**** BEGIN LOGGING AT Fri Jan 25 02:59:57 2008 Jan 25 03:01:53 i am trying to get libshout to build with OE. The configure step completes successfully, however, when I go to compile, a sanity check fails for the autoconf log. anyone know why this would seem to complete successfully on configure, but not on compile? Jan 25 03:05:28 here is the message... Jan 25 03:05:29 ERROR: This autoconf log indicates errors, it looked at host includes. Rerun configure task after fixing this. Path was '/home/dmfrey/dev/embedded/gumstix-oe/tmp/work/armv5te-angstrom-linux-uclibcgnueabi/libshout-2.2.2-r0/libshout-2.2.2' Jan 25 03:05:49 this is for a gumstix build Jan 25 03:08:48 ~pastebin Jan 25 03:08:49 [~pastebin] A "pastebin" is a web-based service where you can paste anything over 3 lines without flooding the channel. Here are links to a few : http://www.pastebin.com , http://pastebin.ca , http://channels.debian.net/paste , http://paste.lisp.org , http://www.rafb.net/paste Jan 25 03:13:43 http://pastebin.com/pastebin.php?dl=m445e1181 Jan 25 03:14:11 Crofton, thanks for the info. Jan 25 03:17:46 just sorry the answer does not jump out at me Jan 25 03:20:58 i have been banging my head over it, off and on, for almost a week. it just seems odd that the configure log seems all in line for it to build correctly Jan 25 03:21:48 sthe guys that may see it quicly should be back in a few hours Jan 25 03:22:05 I am too swamped with other work to take a close look :( Jan 25 03:22:22 should i also pastebin the configure log? Jan 25 03:24:22 that might be helpful Jan 25 03:26:13 configure log: http://pastebin.com/pastebin.php?dl=m7792e110 Jan 25 03:27:31 libshout recipe: http://pastebin.com/pastebin.php?dl=m6bdae834 Jan 25 03:28:09 gl, I need to go to bed myself Jan 25 03:30:47 thanks Jan 25 06:21:22 oh , gerwinin was gone... Jan 25 07:09:17 RP:ping Jan 25 07:42:15 03rwhitby 07org.oe.dev * r874c9950... 10/ (3 files in 3 dirs): slugos-init: Updated leds handling to match latest kernel patches Jan 25 07:42:25 03rwhitby 07org.oe.dev * r52333e2f... 10/ (1 packages/tasks/task-slugos.bb): task-slugos: Enabled IDE modules for the rootfs to match the modulefunctions script Jan 25 07:56:43 good morning Jan 25 07:56:52 any news regarding the dns issues ? Jan 25 07:57:05 still can't reach ml or website Jan 25 08:02:57 steliosk: its all fixed Jan 25 08:05:33 probably it has not propagated then Jan 25 08:17:04 steliosk: UK must be nearer :-D Jan 25 08:20:16 XorA : what's the dns you are using ? Jan 25 08:20:28 steliosk: my ISP Jan 25 08:20:40 XorA : i mean ip Jan 25 08:21:37 steliosk: I dont know, my router doesnt have that info in admin pages :-( Jan 25 08:22:13 * steliosk thinks that XorA needs a new router Jan 25 08:22:20 :) Jan 25 08:22:31 steliosk: I like my linksys leave it alone :-) Jan 25 08:24:39 XorA : heh its now working.... Jan 25 08:24:54 DNS cache running on your network? Jan 25 08:25:49 XorA : yes Jan 25 08:26:13 that'll be why then :-) Jan 25 08:28:01 i had a look at it and there were no entries for OE. but then there it was... easydns ;) Jan 25 08:28:28 probably the updates hit the morning rush our in UK :) Jan 25 08:28:39 s/our/hour Jan 25 08:30:19 * steliosk goes to get some fresh coffee Jan 25 08:31:44 I'm Trying to build images for two different machines. I've looked at the oe wiki but I think I don't know what term I'm looking for: Can someone give me a hint on where to look / what to look for? Jan 25 08:33:46 do I just backup my local.conf then change to the other machine? Jan 25 08:47:01 well I changed the MACHINE entry in my local.conf and it seems to be building binaries for the right architecture so I guess that works Jan 25 08:47:09 is this the officially sanction way to do things? Jan 25 09:04:01 morning Jan 25 09:04:17 yes its the proper way Jan 25 09:12:53 hi all. i'm trying to add libboost-filesystem1.33.1 (IMAGE_INSATLL) to a custom image, but bitbake fails with "No providers of runtime build target libboost-filesystem1.33.1". anybody has an idea what could be wrong? Jan 25 09:15:40 morning all Jan 25 09:17:51 ~hail steliosk Jan 25 09:17:52 * ibot bows down to steliosk and chants, "I'M NOT WORTHY!!" Jan 25 09:20:00 justri: do not add libraries Jan 25 09:20:12 justri: add apps - libs will be autoadded Jan 25 09:20:17 re Jan 25 09:23:43 hrw: the app which depends on the lib is at the moment unfortunately not handled by openembedded Jan 25 09:24:43 add it into own repo? Jan 25 09:24:53 or try to use libboost-filesystem name Jan 25 09:28:47 hrw: adding to repo is not possible (to many conflicts), i tried adding libboost-filesystem1.33.1 and boost-filesystem, which didn't work. maybe i just should add a dummy app which depends on libboost-filesystem? Jan 25 09:29:18 justri: 'add to own repo' == make recipe and keep it local Jan 25 09:30:00 hrw: ah, ok. i'll try Jan 25 09:47:01 morning Jan 25 09:49:12 hi mithro rwhitby Jan 25 10:20:11 /win 12 Jan 25 10:28:28 broonie: /win 2.11 Jan 25 10:31:45 :P Jan 25 11:12:06 Any comments on http://bugs.openembedded.org/show_bug.cgi?id=1856#c3 ? Jan 25 11:12:37 It would life and bb files much easier because it would often make setting S explicitly for native package unnecessary Jan 25 11:12:52 I'll also RFC it, but want to hear your opinions first Jan 25 11:14:14 ok for me Jan 25 11:15:54 hrw: No obvious issues such as forgetting to escape some special characters or some such that you can easily spot? Jan 25 11:31:27 good morning Jan 25 11:31:32 morning all. Jan 25 11:32:55 yo thesing Jan 25 11:34:01 hi thesing Jan 25 11:34:09 thesing: how many collies you have? Jan 25 11:34:26 hrw: only one. Jan 25 11:34:59 want one more? Jan 25 11:35:51 Would be nice to have. So I could compare registers on 2.4 and 2.6. Jan 25 11:36:05 * * OE Bug 3383 has been RESOLVED (FIXED) by dp(AT)xora.org.uk Jan 25 11:36:06 * * zlib: allow debian name to be customized to zlib1g Jan 25 11:36:09 * * http://bugs.openembedded.org/show_bug.cgi?id=3383 Jan 25 11:36:50 thesing: if all will go ok I will send you mine soon then. full box but with other power supply Jan 25 11:36:59 I use sharp power supply to charge usb hub Jan 25 11:38:03 thanks. Jan 25 11:38:19 didn't the usb hub have an own power supply? Jan 25 11:40:29 have and thats what I will include in box Jan 25 11:40:56 I do not have space for usbhub powerplug type charger Jan 25 11:40:58 hi Bernardo Jan 25 11:42:37 hrw: a common problem. Jan 25 11:43:30 * XorA needs to sort power situation here as well Jan 25 11:43:44 might need to get a electrician in Jan 25 11:56:50 Hello.... I'm trying to get the qtopia 4 recipies working for the new Gumstix OE. I was working with Holger earlier in the week (whos done pretty much everything). I can now build images using the new recipies, however I'm having trouble creating a hello world type example bb. The problem seems to be that ${S} is being set wrong somewhere. Does anyone know where ${S} is set? Jan 25 11:57:48 03florian 07org.oe.dev * r1db951a6... 10/ (66 files in 30 dirs): dummy ci Jan 25 11:57:56 03florian 07org.oe.dev * raa327bc6... 10/ (66 files in 30 dirs): disapproval of revision '1db951a66d993086346a502253112d6a52d1ad3f' Jan 25 11:58:03 03florian 07org.oe.dev * r7164da9f... 10/ (4 files in 3 dirs): calendar2: Disable broken tests Jan 25 11:59:48 oh darn... the old mtn did not sync checkins that were disapproved locally Jan 25 12:00:31 florian: hey Jan 25 12:00:49 hi zecke Jan 25 12:00:55 florian: got my mail? Jan 25 12:01:32 zecke: yes, i guessed that my mails were eaten by some spamfiler or somethign like this :) Jan 25 12:02:19 zecke: I'll reply in the next free minute :) Jan 25 12:06:05 * * OE Bug 2471 has been RESOLVED (WONTFIX) by Jan 25 12:06:07 * * Hardcoded PACKAGES lines Jan 25 12:06:09 * * http://bugs.openembedded.org/show_bug.cgi?id=2471 Jan 25 12:19:02 Is a double override like PREFERRED_PROVIDER_glibc_collie = "2.5" supposed to work? Jan 25 12:19:11 or rather an override with two conditions Jan 25 12:54:37 Laibsch: It can, yes Jan 25 12:55:05 Laibsch: It has also been known to break but in theory it should... Jan 25 12:55:12 (work) Jan 25 12:55:17 Well, it did not work when I tested it Jan 25 12:55:31 I added above line and started "bitbake -i" Jan 25 12:55:46 "which glibc" still answered 2.6.1 Jan 25 13:03:02 RP: ? Jan 25 13:05:34 Laibsch: I don't think glibc is an overide Jan 25 13:06:16 RP: Do you have an idea how to specify for collie to use glibc 2.5 instead of the default 2.6.1? Jan 25 13:06:22 I'd like to do it in local.conf Jan 25 13:06:43 so as not to touch the files provided by angstrom-stable Jan 25 13:07:40 Laibsch: I think "linux" is the override you want Jan 25 13:08:39 Aha, thank you Jan 25 13:18:32 RP: linux looks more like the kernel to me Jan 25 13:18:50 Laibsch: Its target-os Jan 25 13:18:57 Thanks Jan 25 13:18:59 linux | linux-uclibc Jan 25 13:19:07 | linux-gnueabi Jan 25 13:19:32 | linux-uclibceabi? Jan 25 13:21:19 XorA, you are the sound guru? Jan 25 13:21:32 I have a sumb question Jan 25 13:21:54 IS 48kHz a widely suported sample rate (in hw) Jan 25 13:22:11 hrw: yes Jan 25 13:22:18 Crofton: It is, over the last few years at least Jan 25 13:22:28 Crofton: I'd say yes too Jan 25 13:22:38 ok Jan 25 13:22:59 I need a sane default that avoids whatever sample rate voodoo alsa can do Jan 25 13:23:30 RP, hrw: I don't think this is it. angstrom-2007-preferred-versions.inc does define PREFERRED_VERSION_glibc ?= "2.6.1" Jan 25 13:23:34 Crofton: IIRC, AC97 requires 48KHz in hw Jan 25 13:23:41 There's nothing that'll *guarantee* you won't get any rate conversion. Jan 25 13:24:00 hrw, RP: But PREFERRED_VERSION_glibc_collie = "2.5" does not seem to have the expected result Jan 25 13:24:33 Laibsch: Thats because glibc isn't an override, its a package name Jan 25 13:25:01 my friend machine has nearly 48kHz but not exactly one - it was hard to get proper crystal for clock Jan 25 13:25:09 thanks guys Jan 25 13:25:26 Laibsch: PREFERRED_VERSION looks up package names as a suffix Jan 25 13:25:52 Laibsch: If that doesn't work for the collie example, it probably means bitbake isn't expanding overrides somewhere when it should Jan 25 13:28:11 RP: Where to fix? Jan 25 13:28:14 Other question: Can sed *remove* a matching line? Jan 25 13:28:59 found the sed thing, thanks Jan 25 13:34:06 * Laibsch hugs sed Jan 25 13:36:19 * hrw search for lost GB of space Jan 25 13:36:37 Filesystem Size Used Avail Use% Mounted on Jan 25 13:36:38 /dev/mapper/home--vg-homefs 396G 322G 59G 85% /home Jan 25 13:41:26 hrw: maybe the rest is reserved for root? Jan 25 13:42:06 ;)) Jan 25 13:42:18 thesing_: I need to burn some dvds with stuff thats sure Jan 25 13:42:47 but I lack dual-layer dvds now Jan 25 13:49:40 * mwester recovered a great deal of space from his /home partiton with "rm -rf ~/.ccache" :) Jan 25 13:54:58 scratchbox fun... rebooting Jan 25 13:56:27 Crofton: if you adrees the hardware device in alsa eg hw:1 as sound device then it wont do any voodoo Jan 25 13:57:03 Crofton: voodoo only happens when you send stuff through the default which allows all the plugins to get a look it Jan 25 13:58:07 so address the hw device directly and it il lfail if you request unsupported rates? Jan 25 14:08:06 anyone know anything about libshout? Jan 25 14:12:16 Crofton: I know you are asking about it and that it is probably shoutcast related? Jan 25 14:13:09 some guy is trying to build it on a gumstix Jan 25 14:13:28 he had a problem with sanity catching something in the config.log Jan 25 14:13:40 mainly I am curious if someone had looked at it Jan 25 14:14:03 BTW: good job with the gumstix Jan 25 14:14:30 ah we don't have it in OE at all, now I understand Jan 25 14:15:26 well, I need to find some time to back port stuff that sakoman has done Jan 25 14:15:38 Crofton: yes Jan 25 14:15:41 but he has structured his repo so that his local