**** BEGIN LOGGING AT Wed Jul 30 02:59:56 2008 Jul 30 04:07:28 Hello, all. Jul 30 04:08:42 I try to follow documentation: http://bec-systems.com/oe/html/gettingoe_getting_oe.html Jul 30 04:08:42 When I chtckout mtn --db=OE.mtn co -b org.openembedded.dev, mtn says: Jul 30 04:08:59 [impatt@kb33 oe]$ mtn --db=OE.mtn co -b org.openembedded.devmtn: misuse: database /tmp/oe/OE.mtn is laid out according to an old schemamtn: misuse: try 'mtn db migrate' to upgrademtn: misuse: (this is irreversible; you may want to make a backup copy first) Jul 30 04:09:36 [impatt@kb33 oe]$ mtn --versionmonotone 0.40 (base revision: 5ccc279f9dea0444b47f03dd5291ecc985fcb7f6) Jul 30 04:10:10 How I can fix it ? Jul 30 04:31:13 impatt: try "mtn db migrate" Jul 30 04:32:18 khem: didn't see your message. still around? Jul 30 04:33:23 dcordes: one minute, plz Jul 30 04:35:15 dcordes: [impatt@kb33 oe]$ mtn --db=OE.mtn db migrate Jul 30 04:35:16 Sucess. Thank you very much ! Jul 30 04:36:05 impatt: no problem Jul 30 05:54:05 * * OE Bug 4458 has been created by victorpalacio(AT)adtelecom.es Jul 30 05:54:07 * * Error with oe_libinstall doing populate_staging of libsdl-x11_1.2 Jul 30 05:54:09 * * http://bugs.openembedded.net/show_bug.cgi?id=4458 Jul 30 06:56:34 03  07org.openembedded.dev * r4577d28bef 10OE.dev/packages/mythtv/ (mythtv-0.21/no-cortex-deadlock.patch mythtv_0.21.bb): mythtv 0.21: update NEON patch to fix cpu deadlocks Jul 30 06:56:40 03  07org.openembedded.dev * r6dc464b253 10OE.dev/packages/ffmpeg/ (ffmpeg_git.bb omapfbplay_git.bb): Jul 30 06:56:40 ffmpeg git: bump SRCREV for cortex deadlock fixes Jul 30 06:56:40 omapfbplay: bump PR Jul 30 06:56:40 03  07org.openembedded.dev * r7a832c7c84 10OE.dev/packages/mozilla/ (14 files in 2 dirs): firefox: move to 3.0.1 Jul 30 06:56:42 03  07master * r7a832c7c84 10OE.dev/packages/mozilla/ (14 files in 2 dirs): firefox: move to 3.0.1 Jul 30 06:56:45 03  07master * r6dc464b253 10OE.dev/packages/ffmpeg/ (ffmpeg_git.bb omapfbplay_git.bb): Jul 30 06:56:48 ffmpeg git: bump SRCREV for cortex deadlock fixes Jul 30 06:56:50 omapfbplay: bump PR Jul 30 06:56:52 03  07master * r4577d28bef 10OE.dev/packages/mythtv/ (mythtv-0.21/no-cortex-deadlock.patch mythtv_0.21.bb): mythtv 0.21: update NEON patch to fix cpu deadlocks Jul 30 07:00:11 hi. bitbaking thttpd (oe/stable/angstrom) and the build fails with "ERROR: QA Issue: non debug package contains .debug directory: thttpd path /work/armv5te-angstrom-linux-gnueabi/thttpd-2.25b-r6/install/thttpd/srv/www/cgi-bin/.debug/ssi" and then later "ERROR: Build of /usr/src/oe/org.openembedded.stable/packages/thttpd/thttpd_2.25b.bb do_package failed". any hints where I should start looking for the issue or how to fix it? Jul 30 07:18:40 * josch|nsn is scared by any file that took his computer 3 weeks to compile Jul 30 07:19:30 uh oh... this happens when you think that things from the backlog just happend instead of a day ago :P Jul 30 07:27:00 03  07org.openembedded.dev * re6cfd6689e 10OE.dev/packages/mozilla/firefox_3.0.1.bb: firefox 3.0.1: remove js hack Jul 30 07:27:06 03  07master * re6cfd6689e 10OE.dev/packages/mozilla/firefox_3.0.1.bb: firefox 3.0.1: remove js hack Jul 30 07:43:34 josch|nsn: doh Jul 30 07:46:07 bonjour Jul 30 07:46:16 salut Genesis :P Jul 30 07:46:54 i created a new machine conf called htcnike.conf which is basicly a copy of the htckaiser.conf .. this conf sets PREFERRED_PROVIDER_virtual/kernel = "linux-kaiser" my build now trys to build linux_2.6.21 vanilla ? Jul 30 07:47:00 guten morgen josch|nsn :) Jul 30 07:47:28 why would a PREFERRED_PROVIDER be overwritten or ignored ? Jul 30 07:48:12 preferred not find , set elsewhere Jul 30 07:49:03 ah i think i got it .. the kernel doesnt include htcnike as compatible .. thanks anyway Jul 30 07:56:26 rob_w: are you working with dcordes? Jul 30 07:56:38 rob_w: he is adding htcpolaris support as well Jul 30 07:57:01 i will touch base with him. .. i just revived my htc genes and see how far i can go with my nike Jul 30 08:15:54 rob_w: That kernel recipe probably sets COMPATIBLE_MACHINE Jul 30 08:29:07 git is so slow :-( Jul 30 08:29:22 morning Jul 30 08:30:14 rob_w: nike is tuoch or touchdual? Jul 30 08:51:33 yummies, acetone Jul 30 09:00:06 XorA: tasty Jul 30 09:00:55 * josch|nsn had an acetone muffin for breakfast Jul 30 09:09:52 acetone liquid dont tast so nice Jul 30 09:10:39 morning Jul 30 09:11:57 03  07org.openembedded.dev * re5c8c47109 10OE.dev/packages/linux/ (3 files in 2 dirs): linux omap2 git: add patch to workaround cortex-a8 deadlock Jul 30 09:12:32 good monring Jul 30 09:12:55 03  07master * re5c8c47109 10OE.dev/packages/linux/ (3 files in 2 dirs): linux omap2 git: add patch to workaround cortex-a8 deadlock Jul 30 09:20:04 ~curse SCSI/SATA subsystem in linux Jul 30 09:20:05 May the fleas of a thousand camels infest your most sensitive regions, SCSI/SATA subsystem in linux ! Jul 30 09:34:59 florian: good monring Jul 30 09:47:33 morning folks Jul 30 09:47:48 morning Jul 30 09:48:10 y0 mickeyl Jul 30 09:48:10 hi mickeyl, hi methril Jul 30 09:49:54 mickeyl: did you put my sphinx and python-sphinxsearch modules into oe yet? Jul 30 09:51:01 not yet, no Jul 30 09:51:08 where can i get these? Jul 30 09:51:41 mickeyl: Marcel_M send you the bb recipes Jul 30 09:51:56 mickeyl: you helped us with creating them, remember? :D Jul 30 09:52:43 not really Jul 30 09:52:48 but my memory is limited Jul 30 09:52:52 uh ^^ Jul 30 09:52:54 too many context switches Jul 30 09:53:00 need to get rid of some things Jul 30 09:53:11 quick someone get mickeyl coffee before he crashes Jul 30 09:53:11 anyways, please have him email me the recipes Jul 30 09:53:16 XorA: ;) Jul 30 09:53:24 picked up zecke at the airport way too early Jul 30 09:53:29 its a bit early in the morning for him Jul 30 09:53:34 ah, that explains it Jul 30 09:53:34 * josch|nsn runs with