08:00:28 it is no error or so Oct 12 08:00:40 the MMU Oct 12 08:01:42 thanks a lot, I will take a look at that site. for the moment, I just use the addr as instructed. Oct 12 08:02:16 the mmu does the mapping, guess u-boot sets up the mapping, assume the kernel is not that relocatable that you can give it a different virt addresss space (and yea, I know that some work is done on relocating) Oct 12 08:03:04 kernels aren't too relocatable... Oct 12 08:03:14 say, take this addr as a magic spell. :) Oct 12 08:19:38 good morning Oct 12 08:45:04 One more question, where are these MTD partitions defined? pls see http://pastebin.com/Ai5Aw8QA Oct 12 08:45:33 Are they created by HW vendor? Oct 12 08:46:09 mfhk, what hardware is this? Oct 12 08:46:18 OK6410 Oct 12 08:47:04 Are they auto created by kenel? But they have names. How kernel know the names? Oct 12 08:47:58 they are created by uboot Oct 12 08:48:08 (or mtd-utils) Oct 12 08:48:59 I did not do it. So it's probably u-boot. Then, what u-boot command to list the partition table? Oct 12 08:51:09 my u-boot gives: Unknown command 'mtdparts' - try 'help' Oct 12 08:52:54 or are hardcoded in kernel Oct 12 08:53:09 too bad, I can't find any relevant mtd commands in my device's u-boot. :( Oct 12 08:53:30 hrw: any pointer to the kernel config? Oct 12 08:53:55 board file you need to check Oct 12 08:56:28 BTW, this kernel is built using machine mini6410.bb in OE. Oct 12 08:57:23 they are not in u-boot code, they are either in u-boot env area or in the kernel itself Oct 12 08:57:52 tried b-boot printenv, found nothing related. Oct 12 08:58:21 then it is kernel related Oct 12 08:58:22 now, searching kernel tree... Oct 12 09:02:11 Thanks! Found in "recipes/linux/linux-2.6.28/mini6410/mini6410.patch" Oct 12 09:03:45 It patches "linux/arch/arm/plat-s3c/include/plat/partition.h" to hardcode that partitions. Oct 12 09:04:29 Do you think this hardcode approach any good? I feel it is not flexible. Oct 12 09:06:32 mfhk, its the way it has been done for a LONG time. Oct 12 09:07:42 How about I want to add another kernel for rescue, then I need to recompile and replace the two? Oct 12 09:10:42 why? Oct 12 09:11:25 because the partition table changed. Oct 12 09:11:47 I need 4 partition. Oct 12 09:12:28 1- bootloader, 2 - rescue mode kernel with initramfs inline , 3 - normal kernel, 4 - rootfs Oct 12 09:13:18 are you going to be doing kernel development? Oct 12 09:14:09 No. I mean that hardcode thing make me have to **well** plan ahead before production. Oct 12 09:14:39 I prefer something like fdisk for mtd :-) Oct 12 09:16:41 if you are concerned that your rootfs might have a problem then make an initramfs for the stock kernel. Oct 12 09:17:16 unless you are going to intercept uboot and redirect it to the "alternative" kernel. Oct 12 09:18:27 thats kinda what ill be doing too, degrade to a "safe-mode" kernel/initrd in case the rootfs gets messed up Oct 12 09:18:27 Yes, that's what I think, field engineer boot into special mode... Oct 12 09:19:09 are you planning on using a SD card for your rootfs or the onboard flash? Oct 12 09:19:43 Not decided yet. There is no custmer yet :) Oct 12 09:21:16 when I bring up a board I typically make a initramfs in case things are "broke" and then change the boot args so it does'n't pivotroot to the rootfs Oct 12 09:21:33 then it comes up on the initramfs and I can fix it. Oct 12 09:21:48 Agree! Oct 12 09:21:57 but its only 1 kernel Oct 12 09:22:03 not 2 Oct 12 09:22:11 and I rarely mess with the partition table Oct 12 09:22:44 do you know the size of the kernel? Oct 12 09:23:30 I know it's doable with 1 kernel. Customize the initramfs to drop you to shell if error, something like that. Oct 12 09:23:52 I tried, about 7MB Oct 12 09:24:05 7MB with the initramfs? Oct 12 09:24:12 let me see.... Oct 12 09:24:29 heh, the only case of playing with 2 kernels and mtdparts is if you have broken/too old bootloader (old-params are now deprecated) Oct 12 09:24:40 it's 5.73MB. Oct 12 09:24:59 it's with mtd utils . Oct 12 09:25:08 ant_work, yep Oct 12 09:25:49 a LONG time ago we messed with the partitions on the Slug and I learned my lesson on messing with that. Oct 12 09:27:11 See you guys. Oct 12 09:28:32 only 500k too big for the partition Oct 12 09:30:16 well you might want 2 kernels to do upgrades too Oct 12 09:30:26 and still have the old one in place in case the new one fails Oct 12 09:32:24 we thought about that too on the spacecraft but realized it took too much storage for a maybe and instead used a firecode to put it in a mode to receive a new kernel being uploaded. Oct 12 09:32:55 it all depends on the machine bootloader Oct 12 09:33:39 with a smart initramfs you can boot the normal rootfs or an emergency one from a single kernel Oct 12 09:34:49 that works nicely and minimizes the use of precious Flash for a "oops" Oct 12 09:35:28 in order to fit his kernel he would have needed another 500k or so of flash space Oct 12 09:35:39 or pare down his initramfs. Oct 12 09:36:05 heh Oct 12 09:36:16 we shrunk kernelèinitramfs to 840kb Oct 12 09:36:41 can still boot SD/CF/MTD Oct 12 09:37:00 no network stack, thoug Oct 12 09:37:20 ah, was wondering Oct 12 09:37:27 lzma :) Oct 12 09:37:28 so module city Oct 12 09:37:38 no modules at all Oct 12 09:38:41 with the slug it was always a balancing act between what went *in* and what was a module Oct 12 09:38:49 they did a good job. Oct 12 09:39:01 in this regard, fot the particular machines I'm building for, 2.6.36's size seems decreased compared to let say 2.6.29 or early .30 Oct 12 09:39:24 cleaning effect? Oct 12 09:39:42 getting rid of cruft is always nice Oct 12 09:40:09 it looks like lot of code has been removed Oct 12 09:40:14 duplicated code Oct 12 09:40:34 great work for arm kernels ! Oct 12 09:50:56 bah bzr Oct 12 09:51:39 XorA: +11 Oct 12 09:52:24 hrw: hehe, you chose to work there :-D Oct 12 09:52:55 XorA: good is that I use bzr very very rarely Oct 12 09:53:17 XorA: git-bzr-ng rulz Oct 12 09:53:35 hrw: ubuntu gave me a whole training course on bzr/launchpad I forgot it all 10 mins later Oct 12 09:56:16 hrw, recipes/subdir idea is cool, would love to have recipes/X11, recipes/modules, recipes/libs etc too Oct 12 09:56:29 bah, earliest train to cambridge gets in at 10:50am Oct 12 09:57:02 ouch Oct 12 09:58:53 * XorA can fly though Oct 12 09:59:56 man thats a long way away. Oct 12 10:00:59 ka6sox-work: what is? Oct 12 10:01:10 Cambridge Oct 12 10:01:30 I was trying to work out if I can do the GA in one day Oct 12 10:02:41 ouch...getting in @ 10:50 might be a bit late. Oct 12 10:02:53 yes, hence looking at flights Oct 12 10:03:09 can be there before 9am Oct 12 10:03:32 * XorA has to be back in scotland on friday night Oct 12 10:03:51 even the 125 is too slow? Oct 12 10:04:48 if only driving at 120MPH were legal Oct 12 10:05:06 if it were in .de maybe Oct 12 10:05:06 XorA: depends on country :) Oct 12 10:05:25 dm8tbr: well taking a detour via germany wouldnt really help Oct 12 10:05:43 :) Oct 12 10:05:49 although a Bus @ 140kms/hr was a bit scary Oct 12 10:06:41 XorA: where are you coming from? (and thanks for reminding me to check the train schedule but from stansted to cambridge seems regular and fast) Oct 12 10:06:55 eFfeM_work: Edinburgh Oct 12 10:07:15 driving would be a bit long Oct 12 10:07:26 ka6sox-work: I bet less scary than a morris minor at 90MPH :-D Oct 12 10:07:45 dm8tbr: guess you'll get a panda Oct 12 10:07:53 oh my, a minor can make it up to 90? Oct 12 10:08:04 maybe downhill with a wind behind its back? Oct 12 10:08:17 eFfeM_work: this is going to be ignored anyway. idea is from 2005 Oct 12 10:08:27 eFfeM_work: looks very much like it Oct 12 10:08:42 hrw, pity, i think it is a good plan Oct 12 10:08:59 