**** BEGIN LOGGING AT Tue Sep 11 02:59:56 2007 Sep 11 04:15:58 hughescr, ping Sep 11 05:28:47 good morning all Sep 11 07:36:55 morning all Sep 11 07:40:15 RP : hi ! Sep 11 07:51:46 hi Sep 11 08:10:57 morning Sep 11 08:13:29 morning Sep 11 08:25:11 morning. when compiling glibc-intermediate-2.6.1-r0 i get a strange Sep 11 08:25:11 | running configure fragment for nptl/sysdeps/pthread Sep 11 08:25:12 | configure: error: compiler support for __thread is required Sep 11 08:25:17 error during configure Sep 11 08:25:23 any1 experienced that? Sep 11 08:30:57 morning Sep 11 08:36:19 hi all Sep 11 08:37:43 I am starting to dislike angstrom images ;( Sep 11 08:41:13 hrw: why? Sep 11 08:41:54 bluez-gnome recently Sep 11 08:43:09 and task-base got too many extra deps Sep 11 08:43:36 bluez-utils req cups and gstreamer so it also gets x11 and many other things Sep 11 09:05:05 bluez-utils -> gst-plugins-base -> esound -> gconf-dbus -> gtk+ :( Sep 11 09:25:41 can someone help me to create a bbfile ? it fails miserably in do_configure, like if there was no configure file (but there is), so I can't figure why... Sep 11 09:26:53 http://rafb.net/p/MHnitj62.html <-- the bb file Sep 11 09:28:25 nud: pwlib? Sep 11 09:28:32 hrw: yes Sep 11 09:28:36 but I need the cvs version Sep 11 09:28:39 we do not have it in OE? Sep 11 09:28:48 not the cvs version Sep 11 09:28:59 also I tweaked the configure options a bit Sep 11 09:29:13 but it's basically a cutnpaste with sources and configure options tweaked Sep 11 09:29:27 nud: drop RDEPENDS, use EXTRA_OECONF += and drop prefix Sep 11 09:30:04 nud: check does S var point to proper place Sep 11 09:30:29 the localdir is pwlib, so yes Sep 11 09:31:55 nud: check log.do_configure and run.do_configure in WORKDIR/temp/ Sep 11 09:32:44 eck Sep 11 09:32:47 I think I found it Sep 11 09:33:07 S = "${WORKDIR)/pwlib" Sep 11 09:33:36 sorry for the noise :( Sep 11 09:33:41 thanks anyway Sep 11 09:34:41 np Sep 11 09:47:59 hi, all! Sep 11 09:48:02 http://maemo.org/news/announcements/view/1189194936.html Sep 11 09:57:35 already discussed Sep 11 10:10:50 Hi, is there a way with bitbake that I can conditionally patch based on a variable? So far I've added a task, I source CONFIG_SITE then I have a condition for the thing I want, but I don't know how to apply a patch from here. Appending to SRC_URI doesn't seem to work. Sep 11 10:11:20 trj: look at dropbear Sep 11 10:11:57 cheers. Sep 11 10:15:20 is openldap broken for someone else ? Sep 11 10:17:43 it was last time Sep 11 10:18:53 mickeyl: Why yaffas2 out of interest? Sep 11 10:28:52 hrw, is it in OE yet? Sep 11 10:30:32 RP: i think because of performance problems w/ large flash sizes Sep 11 10:35:50 slapin: not yet Sep 11 10:35:56 ~logs Sep 11 10:35:58 methinks logs is apt/ibot/infobot/jbot/purl all log daily to http://ibot.rikers.org// where channelname is html encoded ie: %23debian | lines that start with a space are not shown | some channels have stats at http://ibot.rikers.org/stats/.html.gz Sep 11 10:36:01 slapin: check yesterday log Sep 11 10:38:37 mickeyl: mount time of 256M flash in gta02? Sep 11 10:38:55 hrw: for example, yes Sep 11 10:39:24 also some devices with 1G of flash Sep 11 10:39:36 where jffs2 doesn't seem to cut it Sep 11 10:40:24 mickeyl: RIght, I can understand the problem in those size of devices :-( Sep 11 10:51:36 jffs3 ? Sep 11 10:56:57 mickey|uni: hdx8? Sep 11 10:57:41 shhh - that's a secret. Sep 11 10:58:41 rwhitby: hdx8 is not a secret