**** BEGIN LOGGING AT Thu Oct 18 02:59:57 2007 Oct 18 06:09:04 * * OE Bug 3173 has been created by bluelightning(AT)bluelightning.org Oct 18 06:09:06 * * PATCH: libopieobex: disable bluetooth detection Oct 18 06:09:08 * * http://bugs.openembedded.org/show_bug.cgi?id=3173 Oct 18 06:41:03 goodmorning all ! Oct 18 06:41:28 koen : After the last meta-toolchain commits it builds fine on x86 Oct 18 06:41:54 koen : BTW should we bump PR on the meta-toolchain package itself ? Oct 18 06:57:14 morning Oct 18 07:04:49 Henryk: distro package arch are *very* bad Oct 18 07:04:59 Henryk: I'd almost say "forbidden" in OE Oct 18 07:09:45 does someone built openmoko-browser2 r3082? Oct 18 07:13:16 hrw: http://www.angstrom-distribution.org/repo/?action=details&pnm=openmoko-browser2 Oct 18 07:14:37 ok Oct 18 07:18:56 rebuild solved Oct 18 07:27:32 morning everyone Oct 18 07:36:04 * * OE Bug 3171 has been RESOLVED (INVALID) by Oct 18 07:36:06 * *  gcc-cross-sdk-4.1.2-r6-do_unpack Oct 18 07:36:08 * * http://bugs.openembedded.org/show_bug.cgi?id=3171 Oct 18 07:36:15 * * OE Bug 3172 has been RESOLVED (FIXED) by Oct 18 07:36:17 * *  uclibc-0.9.29-r8-do_configure Oct 18 07:36:19 * * http://bugs.openembedded.org/show_bug.cgi?id=3172 Oct 18 07:36:36 03koen 07org.oe.dev * r6a20ffdd... 10/ (2 files in 2 dirs): xserver-kdrive-common: nothing support PACKAGE_ARCH = "", we don't want those, and you can't build 2 distros in the same TMPDIR anyway. Oct 18 07:41:39 hi everyone ... is there a way to clean a OE tmp directory ? do the morgue directories have something to do with it ? Oct 18 07:42:57 Longfield: remove tmp/work and tmp/deploy/*/morgue Oct 18 07:43:07 and stamps Oct 18 07:43:09 and staging Oct 18 07:43:13 and cross Oct 18 07:44:25 koen: clean != remove Oct 18 07:44:41 in OE it is :) Oct 18 07:46:45 hum, I use the cross-compiler in cross as my general cross-compiler too ... so I don't want to remove it Oct 18 07:47:03 * hrw do another attempt to croot gllin Oct 18 07:48:22 hrw: try LD_PRELOADing the oabi glibc Oct 18 07:49:44 koen: and pthread and util and m and... Oct 18 07:52:19 http://www.angstrom-distribution.org/unstable/toolchain/ Oct 18 07:52:35 when last time I played with gllin it atleast gave nmea without fix etc. now it is quiet Oct 18 07:53:30 * gilligan_ googles for gllin Oct 18 07:53:39 ah..some userspace driver Oct 18 07:55:04 oh by the way.. has anyone here done/read some jffs2 endurance tests ? like if one can actually rely on it remaining in a consistent state when switching of power and such Oct 18 07:55:57 koen: how do you build the angstrom toolchain ? what is the recipe for that ? I saw a mail of Mickey about it yersterday on the mailing-list ... I am going to need to provide a toolchain or sdk for my board user and I am very interested in this Oct 18 07:56:11 'bitbake meta-toolchain' Oct 18 07:56:25 but I am pretty sure it's broken atm Oct 18 07:56:45 because the bitbake meta-sdk or so has never worked for me until now Oct 18 08:18:34 koen: yep, it is broken atm: http://pastebin.ca/740812 Oct 18 08:59:02 morning all Oct 18 09:04:33 good morning all Oct 18 09:05:12 hi koen :) Oct 18 09:27:06 "md5sum libbonobo-2.10.0.tar.bz2" someone? Oct 18 09:32:09 f1f0255f94e3354250d142b688013fad /home/koen/OE/downloads/libbonobo-2.14.0.tar.bz2 Oct 18 09:32:09 22bd67a0a8b7f156c02c5fc08fb1fa24 /home/koen/OE/downloads/libbonoboui-2.18.0.tar.bz2 Oct 18 09:32:10 14427a459f6ca355de5f5e47ce95c1c4 /home/koen/OE/downloads/libbonoboui-2.20.0.tar.bz2 Oct 18 09:32:14 that's all I hae Oct 18 