a cup of coffee Jul 30 09:53:40 * josch|nsn slips Jul 30 09:53:53 * josch|nsn drains mickeyl 's trousers with hot coffe Jul 30 09:54:34 now i is awake i assume :P Jul 30 09:54:42 ah... *he Jul 30 09:55:06 mickeyl: i will tell Marcel_M to do so, thanks Jul 30 09:58:09 mickeyl: Say hi to zecke for me Jul 30 09:58:31 zecke must be in the dungeon again Jul 30 09:58:44 ah hey Laibsch Jul 30 09:58:51 Laibsch: greetings back Jul 30 09:59:05 thanks Jul 30 10:15:17 mickeyl: good morning Jul 30 10:18:30 hmm. any idea how to get rrdtool to build with OE? getting complaints about "Your compiler does not support propper IEEE math" (target has no hw FPU) Jul 30 10:26:54 czr_: I guess you'd have to check the logs to find out why it thinks that IEEE math is missing Jul 30 10:27:12 assuming you feel that your target should indeed have ieee math. what platform are you using? Jul 30 10:29:06 xscale atm. Jul 30 10:29:15 that should be fine Jul 30 10:29:44 indeed, but I think the rrdtool configure runs some tests and based on that none of the -ieee-math-and such switches work with the cross compiler bails out Jul 30 10:30:33 which are the particular switches that it's having a problem with, and what is the compiler output in that case? Jul 30 10:30:51 give me a sec, going to do a pastebin Jul 30 10:31:36 http://www.pastebin.org/58466 Jul 30 10:31:58 building cairo worked without a hitch Jul 30 10:32:22 # Jul 30 10:32:22 | Your Compiler does not do propper IEEE math ... Please find out how to Jul 30 10:32:22 # Jul 30 10:32:22 | make IEEE math work with your compiler and let me know (oetiker@ee.ethz.ch). Jul 30 10:32:22 # Jul 30 10:32:25 | Check config.log to see what went wrong ... Jul 30 10:32:27 did you check config.log? Jul 30 10:33:53 it doesn't contain anything interesting to my eye Jul 30 10:34:06 lines like this (no extra info): Jul 30 10:34:10 configure:22961: checking if IEEE math works out of the box Jul 30 10:34:10 configure:23081: result: no Jul 30 10:35:01 ah right, nothing else between those lines? Jul 30 10:35:07 nope Jul 30 10:35:10 in that case you'll have to inspect the configure script itself to see what it's doing Jul 30 10:35:27 it's just a shell script, albeit a large and ugly one Jul 30 10:35:39 * czr_ nods Jul 30 10:35:52 if you start reading from line 22961 you should be able to find the particular gcc invocation that's causing the problem. then, try running that by hand and see what happens. Jul 30 10:36:06 I've done this before :-) Jul 30 10:36:16 just new to OE, or rather, everything else has built fine so far :-) Jul 30 10:36:28 (minus couple of small fixes that I've "backported" from .dev to .stable) Jul 30 10:37:24 heh, right Jul 30 10:38:02 ah. it's a small program that causes different kinds of IEEE exceptions Jul 30 10:47:01 ah, right. Jul 30 10:47:08 that's obviously not going to work when cross compiling Jul 30 10:47:26 probably best to just hack the configure script to assume the answer is yes in that case Jul 30 10:47:30 :-) Jul 30 10:47:50 hmm. where should I hack it? Jul 30 10:48:02 look at configure.ac (or configure.in), find the corresponding line there Jul 30 10:48:03 if I change it under work/rrdtool then I can't use bitbake any longer, can I? Jul 30 10:48:09 it'll probably be some kind of AC_TRY_RUN command Jul 30 10:48:26 you should be able to add an extra argument to AC_TRY_RUN, being the action to take when cross-compiling Jul 30 10:48:27 that's not the problem really Jul 30 10:48:52 oh, you mean how do you integrate the change with oe? Jul 30 10:48:55 yes Jul 30 10:49:05 make a diff between the old and new configure.ac files and add it as a patch to the .bb file Jul 30 10:49:15 ah! Jul 30 10:49:20 thanks, will try it Jul 30 10:49:45 it looks like rrdtool already has one patch listed in its SRC_URI so you can use that as an example Jul 30 10:49:53 indeed Jul 30 10:50:11 I'm trying to bump the version higher as well, while at it. from 1.0.49 to 1.2.28. nothing can fail. right? ;-) Jul 30 10:50:20 heh, right. what could possibly go wrong? Jul 30 10:50:24 indeed! Jul 30 10:52:08 it uses AC_FULL_IEEE macro Jul 30 10:52:14 which comes with autoconf, not rrdtool Jul 30 10:52:15 * czr_ sighs Jul 30 10:52:35 oh right, drat Jul 30 10:52:38 in that case, you have two options: Jul 30 10:52:54 (a) apply your hack to autoconf, make a new version of that, and then make rrdtool depend on a new enough autoconf-native Jul 30 10:53:06 or just modify the configure script directly and patch that, right? Jul 30 10:53:39 (b) copy the AC_FULL_IEEE macro definition into rrdtool's aclocal.m4, rename it to (say) AC_FULL_IEEE_NOT_BROKEN, make the change there, then change configure.ac to reference that macro not the broken version Jul 30 10:54:35 patching the configure script directly is probably not a good idea because it will involve disabling a lot of the autotools.bbclass stuff. you can make it work that way but it tends to be a bit fragile and difficult to maintain in the future. Jul 30 10:57:24 ah Jul 30 10:57:31 so better to hack .ac then Jul 30 10:58:28 yeah Jul 30 10:58:42 autotools.bbclass will automatically regenerate configure from configure.ac when you run a build Jul 30 10:58:53 ah. good to know Jul 30 10:59:14 diff -urN rrdtool-orig rrdtool-patched > patch ok? Jul 30 10:59:22 or should it be in some more specific format? Jul 30 11:00:56 that should be fine Jul 30 11:01:20 bitbake doesn't try to interpret the patches, so anything that "patch" can read ought to be ok Jul 30 11:01:33 right. let's see then :-) Jul 30 11:03:54 hmm. how do I apply "strip" to the patching command? Jul 30 11:04:05 since now I have one extra level of dirs in the patch (bitbake breakes in do_patch) Jul 30 11:09:32 hmm. I guess 1.2.28 is slightly different from 1.0.48 ;-) Jul 30 11:09:58 add "pnum=N" to the SRC_URI entry Jul 30 11:10:04 where N is the