Sep 11 11:00:18 what it's application is is a secret from what I saw in the latest email on the subject on the mailing list. it's technical specs of course are not secret. Sep 11 11:22:15 slapin: btw, does matchbox-kbd work for you at all in recent build? Sep 11 11:34:20 re Sep 11 11:34:38 hrw: what did your davinci kernel build produce yesterday? also 4.4MB image? Sep 11 11:35:24 yes Sep 11 11:35:32 ok.. so it was not just my setup.. Sep 11 11:35:35 mhm Sep 11 11:35:38 weird Sep 11 11:35:47 Jin^eLD: you have that board? Sep 11 11:35:53 hrw: yes Sep 11 11:36:22 Jin^eLD: consider enabling tickless setup and drop initrd Sep 11 11:36:52 ~spell acknowledgement Sep 11 11:36:54 possible spellings for acknowledgement: acknowledgment acknowledgments acknowledgment's acknowledged Sep 11 11:37:15 hrw: I guess I'd have to read up on the tickless thing, first time I hear about it Sep 11 11:37:40 Jin^eLD: CONFIG_NO_HZ is magic word for searchhing Sep 11 11:37:52 thanks! Sep 11 11:38:33 hrw: do you have that board too? Sep 11 11:39:25 no Sep 11 11:40:22 Jin^eLD: you boot from nand? Sep 11 11:40:44 hrw: no, actually I wanted to make a setup and boot from NFS Sep 11 11:40:57 easier for development Sep 11 11:41:03 TFTP/NFS that is Sep 11 11:42:30 Jin^eLD: sure Sep 11 11:42:39 Jin^eLD: davinci kernel config is mess Sep 11 11:43:24 well the whole mvl stuff looked very weird and strange Sep 11 11:43:34 and somewhat confusing Sep 11 11:44:04 no - you have many subsystems enabled and not used Sep 11 11:44:21 1Ge, 10Ge, DAB, V4L etc Sep 11 11:44:34 well thats the other thing but I did not dig that deep yet Sep 11 11:44:52 first I wanted to get it to work with OE Sep 11 11:45:06 problem is that Crofton does not have the board, so he more or less made a "blind" port Sep 11 11:45:11 and could not test anything Sep 11 11:45:36 32G log.do_compile.6792 Sep 11 11:45:38 Jin^eLD, someone in #davinci did boot a board with angstrom Sep 11 11:45:40 now THATS impressive. Sep 11 11:45:56 Crofton: I guess I should ask that person how he built the kernel then.. Sep 11 11:46:08 I think he used the one I built Sep 11 11:46:20 we should redo the defconfig, since it was based on an older git Sep 11 11:46:34 Crofton: ok how did you do it? :) hrw and me get a 4.4MB kernel image Sep 11 11:46:35 then they updated got from .21 to .23 or something like that :) Sep 11 11:46:38 and that kernel does not work for me Sep 11 11:46:47 he used the .21 kernel Sep 11 11:46:48 I also tried a newer defconfig file from their update Sep 11 11:46:53 aha.. hmm Sep 11 11:46:59 -rwxr-xr-x 1 hrw hrw 3,1G IX 10 23:55 Image Sep 11 11:47:10 I would do a make davinic_config Sep 11 11:47:24 or whatever that target is and adjust specific options for angstrom Sep 11 11:48:29 Crofton: you mean tune the .bb to do that? Sep 11 11:48:39 ok, need to shower, ea, shop and hit the road Sep 11 11:48:48 update the dfconfig in oe Sep 11 11:48:53 not use the bb fiel Sep 11 11:49:02 gotta run Sep 11 11:49:08 k Sep 11 11:49:16 well, did update the defconfig already :P did not help Sep 11 11:49:31 * Crofton wonders why he decided to cross from Canada to US today ... Sep 11 11:49:34 bother Sep 11 11:49:37 :) Sep 11 12:12:04 psokolovsky_, I'll test that soon Sep 11 12:14:57 slapin: nice. for me, clicking on taskbar icon give no effect. Sep 11 12:23:41 Jin^eLD: its binutils 2.18 ;( Sep 11 12:24:22 cbrake_away: do you have a hint on how to work with jffs2 on the cmx270 ? Sep 11 12:25:43 03hrw 07org.oe.dev * r86f3076c... 10/ (1 packages/madwifi/madwifi-ng_r2702-20070903.bb): madwifi-ng: make r2702 default for alix (which use 2.6.23-rc kernel) - close #2974 Sep 11 12:28:21 any ixp4xx/openocd expertise around? Sep 11 12:29:51 hrw: uuh.. can you recommend another version that I could try with? Sep 11 12:32:55 2.17.50.0.12 Sep 11 12:33:10 that one gave me 1.4M kernel Sep 11 12:37:09 ok let me try that Sep 11 12:37:17 that is set in angstrom conf file? Sep 11 12:37:26 or can I override it in the kernel recipe? Sep 11 12:38:08 just build it by hand Sep 11 12:38:16 and then check Sep 11 12:38:33 by hand as in "not in OE"? Sep 11 12:40:53 nud: hey, how are things going? Sep 11 12:41:27 cbrake: well, I built the console-image with glibc and made a custom image with a few more packages Sep 11 12:41:37 Jin^eLD: it is in OE but not-default for your config Sep 11 12:41:37 but I can't manage to boot on that jffs2 image... Sep 11 12:42:03 nud: did you program the jffs2 image into NAND using mtd-utils? Sep 11 12:42:09 hrw: oh ok, I'll tune my local conf then Sep 11 12:42:13 nud: from the initramfs? Sep 11 12:42:34 or directly from the kernel options Sep 11 12:43:03 I managed to mount it but it's slow and spits a lot of strange messages (I guess it's normal) Sep 11 12:43:51 nud: the first time you mount it, it may be slow, depending on clean-markers options, etc. Sep 11 12:43:54 also, switch_root needs PID=1, which makes it hard to have a script boot_for_real in the initramfs image Sep 11 12:44:19 JFFS2 warning: (1192) jffs2_get_inode_nodes: Eep. No valid nodes for ino #55. Sep 11 12:44:20 JFFS2 warning: (1192) jffs2_do_read_inode_internal: no data nodes found for ino #55 Sep 11 12:44:20 ls: /jffs2/bin/lsmod.26: Input/output error Sep 11 12:44:20 mtd->read(0x580 bytes from 0x18b280) returned ECC error Sep 11 12:44:20 nud: getting all the jffs2 options matched up correctly takes some work Sep 11 12:44:29 this is the kind of messages I get Sep 11 12:44:39 nud: you should not be seeing anything like that Sep 11 12:44:55 each ls spits its bunch of warnings Sep 11 12:45:05 good morning all Sep 11 12:45:37 hi koen Sep 11 12:45:56 nud: I would not mess with switch_root for now, just add "root=mtd3 rootfstype=jffs2" to kernel command line, and disable the ramdisk loading in the BL, and the kernel should boot directly into the jffs2 rootfs Sep 11 12:46:24 nud: once you get all that working, you can create a custom init script in your initramfs that executes the switch_root Sep 11 12:47:22 hrw: i don't think making a bluez-lite is a good idea Sep 11 12:47:35 nud: did you erase the mtd partition before writing your jffs2 image into it? That may be one thing that would cause jffs2 errors Sep 11 12:47:58 nud: you could also try to erase it, and then mount it without writing anything into it -- that should work. Sep 11 12:48:02 howdy all.. Sep 11 12:48:13 anyone here workin with logicpd boards? Sep 11 12:48:35 fer_luck: I have a PXA270 LPD board Sep 11 12:48:46 cbrake: can we have a lil chat about it? Sep 11 12:48:58 * koen|away stabs gprs Sep 11 12:49:04 koen|away: then make bluez-full and bluez-nocups-nogstreamer Sep 11 12:49:47 koen|away: its 500 tasks just to have one gstreamer plugin and support for rare BT printing Sep 11 12:49:56 hrw: no, there should be only _one_ bluez-utils recipe Sep 11 12:50:05 which suxx Sep 11 12:50:16 hrw: the solution to your problem is package staging, not brutalising recipes Sep 11 12:50:50 cbrake: I think I did