changes are easy to find Jan 25 14:16:34 I actually like the way he has set things up, especially once he starts basing off angstrom stable Jan 25 14:16:56 well, I do not the like the bit about loading the OE meta-data in svn, that is one slow upadte :) Jan 25 14:17:52 Crofton: don't get me started on irc Jan 25 14:17:55 XorA: hey Jan 25 14:18:14 zecke: hey Jan 25 14:19:03 zecke, talk to me at FOSDEM then :) Jan 25 14:22:04 * * OE Bug 3709 has been RESOLVED (FIXED) by thommycheck(AT)gmx.de Jan 25 14:22:06 * * 2.6.23 git kernel for collie does not pass size test Jan 25 14:22:08 * * http://bugs.openembedded.org/show_bug.cgi?id=3709 Jan 25 14:26:18 03Laibsch 07org.oe.dev * r68cf185f... 10/ (64 files in 27 dirs): divers: remove lines of the form "PACKAGES = ${PN}" from 63 files Jan 25 14:32:18 morning Jan 25 15:03:37 03Laibsch 07org.oe.dev * rabdd40ff... 10/ (3 files in 3 dirs): beep and ez-ipupdate: replace 'PACKAGES =' with 'PACKAGES +='. Partly fixes 2471. Jan 25 15:10:08 RP: it looks like the corgi_ts stuff in arch/arm/mach-pxa/poodle.c needs some TLC for 2.6.24 Jan 25 15:24:29 thesing_: Am Fließ address is still proper one? Jan 25 15:24:37 hrw: yes Jan 25 15:24:43 ok Jan 25 15:25:46 hrw: Do you know who did the changes to corgi.c in 2.6.24? i think poodle.c needs some fixes as well Jan 25 15:26:13 is there a way to force bitbake to run a specific task again? Jan 25 15:28:03 -f Jan 25 15:28:10 bitbake -c compile -f hello Jan 25 15:28:47 XorA: Thats to simple. ;) Jan 25 15:28:54 s/o/oo Jan 25 15:42:33 hrw: Why are you scaring unsupecting collie users away? ;-) Jan 25 15:48:20 Laibsch: does he? Jan 25 15:51:52 03Laibsch 07org.oe.dev * r3ba9bb70... 10/ (86 files in 61 dirs): divers: add debug package for 85 files with hardcoded PACKAGES definition. Partly fixes 2471. Jan 25 15:53:44 03Laibsch 07org.oe.dev * r3ba9bb70... 10/ (86 files in 61 dirs): divers: add debug package for 85 files with hardcoded PACKAGES definition. Partly fixes 2471. Jan 25 15:55:00 yes, he seems to: http://hentges.net/tmp/logs/irc/livelogs/%23angstrom.livelog#[20080125%2016:00:37] Jan 25 15:55:32 thesing_: But now I found out that hrw also mentioned you are working hard on it so I don't know how that guy got the impression he got Jan 25 15:58:29 hvontres|work: use git annotate? Jan 25 15:59:29 Laibsch: I am telling hard true Jan 25 16:00:06 I think collie has some life left Jan 25 16:00:13 I prefer it to spitz Jan 25 16:00:39 just because spitz is bulky - right? Jan 25 16:00:41 And I will rather use it once I get that darn image creation problem fixed Jan 25 16:00:50 yes, I don't like the form factor Jan 25 16:00:55 I also prefer the keyboard Jan 25 16:01:16 I prefer c7x0 keyboard Jan 25 16:02:03 * Laibsch has never touched it Jan 25 16:02:29 similar to spitz/akita one but more flat foil touch Jan 25 16:04:39 Collie screen is also better than spitz screen. Ok spitz screen is bigger but collie screen is transreflective. Jan 25 16:06:03 basically spitz has same screen as c7x0 but c7x0 has more backlight levels so it is more visible Jan 25 16:06:09 readable Jan 25 16:06:47 he wasn't scaring me away :P Jan 25 16:08:59 I'd be willing to do some work on the kernel if someone could help me get started Jan 25 16:09:01 brb, cookies Jan 25 16:09:29 wvdschel: did you do kernel hacking before? Jan 25 16:13:08 thesing_, not much Jan 25 16:13:13 some playing around really Jan 25 16:13:40 did some one touch opie lately? Jan 25 16:13:56 I've never really gotten into it because I didn't know where to start looking for things Jan 25 16:14:13 parts that are really missing are usb-gadget support and sound support. Jan 25 16:14:51 zecke: I did. "rm -rf ~/devel/opie" Jan 25 16:15:00 sound support Jan 25 16:15:01 zecke: does it count? Jan 25 16:15:20 but what does usb-gadget do? Jan 25 16:15:32 zecke: sure, why are you asking? Jan 25 16:15:34 hrw: EOL while scanning single-quoted string (, line 1) while parsing /home/ich/source/embedded/org.openembedded.dev/packages/opie-taskbar/opie-taskbar-images_cvs.bb Jan 25 16:15:49 zecke: Let me take a look Jan 25 16:15:54 ${@['', 'opie-taskbar-images (= ${OPIE_CVS_PV}-r6)'][packaged(bb.data.getVar('PN', d, 1), d) == True]} Jan 25 16:15:57 I am doing some automatic changes Jan 25 16:16:10 so, OPIE_CVS_PV might not be defined here Jan 25 16:16:10 And this might break here and there ;-) Jan 25 16:16:15 I'll fix it pronto Jan 25 16:16:27 (if it was me) Jan 25 16:16:35 zecke: Thanks for letting us know Jan 25 16:16:47 lost wifi connection Jan 25 16:16:57 wvdschel: it lets you use the device as usb client. For example ethernet over usb. Jan 25 16:17:05 ah I see Jan 25 16:17:12 Laibsch: well, bitbake stops after parsing on errors ;) Jan 25 16:17:17 thesing_: for collie? Jan 25 16:17:23 well has any work been done on either? Jan 25 16:17:28 zecke: Worked fine for me last time I tried Jan 25 16:17:35 ;-) Jan 25 16:17:40 Newbie question: can someone tell me the appropriate way to reconfigure my kernel? Should I just make menuconfig under work/, or is there a better way which integrates with bitbake? Jan 25 16:18:02 Laibsch: for collie its not done yet. Jan 25 16:18:10 wvdschel: no. Jan 25 16:18:36 thesing_: But what kind of USB options does the collie have? Jan 25 16:18:41 The cradle? Jan 25 16:18:48 Yes, of course. Jan 25 16:18:50 Silly me Jan 25 16:18:53 Laibsch: conf/distro/include/preferred-opie-versions-1.2.4-pre.inc this might not be included by my distro.. ;) Jan 25 16:19:09 Which is? Jan 25 16:19:15 I'm currently getting a vanilla oe environment Jan 25 16:19:22 zecke: I am glad it looks like I did not break anything ;-) Jan 25 16:19:27 wvdschel: you would have to grab the sa1100 manual the read the usb-gadget doc and start writing a driver. Is a not working driver somewhere. Jan 25 16:19:27 what distro/machine/... configs do I use? Jan 25 16:19:44 and for sound? Jan 25 16:20:08 wvdschel: its the same but without a driver