hrw although it is probably tough to define a proper structure Oct 12 10:09:25 ka6sox-work: we got my friends to 90 in exactly those circumstances :-D Oct 12 10:10:13 eFfeM_work: I hope top100fps also gets one, this hardware has even more potential for XBMC Oct 12 10:11:00 I remember being with a crazy person driving from Aberdeen to Edinburgh thru Fife and thinking I was going to die with all the Lorries coming so close Oct 12 10:11:09 (in an OLD mini) Oct 12 10:11:55 dm8tbr: hemost likely will, he is in top5 , only one project that could make this a draw Oct 12 10:12:07 I'd like to replace my openRD client with a panda. Oct 12 10:12:44 good :) Oct 12 10:12:49 panda is nice as it can do 1080p, for other things my beagle still is ok Oct 12 10:13:01 dm8tbr: http://www.omappedia.org/wiki/PandaBoard_Voting Oct 12 10:13:14 eFfeM_work: yes I'm waiting for the last vote to come in :) Oct 12 10:13:21 the Beagle didn't do much more than the openRD client so I did'n't swap it out. Oct 12 10:13:26 put the page on watch with email notification Oct 12 10:14:24 ka6sox actually the openrd client video out sucks (analog, there is a driver, but it is not mainlined); and support sucks too (old u-boot etc) Oct 12 10:14:36 av500 commented that it could probably feed both dvi ports at 1080p while encoding in real time too Oct 12 10:14:47 dm8tbr: that would be very cool Oct 12 10:14:51 video conferencing comes to mind Oct 12 10:15:15 eFfeM_work, yep...its best suited for controlling the telescope and taking the pictures. Oct 12 10:17:49 probably should have gotten the base board so I could add a pci card, client looked interesting because of multiple usb and dual eth Oct 12 10:18:14 dm8tbr, I like your proposal. Oct 12 10:19:17 ka6sox-work: thanks :) - I really hope we find an nice ADC. that would be already worth a ton. Oct 12 10:19:26 eFfeM_work, ya, I got it originally to replace the 3 slugs @ the observatory Oct 12 10:19:34 2mhz? what resolution? Oct 12 10:19:41 and then some QT-3d-EGL flashy-ness Oct 12 10:20:20 antti/adj looked those two up that we mention. would need to read the datasheets but IIRC they were 12 and 14 bit Oct 12 10:20:36 SPI I hope. Oct 12 10:20:40 yes Oct 12 10:20:58 of course no idea if really all the bits are usable or if the lowest ones are pure noise Oct 12 10:21:15 depends on the front end. Oct 12 10:22:31 the club has quite the GHz experiment history and we have people in some of the right EE departments, so a frontend should be possible Oct 12 10:23:21 NF is determined by that. Oct 12 10:24:15 yepp, I took the 'basics or rf-measurements' course with OH3HNY and she made sure we understood NF Oct 12 10:25:12 many years ago I worked with a TMS320C board and the front end was lousy. Oct 12 10:25:37 redesigned the front end and it is still in use today. Oct 12 10:25:47 :) Oct 12 10:26:57 * XorA needs to look at getting another armv5 debian capable smal box :-D Oct 12 10:27:36 XorA, I have quite a few slugs :D Oct 12 10:27:40 I have a pile of PMA430 units Oct 12 10:27:57 OMAP1510/5190 armv4 IIRC Oct 12 10:28:12 need something to run ldap on Oct 12 10:29:13 dm8tbr: omap1510 was armv5te Oct 12 10:29:19 iirc Oct 12 10:29:28 ;D Oct 12 10:30:29 dm8tbr: http://www.tapr.org/kits_evm56k.html Oct 12 10:30:59 56002...I see the book for it I still have Oct 12 10:33:12 ah, nice Oct 12 10:34:45 15yrs old and still cranking. Oct 12 10:38:37 that's amateur radio. things dont get obsolete that fast there Oct 12 10:39:25 dm8tbr, yep...the code is still good and since its RO I've seen no Bit Rot. Oct 12 10:40:06 that'll take another 15 years :) Oct 12 11:15:56 xora get yourself a sheevaplug Oct 12 11:16:34 eFfeM_work: probably Oct 12 11:17:57 yep... might be a good choice. or a dockstar Oct 12 11:18:29 dockstar is pricewise even nicer, but iirc it has a little bit less mem Oct 12 11:18:44 those eur 20 offers seem gone Oct 12 11:22:08 128MB one? Oct 12 11:23:42 those are nice too. Oct 12 11:51:35 anyone know how to debug and resolve libtool issues (with 2.4? Oct 12 11:51:56 current issue is openjade include the work dir in -L paths Oct 12 11:53:50 Crofton|work: do you also see lots of python-* recipes failing? (having -L/usr/lib -lpython26 while building target libs? Oct 12 11:56:25 JaMa, I haven't got that far yet :( Oct 12 12:54:14 I'd like to have OE build an HD image with a DOS partition table and specially-crafted first boot partition. Oct 12 12:54:41 image.bbclass looks like a good place to start, but doesn't appear used anywhere. Is anyone using image.bbclass? Oct 12 13:02:52 ...oh. Right. Guess it'd help if I spelled "inerit" correctly. Oct 12 13:03:47 inherit? Oct 12 13:04:09 ...right. Small wonder my grep turned up naught. Oct 12 13:11:07 openjade is fialing with libtool 2.4 Oct 12 13:11:10 any thoughts? Oct 12 13:12:35 make that openjade-native Oct 12 14:01:12 03Tom Rini  07org.openembedded.dev * rb6e9c76028 10openembedded.git/recipes/freetype/ (3 files in 2 dirs): Oct 12 14:01:13 freetype: Add 2.4.3 Oct 12 14:01:13 This fixes a number of security issues over previous versions including Oct 12 14:01:13 CVE-2010-3054 and CVE-2010-3311 Oct 12 14:01:13 Signed-off-by: Tom Rini Oct 12 14:01:13 03Tom Rini  07org.openembedded.dev * r0946177ef3 10openembedded.git/recipes/samba/ (samba-ads_3.5.6.bb samba_3.5.6.bb): Oct 12 14:01:14 samba, samba-ads: Update to 3.5.6 for security fixes Oct 12 14:01:14 There's a large number of issues fixed over 3.3.9 (and even more over Oct 12 14:01:15 the older versions) here. Still D_P = -1 for the tdb thing but if Oct 12 14:01:15 your distro was on 3.3.x 3.5.6 should be a safe upgrade. Oct 12 14:01:16 Signed-off-by: Tom Rini Oct 12 14:01:16 03Tom Rini  07org.openembedded.dev * ra5d76edf32 10openembedded.git/recipes/samba/ (8 files): Oct 12 14:01:17 samba: Drop name=src from checksums Oct 12 14:01:17 Signed-off-by: Tom Rini Oct 12 14:08:02 Tartarus: wrt the security issues fixed with freetype 2.4.3: shouldn't we then get rid of the insecure versions ? Oct 12 14:09:25 at least opie pins Oct 12 14:10:30 But getting rid of older versions is your bag ;) Oct 12 14:12:27 Tartarus, add the old version to removal.txt Oct 12 14:12:56 does someone know how to fix that: http://pastebin.com/4LQ4sB4q Oct 12 14:13:20 machine: om-gta02, image fso2-demo-image Oct 12 14:13:29 cbrake, www.openembedded.org points to melo now Oct 12 14:13:29 FATAL: This autoconf log indicates errors, it looked at host includes. <- does this mean I'm missing something on my host system? Oct 12 14:13:38 tasslehoff, no Oct 12 14:13:44 something is broke in the recipe Oct 12 14:14:03 cbrake, but www.openembedded.org returns "it works" Oct 12 14:14:22 * GNUtoo|laptop remembers somehting about python and multibytes Oct 12 14:14:31 Crofton|work: ok. it's apr-util_1.3.4.bb if someone wants to take a look. Oct 12 14:14:32 * GNUtoo|laptop will look on ml Oct 12 14:15:01 hmm Oct 12 14:15:10 I am stuck on openjade-native atm Oct 12 14:16:32 mickey|office, any ideas? Oct 12 14:16:35 hi btw Oct 12 14:16:47 because I don't find the multi-byte mails Oct 12 14:18:00 hmmm Oct 12 14:18:04 there is a python bug Oct 12 14:19:21 with a patch Oct 12 14:19:24 I'll try it Oct 12 14:21:33 mickey|office, you're still the python maintainer? Oct 12 14:21:43 if so: http://bugs.python.org/issue1704287 Oct 12 14:22:31 MAINTAINERS say so Oct 12 14:26:51 hmmm Oct 12 14:26:52 PR = "${INC_PR}.2" Oct 12 14:27:00 in python_2.6.5.bb Oct 12 14:27:11 and it keep saying: Oct 12 14:27:13 NOTE: package python-native-2.6.5-ml12.1: task do_install: Failed Oct 12 14:27:24 I'll update bitbake Oct 12 14:27:43 i highly doubt python-native failing to install has anything to do with bitbake Oct 12 14:27:45 * kergoth yawns Oct 12 14:27:51 ok Oct 12 14:27:57 but it doesn't