09:32:52 wait Oct 18 09:32:56 68179b91f5043ce0ef050dbbed04979b libbonobo-2.10.0.tar.bz2 Oct 18 09:32:57 34046c6880f4c90e47a1acae34b74d26 libbonobo-2.13.0.tar.bz2 Oct 18 09:32:57 f1f0255f94e3354250d142b688013fad libbonobo-2.14.0.tar.bz2 Oct 18 09:32:58 bd4fb92f993b7fb7e660bb999465ef3b libbonoboui-2.10.0.tar.bz2 Oct 18 09:32:58 dc26dc17cddc625cac37ecfab263a51a libbonoboui-2.14.0.tar.bz2 Oct 18 09:32:59 22bd67a0a8b7f156c02c5fc08fb1fa24 libbonoboui-2.18.0.tar.bz2 Oct 18 09:32:59 14427a459f6ca355de5f5e47ce95c1c4 libbonoboui-2.20.0.tar.bz2 Oct 18 09:36:26 ~fish lrg Oct 18 09:36:27 * ibot slaps lrg around with a large trout Oct 18 09:38:10 hey Oct 18 09:38:40 hey XorA, zecke Oct 18 09:39:01 moin zecke Oct 18 09:40:03 zecke: not going to make foss.in, submission date is too soon for papers - too much customer support :-/ Oct 18 09:40:14 68179b91f5043ce0ef050dbbed04979b libbonoboui-2.10.0.tar.bz2 Oct 18 09:40:34 "using wolfson codecs to make sound" Oct 18 09:40:46 koen: so my today morning script will be usable for you when I push it into contrib Oct 18 09:40:49 hey gilligan_, lrg & zecke Oct 18 09:41:30 lrg: just turn something in? Oct 18 09:41:58 lrg: "Do ASoC development, okaay?!" would be a nice title Oct 18 09:42:39 my university is a strange place, everything renovated... and people try to sell you stuff Oct 18 09:42:52 hey koen Oct 18 09:43:16 zecke: it was a PM talk for SoC CPU power saving Oct 18 09:43:29 * gilligan_ whispers a little prayer to himself thanking god that his university time is over... Oct 18 09:54:38 good morning Oct 18 09:54:50 hi florian Oct 18 09:54:56 hey florian, hrw Oct 18 09:55:16 hi liam Oct 18 09:57:11 lrg: please turn something in, they will probably just accept you Oct 18 09:57:46 XorA: I'm going to hold off adding these to OE untill after the weekend: http://lists.openmoko.org/pipermail/gsmd-devel/2007-October/000257.html Oct 18 09:58:19 zecke: /me did miss your talk at T-DOSE ( it was first talk on sunday morning ). Was your talk recorded on video? Oct 18 09:59:01 it wasn't Oct 18 09:59:44 zecke: nokia dumped opera, but they sadly went for mozilla Oct 18 09:59:50 A pity. Will the sheets be published? Oct 18 10:00:26 stappers: yes, but sheets are only one thing Oct 18 10:01:11 koen: for the n810? Oct 18 10:01:23 for it2008 OS Oct 18 10:01:42 ah right works on both n8x0 because nokia is so nice to support old devices Oct 18 10:01:46 they finally figures out how to build an image for two devices Oct 18 10:01:56 (by making the devices identical) Oct 18 10:02:16 the n810 is just n800 + gps + keyboard - sd Oct 18 10:02:16 okay, I head to class now Oct 18 10:02:17 n810 is faster according to linuxdevices Oct 18 10:02:19 ~seen woglinde Oct 18 10:02:20 woglinde was last seen on IRC in channel #oe, 1d 18h 23m 17s ago, saying: 'bye'. Oct 18 10:02:39 hrw: according to nokia it's the same cpu and ram Oct 18 10:23:00 koen: prepare for oe-source-checker.py Oct 18 10:28:16 koen: pull and look into contrib/source-checker - run it against dl_dir Oct 18 10:31:29 koen@bitbake:~/OE/monotone/org.openembedded.dev/contrib/source-checker$ python oe-source-checker.py ~/OE/downloads/ Oct 18 10:31:29 koen@bitbake:~/OE/monotone/org.openembedded.dev/contrib/source-checker$ Oct 18 10:31:57 03hrw 07org.oe.dev * r714c316a... 10/ (1 packages/mathomatic/mathomatic_unstable.bb): mathomatic: mark unstable version as BROKEN Oct 18 10:32:05 03hrw 07org.oe.dev * rd0591d73... 10/ (1 conf/checksums.ini): checksums: removed am.tgz which is generated daily Oct 18 10:32:10 03hrw 07org.oe.dev * r72dc4c03... 