**** BEGIN LOGGING AT Thu Sep 01 02:59:56 2005 Sep 01 02:59:58 Hello every one - as I just see the python update, I have encountered a strange python "bug ?". *** glibc detected *** python: free(): invalid pointer --- When I google for it I find references to a memory problem, has any one got an idea how to fix that ? Sep 01 03:00:44 hi Pigi Sep 01 03:01:05 hmmm Sep 01 03:01:24 oho: no Sep 01 03:01:26 Pigi: does ipkg handle "ipkg install hostap-modules-cs" where I have only hostap-modules-cs-2.6.13 which provide hostap-modules-cs? Sep 01 03:01:40 oho: how did you build these modules? Sep 01 03:02:32 zecke: they are from the 0.8.2 stable feed Sep 01 03:02:47 oho: sorry, then you asked in the wrong channel Sep 01 03:03:12 zecke: ok - I will go to handhelds.org !? Sep 01 03:03:22 oho: your issue could be caused by many different problems Sep 01 03:03:47 hrw|work, it should handle it correctly since last patch. Sep 01 03:03:47 I've seen a lot of those '*** glibc detected ***' latetly Sep 01 03:04:02 Pigi: .152 or later? Sep 01 03:04:02 hey zecke Sep 01 03:04:12 let me see Sep 01 03:04:21 zecke: the strange thing is that my prog runns anything between 1 minutes and 1 day and then it falls over Sep 01 03:04:53 hrw|work, .153 Sep 01 03:04:54 Pigi: .153 - will add it into OE Sep 01 03:05:04 zecke, ping Sep 01 03:05:28 * zecke hides Sep 01 03:05:47 Pigi: I'm so freaking busy, I have not booted my desktop system the last two weeks :} Sep 01 03:06:33 hmm.. someone force ipkg older then 152? Sep 01 03:06:45 heh. Ok. I will wait a bit yet . Or evntually, I will commit that patch ( that is working on Linux ) and eventually fix some slight difference after your tests :) Sep 01 03:07:03 hrw|work, ? Sep 01 03:07:19 Pigi: right Sep 01 03:07:39 Pigi: we have 12 versions of ipkg in OE - little too much I think ;) Sep 01 03:08:48 I'm not sure how did the OE logic works for the versions. In effect, there should be only the latest. Sep 01 03:09:16 that could also save some spaces on disks. :) Sep 01 03:09:28 ;) Sep 01 03:09:34 NOTE: package ipkg-0.99.153-r0: task do_build: completed Sep 01 03:09:41 RP: ok, cheers for info Sep 01 03:10:00 ~heil hrw|work Sep 01 03:10:06 ~hail hrw|work Sep 01 03:10:08 * ibot bows down to hrw|work and chants, "I'M NOT WORTHY!!" Sep 01 03:11:48 ;) Sep 01 03:26:10 03hrw 07org.oe.dev * r3ca788a6... 10/packages/ipkg/ (ipkg-native_0.99.153.bb ipkg_0.99.153.bb): added ipkg 0.99.153 which fix problem with Provides: support (needed when using PARALLEL_INSTALL_MODULES) Sep 01 03:31:52 03hrw 07org.oe.dev * r8679a2e7... 10/packages/ipkg/ (28 files in 8 dirs): Sep 01 03:31:52 drop all old versions of ipkg Sep 01 03:31:52 - 0.99.153 is latest one Sep 01 03:31:52 - 0.99.152 is latest tested one Sep 01 03:31:52 - older versions are removed Sep 01 03:33:29 hrw|work: I still see the old ipkg- ok Sep 01 03:34:15 NOTE: package bootstrap-image-1.0-r0: task do_build: completed Sep 01 03:34:22 Pigi: thx again for ipkg .153 Sep 01 03:34:36 hrw|work, you're welcome ;) Sep 01 03:35:24 koen: will check Sep 01 03:36:12 hrw|work, your message about bootstrap-image means that ipkg .153 is working as expected ? Sep 01 03:36:18 Pigi: yes Sep 01 03:36:37 good. Another one step to ipkg 1.0 :) Sep 01 03:36:44 Pigi: I had a problem with provides Sep 01 03:37:01 before 153 or now ? Sep 01 03:37:05 before Sep 01 03:37:45 yes, there were a bit of confusion in that code before 153. hopefully now is all ok :) Sep 01 03:38:47 Malaga.. I know one OZ user ther Sep 01 03:38:48 e Sep 01 03:45:34 koen: hmm.. dont know how to tell monotone to remove empty dir.. Sep 01 03:45:35 12:45 hrw@work:ipkg$ mt rm ipkg-0.99.129 Sep 01 03:45:35 monotone: skipping packages/ipkg/ipkg-0.99.129, not currently tracked Sep 01 03:48:48 weird Sep 01 03:49:45 yep Sep 01 03:52:50 heh.. opie-image build perl.. something changed since last attempt Sep 01 03:59:50 Time to sleep here. Sep 01 03:59:52 bbl Sep 01 03:59:58 * Pigi is away: I'm busy Sep 01 04:05:42 found.. opie-image -> obexftp -> libgsm -> intltool-native -> libxml-parser-perl-native -> perl-native Sep 01 04:08:16 anyone having probs connecting to cvs.handhelds.org(192.58.209.90):2401 failed: Connection timed out Sep 01 04:13:14 I fetched cvs.handhelds.org today Sep 01 04:14:35 NOTE: package sqlite3-3.2.5-r0: task do_build: completed Sep 01 04:15:37 hh.org is down atm Sep 01 04:15:48 unscheduled Sep 01 04:21:42 koen: ok thanx Sep 01 04:22:23 * koen looks at france Sep 01 04:52:32 how to avoid problem with software fpu vs hardware fpu ? Sep 01 04:53:26 help glibc people with EABI Sep 01 04:53:42 hey rwhitby, do you have a minute to help me out with some nslu2 stuff? Sep 01 04:54:43 openslug? Sep 01 04:55:06 03hrw 07org.oe.dev * r8f24f9e5... 10/packages/python/python-gnosis_1.1.1.bb: python-gnosis: fixed SRC_URI - partially close bug#279 Sep 01 04:57:02 actually ucslugc :P more with the best way to add new .bb files? Sep 01 04:58:20 03hrw 07org.oe.dev * rdb64b489... 10/packages/mtd/mtd-utils_20050801.bb: mtd-utils: switched back to CVS with fixed CVSDATE Sep 01 04:59:46 mithro: go for it Sep 01 04:59:50 normally i would use btshell etc, but the makefile makes it a bit different Sep 01 05:02:48 mithro: as far as writing .bb files, anyone here can probably help you better than I can (assuming it's not slug-specific) ... Sep 01 05:03:24 mithro: any news from the CELF people? Sep 01 05:03:27 rwhitby: i can write the bb fine, it's testing/building it Sep 01 05:03:32 koen: ?? Sep 01 05:03:47 how do i do it in the context of the stuff the MasterMakefile set up Sep 01 05:04:01 mithro: you are being mentioned on their wiki Sep 01 05:04:21 mithro: you have two choices - either add it to conf/distro/ucslugc-packages.conf and packages/meta/ucslugc-packages.bb Sep 01 05:04:35 or run bb unfreeze and then just add it to ucslugc-packages.bb Sep 01 05:04:46 bb unfreeze? Sep 01 05:05:21 yeah, instead of parsing all the OE bb files, we keep a list of those ones that are relevant to nslu2-linux. It's called freezing. Sep 01 05:05:40 okay Sep 01 05:07:39 so just add it to conf/distro/ucslugc-packages.conf then do a make ?? Sep 01 05:11:06 cd ucslugc ; . ./setup-env ; bb Sep 01 05:11:36 need mtd-snapshot-20050715.tar.bz2 can anyone help Sep 01 05:11:37 then add it to packages/meta/ucslugc-packages.bb when it's ready to go to the feed Sep 01 05:12:03 gruad: update metadata Sep 01 05:12:07 01 13:58 < CIA-9> hrw org.oe.dev * rdb64b489... /packages/mtd/mtd-utils_20050801.bb: mtd-utils: Sep 01 05:12:11 switched back to CVS with fixed CVSDATE Sep 01 05:13:44 adding btsco Sep 01 05:14:54 hh.org up again Sep 01 05:18:44 removing ipsec and ipv6 gives ~3M in rootfs Sep 01 05:19:05 is ipv6.ko that big? Sep 01 05:19:34 ipsec is big Sep 01 05:20:00 it is Sep 01 05:20:14 and if your name is alwin you remove it without asking Sep 01 05:20:35 and if your name is groucho you state that pda don't need networking Sep 01 05:20:58 I plan to flash collie with 3.5.4-pre and write howto about getting free space on / by using 'ipkg remove' Sep 01 05:22:41 hrw|work: using the dreambox branch... Sep 01 05:22:47 hrw|work: like http://handhelds.org/moin/moin.cgi/FamiliarSlimDownHowto ? Sep 01 05:23:11 koen: yep - maybe will just update that and point to this on oz.org Sep 01 05:23:29 gruad: you can populate org.openembedded.dev into dreambox Sep 01 05:24:28 hrw|work: on which machine? Sep 01 05:24:54 mithro: ? Sep 01 05:25:07 I think he missed 'collie' Sep 01 05:26:05 koen: IPsec is generally unneeded! Sep 01 05:26:13 koen: I need MPPE not IPsec Sep 01 05:26:18 koen: I need WPA not IPsec Sep 01 05:26:23 hrw|work: you mean i should pull the whole dev branch? Sep 01 05:26:26 koen: I need nfsd and portmap not IPsec.. Sep 01 05:26:27 * koen doesn't need ipsec at all Sep 01 05:26:39 koen: I need an icecream/shoutcast server not IPsec Sep 01 05:26:59 koen: all these are network applications familiar is not shipping by default... Sep 01 05:27:01 *sigh* Sep 01 05:27:31 anyway now I need to look into meta-opie.bb, opie-collections.inc and opie-image.bb to get them all in sync and nicely done Sep 01 05:27:51 hehe Sep 01 05:27:52 i don't need IPSec or ipv6 or portmap :P Sep 01 05:28:01 * zecke is overworked Sep 01 05:28:07 currently meta-opie builds all opie and qpe stuff, opie-image use opie-collections.inc Sep 01 05:29:00 I probably do opie-image depend on new meta-opie (opie stuff only) and add meta-opie-all which will contain opie and qpe stuff Sep 01 05:32:02 How do I see qDebug messages fom qte? Sep 01 05:32:39 RP: well, compile Opie with debug Sep 01 05:32:55 RP: these messages get filtered when libqpe-opie is not compiled with debug Sep 01 05:33:03 zecke: Ah, I have to recompile it :-/ Sep 01 05:33:15 RP: in library/qpeapplication.cpp Sep 01 05:33:17 i need some help with a bb file Sep 01 05:33:38 RP: look for MsgHandler Sep 01 05:33:38 it needs to produce some stuff via automake and some stuff via the kernel module stuff Sep 01 05:34:09 mithro: that sounds tricky, it all depends how these classes export their methods Sep 01 05:34:57 zecke: thanks Sep 01 05:35:03 zecke: i guess i could just have to files Sep 01 05:36:32 s/to/two Sep 01 05:55:17 the slug has 2.6 kernel right? Sep 01 05:56:19 openslug and ucslugc do Sep 01 05:56:26 unslung uses 2.4 Sep 01 05:56:49 i'm getting the following errors "include/linux/interrupt.h:123: warning: `regparm' attribute directive ignored" Sep 01 05:57:12 ahh, it's trying to use my kernel instead of the ucslugc one Sep 01 05:57:16 make[1]: Entering directory `/usr/src/ultralight/linux-2.6.12.1' Sep 01 05:59:17 http://www.linuxdevices.com/articles/AT9034569280.html - another device which base on Familiar Sep 01 06:02:12 morning folks Sep 01 06:02:30 mickeyl: I'm 'cleaning' opie-image/opie-collections/meta-opie Sep 01 06:02:39 bblr Sep 01 06:02:40 ok, cool Sep 01 06:06:10 mickeyl: hey Sep 01 06:13:08 hi zecke, how are things Sep 01 06:16:26 hi mickeyl Sep 01 06:17:30 mickeyl: Does opie use the calibration functions from qwsmouse.cpp in qte-2.3.10 or do these get replaced by something somewhere else? Sep 01 06:18:39 * mickeyl hands this question over to zecke Sep 01 06:19:05 RP: i _think_ we are using the calibration from the calibrate subdirectrory Sep 01 06:19:14 this one gets compiled into the launcher Sep 01 06:19:40 re: your question w/ kergoth yesterday. I would love to