take into account the PR Oct 12 14:28:12 and anyway I was already up to date Oct 12 14:28:39 hmmm Oct 12 14:28:52 ahhh Oct 12 14:28:57 it's not converted Oct 12 14:29:13 there is a -native recipe Oct 12 14:34:00 03Koen Kooi  07org.openembedded.dev * raaeefe4ecc 10openembedded.git/recipes/linux/multi-kernel.inc: Oct 12 14:34:00 multi-kernel: adapt to kernel_do_compile_kernelmodules change in kernel.bbclass Oct 12 14:34:00 Signed-off-by: Koen Kooi Oct 12 14:34:00 Acked-by: Frans Meulenbroeks Oct 12 14:34:00 Acked-by: Graeme Gregory Oct 12 14:34:01 03Koen Kooi  07org.openembedded.dev * r615876fe21 10openembedded.git/classes/kernel.bbclass: Oct 12 14:34:02 kernel bbclass: split do_compile into do_compile and do_compile_modules Oct 12 14:34:02 This allows recipes to insert a custom task in between building *Image and modules Oct 12 14:34:03 Signed-off-by: Koen Kooi Oct 12 14:34:03 Acked-by: Frans Meulenbroeks Oct 12 14:34:04 Acked-by: Graeme Gregory Oct 12 14:34:09 03Steve Sakoman  07org.openembedded.dev * r78dd26f49e 10openembedded.git/recipes/linux/ (9 files in 8 dirs): Oct 12 14:34:09 linux: Add recipe to build from sakoman.com linux repository Oct 12 14:34:09 Signed-off-by: Koen Kooi Oct 12 14:34:25 03Koen Kooi  07org.openembedded.dev * r0fe50ea6a3 10openembedded.git/conf/machine/include/omap3.inc: omap3.inc: set TARGET_ARCH to arm Oct 12 14:34:29 03Koen Kooi  07org.openembedded.dev * r3ac2460b92 10openembedded.git/classes/kernel.bbclass: Oct 12 14:34:29 kernel bbclass: make 'kernel-modules' meta-package use RRECOMMENDS to please insane.bbclass Oct 12 14:34:29 Signed-off-by: Koen Kooi Oct 12 14:34:29 Acked-by: Frans Meulenbroeks Oct 12 14:34:30 Acked-by: Graeme Gregory Oct 12 14:34:30 03Steve Sakoman  07org.openembedded.dev * re4fd90fc53 10openembedded.git/recipes/x-load/ (x-load-sakoman_git.bb x-load.inc): Oct 12 14:34:31 x-load: Add recipe to build from sakoman.com x-load repository Oct 12 14:34:31 Signed-off-by: Koen Kooi Oct 12 14:34:32 03Koen Kooi  07org.openembedded.dev * r200e6057a3 10openembedded.git/conf/machine/omap4430-panda.conf: omap4430-panda: work around image depends Oct 12 14:34:44 03Koen Kooi  07org.openembedded.dev * rb9e0261ad8 10openembedded.git/recipes/u-boot/u-boot-sakoman_git.bb: u-boot-sakoman: fix build Oct 12 14:34:44 03Koen Kooi  07org.openembedded.dev * r3f32b17e6c 10openembedded.git/recipes/linux/ (linux-sakoman.inc linux-sakoman_2.6.35.bb): linux-sakoman 2.6.35: fix build Oct 12 14:35:02 03Steve Sakoman  07org.openembedded.dev * rdb4c0d8df9 10openembedded.git/recipes/powervr-drivers/ (2 files in 2 dirs): Oct 12 14:35:02 omap3-sgx-modules: Fix to build for recent linux-omap kernels Oct 12 14:35:02 Signed-off-by: Koen Kooi Oct 12 14:35:02 03Steve Sakoman  07org.openembedded.dev * r36b79858bd 10openembedded.git/conf/machine/omap4430-panda.conf: Oct 12 14:35:03 machine conf: Add Panda Oct 12 14:35:03 Signed-off-by: Koen Kooi Oct 12 14:35:04 03Koen Kooi  07org.openembedded.dev * rc51cc77038 10openembedded.git/recipes/linux/ (10 files in 3 dirs): linux-omap: add 2.6.36rc7, only for beagleboard at the moment Oct 12 14:35:20 03Steve Sakoman  07org.openembedded.dev * rb19bdfd5ee 10openembedded.git/recipes/u-boot/u-boot-sakoman_git.bb: Oct 12 14:35:20 u-boot: Add recipe to build from sakoman.com u-boot repository Oct 12 14:35:20 Signed-off-by: Koen Kooi Oct 12 14:41:03 hmmm wrong fix Oct 12 14:41:07 the fix was already there Oct 12 14:48:27 Crofton|work: openjade-native/libtool 2.4 issue is known to me and its on my list after I fix python-numpy and gnuradio Oct 12 14:52:02 I abandon and rebuild from scratch as advised by JaMa Oct 12 14:52:03 poor khem :) Oct 12 14:53:28 khem: python-numpy has problem with -L/usr/lib -lpython2.6 or something else? Oct 12 14:53:32 khem, let me know if you need any help testing anything Oct 12 14:53:57 JaMa, on some machine it tries to include a host file Oct 12 14:54:04 I have access to such a beast ... Oct 12 14:54:42 oh man, i tried killing the native prefix override last night Oct 12 14:54:49 the explosion was even more fiery than i expected :) Oct 12 14:54:52 Crofton|work: I have lots of those on mine host .. http://pastebin.com/1WmukW1m Oct 12 14:55:22 I want to make a simple qt apps for embedded boards Oct 12 14:55:23 this one predates the libtool switch though Oct 12 14:55:50 should I use meta-toolchain-qte? Oct 12 14:56:28 Crofton|work: I have it only after libtool switch.. Oct 12 14:58:51 or should I build the qt4 recipe? Oct 12 14:59:11 TheNoise: that depends on whether you want to build your qt apps within oe or outside of it. Oct 12 14:59:24 its likely you'd want to use the meta-toolchain to develop it, then create a recipe for it when you need it Oct 12 15:00:30 kergoth, ideally I'd like to develop the apps like a desktop utility and then make a recipe for it Oct 12 15:01:58 kergoth, I mean that the initial development is done on the desktop using the host qt4, then switch to the embedded compilation Oct 12 15:03:12 kergoth_, is there another method do do it? Oct 12 15:04:00 03Koen Kooi  07org.openembedded.dev * r18458d0da5 10openembedded.git/recipes/linux/linux-sakoman_2.6.35.bb: linux-sakoman: remove autorev and non OE machines Oct 12 15:04:03 03Koen Kooi  07org.openembedded.dev * r7fb24d2378 10openembedded.git/recipes/x-load/x-load-sakoman_git.bb: x-load-sakoman: remove autorev Oct 12 15:04:55 TheNoise: that should work just fine. no need to bake qte4 at this time, in that case, just ensure its in the deps of the recipe when you create it Oct 12 15:05:38 damnit, i hate it when someone has a gitweb instance set up and don't put the actual link to *clone* the thing anywhere Oct 12 15:06:09 03Koen Kooi  07org.openembedded.dev * r6dd461ca40 10openembedded.git/recipes/u-boot/u-boot-sakoman_git.bb: u-boot-sakoman: remove autorev Oct 12 15:07:49 kergoth_, thanks. So I have to use qte in the embedded world right? (I read is not maintained anymore) If not, is there another qt4 version should I use in the embedded systems? Oct 12 15:08:26 no idea, i'm sorry to say -- haven't done qt development for embedded in a few years now, i'm sure someone else in here will know that, or you might try #edev or something Oct 12 15:10:03 ok, thanks kergoth. Oct 12 15:49:25 03Koen Kooi  07org.openembedded.dev * r88335ada41 10openembedded.git/recipes/linux/ (linux-sakoman_2.6.35.bb linux.inc): Oct 12 15:49:25 linux.inc: add knob to turn off cortex a8 fixups Oct 12 15:49:25 linux-sakoman 2.6.35: use above knob Oct 12 16:22:41 * foobaz hugs kergoth. My project cross compiles from ubuntu to angstrom! I just scp'ed the ipk and installed and ran "walker" and it worked. :) Oct 12 16:22:56 little hello world xlib thing Oct 12 16:23:48 grats Oct 12 16:29:36 * Tartarus rfc's killing all the old freetypes Oct 12 16:44:21 Tartarus, add the freetype versions to removal.txt Oct 12 16:54:50 I'm trying to compile a simple qt4 helloword program Oct 12 16:55:44 The program builds just fine on the desktop machine using qmake-qt4 Oct 12 16:56:14 I tried to make a recipe for the program but I am stuck on the configure step Oct 12 16:56:23 I get this error: Oct 12 16:56:35 NOTE: using qmake spec in /oe2/angstrom-setup-scripts/build/tmp-angstrom_2008_1/sysroots/armv5te-angstrom-linux-gnueabi/usr/share/qtopia/mkspecs/linux-gnueabi-oe-g++, using profiles '' Oct 12 16:56:35 ls: cannot access *.pro: No such file or directory Oct 12 16:56:35 ERROR: Function do_configure failed Oct 12 16:56:45 -- Oct 12 16:56:51 your S isn't set right, or your sourcet ree has no .pro file Oct 12 16:57:53 the .pro file is there in the WORKDIR Oct 12 16:58:19 i'm not sure how i can be any cleaner. Oct 12 16:58:23 your S isn't set right. Oct 12 16:58:29 S points to the source tre, usually a subdir within WORKDIR Oct 