at all. Jan 25 16:20:09 zecke: OPIE_CVS_PV is defined only in opie.bbclass Jan 25 16:20:26 zecke: and that recipe does not inherit it Jan 25 16:20:55 hjm200: menuconfig is best Jan 25 16:21:17 Laibsch: with the crandle or a usb cable. (trisoft had them) Jan 25 16:21:39 thesing_: I will add zaurus->usb cable if you lack it Jan 25 16:22:25 hrw: I don't have one. Thanks. Jan 25 16:22:26 zecke: No, I did not touch that file Jan 25 16:23:50 hrw: thanks Jan 25 16:23:57 wvdschel: There are other things to be done on collie. See the bugtracker. Jan 25 16:24:07 Laibsch: want to add "inherit opie" to opie-taskbar-images? Jan 25 16:25:31 hrw: where the oz collie images compiled with serial console? Jan 25 16:26:16 and does someone here still have them? I didn't find them at the website. Jan 25 16:26:42 zecke: I will Jan 25 16:28:09 thesing_: you can build such ones with OE Jan 25 16:28:16 thesing_: bitbake openzaurus-sa Jan 25 16:28:57 hrw: I will try but I suspect that they are bitrotten. Jan 25 16:32:21 thesing_: I think I might have a few backups Jan 25 16:32:27 But I cannot check right now Jan 25 16:32:31 gotta go Jan 25 16:33:09 Laibsch: It would be nice if you could check later. Jan 25 16:33:20 thesing_: CoreDump might have some laying around Jan 25 16:33:43 hvontres|work: good idea. Jan 25 16:35:14 zecke: 8725fc715529f6dba921a0a92e01e4e68124bfdb Jan 25 16:37:37 Crofton: I will add libshout tonight, I'm currently on it Jan 25 16:38:18 hrw: thanks. copied the .config over the package defconfig so bitbake can rebuild it Jan 25 16:42:13 thesing_, where do I get the latest collie kernel source / patches? Jan 25 16:43:40 wvdschel: from oe. bitbake -c patch virtual/kernel will give you the sources with applied patches in TMPDIR/work/collie*/linux-rp* Jan 25 16:44:30 and how do I tell OE to try them out? Jan 25 16:46:12 or does it just use whats available locally when I tell it to build virtual/kernel? Jan 25 16:46:49 wvdschel: it builds the most recent version it knows Jan 25 16:47:21 wvdschel: did you already configure OE? Jan 25 16:52:36 thesing_, yes, why? Jan 25 16:52:50 i just picked angstrom-2008.1 as distro and collie as machine Jan 25 16:52:55 thats correct right? Jan 25 16:53:03 wvdschel: Yes. Jan 25 16:53:19 im having it build nano now, just so that I have the toolchain and everything handy Jan 25 16:54:06 I really should be doing this on my desktop, its a lot faster :( Jan 25 16:56:13 wvdschel: to get only toolchain "bitbake gcc-cross" is enough Jan 25 16:56:31 bye Jan 25 16:56:35 bye :) Jan 25 16:57:12 Angstrom 2008 is broken for collie, I think Jan 25 16:57:12 thesing_, how did OZ and friends play sound on the collie before then? Jan 25 16:57:26 OZ was kernel 2.4 Jan 25 16:57:40 is there a sonkei distro file then? Jan 25 16:57:56 erm, wasn't alsa already around back then? Jan 25 16:58:05 or isn't it an alsa vs oss thing? Jan 25 16:59:56 try 2008 first Jan 25 17:00:03 if it works, all the better Jan 25 17:00:26 wvdschel: it an alsa vs oss thing. Jan 25 17:01:01 alright then Jan 25 17:01:30 erm, its bugging me about /opt/openembedded/org.openembedded.dev/packages/opie-taskbar/opie-taskbar-images_cvs.bb, anyone that knows a quick fix? Jan 25 17:01:47 wvdschel: you get the souce for the 2.4 kernel if you 'bitbake openzaurus-sa -c patch' Jan 25 17:02:40 I was told its dirty, is it? Jan 25 17:02:56 is anyone familiar with the ltp package? Jan 25 17:09:05 wvdschel: yes its quite dirty. But one gets used to it. Jan 25 17:15:34 its compiling some stuff, I'm going to get back to my course first, I'll try and do something useful before FOSDEM Jan 25 17:32:13 zecke, awesome Jan 25 17:38:45 Crofton: added and I'm pushing some local changes Jan 25 17:46:26 03freyther 07org.oe.dev * ra7748351... 10/ (1 packages/libogg/libtheora_0.9+1.0alpha7.bb): Jan 25 17:46:26 packages/libogg/libtheora: The encoder example is using vorbis add it to the DEPENDS Jan 25 17:46:26 Add libvorbis to the DEPENDS as the encoder example is using libvorbis. Jan 25 17:46:36 03freyther 07org.oe.dev * r32bd256d... 10/ (5 files in 2 dirs): packages/speex/speex: Stage the speex.m4 file as well, copied the line from the vorbis recipe Jan 25 17:46:39 03freyther 07org.oe.dev * r644e6a94... 10/ (6 files in 4 dirs): Jan 25 17:46:39 packages/libshout/libshout: Add initial version 2.2.2 Jan 25 17:46:39 Remove ogg.m4, speex.m4, vorbis.m4 from the source directory before Jan 25 17:46:39 starting to regenerate the configure script. Jan 25 17:46:40 Do not add $prefix/include to the CPPFLAGS as this is evil Jan 25 17:46:49 03freyther 07org.oe.dev * r583c5bfa... 10/ (1 conf/checksums.ini): conf/checksums.ini: Add shoutcast to the list Jan 25 17:46:57 zecke, I owe you beer Jan 25 17:47:00 03freyther 07org.oe.dev * r089b70d9... 10/ (1 conf/checksums.ini): conf/checksums.ini: Add fakeroot Jan 25 17:47:05 there we go Jan 25 17:47:05 03freyther 07org.oe.dev * r4fa8a98f... 10/ (1 conf/checksums.ini): conf/checksums.ini: Add theora 1.0alpha7 Jan 25 17:47:11 I even played your janitor... Jan 25 17:47:13 03freyther 07org.oe.dev * r1a282ca1... 10/ (1 conf/checksums.ini): conf/checksums.ini: Add libvorbis 1.2.0 Jan 25 17:47:19 03freyther 07org.oe.dev * r111c34bd... 10/ (1 conf/checksums.ini): conf/checksums.ini: Add MesaLib 7.0.2 Jan 25 17:47:32 03freyther 07org.oe.dev * r0f70c831... 10/ (1 conf/checksums.ini): conf/checksums.ini: Add libogg 1.1.3 Jan 25 17:47:42 03freyther 07org.oe.dev * r9953967e... 10/ (1 conf/checksums.ini): conf/checksums.ini: Add gcc-4.2.2 Jan 25 17:48:00 03freyther 07org.oe.dev * r5a920543... 10/ (1 conf/checksums.ini): conf/checksums.ini: Add libXi-1.1.3 Jan 25 17:48:09 03freyther 07org.oe.dev * rc5184674... 10/ (1 conf/checksums.ini): conf/checksums.ini: Add busybox 1.7.2 and the patches we apply Jan 25 17:48:20 03freyther 07org.oe.dev * r01da10b8... 10/ (1 conf/checksums.ini): conf/checksums.ini: Add Xrandr 1.2.2 Jan 25 17:48:27 03freyther 07org.oe.dev * r1e7c7ad5... 