have calibration API in tslib, also for my E stuff Sep 01 06:20:37 mickeyl: Ok, I'll keep that in mind. I'm trying to hack qte to get a working algorithm for spitz as a first step :) Sep 01 06:20:57 The trouble is opie doesn't appear to be using the code I'm changing Sep 01 06:20:58 RP: cool Sep 01 06:21:59 A call to QWSServer::mouseHandler() ->calibrate(); would go to qte ? correct? Sep 01 06:23:36 let me have a look Sep 01 06:23:49 RP: right Sep 01 06:24:01 RP: QTSlibHandler Sep 01 06:24:11 RP: in src/kernel/qwsmouse_qws.cpp Sep 01 06:24:16 RP: what issue do you see? Sep 01 06:24:27 RP: I'm aware of two issues with calibration Sep 01 06:24:54 1) the calibration sanity check is wrong, a Touchscreen can be diagonally mirrored Sep 01 06:25:14 2) the writing and calculation logic in QtE only supports one orientation of the touchscreen Sep 01 06:25:59 zecke: I suspect my touchscreen X and Y are rotated compared to my screen X and Y Sep 01 06:26:30 zecke: becuase b=0 and d=0 are assumed by QtE, I can't get a good calibration Sep 01 06:27:04 I'm trying to replace the calculation in QtE with something better. Its as if my code doesn't make it onto the device though Sep 01 06:27:20 Is anything statically linked against qte? Sep 01 06:27:27 RP: no Sep 01 06:27:36 RP: but be sure to reimplement the right handler! Sep 01 06:27:43 RP: calibrate() is a virtual method Sep 01 06:27:58 and somehow writeCalibration gets called as well Sep 01 06:28:13 cia mallum Sep 01 06:28:17 ciao even Sep 01 06:28:20 zecke: I didn't notice there were two :) Sep 01 06:28:32 zecke: This would explain things... Sep 01 06:28:40 hey zecke Sep 01 06:29:10 pizza... Sep 01 06:29:14 the joys of vtables Sep 01 06:29:25 Good afternoon Sep 01 06:29:32 vtables sounds like 'break my ABI' Sep 01 06:30:02 hi sirfred Sep 01 06:30:19 mickeyl: Hello. Sep 01 06:31:08 hi sirfred Sep 01 06:31:21 hi, RP Sep 01 06:35:57 03florian 07org.oe.dev * r79422479... 10/packages/xserver-common/xserver-common_1.1.bb: Sep 01 06:35:57 Remove useless patch breaking several other applications. Sep 01 06:35:57 I really explained why it is there. Sep 01 06:38:45 mickeyl: Now opie's pcmciaapp is doing nice things like suspending my /dev/hda :-/ Sep 01 06:40:52 RP: sounds like you should configure it properly Sep 01 06:41:22 "activate" = "doing nothing" Sep 01 06:41:52 mickeyl: a 'assign driver' like gpe has would be a nice addition Sep 01 06:43:17 mickeyl: Yes, I hit anyold thing with a none working touchscreen when trying to get rid of a dialog. Now I can't get into Opie (or let the device boot) as it will crash as soon opie starts :) Sep 01 06:44:01 * RP rm PCMCIA.conf Sep 01 06:44:25 koen: already implemented for unknown devices. Sep 01 06:44:40 JustinP: ping Sep 01 06:45:15 ERROR: Cannot satisfy the following dependencies for task-enlightenment-x11-core: Sep 01 06:45:15 rxvt-unicode xstroke xtscal xrandr xmodmap xdpyinfo ttf-bitstream-vera chkhinge entrance e-wm Sep 01 06:45:15 ERROR: Cannot satisfy the following dependencies for task-enlightenment-x11: Sep 01 06:45:15 entice eclair examine evidence e-modules e-utils elitaire Sep 01 06:45:31 guess that bug was introduced by splitting e-image and e-image-core Sep 01 06:45:47 mickeyl: ok Sep 01 06:45:54 I should reflash a newer image Sep 01 06:46:02 this one still has the ljp unfluffing Sep 01 06:46:10 * koen resizes another window Sep 01 06:46:46 heh Sep 01 06:46:52 that's pretty annoying Sep 01 06:47:04 upgrading libqpe will do Sep 01 06:47:11 mickeyl: for me pcmcia applet does not work at all.. it show in taskbar but does not react to press (I removed PCMCIA.conf) Sep 01 06:47:34 hrw|work: that may be a consequence of the latest pcmcia bits in 2.6.13-mm Sep 01 06:47:39 mickeyl: 0830 build Sep 01 06:47:56 mickeyl: and I dont have pcmcia-cs Sep 01 06:48:08 right Sep 01 06:48:10 that would be a problem Sep 01 06:48:12 mickeyl: the paranoid voice inside me insists that TT is sabotaging opie Sep 01 06:49:00 hrw|work: opie pcmciaapplet still needs cardmgr to check if a device is known/unknown by cardmgr. if it doesn't exist, it finds nothing. i guess i have to see how that's handled nowadays :/ Sep 01 06:49:08 koen: i can slightly understand that :)) Sep 01 06:49:20 "of course I'm paranoid... all are after me!" Sep 01 06:49:39 "being paranoid doesn't mean nobody is after you" ;) Sep 01 06:49:43 hehe Sep 01 06:49:52 mickeyl: http://kernel.org/pub/linux/utils/kernel/pcmcia/howto.html Sep 01 06:50:56 mickeyl: make it rdeneds on pcmcia-cs ;) Sep 01 06:51:07 rdepends Sep 01 06:51:28 well Sep 01 06:51:32 what TT work for GPE :-D Sep 01 06:51:36 although i really welcome the new stuff Sep 01 06:51:52 XorA: no, SCO Sep 01 06:51:53 but i have absolutely no motivation to add yet another 2.4/2.6 extra handling :/ Sep 01 06:52:11 XorA: the next thing you know they will be selling licenses Sep 01 06:53:08 Hi zecke Sep 01 06:53:21 mickeyl: well if its a hotplug pcmcia device then you dont need to eject/insert and all that stuff so you could claim your applet isnt needed Sep 01 06:55:03 XorA: I would use it to emulate eject/insert on resume of my bt card Sep 01 06:55:30 XorA: nope. the applet is not just inserting and ejecting. it features driver configuration for unknown devices and configurable action handling on insert and suspend Sep 01 06:56:11 mickeyl: does it call cardctl directly? Sep 01 06:56:36 XorA: even better. it does the ugly ioctl's and talks to the kernel directly. i am not a fan of userland helpers. Sep 01 06:57:01 s/userland helpers/parsing output of external processes/ Sep 01 06:57:19 mickeyl: cool Sep 01 06:57:33 Hmm, I copied my OE stuff across to my FC4 laptop, did a fresh build (kill tmp), and now it fails in base-passwd with 'compiler can't create executables'. It seems cross-gcc isn't made at that point.. Anyone care to guess why? :) Sep 01 06:58:08 * koen blaims RPM and/or redhat Sep 01 06:59:11 I'd think it's something like that yeah.. guess I have to compare the lists of installed rpms Sep 01 06:59:28 (the system that works is FC4 too, but x86_64) Sep 01 06:59:59 wrobbie: then your cross compiler is in staging/i686, but OE thinks it has been built Sep 01 07:00:02 though i don't see how that could influence OE dependencies.. Sep 01 07:00:13 staging/x86_64 I mean, but looks in staging/i686 Sep 01 07:00:31 should be in cross/bin, not? Sep 01 07:00:44 wrobbie: ignore me, I missed a line in what you said Sep 01 07:03:11 manual 'build gcc-cross' fails on glibc build, possibly due to missing headers (glibc-initial?). Seems dependencies are screwy.. but all conf & bb files are same Sep 01 07:28:11 heh.. I know why I have not working autorotate :) lack of opie-autorotateapplet on my husky ;) Sep 01 07:31:11 *cough* Sep 01 07:31:29 s/opie-autorotate/opie-noautorotate/ Sep 01 07:31:33 ah.. Sep 01 07:31:42 god i hate debugging cgi. Sep 01 07:31:47 ~lart treke for a misleading name Sep 01 07:31:47 * ibot eats treke and falls over dead for a misleading name Sep 01 07:32:01 so I dont know why it does not want to autorotate with 2.6.13-rc6-mm2-r20 and 20050901 libopie Sep 01 07:32:11 yeah. Sep 01 07:32:24 and i can no longer build opie against OE for some odd reasons Sep 01 07:32:30 otherwise i would've looked into it Sep 01 07:37:29 heh.. I should not shutdown my home machine.. here I have to build whole opie as builddir is too old Sep 01 07:42:06 bbl Sep 01 07:42:11 cu mickeyl Sep 01 08:14:14 03koen 07org.oe.dev * rc0088957... 10/ (3 files in 3 dirs): gpe-today: add 0.09 + patch and update preferred-gpe-versions-2.7.inc Sep 01 08:35:45 Is there an easy way to regenerate the Packages.gz file in tmp/deploy/ipk (as happens when one builds an image)? Sep 01 08:36:02 ipkg-make-index Sep 01 08:36:24 cd tmp/deploy/ipk ; ipkg-make-index -m -p Packages . Sep 01 08:38:11 or rebuild bootstrap-image Sep 01 08:40:45 hmm.. new meta-opie goes quite nice Sep 01 08:41:17 lardman: tmp/staging/i686-linux/bin/ipkg-make-index -p tmp/deploy/ipk/Packages tmp/deploy Sep 01 08:41:20 /ipk/. Sep 01 08:43:26 hi sirfred Sep 01 08:43:34 hrw|work: Hi. Sep 01 08:44:18 2 bitbakes (one -nv, one normal) makes monotone even slower.. Sep 01 08:45:05 monotone slow? Sep 01 08:45:09 never noticed that? Sep 01 08:45:11 strangest spam ever Sep 01 08:45:16 :) Sep 01 08:45:27 * koen wonders why people think complaining about it will make it go faster Sep 01 08:45:45 That would be nice. Sep 01 08:49:22 03hrw 07org.oe.dev * r0875a62d... 10/packages/libopieobex/ (libopieobex0.inc libopieobex0_1.2.0.bb libopieobex0_cvs.bb): Sep 01 08:49:22 simplify libopieobex recipes Sep 01 08:49:22 - moved common part to libopieobex.inc Sep 01 08:49:23 We're probably also a victim of our own popularity - on a normal day, we'll see what, 50 - 100 changes Sep 01 08:49:40 We're probably one of the more active montone based projects Sep 01 08:50:56 CosmicPenguin: very true Sep 01 08:51:17 I got used to pull/update at least once a day. Sep 01 08:51:30 That way I can live with its slowness. Sep 01 08:51:52 hmm.. maybe I will add cronjob to pull on 06:00 local time.. Sep 01 08:51:53 * RP has working ts calibration on opie+spitz+2.6 Sep 01 08:52:37 hmm.. meta-opie has task-opie-todayplugins in PACKAGES so it provides it - I'm right? Sep 01 08:53:03 RP: I'm starting to understand why you don't have time to work in the w100, ;-) (Congrats) Sep 01 08:53:13 * koen pulls every 10 minutes Sep 01 08:53:34 sirfred: 2.6 is nearly finished on spitz ;-) Sep 01 08:53:39 RP: yay! Sep 01 08:53:44 koen: your machine is fast enough to pull so often Sep 01 08:54:09 If I pulled every 10 minutes, they will overlap Sep 01 08:54:19 RP: it looks like giving you zaurus making 2.6 support almost done in very quick time Sep 01 08:54:31 while true ; do for host in monotone.vanille.de ewi546.ewi.utwente.nl monotone.nslu2-linux.org monotone.vanille.de; do monotone --db=~/OE-koen.db pull $host "org.openembedded.{dev,dreambox}" -kkoen@openembedded.org ; monotone --db=~/OE-koen.db push $host "org.openembedded.