**** BEGIN LOGGING AT Thu Jan 25 02:59:57 2007 Jan 25 04:32:24 morning Jan 25 04:33:29 morning XorA Jan 25 04:35:02 hey rwhitby Im working on korean time just for this morning :-) Jan 25 04:35:18 something special about today? Jan 25 04:35:33 conference call with Samsung Jan 25 04:52:20 rwhitby: re your pivot, collie in OE uses altboot to achive the same thing Jan 25 04:53:35 can altboot work with no GUI or console? Jan 25 04:53:43 rwhitby: no Jan 25 04:54:07 what is the backend technology it uses to change the boot? Jan 25 04:54:23 rwhitby: I think it pivots Jan 25 04:55:02 collie is in similar problems, only 14M of flash Jan 25 04:55:47 can it work with only a serial console for configuration, and no user input when it boots? Jan 25 04:56:53 rwhitby: I think it can, by default on collie of course it uses tty0 Jan 25 04:59:27 XorA: ok, I'll look into it. Jan 25 05:48:26 03lenehan 07org.oe.dev * rf3a12e85... 10/ (1 packages/arpwatch/arpwatch_2.1a15.bb): (log message trimmed) Jan 25 05:48:26 arpwatch 2.1a15: Run do_install under fakeroot because the Makefile Jan 25 05:48:26 attempts to set the ownership of the installed files (install Jan 25 05:48:26 -o/-g): Jan 25 05:48:26 | NOTE: make install DESTDIR=/home/lenehan/devel/oe/build/titan-glibc-25/tmp/work/arpwatch-2.1a15-r1/image Jan 25 05:48:27 | /usr/bin/install -c -m 555 -o bin -g bin arpwatch /home/lenehan/devel/oe/build/titan-glibc-25/tmp/work/arpwatch-2.1a15-r1/image/usr/sbin Jan 25 05:48:32 | /usr/bin/install: cannot change ownership of /home/lenehan/devel/oe/build/titan-glibc-25/tmp/work/arpwatch-2.1a15-r1/image/usr/sbin/arpwatch': Operation not permitted Jan 25 05:59:36 03lenehan 07org.oe.dev * rd1066bdb... 10/ (4 files in 2 dirs): Jan 25 05:59:36 quagga: Don't -RDEPENDS on ncurses and readline - the shlibs code Jan 25 05:59:36 adds those dependencies automatically. Jan 25 06:41:07 03lenehan 07org.oe.dev * rde7f242d... 10/ (5 files in 3 dirs): Jan 25 06:41:07 quagga: Modify the package to use the volatiles system to create the Jan 25 06:41:07 /var/run/quagga directory so that it'll exist after reboot on systems Jan 25 06:41:07 that use a tmpfs /var. Jan 25 06:56:11 koen|away: pulse isn't working for me....I copied a few recipes into .oz354x and built pulse....and it refuses to start. keep aborting. Jan 25 08:43:39 ~lart AMD Jan 25 08:43:39 * ibot slaps AMD around with a large trout Jan 25 09:17:21 re Jan 25 09:25:22 i am having again the problem that angstrom is not building the compulab machine, where did i add that machine last time Jan 25 09:27:17 ah my local.conf was old Jan 25 10:26:53 morning all Jan 25 10:27:05 HI Liam Jan 25 10:27:14 and morning all :) Jan 25 10:27:14 hey Richard Jan 25 10:27:57 * lrg wonders if XorA remembered to wake up at 5.30am for conference call Jan 25 10:32:04 hehe Jan 25 10:32:47 lrg: yup Jan 25 10:32:54 yay Jan 25 10:33:07 hey XorA Jan 25 10:43:45 * RP doesn't envy Xora :) Jan 25 10:46:05 good thing my wife loves samsung kit :-) Jan 25 10:55:28 Is it possible to use autoconf later than 2.59 on a case-by-case basis? Later versions seem to break a lot of stuff but I need at least 2.60b for one package I'd like to build. DEPENDS += "autoconf-native (>= 2.60b)" didn't seem to cut it. Jan 25 10:56:08 morning all Jan 25 10:57:23 hey ade|desk Jan 25 10:58:03 morning guys Jan 25 10:58:30 Laibsch: i don't see how. might ask pb_ and/or on the mailing list Jan 25 11:00:36 hi all Jan 25 11:01:00 :-/ Jan 25 11:08:26 hi mickeyl , florian Jan 25 11:08:40 hi RP Jan 25 11:22:50 cheers RP Jan 25 11:28:35 morning Jan 25 11:29:14 hi chouimat Jan 25 