10/ (1 conf/checksums.ini): conf/checksums.ini: Add fakeroot 1.8.3 Jan 25 17:48:38 03freyther 07org.oe.dev * r80414ca6... 10/ (1 conf/checksums.ini): conf/checksums.ini:Add gtk+ 2.12.0 Jan 25 17:48:51 03freyther 07org.oe.dev * rcbc49bb6... 10/ (1 conf/checksums.ini): conf/checksums.ini: Add libXcursor 1.1.9 Jan 25 17:49:04 03freyther 07org.oe.dev * rfc00a494... 10/ (1 conf/checksums.ini): conf/checksums.ini: Add cairo 1.4.8 Jan 25 17:49:12 03freyther 07org.oe.dev * r7208ce85... 10/ (1 conf/checksums.ini): conf/checksums.ini: Add 0.9.4 of libXrender Jan 25 17:49:22 03freyther 07org.oe.dev * r2a7c4c19... 10/ (1 conf/checksums.ini): conf/checksums.ini: Add libICE 1.0.4 Jan 25 17:49:35 03freyther 07org.oe.dev * r1d9b1d2a... 10/ (1 conf/checksums.ini): conf/checksums.ini: Add glib-2.14.0 Jan 25 17:49:46 03freyther 07org.oe.dev * rdbdba935... 10/ (1 conf/checksums.ini): conf/checksums.ini: Add glibc 2.6.1 Jan 25 17:49:57 03freyther 07org.oe.dev * r2c69f2ad... 10/ (1 conf/checksums.ini): conf/checksums.ini: Add quilt 0.46 Jan 25 17:50:07 03freyther 07org.oe.dev * rbdc91fac... 10/ (1 conf/checksums.ini): conf/checksums.ini: Add inputproto 1.4.2.1 Jan 25 17:50:24 03freyther 07org.oe.dev * r7ad89e84... 10/ (1 conf/checksums.ini): conf/checksums.ini: Add renderproto 0.9.3 Jan 25 17:50:59 03freyther 07org.oe.dev * r39cd01d5... 10/ (1 conf/checksums.ini): conf/checksums.ini: Add libX11 1.1.3 Jan 25 17:50:59 checksums eh, cool Jan 25 17:51:19 03freyther 07org.oe.dev * r9f6c6e06... 10/ (1 packages/libshout/libshout.inc): Jan 25 17:51:19 packages/libshout/libshout.inc: Stage the headers, libs and m4 file Jan 25 17:51:19 Like all xiph.org packages the .pc files are a bit odd and not staged. Jan 25 17:51:28 kergoth: we have plenty of them! Jan 25 17:51:39 apparently :) Jan 25 17:51:39 03Laibsch 07org.oe.dev * r8725fc71... 10/ (1 packages/opie-taskbar/opie-taskbar-images.inc): opie-taskbar-images: add "inherit opie" stanza. Thanks, zecke for spotting it. Jan 25 17:51:40 kergoth: md5 and sha256 Jan 25 18:08:42 hmmm when I just run bitbake nano, it tries to build qemu, and the config fails because it wont accept a gcc 4.x compiler Jan 25 18:08:47 is this normal? Jan 25 18:10:50 bbl Jan 25 18:11:18 wvdschel, yes Jan 25 18:12:20 wvdschel: Try bitbake gcc-native first Jan 25 18:12:38 (gcc-native is experimental but it might help) Jan 25 18:17:20 doh... Jan 25 18:17:26 this ltp make env is driving me crazy Jan 25 18:17:48 I really hate custom handcrafted make stuff with some additional perl processing and who knows what Jan 25 18:19:38 I am trying to update ltp to the newest version Jan 25 18:19:57 but somewhere along the lines for one test they seem to be picking native gcc and not the crosscompiler Jan 25 18:20:08 and it seems that finding that place is a nightmare Jan 25 18:26:30 RP: is there a DEPMOD variable? because there is depmod for 2.4 and depmod-2.6 Jan 25 18:27:41 thesing_: I think update-alternatives handles that in the target rootfs Jan 25 18:28:49 RP: yes but I want to run it on image generation. and in TMP/cross/bin there is only arm-angstrom-linux-depmod-2.6 Jan 25 18:29:19 thesing_: There is no variable that I know of Jan 25 18:33:01 RP: is there a way to get the output of commands run by bitbake? Jan 25 18:34:08 mickeyl: you alive? Jan 25 18:45:45 thesing_: you mean the scripts in the WORKDIR/temp directory? Jan 25 18:46:23 RP: yes. I discovered them. Jan 25 18:48:14 what exactly is the devshell doing? what is so special about the terminal that is opened? Jan 25 18:49:06 re Jan 25 18:49:27 wb zecke Jan 25 18:53:05 RP: I need some training between the difference of native and cross Jan 25 18:56:39 ggilbert_: kind of. busy getting rid of the broken ubuntu and setting up Mandriva 2008 Jan 25 18:57:06 hehe Jan 25 18:57:38 broken ubuntu this is not possible :) Jan 25 18:57:42 Jin^eLD: it populates your environment, so you find yourself in an OE environment making it very convenient to develop Jan 25 18:57:57 mickeyl: I was planning to get rid of the oesources.org domain when it expires this year. If you'd like it, I can transfer it to you. if not I'll just let it expire when it comes up in september Jan 25 18:58:02 Jin^eLD: i.e. you can edit the sourcecode and just 'make' Jan 25 18:58:27 ggilbert_: feel free to let it expire. I guess we want to setup sources.openmoko.org then. thanks! Jan 25 18:58:46 mickeyl: meaning that CC is already set to the cross compiler and so on? Jan 25 18:59:05 Jin^eLD: exactly Jan 25 18:59:10 Sure thing. Just figured I'd make the offer Jan 25 18:59:13 cool, thanks Jan 25 18:59:16 ggilbert_: oops. sources.openembedded.org, of course Jan 25 18:59:22 lol Jan 25 18:59:27 habits... Jan 25 18:59:32 been typing that too much Jan 25 18:59:57 I've not pointed it at the new server, since bandwdith is a bit pricier than it used to be :p Jan 25 19:01:11 where do I put that file? Jan 25 19:01:13 oops wrong window Jan 25 19:01:14 :) Jan 25 19:02:59 mickeyl: I guess you are the one to talk to about packages/lua/lua.inc? Jan 25 19:03:08 I am working on Jan 25 19:03:14 !oebug 2317 Jan 25 19:03:18 * * Bug 2317, Status: NEW, Created: 2007-05-14 15:11 Jan 25 19:03:19 * * : incorrect use of FILES_$\{PN\}-dbg Jan 25 19:03:20 * * http://bugs.openembedded.net/show_bug.cgi?id=2317 Jan 25 19:03:51 I wonder if there ever is the need for more than one -dbg ipk being created Jan 25 19:04:05 lua.inc is such a case and I wonder about the purpose Jan 25 19:04:31 hmm lua? you mean because i'm mr.scripintg? Jan 25 19:04:39 mr python Jan 25 19:04:47 I thought lua was about python Jan 25 19:04:47 feel free to make just one big dbg package Jan 25 19:04:49 no? Jan 25 19:04:50 no Jan 25 19:04:53 lua is a scripting language Jan 25 19:04:53 OK Jan 25 19:04:59 that's about it :) Jan 25 19:04:59 Aha Jan 25 19:05:23 back to your question, fine granular packaging makes sense, but IMO not for -dbg or -dev packages Jan 