{dev}" ; done ;sleep 10m ; done Sep 01 08:54:35 hrw|work: I always said it would ;-) Sep 01 08:54:36 no overlap Sep 01 08:55:42 koen: On what machine? Sep 01 08:56:30 amd 1800+ Sep 01 08:56:31 My server could probably cope with that. I might have to try it... Sep 01 08:56:32 NOTE: package opie-image-1.0: completed Sep 01 08:56:41 with my own meta-opie and opie-image Sep 01 08:56:46 yeah! :) Sep 01 08:57:01 hrw|work: Yay :) Sep 01 08:57:43 I'm on a PIII 700, I'd better not try it. Sep 01 08:59:53 CoreDump|home: hi Sep 01 09:00:08 hey Sep 01 09:00:24 CoreDump|home: I had to ask you something but forgot what it was... Sep 01 09:00:26 CoreDump|home: Do I remember you correctly as an akita owner? Sep 01 09:01:15 hrw|work: heh Sep 01 09:01:19 RP: yeah Sep 01 09:01:48 CoreDump|home: I might have a 2.6 kernel for you to play with... Sep 01 09:01:55 ah interesting Sep 01 09:02:17 CoreDump|home: Basically the kernel in OE should run on akita with the right commandline Sep 01 09:02:20 how high is the chance of bricking the Z? =) Sep 01 09:02:47 CoreDump|home: One of the people on the users list told me it boots - they've done the nasty bit Sep 01 09:02:53 ah Sep 01 09:03:01 clamshells have a ROM, they can't be bricked Sep 01 09:03:01 very nice Sep 01 09:03:06 CoreDump|home: you can't brick them anyway Sep 01 09:03:11 =) Sep 01 09:03:16 koen: 5600 and 6000 too Sep 01 09:03:28 hrw|work: ok, I wasn't sure about those Sep 01 09:03:32 or in other words: PXA zauruses cannot be bricked Sep 01 09:03:33 You can just corrupt the /home and / flash partitions Sep 01 09:03:39 bricking == jtag needed Sep 01 09:03:51 dont know about A300 ofcourse but we dont support it at all Sep 01 09:03:56 IMO bricking = normal flashing doesn't work any longer Sep 01 09:04:17 nope Sep 01 09:04:20 that isnt the accepted defintiion of bricking Sep 01 09:04:22 The mtd partitioning looked fine in the garbled message I got back from the tester so I'm confident Sep 01 09:04:22 by anyone Sep 01 09:05:00 It's mine, for me a PDA is bricked if I have to solder some wires to it heh Sep 01 09:05:07 hey kergoth Sep 01 09:05:19 CoreDump|home: Is it possible to end up in a state where you have to use a different flashing mechanism. No need to solder wires though Sep 01 09:05:30 hey Sep 01 09:05:39 unless you manage to mess up the real ROM Sep 01 09:05:45 which actually is ROM Sep 01 09:06:11 which propably requires something sharp with stabbing motions Sep 01 09:06:26 I didn't know. I thought it could be bricked. I will sleep better. Sep 01 09:06:30 right, anyway, I gotta go :\ RP, please post a link and I'll try it on the weekend ok? Sep 01 09:06:39 CoreDump|home: I discovered that i2c chip is a Maxim 7310 btw Sep 01 09:06:47 CoreDump|home: Ok, will do Sep 01 09:06:51 later! Sep 01 09:07:59 cu too Sep 01 09:08:38 koen & XorA: a belated thanks Sep 01 09:08:58 <_alwin_> hi Sep 01 09:26:27 monotone update gives me this: "... monotone: warning: missing packages/ipkg/ipkg_0.99.151.bb monotone: misuse: 28 missing files" Sep 01 09:26:38 should I just mt drop all those? Sep 01 09:26:43 yes Sep 01 09:26:48 hi koen Sep 01 09:26:52 hey mreimer Sep 01 09:27:06 should I drop files whenever I see that message? Sep 01 09:27:17 most of the times, yes Sep 01 09:27:51 you can always check http://handhelds.org/hypermail/oe-commits/ or http://cia.navi.cx/stats/project/openembedded Sep 01 09:34:22 hey people Sep 01 09:37:38 mickeyl: PONG Sep 01 09:37:47 mickeyl: yes, I see that too. It's very frustrating. Sep 01 09:38:09 mickeyl: and *NO* this has nothing to do with -core. It worked fine when I first split it. Somethine more recent has caused this but I don't know what/ Sep 01 09:38:34 mickeyl: bb is simply not building the EFL and EFL-dependant programs. It's ignoring the dependencies...not even complaining. Sep 01 09:43:24 mickeyl: I was futzing with the bb files a while yesterday and even removed my tmpdir but it's just assuming these packages are built somehow...a bitbake -nv doesn't show the efl being built Sep 01 09:43:36 mickeyl: if I "bitbake entrance" is starts up with edb... Sep 01 09:44:47 hey mickeyl, you alive? Sep 01 09:46:17 hi...i'm having a problem compiling gpe-timesheet Sep 01 09:46:26 JustinP, did you remove your stamps dir then? Sep 01 09:46:41 it says there is an undefined reference to pango_font_get_font_map Sep 01 09:46:43 usually that's in the tmp folder... Sep 01 09:46:46 anyone else had this problem? Sep 01 09:47:38 ldc: I removed the entire tmpdir Sep 01 09:48:05 ldc: I had tried previously removing all of the e* tmp/work and tmp/stamps files but it just went on using old version of the ipks Sep 01 09:48:18 mickeyl: hmmm....perhaps is *could* have been the -core switch.... Sep 01 09:51:55 I don't understand how it would be, though.... Sep 01 10:09:51 03justinp 07org.oe.dev * r6cb03263... 10/packages/e17/entrance_20050810.bb: Fix typo Sep 01 10:09:55 03justinp 07org.oe.dev * r29134267... 10/packages/xserver-common/files/remove-run-calibrate-sh.patch: Remove unused patch Sep 01 10:10:03 damn Sep 01 10:10:07 I forgot the package names Sep 01 10:10:09 oh well Sep 01 10:15:55 JustinP: I'm sorry, but it _is_ the split Sep 01 10:16:02 I've downloaded the revision right before your patch Sep 01 10:16:06 and that worked Sep 01 10:16:10 mickeyl: ok, I understand Sep 01 10:16:21 mickeyl: like I said, I tested building the images but not from scratch Sep 01 10:16:24 (after the split) Sep 01 10:16:30 I think I fixed it, give me a minute Sep 01 10:16:32 ya, no worries, we just need to find out how to repair it Sep 01 10:16:35 sure Sep 01 10:17:05 mithro: yea, i'm here Sep 01 10:17:22 I don't know how it was working before....if this fixes it Sep 01 10:17:23 mickeyl: how small did/can you get python down to? Sep 01 10:17:57 LoDown: rebuild from scratch Sep 01 10:18:03 hrw|gone: what's in PACKAGES is just RPROVIDES. Sep 01 10:18:20 oh, BTW, I got synergy working on my Zaurus ^_^ I had to install full X, though, diet-x11 has missing parts for it. Sep 01 10:18:56 not that it's excessively useful, but it's still cool to be mousing over to my zaurus from my workstation Sep 01 10:18:59 mithro: that depends on whether it should be useful or not. I can stuff the core package into <= 2MB but without the library it won't be that useful. a reasonable complete image needs at least 4MB, better 6MB Sep 01 10:19:02 whats synergy? Sep 01 10:19:08 s/image/distribution/ Sep 01 10:19:14 that is _without_ the bootstrap stuff etc. Sep 01 10:19:26 anyone playing with oe uclibc images nowadays? Sep 01 10:19:29 http://synergy2.sourceforge.net/ Sep 01 10:19:59 kergoth: iirc the nslu2 folks a bit. no time @ my side to further pursue it Sep 01 10:20:04 nod Sep 01 10:20:29 right now at work i'm doing new features on the existing product, so cant make substantial changes Sep 01 10:20:38 but the new product i'll be working on, i'll likely get them to use oe Sep 01 10:20:50 at least for the PDA models I think it's not that interesting any more. there are other device classes that can benefit more from uclibc Sep 01 10:20:56 cool Sep 01 10:21:05 ucslugc is pretty cool Sep 01 10:21:16 * mithro is playing with it atm Sep 01 10:21:29 mickeyl: well, using uclibc does give you the advantage of a damn quicker build time ;) Sep 01 10:21:38 and of using something with a maintainer we can actually pester Sep 01 10:21:49 true enough Sep 01 10:21:51 that's worthwhile Sep 01 10:21:56 * CosmicPenguin throws things at andersee just for the hell of it Sep 01 10:22:00 hehe Sep 01 10:22:02 Because we can! Sep 01 10:22:12 even for us C++ weenies there's hope .) Sep 01 10:22:17 CosmicPenguin: have you been using glibc or uclibc on that one alchemy board? Sep 01 10:22:24 uclibc Sep 01 10:22:28 i would still like to see Zaurus on uclibc :P Sep 01 10:22:32 mickeyl: have you checked out that micro c++ lib? Sep 01 10:22:37 kergoth: works awesome Sep 01 10:22:46 kergoth: just a quick glance. it's infant but promising Sep 01 10:22:47 naturally it doesnt really benefit opie, since its stl, but it looks sweet for that Sep 01 10:22:49 * kergoth nods Sep 01 10:22:54 CosmicPenguin: nice Sep 01 10:23:19 damn hci-usb kernel panics on my slug :/ Sep 01 10:23:20 I'm happy today, I finally found a small C++ type safe callback minilibrary I can use. good bye sigc Sep 01 10:23:30 nice Sep 01 10:23:36 mickeyl: url? Sep 01 10:23:41 http://www.mark13.org/toad/signalnslot.html Sep 01 10:23:46 two nice things: Sep 01 10:23:56 a) signal receivers don't need to inherit a common base object Sep 01 10:24:09 b) you don't need cryptic templates for most things Sep 01 10:24:14 and it's _small_ Sep 01 10:24:15 18k Sep 01 10:24:19 nice Sep 01 10:24:21 very cool Sep 01 10:24:33 we actually had someone in #gpe who build gpe for uclibc to save some space Sep 01 10:24:35 mickeyl: thats huge! ;) Sep 01 10:24:44 mithro: nah. not even for a phone ;) Sep 01 10:24:57 bah! Sep 01 10:24:58 but it was a gentoo moron, because his image is 200+ MB Sep 01 10:25:09 lol Sep 01 10:25:30 bbl tonite, dinner and dvd forthcoming Sep 01 10:25:45 mickey_away: have fun Sep 01 10:25:53 it seems that i no longer hack on oe just to hack on oe, but now only when i actually have a need to use it for a project Sep 01 10:26:05 which i suppose was the natural next step, since i was getting bored Sep 01 10:26:11 anyone know where to get a "pass through" SD card reader? Sep 01 10:26:12 bitbake2/bitbake-ng would hold my interest though Sep 01 10:26:36 is PROVIDES for the file or is it PACKAGE specific? Sep 01 10:26:48 PACKAGES isn't working as I would expect.... Sep 01 10:26:50 ie a SD card reader with inbuilt usb hub Sep 01 10:26:51 kergoth: call it bitbake 3000 Sep 01 10:26:56 bitbake xp! Sep 01 10:26:58 * kergoth runs Sep 01 10:27:12 eXtreme Performance Sep 01 10:27:27 mithro: sounds more like a USB hub with a built-in SD reader to me Sep 01 10:27:32 Bitbake EC: now with Extra Cluebatting Sep 01 10:27:39 lol Sep 01 10:27:41 i like that! Sep 01 10:28:40 is this not enough? PACKAGES = "task-enlightenment-x11 task-enlightenment-x11-core" bb can't find these packages for another's DEPENDS Sep 01 10:29:05 not += ? Sep 01 10:29:15 JustinP: PACKAGES Is used to control what ipk files or whatever get generated in the do_package process Sep 01 10:29:20 thats it Sep 01 10:29:30 ok, so it doesn't add the packages to PROVIDES then Sep 01 10:29:39 mickeyl told you it was RPROVIDES Sep 01 10:29:41 JustinP: i just said that :P Sep 01 10:29:41 not PROVIDES Sep 01 10:29:46 erm mithro Sep 01 10:29:47 heh Sep 01 10:29:53 ~fishslap JustinP Sep 01 10:29:55 * ibot slaps JustinP up side the head with a wet fish. Sep 01 10:30:15 mithro: no you didn't. You said is the other way Sep 01 10:30:30 hrw|gone: what's in