10/ (1 conf/checksums.ini): (log message trimmed) Oct 18 10:32:10 checksums: integrate changes from Poky Oct 18 10:32:10 r2906 | hrw | 2007-10-18 11:44:17 +0200 (Cz, 18 X 2007) | 1 line Oct 18 10:32:10 checksums: drop daily generated pa_snapshot Oct 18 10:32:10 ------------------------------------------------------------------------ Oct 18 10:32:11 r2905 | hrw | 2007-10-18 11:43:47 +0200 (Cz, 18 X 2007) | 1 line Oct 18 10:32:13 checksums: fix fakechroot checksum Oct 18 10:32:17 03hrw 07org.oe.dev * rb5ad5b3b... 10/ (3 files in 3 dirs): source-checker: small Python script to check local sources against conf/checksums.ini Oct 18 10:32:36 koen: read script before use? Oct 18 10:32:57 hrw: implement better feedback? Oct 18 10:33:11 ok Oct 18 10:33:22 "koen: pull and look into contrib/source-checker - run it against dl_dir" Oct 18 10:33:35 ;) Oct 18 10:34:34 koen@bitbake:~/OE/monotone/org.openembedded.dev/contrib/source-checker$ python oe-source-checker.py ../../conf/checksums.ini ~/OE/downloads/ Oct 18 10:34:34 koen@bitbake:~/OE/monotone/org.openembedded.dev/contrib/source-checker$ Oct 18 10:34:41 I assume that's a good sign? Oct 18 10:34:59 yes Oct 18 10:37:58 better than the results on my machine or the poky autobuilder :) Oct 18 10:38:24 but then my DL_DIR has files from 2002 in it... Oct 18 10:39:11 hrw: how goes your project to get Marvell CF wifi working? Oct 18 10:39:21 RP: how to check size of list in python? Oct 18 10:39:29 a = [1,2,3] Oct 18 10:39:44 "print a.len()" == 3 Oct 18 10:39:54 RP: what to use as .len()? Oct 18 10:40:17 XorA: its rather Schurig project - I just started it Oct 18 10:41:04 * XorA needs 802.11g on his Zaurus Oct 18 10:42:04 hrw: Don't you just want listname.len() ? Oct 18 10:42:28 XorA: 2.6.24-rc will have it Oct 18 10:42:43 hrw: it returns almost instantly Oct 18 10:42:48 RP: AttributeError: 'list' object has no attribute 'len' Oct 18 10:42:58 angstrom@serenity:~/website/unstable/sources$ python /tmp/oe-source-checker.py /tmp/checksums.ini . Oct 18 10:42:59 angstrom@serenity:~/website/unstable/sources$ Oct 18 10:43:09 hrw: Marvell stuff got that far advanced, wow Oct 18 10:43:15 real 0m0.199s Oct 18 10:43:15 koen: I probably know why Oct 18 10:43:15 user 0m0.188s Oct 18 10:43:16 sys 0m0.009s Oct 18 10:43:28 koen: lack of oe_sha256sum in PATH? Oct 18 10:43:45 yes Oct 18 10:44:24 hrw: Try len(listname) Oct 18 10:44:40 thx this works Oct 18 10:44:44 * hrw is python noob Oct 18 11:00:30 someone has idea how to check does command exists in PATH? Oct 18 11:00:48 My OE refuses to make images: Task do_rootfs always fails :-( Oct 18 11:02:47 http://www.pastebin.ca/740946 that's the main error Oct 18 11:03:41 and that's the do_rootfs log http://www.pastebin.ca/740949 Oct 18 11:04:28 dcordes: you play with IMAGE_FSTYPES? Oct 18 11:05:36 gm Oct 18 11:06:44 hi likewise Oct 18 11:06:47 good morning likewise Oct 18 11:07:05 dcordes: still chasing the jffs2 issue? Oct 18 11:07:13 hrw: yes I change the IMAGE_FSTYPES from jffs2 to tar.gz in order to make sure it is not the jffs2 Oct 18 11:07:20 but it didn't work with both Oct 18 11:07:24 likewise: yep Oct 18 11:07:57 IMAGE_FSTYPES ?