12 16:58:38 if your sources are sitting in the root of workdir, set S = "${WORKDIR}" Oct 12 16:58:56 ok Oct 12 17:00:39 thanks, kergoth now works, I was wrong because the helloworld.bb recipe didn't have that assignment Oct 12 17:00:47 np Oct 12 17:03:31 The build was successful! Oct 12 17:05:23 However only the -dev and -dbg packages are created Oct 12 17:08:04 (... studying the docs ...) Oct 12 17:08:04 sounds like you need a do_install Oct 12 17:08:42 I'll try Oct 12 17:09:59 Yes!! Oct 12 17:10:38 Thanks a lot kergoth ;) Oct 12 17:13:27 np Oct 12 17:17:02 Running the test program on the embedded system I get a missing lib: libQtGuiE.so.4 Oct 12 17:17:38 I've just installed qt4-embedded but it's not enougth Oct 12 17:17:44 well, installing just that ipk isn't going to magically install its dependencies Oct 12 17:17:51 unelss you're installing from a feed Oct 12 17:19:47 while installing qt4-embedded I got this line: Oct 12 17:19:49 qt4-embedded: unsatisfied recommendation for libqtguie4 Oct 12 17:20:51 how can I seach for this package? Oct 12 17:31:03 can I override a package's EXTRA_OECONF from the distro file? Oct 12 17:31:16 Yes. Oct 12 17:32:18 Don't recall the exact syntax, but it involves the use of the "pn" string in the override string to tell OE that this is a package name override, vs a machine name or such like. Oct 12 17:32:20 I was wondering - maybe having openGL in DISTRO_FEATURES would be a nice thing? Oct 12 17:32:56 * mwester doesn't know, none of the machines he works with have any sort of graphics. Oct 12 17:33:06 I am a developer of jlime and we support four machines, none of which provides any sort of DRI/openGL acceleration Oct 12 17:33:18 hm Oct 12 17:33:19 it's an override, use as any other -- the override is pn-${PN} Oct 12 17:33:25 see bitbake -e | grep \^OVERRIDES= Oct 12 17:35:28 kergoth_: you mean EXTRA_OECONF? Oct 12 17:35:44 kergoth_: I thought it would be more elegant to add support for opengl in DISTRO_FEATURES? Oct 12 17:36:07 kergoth_: what do you think? Oct 12 17:36:10 huh? Oct 12 17:36:15 no, i mean what i say Oct 12 17:36:28 if you don't know what overrides is, you need to read the manual Oct 12 17:37:45 kergoth_: I asked if you meant EXTRA_OECONF_pn-xserver-xorg Oct 12 17:37:56 huh? Oct 12 17:38:02 i don't give a shit waht you use the override for Oct 12 17:38:11 you can use it, once again ,the same way you use every other override Oct 12 17:38:36 yes, that would be how you would override it with that override, just as you would with any other, and as the documentation says Oct 12 17:38:40 kergoth_: whoah, having a bad day today? I'm just discussing a design choice here, why are you all nervous? Oct 12 17:38:53 I'm not nervous, but i'm not going to sit here and hold your hand while you cross the street Oct 12 17:39:02 i told you what the override is, do with it what you will Oct 12 17:39:43 i haven't commented on the design decision at all, i didn't answer that question. you wanted to know how to override a recipe, now you do Oct 12 17:39:50 look, I am going to push some commits and I am trying to discuss them with the community. My question is: will a commit adding opengl support to DISTRO_FEATURES be accepted? Oct 12 17:39:56 kergoth_: how is this hand holding? Oct 12 17:40:31 you asked a question, i answered it. period. i have zero interest in opengl, so i'm not going to address this Oct 12 17:40:35 kergoth_: I can either override EXTRA_OECONF or change the xserver recipe to alter it depending on FEATURES Oct 12 17:40:36 it's not that hard to understand Oct 12 17:40:55 kergoth_: ok, but I see no reason to be rude about it Oct 12 17:41:34 the reason is quite a valid one: you're being dense. Oct 12 17:42:02 * kergoth_ ignores Oct 12 17:55:54 cbrake, ping Oct 12 17:59:04 * TheNoise pulls the plug Oct 12 18:11:55 Crofton|work: hi Oct 12 18:12:37 Crofton|work: Did numpy ever compile even before libtool 2.4 Oct 12 18:12:50 its little adamant on cross compiling Oct 12 18:13:07 it does on my machine here Oct 12 18:13:22 on another machine there is an error on something like gstubs Oct 12 18:13:24 hmmm it does not use libtool at all Oct 12 18:13:31 it tries to get into host includes Oct 12 18:13:36 so I wonder whats causing it to fail now Oct 12 18:13:43 I see that Oct 12 18:13:47 but nothing to do with libtool Oct 12 18:13:57 -L/usr/lib is the real problem Oct 12 18:14:01 how about we look at the libtool issues firsrt :) Oct 12 18:14:05 yes Oct 12 18:14:06 exactly Oct 12 18:14:19 well I was trying to get to gnuradio Oct 12 18:14:24 and thats where I stumbled Oct 12 18:14:27 so Oct 12 18:14:28 ouch Oct 12 18:14:29 bither Oct 12 18:14:35 anyone have any recommendations for a freelance graphic designer? Oct 12 18:14:35 opanjade? Oct 12 18:14:45 maybe that will give us some ideas Oct 12 18:14:50 let me run numpy locally Oct 12 18:14:55 openjade: has a different issue Oct 12 18:14:59 its a QA issue Oct 12 18:15:10 where the .la files have paths into workdir Oct 12 18:15:16 so in theory it does not stop you Oct 12 18:15:17 yes Oct 12 18:15:40 you can disable the check temporarity its harmless Oct 12 18:15:43 can we just remove la files? Oct 12 18:15:45 heh Oct 12 18:15:56 until you try building on the target ... Oct 12 18:15:58 well .la files can be removed Oct 12 18:16:05 but static linking will suck Oct 12 18:16:23 we have to use something like as-needed etc. Oct 12 18:16:28 before we dump them Oct 12 18:16:30 things with .pc's should be okay, no? Oct 12 18:16:35 yes Oct 12 18:16:46 numpy builds for me Oct 12 18:17:06 what is the file it look sfor? Oct 12 18:17:08 I would like to have a look at log.do_compile ? Oct 12 18:17:36 the build system sticks in -L/usr/lib and the linker looks for libm.so and it finds it there Oct 12 18:17:59 and obviously later on linker bails out because that not an arm library Oct 12 18:18:16 yes, I also have that Oct 12 18:18:23 phizz: http://www.deleket.com/ or http://deleket.deviantart.com/ Oct 12 18:18:25 is distutils supposed to know cross compiling Oct 12 18:18:39 I'm not sure Oct 12 18:18:47 python building frightens me Oct 12 18:18:50 if you have that then how does it compile I wonder Oct 12 18:19:33 64 versus 32 bit hosts? Oct 12 18:19:34 and the build system is all in python Oct 12 18:19:50 is wont matter its arm vs x86 Oct 12 18:22:08 openjade is so jaded that it uses really old libtool Oct 12 18:22:19 and if I give it shiny new macros it pukes Oct 12 18:22:27 lovely Oct 12 18:22:34 any idea why it gets sucked in? Oct 12 18:22:43 and it does not fully use autotools Oct 12 18:22:45 what a joy Oct 12 18:23:10 so libtool 2.4 is going to discover such lame asses Oct 12 18:23:22 :( Oct 12 18:23:45 * khem needs a coffee Oct 12 18:27:55 ynezz: thanks Oct 12 18:30:50 god I hate the boost build system Oct 12 18:33:28 * Crofton|work hates Oct 12 18:33:46 * Crofton|work hates some more Oct 12 18:33:54 * Crofton|work and some more Oct 12 18:34:18 * khem waits for hate balloon to blast Oct 12 18:38:36 03Martin Jansa  07master * re0cb7893e0 10openembedded.git/recipes/openmoko-3rdparty/mokojeweled_git.bb: Oct 12 18:38:37 mokojeweled: bump SRCREV Oct 12 18:38:37 Signed-off-by: Martin Jansa Oct 12 18:38:37 03Martin Jansa  07master * r9dd4478527 10openembedded.git/recipes/dbus/ (14 files in 6 dirs): Oct 12 18:38:37 dbus: move initscript and patches shared by all dbus versions to shared dbus directory Oct 12 18:38:37 Signed-off-by: Martin Jansa Oct 12 18:38:38 03Martin Jansa  07master * r20d6a51b93 10openembedded.git/recipes/dbus/dbus_git.bb: Oct 12 18:38:39 dbus: add recipe for git version Oct 12 18:38:39 Signed-off-by: Martin Jansa Oct 12 18:38:40 03Łukasz Pankowski  07master * r219b605104 10openembedded.git/recipes/ffphonelog/ffphonelog_git.bb: Oct 12 