25 19:05:39 yes, that is waht I thought Jan 25 19:05:48 Those packages are probably rarely used Jan 25 19:05:59 And when they are used, space is not the biggest concern Jan 25 19:06:22 thesing_: Are you short on CPU cycles? Jan 25 19:06:54 If you have a few cycles to spare, maybe you can move your tmp dir to a safe place and retry a "bitbake minimal-image"? Jan 25 19:08:06 Laibsch: I did the last night "bitbake console-image" from scratch. But I use angstrom-2007.1 as distro (copied from stable) Jan 25 19:08:37 And it worked? Jan 25 19:09:07 I tried with angstrom stable (the branch, checked out to a different directory) and that failed real early Jan 25 19:09:14 building glibc-intermediate 2.6.1 Jan 25 19:09:28 What exactly did you copy? Jan 25 19:11:31 cd conf/distro; cp include/angstrom-2007-for-openmoko.inc angstrom-2007.1.conf Jan 25 19:13:08 RP: How much do you know about the changes in corgi_ts_machinfo between 2.6.23 and 2.6.24? Jan 25 19:13:45 hi mickeyl Jan 25 19:13:47 hi zecke_ Jan 25 19:13:57 thesing_: I guess you omit the destination at the end right or do you really intend to overwrite angstrom-2007.1.conf with angstrom-2007-for-openmoko.inc Jan 25 19:13:59 ? Jan 25 19:14:26 angtrom-2007.1.conf was dropped in .dev Jan 25 19:14:43 hi pb_ Jan 25 19:14:52 I didn't copy the distro file from stable. Jan 25 19:15:19 I looked at the commit that removed 2007.1 and reversed it locally ;) Jan 25 19:16:01 I think I did something similar for Sonkei, but I don't recollect the step 100% Jan 25 19:17:51 thesing_: Just to verify once more, your compilation of the image ended fine, including do_rootfs? Jan 25 19:18:26 yes. Jan 25 19:18:40 thanks Jan 25 19:19:12 Its possible that I changed something to make it work but I don't think so. Jan 25 19:20:01 I'm running mtn diff to be sure. Jan 25 19:20:06 Thanks Jan 25 19:36:47 Laibsch: all my changes: http://www.pastebin.ca/872440 Jan 25 19:39:30 rotfl, I'm an idiot :) all the way I was looking why binaries are generated for the wrong arch Jan 25 19:39:37 and then it turns out they were included with the source package Jan 25 19:39:41 baaaah Jan 25 20:24:41 is it possible to pass an option (-l or --ignore-whitespace) to patch? It keeps choking on a whitespace change Jan 25 20:37:34 i need help creating a bb file from scratch, the bb file compiles the source file, only 1 library needs to be installed, but the bb file fails on the make install phase Jan 25 20:38:47 scant: you could override the install stage and do it manually Jan 25 20:39:00 i.e. write your own do_install{ Jan 25 20:39:03 in the bb file Jan 25 20:39:51 Jin^eLD: i have to do that, but don't know how to write my own do_install Jan 25 20:40:08 look at packages/ltp_bl.bb Jan 25 20:40:17 aem, not bl but bla.. :) I do not remember the version number Jan 25 20:40:28 they have a large do_install section, you could check that as an example Jan 25 20:42:18 scant: pastebin your .bb file so people can see it. Jan 25 20:45:38 rwhitby: http://pastebin.ca/872536 Jan 25 20:46:29 it should compile libsane-pixma.so and libsane-pixma.so needs to be renamed to libsane-pixma.so.1, and moved to /usr/lib/sane Jan 25 20:46:48 ok, now pastebin the error messages when install fails Jan 25 20:48:28 rwhitby: http://pastebin.ca/872541 Jan 25 20:49:47 rwhitby: i looked in 'tmp/work/armv5teb-linux/libsane-pixma-0.13.1-r1/mp150-0.13.1' where the source files are and i don't see any compiled files though Jan 25 20:51:48 scant: your makefile needs an install target Jan 25 20:52:02 or you need to write a custom do_install which uses the existing makefile target Jan 25 20:53:02 rwhitby: only 1 file needs to be installed, i think i would rather implement a do_install Jan 25 20:53:12 rwhitby: how come i don't see any compiled files though Jan 25 20:53:47 scant: no idea - hopefully an OE packaging guru will help you now that all the info is there for them to see Jan 25 20:54:28 rwhitby: are likes 14, 15, 16 correct in the bb file? Jan 25 20:57:00 scant: I no nothing about lib renaming, so I'm the wrong person to ask. Jan 25 20:59:29 rwhitby: ok, thanks for your advice Jan 25 20:59:44 gotta go, but i'll ask again when there are more people around Jan 25 21:11:59 hmm, interesting Jan 25 21:12:05 ltp is 22MB, the ipk that is Jan 25 21:12:27 when I try to install it, after some long waiting I get errors like ipkg: extract_archive: /usr/libexec/ltp/teblabla: No such file or directory Jan 25 21:12:44 if I manually do ar x and then untar the data.tar.gz then I see all stuff is there Jan 25 21:12:49 so ipkg bug? Jan 25 21:12:53 or is it running out of mem or something= Jan 25 21:12:58 did anyone see something like that before? Jan 25 22:15:51 ERROR: Build of /opt/openembedded/org.openembedded.dev/packages/glibc/glibc-intermediate_2.6.1.bb do_compile failed Jan 25 22:15:57 does this happen to anyone but me? Jan 25 22:23:28 wvdschel: I use angstrom 2007.1 for .dev. Jan 25 22:23:41 ah ok Jan 25 22:24:10 cd conf/distro; cp include/angstrom-2007-for-openmoko.inc angstrom-2007.1.conf to get in in .dev Jan 25 22:24:46 d Jan 25 22:26:15 copy that include file over the conf file? Jan 25 22:26:25 and that should make it pick the right stuff for me? Jan 25 22:31:34 re Jan 25 22:32:30 wvdschel: yes. Jan 25 22:36:03 yey Jan 25 22:57:22 zecke: Still there? Jan 25 22:57:59 hvontres|work: I'm not entirely sure about the changes you mean... Jan 25 22:58:02 RP: sort of Jan 25 22:58:13 zecke: time for training or another time? Jan 25 22:58:39 go ahead Jan 25 22:58:54 should not be more than a line :) Jan 25 22:59:07 I think gcc-native and gcc-cross would be two examples? Jan 25 22:59:36 zecke: cross has something to do with the target i.e. it touches target staging and needs to rebuilt for each target. native works for all targets Jan 25 23:00:26 zecke: You can think of it in terms of the gnu triplet - native has the same build, host and target, cross has a different