= "tar.gz" Oct 18 11:11:46 hrw: any idea? Oct 18 11:16:01 dcordes: IMAGE_FSTYPES += "tar.gz" Oct 18 11:16:12 ah that's a problem in the tree then Oct 18 11:16:14 zaurus machine config require jffs2 image format Oct 18 11:16:29 hrw: Oct 18 11:16:29 hrw: 13:06 < likewise> dcordes: still chasing the jffs2 issue? Oct 18 11:16:29 13:06 < dcordes> hrw: yes I change the IMAGE_FSTYPES from jffs2 to tar.gz in order to make sure it is not the jffs2 Oct 18 11:16:33 13:07 < dcordes> but it didn't work with both Oct 18 11:16:46 oh no putty... sorry. http://www.openembedded.org/repo/org.openembedded.dev/conf/machine/akita.conf Oct 18 11:17:31 dcordes: then IMAGE_FSTYPES = "jffs2 tar.gz" in local.conf and build Oct 18 11:17:44 w/o the plus? Oct 18 11:17:56 order of incldued files matter Oct 18 11:19:07 hrw: NOTE: build 200710181316: completed. should I make a bug? Oct 18 11:22:52 dcordes: bug about? Oct 18 11:23:29 hrw: about the ? thing in akita.conf Oct 18 11:23:43 it is proper set in akita.conf Oct 18 11:23:54 as akita use jffs2 for rootfs Oct 18 11:24:11 but it didn't build with the ?= Oct 18 11:24:14 if user wants to build other formats then let remember that jffs2 is req too Oct 18 11:34:19 koen: does new version works for you? Oct 18 11:34:35 didn't try yet Oct 18 11:35:14 koen : looks like there is something warped with package QA on meta-toolchain Oct 18 11:35:37 meta-toolchain needs more fixes (see ml) Oct 18 11:37:03 koen : efika build failed at binutils-cross-sdk QA and then when i re-issue bitbake meta-toolchain.... it passed without errors... Oct 18 11:37:25 i think i have seen that behaviour before Oct 18 11:37:30 with QA Oct 18 11:38:39 steliosk: yes, QA behaviour. Oct 18 11:38:54 steliosk: its a limitation of insane.bbclass it doesnt undo what was borked, so when you restart you are past that point Oct 18 11:40:20 that's not good :( Oct 18 11:41:12 Probably we need to add some marker at the start of the check, and remove it after the checks are done Oct 18 11:42:35 because what is actually happening is that insane.bbclass blows up Oct 18 11:42:47 that does not look a proper way to exit -> http://rafb.net/p/Qj8LS449.html Oct 18 11:45:23 koen : any idea why binutils-cross-sdk will fail again at the Qa, even after your last patches ? Oct 18 11:45:35 nope Oct 18 11:45:40 it works here Oct 18 11:45:54 http://www.angstrom-distribution.org/unstable/toolchain/ Oct 18 12:08:29 Filesystem Size Used Available Use% Mounted on Oct 18 12:08:29 /dev/mtdblock4 61.1M 60.5M 652.0k 99% / Oct 18 12:08:33 AU!! Oct 18 12:08:50 hrw: laptop? Oct 18 12:09:00 hrw: ah flashrom.. Oct 18 12:09:17 hrw: I thought 60 GB :-) Oct 18 12:09:52 not ROM! Oct 18 12:09:55 likewise: laptop has 80GB Oct 18 12:10:31 +apple_iphone MACH_APPLE_IPHONE APPLE_IPHONE 1506 Oct 18 12:10:33 heh Oct 18 12:16:24 hi, all! Oct 18 12:16:53 do anybody knows some free compile farm/build host serivices? Oct 18 12:21:49 hi cbrake Oct 18 12:28:05 * * OE Bug 3155 has been RESOLVED (FIXED) by dp(AT)xora.org.uk Oct 18 12:28:07 * *  Patch for libgstmad to generate 16-bit output instead of 32-bit Oct 18 12:28:08 * * http://bugs.openembedded.org/show_bug.cgi?id=3155 Oct 18 12:29:55 hrw: hello Oct 18 12:32:10 aaargh Oct 18 12:32:27 gail depends on libgnomecanvas which depends on gail Oct 18 12:32:46 solution: build libgnome-canvas 2.14 by hand, then build gail Oct 18 12:34:13 argh one Oct 18 12:34:29 if i want to download the sources alone, and then do an offline build, what to use with bitbake? (using OM, if i can mention that) Oct 18 12:34:41 mbuf: bitbake -cfetchall your-target Oct 18 12:34:45 bitbake -c fetchall Oct 18 12:34:50 thanks Oct 18 12:35:14 koen: can I abuse network of ewi with lot of fetching? Oct 18 12:35:24 hrw: sure Oct 18 12:35:42 then I will do fetchall world of OE to gain new checksums Oct 18 12:36:53 koen: did you checked angstrom source mirror quality? Oct 