18:38:43 ffphonelog: new recipe from SHR Oct 12 18:38:43 Signed-off-by: Martin Jansa Oct 12 18:38:43 03Justus Winter <4winter@informatik.uni-hamburg.de> 07master * r13c14c750c 10openembedded.git/recipes/inkspill/inkspill_0.1.bb: Oct 12 18:38:43 inkspill: add recipe for nice little game written by Al Sweigart. Oct 12 18:38:43 * I merely optimized the animations until it ran smoothly on my openmoko, Oct 12 18:38:44 tweaked the user interface for the mokos screen and packaged it. Oct 12 18:38:44 Signed-off-by: Martin Jansa Oct 12 18:38:45 03Martin Jansa  07master * ra73a124c5b 10openembedded.git/recipes/tasks/task-shr-feed.bb: Oct 12 18:39:16 mokosuite2: bump SRCREV Oct 12 18:39:16 Signed-off-by: Martin Jansa Oct 12 18:51:09 oh god, I just found a git repo with boost 1.44 + cmake Oct 12 18:51:20 I never though I would be glad to see anything using cmake Oct 12 18:51:55 eFfeM_work, I apologize in advance if I commit yet another crappy boost recipe and do not delete any old ones .... Oct 12 19:04:37 hm, just trying again, anyone using OSX as a build host for OE? Oct 12 19:04:47 ynezz: it does not work Oct 12 19:05:03 ynezz: there is a branch in git Oct 12 19:05:15 which was to fix it for OSX Oct 12 19:05:19 but its not complete Oct 12 19:05:29 and I am sure over time it has rotted Oct 12 19:05:46 i rebased it on master a while back, but there's more to be done Oct 12 19:05:58 yes Oct 12 19:06:20 well just got remote access to decent macmini and wanted to use it for building testing Oct 12 19:07:12 03Martin Jansa  07master * rb240989fc5 10openembedded.git/recipes/vala-terminal/vala-terminal_git.bb: vala-terminal: bump SRCREV for new vala detection Oct 12 19:07:46 ynezz: install lnx on it Oct 12 19:08:03 well, I can't, otherwise I would do it asap Oct 12 19:08:30 ynezz: if you feel motivated then use the branch and see what breaks Oct 12 19:08:30 just trying ubuntu inside virtualbox 4GB RAM, but even the parsing of the recipes is slow as hell Oct 12 19:09:16 ynezz: virtual machines are slow Oct 12 19:09:29 unless you have a machine farm Oct 12 19:09:40 seems so Oct 12 19:10:12 I think for OE we should not hesitate to demand a decent machine Oct 12 19:12:13 I wonder if it's possible to build OE on some BSD Oct 12 19:12:29 ynezz: should be easier Oct 12 19:12:35 provided you can install ports Oct 12 19:12:57 I have not tried it myself Oct 12 19:14:30 hm,looks like there's "darwin freebsd port collection" Oct 12 19:15:14 4260 ports, going to try it Oct 12 19:37:34 JaMa: around ? Oct 12 19:37:55 JaMa: the samba problem you can workaround by disabling thumb for samba Oct 12 19:38:11 add ARM_INSTRUCTION_SET = "arm" to the samba recipe Oct 12 19:38:11 khem: OK, trying now Oct 12 19:38:26 its a generic gcc problem Oct 12 19:38:40 with -fPIC -mthumb Oct 12 19:39:05 03Koen Kooi  07org.openembedded.dev * rdaafb3282c 10openembedded.git/recipes/x-load/x-load-sakoman_git.bb: x-load-sakoman: fix typo Oct 12 19:39:08 so should I put it to samba.inc for all versions, right? Oct 12 19:46:41 khem: still ICE here :/ http://paste.pocoo.org/show/274675/ Oct 12 19:47:38 anyone have a favorite VM? Oct 12 19:49:08 phizz: I use kvm Oct 12 19:50:25 03Michael Smith  07master * r5fc5e69f02 10openembedded.git/classes/package.bbclass: Oct 12 19:50:25 package.bbclass: copy dotfiles in root D to PKGD Oct 12 19:50:25 Use shutil.copytree() to copy D ("image") to PKGD ("package"). The Oct 12 19:50:25 previous system("cp %s/* ...") missed dotfiles/dirs at the top-level. Oct 12 19:50:25 Signed-off-by: Michael Smith Oct 12 19:50:25 Acked-by: Chris Larson Oct 12 19:51:16 filip: thanks Oct 12 19:51:38 phizz: what do you need it for? Oct 12 19:54:54 to run windows 95/98/ME/2000/etc for testing Oct 12 19:55:05 JaMa: hmmm thats another problem I guess Oct 12 20:09:05 JaMa: you could add -O0 to CFLAGS Oct 12 20:09:10 that should fix it Oct 12 20:12:27 khem: yes, -O0 works Oct 12 20:14:07 Anyone else running into build problem for libprelude? Oct 12 20:14:45 JaMa: yeah woraround upon workaround Oct 12 20:15:17 JaMa: although I would prefer if you could tweak the make system of samba so O0 is only passed to this one file Oct 12 20:16:42 ok will try Oct 12 20:44:53 gm all Oct 12 20:45:35 my fetch on a CVS server using ext=ssh asks for my password, whereas I should be able to login using identities. Works manually. Oct 12 20:45:59 Someone else notices this behavior? Oct 12 20:46:53 hmmm did you specify your username in src_uri Oct 12 20:47:24 hi likewise Oct 12 20:47:36 khem: Oct 12 20:47:38 NOTE: Fetch cvs://leon@cvs.sidebranch.com/repository;module=Lancero/C/driver;method=ext;rsh=ssh;tag=HEAD Oct 12 20:47:39 leon@cvs.sidebranch.com's password: Oct 12 20:47:48 cvs://leon@ Oct 12 20:47:52 likewise: env var being filtered out? Oct 12 20:48:05 kergoth_: yes but which one? HOME maybe? Oct 12 20:48:13 no, HOME is in the whitelist Oct 12 20:48:49 i'd check the ssh docs to be sure Oct 12 20:49:45 hi eFfeM, all Oct 12 20:50:25 likewise when/how are you going to elc/oedem ? Oct 12 20:51:03 eFfeM: Tue, 26Oct10 Flight FR9272 Depart EIN at 09:25 Oct 12 20:51:37 ah ok, we (walter & me) are on the evening flight Oct 12 20:52:07 I'm dropping in on the afternoon tut (android if my mem is correct) Oct 12 20:54:44 12.00 GStreamer and OMAP4 - Rob Clark, Texas Instruments could be also interesting Oct 12 20:55:15 can't find anything about the android talk Oct 12 20:55:40 ynezz: it's not ELCE2010, but let me look it up for you Oct 12 20:55:52 ynezz: it is : http://www.embeddedlinuxconference.com/elc_europe10/tutorial.html Oct 12 20:56:07 ynezz: "Workshop 2, "What else can you do with Android?" " Oct 12 20:56:11 ah Oct 12 20:56:16 workshop Oct 12 20:56:20 as if I know what I can do with Android in the first place... Oct 12 20:56:46 ynezz: ah well, the URL reads tutorial :-) guess not much work for me then :-) Oct 12 20:57:36 hm, but it might be interesting also Oct 12 20:58:14 but for 60EUR, I'll buy rather some book :p Oct 12 20:58:43 khem: found something wrt failing python packages after moving to newer libtool? Oct 12 20:58:58 JaMa: ok Oct 12 21:00:54 ynezz: tss, you know what this whole ELCE2010 things costs me!? ELCE2010 is not to replace books, trust me. Oct 12 21:01:13 I don't meet people in a book :-) Oct 12 21:01:20 I cannot drink beer in a book? Oct 12 21:01:27 ?=! Oct 12 21:01:31 :-) Oct 12 21:02:11 khem: http://paste.pocoo.org/show/274722/ Oct 12 21:10:11 kergoth_: hmm, when I manually run "env -i " it ask for my passwd, if I run "env -i HOME=/home/leon " it does not. Seems my HOME gets lost. I'm probably breaking the WHITELIST somehow. Thanks for the hint. Oct 12 21:11:40 why doesn't do_fetch run from a generated shell script? Oct 12 21:12:48 likewise: bitbake -e yourrecipe | grep HOME= Oct 12 21:12:59 do_fetch is python.. Oct 12 21:13:28 bitbake/lib/bb/fetch/ is run from it, and it uses runfetchcmd, which is its own sort of thing, it doesn't pull all the exported vars from the metadata Oct 12 21:18:31 anyone using uClibc > 0.9.30.2 ? Oct 12 21:19:33 hm, so it's gcc-cross problem Oct 12 21:20:40 I've rebased like few hours before and gcc-cross doesn't compile http://pastebin.org/158313 Oct 12 21:21:11 it's know problem? Oct 12 21:21:38 JaMa: yes I think we do not need those -L in LDFLAGS Oct 12 21:21:59 JaMa: but its proving to be not that much true Oct 12 21:22:44 CONFIG_ARGS change is ok Oct 12 21:23:03 we can not avoid it its needed for libtool 2.4 sysroot feature Oct 12 21:24:08 JaMa: that also explains why python-numpy built Oct 12 21:25:09 khem: I'm trying to find where TARGET_LDFLAGS are prepared isn't it related to staging.bbclass sysroot_stage_libdir() changes? Oct 12 21:26:48 JaMa: bitbake.conf Oct 12 21:28:11 khem: ah I see Oct 12 21:31:03 khem: I'm trying to