Sep 11 12:51:03 I will flash it again to be sure Sep 11 12:51:42 XorA: I would prefer bluez-gstreamer.bb but this is upstream part not our Sep 11 12:54:53 cbrake: by erasing you mean "nand flash format", don't you ? Sep 11 12:54:58 in the BL Sep 11 12:56:21 nud: no, use the mtd-utils in the initramfs ... Sep 11 12:56:32 nud: let me boot my system ... Sep 11 12:56:35 ah ok Sep 11 12:56:43 I thought they were doing the same thing Sep 11 12:59:51 HI I get a strange nptl error when compiling glibc 2.5 Sep 11 12:59:53 or 2.6.1 Sep 11 12:59:56 | checking size of long double... (cached) 8 Sep 11 12:59:56 | running configure fragment for ports/sysdeps/arm/elf Sep 11 12:59:56 | checking for ARM TLS support... no Sep 11 12:59:56 | running configure fragment for nptl/sysdeps/pthread Sep 11 12:59:58 | configure: error: compiler support for __thread is required Sep 11 13:00:01 | + oefatal 'oe_runconf failed' Sep 11 13:00:03 | + echo FATAL: 'oe_runconf failed' Sep 11 13:00:06 any1 got a clue? Sep 11 13:00:29 nud: I have no idea what the compulab BL is doing -- that is why I do everything with mtd-utils. Sep 11 13:00:30 nik0n: machine? distro? host machine? host distro? Sep 11 13:01:02 cbrake: "format Sep 11 13:01:14 cbrake: "low level formatting of the nand flash" Sep 11 13:01:26 machine=gumstix (arm) distro=angstrom host=i686 debian Sep 11 13:01:36 nud: the sequence is: boot into initramfs; flash_eraseall /dev/mtdx; wget image.jffs2; nandwrite /dev/mtdx image.jffs2 Sep 11 13:01:56 nud: or you can get your jffs2 image from a USB flash drive Sep 11 13:02:29 nud: skip the compulab BL flash commands as we have no ideas what their nand "format" is Sep 11 13:02:37 cbrake: btw how do you tell OE the size of the image partition you want ? Sep 11 13:02:55 nud: you don't need to with jffs2 as long as you erase the flash partition first Sep 11 13:03:01 ah ok Sep 11 13:03:29 nud: this is nice because your update images only need to be as big as the contents, instead of padded to the parition size like ext2/3 Sep 11 13:03:46 cutecom is not very convenient to interact with a bash term Sep 11 13:11:46 bye Sep 11 13:15:31 koen: btw, any news on blackfin? Sep 11 13:16:36 ~~waiting for the new board to try i t out Sep 11 13:16:48 but were you finally able to build the whole thing? Sep 11 13:16:52 I remember you were cursing about that Sep 11 13:17:34 i got uclibc to build Sep 11 13:24:36 only 98 minutes of boredom left Sep 11 13:24:48 96 even Sep 11 13:25:42 cbrake: root@compulab-pxa270:~$ nandwrite /dev/mtd3 console-image.jffs2 Sep 11 13:25:43 Input file is not page aligned Sep 11 13:25:43 Data was only partially written due to error: Success Sep 11 13:25:50 koen: hmm, what are you waiting for? Sep 11 13:26:28 hrw|gone: yep, switched to the binutils version that you suggested and - the kernel image is only 1.4MB now, seems more reasonable Sep 11 13:26:35 nud: you need to pad the image out to the block size when it is created Sep 11 13:27:03 me is waiting for work to end Sep 11 13:27:18 heh.. Sep 11 13:27:29 doh.. crap.. still the davinci board won't boot Sep 11 13:27:55 wtf.. Uncompressing Linux..... (more dots) .. done, booting the kernel. - and thts it Sep 11 13:28:21 <- using nokia 800 atm Sep 11 13:29:30 cbrake: by adding EXTRA_IMAGECMD_jffs2 = "--pad" in local.conf ? Sep 11 13:29:30 koen: you could finally try some UPnP then ;) Sep 11 13:29:36 heh Sep 11 13:29:38 nud: yes Sep 11 13:29:54 ok Sep 11 13:29:55 unp over gprs Sep 11 13:29:59 