number of levels you want to strip Jul 30 11:10:26 ah, thanks. ended up editing the patch manually to strip the levels and did the renames ;-) Jul 30 11:10:32 heh, fair enough Jul 30 11:12:39 argh. 1.2 uses libart. Jul 30 11:13:01 (1.3 moves to cairo & fontconfig. and Pango for text layout). Jul 30 11:13:10 * czr_ decides to stick with 1.0.43 Jul 30 11:20:06 pb_, is the patch=1 a ordering variable or it means that "this uri is a patch" Jul 30 11:25:28 it just means "this uri is a patch" Jul 30 11:25:39 the ordering in SRC_URI determines the order in which the patches are applied Jul 30 11:26:23 right, figured that out too Jul 30 11:26:24 :-) Jul 30 11:26:46 in 1.0.49 the IEEE test was inline coded (the whole C program was there). So, I removed the test completely, and now rrdtool builds Jul 30 11:26:53 not technically very elegant solution, but.. Jul 30 11:26:58 (from configure.ac) Jul 30 11:28:49 ok, very good Jul 30 11:29:00 hmm. actually it doesn't work.. Jul 30 11:29:12 now my patch isn't even applied anymore.. /me scratches head Jul 30 11:36:29 hmm. how do I reset a package? seems that oe now cached the results of some previous run somewhere and even bitbake -c clean rrdtool doesn't get it updated (I updated the .bb file to include my patch but it isn't taken from the .bb file) Jul 30 11:37:40 * czr_ hits himself with a yardstick and goes to the corner to cry in shame Jul 30 11:40:53 that's odd, bitbake -c clean should be fine Jul 30 11:40:58 you're sure you're using the right .bb file? Jul 30 11:41:18 yeah. I have an "backup" of the original rrdtool packages dir (rrdtool.original). ;-) Jul 30 11:41:40 so the same version of rrdtool was defined in two diff packages-dirs. Jul 30 11:41:55 ah Jul 30 11:42:31 hmm. weird. now it's trying to link rrdtool stuff against my local x86-64 version of libz Jul 30 11:42:43 "unrecognized file format". I wonder why... Jul 30 11:46:18 can anyone test building rrdtool in stable? Jul 30 11:46:26 (against a non-x86-target) Jul 30 11:55:25 ah. it's building some perl stuff and the perl module uses local perl stuff, although the build uses cross compiler. Jul 30 11:55:36 and there's no easy way to disable perl module building in 1.0.49 Jul 30 12:06:48 03  07org.openembedded.dev * r784d60ea32 10OE.dev/: Jul 30 12:06:48 merge of '87e3cc07eb2fd42d7850e6b0c1f67268404695d0' Jul 30 12:06:48 and 'f572f61df65c44a6ea6b9b987560f1edd3171a1b' Jul 30 12:06:53 03  07master * r5f2ac7ab23 10OE.dev/conf/distro/include/sane-srcrevs.inc: update openmoko packages for fso milestone2 Jul 30 12:06:54 03  07master * r784d60ea32 10OE.dev/: Jul 30 12:06:54 merge of '87e3cc07eb2fd42d7850e6b0c1f67268404695d0' Jul 30 12:06:54 and 'f572f61df65c44a6ea6b9b987560f1edd3171a1b' Jul 30 12:06:56 03  07org.openembedded.dev * r5f2ac7ab23 10OE.dev/conf/distro/include/sane-srcrevs.inc: update openmoko packages for fso milestone2 Jul 30 12:25:04 pb_, backported a --disable-perl support from 1.2 series back to 1.0.49 and now I get it to build properly (with the IEEE-check-disable too) Jul 30 12:43:25 03  07org.openembedded.dev * r78f1d79f91 10OE.dev/packages/linux/ (linux-omap2-git/beagleboard/defconfig linux-omap2_git.bb): linux omap2 git: turn of MUSB dma to avoid crashes Jul 30 12:43:32 03  07master * r78f1d79f91 10OE.dev/packages/linux/ (linux-omap2-git/beagleboard/defconfig linux-omap2_git.bb): linux omap2 git: turn of MUSB dma to avoid crashes Jul 30 12:53:09 czr_: very good! Jul 30 12:59:26 am I the only one thats gets all GIT commit messages twice? Jul 30 13:02:11 XorA: You mean the mails? One form org.oe.dev branch the othe one from master. Jul 30 13:02:33 XorA: A rule tofilter duplicates on branches out would be good Jul 30 13:02:46 Hello Jul 30 13:03:09 I have just uploaded a new feed for JLime, and I'm trying it out. Jul 30 13:03:28 Running opkg update with the new feed complains about a missing Packages.sig Jul 30 13:03:46 And I have not found how to generate that on Google Jul 30 13:03:53 Any help, please? Jul 30 13:04:44 XorA: Something similar for the CIA script: http://joey.kitenet.net/blog/entry/lazyweb:_git_cia_hooks/ Jul 30 13:04:55 * stefan_schmidt -> food Jul 30 13:06:20 stefan_schmidt: possibly, I didnt look at them closely enough to realise we were using two branches Jul 30 13:57:48 03  07org.openembedded.dev * r54dcc89306 10OE.dev/conf/machine/davinci-dvevm.conf: davinci-dvevm.conf : Build u-boot 1.3.4-rc1 for testing purposes. Jul 30 13:57:49 03  07org.openembedded.dev * rb27e088f3e 10OE.dev/packages/dsplink/ (4 files in 2 dirs): Jul 30 13:57:49 dsplink : Commit work in progress. Jul 30 13:57:49 * Add support for SFFSDR board with 128M of ram Jul 30 13:57:49 * Build more DSP/ARM transfer methods Jul 30 13:57:53 * Install dsp samples in /usr/share/dsplink Jul 30 13:57:55 03  07org.openembedded.dev * rf506d3b065 10OE.dev/packages/linux/linux-davinci_2.6.x+git-davinci.bb: linux-davinci : use latest git rev. Jul 30 13:57:58 03  07org.openembedded.dev * r8858c6b990 10OE.dev/: Jul 30 13:58:00 merge of '648476e08584946c4059307fb756f1cd7f7fa603' Jul 30 13:58:02 and 'eb28dc45b262f9456efc6bd50fbcf80645179f8f' Jul 30 13:58:04 03  07master * r54dcc89306 10OE.dev/conf/machine/davinci-dvevm.conf: davinci-dvevm.conf : Build u-boot 1.3.4-rc1 for testing purposes. Jul 30 13:58:09 03  07master * rb27e088f3e 10OE.dev/packages/dsplink/ (4 files in 2 dirs): Jul 30 13:58:11 dsplink : Commit work in progress. Jul 30 13:58:13 * Add support for SFFSDR board with 128M of ram Jul 30 13:58:15 * Build more DSP/ARM transfer methods Jul 30 13:58:17 * Install dsp samples in /usr/share/dsplink Jul 30 13:58:19 03  07master * r8858c6b990 10OE.dev/: Jul 30 13:58:21 merge of '648476e08584946c4059307fb756f1cd7f7fa603' Jul 30 13:58:25 and 'eb28dc45b262f9456efc6bd50fbcf80645179f8f' Jul 30 13:58:27 03  07master * rf506d3b065 10OE.dev/packages/linux/linux-davinci_2.6.x+git-davinci.bb: linux-davinci : use latest git rev. Jul 30 14:08:44 what packages from OE do I need to be able to mount NFS shares? Jul 30 14:09:19 it works out of the box with the openmoko distribution but not with angstrom on the beagleboard Jul 30 14:09:45 i've seen phillip had the same problem lately: http://de.pastebin.ca/1078855 Jul 30 14:13:42 do you mean nfs4 specifically, or all nfs? Jul 30 14:18:38 pb_: all nfs Jul 30 14:18:52 that's probably a different bug to the one in the pastebin, then. Jul 30 14:18:57 that looks like an nfs4 specific failure Jul 30 14:19:06 what exactly is the output that you get? Jul 30 14:19:17 (you verified that nfs support is in your kernel, right?) Jul 30 14:22:11 pb__: http://de.pastebin.ca/1087200 Jul 30 14:22:27 pb__: the kernel has nfs support compiled in Jul 30 14:22:47 what does strace show in your case? Jul 30 14:23:11 CONFIG_NETWORK_FILESYSTEMS=y Jul 30 14:23:11 CONFIG_NFS_FS=y Jul 30 14:23:11 CONFIG_NFS_V3=y Jul 30 14:23:24 and do you have mount.nfs installed? Jul 30 14:23:31 nfs-utils and portmap installed? Jul 30 14:24:13 strace is here: http://de.pastebin.ca/1087208 Jul 30 14:24:30 hrw: do I need nfs-utils and portmap on the client? Jul 30 14:24:48 yeah, you lack mount.nfs Jul 30 14:25:13 pb__: yeah, but there is no mount.nfs in openmoko's busybox either Jul 30 14:25:46 busybox does have the option to handle nfs mounting internally Jul 30 14:25:52 CONFIG_FEATURE_MOUNT_NFS or something Jul 30 14:25:59 I would guess it's on in the openmoko case and off in yours Jul 30 14:26:53 from busybox defconfig: CONFIG_FEATURE_MOUNT_NFS=y Jul 30 14:27:20 and that's definitely the config file you're using? Jul 30 14:27:48 (and your /sbin/mount really is busybox, right, not the one from util-linux?) Jul 30 14:28:02 or /bin/mount even Jul 30 14:28:44 pb__: good point. i'll check. Jul 30 14:29:49 pb__: I think thats it: Jul 30 14:29:50 root@beagleboard:/home# mount -V Jul 30 14:29:50 mount from util-linux-ng 2.14 (with libblkid support) Jul 30 14:29:58 ah, that would be a problem Jul 30 14:30:15 you either need to switch to busybox mount, or install the mount.nfs helper. Jul 30 14:30:20 I think the latter is in nfs-utils but I'm not certain Jul 30 14:32:01 ah, no, it's in util-linux as well Jul 30 14:36:48 pb__: thanks for your help. I think I should file this as a bug against angstrom. Jul 30 14:36:58 pb__: it works now :) Jul 30 14:36:59 root@beagleboard:/home/rob/oe# mount | grep nfs Jul 30 14:36:59 192.168.0.200:/home/rob on /home/rob type nfs (rw,vers=3,rsize=32768,wsize=3276) Jul 30 14:37:18 very good Jul 30 14:37:35 I can't think why angstrom would be wanting to use util-linux mount offhand but I guess they must have their reasons Jul 30 14:37:50 pb__: old busybox had broken mount Jul 30 14:37:59 pb__: I guess no-one ever retested Jul 30 14:39:59 the bitbake user manual explains variable overrides fairly well (http://bitbake.berlios.de/manual/ch02.html#id868981) however I'm not sure I understand the difference between something like 'SRC_URI +=' 'SRC_URI_append =' - is there a difference? Jul 30 14:41:44 tharvey: simple Jul 30 14:42:20 tharvey: FOO = "baz" and later FOO += "bar" creates "baz bar" Jul 30 14:42:34 it automatically adds a space Jul 30 14:42:50 _append does not do that Jul 30 14:43:49 tharvey: it is often used for the DEPENDS variable where the items are space separated (in SRC_URI as well, but its not so obvious) Jul 30 14:43:55 that's not actually the most significant difference between the two. Jul 30 14:44:00 oh... ok then whats the diff between 'FOO .= "baz"' and 'FOO_append = "baz"'? Jul 30 14:44:04 if you were just interested in spaces, you could use ".=" which doesn't add a space either Jul 30 14:44:24 the important thing is that += is an immediate append, whereas "_append" is a delayed append that happens when the overrides are expanded. Jul 30 14:44:45 in particular, that means you can use _append to do conditional appending, which you can't do with += or .= Jul 30 14:44:46 oh ok. :) Jul 30 14:44:53 ok thx for clearning that up Jul 30 14:44:56 bitbake is so mighty! ;) Jul 30 14:48:12 I'm still trying to get a handle on how to create an image that does not use busybox - it was suggested that I need to user coreutils instead and perhaps do something with INITSCRIPTS - can anyone provide more details? Jul 30 14:50:55 how to inhebit automatic autotools/whatever in a recipe when the project has some Makefile ? Jul 30 14:53:05 * * OE Bug 4459 has been created by thebohemian(AT)gmx.net Jul 30 14:53:07 * * mounting nfs shares does not work without hassles Jul 30 14:53:09 * * http://bugs.openembedded.net/show_bug.cgi?id=4459 Jul 30 14:53:50 Genesis: if there is no 'inherit autotools' somewhere then it does not care about autotools files Jul 30 14:54:26 Genesis: you could also override the functions (have a look at autotools.bbclass to find out which that are) Jul 30 14:58:09 03  07org.openembedded.dev * r0be6859d88 10OE.dev/packages/classpath/classpath-minimal_0.97.2.bb: classpath-minimal: add javanet-local.patch for fixing build Jul 30 14:58:18 03  07master * r0be6859d88 10OE.dev/packages/classpath/classpath-minimal_0.97.2.bb: classpath-minimal: add javanet-local.patch for fixing build Jul 30 14:59:26 rschuster : there is not but i should override i think Jul 30 15:02:25 tharvey: it rather depends on what you're trying to achieve with your image. You'll need to establish what set of programs you need and find non-busybox replacements for each of them. Jul 30 15:02:56 for example, you will need some kind of shell, probably some util-linux and coreutils equivalent, maybe the networking tools, ... Jul 30 15:04:23 bye Jul 30 15:05:10 pb_, so simply including the desired packages will 'override' the busybox alternatives? Jul 30 15:07:05 * * OE Bug 4460 has been created by robert.piasek(AT)member.fsf.org