PACKAGES is just RPROVIDES. Sep 01 10:30:34 heh Sep 01 10:30:42 kergoth: I am *not* hrw|gone Sep 01 10:30:57 kergoth: mickeyl said nothing to *me* about *my* problem about RPROVIDES Sep 01 10:31:00 JustinP: and? learn to read more than whats right in front of you Sep 01 10:31:06 you'll be less ignorant that way Sep 01 10:31:22 JustinP: bah :P Sep 01 10:31:32 i'm after one which is really really small too :P Sep 01 10:34:09 CosmicPenguin: ouch Sep 01 10:36:17 andersee: actually, you are a folk hero around here Sep 01 10:39:07 RPROVIDES is not defined in the manual. I also do not see how RPROVIDES could help me as I don't need to fix install-time. Bitbake isn't building the packages. It's ignoring the DEPENDS set. Sep 01 10:39:16 Read that again. It's *ignoring* the depends. Sep 01 10:39:27 It's not trying to *build* the depends. Sep 01 10:39:42 And it doesn't complain about *not building* the depends. Sep 01 10:40:07 it just fails on install-time because the DEPENDS have not been built Sep 01 10:44:24 then why are you wasting your time with PACKAGES? if the .bb's themselves arent being built, why would you expect packages to be emitted? Sep 01 10:44:29 packages is irrelevent. fix the real problem. Sep 01 10:45:02 the ipks for the tasks are being built but their DEPENDS are not Sep 01 10:45:04 try again Sep 01 10:45:47 why dont you start over, and actually describe your problem in english. Sep 01 10:46:04 "their DEPENDS"? whose? the ipks dont have DEPENDS, they have Depends/RDEPENDS. Sep 01 10:48:42 kergoth: when's the product going to be released? Sep 01 10:48:51 the new one? not sure yet Sep 01 10:50:22 CosmicPenguin: heh Sep 01 10:50:56 * koen notices uclibc 0.9.28 isn't in OE yet Sep 01 10:51:19 koen: should try it Sep 01 10:52:14 koen: 0.9.28 should do much better with the obscure dynamic linking wierdness needed for qt Sep 01 10:52:29 qt? I don't need no stinking qt Sep 01 10:53:04 * koen reads http://uclibc.org/FAQ.html Sep 01 10:54:52 koen: mentioned it since I've only used oe to (attempt to) build opie for my Zaurus Sep 01 10:58:00 we do have uclibc_svn, so qt loving people can try it Sep 01 10:59:45 koen: and of course non qt lovers as well. uclibc doesn't much care, and in theory should work perfectly now regardless of ones qtological preference Sep 01 10:59:54 :) Sep 01 11:00:39 I'm thinking of running ucslugc instead of openslug Sep 01 11:00:56 and of course look into building gpe and opie with uclibc Sep 01 11:02:31 koen: if you run into trouble, please let me know Sep 01 11:02:48 sure Sep 01 11:03:24 koen: I really, really would like to have gpe and open and all other stuff in oe work perfectly with uclibc Sep 01 11:04:06 I think about 80% of gpe (the core platform) builds with uclibc now Sep 01 11:04:26 the other 20% is mostly a matter of ifdeffing gettext stuff Sep 01 11:04:38 koen: or including gettext Sep 01 11:04:49 yeah Sep 01 11:05:03 if you go hardcore non-locale you run into problems Sep 01 11:05:04 when I build debian woody with uclibc, I included an external libintl and gettext Sep 01 11:05:17 s/build/built/ Sep 01 11:05:20 just booting simpad with an image i built yesturday, got themost bizarra mess of thivk white and black stripes on the screen! Sep 01 11:06:36 starting opie hasnt ceared it either Sep 01 11:12:19 hey pb_ Sep 01 11:12:32 hi koen Sep 01 11:17:40 any ideas how to fix this? look slike it may be set to 320x240, i could make out a dlashing curser at one point Sep 01 11:17:56 fbset reports 800x600 tho Sep 01 11:18:59 <[cc]smart> hi. do i read this correct, that autotools_do_configure recreates the configure script ? Sep 01 11:20:41 correct Sep 01 11:22:42 [cc]smart: if you need to alter configure / Makefile you should be altering the .am or .in files or adding a _prepend task to configure/compile Sep 01 11:24:53 note that modifying configure.in/configure.ac or makefile.in/makefile.am is the correct solution. hacks with sed are only tolerated if there's no other solution Sep 01 11:24:56 heh Sep 01 11:25:17 using sed is never a hack! Sep 01 11:25:22 it's pure beauty Sep 01 11:25:27 haw Sep 01 11:26:39 what stamp should i delete to rebuild the kernel? Sep 01 11:27:14 'rebuild virtual/kernel' in bitbake shell Sep 01 11:27:43 hi kergoth Sep 01 11:27:49 hey pb_ Sep 01 11:27:52 how goes it? Sep 01 11:28:08 koen, k Sep 01 11:29:28 <[cc]smart> JustinP: tried to patch the configure file in configure_prepend and i wondered why it never had an effect Sep 01 11:29:42 <[cc]smart> as i see it, the regenration of configure is done after prepend Sep 01 11:29:44 patching configure is -never- the correct solution Sep 01 11:29:46 <[cc]smart> so i should not need to wonder Sep 01 11:29:49 heh Sep 01 11:30:18 <[cc]smart> but then, i now wonder where the actualy failure sits, now knowing that OE recretes it cause: Sep 01 11:30:21 <[cc]smart> autoreconf: configure.ac: tracing Sep 01 11:30:21 <[cc]smart> configure.ac:475: warning: AC_RUN_IFELSE called without default to allow cross compiling Sep 01 11:30:21 <[cc]smart> autoconf/general.m4:2282: AC_RUN_IFELSE is expanded from... Sep 01 11:30:21 <[cc]smart> configure.ac:475: the top level Sep 01 11:30:37 those arent errors Sep 01 11:30:40 <[cc]smart> which might just mean that OE doesn't give all parameters Sep 01 11:30:42 perhapss you should paste the error Sep 01 11:31:12 <[cc]smart> checking if a program can be linked against Berkeley DB and run... configure: error: cannot run test program while cross compiling Sep 01 11:31:25 <[cc]smart> which does point to above for ,e Sep 01 11:31:27 your site file is missing that test result Sep 01 11:31:59 <[cc]smart> er, that test is cancelled Sep 01 11:32:05 <[cc]smart> due to crosscompile situation Sep 01 11:32:38 <[cc]smart> that test would be compiling and running a mini app, which doesn't make sense since the compiled executable would not match platform Sep 01 11:32:43 once again, your site file is missing that test result Sep 01 11:32:55 <[cc]smart> then, what's site file. Sep 01 11:32:58 <[cc]smart> but hey Sep 01 11:32:58 we presupply results for various arch-os combinations so it doesnt need to run the tests. Sep 01 11:32:59 <[cc]smart> wait a minute Sep 01 11:33:02 look in openembedded/site/ Sep 01 11:33:03 <[cc]smart> i know where this goes Sep 01 11:33:06 <[cc]smart> somebody else please Sep 01 11:33:28 i'd recommend you actually try listening to what i'm saying. Sep 01 11:33:35 maybe you'll learn something Sep 01 11:34:14 <[cc]smart> maybe, and maybe somebody prefers pumpin himself even more than explaining Sep 01 11:34:26 we presupply results for various arch-os combinations so it doesnt need to run the tests. Sep 01 11:34:40 that looks like an explanation to me Sep 01 11:36:19 i'm perfectly willing to explain, but if you can't be bothered to actually listen and follow instructions, it's just going to frustrate us both. Sep 01 11:39:04 <[cc]smart> so a resultsfile would replace a file which would normally be generated by configure, correct ? Sep 01 11:39:30 autoconf has a mechanism for caching the results of its tests. Sep 01 11:39:51 we leverage its caching mechanism to supply the test results so that it believes it's using the cached values, and does not run the tests. Sep 01 11:40:59 kergoth: Hihi ! Sep 01 11:41:09 hey schurig Sep 01 11:41:11 how ya been? Sep 01 11:41:12 kergoth: ah, HiHo (no intent to laught :-) Sep 01 11:41:47 kergoth: ill, busy and much more Sep 01 11:49:29 heh, been pretty busy myself Sep 01 11:49:31 kergoth: and, how're you? Still in TX ? Sep 01 11:49:41 nope, i quit my job at ti, wasnt happy in that position Sep 01 11:49:52 moved back to MN, living in my old townhouse with my dad to keep expenses down Sep 01 11:50:01 kergoth: any work there? Sep 01 11:50:15 recently found a job here hacking on linux network appliances, though its $15k less than i made in texas, given my low expenses, its okay.. plus i dont hate it ;) Sep 01 11:50:30 learning lots of networking stuff Sep 01 11:50:50 kergoth: sounds ok :-) Sep 01 11:51:05 for now anyway. trying to decide whether i want ot stay in computers or not Sep 01 11:51:43 kergoth: what would you do, besides computer? Sep 01 11:51:53 professional gambler Sep 01 11:51:54 i have no idea :) Sep 01 11:51:56 hehe Sep 01 11:52:07 rouge && vagabond? Sep 01 11:52:08 :-D Sep 01 11:52:16 if i dont stay in computers i'll likely go back to college, get a degree, since i dont have one now Sep 01 11:52:19 dunno what i'd do though Sep 01 11:52:40 the degree is really a must in the usa to get a big company to hire you. Sep 01 11:52:44 maybe an enterprise architect type position, major in management and computing Sep 01 11:52:46 no matter how talented Sep 01 11:53:02 kergoth: the problem is that you know to much to be a normal Manager Sep 01 11:53:09 in some places it is phd only. Sep 01 11:53:12 france_: yeah, i only got into TI because of open minded technically able managers there, one of which i'd know for quite a while (tim riker) Sep 01 11:53:21 s/know/known/ Sep 01 11:53:33 s/to/too/ Sep 01 11:53:33 * france_ is scared Sep 01 11:53:38 did tim die? Sep 01 11:53:55 at the time i'd known him for a while, now i've known him longer than that ;) Sep 01 11:54:03 ahh, ok Sep 01 11:54:04 he's not at ti anymore, moved back to utah Sep 01 11:54:13 now he's on huge clusters instead of embedded Sep 01 11:54:14 quite the change Sep 01 11:54:37 thats what I did for several years after starting hh.org Sep 01 11:54:46 but much of the super cluster stuff has similiar isssues Sep 01 11:54:55 as the embedded stuff... Sep 01 11:55:58 hmm, really? how so? Sep 01 11:56:04 france_: you think so? When you deal with clusters, you needs something like FAI, not OE Sep 01 11:56:43 I know so, as linux on the alpha arch was my job at dec/compaq for a while until the alpha arch was discontinued. Sep 01 11:56:48 ah Sep 01 11:57:11 think of have to build building the size of foot ball fields to hold the clusters. Sep 01 11:58:02 france_: as long as it's not the size of a Golf resort ... Sep 01 11:58:06 dang...thats some building :) Sep 01 11:58:16 re Sep 01 11:58:35 * kergoth is having fun with vlan atm Sep 01 11:59:56 * koen realises this is the last week of guiltless idling Sep 01 12:00:38 ? Sep 01 12:00:39 * LoDown hates school so much Sep 01 12:01:06 next week I will start my fifth year doing EE Sep 01 12:01:20 ah