thanks **** BEGIN LOGGING AT Tue Sep 11 13:31:46 2007 Sep 11 13:31:57 heh.. no fun indeed then Sep 11 13:33:14 should have brought my laptop Sep 11 13:35:42 why does bitbake rebuild the cache each time local.conf is edited ? Sep 11 13:36:20 because you changed your config Sep 11 13:36:51 nud: because *.conf files change global variables, which can affect any or all bb files, so they must be reparsed. Sep 11 13:41:16 later all Sep 11 13:42:32 * nud promises he will write a "how I (and cbrake) made linux run on my x270" Sep 11 13:59:51 03mickeyl 07org.oe.dev * rf54b6b18... 10/ (7 files in 4 dirs): add yaffs2-utils[-native] Sep 11 13:59:56 03mickeyl 07org.oe.dev * rd8942575... 10/ (1 conf/bitbake.conf): bitbake.conf: enable yaffs2 image generation Sep 11 14:00:00 03mickeyl 07org.oe.dev * r3fcc49d8... 10/ (3 files in 2 dirs): Sep 11 14:00:00 motorola-ezx.inc: remove Sep 11 14:00:00 motorola-ezx-base.inc: reorder some entries Sep 11 14:06:00 morning Sep 11 14:12:19 openezx makes bitbake fail Sep 11 14:12:40 OE looks much dependant on external websites :-/ Sep 11 14:18:09 koen|away: you flying out? Sep 11 14:37:11 XorA: You use a 64 bit machine? Have you noticed problems with 4GB vmlinux files with binutils 2.18? Sep 11 14:37:30 RP: I havent fixed binutils building yet Sep 11 14:37:55 XorA: Poky has a fix from the OE bugzilla which seems to make it build for people Sep 11 14:38:17 RP: yeah, I keep meaning to pull in fix, havent built on my home machine for a while Sep 11 14:38:37 re Sep 11 14:39:45 XorA: sameo@OH is seeing insane zimage sizes with it on x86_64 :( Sep 11 14:40:18 RP: ouch Sep 11 14:40:47 fixes welcome :/ Sep 11 14:40:52 3.1GB image Sep 11 14:40:59 I had it too (on x86-64) Sep 11 14:41:15 XorA, RP, zImage is actually sort of fine (4.5M), besides the fact that it doesn't boot. But arch/arm/boot/Image is 3.1 Gbytes Sep 11 14:41:34 sameo: so there is a whole hell of a lot of zero in it? Sep 11 14:42:01 XorA: I was expecting to see zeros by the end of the image, but it's not so... Sep 11 14:42:04 time for Neo to inclease flash size then, 8gig should be ok :-) Sep 11 14:42:13 XorA: :-) Sep 11 14:42:27 sameo: must be a repeating pattern to comrpess from 3.1G to 4.5M Sep 11 14:42:42 cbrake: how do you know which values you have to pad for the device ? I tried several settings without success Sep 11 14:42:54 nud: let me look ... Sep 11 14:43:15 flash_info doesn't say anything Sep 11 14:43:36 well, it says "/dev/mtd3 has 0 erase regions" Sep 11 14:43:51 XorA: the thing is, vmlinux is fine (3.5 Mb), then Image is insanely big, and it gets compressed to 4.3 M Sep 11 14:45:01 XorA: and Image is just an objcopy from vmlinux Sep 11 14:46:23 nud: I get 0 erase regions as well here, so that must be benign Sep 11 14:47:42 cbrake: yes but nandwrite still complains with EXTRA_IMAGECMD_jffs2="--pad=0x1000000 --eraseblock=0x40000" Sep 11 14:47:45 nud: look in conf/machine/compulab-pxa270.conf -- I've already included EXTRA_IMAGECMD_jffs2 Sep 11 14:47:47 also with --pad alone Sep 11 14:48:47 cbrake: why does it complain then ? Sep 11 14:48:59 nud: what is the jffs2 image size in bytes? Sep 11 14:49:40 7999800 Sep 11 14:50:43 nud: 7999800 % 0x20000 != 0 therefore, something is still wrong with your jffs2 image generation Sep 11 15:06:56 XorA: you around? Sep 11 15:07:17 dcordes: yes Sep 11 15:08:08 XorA: did you speak to the #htc-linux people about a possible htcwizard kernel for OE? Sep 11 15:08:34 dcordes: no, as i have no