Jul 30 15:07:07 * * zhone compilation fails Jul 30 15:07:09 * * http://bugs.openembedded.net/show_bug.cgi?id=4460 Jul 30 15:08:00 RP, git is not the perfect SCM? Jul 30 15:08:01 :) Jul 30 15:08:21 Crofton: I have never ever claimed that Jul 30 15:08:36 I know Jul 30 15:08:43 I am poking fun at you Jul 30 15:09:04 Crofton: Just be warned that your stick may catch fire ;-) Jul 30 15:09:10 tharvey: if you mean you want to mix and match, no, I don't think busybox is currently set up right for that Jul 30 15:09:13 * RP is not in the best of mindsets today :/ Jul 30 15:09:16 pb__, busybox mount does not support some options used in fstabs used in the image Jul 30 15:09:30 asbestos stick :) Jul 30 15:10:45 tharvey: but, if you don't want busybox at all, you can just remove it from your image and install the replacements instead. Jul 30 15:10:54 Crofton: as we are now 4 versions later in busybox, are we still sure it doesnt? Jul 30 15:11:32 Crofton: Now soaked in petrol ;-) Jul 30 15:11:46 * RP plays with matches Jul 30 15:11:47 what are the options you need? if they're just flags then it should be trivial to add support for them to busybox. Jul 30 15:12:01 pb_, currently I don't need to mix and match - is there anything else I should have to change like initscripts? Jul 30 15:12:24 watch out the biker guy got violent again :-D Jul 30 15:12:31 tharvey: not that I can think of. on the whole the busybox tools are pretty much upwards-compatible with their "big" equivalents. Jul 30 15:13:02 XorA, no, but we are too lazy to check Jul 30 15:13:06 pb_, yes, thats been my experience as well - thanks I'll try what your suggesting Jul 30 15:13:15 the only real breakage swapping busbox to big versions seems to be ps which takes different args Jul 30 15:13:27 which really upsets some initscripts Jul 30 15:13:45 that's just a sysv vs bsd defaulting thing though, right? Jul 30 15:13:58 pb__: yes Jul 30 15:13:59 if so then I guess the scripts should be fixed to dtrt. Jul 30 15:14:10 I always install procps, but I am not doing hand held device work Jul 30 15:14:12 pb__: they are bad initscripts Jul 30 15:14:47 right Jul 30 15:15:27 * XorA needs to work up the excitement to clean that up for openmoko Jul 30 15:15:58 what about introduice a provider for httpd ? Jul 30 15:16:40 hi likewise Jul 30 15:16:50 HopsNBarley: hey howdy Jul 30 15:16:57 hey khem ! Jul 30 15:17:07 HopsNBarley: how are you ? Jul 30 15:17:10 khem, we should get lunch again sometime (-; Jul 30 15:17:20 HopsNBarley: sure any time :) Jul 30 15:17:44 khem, what the heck - how's today or tomorrow? Jul 30 15:18:04 today will be good Jul 30 15:18:12 my wife is not cooking Jul 30 15:18:46 rivermark again. you are at tasman? Jul 30 15:19:05 yes rivermark is good Jul 30 15:19:22 say at 11:30ish Jul 30 15:20:09 I am on Maccarthy Jul 30 15:20:15 sounds good! where to meet? japanese place? Jul 30 15:20:22 i'm at 1st and component. Jul 30 15:20:35 Yoyo is ok Jul 30 15:20:43 any preference? Jul 30 15:21:02 well, we can meet there and decide :-) Jul 30 15:21:08 yeah sure Jul 30 15:21:21 see you then, i better run to work. Jul 30 15:21:25 ok see you Jul 30 15:21:52 what's difference between PROVIDER & PROVIDER_virtual ? Jul 30 15:23:36 03  07master * r292bc88e55 10OE.dev/: Jul 30 15:23:36 merge of '4d6ee85cc3ba867945ab4763b494533e79d9e0ac' Jul 30 15:23:36 and '750fe5c3f11c985f4ea63c4d0e6a98d0326f40b7' Jul 30 15:23:38 03  07master * red4e563365 10OE.dev/conf/distro/include/sane-srcrevs.inc: opkg*: bump SRCREV to 4564 to fix memory leaks and segfaults Jul 30 15:23:39 03  07org.openembedded.dev * rb572f0f53d 10OE.dev/packages/dsplink/ (3 files in 2 dirs): Jul 30 15:23:41 dsplink: use 128mb patch for beagle as well Jul 30 15:23:43 codec-engine: more tweaks Jul 30 15:23:45 03  07org.openembedded.dev * r292bc88e55 10OE.dev/: Jul 30 15:23:47 merge of '4d6ee85cc3ba867945ab4763b494533e79d9e0ac' Jul 30 15:23:49 and '750fe5c3f11c985f4ea63c4d0e6a98d0326f40b7' Jul 30 15:23:51 03  07org.openembedded.dev * red4e563365 10OE.dev/conf/distro/include/sane-srcrevs.inc: opkg*: bump SRCREV to 4564 to fix memory leaks and segfaults Jul 30 15:23:54 03  07master * rb572f0f53d 10OE.dev/packages/dsplink/ (3 files in 2 dirs): Jul 30 15:23:56 dsplink: use 128mb patch for beagle as well Jul 30 15:23:58 codec-engine: more tweaks Jul 30 15:36:04 pb__: on EABI systems libgcc is NEEDED by /bin/sh or /bin/busybox to satisfy __aeabi_unwind_cpp_pr0 Jul 30 15:36:34 and that pulls in libgcc into image Jul 30 15:36:43 khem: ah, right Jul 30 15:36:44 but same does not happen for oabu Jul 30 15:36:48 oabi Jul 30 15:36:54 yeah, that symbol doesn't exist for oabi Jul 30 15:36:57 as you might guess from the name :-) Jul 30 15:37:40 yeah thats the problem thesing is facing and the reason for creating the RRECOMMEND Jul 30 15:38:10 but I grepped through the packages and saw that dbus uses pthread_cancel Jul 30 15:38:18 I'm not entirely sure that the binaries ought to be getting that NEEDED even in the eabi case, actually. Jul 30 15:38:46 my memory of how the exception handling personality stuff works is pretty rusty, though. I would have to go check the documentation again. Jul 30 15:39:13 ok, so in his case dbus should be getting a dependency on libgcc. Jul 30 15:39:50 pb__: pb__ EHABI requires it Jul 30 15:39:51 (it should be safe to make it a hard dependency not a recommend, and I think that's probably a better idea in this case.) Jul 30 15:40:08 khem: can you remind me why? Jul 30 15:40:26 I thought that was meant to be an object-file-level only thing. Jul 30 15:41:23 well not always only if necessary Jul 30 15:41:48 what does "necessary" mean in that context? Jul 30 15:43:59 if unwinding is needed by the program Jul 30 15:44:11 okay, so why is busybox doing it? Jul 30 15:44:14 busybox doesn't use unwinding Jul 30 15:44:30 neither does bash as far