Sep 01 12:01:50 koen: you start late Sep 01 12:01:57 i've been in for 2 weeks now Sep 01 12:02:00 lucky Sep 01 12:08:26 kergoth: is tim still running ibot?? I have lost track Sep 01 12:08:52 yep Sep 01 12:08:53 I thgouth tim has always run ibot Sep 01 12:09:14 yep, afaik Sep 01 12:09:43 CosmicPenguin: it has changed hands a few times Sep 01 12:10:20 heh - to be fair, I've only been around since late 1999 anyway Sep 01 12:11:00 <[cc]smart> i don't see where configure or configure.ac would have a conditional that would make the test skip, aka which variable that is which would be missing from the resultsfile Sep 01 12:11:06 * CosmicPenguin knows nothing of the heady days before #tuxscreen Sep 01 12:11:22 [cc]smart: AC_TRY_RUN/AC_RUN_IFELSE has the code to do so in the macro itself. Sep 01 12:11:30 [cc]smart: so no, you owuldnt see it in configure.in orconfigure.ac Sep 01 12:11:49 if you really want to read it, for whatever reason, look in the .m4 files included with autoconf Sep 01 12:11:59 or just read the autoconf documentation, which covers caching Sep 01 12:13:43 <[cc]smart> hm, i would think i'd have to read the conditional to find out which variable to set in the resultsfile (short of knowing right now to what value) to avoid the test. Sep 01 12:14:42 oh, you're right, that AC_TRY_RUN() needs to be surrounded by an AC_CACHE_CHECK() or similar. Sep 01 12:14:54 remembered incorrectly, havent dealt with buildsystems in months Sep 01 12:20:26 <[cc]smart> then i don't think this using the resultsfile is not going to work in this case. the last AC_CACHE_CHECK before that test is miles away and for as far as i can follow the brackets, the closing one is not much closer. Sep 01 12:20:35 <[cc]smart> topic of that is also different Sep 01 12:21:07 what exactly is it testing? if it's just a sanity check, and the result of the test isn't used, you can simply not run the test when crosscompiling Sep 01 12:21:34 if the result is used, you'll want to add a call to AC_CACHE_CHECK to ensure the test results are cached, adn then add the correct test result to your site file Sep 01 12:21:37 <[cc]smart> well, this is what i tried to achieve patching the configure Sep 01 12:21:49 <[cc]smart> er that is skipping the test Sep 01 12:21:50 yes, and that isnt the right fix Sep 01 12:21:57 patching the configure, that is Sep 01 12:22:06 if thats really what you want to do, and the result isnt used, just do: Sep 01 12:22:11 <[cc]smart> got that, just didnt see another option yet Sep 01 12:22:17 if test x"$cross_compiling" = "xyes"; then Sep 01 12:22:20 before hte AC_TRY_RUN Sep 01 12:22:21 and fi Sep 01 12:22:23 after it Sep 01 12:22:28 or whatever Sep 01 12:22:42 then the generated configure will haev that shell code to skip it when crosscompiling Sep 01 12:22:54 <[cc]smart> well, if i'dcome to patch the configure.ac then i could just dump that part. Sep 01 12:23:07 that isnt the right fix either. Sep 01 12:23:20 thats a hack. if you're going to do something, do it properly Sep 01 12:23:34 i just gave you the right method for an AC_TRY_RUN() whose value isnt used Sep 01 12:24:31 <[cc]smart> for me this look like it has no other result than a go/no go desicion Sep 01 12:25:09 * kergoth nods Sep 01 12:25:33 if that's the case, i'd recommend patching the configure to do the check for crosscompiling like i said. if you patch it that way, it's not only clean, but the upstream developers might even accept the patch Sep 01 12:25:59 since it doesn't interfere with the non-crosscompile case Sep 01 12:26:13 isnt the gfx in the simpad sa1100 not mq200? Sep 01 12:27:28 no Sep 01 12:40:33 <[cc]smart> at least error condition moved from configure to compile state Sep 01 12:40:52 <[cc]smart> s/state/phase/ Sep 01 12:42:00 <[cc]smart> tomorrow's another day Sep 01 12:42:10 <[cc]smart> n8 Sep 01 12:46:48 03nail 07org.oe.dev * r55e352be... 10/packages/zd1211/zd1211_20050822.bb: Sep 01 12:46:48 zd1211: Updated to new version Sep 01 12:46:48 Supports bigendian arch. Sep 01 12:46:48 Fix install part Sep 01 12:46:48 Add wireless-tools dependency Sep 01 13:05:35 good evening Sep 01 13:07:09 anyone building for mipsel here? Had trouble with glibc today. Sep 01 13:07:23 I'm using gcc 3.4.4 Sep 01 13:07:52 had trouble is pretty vague Sep 01 13:10:16 true. some opcode not supported by mips. Sep 01 13:11:07 which gcc / glib versions are stable for mipsel Sep 01 13:12:10 glib or glibc Sep 01 13:12:16 you're being way too vague Sep 01 13:12:17 glibc Sep 01 13:12:28 error message are good. use pastebin Sep 01 13:12:34 s/message/messages/ Sep 01 13:13:09 * koen loves the inegrated pastebin in bbshell Sep 01 13:13:31 * RP always totally crashes bbshell :-/ Sep 01 13:13:44 just found out about the bbshell and pastebin last week Sep 01 13:14:55 my problem at the moment is that i run into problems during my workday and i'm behind a firewall which does not give me access to irc and cvs Sep 01 13:15:20 hm cant you use the web ircclient? Sep 01 13:16:20 i tried one, but its a java applet that also needs a transparent connection Sep 01 13:16:39 do you know one where the webserver plays for client? Sep 01 13:17:03 http://vengefuldeath.com/cgi-bin/irc.cgi Sep 01 13:18:35 thanks will try it tommorow Sep 01 13:19:13 the daily tarballs of oe are of great help for us Sep 01 13:19:57 epronk do oyu work on a nda-projekt? Sep 01 13:20:28 yes Sep 01 13:21:04 at Philips Research, but we got permission now to contribute bb files to oe Sep 01 13:21:07 epronk dont forget the GPL *g* Sep 01 13:21:30 hm philips and mips Sep 01 13:21:45 maybee a new settopbox Sep 01 13:21:56 they are very aware of gpl Sep 01 13:22:58 there is a COPYING.MIT file in the oe tree Sep 01 13:23:18 .bb files are nominally MIT Sep 01 13:24:27 Philips can contribute .bb files with a reference to the MIT license Sep 01 13:24:39 philips? Sep 01 13:24:45 isn't that montavista land? Sep 01 13:25:00 koen? Sep 01 13:25:08 epronk: eindhoven, nijmegen? Sep 01 13:25:17 eindhoven Sep 01 13:25:58 with oe it's much easier to build a root filesystem than with the montavista tools available Sep 01 13:26:00 It's been a while since I visited philips there Sep 01 13:26:37 there is a possibility to organise an oe meeting at philips research Sep 01 13:26:39 hm I something remeber that philips had some job offers at the CEBIT for embbeded opensource people Sep 01 13:27:23 btw. I'm not an employer of Philips Sep 01 13:27:41 at CEBIT? Sep 01 13:27:50 ahhhhhhhh i have to change work !!!!!!!!!!!!!!! Sep 01 13:28:10 with oe it's much easier to build a root filesystem than with the montavista tools available Sep 01 13:28:15 That should be on the OE home page Sep 01 13:28:16 epronk you dont know the greatest eurpean compuer fair? Sep 01 13:28:50 yes, but it's not a job fair :) Sep 01 13:29:02 tell me about it Sep 01 13:29:10 epronk lol, cebit is also one of the greatest job fair Sep 01 13:29:23 CosmicPenguin: http://oe.handhelds.org/cgi-bin/moin.cgi/SuccesStories Sep 01 13:30:18 koen|work: yeah, but thats such a great marketing slogen Sep 01 13:30:20 slogan even Sep 01 13:30:28 it sure is Sep 01 13:30:32 * koen|work really gone now Sep 01 13:34:15 epronk: you mean "employee" Sep 01 13:35:07 yes. did a lot of typing today, and it all compiled. Sep 01 13:35:50 any views on montavista ? Sep 01 13:37:10 our 3.3 code just didn't build on the 2.95 for ARM of montavista Sep 01 13:37:58 it was easy to build a 3.4.4 and bootstrap-image was already good enough to make a flying start Sep 01 13:43:24 i figured out why it was taking so long to compile an gpe-image on my box Sep 01 13:43:29 geez am I dumb... Sep 01 13:44:16 i hadn't uncommented the PARALLEL_MAKE option Sep 01 13:44:28 so it was using only one processor... Sep 01 13:44:31 heh, oops Sep 01 13:44:49 so now...it is progressing much more rapidly Sep 01 13:44:50 heh Sep 01 13:45:01 nood niht to all Sep 01 13:45:13 gremlin[it]: nood niht Sep 01 13:45:18 are there any known issues with the gcc_csl-arm-20050603 compiler? Sep 01 13:45:33 the PARALLEL_MAKE also caused problems with building xerces-c Sep 01 13:45:47 is it possible to disable it for a single package? Sep 01 13:47:28 you can set PARALLEL_MAKE = "" in the particular .bb file Sep 01 13:48:07 gb2: yes, there are. Some things won't compile with the CSL compiler Sep 01 13:48:44 pb: thanks, will try this Sep 01 13:48:55 * CosmicPenguin isn't really convinced that PARALLEL_MAKE makes that much difference in the grand scheme of things anyway Sep 01 13:49:30 * Philippe is convinced. Build time went down with 30% (once I moved to a raid array too keep the CPU's fed) Sep 01 13:49:41 20 to 30% so i've heard Sep 01 13:49:47 it seems to be helping with my build... Sep 01 13:49:54 that and installing psyco... Sep 01 13:49:58 JustinP: hrmmm Sep 01 13:50:14 I don't have solid figures...but they both helped (installed/turned on at seperate times) Sep 01 13:50:26 Good night. Sep 01 13:50:36 heh - thats the sort of report I always hear.. "It worked for me, but I added a new processor" Sep 01 13:50:50 heh Sep 01 13:50:56 anyone tried distcc with oe? Sep 01 13:50:56 haha Sep 01 13:51:09 epronk: zecke uses icecream with oe, I think Sep 01 13:51:17 CosmicPenguin, good to have solid numbers, eh? Sep 01 13:51:21 interesting, 2 the two dated gcc_csl-arm .bb files both contain DEFAULT_PREFERENCE settings, one for _arm-linux and _arm-linuxeabi, the other for _arm Sep 01 13:51:35 seems it's choosing the one for _arm for my setup here. Sep 01 13:52:44 pb: ah, if from SUSE Sep 01 13:53:11 i mean is Sep 01 13:53:18 its important to remember that in a large oe build, compilation isnt the main consumer of time. Sep 01 13:53:23 extraction/patching/etc is Sep 01 13:53:31 I wonder if that's intentional Sep 01 13:53:38 no, I think that's a bug Sep 01 13:54:22 i really wish monotone was faster Sep 01 13:54:42 :) Sep 01 13:54:46 monotone is a big reason why i haven't really spent time with oe since getting employed. Sep 01 13:54:47 heh Sep 01 13:54:55 yeah, same here Sep 01 13:55:01 We're all bound by our hate of Monotone Sep 01 13:55:03 gb2, that is definitely a good reason to have 2 cpu's :-) one for the oe build, one for monotone updates Sep 01 13:55:03 it's just too painful to check out the code nowadays Sep 01 13:55:04 kergoth: it works fine Sep 01 13:55:05 everyone is b*tchin about monotone... Sep 01 13:55:19 