target Jan 25 23:01:00 RP: okay, then your change is not correct. qmake is supposed to have all the mkspecs installed but I will look at marcin's initial bug report Jan 25 23:02:11 zecke: It stages files into CROSS_DIR which is target specific Jan 25 23:02:31 RP: then this is the bug :) Jan 25 23:02:43 zecke: if I bitbake qmake-native which qemuarm and then try and build something on qemux86 using qmake, it breaks Jan 25 23:02:53 zecke: right, this is what I wondered :) Jan 25 23:03:19 s/which/with/ Jan 25 23:03:20 RP: one qmake-native should be able to do the job to build Makefiles for s390/uclibc and mips/eglibc Jan 25 23:03:33 zecke: right, it can't :( Jan 25 23:04:47 RP: let me take a look at marcin's bug report tomorror or on sunday :) Jan 25 23:05:17 zecke: The problem is the staging into CROSS_DIR. If its really a native package it shouldn't need to touch that Jan 25 23:07:11 * stefan_schmidt thinks zecke should make himself a nice saturday. Old people need to rest at the weekend Jan 25 23:10:22 stefan_schmidt: zecke isn't old ;-) Jan 25 23:10:42 not that I have any objection to zecke taking a rest mind... Jan 25 23:10:42 RP: qtopia puts years on you Jan 25 23:10:42 RP: 11 minutes ago he at least got one year older. ;) Jan 25 23:10:52 RP: he knows better (slight advantage to be in #opie.de) Jan 25 23:11:09 zecke: Many happy returns ;-) Jan 25 23:11:18 which is the OpenMoko Devels Germany channel Jan 25 23:11:37 I doubt I'd understand a word in there ;-) Jan 25 23:11:45 zecke: The nice channel name makes it a nice secret. :) Jan 25 23:12:01 RP: heh, better ranting for us. ;) Jan 25 23:21:43 * RP wishes he could remember more from school at times like this... Jan 25 23:22:33 ya, if you don't practice, it slowly vanishes Jan 25 23:22:38 goes for a lot of knowledge areas Jan 25 23:22:59 there are some things you don't forget but yes Jan 25 23:23:21 I was taught German then never met a German or used the language for 10 years :/ Jan 25 23:25:15 RP: same here Jan 25 23:25:34 then suddenly they are everywhere :-) Jan 25 23:26:13 It was strange landing in a German airport 3 years ago and realising I knew what an ausgang was :) Jan 25 23:27:15 its the same with me and Russian ;) Jan 25 23:27:57 but there aren't much Russian speaking people in FOSS Jan 25 23:29:37 thesing_: I wouldn't have the slightest clue with Russian ;-) Jan 25 23:30:23 RP: compared to german its an easy language. Jan 25 23:31:39 RP: what do you think about http://www.pastebin.ca/872749 Jan 25 23:31:48 thesing_: I find that hard to believe somehow. I found German kind of similar to English, I could see the common roots for both... Jan 25 23:32:00 RP: this runs update-modules at image generation Jan 25 23:34:12 thesing_: It looks good to me at a quick glance Jan 25 23:34:47 RP: once you get used to the russian alphabet its not that hard. Jan 25 23:35:18 RP: should I open a bug for review? Jan 25 23:35:25 thesing_: yes please Jan 25 23:35:48 RP: collie does not work atm because the unix domain sockets aren't build into the kernel. Jan 25 23:36:02 and so the device hangs at boot. Jan 25 23:36:06 thesing_: :/ Jan 25 23:36:15 thesing_: We have real space problems on collie Jan 25 23:36:25 I'm trying to see if we can find ways to save some... Jan 25 23:36:47 Preferably ways we can merge back upstream.... Jan 25 23:37:05 with modutils.sh moved before udev it should work but for some reason the unix module doen't get included in images. Jan 25 23:37:30 A RRECOMMENDS for it will be missing... Jan 25 23:37:41 what do I have to rebuild when I change MACHINE_EXTRA_RRECOMMENDS and module_autoload? Jan 25 23:37:56 thesing_: probably task-base Jan 25 23:45:14 RP: is there a way to see if a module will end up in an image without building the image? I tried bitbake -g but this only shows tasks not packages. Jan 25 23:45:48 thesing_: Not really, no... Jan 25 23:46:00 thesing_: Its up to the package manager, not bitbake Jan 25 23:46:11 You can look at the Packages file to have an idea Jan 25 23:47:45 the update-modules patch is bug 3714 Jan 25 23:48:24 RP: do you know how koen generated the image on his blog? Jan 25 23:50:03 RP: this image http://dominion.thruhere.net/koen/images/angstrom-minimal-image-graph-nq8-small.png Jan 25 23:50:32 RP: it looks like some stuff got re-named. I'll bang on it some more on the way home and post any patches on bugzilla Jan 25 23:52:47 thesing_: I don't... Jan 25 23:53:08 hvontres|work: ok, thanks. I'll try and have a look if/when I get a chance Jan 25 23:54:04 * * OE Bug 3714 has been created by thommycheck(AT)gmx.de Jan 25 23:54:06 * * run update-modules at image generation time Jan 25 23:54:08 * * http://bugs.openembedded.net/show_bug.cgi?id=3714 Jan 25 23:54:53 time to go home. Have a good weekend Jan 25 23:56:48 RP: you were right I only had to rebuild task base. Maybe config should have revisions too. I tried to figure this out for several hours. Jan 26 00:01:32 thesing_: sadly its not quite as simpler problem as that :/ Jan 26 00:02:54 RP: I know. Its hard to decide what needs to be rebuild as it depends on the change one does. Jan 26 00:04:22 RP: but I think one could come up with some rules. For example rebuild task-base if MACHINE_EXTRA_RRECOMMENDS was changed. Jan 26 00:05:00 thesing_: It would be good to document that kind of thing Jan 26 00:07:44 RP: yes. Maybe something like "common pitfalls" ? I will write an email to the ml. Jan 26 00:10:09 RP: do you remember me? Jan 26 00:11:26 scant: remind me ;-) Jan 26 00:11:54 RP: you helped me out the other day, with mpd_svn.bb Jan 26 00:12:34 scant: ah, yes :) Jan 26 00:13:09 RP: i'm trying to create a bb file from scratch, and i'm in desperate need of some help Jan 26 00:13:32 scant: Here would be a good place to ask Jan 26 00:15:07 RP: what would you think about something like MACHINE_MODULES wich would contain modules which are RRECOMMENDed and autoloaded? ATM one