18 12:37:34 your script reported nothing Oct 18 12:39:09 cool Oct 18 12:41:10 koen: did it finally do anything though? Oct 18 12:41:16 no idea Oct 18 12:42:29 koen: How long did it run for? Oct 18 12:42:33 I can add counter into it Oct 18 12:42:46 0.2 seconds according to time Oct 18 12:43:25 koen: Its supposed to be running md5sum on every file in the directory so it isn't working Oct 18 12:48:09 I am adding progressbar into it Oct 18 12:50:18 Checking files: \ (0065/3181) [ 2 %] Oct 18 12:53:29 ~lart python for lack of "var++" Oct 18 12:53:29 * ibot makes a balloon animal out of python for lack of "var++" Oct 18 12:53:55 hrw: it has +=1 though Oct 18 12:54:51 hrw: how about making the checker script download the files that throw errors and checksum those automatically? Oct 18 12:55:04 koen: thats for later Oct 18 12:56:23 koen: it had to be simple script... Oct 18 12:59:28 koen: and because of you it is getting more and more fat Oct 18 12:59:30 hm.. i just got an bitbake exception executing -cdo_rootfs for a (self declared, but basically devimage-image) image package Oct 18 12:59:54 ah..oops Oct 18 13:00:13 my fault obviously :) Oct 18 13:00:17 * gilligan_ shupts up Oct 18 13:00:19 :) Oct 18 13:00:22 shuts even Oct 18 13:01:47 gilligan_: I have done some power fail testing with jffs2 -- a small app that writes to a file, and then yank power with a prog power supply. I ran this for a couple days and never had any jffs2 problems. Oct 18 13:02:34 gilligan_: but I was able completely kill a sandisk SD card with that method. It would not even respond to the initial SD commands. Oct 18 13:03:26 gilligan_: however, I still recommend putting a backup battery in your design if you are going to be yanking power on a regular basis -- Varta makes some nice Lith-poly bats that work well. Oct 18 13:03:54 Checked 1634 files. 1634 was OK and 0 had wrong checksums. Oct 18 13:03:54 real 6m41.131s Oct 18 13:04:04 cbrake: i see.. good to know Oct 18 13:08:33 koen: pushed new version Oct 18 13:19:58 when adding "python" to IMAGE_INSTALL I get an error from ipk during do_roofs -- 2 matching providers, python-core and python-curses .. not quite sure what this implies for me now to be honest.. any hints please? :) Oct 18 13:20:46 ah.. okay.. think i got it Oct 18 13:21:11 python-core and python-curses both provide the virtual "python" and i have to pick one Oct 18 13:22:47 select core Oct 18 13:23:02 iirc we fixed it? Oct 18 13:23:59 does the coreutils fully replace busybox functionality ? Oct 18 13:24:29 no - parts Oct 18 13:25:30 but it is safe to have both installed? Oct 18 13:25:46 yes Oct 18 13:29:55 hrw: http://rafb.net/p/a7dWwG79.html Oct 18 13:34:05 thx Oct 18 13:38:00 pushing new ver anyway Oct 18 13:39:41 15:39:04 (380.56 MB/s) - `xserver-common-1.11.tar.gz' saved [9178/9178] Oct 18 13:39:48 small files flies Oct 18 13:40:09 oe.org: weeh! FATAL: connection limit exceeded for non-superusers Oct 18 13:40:26 * likewise hopes the limit is 100.000 users :-) Oct 18 13:41:43 Should PACKAGE_ARCH adhere to some toolchain-defined architecture? Oct 18 13:41:57 More correctly: Should PACKAGE_ARCH adhere to some toolchain-defined architecture namespace? Oct 18 13:46:26 koen: few entries fixed, few are wrong on your side Oct 18 13:48:20 koen: ewi is nice machine for such tests Oct 18 13:53:27 koen : just noticed that angstrom for x86 makes only jffs2 image. is that by choice ? Oct 18 13:53:58 probably a bug in local.conf or something Oct 18 13:55:23 koen: please use root powers and remove /home/hrw/tmp/CONTENTS on ewi Oct 18 13:55:31 koen : i don't change the rootfs_type in local.conf..... Oct 18 13:58:42 koen: thx Oct 18 13:59:24 morning Oct 18 14:04:19 hi Gerwin Oct 18 14:04:33 * koen discovers http://geocanvas.sourceforge.net/ Oct 18 14:10:05 * hrw gave that url for few people Oct 18 14:16:03 cbrake: hm.. can't mount usb disks in the angstrom minimal installation running from NOR can I ? Oct 18 14:16:31 gilligan_: you should be able to Oct 18 14:16:57 gilligan_: I just updated my build and I'm working through a few issues with that right now Oct 18 14:17:39 cbrake: usb-storage is built into kernel? Oct 18 14:17:53 gilligan_: I think so Oct 18 14:19:49 cbrake: ah, right.. it actually works just fine Oct 18 14:21:54 ERROR: No buildable providers available for required build target compositeext ('['fixesext', 'xextensions']') Oct 18 14:22:01 someone has idea? Oct 18 14:22:12 c7x0 + angstrom-2007.1 Oct 18 14:24:45 ok - this is because of world Oct 18 14:27:22 cbrake: what do you usually base your full (nand or hdd or whatever) installations on ? fully self defined image or "console-image" or.. ? Oct 18 14:28:17 gilligan_: extension of console-image, but that is going to change shortly as I need to remove a bunch of stuff from it. It has just not been a priority yet with 128MB of NAND flash ... Oct 18 14:34:20 cbrake: hmm.. isn't /lib/modules automatically populated in images from the contents of the appropraite kernel provider ? Oct 18 14:36:25 gilligan_: there has to be a RRECOMMENDS for each module to be included. Oct 18 14:37:51 oh Oct 18 14:38:46 gilligan_: grep for RRECOMMENDS in the conf/machine dir -- I'm not sure if this is the best way -- need to think through it some. Oct 18 14:40:03 cbrake: maybe a sane set of default modules could always be included Oct 18 14:40:24 gilligan_: yes, that is what I would eventually like to get to Oct 18 14:40:50 gilligan_: for the cm-x270, I would like to include audio/ts, modules by default Oct 18 14:42:38 cbrake: well any set of defaults is better than /lib/modules : no such file or directory ;] And as long as it remains possible to tweak the default set everyone should be fine with it I guess Oct 18 15:02:07 hm, is there a way to retrieve all packages that will be installed to an image? the only related option that bitbake offers is the creation of a .dot file with the dependencies.. suppose with some sed,grep,sort,uniq etc a list could be created from that -- but that's not exactly convenient ;] Oct 18 15:03:15 gilligan_: pipe the dot file through dot to normalise it, then just awk it up. Oct 18 15:03:23 that's probably the quick and dirty way Oct 18 15:04:25 bit stupid really.. bitbake constructs a .dot and then we take it apart again to get the basic information... ;] Oct 18 15:05:15 i'll write something about that and other bitbake related questions I have on mind on the ml Oct 18 15:16:20 bye Oct 18 15:18:00 should i have a /dev/ts0 or something for my touchscreen? Oct 18 15:18:44 rater /dev/input/touchscreen0 Oct 18 15:18:47 /dev/input/touchscreen0 Oct 18 15:18:49 ;) Oct 18 15:20:16 ah - there it is! Oct 18 15:20:22 thanks hrw! Oct 18 15:20:42 (and thanks koen) Oct 18 15:28:04 ts0 is old fasion Oct 18 15:29:34 at least here :) . we use the input system and that defines /dev/input/event0 stuff Oct 18 15:34:45 so, is this bad? - Pointer option key (/dev/input/event0) of value ((null)) not assigned! Oct 18 15:35:12 i'm invoking with: /usr/bin/Xfbdev -mouse evdev,,/dev/input/event0 Oct 18 15:41:24 has