Philippe: lol Sep 01 13:55:36 i never used BK, but i'm guessing from the talk that it was _much_ faster, eh? Sep 01 13:55:39 JustinP: the first time i did a checkout, i watched it crash and tell me to contact the monotone developers. Sep 01 13:55:48 that wasnt the start of a beautiful relationshipo Sep 01 13:55:51 Why are there so many revision control systems these days? Sep 01 13:55:52 not to mention all the other stuff Sep 01 13:56:03 is svn/cvs not good enough? Sep 01 13:56:04 it's taking about 30 seconds per revision Sep 01 13:56:09 JustinTodd: no Sep 01 13:56:14 JustinTodd: svn/cvs doesnt meet our development model. Sep 01 13:56:19 s/meet/meet the needs of/ Sep 01 13:56:38 kergoth: well it works pretty flawlessly for me nowadays Sep 01 13:56:43 kergoth, why is that? need of 3 way merges? (i am completely uninformed) Sep 01 13:57:00 distributed development model Sep 01 13:57:03 kergoth: I'm still waiting for your official "why svk is ok for OE" propoasition Sep 01 13:57:05 svn and cvs are a central model Sep 01 13:57:07 LoDown: we have a more peer to peer development model. changes flow between developers repositories, not from client->server Sep 01 13:57:10 ah Sep 01 13:57:11 LoDown: look up distributed SCMs Sep 01 13:57:17 JustinP: yeah i know, i havent had time to get an svk repo up with our history Sep 01 13:57:17 * Philippe is trying out mercurial now.... Sep 01 13:57:17 aight Sep 01 13:57:23 LoDown: yes, bitkeeper was much faster than monotone. Not that it seemed particularly fast at the time, but monotone takes slowness to a whole new level. Sep 01 13:57:26 JustinP: girlfriend from out of town is in town this week.. :) Sep 01 13:57:37 kergoth: ah yes, I remember you saying that earlier Sep 01 13:57:38 is bk distributed? Sep 01 13:57:41 JustinTodd: yes Sep 01 13:57:49 does the monotone web interface let me get history of a single file? Sep 01 13:58:06 gb2: there is no such thing in monotone, really... Sep 01 13:58:14 JustinP: you can walk the graph Sep 01 13:58:15 heh Sep 01 13:58:17 gb2: and no, it doesn't ;-) Sep 01 13:58:20 same way git does it Sep 01 13:58:27 bk used a changeset model too, but it still let you look at the changes done to a file by the changesets Sep 01 13:58:37 as does git, and svk, and .. Sep 01 13:58:41 maybe I'm retarded...but what was the impetus behind going with monotone vs another distrubted SCM (or even git, or something) Sep 01 13:58:47 havent used darcs or arch/tla or bazaar lately Sep 01 13:59:07 git didn't fully exist at the time Sep 01 13:59:09 Its important to note that git was about a week old when we had to switch from BK Sep 01 13:59:12 ah Sep 01 13:59:17 true Sep 01 13:59:24 me forget :-D Sep 01 13:59:25 and Mercurial was not very stable Sep 01 13:59:35 CP: it was also about a week old :) Sep 01 13:59:41 arch/tla was a pain in the ass, svk didnt look sufficient, bazaar/bazaar-ng didn't seem to fit our needs, .. Sep 01 13:59:44 seeing as they started at the same time iirc Sep 01 13:59:52 svk is scary to me Sep 01 13:59:56 perl wrapped svn Sep 01 13:59:59 :-D Sep 01 14:00:06 ah hell, stupid xserver on my pc has crashed. Sep 01 14:00:12 taking all my work from the last hour with it, presumably Sep 01 14:00:14 it doesnt actually call out svn, it links to the svn library itself and thereby bypasses issues that the svn client itself has Sep 01 14:00:17 * pb_ stomps around Sep 01 14:00:17 but yeah, perl is icky Sep 01 14:00:22 pb_: doh :( Sep 01 14:00:28 does the monotone tool atleast let me see history of a file? Sep 01 14:00:34 ouch pb_ Sep 01 14:00:36 has anybody else built a simpad image recently? if so did the display work corectly? Sep 01 14:00:38 gb2: yes, you can ask for the log of a single file Sep 01 14:00:44 pb_: sorry to hear Sep 01 14:01:03 oh well Sep 01 14:01:07 * pb_ dinner time now Sep 01 14:01:14 netlink is pretty cool. never really spent much time with it before Sep 01 14:01:28 kergoth: what do you work on now a days? Sep 01 14:01:34 dan2003 you mean with the handhelds.org 2.6 kernel? Sep 01 14:01:41 gb2: working on the products at http://www.astrocorp.com/ Sep 01 14:01:42 woglinde, yup Sep 01 14:01:50 gb2: linux load balancing network appliances, basically Sep 01 14:01:54 cool Sep 01 14:02:08 yeah, lots of networking related resume fodder ;) Sep 01 14:02:13 dan2003 it works only on zecke's simpad Sep 01 14:02:30 !! whys that? Sep 01 14:02:31 dan2003: ?? Sep 01 14:02:39 dan2003 dont ask me Sep 01 14:02:48 what model does he have? Sep 01 14:02:55 dan2003 a slc like me Sep 01 14:03:04 dan2003 args sl4 Sep 01 14:03:07 dan2003 sorry Sep 01 14:03:12 mines sl4 Sep 01 14:03:21 I wonder how my gf's job interview went today Sep 01 14:03:31 where'd she interview at? Sep 01 14:03:46 some small company that does product prototyping, she's a mechE Sep 01 14:03:51 nice... Sep 01 14:03:54 i'm a ME too Sep 01 14:03:55 woglinde, u and zecke both flashed form the exact same image? Sep 01 14:03:57 :-D Sep 01 14:04:03 dan2003 dan yes Sep 01 14:04:06 she's finding jobs pretty scarse unfortunately Sep 01 14:04:14 oh...not good for me Sep 01 14:04:14 ah Sep 01 14:04:14 she's been looking for about 3 months Sep 01 14:04:20 i have 2 more years of school Sep 01 14:04:28 dan2003 because zecke was sitting next to me Sep 01 14:04:30 where are you located, if you don't mind me asking? Sep 01 14:04:34 Boston Sep 01 14:04:36 woglinde, ahha! how very odd Sep 01 14:04:46 hm Sep 01 14:04:52 u looked at the hardwarr? ie chip revisions ? Sep 01 14:04:52 well...I'm in Louisville... Sep 01 14:05:03 we'll see what the market looks like when I get out of school finally Sep 01 14:05:05 dan2003 not really Sep 01 14:05:13 she doesn't want to have to commute by car, so she's been limited in what she's applied for Sep 01 14:05:20 ah Sep 01 14:05:20 dan2003 zecke didnt have time the whole summer for the simpad Sep 01 14:05:26 i have another simpad, ill di that out and see if it works on that Sep 01 14:05:39 we went through and moved her over onto my health insurance though since her insurance through school ran out at the end of august Sep 01 14:05:40 dan2003 and me either Sep 01 14:05:42 the otuch screen is broke, but the lcd is intact Sep 01 14:06:10 it's great living somewhere where benefits for unmarried domestic partners is common Sep 01 14:06:16 any really cool uber kernel hackers about that want to figure out why my SLUG gives a kernel panic on inserting my hci-usb device? Sep 01 14:06:22 haha Sep 01 14:06:28 yep... Sep 01 14:06:39 gb2, how are the gas prices up there? Sep 01 14:06:48 they shot up to 3.10 here in one afternoon Sep 01 14:06:55 3-3.50 Sep 01 14:07:01 i don't own a car though :) Sep 01 14:07:15 gb2, one way to get around expensive gas Sep 01 14:07:16 heh Sep 01 14:07:18 if i could get away with not owning a car, i'd probably do it Sep 01 14:07:18 heh Sep 01 14:07:26 on the other hand, maybe not, since i love my car Sep 01 14:07:34 terps% monotone log gcc_csl-arm-20050603.bb Sep 01 14:07:34 terps% Sep 01 14:07:38 did i do something wrong? Sep 01 14:09:15 * LoDown rips hair out over homework Sep 01 14:13:26 wb mickeyl Sep 01 14:14:06 RP: got your mail. I'll take a look into the patches tomorrow. Sep 01 14:14:27 mickeyl: ok. It could be better, could be a lot worse ;-) Sep 01 14:14:39 damnit damnit damnit Sep 01 14:14:44 i lost the kernel tree i was working on last week Sep 01 14:14:49 just went and misplaced it Sep 01 14:14:53 bummer Sep 01 14:14:53 * kergoth looks around his drive Sep 01 14:15:01 i need that patchset, too Sep 01 14:15:02 bakcups? Sep 01 14:15:10 yeah, i should start making those :P Sep 01 14:15:14 heh Sep 01 14:15:16 sounds familiar Sep 01 14:15:33 probably just set up a periodic rsync to one of the servers here or something Sep 01 14:16:03 I'm a disk down on my server. Can't find a quiet time to swap it out though :-/ Sep 01 14:16:14 * RP should make better backups Sep 01 14:20:11 heh Sep 01 14:21:35 argh. When bitbake links glib-2.0 it complains it cant find libiconv functions. (familiar083, ipaq 4705) Sep 01 14:27:09 yeah, looks like that gcc is definately wrong Sep 01 14:27:15 and has been for a long time Sep 01 14:29:16 gb2. was that directed at me? Sep 01 14:30:16 no Sep 01 14:41:34 03mickeyl 07org.oe.dev * rc0acc534... 10/packages/efl/esmart.inc: esmart: fix typo in package name Sep 01 14:45:47 who is our recent uclibc guru? the uclibc_svn patch doesn't apply. a quick try at renaming 0.27 to 0.28 brought no joy, since all those patches don't apply either. Sep 01 14:47:29 damn it.... Sep 01 14:47:57 I'm probably going to have a merge conflict now... Sep 01 14:49:56 strange....it worked... Sep 01 14:50:10 maybe I didn't touch that file.... Sep 01 14:50:35 Monotone. It's better than its reputation (TM) Sep 01 14:50:40 * JustinP pushes Sep 01 14:50:43 yes, I agree Sep 01 14:50:49 * mickeyl founds a marketing company for monotone Sep 01 14:50:50 heh Sep 01 14:51:09 get ready for a bunch of commits Sep 01 14:51:15 uh oh ;) Sep 01 14:51:20 * mickeyl puts on the helmet Sep 01 14:51:31 I *did* test Sep 01 14:51:34 hehe Sep 01 14:51:46 I'll be testing from scratch in a minute, once my latest test finishes up Sep 01 14:51:56 they can't be more broken than they were, so.... Sep 01 14:52:36 / Sep 01 14:52:38 - Sep 01 14:52:40 \ Sep 01 14:52:42 - Sep 01 14:52:43 | Sep 01 14:52:48 :D Sep 01 14:52:51 oops Sep 01 14:52:56 I screwed that up Sep 01 14:53:05 /-\| Sep 01 14:53:17 ok...if a package fails to compile...where do I report that to? Sep 01 14:53:20 03justinp 07org.oe.dev * rb4ac82bd... 10/packages/meta/ (e-image-core.bb e-image.bb meta-e-x11.bb): Sep 01 14:53:20 task-e-x11, e-image: Rename e tasks, fix dependency problems after -core split Sep 01 14:53:20 - task-enlightenment-x11(-core) has been renamed to task-e-x11(-core) Sep 01 14:53:20 - Add PROVIDES since bb can't figure it out from PACKAGES Sep 01 14:53:20 - use DEPENDS instead of DEPENDS_package Sep 01 14:53:24 03justinp 07org.oe.dev * r70b37819... 10/packages/e17/entrance_20050810.bb: Sep 01 14:53:24 entrance: add edb-native to depends Sep 01 14:53:24 -We can remove this when we update to a newer entrance which uses ecore_config instead of edb_ed Sep 01 14:53:27 03justinp 07org.oe.dev * r34a24ff2... 10/packages/meta/meta-e-x11.bb: mete-e-x11: Remove task-e-x11-core from DEPENDS as it will already be built Sep 01 14:53:32 03justinp 07org.oe.dev * r8ae0fb14... 