as I know Jul 30 15:45:33 yeah it could be other libgcc functions like udivdi3 called which inturn then calls the personality routines Jul 30 15:45:50 but I did not find other functions in busybox so .. Jul 30 15:46:05 * * OE Bug 4460 has been RESOLVED (FIXED) by Jul 30 15:46:07 * * zhone compilation fails Jul 30 15:46:09 * * http://bugs.openembedded.net/show_bug.cgi?id=4460 Jul 30 15:46:24 udivdi3 and so on don't do any unwinding either Jul 30 15:46:57 pb__: it might be then the toolchain produced the references wrongly ? Jul 30 15:46:59 I think it is probably a bug that the linker is outputting that NEEDED for busybox. but, anyway, that is a side issue and I don't think it has any bearing on the oabi case. Jul 30 15:47:11 khem: yes, I think so Jul 30 15:47:38 pb__: for OABI I think if we create DEPENDS for dbus that should work Jul 30 15:47:54 right Jul 30 15:48:04 this is what I have been saying for three days. I am glad we agree. :-) Jul 30 15:48:18 heh Jul 30 15:49:26 now I am able to reproduce the exact problem that thesing saw here so I might get better feel of the issue Jul 30 15:53:57 ok, great Jul 30 15:56:51 hmm, i should pick up an embedded board to play with again. those beagles look pretty nifty Jul 30 15:57:15 morning kergoth Jul 30 15:57:18 yeap Jul 30 15:57:36 morning Jul 30 15:57:39 * kergoth`work yawns Jul 30 15:57:48 but the chip is not cheap when you want to build your own board Jul 30 15:58:01 which chip is that? Jul 30 15:58:15 cortex-a8 Jul 30 15:58:19 if i remember Jul 30 15:58:29 50$ the chip Jul 30 15:58:33 And still not that easy to get... Jul 30 15:59:07 * khem office Jul 30 16:05:52 ah right Jul 30 16:19:30 the beagle is rather nice. My video lead for it arrived this morning so perhaps I can play with it properly now :) Jul 30 16:20:39 rp hehe Jul 30 16:20:44 dev board are nices ; but design such card you'll not sell it 150$ Jul 30 16:22:20 well, no, but if you were buying the thing in quantities you wouldn't be buying it from digi-key either Jul 30 16:30:36 03  07org.openembedded.dev * r295a5a4abf 10OE.dev/ (3 files in 2 dirs): add numptyphysics (ugly ugly build), an insanely cool crayon game Jul 30 16:30:37 03  07org.openembedded.dev * r03ec8bc712 10OE.dev/packages/images/fso-image.bb: add numptyphysics to fso image Jul 30 16:30:38 03  07org.openembedded.dev * rd011215b17 10OE.dev/: Jul 30 16:30:38 merge of '0521752c751b5734b320120045655229b2faff4f' Jul 30 16:30:39 and 'f75d239d7378665c50aea95b24abbfc6a815bd43' Jul 30 16:30:43 03  07master * r03ec8bc712 10OE.dev/packages/images/fso-image.bb: add numptyphysics to fso image Jul 30 16:30:44 03  07master * r295a5a4abf 10OE.dev/ (3 files in 2 dirs): add numptyphysics (ugly ugly build), an insanely cool crayon game Jul 30 16:30:47 03  07master * rd011215b17 10OE.dev/: Jul 30 16:30:49 merge of '0521752c751b5734b320120045655229b2faff4f' Jul 30 16:30:53 and 'f75d239d7378665c50aea95b24abbfc6a815bd43' Jul 30 17:33:10 khem: having re-read the eabi documentation I am now more convinced that this external reference to __aeabi_unwind_cpp_pr0, and hence the DT_NEEDED for libgcc, is spurious. I think the linker should probably be stripping them in this situation. Jul 30 17:39:45 03  07org.openembedded.dev * rb263817305 10OE.dev/packages/vlc/vlc-davinci_0.8.6h.bb: vlc-davinci: bump SRCREV Jul 30 17:39:46 03  07org.openembedded.dev * r75b03f903d 10OE.dev/packages/images/neuros-osd-base-image.bb: neuros-osd base image: add more neuros apps Jul 30 17:39:47 03  07org.openembedded.dev * r86c16f31d3 10OE.dev/packages/linux/ (linux-neuros/neuros-osd2/defconfig linux-neuros_git.bb): linux-neuros: bump SRCREV Jul 30 17:39:52 03  07master * r75b03f903d 10OE.dev/packages/images/neuros-osd-base-image.bb: neuros-osd base image: add more neuros apps Jul 30 17:39:53 03  07master * rb263817305 10OE.dev/packages/vlc/vlc-davinci_0.8.6h.bb: vlc-davinci: bump SRCREV Jul 30 17:40:04 03  07master * r86c16f31d3 10OE.dev/packages/linux/ (linux-neuros/neuros-osd2/defconfig linux-neuros_git.bb): linux-neuros: bump SRCREV Jul 30 17:46:57 What to do in case of sanity test errors? (do_qa_staging task failed with libform.la libpanel.la libncurses.la libmenu.la) Jul 30 17:47:21 Could somebody point how to resolve this things? Jul 30 18:36:49 03  07org.openembedded.dev * r9fe8165364 10OE.dev/packages/python/python-sphinxsearch_0.1.bb: add python-sphinxsearch, a module for querying the sphinx search daemon Jul 30 18:36:49 03  07org.openembedded.dev * r979fa0dd66 10OE.dev/: Jul 30 18:36:49 merge of '981cc5dae86118a7c743b3f8ad711e7763194aae' Jul 30 18:36:49 and 'e991ee5c3b5bbf092467b905481b36d1fde8d2de' Jul 30 18:36:50 03  07org.openembedded.dev * r1f7365553c 10OE.dev/packages/python/python-mako_0.2.2.bb: add python-mako, a super-fast templating language that borrows the best ideas from the existing templating languages Jul 30 18:36:54 03  07org.openembedded.dev * re13eea6cbe 10OE.dev/packages/tasks/task-python-everything.bb: task-python-everything: add python-mako and python-sphinxsearch Jul 30 18:36:57 03  07master * r9fe8165364 10OE.dev/packages/python/python-sphinxsearch_0.1.bb: add python-sphinxsearch, a module for querying the sphinx search daemon Jul 30 18:37:00 03  07master * r1f7365553c 10OE.dev/packages/python/python-mako_0.2.2.bb: add python-mako, a super-fast templating language that borrows the best ideas from the existing templating languages Jul 30 18:37:05 03  07master * r979fa0dd66 10OE.dev/: Jul 30 18:37:07 merge of '981cc5dae86118a7c743b3f8ad711e7763194aae' Jul 30 18:37:09 and 'e991ee5c3b5bbf092467b905481b36d1fde8d2de' Jul 30 18:37:11 03  07master * re13eea6cbe 10OE.dev/packages/tasks/task-python-everything.bb: task-python-everything: add python-mako and python-sphinxsearch Jul 30 18:45:01 03  07org.openembedded.dev * rfc67741b7f 10OE.dev/: Jul 30 18:45:01 merge of '368c3f8ff3b1b9dcaff963660d353cc8e2739101' Jul 