rebuild python-native with those -L to see if it's then passed ok through distutils to python recipes Oct 12 21:31:29 khem: but now I have to sleep, gnite all Oct 12 21:31:34 JaMa: ok Oct 12 21:31:36 gnight Oct 12 21:32:03 we need to remove those from bitbake.conf and may be put into distutils.bbclass if they are so needed by distutils Oct 12 21:38:33 hm curl-native fails: http://tinderbox.openembedded.net/builds/95436/ Oct 12 21:38:43 but bedtime now, Oct 12 21:38:54 eFfeM: I fixed it last night Oct 12 21:39:10 khem, did a git pull one hour ago Oct 12 21:39:15 khem, still no luck trying to build with TMPDIR pointing to a tmpfs filesystem - attempted to do so on a 2GB tmpfs with deploy and pstage bound from disk fs - ran out of space on tmpfs after only 153 of 6029 tasks Oct 12 21:39:40 tharvey: ok Oct 12 21:39:48 with inherit += "rm_work" in my conf as well... I'm not sure what went wrong Oct 12 21:39:51 this is ridiculous, now I am writing a recipe that pulls from a git repo to get a copy of boost that build with cmake Oct 12 21:39:59 eFfeM: hmmm Oct 12 21:40:56 * tharvey gives up trying to figure out how to minimize image build time Oct 12 21:41:17 tharvey: you need atleast 6GB Oct 12 21:41:25 tharvey, get a faster machine :) Oct 12 21:41:45 oh I have upgraded to a faster machine :) still takes hours to build a stinking image Oct 12 21:41:46 tharvey INHERIT += "rm_work" Oct 12 21:41:54 is right syntax if you meant so Oct 12 21:41:59 there was an excellent post on the gumstix list a number of months back by a guy who described his cheap and fast build machine Oct 12 21:42:04 of course I have a 64bit system now so I can't use psyco - not sure how much thats hurting me Oct 12 21:42:17 khem seems libtool related, this build is still on libtool 2.2.6 Oct 12 21:42:37 /bin/sh ../i686-linux-libtool --tag=CC --mode=link gcc -isystem/home/frans/oe/tmp_minimal/sysroots/i686-linux/usr/include -O2 -g0 -Wno-system-headers -L/home/frans/oe/tmp_minimal/sysroots/i686-linux/usr/lib -Wl,-rpath-link,/home/frans/oe/tmp_minimal/sysroots/i686-linux/usr/lib -Wl,-rpath,/home/frans/oe/tmp_minimal/sysroots/i686-linux/usr/lib -Wl,-O1 -L/home/frans/oe/tmp_minimal/sysroots/i686-linux/usr/lib/..//lib -L/home/frans/oe/tmp_minimal/sysroots Oct 12 21:42:50 oops wanted to paste this: http://tinderbox.openembedded.net/public/logs/task/8668587.txt Oct 12 21:42:58 anyway, calling it a day, cya Oct 12 21:43:35 Crofton|work, thx I'll dig for that post - not clear what people consider 'fast' Oct 12 21:44:49 khem: test build finished, seems to build OK, with -L and rpath back in TARGET_LDFLAGS in distutils.bbclass Oct 12 21:45:21 tharvey, basically get a cheap quad Oct 12 21:46:20 thats what I have Oct 12 21:46:45 I don't think psyco actually help[s much Oct 12 21:46:56 Core2 Quad @ 2.66GHz, 8GB mem and 250GB SSD - still takes 3+ hours to build a filesystem image like console-image Oct 12 21:47:13 tharvey: go i7-950 and ssd Oct 12 21:47:16 and perhaps thats 'fast' I don't know Oct 12 21:47:17 from scratch though Oct 12 21:47:46 haven't seen anyone do any testing to see how much diff psyco makes Oct 12 21:47:59 seems like I spend a lot of time waiting for BB to cache on conf changes... Oct 12 21:48:02 tharvey: the real thing to minimize image build time is to customize your own distro to get all the cruft out Oct 12 21:48:08 JaMa|Zzzz: ok Oct 12 21:48:31 likewise, went that route but maintaining an out-of-tree private distro has quite a bit of cost associated with it as well... Oct 12 21:48:32 JaMa|Zzzz: remove the rpath thats not needed only -L should be enough Oct 12 21:48:41 tharvey: psyco is hardly noticable here Oct 12 21:49:10 thats good to know... I was considering re-installing 32bit OS to test pyco difference - perhaps I'll skip that Oct 12 21:49:22 but you should be making conf changes that often Oct 12 21:49:37 Crofton|work, now I'm seeing your complaint regarding searching of gmane - can't find that posting on gumstix list :P Oct 12 21:50:06 tharvey: yes, skip that. Oct 12 21:51:44 just seem to be caught up in a lot of build host issues lately: not clear if a 64bit host can build a toolchain appropriate for a 32bit host (unlikely as x64 is plastered all over the paths), ran into some obscure issues because on of our build machines here is running an ancient ubuntu 7.10 Oct 12 21:52:11 so I find myself recently doing a lot of git pulls, rm -rf tmp, rebuild, wait until the next day to see what blows up Oct 12 21:52:52 tharvey, I can't find the email Oct 12 21:52:55 tharvey: I know what you mean, cost me hours and hours as well. Oct 12 21:53:00 but it sounds like you are trying to use silly hw Oct 12 21:53:10 dev has been no fun lately Oct 12 21:53:32 silly hardware - ubuntu 7.10 completely agree - 'management' here doesn't understand what the issue is :P Oct 12 21:54:24 Crofton|work: here is a fix for python-numpy http://pastebin.com/0AXPHrky Oct 12 21:54:25 64bit host - caught by suprise on that one, have read in past that a lot of you here use 64bit so perhaps I misunderstood whether or not thats an issue for building 32bit toolchains Oct 12 21:54:54 khem, do you want me to test on the test case I have> Oct 12 21:55:14 Crofton|work: sure Oct 12 21:55:58 khem, regarding python-numpy - someone asked me the other day if there is an issue building that on x86 host destined for armv7 target as well as other python libs that provide wrappers around native code Oct 12 21:57:14 khem, by the way I am using INHERIT += "rm_work" - are you saying even with that it should require 6GB TMPDIR (excluding pstage/deploy) building something like console-image? Oct 12 22:00:15 likewise, on your i7-950 system with SSD - how long does it take to build something like console-image from scratch? Oct 12 22:00:24 or any other benchmark you may have? Oct 12 22:00:46 khem, cross yuor fingers Oct 12 22:01:41 | /usr/include/gnu/stubs.h:7:27: fatal error: gnu/stubs-32.h: No such file or directory Oct 12 22:01:41 | compilation terminated. Oct 12 22:01:46 still happens Oct 12 22:04:04 there is a -I/usr/include also Oct 12 22:04:50 prepepnd STAGING_INCDIR maybe? Oct 12 22:06:05 tharvey: let me check in tinderbox... hold on Oct 12 22:08:35 tharvey: tinderbox: builder likewise machine openrd-base Oct 12 22:08:54 tharvey: the oldest build is minimal-image, the one but oldest is console-image Oct 12 22:10:18 adding the two up (I built them one after the other): 51 minutes and that was WITHOUT SSD Oct 12 22:10:41 actually I do not quite trust these numbers... Oct 12 22:11:02 I should really do a "time bitbake console" for you from scratch. Oct 12 22:11:34 tharvey: you are using parallel bitbake and make, right? Oct 12 22:12:07 oh, the above was on a i5-750, not my (new) i7-950 Oct 12 22:12:23 tharvey: still with me? :-) Oct 12 22:12:33 not really following tinderbox there - don't see how it shows 'what' you actually built Oct 12 22:13:04 your saying that with a clean TMPDIR you build minimal-image for some machine in 51 minutes? I can't believe that Oct 12 22:13:50 likewise, I'm using BB_NUMBER_THREADS = "2" PARALLEL_MAKE="-j4" - perhaps those are not the best args for my current build system? Oct 12 22:14:26 jack up the number of threads .... Oct 12 22:15:34 what is recommended for a quad core? Oct 12 22:16:01 tharvey: I use 3 and 5 Oct 12 22:16:40 what do you mean by that? do you use 3, or do you use 5? :) Oct 12 22:16:40 tharvey: that's why I mentioned what target I built, tinderbox only shows the packages. Oct 12 22:16:46 3 resp. 5 Oct 12 22:16:56 BB_NUMBER_THREADS = "3" PARALLEL_MAKE="-j5" Oct 12 22:17:23 I'll give that a try in tonight's build Oct 12 22:18:03 tharvey: I guess a 5-15% speed-up. i7 gives you another 30%, and a fast disk another 20% or 30% Oct 12 22:18:42 