11:57:16 how do i genreate he Packages.gz again ?? Jan 25 11:59:24 rob_w: bitbake package-index Jan 25 12:01:15 thanks cbrake Jan 25 12:01:49 btw cbrake the optrex we where talking last week was broken indeed .. now i ve a working .. maybe i killed it when i did put the cables wrong in the first place Jan 25 12:02:40 rob_w: sounds likely -- glad to hear it is working. Jan 25 12:44:26 hi, all! Jan 25 12:44:50 is ARM thumb build is supported by OE? Jan 25 12:47:54 slapin_nb: supported is a tough question. there is ucslugc which is suing uclibc and thumb Jan 25 12:48:21 slapin_nb: depending on what uclibc gets build you might succeed or fail Jan 25 12:48:52 zecke_, I think more about glibc ATM Jan 25 12:49:13 zecke_, and Angstrom Jan 25 12:51:29 slapin_nb, there's a config vars to select thumb. Apparently, it wasn't tested on a full rootfs. You're welcome to test it - I know plam devices in a clear need for it. Jan 25 12:58:59 hi zecke_ Jan 25 13:04:30 psokolovsky_, any details/links? Jan 25 13:05:19 slapin_nb, read oe-devel ML archives for last 1.5months Jan 25 13:16:04 well as I see it is via ARM_INSTRICTION_SET="thumb" in local.conf. but how could I set some package to be built using "arm" instruction set? Jan 25 13:22:57 slapin_nb, you read archive? I at least raised that question there, and I belive Koen though about that. (Besides, generic answer is simple - override that on package level) Jan 25 13:31:12 psokolovsky_, I read my local mail, 23K messages left... so if you remember point or subject line, that'd be cool Jan 25 13:33:40 slapin_nb, "thumb"? Jan 25 13:34:17 "[oe] [RFC] Enable tune-thumb.inc for all armv5t and armv6t machines" Jan 25 13:36:06 psokolovsky_, thanks! Jan 25 13:44:23 re Jan 25 13:46:11 what tells me a sefault in g_atomic_int_get () Jan 25 13:46:52 rob_w: I was not following the discussion, just returned; but if you are talking about probolems with the atomic stuff, from glib, right? Jan 25 13:47:21 we had problems and needed atomic code in our project, so we looked at glib and it simply did not work on some architectures Jan 25 13:48:19 Jin^eLD, hmm tell me more Jan 25 13:48:23 Jin^eLD: which architectures and version of glibc? Jan 25 13:48:39 cbrake: not glibc, the glib, this gnome library that has atomic code Jan 25 13:48:48 I think its called glib, isnt it? Jan 25 13:48:55 cbrake, compulab libc-2.3.90.so Jan 25 13:49:09 rob_w: well, we needed atomic increment and decrement functions, so we were looking at libraries that may provide that Jan 25 13:49:21 Jin^eLD: right, but I'm still curious which version of glibc you are running underneath glib. Jan 25 13:49:37 someone pointed us to glib Jan 25 13:49:42 cbrake: oh I see, sorry :) Jan 25 13:49:55 so we wrote a test program and tried it on ARM Jan 25 13:49:59 it tells me it segfaults in usr/lib/glib-2.0 Jan 25 13:50:27 basically we had two threads, that tried to inc/dec a variable at the same time Jan 25 13:50:42 while it worked correctly on PC, using glib atomic operations, it failed miserably on ARM Jan 25 13:50:54 after looking at the glib code we could not find anything ARM related, so its just not there I guess... Jan 25 13:51:03 so oyu think i do this also by accident ? Jan 25 13:51:24 rob_w: what arch are you on? Jan 25 13:51:38 the app works on x86 but fails on pxa Jan 25 13:51:53 and the app uses glib? Jan 25 13:52:02 well, could be the same problem Jan 25 13:52:11 glib does not seem to have atomic stuff for other platforms Jan 25 13:52:17 at least they should have use a pthread locking fallback Jan 25 13:52:56 i dont plan to use atomic .. so i did this by accident .. what code do i need to look at ? Jan 25 13:53:24 I do not remember exactly anymore, gatomic.h if I recall correctlyl Jan 25 13:53:50 rob_w: set a breakpoint there and do bt? Jan 25 13:54:04 rob_w: at least thats what we were looking at Jan 25 13:54:58 bt ? Jan 25 13:55:09 backtrace (shortcut) Jan 25 13:56:19 ah right Jan 25 13:56:39 how can tell gdb where the sources are so it can look up the code ?? Jan 25 13:56:53 rob_w: use gdbserver? Jan 25 13:57:05 rob_w: install the dbg packages, you only want to have a stacj trace Jan 25 13:58:15 ok will try Jan 25 13:59:49 ahhh i cracked a path Jan 25 14:03:09 it cracks when loading images .. and i dont have any gdk-pixbuf-loader installed Jan 25 14:07:30 hey Jan 25 14:07:42 anyone know where to get replacement styluses? Jan 25 14:50:59 maybe a stupid question, but how could I build kernel image not at /boot/zImage-xxx, but (or in addition to) just as additional image? Jan 25 14:51:37 slapin_nb, that's how it builds anyway Jan 25 14:52:31 psokolovsky_, yeah, but I'm interested in rsync-based updates to images on my site, but to boot on palm you need separate kernel... Jan 25 14:52:56 psokolovsky_, so is it possible oe-wise? Jan 25 14:53:30 or I just need to take it from rootfs dir and forget about it? Jan 25 14:54:11 slapin_nb, I'd say, better leave it where it is and forget about ;-) Jan 25 14:55:07 psokolovsky_, I need to pote it separately for people to be able to mindlessly put stuff on their flash and have it working Jan 25 14:55:42 s/pote/put Jan 25 14:56:17 slapin_nb, kernel image is always built separately, in addition to be packaged, to later go into rootfs. Jan 25 14:56:48 $ ls -l tmp/deploy/glibc/images/h4000/ Jan 25 14:56:54 -rw-r--r-- 1 pfalcon pfalcon 1152812 Jan 22 04:11 zImage-2.6.19-hh8-h4000-20070122035702 Jan 25 15:01:59 psokolovsky_, only difference I see is not definind PREFERRED_PROVIDER_virtual/kernel-image... Jan 25 15:02:26 slapin_nb, difference for what? Jan 25 15:02:47 psokolovsky_, for machine conf Jan 25 15:03:14 slapin_nb, how that relates to what we've talked about 5 secs ago? Jan 25 15:04:04 psokolovsky_, so it probably is related to zImage-xxx absent in my images dir. Jan 25 15:04:18 slapin_nb: There was some dicussion about this on the mailing list recently, I suggest reading the archives Jan 25 15:04:42 (about kernels in /boot that is) Jan 25 15:04:53 slapin_nb, you never told it is absent in your images dir Jan 25 15:05:37 psokolovsky_, it was actually my problem :) probably I explained it in a bid weird way... Jan 25 15:05:45 If its absent from the images dir, the kernel .bb is probably missing a deploy task Jan 25 15:06:31 RP, thanks! I'll look into it then... Jan 25 15:08:07 slapin_nb, yep, please just don't assume I can read you mind ;-) I may imagine you don't have it there, but as well I can imagine you're looking in wrong place ;-). So, where do you look for it? Jan 25 15:10:14 psokolovsky_, in same place as all images, as you did. Jan 25 15:10:40 slapin_nb, can you pastebin ls -l of it? Jan 25 15:13:18 psokolovsky_, http://ossfans.org/paste?id=15266d Jan 25 15:15:15 slapin_nb, ok, as RP said, linux-hackndev misses deploy task Jan 25 15:15:57 psokolovsky_, I'm preparing new version for zire72 and linux-hackndev Jan 25 15:16:10 slapin_nb, ok Jan 25 15:21:07 Hi all, anyone had a chance to look at bug 1815? Or can anyone give me clue as to how to fix? Jan 25 15:47:52 hrw|gone: Hi! I've just got a link to a document Marvell published about their WLAN chip, it's the "Firmware specification 5.1" and quite nice. Jan 