has to add modules to MACHINE_EXTRA_RRECOMMENDS in machine config and to module_autoload in the kernel recipe. Jan 26 00:15:46 RP: of course this is only important for machines with size limited kernel. Other machines would build those things into the kernel. Jan 26 00:15:58 thesing_: In general these should be added to task-base rather than being machine specific Jan 26 00:17:03 RP: the kernel modules I added are all machine specific (keyboard, touchscreen, leds, mmc, pcmcia) Jan 26 00:17:32 thesing_: Not domain sockets though... Jan 26 00:19:28 RP: no but that is the easiest way to test if this helps. At least so I thought this morning... Jan 26 00:21:20 RP: btw. please add linux-rp_2.6.24 soon. it takes ages to apply the rc patch to 2.6.23 here ;) Jan 26 00:22:22 thesing_: I think its been added ;-) Jan 26 00:28:30 hmm, maybe not... Jan 26 00:28:40 if not, it will be soon ;-) Jan 26 00:30:17 RP: thanks. Jan 26 00:36:47 ok time to get some sleep. Jan 26 00:36:50 bye Jan 26 00:39:05 hey RP, u still there? Jan 26 00:44:56 scant: kind of Jan 26 00:45:00 RP: http://pastebin.ca/872541 Jan 26 00:45:07 RP: http://pastebin.ca/872536 Jan 26 00:46:06 scant: My guess would be S isn't correct Jan 26 00:46:18 hi Jan 26 00:46:27 scant: Its not finding anything to configure/compile Jan 26 00:46:30 hi timtimred Jan 26 00:46:36 RP: can you help me fix it? Jan 26 00:46:38 a question... :) Jan 26 00:46:45 NOTE: consider defining PREFERRED_PROVIDER_virtual/arm-angstrom-linux-uclibcgnueabi-libc-for-gcc Jan 26 00:46:50 its set.... Jan 26 00:47:01 known issue ? Jan 26 00:47:32 oh my distro is a heavily bastardised and updated angstrom, not a vanilla one, but hey this shouldnt make a difference Jan 26 00:47:44 oops 2007 Jan 26 00:48:02 it gos off trying to build an external toolchain Jan 26 00:48:34 timtimred: Did you pull in the change to the external toolchain from OE.dev? Jan 26 00:48:47 ohh possibly not to external toolchain... Jan 26 00:49:03 let me go do that, thx :) Jan 26 00:51:01 RP: can you step me through it? Jan 26 00:52:10 scant: I'm trying to build that now ;-) Jan 26 00:52:28 RP: this bb file isn't in the repository, i created it myself Jan 26 00:53:18 unfortunately it does the same thing. Jan 26 00:53:19 RP: the make should produce: libsane-pixma.so, which needs to be renamed to: libsane-pixma.so.1 and moved here: /usr/lib/sane Jan 26 00:53:53 RP: i think rwhitby helped me determine that I need to do a do_install to do this renaming and copying of the lib Jan 26 00:54:28 scant: The problem is that S is set incorrectly Jan 26 00:54:44 if you download that tarball, what directory does it contain? Jan 26 00:55:12 RP: mp150-0.13.1 Jan 26 00:55:38 scant: yet the .bb says S = "${WORKDIR}/something else" ;-) Jan 26 00:58:46 RP: what about the rest of it? Jan 26 00:59:01 well, fix that and start from there Jan 26 01:01:13 scant: Fix that, then rebuild and you should see something different happen Jan 26 01:03:13 RP: is this correct: S = "${WORKDIR}/mp150-${PV}" Jan 26 01:03:29 scant: I think si Jan 26 01:03:31 so Jan 26 01:04:04 RP: is it the case that we can override the 'make install' with do_install? Jan 26 01:04:30 scant: yes Jan 26 01:04:46 RP: cool Jan 26 01:04:57 my bitbake cache is rebuilding, about to rebuild Jan 26 01:07:55 RP: http://pastebin.ca/872825 Jan 26 01:08:27 scant: So a header is missing and it can't compile Jan 26 01:08:34 incorrect DEPENDS? Jan 26 01:09:54 RP: it depends on libjpeg and libtiff - I THINK Jan 26 01:10:01 RP: do i just add that to the depends line? Jan 26 01:10:32 scant: yes Jan 26 01:12:36 RP: i added: DEPENDS = "libtiff" Jan 26 01:13:58 RP: Missing or unbuildable dependency chain was: ['libsane-pixma', 'libtiff'] Jan 26 01:14:58 does that mean libtiff isn't available for openslug? Jan 26 01:15:12 the .bb file is called tiff, not libtiff Jan 26 01:25:04 RP: http://pastebin.ca/872841 Jan 26 01:26:38 scant: Looks like it wants kernel headers... Jan 26 01:26:58 RP: what do I do? Jan 26 01:27:45 I'd be tempted to see if it really needs that file and comment out that include Jan 26 01:28:18 RP: all i can tell you is that this is version 0.13 and 0.11 is included in sane-backends which compiles fine via bb on a cross dev env Jan 26 01:28:59 Like I said try commenting out the include Jan 26 01:29:59 RP: i commented it out, rebuilding Jan 26 01:29:59 This is all basic cross compiling though, you just need to iterate over this process of solve problem, move forward, solve problem... Jan 26 01:30:19 RP: yes, thank you for stepping through it with me, i REALLY appreciate it Jan 26 01:30:36 scant: I need to sleep now I'm afraid... Jan 26 01:31:24 oh Jan 26 01:31:31 cause i think i'm almost there Jan 26 01:31:34 http://pastebin.ca/872847 Jan 26 01:32:15 What does the install log say? Jan 26 01:32:21 RP: http://pastebin.ca/872850 Jan 26 01:32:26 and what did you set the install function to? Jan 26 01:33:06 Looks like you need to write a do_install function Jan 26 01:33:20 RP: do_stage(){ autotools_stage_all } Jan 26 01:33:43 RP: I need to prevent install.sh from running i think Jan 26 01:34:39 can anyone here help me write the do_install function? Jan 26 01:34:44 scant: do_stage and do_install are different things Jan 26 01:34:58 scant: Have a look at other .bb files for examples Jan 26 01:35:07 and on that note, 'night! Jan 26 01:35:11 night Jan 26 01:35:51 g'night RP Jan 26 01:35:53 good idea, btw. Jan 26 01:47:18 does do_install replace a 'make install' and does do_install_append execute after 'make install'? Jan 26 02:10:04 * * OE Bug 2317 has been RESOLVED (FIXED) by Jan 26 02:10:06 * * incorrect use of FILES_${PN}-dbg Jan 26 02:10:09 * * http://bugs.openembedded.org/show_bug.cgi?id=2317 Jan 26 02:25:02 03Laibsch 07org.oe.dev * r10d83984... 10/ (44 files in 33 dirs): divers: replace '=' definition of dbg packages with '+=' for 43 files. fixes 2317. **** ENDING LOGGING AT Sat Jan 26 02:59:57 2008