real time for another device Sep 11 15:10:04 another device among akita? Sep 11 15:10:32 dcordes: I have every Z apart from akita Sep 11 15:10:40 oh ok Sep 11 15:10:50 dcordes: + Neo Sep 11 15:10:53 x2 Sep 11 15:11:05 could you think of other OE ppl possibly interested in htcwizard? Sep 11 15:12:48 dcordes: post on email list Sep 11 15:13:49 hm ok Sep 11 15:15:18 hrm Sep 11 15:15:49 cbrake: do you have an idea on what I could look at ? my manual tweaks just don't change anything Sep 11 15:17:18 XorA: soon you will not have tosa, poodle afaik.. Sep 11 15:17:46 hrw: plan to take them as lrg says he will most likely not use them Sep 11 15:17:55 ok Sep 11 15:18:10 hrw: and at moment he is not even an OE developer :-D Sep 11 15:18:46 nud: run bitbake with -DDD so you can watch the commands Sep 11 15:22:10 nud: beyond that, run bitbake -i, and then "peek angstrom-console-image EXTRA_IMAGECMD_jffs2" Sep 11 15:26:32 cbrake: --little-endian --eraseblock=0x20000 --pad=0x20000 --no-cleanmarkers;sumtool -i /stuff/build/tmp/deploy/glibc/images/compulab-pxa270/Angstrom-console-image-glibc-ipk-2007.9-test-20070911-compulab-pxa270.rootfs.jffs2 -o /stuff/build/tmp/deploy/glibc/images/compulab-pxa270/Angstrom-console-image-glibc-ipk-2007.9-test-20070911-compulab-pxa270.rootfs-summary.jffs2 -e 0x20000 -p; cd /stuff/build/tmp/deploy/glibc/images/compulab-pxa270; rm -f /stuff/build/tmp/dep Sep 11 15:26:32 loy/glibc/images/compulab-pxa270/console-image-compulab-pxa270.rootfs-summary.jffs2; ln -s Angstrom-console-image-glibc-ipk-2007.9-test-20070911-compulab-pxa270.rootfs-summary.jffs2 console-image-compulab-pxa270.rootfs-summary.jffs2 Sep 11 15:26:45 so "looks alright" Sep 11 15:27:28 nud: so, now try bitbake -DDD and watch very carefully what its doing when it gets to mkfs.jffs2 Sep 11 15:27:40 nud: if you need, you can run the command yourself to debug Sep 11 15:30:51 nud, cbrake: it is important in which conf you set EXTRA_IMAGECMD. I do it in the image bb Sep 11 15:35:55 the options are effectively passed to mkfs.jffs2 Sep 11 15:36:49 nud: now try running the command yourself to debug why the padding is not working. You should be able to copy-n-paste from the -DDD output. Sep 11 15:43:00 http://rafb.net/p/l3wnUo52.html Sep 11 15:43:18 looks like the image is compressed Sep 11 15:43:32 should it be padded anyway ? Sep 11 15:44:30 nud: I think so -- let me look at some of my images Sep 11 15:45:44 hi likewise! Sep 11 15:46:43 hm sorry but I have to go, cbrake Sep 11 15:46:48 I'll be back in a couple of hours Sep 11 15:46:58 thanks for your help :-) Sep 11 15:47:42 nud: ok, later Sep 11 15:47:48 later Sep 11 15:49:04 03mickeyl 07org.oe.dev * r48a8d401... 10/ (1 packages/openmoko2/neod_svn.bb): neod svn: speficy platform for FIC Neo1973 and Motorola EZX devices Sep 11 16:04:22 ~lart monotone Sep 11 16:04:22 * ibot shoves a crumpet down monotone's throat, happy now?! Huh? Want some JAM with that? Sep 11 16:04:23 ~lart monotone Sep 11 16:08:29 koen, what is the N800 bootloader like to work with? Sep 11 16:13:34 bye Sep 11 16:23:02 hey everybody. I think I did something bad and not sure how to get it back. Sep 11 16:23:27 I did a bitbake -crebuild angstrom-image-file that was working before...but now, there is no file 'install' for make Sep 11 16:23:29 and it fails Sep 11 16:23:49 make: *** No rule to make target `install'. Stop. Sep 11 16:23:49 FATAL: oe_runmake failed Sep 11 16:24:12 is there something else I need to run after doing that rebuild? it was building and working just fine before! :( Sep 11 16:39:01 koen: I'm going to upgrade linux-hh to -hh16, and consequently, defconfigs. I've merged your h2200 pcmcia-wifi changes to defconfigman, but only that so far. I'm adding merging other stuff (ether, etc) to my queue. If you need something with priority, pls let me know. Sep 11 17:01:10 koen: should preferred-xorg-versions.inc be changed to use "?