25 15:48:00 hrw|gone: I've put the link to it into the README Jan 25 15:57:08 morning Jan 25 15:57:11 schurig: thx Jan 25 15:57:21 schurig: I like that 'CONFIDENTIAL' sign on each page ;D Jan 25 16:05:08 hi all Jan 25 16:06:10 psokolovsky_, could you look at bug 1818 (Zire 72 related)? Jan 25 16:09:04 could anybody look at #1818 please? it is about replacement of 2 small files Jan 25 16:09:15 it is about Palm Zire 72 only Jan 25 16:10:21 hrw: i don't care for this sign, because an employee of Marvell posted it to a public mailing list Jan 25 16:10:31 :-) Jan 25 16:10:49 slapin_nb, I'll do as soon as I can Jan 25 16:11:25 psokolovsky_, it is just notification :) Jan 25 16:14:10 schurig: ;) Jan 25 16:21:42 I'm trying to build the ep93xx distro but I don't know what it's output is can anyone help? Jan 25 16:22:35 It creates zImage-2.6.17+2.6.18-rc1-ep93xx-20070124224810 which is great but I also need a ramdisk don't I? Jan 25 16:22:54 I'm a newbie to oe btw Jan 25 16:27:36 please someone give me a nod, I'm fairly new to IRC discussions and want to make sure I'm connected properly Jan 25 16:27:39 hi mickeyl hvontres|poodle Jan 25 16:27:49 NZG: bitbake bootstrap-image Jan 25 16:28:13 thank you Jan 25 16:28:20 I'll try that Jan 25 16:28:48 hrw: morning :) Jan 25 16:29:16 yo hrw Jan 25 16:29:27 hey schurig, long time no see Jan 25 16:29:31 hrw: I tested gpsd 2.34 on my 5600 using gpsfake and it works. :) Jan 25 16:29:34 schurig: everything alright ? Jan 25 16:29:48 mickeyl: morning, almost Dr. mickeyl :) Jan 25 16:30:24 hvontres|poodle: on OEDEM'06 mickeyl told that according to German law we can call him Dr. but he cannot Jan 25 16:30:48 he can't call himself Dr.? Jan 25 16:31:21 raduga: he is not Dr. yet Jan 25 16:31:21 well, actually it's like that Jan 25 16:31:36 i guess, it would be legal for us to give him a shirt that says "Dr. Mickeyl", but not legal for him to wear it? Jan 25 16:31:37 until the defence, i need to say "no, sorry, you are not allowed to call me Dr." Jan 25 16:31:54 after the defence before the publication i don't need to say "no sorry,...", but i am not yet allowed to call myself Dr. Jan 25 16:31:54 mickeyl: ok. I won't call you Dr. Jan 25 16:31:57 until the defense :) Jan 25 16:32:07 after the publication, I am allowed to say "My name is Lauer. Dr. Lauer " Jan 25 16:32:08 heh Jan 25 16:32:12 strange german laws :) Jan 25 16:32:29 I suspect it would not be legal for me to say the same. Jan 25 16:32:32 mickeyl: heh... very james bond..:) Jan 25 16:32:35 ;) Jan 25 16:32:36 Since I'm neither a Dr. nor named Lauer. Jan 25 16:35:04 hrw: if you get a chance, could you please push the fixes in #1806? Thanks :) Jan 25 16:35:18 hvontres|poodle: ok Jan 25 16:35:43 * hvontres|poodle is happ to have finally fixed a bug himself :) Jan 25 16:35:44 if my umbaumba 17.3 build is working Jan 25 16:36:22 hrw: so where can I get an umbaumba 17.3 image ? :) Jan 25 16:36:37 hvontres|poodle: nowhere - its my private distro Jan 25 16:37:16 hrw: heh :) Jan 25 16:41:05 I have some newbie oe questions, is this the right place to ask them? Jan 25 16:41:22 NZG: a, you are here as well, sure go ahead Jan 25 16:41:39 NZG: ask questions and someone might be able to answer Jan 25 16:42:06 hi mickeyl Jan 25 16:42:15 cheers pb__ Jan 25 16:42:22 hvontres|poodle: testing will take lot of time - need to build whole opie etc Jan 25 16:42:27 pb_: could you do me a favour, if you have a bit of time (no hurry) Jan 25 16:42:39 sure Jan 25 16:42:39 NZG: have you followed GettingStarted? Jan 25 16:42:40 essentially I want to know about the build