so that's about factor 1.86 speedup Oct 12 22:19:09 clock speed can add some Oct 12 22:19:39 however, the best speed-up is by creating your own distro (if you build small images, not much software in them) Oct 12 22:19:54 curious... in tinderbox all these distros of 'angstrom' - thats not a distro (angstrom-2008.1 and angstrom-2010.x are) - where does that value come from in tinderbox? Oct 12 22:20:03 Angstrom build ~3600 packages just for a console image Oct 12 22:20:15 tharvey: dunno Oct 12 22:20:22 not understanding how the distro affects your image so much, its your 'image' recipe that would affect that correct? Oct 12 22:21:17 I've got a SRC_URI that fetches a file called cmake-1.44.0 Oct 12 22:21:27 how do I persuade bb to unpack it Oct 12 22:21:30 it is tgz Oct 12 22:22:32 yuck, they omitted the filename extension? Oct 12 22:22:53 yeah Oct 12 22:23:06 it comes from a gitorious git repo Oct 12 22:23:24 kergoth_: ah, cvs.py doesn't use runfetchcmd like the others .py do... that's the reason for my regression Oct 12 22:23:43 ah. that should definitely be fixed then Oct 12 22:23:49 kergoth_: none of the exports are in place. Oct 12 22:24:12 it uses os.system on the cmd Oct 12 22:25:11 how do I tell bitbaek the file extension is wrong Oct 12 22:25:23 Crofton|work: ? Oct 12 22:25:26 Crofton|work, dunno if that's in place, but should be.. Oct 12 22:25:44 the tarball name does not have a tgz extension Oct 12 22:26:13 bb can't unpack, just copies it into work Oct 12 22:26:26 Crofton|work: hmm, custom do_unpack() .. (yikes...) Oct 12 22:26:37 yeah, I was headed there Oct 12 22:26:40 likewise: like i said, definitely needs to be fixed. ideally, we'd drop all system() usage everywhere in bitbake and oe.. Oct 12 22:27:43 kergoth_: was already replacing os.system by runfetchcmd (ignoring the different calling conventions). I will search for my python books (again). Oct 12 22:28:00 Crofton|work, perhaps fetch it in a function, rename it and ref it as local file in SRC_URI? Oct 12 22:28:11 cool, would be happy to apply a patch -- don't know if i'd have time to do it myself until the weekend Oct 12 22:28:49 Crofton|work, a do_unpack2 or so would be easiest Oct 12 22:28:55 kergoth_: wish I could be more of service on the .py side of life... I always fail there Oct 12 22:29:00 see the csl gcc recipes Oct 12 22:29:20 Crofton|work: I'd guess we should enhance do_unpack and unpack_file to handle an override of the handler Oct 12 22:29:29 shouldn't be tterribly difficult Oct 12 22:29:32 I think I am goign to change the file name Oct 12 22:29:43 and see if this makes life any better Oct 12 22:29:50 I hate boost Oct 12 22:30:06 who likes it? Oct 12 22:30:42 for me it's comparable to technologies like .Net Oct 12 22:31:08 some people like it Oct 12 22:31:22 I wouldn't mind it, if they used a sane build system Oct 12 22:33:19 rofl Oct 12 22:33:30 hacking the file name worked Oct 12 22:33:50 I should push this recipe, just to annoy people :) Oct 12 22:34:41 also, it isn't bitbake unpacking, its entirely in lib/oe/unpack.py and base.bbclass Oct 12 22:34:57 sorry for my confusoin :) Oct 12 22:35:13 I was wondering why I couldn't find unpack in the bitbake manual Oct 12 22:35:23 well, now you know :) Oct 12 22:35:39 * kergoth_ still thinks fetch+unpack could be merged in a nifty way Oct 12 22:53:47 likewise, what distro are you using? why wouldn't the image be the thing dictating the deps built? I don't see angstrom building things not required for images Oct 12 22:53:50 khem, how hard is it for me to ggo back to the older libtool? Oct 12 22:53:57 I need to test some stuff. Oct 12 22:54:21 Crofton|work, just don't set P_V for 2.4? Oct 12 22:54:24 Or you mean totally back out? Oct 12 22:54:42 just get something to work Oct 12 22:55:06 2.4 is opt-in, it's controlled by a variable -- adjust LIBTOOL_HAS_SYSROOT and prefer the other version Oct 12 22:56:28 got it Oct 12 22:57:25 looks like we used to use 2.2.6b ... Oct 12 23:01:37 ok back to libtool 2.2.6b for a day Oct 12 23:01:49 need to see if recent boost solves my gnuradio mutex headache Oct 12 23:06:46 what are you doing on usrp? Oct 12 23:08:03 me? Oct 12 23:08:12 yep Oct 12 23:08:28 getting gnuradio running on omap3 Oct 12 23:08:36 for usrp embedded Oct 12 23:09:39 I'm just considering usrp2 + rfx900 for playing in depth with UHF RFID - EPC GEN2 tags/readers Oct 12 23:09:48 cool Oct 12 23:10:11 people have done a lot of neat things with usrps Oct 12 23:10:49 yep, that's why I asked Oct 12 23:11:23 the really cool thing is the range of things people do Oct 12 23:11:47 I am amused there seems to be a talk a year by someone who did some neat hacking with a usrp Oct 12 23:11:59 a talk a year at DEFCON Oct 12 23:13:07 other people do the really cool stuff Oct 12 23:13:15 I just try to make it possibel for them Oct 12 23:13:55 Crofton|work, there are other in tree examples of files in SRC_URI with bad extensions. grep do_unpack, i forget where i saw it Oct 12 23:14:10 cool Oct 12 23:14:11 thanks Oct 12 23:20:46 Crofton|work, flnx/flnx_0.18.bb is an interesting one Oct 12 23:33:47 yikes Oct 13 00:07:09 anybody can help-me with this problem? i made a image based on console-image named console-mplayer-image. on this new image,i only add mplayer to be compiled,but i get a error on do_rootfs Oct 13 00:07:40 http://pastebin.com/mYa4Zkab Oct 13 00:07:51 anybody can help-me ? Oct 13 00:10:31 i need pastebin my new bb file ? Oct 13 00:11:24 angelox_123, yes Oct 13 00:11:35 i minute please Oct 13 00:11:38 1* Oct 13 00:11:48 I'm guessing you've copied console-image.bb but did not change the IMAGE_NAME defined in the recipe so bitbake is confused Oct 13 00:12:25 http://pastebin.com/MZb7JwXi Oct 13 00:12:39 i had did that,but after i changed but the same error Oct 13 00:13:07 you mean IMAGE_BASENAME isn't? Oct 13 00:13:31 You've got 2 problems there in that first pastebin Oct 13 00:13:40 That conffile one will bite you in the butt and make your rootfs not work well Oct 13 00:13:52 What MACHINE? Oct 13 00:13:58 Is it one that sets IMAGE_DEVICE_TABLES ? Oct 13 00:14:08 machine = a1200 Oct 13 00:14:14 distro = angstrom-2008.1 Oct 13 00:14:27 ok Oct 13 00:14:34 So now, add mplayer to DEPENDS Oct 13 00:14:39 and to IMAGE_INSTALL Oct 13 00:15:11 only this? Oct 13 00:15:25 That'll fix the not building of mplayer problem Oct 13 00:15:28 sec for the other one Oct 13 00:15:38 what file? Oct 13 00:15:45 console-mplayer-image.bb Oct 13 00:16:10 no,you say: sec for the other one...what other one? Oct 13 00:16:15 sec Oct 13 00:16:20 line 181 Oct 13 00:16:39 ah Oct 13 00:16:42 what i do on that? Oct 13 00:16:52 That lacks a clean fix atm Oct 13 00:17:04 The hack is to open up recipes/initscripts/initscripts_1.0.bb Oct 13 00:17:07 look for CONFFILES Oct 13 00:17:09 and remove that line Oct 13 00:17:13 bitbake -c clean initscripts Oct 13 00:17:14 try again Oct 13 00:17:22 ok thanks by help i'll try Oct 13 00:17:36 * angelox_123 says thanks to Tartarus Oct 13 00:17:43 :) Oct 13 00:17:43 np :) Oct 13 00:18:35 i need bump pr (on initscripts_1.0.bb) ? Oct 13 00:18:53 Locally, won't hurt Oct 13 00:18:59 :) Oct 13 00:20:45 * angelox_123 says thanks again to Tartarus Oct 13 00:29:00 Tartarus,another error Oct 13 00:29:05 http://pastebin.com/zRA4yEUy Oct 13 00:30:49 any help? Oct 13 00:32:26 hm Oct 13 00:33:03 Well, that's the same error :) Oct 13 00:33:13 Do a bitbake -g console-mplayer-image | grep ^DEPENDS Oct 13 00:33:15 er Oct 13 00:33:16 bitbake -e Oct 13 00:33:38 hum? bitbake -e console-mplayer-image | grep ^DEPENDS ? Oct 13 00:35:15 yes Oct 13 00:35:26 that will show what that image says it must have to build Oct 13 00:35:27 doing :) Oct 13 00:37:09 you have any idea of why that is occurring ? Oct 13 00:37:14 ops