30 18:45:01 and 'f350641f3893143efec22c74299cdeeae73af124' Jul 30 18:45:02 03  07org.openembedded.dev * r1529744350 10OE.dev/: Jul 30 18:45:02 merge of '1d6f127aae1dbce4a91a9955218de3c4f5ea5f7b' Jul 30 18:45:02 and '9a389e1e5a5eaac66e8c921c1d1196c323a5030c' Jul 30 18:45:04 03  07org.openembedded.dev * r2d81168ab0 10OE.dev/packages/tangogps/ (tangogps-fso_0.9.2.bb tangogps.inc tangogps_0.9.0.3.bb): tangogps: outfactor .inc file Jul 30 18:45:09 03  07master * r2d81168ab0 10OE.dev/packages/tangogps/ (tangogps-fso_0.9.2.bb tangogps.inc tangogps_0.9.0.3.bb): tangogps: outfactor .inc file Jul 30 18:45:12 03  07master * r1529744350 10OE.dev/: Jul 30 18:45:14 merge of '1d6f127aae1dbce4a91a9955218de3c4f5ea5f7b' Jul 30 18:45:16 and '9a389e1e5a5eaac66e8c921c1d1196c323a5030c' Jul 30 18:45:18 03  07master * rfc67741b7f 10OE.dev/: Jul 30 18:45:22 merge of '368c3f8ff3b1b9dcaff963660d353cc8e2739101' Jul 30 18:45:24 and 'f350641f3893143efec22c74299cdeeae73af124' Jul 30 19:04:11 03  07org.openembedded.dev * r6c7403aacd 10OE.dev/: Jul 30 19:04:11 merge of '29d752d6760fe8971e4d37f73f5bad5ed1b0abd7' Jul 30 19:04:11 and 'c60a4268168f4df8d07b333ec16af3ae2aed261f' Jul 30 19:04:20 03  07org.openembedded.dev * r2c9b169f45 10OE.dev/packages/ (9 files in 3 dirs): add sphinx, a full text indexer Jul 30 19:04:20 03  07master * r6c7403aacd 10OE.dev/: Jul 30 19:04:20 merge of '29d752d6760fe8971e4d37f73f5bad5ed1b0abd7' Jul 30 19:04:20 and 'c60a4268168f4df8d07b333ec16af3ae2aed261f' Jul 30 19:04:21 03  07master * r2c9b169f45 10OE.dev/packages/ (9 files in 3 dirs): add sphinx, a full text indexer Jul 30 19:28:17 morning all Jul 30 19:28:53 hi thesing Jul 30 19:36:23 bye Jul 30 19:36:55 How does cmus run on the 720 with mp3 playback? Jul 30 19:40:51 Whoops, wrong channel Jul 30 19:45:25 03  07org.openembedded.dev * r1472e3bab2 10OE.dev/ (2 files in 2 dirs): Jul 30 19:45:25 xserver-kdrive-glamo git fetch latest from git.openmoko.org, fixes (among other things) the Jul 30 19:45:25 infamous "can't read from one accelerometer while Xserver is running" bug on om-gta02 Jul 30 19:45:31 03  07master * r1472e3bab2 10OE.dev/ (2 files in 2 dirs): Jul 30 19:45:31 xserver-kdrive-glamo git fetch latest from git.openmoko.org, fixes (among other things) the Jul 30 19:45:31 infamous "can't read from one accelerometer while Xserver is running" bug on om-gta02 Jul 30 19:58:09 03  07org.openembedded.dev * r1a9a749cd7 10OE.dev/conf/machine/simpad.conf: simpad.conf: add MACHINE_DISPLAY_WIDTH/HEIGHT_PIXELS lines Jul 30 19:58:17 03  07org.openembedded.dev * r76b369760a 10OE.dev/: Jul 30 19:58:17 merge of 'ac7218640853632ccf63747d5222d3db90378a56' Jul 30 19:58:17 and 'e82644f718a3dd0995659c6a5818c83ebc731f8e' Jul 30 19:58:17 03  07master * r1a9a749cd7 10OE.dev/conf/machine/simpad.conf: simpad.conf: add MACHINE_DISPLAY_WIDTH/HEIGHT_PIXELS lines Jul 30 19:58:18 03  07master * r76b369760a 10OE.dev/: Jul 30 19:58:20 merge of 'ac7218640853632ccf63747d5222d3db90378a56' Jul 30 19:58:22 and 'e82644f718a3dd0995659c6a5818c83ebc731f8e' Jul 30 19:59:17 anyone use xvkbd? Jul 30 20:30:22 mickey|tv: so, regarding your ISO8859-1.so vs libc.so thing, I see the same problem here in my build tree. But, oddly, if you clean and rebuild glibc the problem goes away. Jul 30 20:30:33 so, it must be something to do with the bootstrapping environment. Jul 30 20:32:20 how do you configure the block size used in genext2fs when an ext2.gz filesystem is created? I can't seem to find where genext2fs is run from Jul 30 20:33:26 look for IMAGE_CMD_ext2 in conf/bitbake.conf Jul 30 20:34:28 CosmicPenguin, cool - thats what I was looking for - IMAGE_ROOTFS_SIZE Jul 30 20:55:42 Hi, could someone else please ack my recent RFC on the oe-stable mailinglist? koen has already, so I just need one more. Thanks! Jul 30 21:01:52 I'm trying to build u-boot-cvs and am getting the error "fatal: Not a valid object name 1" from git Jul 30 21:02:04 it seems to be running ' git read-tree 1 ' Jul 30 21:02:21 anyone any ideas what the '1' is supposed to be and where it's supposed to be set? Jul 30 21:08:31 bluelightning: I just did. But I don't know if I'm allowed to do this. Jul 30 21:08:59 I'm not really sure who's allowed either to be honest... :( Jul 30 21:10:46 Crofton: you can now see my dodgy photos :_) Jul 30 21:30:29 I suggest you wait some days and if nobody objects to my ack. I'm allowed to ack ;) Jul 30 21:34:52 XorA|gone, awesome! Jul 30 21:36:12 Crofton: should we start an OE network? Jul 30 21:37:53 Sure! Jul 30 21:38:10 I wonder if I already made one Jul 30 21:38:14 * XorA|gone wonders how to Jul 30 21:38:42 facebook still mystifies me Jul 30 21:39:00 facecrack Jul 30 21:39:28 * timtimred likes a bit of lunchtime facebooking Jul 30 21:39:29 ah, it must be tatoo test night Jul 30 21:39:34 I can hear explosions Jul 30 21:40:06 basically, you update your status a lot Jul 30 21:40:26 generally in ways that baffle most of the people in your network Jul 30 21:40:59 :) Jul 30 21:41:16 dont forget poking Jul 30 21:41:40 mostly i use it like that friends reunited site Jul 30 21:41:48 thats what its best at Jul 30 21:43:11 Crofton|work: we need to use xxx@openembedded.org email addresses to create one Jul 30 21:43:20 Crofton|work: http://www.facebook.com/help.php?page=403 Jul 30 21:43:21 hmmm Jul 30 21:43:47 we should work this into the git id discussion :) Jul 30 21:44:17 ROFL "I'm home schooled" Jul 30 21:44:43 We can create a "group" Jul 30 21:51:13 yes u should do that :) takes about a min Jul 30 21:54:10 ok, OpenEmbedded exists on crackbook Jul 30 21:54:16 * XorA|gone wonders how to join Jul 30 21:54:33 I think the tmieline thingy will have me creating it Jul 30 21:54:41 ahah, got it Jul 30 21:55:56 and I invited you Jul 30 21:58:29 I beat the invite Jul 30 21:59:24 awesome Jul 30 21:59:30 you are catching on Jul 30 21:59:32 * XorA|gone wishes he could persuade KDE to not manage his second monitor Jul 30 21:59:52 Crofton|work: there seems to be a lack of pr0n on facebook, the barrier to learning is therefore increased Jul 30 22:00:02 hahaha Jul 30 22:00:15 my father and step mother are on facebook Jul 30 22:00:27 I need to find college photos on my father and post them Jul 30 22:02:42 heh Jul 30 22:02:52 * XorA|gone has too many social network accounts Jul 30 22:03:28 I use linkedin and crackbook Jul 30 22:03:34 the erst are useless Jul 30 22:03:46 linkedin is the only one I dont have Jul 30 22:03:50 I think Jul 30 22:04:08 heh Jul 30 22:04:14 no dodgy photos Jul 30 22:04:18 * XorA|gone kicks KDE in the knackers Jul 30 22:04:31 Crofton|work: heh, well I like the one of you being trappaned Jul 30 22:04:49 heh **** BEGIN LOGGING AT Wed Jul 30 22:09:09 2008 Jul 30 22:28:39 do we have a png of the OE logo anywhere Jul 30 22:28:47 for the crackbook group? Jul 30 22:31:09 Crofton: only and svg in contrib Jul 30 22:31:26 yeah, and gimp is screwing up the text Jul 30 22:32:42 its not gimp, the text is actually buggered, it has some wierd offset Jul 30 22:32:50 ah Jul 30 22:33:07 I had to fix it in inkscape when I did the take off with OE logo Jul 30 22:34:31 XorA|gone: hi, do_cfongiure_prepend_htckaiser/polaris didn't work. do I have to specify somewhere that htckaiser/polaris are the select $MACHINE ? Jul 30 22:34:39 Crofton: http://wiki.openembedded.net/images/logo.png Jul 30 22:35:05 dcordes: you do that with MACHINE = "htckaiser" Jul 30 22:35:18 it might be do_configure_htckaiser_prepend Jul 30 22:35:34 dcordes: BTW rob_w is working on htcnike using same kernel Jul 30 22:35:45 XorA|gone: yea that's what I normally have in local.conf. no I have seen similar things grepping packages/linux Jul 30 22:36:03 I've heard of it :) Jul 30 22:36:14 let me retry real quick.. Jul 30 22:36:25 its a pity its not defconfig_htckaiser then you couple just to cp defconfig_${MACHINE} Jul 30 22:36:51 it is Jul 30 22:37:01 dcordes: then do the cp that way :-) Jul 30 22:37:04 queer, it display with a black background Jul 30 22:37:14 Crofton|work: odd, is where on firefox Jul 30 22:37:15 arch/arm/configs/defconfig_htckaiser and arch/arm/configs/defconfig_htcpolaris Jul 30 22:37:19 ok Jul 30 22:37:45 it was fine until I saved locally and uploaded to crackbook Jul 30 22:37:54 Crofton|work: maybe its actually transparent and facebook puts it on black background Jul 30 22:38:10 hmm Jul 30 22:38:31 yes its transparent Jul 30 22:39:12 gimp layer->transparency->remove alpha channel Jul 30 22:39:58 XorA|gone: do you think we should change linux-kaiser do linux-msm7200 if even more will join (nike) Jul 30 22:40:07 and also sedna is on the horizon Jul 30 22:40:22 dcordes: sounds possibly sensible Jul 30 22:40:44 dcordes: probably sensible to syncronise with rob_w first Jul 30 22:41:08 just some gpio tracing to get the keypad working Jul 30 22:42:52 ok I did the cp defconfig_$MACHINE let's see if it will build htc-msm with polaris defconfig now when I set MACHINE = htcpolaris Jul 30 22:58:56 XorA|gone: that built 2.6.21 Jul 30 22:59:59 XorA|gone: while htcpolaris.conf has PREFERRED_PROVIDER_virtual/kernel = "linux-kaiser" Jul 30 23:00:30 NOTE: consider defining PREFERRED_PROVIDER_virtual/kernel Jul 30 23:08:02 dcordes: dont forget to add htcpolaris to COMPATIBLE_MACHINE in linux-kaiser Jul 30 23:10:25 if I'm trying to provide an 'SDK' for developers to build/package external apps consisting of a toolchain/libs/headers/packager where should I be looking? Jul 30 23:10:34 openmoko has this correct? Jul 30 23:11:08 is time for me to sleep Jul 30 23:12:26 XorA|gone: gnight Jul 30 23:34:07 * * OE Bug 4461 has been created by lukas.gorris(AT)gmx.de Jul 30 23:34:09 * * Add machine htcpolaris Jul 30 23:34:11 * * http://bugs.openembedded.net/show_bug.cgi?id=4461 Jul 30 23:37:13 tharvey: at openembedded.org I would look Jul 30 23:56:18 dcordes, well specifically I'm asking more about how the sdk class works and what it provides and who is using it - I'll dig into the code a bit Jul 30 23:57:00 I'm using it Jul 30 23:57:19 many devices are preconfigured. it's easy to modify Jul 30 23:57:50 and an ever growing huge amount of packages Jul 30 23:58:11 currently I'm building 'meta-toolchain' which provides a tarball of the toolchain and a tarball of 'extras' - I'm not really sure what extras is for (libs/headers?) - I'm trying to figure out how I would provide an sdk that provided meta-toolchain and support for creating packages outside of OE Jul 30 23:58:12 ..is available :) Jul 30 23:58:32 I'm very familiar with OE by now... just not the 'sdk' class and its use Jul 30 23:58:55 oh sorry. what is sdk class? Jul 30 23:59:13 I use the OE toolchain to build all arm stuff Jul 30 23:59:19 algo outside of OE Jul 31 00:01:01 also Jul 31 00:10:25 dcordes, well... thats what I was asking heh... for a good summary of what the sdk class is all about - from what I can tell it looks like its used as the base for all things that would export something appropriate for an SDK - such as compiler, toolchain etc Jul 31 00:10:53 I'm just not sure how to add external packaging support - I want users to be able to cross-compile apps, then create an ipkg for them (outside of OE) Jul 31 00:14:38 tharvey: I never tried that. I think there was something in the openmoko wiki on using the OE toolchain to manually (outside, not using bitbake) compile and package Jul 31 00:14:42 let me see if I can find it Jul 31 00:18:08 tharvey: http://wiki.openmoko.org/wiki/Toolchain Jul 31 00:22:39 dcordes, cool - I should be able to figure out what I need from http://wiki.openmoko.org/wiki/Toolchain#Packaging_your_application Jul 31 00:22:40 thanks Jul 31 00:24:29 night Jul 31 00:26:04 tharvey: no problem, night **** ENDING LOGGING AT Thu Jul 31 02:59:57 2008