process Jan 25 16:42:46 o yes, I'm beyond that Jan 25 16:42:55 I'm making a bootstrap image now Jan 25 16:42:55 pb_: it would be very helpful, if you had a chance to port your "integer-ops-for-pixmap-engine" patch for both gtk 2.6.10 and 2.10.6 Jan 25 16:43:03 NZG: okay, then try to ask a concrete question Jan 25 16:43:09 kk Jan 25 16:43:20 mickeyl: ah right, yeah. I keep meaning to look at that. Jan 25 16:43:49 cool. Jan 25 16:45:22 I just built task-base for the ep93xx, now I'm building bootstrap-image, it appears to be rebuilding from scratch. Is this the desired behavior or have I done something wrong? Jan 25 16:46:06 NZG: it is not desired, it should just build kernel, and tools to create the the image Jan 25 16:46:18 NZG: how do you determine that it is building from scratch? Jan 25 16:46:45 It's first step was to download and build ipkg Jan 25 16:46:54 I'm assuming that was already done for task-base Jan 25 16:48:03 please for give and correct any IRC protocol blunders by the way, newbie Jan 25 16:49:25 build just completed, it wasn't building from scratch, seems strange it downloaded ipkg though Jan 25 16:49:27 are you sure you started bitbake in the same directory? Jan 25 16:49:41 besides, there may be different versions for ipkg-native and ipkg-cross Jan 25 16:49:47 so i wouldn't be surprised Jan 25 16:49:48 ah Jan 25 16:50:20 NZG: there are two things ipkg and ipkg-native :) Jan 25 16:50:29 NZG: ipkg-native is used to create the packages Jan 25 16:50:38 NZG: ipkg is installed to the image and executed on start Jan 25 16:51:39 that makes sense, ok good Jan 25 16:51:50 now, I have 3 bootstrap images and the zImage Jan 25 16:52:05 Here comes the newbness Jan 25 16:52:22 are these images usable for creating a ramdisk for the ep93xx? Jan 25 16:52:24 I can't comment on the '3' but one could be jffs2, one an ext2 and the other just a tar.gz of the files? Jan 25 16:52:45 zecke_:yes Jan 25 16:53:04 NZG: these images should be bootable (well I don#t have too much ep93xx knowledge) Jan 25 16:53:30 I ask because I'm using redboot, which wants to use it's exec command to start a kernel and point at a ramdisk for it Jan 25 16:53:57 I'll try just handing it the ext2 and see if it chokes, brb Jan 25 16:54:06 NZG: it will choke Jan 25 16:54:18 zecke_:o? Jan 25 16:54:23 NZG: these images are generally meant to be flashed or used on a harddisk Jan 25 16:54:56 zecke:alright Jan 25 16:55:11 zecke:well, I'll try to build a ramdisk from the tarball then Jan 25 16:55:27 NZG: don't you have a disk or flash? Jan 25 16:55:45 zecke_:I have a little flash Jan 25 16:56:00 aynway the booting process of the ep93xx is beyond my knowledge, so good lucj Jan 25 16:56:14 thank you Jan 25 16:59:14 pb__: mythtv is your baby - right? Jan 25 17:01:12 pb__: http://pastebin.ca/328262 - attempt to get mythtv recipes share common stuff in one file. also should remove message from bitbake about 'mythtv_arch already seen' Jan 25 17:03:56 hrw: looks good to me, go for it. Jan 25 17:04:04 [if it builds] Jan 25 17:05:41 mickeyl: getting it build would take me few hours - thats why I'm asking Jan 25 17:05:52 ah Jan 25 17:06:04 go commit and let me check later tonite Jan 25 17:11:08 hrw: mythtv is mickeyl's baby really, I just abduct it from time to time Jan 25 17:11:19 ok Jan 25 17:11:59 that patch does look basically ok to me, though Jan 25 17:12:09 you might want to change that RDEPENDS to RDEPENDS_${PN} while you're at it Jan 25 17:12:43 I guess we should train zecke to identify and destroy the .bb files that contain "RDEPENDS = '...'" type constructs. They're all wrong. Jan 25 17:13:12 pb__: insane.bbclass is trying that Jan 25 17:13:20 excellent Jan 25 17:13:20 pb__: and I always like destroy and punishment Jan 25 17:13:35 indeed, you were the first man I thought of Jan 25 17:13:36 pb__: I'm confused, bittest is looking at RDEPENDS Jan 25 17:17:41 ah right Jan 25 17:20:25 03hrw 07org.oe.dev * re84fca35... 