="? Sep 11 17:01:50 pH5: In my view, yes fwiw Sep 11 17:02:05 currently angstrom-2008.1 includes preferred-xorg-version-X11R7.3.inc (which does what we want with "?=") and then angstrom-2007.1.conf, which includes preferred-xorg-versions.inc, which overwrites all the previously defined PREFERRED_VERSIONs with "=" Sep 11 17:03:24 RP: ok Sep 11 17:35:37 anyone have a howto on building OE from binaries packages and without the source step? Sep 11 18:16:16 mranostay_work: well you just write a .bb recipe without inheriting autotools and simply write your own do_stage / do_install things Sep 11 18:16:22 re Sep 11 18:16:29 wb Sep 11 18:17:01 does anyone know what I can delete to get an image to build that now has the following error after I did a -crebuild on it? Sep 11 18:17:04 ERROR: function do_install failed Sep 11 18:17:04 ERROR: log data follows (/oe/build/tmp/work/sarge-at91-angstrom-linux-gnueabi/angstrom-dci-image-1.0-r8/temp/log.do_install.32624) Sep 11 18:17:04 | NOTE: make DESTDIR=/oe/build/tmp/work/sarge-at91-angstrom-linux-gnueabi/angstrom-dci-image-1.0-r8/image install Sep 11 18:17:04 | make: *** No rule to make target `install'. Stop. Sep 11 18:17:04 | FATAL: oe_runmake failed Sep 11 18:17:45 christopher: bitbake -c clean angstrom-dci-image and try again Sep 11 18:19:13 Darth_Wader: same error Sep 11 18:19:30 maybe then paste /oe/build/tmp/work/sarge-at91-angstrom-linux-gnueabi/angstrom-dci-image-1.0-r8/temp/log.do_install.32624 to pastebin.ca? Sep 11 18:19:34 Darth_Wader: doesn't the -crebuild do a clean, the build? Sep 11 18:19:39 so we can see what actually happens there Sep 11 18:19:41 unsure Sep 11 18:20:05 NOTE: make DESTDIR=/oe/build/tmp/work/sarge-at91-angstrom-linux-gnueabi/angstrom-dci-$make: *** No rule to make target `install'. Stop. Sep 11 18:20:05 FATAL: oe_runmake failed Sep 11 18:20:11 that's all that is in there... Sep 11 18:20:35 guess that first line got truncated... Sep 11 18:20:41 it is doing an image install Sep 11 18:20:42 in that log? Sep 11 18:20:50 yep Sep 11 18:21:03 I was hoping for more there, but that's all it showed Sep 11 18:21:34 Other images still work just fine. Sep 11 18:21:48 so, I am worried about trying a rebuild on them knowing they could fail in the same way Sep 11 18:22:29 i can only think about examining contents of /oe/build/tmp/work/sarge-at91-angstrom-linux-gnueabi/angstrom-dci-image-1.0-r8 Sep 11 18:25:12 in the temp directory, there are less scripts than the others, but I guess those get dynmically built and haven't been built in the sequence yet Sep 11 18:28:15 I don't even see how the other images are built. there is nothing in the image directory for them either Sep 11 20:05:17 is a '.' acceptable in PR? Sep 11 20:09:45 Technically, it works. I think it might be discouraged unless there are specific reasons for doing so... Sep 11 20:15:05 So OpenEmbedded has no set way to build an image from a bunch of ipkgs? Sep 11 20:15:53 OE *does* build an image from ipkgs. Sep 11 20:16:06 It just may take a bit more than simply ipkgs to do it. Sep 11 20:18:14 ah i see Sep 11 20:39:19 03mickeyl 07org.oe.dev * rb6a14ede... 10/ (1 packages/puzzles/oh-puzzles_svn.bb): oh-puzzles: convert to SRCREV Sep 11 20:39:24 03pfalcon 07org.oe.dev * rfaa846d8... 