Oct 13 00:37:15 done Oct 13 00:37:33 What was the output? Oct 13 00:37:37 http://pastebin.com/FCYkXfNt Oct 13 00:38:38 OK, so it should be building mplayer Oct 13 00:38:49 and I bet if you do a bitbake mplayer, it says you're all built Oct 13 00:38:50 right? Oct 13 00:39:07 So, find tmp/deploy/glibc/ipk -name mplayer* Oct 13 00:39:17 ok Oct 13 00:39:44 i think i know what is the problem Oct 13 00:39:46 see: Oct 13 00:39:48 deploy/glibc/ipk/all/mplayer-common-dev_0.0.1-r6.6_all.ipk Oct 13 00:39:48 deploy/glibc/ipk/all/mplayer-common_0.0.1-r6.6_all.ipk Oct 13 00:39:48 deploy/glibc/ipk/all/mplayer-common-dbg_0.0.1-r6.6_all.ipk Oct 13 00:39:48 deploy/glibc/ipk/iwmmxt/mplayer-doc_0.0+1.0rc3+svnr30165-r26.6_iwmmxt.ipk Oct 13 00:39:51 deploy/glibc/ipk/iwmmxt/mplayer-dev_0.0+1.0rc3+svnr30165-r26.6_iwmmxt.ipk Oct 13 00:39:53 deploy/glibc/ipk/iwmmxt/mplayer_0.0+1.0rc3+svnr30165-r26.6_iwmmxt.ipk Oct 13 00:39:55 deploy/glibc/ipk/iwmmxt/mplayer-dbg_0.0+1.0rc3+svnr30165-r26.6_iwmmxt.ipk Oct 13 00:39:59 dont have mplayer on armv5te directory Oct 13 00:40:10 can be isn't? Oct 13 00:40:13 Right, so that's fine Oct 13 00:40:17 Now, examine ... Oct 13 00:41:00 what? Oct 13 00:41:13 Looking :) Oct 13 00:41:21 the opkg-cl command wasn't in your pastebin Oct 13 00:41:31 but up in your log.do_rootfs, just before what you did pastebin Oct 13 00:41:37 will be opkg-cl -f .../opklg.conf Oct 13 00:41:39 er, opkg.conf Oct 13 00:41:51 examine that file, see if it knows about your ipk/iwmmxt directory Oct 13 00:42:10 examine opkg.conf ? Oct 13 00:43:47 Yes Oct 13 00:43:48 on my tmp directory i have various opkg.conf Oct 13 00:43:52 Right Oct 13 00:44:00 But log.do_rootfs that you made that pastebin from Oct 13 00:44:04 Will have an opkg-cl command in it Oct 13 00:44:09 That says -f /path/to/one/opkg.conf Oct 13 00:44:15 That's the one to look at Oct 13 00:44:33 ok i'll search Oct 13 00:46:20 oh Oct 13 00:46:32 i think it don't have of iwmmxt directory Oct 13 00:46:37 see: Oct 13 00:46:38 http://pastebin.com/gM5CpLkx Oct 13 00:46:59 So, there we go :) Oct 13 00:47:00 Now, why... Oct 13 00:47:05 digging Oct 13 00:47:50 Did you modify conf/machine/include/motorola-ezx-base.inc ? Oct 13 00:48:01 iirc no Oct 13 00:48:14 OK Oct 13 00:48:48 Thinking.. Oct 13 00:49:12 :) Oct 13 00:49:15 me too Oct 13 00:49:23 So, mplayer_svn.bb forces it to be an iwmmxt arch package Oct 13 00:49:33 But, how is this supposed to work when that's not normally the arch? :) Oct 13 00:49:34 * Tartarus digs Oct 13 00:49:50 lol Oct 13 00:51:00 it what i want is too hard Oct 13 00:51:19 No, there's gonna be a few ways to fix this :) Oct 13 00:51:53 if i copy the compiled ipk to armv5te directory will work? Oct 13 00:51:57 OK Oct 13 00:51:58 So, try this Oct 13 00:52:00 edit Oct 13 00:52:04 conf/machine/include/motorola-ezx-base.inc Oct 13 00:52:05 add Oct 13 00:52:26 PACKAGE_EXTRA_ARCHS += "iwmmxt" Oct 13 00:52:34 Then bitbake console-mplayer-image again Oct 13 00:52:39 If that works, lemme know and I'll post a patch Oct 13 00:53:02 hi , i m having trouble on nohup.out , my nohup.out grows so much and takes my all HDD , how can i put limit on nohup.out by nohup ? Oct 13 00:53:18 Find out why it's growing so large :) Oct 13 00:53:23 That's probably a bug somewhere Oct 13 00:53:35 yes but i cant put a limit on nohup.out ? Oct 13 00:54:09 sometimes my alpha version java code loops on some bug Oct 13 00:54:14 and grows nohup.out much Oct 13 00:54:18 where i put "PACKAGE_EXTRA_ARCHS += "iwmmxt"" ? Oct 13 00:54:20 on final? Oct 13 00:54:39 angelox_123 in conf/machine/include/motorola-esx-base.inc, anywhere in there will do Oct 13 00:54:42 I'm doing.. Oct 13 00:54:47 , device ? Oct 13 00:54:51 whats that device Oct 13 00:55:01 angelox_123 , can u show ur board ? Oct 13 00:55:02 me? a1200 Oct 13 00:55:08 http://pastebin.com/vtw60x2U Oct 13 00:55:38 ah Oct 13 00:55:40 these old phones Oct 13 00:55:45 :( Oct 13 00:56:00 can we install oe on iphone ? hehe Oct 13 00:56:02 * i only have money to these old phones LOL Oct 13 00:56:11 JDuke128: Nope, nohup takes no such options Oct 13 00:56:42 Try using less and 'F' instead? possibly while invoked under a 'screen' instance if you need to detach from the terminal or whatever Oct 13 00:58:59 ok i changed motorola ezx base . inc ! openembedded will need rebuild all ? Oct 13 01:00:55 tartarus ? Oct 13 01:01:05 It will rebuild that cache file Oct 13 01:01:14 But no, it shouldn't rebuild any programs Oct 13 01:02:07 ok Oct 13 01:02:09 thanks Oct 13 01:04:03 Lemme know if it works so I can post the fix :) Oct 13 01:05:46 i'll say to you,now it is doing: Handling BitBake files: \ (3319/7141) [46 %] Oct 13 01:05:49 :) slow Oct 13 01:06:39 my pc is toooooooo oollldddd Oct 13 01:06:51 intel p4 3.ghz 1gb ram Oct 13 01:07:00 heh Oct 13 01:07:18 So, if this works, can I add a Tested-by line from you? Need a name and an email address :) Oct 13 01:08:02 yes..if works of course! Oct 13 01:10:18 you pc is good? Oct 13 01:10:32 Yeah, but also always busy :) Oct 13 01:10:41 * Tartarus has something going on the latop and something else going on the desktop Oct 13 01:11:40 * angelox_123 needs a notebook...him pc is too noisy Oct 13 01:11:58 haha Oct 13 01:12:13 my cooler is too noisy and i can't use pc at night Oct 13 01:15:15 Put your build machine in the basement, where it can make as much noise as it wants... :) Oct 13 01:15:54 heh Oct 13 01:19:12 NOTE: package console-mplayer-image-1.0-r0: task do_rootfs: Started Oct 13 01:19:48 now we need luck Oct 13 01:22:19 Lol Oct 13 01:22:21 worked Oct 13 01:22:27 great thanks again Tartarus Oct 13 01:22:34 * angelox_123 says thanks again to tartarus Oct 13 01:23:13 np Oct 13 01:23:14 you'll need my email and name isn't ? Oct 13 01:23:18 Yeah Oct 13 01:23:21 okay: Oct 13 01:23:24 name: Oct 13 01:23:31 Angelo Stefanos Mavridis Bartolome Oct 13 01:23:33 email: Oct 13 01:23:36 angelo@pynell.com Oct 13 01:23:58 thanks Oct 13 01:24:13 no,i say thanks by help :) Oct 13 01:24:36 03Tom Rini  07org.openembedded.dev * r6633370f29 10openembedded.git/conf/machine/include/motorola-ezx-base.inc: Oct 13 01:24:37 motorola-ezx-base.inc: Add iwmmxt to PACKAGE_EXTRA_ARCHS Oct 13 01:24:37 Since we have this as a MACHINE_FEATURES we need this to be Oct 13 01:24:37 able to add contents to the rootfs that use the feature. Oct 13 01:24:37 Tested-by: Angelo Stefanos Mavridis Bartolome Oct 13 01:24:37 Signed-off-by: Tom Rini Oct 13 01:25:04 now i know how to add packages in images with work :) Oct 13 01:26:53 so,i'll test on phone now,but i think it will work :) Oct 13 01:27:01 thanks again! Oct 13 01:27:16 np :) Oct 13 01:27:19 Thanks for helping out Oct 13 01:27:32 :) Oct 13 01:57:08 Bye Oct 13 01:57:33 03Khem Raj  07master * rd8ce8acea7 10openembedded.git/recipes/eglibc/ (eglibc_2.12.bb eglibc_svn.bb): Oct 13 01:57:33 eglibc_2.12/eglibc_svn: Update to latest svn tip Oct 13 01:57:33 Signed-off-by: Khem Raj Oct 13 02:00:59 khem: can you remember what the problem was with the use of exec_task in srctree, offhand? Oct 13 02:01:02 i'm drawing a blank Oct 13 02:07:45 kergoth: threading issues Oct 13 02:08:55 Crofton|work: its easy to roll back. In you local.conf set the version of libtool to 2.2.6b Oct 13 02:09:28 khem: hmm, can you recall the details? Oct 13 02:13:23 I think it operated on a copy of datastore Oct 13 02:13:33 and we wanted to operate on common data store Oct 13 02:14:25 and it lost task specific overrides Oct 13 02:15:05 and RP also mentioned that bitbake has some assumptions about threading and they break too if we use exec_task Oct 13 02:15:25 I did not dig into what those assumptions were. prolly you might know more Oct 13 02:16:25 hmm. Oct 13 02:27:46 * kergoth_ kicks bitbake **** ENDING LOGGING AT Wed Oct 13 02:59:57 2010