10/ (4 files in 2 dirs): mythtv: unified, need testing Jan 25 17:20:25 03hrw 07org.oe.dev * ra168d537... 10/ (1 packages/keymaps/keymaps_1.0.bb): keymaps: fix typo so htcuniversal will get keymap working Jan 25 17:23:35 hrw, can you commit #1367? Jan 25 17:25:38 psokolovsky: probably can Jan 25 17:26:08 hrw, please do, it's tested to work ok Jan 25 17:32:04 03pb 07org.oe.dev * r0e8399ae... 10/ (1 packages/tasks/task-mythfront.bb): task-mythfront: need more rdepends Jan 25 17:32:08 03pb 07org.oe.dev * ree5fc760... 10/ (4 files in 2 dirs): xserver-xorg: fewer rdepends, rprovides Jan 25 17:51:34 could somebody look at 1819 - I added deploy task to linux-hackndev-2.6 Jan 25 18:15:17 slapin_nb: for future: attach diffs not new files Jan 25 18:17:52 this has puzzled me for a while. bitbake says "multiple providers are available ..." and "consider defining PREFERRED_PROVIDER ..." when building my project. Now, it apprently made some choice as to a provider and the project is building ok. How can I find out what that provider choice was? Jan 25 18:18:23 it is for "glibc-intermediate", I have no ipk file by any such name.. Jan 25 18:18:57 so, how can I determine what entry to make in the .bb file for the PREFERRED_PROVIDER ? Jan 25 18:20:30 hi Jan 25 18:20:51 T0mW: set it in the conf/local.conf and it lists the possible ones Jan 25 18:21:12 looking... Jan 25 18:21:50 well, yeah, I know how to set it, but what do I set it for? Jan 25 18:22:11 bitbake made the determination to use some version, but how do I discover what version it used? Jan 25 18:22:30 03Sergey 07org.oe.dev * rb047443e... 10/ (1 packages/linux/linux-hackndev-2.6_svn.bb): linux-hackndev: added do_deploy - close #1819 Jan 25 18:22:30 T0mW: the one it starts building was used Jan 25 18:23:08 isn't there a listing of some sort that shows what packages were used? Jan 25 18:23:25 T0mW: bitbake --help Jan 25 18:23:38 did that, nothing in there for this Jan 25 18:23:40 T0mW: actually I don't remember if it lists virtual/foo packages Jan 25 18:24:02 T0mW: but bitbake printed all the information you need Jan 25 18:24:06 I tried "print providers" and "match provider X" Jan 25 18:24:31 T0mW: it listed the available providers, and picked one and immediately built it Jan 25 18:24:35 "print providers" gives a list of everything available, but not what was used. Jan 25 18:24:57 T0mW: the one that was used is the build package, is that too obvious Jan 25 18:25:39 Multiple Providers consider definig... List (a,b,c) Jan 25 18:25:41 Building c Jan 25 18:26:13 sigh, he didn't understand the question. Jan 25 18:27:29 so, I guess the only way to know the package version is to tee the bitbake process, then examine the log of what it did. Jan 25 18:27:36 the brute-force method. Jan 25 18:55:09 T0mW: try grepping the OE tree for PROVIDES to see what all provides a virtual package. The bb file name does not have to match. Jan 25 18:57:13 well, the bitbake gives that warning when it builds my project. I just thought that I'd better define a version I wanted. And, since the version *it* chose was okay with me, I wanted to say "yes, use this one". I know that I am free to chose any old version at all... But, I wanted to use the *specific* version