10/packages/efl/ (10 files): Sep 01 14:53:32 ecore, evas, imlib2, esmart: Update vitrual and native PROVIDES Sep 01 14:53:33 - Moved all virtual PROVIDES into inc files Sep 01 14:53:35 - Made inc file PROVIDES always += and removed any extra PROVIDES in main files Sep 01 14:53:37 - Made native builds override PROVIDES (native builds should not provide virtual) Sep 01 14:53:41 03justinp 07org.oe.dev * rc2f10032... 10/packages/e17/entice_20050810.bb: entice: change esmart dependency to esmart-x11 Sep 01 14:53:44 03justinp 07org.oe.dev * r7554e482... 10/packages/e17/e-modules_20050810.bb: e-modules: change DEPENDS to use -x11 instead of virtual as they are part of the WM and need X11 Sep 01 14:53:47 03justinp 07org.oe.dev * rbb7ad680... 10/conf/distro/openzaurus-3.5.4.conf: openzaurus-3.5.4: no need for preferred-e-versions currently Sep 01 14:53:50 03justinp 07org.oe.dev * r42a23a2b... 10/packages/elitaire/elitaire_0.0.4.bb: elitaire: change esmart to esmart-x11 Sep 01 14:53:53 03justinp 07org.oe.dev * r6dd4b1d7... 10/packages/meta/ (meta-e-x11-core.bb meta-e-x11.bb): meta-e-x11: break task-e-x11-core out into its own file so that e-image-core builds don't have to build all the extra stuff Sep 01 14:53:59 LoDown: here and/or bugs.openembedded.org Sep 01 14:54:01 woglinde, its working on this 1!! Sep 01 14:54:11 yay, e-image(-core) should now work Sep 01 14:54:15 again Sep 01 14:54:24 ok... Sep 01 14:54:46 LoDown: you can ask here and if no one fixes it report to bugs.openembedded.org Sep 01 14:55:02 mickeyl: I would appreciate a test on your end as well :-) Sep 01 14:55:03 ibgpevtype-0.12-r0 fails to build due to missing header mimedir-recurrence.h Sep 01 14:55:26 JustinP: definitly :) Sep 01 14:55:30 hi mickeyl Sep 01 14:55:37 hey pb_ Sep 01 14:55:43 bloody s/ibgpevtype-0.12-r0/libgpevtype-0.12-r0 Sep 01 14:55:56 woglinde, only diff software wise is that i only just installed hh bootldr on it, and then flashd this image, the other had ahd 0.9.0 on it Sep 01 14:56:28 illl reflash the other just in case that has anyeffect , if not ill open em up and check the hardware out Sep 01 14:57:35 hmm... guess i should try to sleep Sep 01 14:57:38 g'night al Sep 01 14:57:38 l Sep 01 14:59:13 qpe silently does nothing and exits after sauing its going into demon mode tho :( Sep 01 15:02:57 I think you can do -nodaemon (or similar) to see if it has anything else to say Sep 01 15:03:53 anyone know why mimedir-recurrence.h would have compiled, but not be installed in /oe/tmp/staging/arm-linux/include/mimedir-1.0/mimedir/ Sep 01 15:03:55 ??? Sep 01 15:04:20 because libgpevtype-0.12 needs it, but can not find it Sep 01 15:05:10 i guess not... Sep 01 15:05:17 time for a bug report... Sep 01 15:05:48 sounds like a bug in libmimedir. Sep 01 15:06:27 hm Sep 01 15:08:31 can anybody tell me waht DC00R is in this context? MQ-200: DC00R Sep 01 15:08:43 or even MQ-200: DC00R = EF2082AX Sep 01 15:11:06 DC00R is the first device control register. Sep 01 15:11:54 would this be a fixed value then in harware? or can it be dynaically assigned? cs my 2 simpads show a diff value for this and the display works on one and not the other, just wondered if it was relevant Sep 01 15:13:52 well...i fixed it by copying over mimedir-recurrence.h over to where it could be found Sep 01 15:14:00 and now everything is going smooth again Sep 01 15:14:00 off to work l 8r all Sep 01 15:14:58 oh im being stupid, they are the smae, just for some reason one simapd all the outpput text in the consle is capitalised! including normal commands ! Sep 01 15:16:21 oh, heh Sep 01 15:16:42 to answer your question, that register can be written by the software, but for any given hardware configuration there is only likely to be one correct value for it. Sep 01 15:17:20 ok thanks ;) Sep 01 15:17:52 and now i switched back to the othersimpad again the capitlas have returned to normal! very odd! Sep 01 15:23:13 hmm, the two MQ200 IC's are clearly from a difemrent manufactiring plant, but do have identical part numbers, would anbody ekse be able to check there parts to see if this has any signifcacne? Sep 01 15:29:29 * JustinP starts a new e-image compile form scratch just to be totally sure it works Sep 01 15:30:17 k, ive reset the working one and now its not working! no strange pattern, but a clear blank screen Sep 01 15:36:28 we have joy we have fun coding in the sun... Sep 01 15:36:38 dan hm only zecke knows or not Sep 01 15:37:30 dan and my simpad is at zecke's home Sep 01 15:37:43 when i have my distro set to familiar-0.8.3 does it not reflect that when i begin compilling gpe-image? Sep 01 15:38:30 the image will be for/of familiar .8.3 Sep 01 15:38:32 LoDown: it should....why do you ask? Sep 01 15:38:46 not sure what you mean by reflect Sep 01 15:39:23 because its not... Sep 01 15:39:28 i have the line: Sep 01 15:39:38 DISTRO = "familiar-0.8.3" Sep 01 15:39:49 i noticed that too Sep 01 15:39:59 but it is saying Sep 01 15:40:00 DISTRO = "familiar-0.8.3" Sep 01 15:40:04 when it begins to compile Sep 01 15:40:08 did I do something wrong? Sep 01 15:40:24 am i missing something here? Sep 01 15:40:30 DISTRO = "familiar" Sep 01 15:40:32 dang it Sep 01 15:40:42 i can't copy and paste worth a crap Sep 01 15:41:02 did you change versions after you started compiling the image? Sep 01 15:41:06 it says "familiar", shouldn't it say "familiar-0.8.3" Sep 01 15:41:07 no Sep 01 15:41:09 i think if you look in the distro/familiar-0.8.3.conf it sets DISTRO="familar" Sep 01 15:41:35 probably inherits familiar Sep 01 15:42:01 ah...so it does Sep 01 15:42:21 it sets it in familiar-0.8.3.conf to "familiar" Sep 01 15:42:48 then it's as expected Sep 01 15:42:51 stop freaking out ;-) Sep 01 15:42:56 haha Sep 01 15:43:16 i was worried i had messed up...I'm having the hardest time of this... Sep 01 15:43:17 indeed, mone says "openzaurus" even though I set it as -3.5.4 Sep 01 15:43:22 mine Sep 01 15:43:36 I can't seem to get the image to compile worth a damn... Sep 01 15:43:48 trying again :( Sep 01 15:44:13 i will figure this out :-D Sep 01 15:44:22 does anyone know why my glib-2.0 link crashes due to a unresolved external located in the libiconv library? Sep 01 15:44:47 JustinTodd: do you really mean crashes? Sep 01 15:44:57 JustinTodd: is this during build or running on your device? Sep 01 15:44:59 for some reason when i bitbake glib-2.0 it doesnt download libiconv even t hough it includes and links to it Sep 01 15:45:30 my linker complains. its not a crash Sep 01 15:45:45 ok Sep 01 15:45:55 are you sure libiconv was built? Sep 01 15:46:00 during the compilation/link of the glib so Sep 01 15:46:05 Im sure it was not Sep 01 15:46:13 for some reason, bitbake doesnt fetch it. Sep 01 15:46:21 are you doing any strange ASSUME_PROVIDED or PREFERRED_VERSION stuff? Sep 01 15:46:35 did you alter you bb file in any way? Sep 01 15:46:59 I see virtual/libiconv in glib-2.0's DEPENDS and it built fine for me earlier today from scratch Sep 01 15:47:03 No I do not. and No. I browsed it for references to libiconv (none existed)... but i didnt modify Sep 01 15:47:14 let me recheck mine Sep 01 15:47:16 which version? Sep 01 15:47:23 2.6.5 Sep 01 15:47:42 there is no glib-2.0_2.6.5 Sep 01 15:47:43 03rpurdie 07org.oe.dev * r43feaf25... 10/packages/linux/ (12 files in 2 dirs): linux-oz-2.6: Add 2.6.13-mm1 as a new development branch (for now). Several patches merged with mainline, others needed updates. Also some pm changes to add fatal battery and batter cover interrupt functionality. Sep 01 15:48:11 at least not in my checkout (which is the latest) Sep 01 15:48:21 glib-2.0-native_2.6.5 Sep 01 15:49:37 cat glib-2.0-native_2.6.5.bb | grep libiconv Sep 01 15:49:37 (nothing) Sep 01 15:49:38 last i remember iconv is part of glib Sep 01 15:50:27 atleast for familar Sep 01 15:50:37 JustinTodd: I think the expectation is that your host system will have iconv as standard. Sep 01 15:50:49 ah....-native Sep 01 15:51:20 well, there is no libiconv-native bb file, so.... Sep 01 15:51:51 no, i think kergoth fixed everything that needed it with uclibc Sep 01 15:52:09 obviously not since JustinTodd is having problems.... Sep 01 15:52:11 sorry. could i get a bit of clarification on naming conventions? -native would indicate what exactly? Sep 01 15:52:20 is JustinTodd building for uclibc? Sep 01 15:52:28 JustinP, you read what pb_ wrote? Sep 01 15:52:37 just regular ole' familiar linux 0.8,3 Sep 01 15:52:47 emte: of course. Looks like JustinTodd doesn't have it, though Sep 01 15:52:54 in his host Sep 01 15:53:02 emte: it would be prudent then to possibly add a -native variant of libiconv Sep 01 15:53:11 Host as in my cross compiler environmnet? Sep 01 15:53:15 JustinTodd: "-native" indicates that the package is intended to run on the machine where you're building. Sep 01 15:53:23 ikes!!! Sep 01 15:53:23 JustinTodd: as in installed on your host system Sep 01 15:53:54 does that indicate that my local.conf is incorrect? Sep 01 15:54:36 JustinTodd: some native tools always have to be compiled by OE to finish things. I'm not sure what depends on glib-2.0 native but I also have it in my work directory Sep 01 15:54:39 generally, there is an assumption in oe that the build system will be a (more or less) fully featured host. If your build host really is running uclibc natively, I can imagine that you would indeed be having trouble. Sep 01 15:54:39 hence, no, nothing wrong Sep 01 15:54:56 is JustinTodd's host system running uclibc? Sep 01 15:55:16 No it is not Sep 01 15:55:32 I'm not sure that glib is a part of a fully featured host Sep 01 15:55:40 he may be using a stripped version of glib that is missing iconv tho Sep 01 15:55:43 CosmicPenguin: iconv, not glib Sep 01 15:55:43 for clarification: my host system requires libiconv, correct? Sep 01 15:55:45 I mean, it will probably be there Sep 01 15:55:59 JustinTodd: that's the gist of what we're saying, yes Sep 01 15:56:05 JustinTodd: not exactly. the iconv functionality is part of libc on "normal" systems. Sep 01 15:56:05 JustinTodd: although it doesn't have to be that way Sep 01 15:56:14 oh, iconv ... Sep 01 15:56:19 Ok, thats different Sep 01 15:56:35 oh no. this sounds like trouble Sep 01 15:56:47 JustinTodd: what is your host system? Sep 01 15:57:15 CosmicPenguin: oe does include glib-2.0-native, specifically because