10/ (1 packages/linux/linux-handhelds-2.6/h2200/defconfig): Sep 11 20:39:24 linux-handhelds-2.6: Update defconfig for h2200 from defconfigman r377. Sep 11 20:39:24 * This temporarily undos changes to enable USB and PCMCIA-Ether drivers, Sep 11 20:39:24 they will be re-added later (consistently for all devices which offer such Sep 11 20:39:24 functionality). PCMCIA-WiFi support should be intact. Sep 11 20:39:29 03pfalcon 07org.oe.dev * r62ccc1ee... 10/ (26 files in 26 dirs): linux-handhelds-2.6: Add 2.6.21-hh16, update defconfigs to defconfigman r377. Sep 11 20:39:33 03pfalcon 07org.oe.dev * r41bc7cd1... 10/ (1 conf/distro/angstrom-2007.1.conf): angstrom-2007.1.conf: Bump PREFERRED_VERSION_linux-handhelds-2.6 to "2.6.21-hh16". Sep 11 20:39:38 03koen 07org.oe.dev * r25b0d618... 10/ (1 packages/linux/linux-handhelds-2.6/h2200/defconfig): disapproval of revision 'faa846d8f78b2a5a405633e96ac2d3c62bd9240a' Sep 11 20:39:45 03pH5 07org.oe.dev * rf65084f9... 10/ (1 packages/tasks/task-openmoko.bb): task-openmoko: make magician use qvga theme Sep 11 20:39:49 03pH5 07org.oe.dev * rf2c300ba... 10/ (1 conf/distro/include/preferred-xorg-versions.inc): preferred-xorg-versions.inc: use "?=", not "=" Sep 11 20:39:55 03pH5 07org.oe.dev * r886e55ab... 10/ (1 conf/distro/angstrom-2008.1.conf): angstrom-2008.1.conf: use 2.6.21-hh16 version of handhelds.org kernel Sep 11 21:39:24 Darth_Wader: thanks for the help earlier, it turned out to be an issue in the bb file. I started over from one I knew that worked and the image install worked just fine. guess it was a dependency that wasn't met Sep 11 21:39:46 nice issue :) Sep 11 21:46:27 yeah, not sure what happened there, but whatever... Sep 11 21:46:54 have you by chance every used rhythmbox using oe? Sep 11 21:47:07 I was hoping to be able to use and play files from my ipod Sep 11 21:47:12 and that works on my pc Sep 11 21:47:56 03pH5 07org.oe.dev * rc3f0afcb... 10/ (1 conf/distro/include/preferred-xorg-versions-X11R7.3.inc): preferred-xorg-versions-X11R7.3.inc: fix some versions (drop X11R7.x prefix) Sep 11 21:49:04 configure: error: GStreamer not found, or older than version 0.8.1 Sep 11 21:49:15 I'm getting this even though Gstreamer is 0.10 Sep 11 21:50:09 my angstrom target is opie-image for ipaq pda :) Sep 11 21:53:12 ok. so, no need for an ipod then and no usb interface Sep 11 21:55:48 also no ipod at all :) Sep 11 21:56:36 hehe. do you use a package to play mp3s, m4as ? do you just use mplayer. I used that and it worked, just wondering if there was something else Sep 11 21:56:55 i`m using opie`s native opie-mediaplayer2 Sep 11 21:57:58 unfortunately, never seen something like that for gpe enviropment Sep 11 22:00:15 ok Sep 11 22:00:19 how big is that rootfs? Sep 11 22:00:34 is it around 20 meg or so? Sep 11 22:01:01 that is, I am not sure how much I care if I use opie or gpe. I just want the best environment. Sep 11 22:01:07 yup Sep 11 22:01:46 opie currently lacks IM client afaik, and there`s something unclear with browser Sep 11 22:02:11 Darth_Wader: there's a version of gaim for opie Sep 11 22:02:18 qpe-gaim or such, from memory Sep 11 22:02:20 but my ipaq is just strange kind of mp3 player, so i don`t care Sep 11 22:02:35 NAbyss: yup. it`s as ancient as windows 3.11 Sep 11 22:03:00