that bitbake defaulted to Jan 25 18:57:32 the problem is discovering which version bitbake chose Jan 25 18:59:02 as it stands right now, I'd have to do "bitbake bootzipit-image | tee the_log" and later grep "the_log" to see what version it decided upon. Then set my conf files to match it. Jan 25 18:59:52 I just thought that this was a common issue and all I needed to do was something like "bitbake -c show_me_what_you_did" Jan 25 19:00:26 err, "bitbake -c show_me_what_you_did bootzipit-image" Jan 25 19:01:20 T0mW: I would probably grep the OE tree for glibc-intermediate, understand what the options are, and then pick the correct option. Jan 25 19:01:47 yeah Jan 25 19:03:04 My project did build ok until I updated the OE tree, now I'm running into all sorts of issues that I did not have before. I am slogging my way through the new conf format trying to figure out what is now borked Jan 25 19:04:08 some sort of issue with the "*-image.bb" and "task-*.bb" files. It cannot find my task provider now. Jan 25 19:04:33 T0mW: the main issues w/ glibc are picking the correct version (2.3 or 2.5, at least for ARM). And if you are building 2.5, make sure you set it up for NPTL. Jan 25 19:04:48 NPTL? Jan 25 19:05:12 nicolous pitre threaded lib? Jan 25 19:05:54 T0mW: heh Jan 25 19:06:08 T0mW: http://www.openembedded.org/wiki/OeFaq Jan 25 19:06:19 T0mW: Native POSIX Thread Library Jan 25 19:06:20 alright , more reading on the wiki site then Jan 25 19:06:37 * cbrake still needs to finish that faq entry ... Jan 25 19:07:10 and spell NPTL correctly :-\ Jan 25 19:07:35 heh Jan 25 19:08:44 cbrake: thanks, one thing I've learned about OE is that it is never quick to get things done. You always have to spend time to get things accomplished. Jan 25 19:08:59 I fight more with OE than I do my source code Jan 25 19:10:23 T0mW: yeah, grep and the OE source tree are your friends -- I spend lots of time there :-) Jan 25 19:10:47 T0mW: "bitbake --dry-run --force --verbose " might help with what you're looking for. Jan 25 19:13:19 pb_: EXACTLY IT! thank you! Jan 25 19:13:28 * T0mW writes that down Jan 25 19:14:50 T0mW: write it down in the faq :-) Jan 25 19:22:43 cu Jan 25 19:35:11 T0mW: ok, NPTL section in the faq updated. http://www.openembedded.org/wiki/OeFaq Jan 25 19:44:49 cbrake: thanks for the update, and your article at bec-systems Jan 25 20:15:30 03justinp 07org.oe.dev * r04039f8e... 10/ (1 conf/distro/angstrom-2007.1.conf): angstrom: set PREFERRED_VERSION_freetype-native as well as PREFERRED_VERSION_freetype Jan 25 20:45:40 Hello Jan 25 20:45:52 Can someone help me with a output error? Jan 25 20:46:13 I'm trying to build firefox for SH3, and I'm having some problems Jan 25 20:46:43 output error? Jan 25 20:46:47 All versions give me the same problems Jan 25 20:46:54 Yes, half a sec please Jan 25 20:46:59 I'm putting it on pastebin Jan 25 20:48:03 I just need to know what to fix Jan 25 20:48:10 I'm not that good with C code Jan 25 20:48:15 http://pastebin.com/867426 Jan 25 20:49:06 There's a crapload of (non-fatal, obviously) errors with the header files Jan 25 20:49:09 I have a build problem too. Jan 25 20:49:26 Which, I guess, lead to the problem compiling the .c file Jan 25 20:50:27 I'm guessing it's not due to gcc, since I get the same thing with firefox 2 Jan 25 20:50:34 and 1.5 Jan 25 20:50:50 B_Lizzard: sorry, no idea. Looks like firefox doesn't understand your achitecture... Jan 25 20:51:05 Damn.... Jan 25 20:51:20 in my case, perl does not build with MACHINE="native" Jan 25 20:51:49 Well, the problem *is* part of nspr Jan 25 20:52:11 