glib isn't guaranteed to be installed on every host. Sep 01 15:57:20 The cross compiler is being run on a of-the-shelf version of slackware 10.0 Sep 01 15:57:25 pb_: indeed Sep 01 15:57:42 pb_: except that there is no libiconv-native bb file Sep 01 15:57:54 night all Sep 01 15:57:59 JustinTodd: that should be fine. what's the actual error? Sep 01 15:58:04 one second Sep 01 15:59:10 there isnt supposed to be a "libiconv.bb" its part of glib now Sep 01 15:59:25 or is it gettext Sep 01 15:59:28 one or the other Sep 01 15:59:54 gconvert.c:47:2: #error GNU libiconv not in use but included iconv.h is from lib Sep 01 15:59:54 make[4]: *** [gconvert.lo] Error 1 Sep 01 16:00:33 oh dear. it's finding the wrong iconv.h. Sep 01 16:00:37 if you take out that error macro, it results in some link errors about unresolved externals (libiconv_*) Sep 01 16:00:42 what did the compile command that caused that error look like? Sep 01 16:02:00 sorry for the spaM: Sep 01 16:02:00 make[4]: Entering directory `/root/downloads/openembedded/oe_temp/work/glib-2.0-native-2.6.5-r0/glib-2.6.5/glib' Sep 01 16:02:00 if /bin/sh ../i686-linux-libtool --tag=CC --mode=compile ccache gcc -DHAVE_CONFIG_H -I. -I/root/downloads/openembedded/oe_temp/work/glib-2.0-native-2.6.5-r0/glib-2.6.5/glib -I.. -I/root/downloads/openembedded/oe_temp/work/glib-2.0-native-2.6.5-r0/glib-2.6.5 -DG_LOG_DOMAIN=\"GLib\" -DG_DISABLE_CAST_CHECKS -DG_DISABLE_ASSERT -DG_DISABLE_CHECKS -DG_DISABLE_DEPRECATED -DGLIB_COMPILATION -I/root/downloads/openembedded/oe_temp/staging/i686-linux/include -pthread Sep 01 16:02:00 then mv -f ".deps/gconvert.Tpo" ".deps/gconvert.Plo"; else rm -f ".deps/gconvert.Tpo"; exit 1; fi Sep 01 16:02:03 ccache gcc -DHAVE_CONFIG_H -I. -I/root/downloads/openembedded/oe_temp/work/glib-2.0-native-2.6.5-r0/glib-2.6.5/glib -I.. -I/root/downloads/openembedded/oe_temp/work/glib-2.0-native-2.6.5-r0/glib-2.6.5 -DG_LOG_DOMAIN=\"GLib\" -DG_DISABLE_CAST_CHECKS -DG_DISABLE_ASSERT -DG_DISABLE_CHECKS -DG_DISABLE_DEPRECATED -DGLIB_COMPILATION -I/root/downloads/openembedded/oe_temp/staging/i686-linux/include -pthread -I/root/downloads/openembedded/oe_temp/staging/i686-linux/ Sep 01 16:02:14 oh no Sep 01 16:02:27 oh no Sep 01 16:02:43 mm, that looks ok to me Sep 01 16:02:52 ~lart JustinTodd for running bitbake as root Sep 01 16:02:53 * ibot shoves a crumpet down JustinTodd's throat, happy now?! Huh? Want some JAM with that? for running bitbake as root Sep 01 16:03:53 mmm crumpets! Sep 01 16:04:15 JustinTodd: can you paste the first few lines from your /usr/include/iconv.h here? Sep 01 16:04:23 so now the question is did BB escape and install a wrong iconv in the host? Sep 01 16:04:34 sure one sec Sep 01 16:04:40 JustinTodd: do *NOT* run bitbake as root Sep 01 16:05:02 of course it may be way too late for that now... Sep 01 16:05:16 crap :( Sep 01 16:05:34 it *shouldn't* be....but it may Sep 01 16:05:45 night Sep 01 16:05:50 night woglinde Sep 01 16:05:57 bitbake *should* stay sandboxed but it's never a good idea to run it as root...just in case it escapes Sep 01 16:06:07 then you have familiar files all over your host system Sep 01 16:06:38 #include #define __need_size_t #include __BEGIN_DECLS Sep 01 16:06:38 /* Identifier for conversion method from one codeset to another. */ typedef void *iconv_t; Sep 01 16:07:30 i have one in /usr/local/include too :S. could this be a problem? Sep 01 16:07:38 ah Sep 01 16:07:44 yes, that would be a problem Sep 01 16:08:06 * JustinP sings "One of these things is not like the other" Sep 01 16:08:44 the stuff you pasted looks fine, but gcc will use the file in /usr/local in preference to the one in /usr. Sep 01 16:09:18 so, if the file in /usr/local is from libiconv, it would definitely cause the problem you are seeing. Sep 01 16:09:19 as a temporary solution, can i rename my /usr/local/include? im really not sure why its there. i think an application once complained i didnt have it so i reinstalled it, or... installed a second cooy Sep 01 16:09:32 yeah, that should be fine Sep 01 16:09:39 rename my header file in /usr/local/include* Sep 01 16:09:55 ok thanks Sep 01 16:12:01 mmm beer Sep 01 16:12:05 i am hungry ... Sep 01 16:14:46 ttf-fonts/ttf-dejavu_1.13-20050820 needs to be updated to 20050828 because 20050820 is no longer available Sep 01 16:15:39 I made a new bb for it (like it takes more than ten seconds) who do I ask to update it? Sep 01 16:16:20 Zero_Chaos: put it in bugs.openembedded.org Sep 01 16:16:48 JustinP: I figured that would be rude, but, okay Sep 01 16:16:59 why would that be rude? Sep 01 16:17:16 that's why it's there....to keep track of bugs Sep 01 16:17:32 I didn't really think of it as a bug... Sep 01 16:17:45 "can't download tarball" qualifies as a bug Sep 01 16:17:59 wow, so many bugs to submit. Sep 01 16:18:10 ? Sep 01 16:18:19 JustinP: what if I made a new .bb for aircrack and kismet, where do I put those? Sep 01 16:18:30 same place Sep 01 16:18:48 why, are they not working? Sep 01 16:18:58 new programs should also go there Sep 01 16:18:59 JustinP: they do work, but they are ancient Sep 01 16:19:13 basically if you don't have commit access and you want to get something in submit it to the bug tracker Sep 01 16:19:44 just make sure these things work before submitting them Sep 01 16:19:55 make sure to fill out all of the normal vars in the bb file Sep 01 16:19:59 JustinP: they do, but thanks for the tip Sep 01 16:20:29 and remember that OE is for multiple distributions / devices. If these are device specific you'll have touble Sep 01 16:20:37 JustinP: and If I wrote the new .bb file, but I am not planning on being the maintainer, where do I put my name? Sep 01 16:20:43 you don't Sep 01 16:20:58 it *should* be in the commit message but may not be Sep 01 16:21:09 you'll have your name in the bug report at least Sep 01 16:21:18 if you get enough in and want it you may end up with commit access Sep 01 16:21:39 JustinP: and what if my .bb breaks and someone complains to the maintainer even though it's my file :-) Sep 01 16:22:03 Zero_Chaos: feel free to put "this file contributed by " in a comment in the .bb file, if you like. Sep 01 16:22:13 Zero_Chaos: if the person who commits doesn't want to be the maintainer they'll just leave it blank Sep 01 16:22:19 that too ;-) Sep 01 16:22:31 and if there's no MAINTAINER it will come back to the bug tracker Sep 01 16:22:43 pb_: that's a good idea, I just don't want to submit a file without giving people a place to complain Sep 01 16:23:03 JustinP: pb_: thank you gentlemen. Sep 01 20:09:13 is anyone in here? Sep 01 20:44:46 perhaps Sep 01 20:44:54 whtcha need? Sep 01 21:03:32 hey raster Sep 01 21:03:53 blah Sep 01 21:04:05 yeah i hear ya Sep 01 23:08:31 arrrgh **** BEGIN LOGGING AT Thu Sep 01 23:13:44 2005 Sep 02 00:28:14 03nail 07org.oe.dev * r75e0c24b... 10/packages/meta/openslug-packages.bb: openslug-packages: Remove broken watchdog from feed Sep 02 00:48:45 morning Sep 02 00:59:28 hi Sep 02 01:05:59 morning all Sep 02 01:06:32 morning all Sep 02 01:07:06 morning ade|desk Sep 02 01:08:03 time to 2.6.13-mm1 (which is older then 2.6.13-rc6-mm2 for bitbake) Sep 02 01:20:40 woohoo my first opie & gpe builds worked.... :) Sep 02 01:27:22 someone use(d) Linspire? (Linux distro) Sep 02 01:49:12 d'oh opie image with opie build kernel keeped saying no space left on the device then started opie but didn't display the colorful bit just a black'ish screen Sep 02 01:50:21 ZonE-HaKin9: what device? Sep 02 01:50:42 5500 sory Sep 02 01:51:08 just about to try the 64-0 kernel see if it works with that Sep 02 01:51:30 ZonE-HaKin9: collie images are full Sep 02 01:51:43 do u think that will help?... Sep 02 01:52:01 64-0 are best when have sd or cf storage Sep 02 01:52:30 what do u mean...? it was the first boot.. i'll be running it off a 1gb sd when i get it installed Sep 02 01:52:41 or a 256mb sd Sep 02 01:54:11 so whats the go... do i need to install the bootstrap or gpe image.. then altboot to the sd so opie will work due to lack of space?? Sep 02 01:55:12 you can remove: opie-style-liquid opie-style-metal - this will give ~500K Sep 02 01:55:31 then remove kernel-module-ipsec openswan - if you dont use ipsec (~1.5M) Sep 02 01:55:48 if you dont need ipv6 then remove kernel-module-ipv6 Sep 02 01:56:15 then remove wifi drivers which you dont need (hostap-modules* or/and orinoco-modules*) Sep 02 01:56:21 nar its kool, just need to know that.. i run off a sd card anyway.. and just use the int mem just as a backup os Sep 02 01:56:42 ThanX tho Sep 02 01:59:31 btw got a bit of a laugh for ya... we were watching hitch hickers guide to the galaxacy wide screen version mpeg-1 on my bosses c3100 todday at work hehe lol Sep 02 02:01:41 ;)) Sep 02 02:01:58 shit.. looks like I got ipkg problem... Sep 02 02:02:19 0.99.152 has a problem with provides: field.. but 0.99.153 has a problem with depends: field ;( Sep 02 02:03:05 speak of the devil... hey kev Sep 02 02:15:55 hi Phil Sep 02 02:16:31 hi hrw|work Sep 02 02:19:46 shit.. ipkg 152 has one problem, 153 has other.. ;( Sep 02 02:24:28 shit.. ipkg 152 has one problem, 153 has other - looks like it does not look into depends: :( Sep 02 02:25:36 745696 opie-image-c7x0-20050902084603.rootfs.tar.bz2 Sep 02 02:25:55 15047806 opie-image-c7x0-20050728152428.rootfs.tar.bz2 Sep 02 02:26:02 'small' difference.. Sep 02 02:27:46 hrw|work: do u guys use that uclx program to compress your apps? Sep 02 02:27:58 no Sep 02 02:28:10 hrw|work: any reason? Sep 02 02:29:35 I never used it so dont know Sep 02 02:31:25 check out http://inter.zon.free.fr/zaurus_index.html#uclx Sep 02 02:31:35 I know what it is Sep 02 02:32:24 oh sorry, i use it for a few of the sec apps i was building on my 5500 and it was good from what i could tell Sep 02 02:33:11 I can add it into OE anyway Sep 02 02:33:49 hrw|work: kool maybe useful for the 5500 opie image Sep 02 02:37:33 ZonE-HaKin9: jffs2 is already compressed Sep 02 02:37:53 heh.. it looks like getting it to build will be a challenge.. first find which version of libucl author used, then try to build it, then try to build uclx with it.. Sep 02 02:37:56 argh.. Sep 02 02:38:04 first I have to catch ipkg .153 bug Sep 02 02:39:28 it looks like if there is only 1 provider for package it ignore that package Sep 02 02:40:55 hrw|work: ooops **** ENDING LOGGING AT Fri Sep 02 02:59:56 2005