Which seems to be pretty low-level stuff Jan 25 20:55:19 On the other hand, how does this exist? Jan 25 20:55:20 http://rpm.sh-linux.org/rpm-fc5/RPMS/sh3/firefox-1.5.0.1-9.sh3.rpm Jan 25 20:55:25 :/ Jan 25 20:56:04 B_Lizzard: find the src.rpm package and see if they have any patches supplied with it :> just a guess.. but that might give you a hint Jan 25 20:56:17 Allright, will do Jan 25 21:28:43 I want to use open embedded to build my local kernel source (that I'm developing) into a build. Can this be done simply? Jan 25 21:29:02 What would you recommend? Jan 25 21:29:38 sure it can. you can either have it build directly in your local source tree, or copy it into its area and build from there Jan 25 21:30:54 Do I declare the source with file: ? Jan 25 21:32:31 i.e. SRC_URI = "file://linux-2.6.8.1-crus2.0.8-e1.0/*" Jan 25 21:32:34 you can, or you can write a custom do_unpack to copy the sources over yourself, or if you want to build them where they are, leave SRC_URI empty and point S= to the source tree Jan 25 21:32:56 ah! of course Jan 25 21:33:07 thx Jan 25 21:33:40 np Jan 25 21:44:12 hi Jan 25 21:45:28 i'm compiling oe. After build some packages for my embedded distribution, is possibile to test that the programs work ? Jan 25 21:45:53 like launching in emulation or.. don't know.. testing Jan 25 21:56:17 alvise: You can run certain things under qemu Jan 25 22:02:20 I found many patches that I can use Jan 25 22:02:30 I'm fairly sure I'll get firefox working Jan 25 22:02:52 Thanks, Jin^eLD Jan 25 22:03:03 And JustinP Jan 25 22:03:30 make sure to post the patches to bugzilla once you get them working Jan 25 22:03:37 Aye Jan 25 22:03:39 (or submit to OE if you have a key) Jan 25 22:03:48 Kristoffer has a key Jan 25 22:03:57 But i'd rather post a bug report Jan 25 22:07:14 :-) Jan 25 22:11:03 RP: well... I think i didn't understand what exactly OE does. It permits to compile packages for our device isn't it ? From any sources ? Jan 25 22:15:05 alvise: what's your question? Jan 25 22:15:32 yay Jan 25 22:15:34 * mickeyl dances a jig Jan 25 22:15:44 NOTE: package gtk+-2.6.10-r5: task do_build: completed Jan 25 22:15:48 on top of glib 2.12.3 Jan 25 22:15:52 *sigh of relief* Jan 25 22:16:10 mickeyl: yay :) Jan 25 22:16:28 alvise: That is one of the things it can do... Jan 25 22:16:30 "we" need eds 2007 Jan 25 22:16:43 which in turn needs the g_slice stuff Jan 25 22:16:57 but for now i can't upgrade to gtk+ 2.10 Jan 25 22:17:01 at least not until april Jan 25 22:17:13 * RP knows the problem all too well Jan 25 22:17:18 sure Jan 25 22:17:21 * mickeyl preaching to the choir Jan 25 22:17:24 sorry :) Jan 25 22:21:07 mickeyl: Its a source of frustration and has been for a while now :-/ Jan 25 22:21:44 It will get resolved eventually... Jan 25 22:22:35 by now i regained faith in cairo and gtk team to make 2.10 eventually usable Jan 25 22:22:39 i think they nearly have it Jan 25 22:22:47 what do you think? Jan 25 22:23:50 mickeyl: We ran some tests at Xmas and its ever improving but not there yet Jan 25 22:24:01 *nod* Jan 25 22:24:12 mickeyl: I've wanted to upgrade poky for a while but I'm not allowed ;-) Jan 25 22:24:15 * mickeyl tries building eds-20070120 against glib 2.12.3 and gtk+ 2.6.10 Jan 25 22:24:28 hehe, i can imagine Jan 25 22:24:45 I agree we can't until is performs better Jan 25 22:25:10 When it says "Error: pcrel too far", it's a -O problem, eh? Jan 25 22:25:16 Optimisation too high? Jan 25 22:26:02 RP: i have tons of performance problems with 2.6.10 alone. I can't imagine using anything even 5% slower Jan 25 22:26:19