anyone run x86/angstrom x11-image recently ? I am trying it on a p4 but the xserver crashes Oct 18 15:41:29 during boot Oct 18 15:43:40 Ah - was missing device= in front of /dev/input path. Oct 18 15:44:02 but still no touchscreen love. any debugging tips? Oct 18 15:44:55 Did you buy it flowers first? Oct 18 15:45:20 yes. breakfast in bed, too. Oct 18 15:45:23 (-; Oct 18 15:56:18 Crofton: Hey there! Oct 18 16:03:33 HopsNBarley: if you have a ts0 you probably enabled some legacy touchscreen interface in kernel, that own uses a differen events to send the touchscreen events. Oct 18 16:17:31 quickie for you gurus. I'm trying some changes for testing in inflate.c for my board. How should I get them to compile? I've been using the sledgehammer approach 'bitbake -c rebuild virtual/kernel' but of course, now that I'm getting into it, that silly technique replaces my sources in linux-2.6.21-r7/image/lim/module/2.6.21/source/lib/ .... Oct 18 16:22:22 russf, i'm no guru, but i just remove the compile stamp, and then rerun bitbake. all the right stuff happens. Oct 18 16:22:49 it's a little faster if you bitbake -b path/to/your/ikernel.bb Oct 18 16:25:06 HopsNBarley: thanks! Oct 18 16:26:18 HopsNBarley: is that path with in the tmp/work space? or where? Oct 18 16:27:15 the openembedded/packages/linux/your-kernel.bb Oct 18 16:27:58 k. Thx. Oct 18 16:51:08 HopsNBarley: -mouse tslib ? Oct 18 16:53:17 koen, bingo! Oct 18 16:53:41 * mwester threw a mouse in his trash this morning... Oct 18 16:54:54 t'was of the grey, non-electrical variety, however - so I don't think it would help HopsNBarley ;-) Oct 18 16:55:49 mwester, yeah verily, i did some similar mouse chucking last winter. (-; Oct 18 16:58:22 03likewise 07org.oe.dev * r34ae15e7... 10/ (3 files in 3 dirs): gcc-4.1.2: Added support for powerpc e300c3 core. Bumped PV for your pleasure. Oct 18 16:58:28 03likewise 07org.oe.dev * rf3e14b4a... 10/ (3 files in 3 dirs): binutils-2.16: Added support for powerpc e300c3 core. Bumped PV for your pleasure. Oct 18 16:58:33 03likewise 07org.oe.dev * r934c2d01... 10/ (1 conf/machine/mpc8313e-rdb.conf): mpc8313e-rdb: added machine for Freescale MPC8313E-RDB. Oct 18 16:58:38 03likewise 07org.oe.dev * r1d0e13e8... 10/ (1 conf/machine/include/tune-ppce300c3.inc): machine/tune-ppce300c3: Added PowerPC core, used in MPC83xx. Oct 18 16:58:43 03jluebbe 07org.oe.dev * r4751c499... 10/ (1 conf/distro/include/sane-srcrevs.inc): Oct 18 16:58:43 sane-srcrevs: update openmoko-mediaplayer2 to 3081 Oct 18 16:58:43 This fixes OpenMoko bug #947 Oct 18 16:58:49 03cbrake 07org.oe.dev * r1146b968... 10/ (1 conf/machine/em-x270.conf): em-x270.conf: fix include file names Oct 18 16:58:53 03hrw 07org.oe.dev * r40ef55d0... 10/ (1 contrib/source-checker/oe-source-checker.py): Oct 18 16:58:53 oe-source-checker: some improvements (usage, less exceptions) Oct 18 16:58:53 Calling with less then 2 parameters == usage info. Oct 18 16:58:53 Simplified code a bit - os.popen() does not generate exceptions if shell can Oct 18 16:58:54 not find command. Oct 18 16:58:58 03xora 07org.oe.dev * rada400c1... 10/ (4 files in 4 dirs): Oct 18 16:58:58 packages/gstreamer/gst-plugins-ugly_0.10.6.bb : add patch to make mp3->codec Oct 18 16:59:00 16bit all the way through. Increase openmoko-player2 svnrev to match Oct 18 16:59:02 the patch Oct 18 16:59:06 03koen 07org.oe.dev * raed96db0... 10/ (1 packages/gstreamer/gst-meta-base_0.10.bb): gst-meta-base: change some stuff to RRECOMMENDS, not everybody wants X Oct 18 16:59:10 03koen 07org.oe.dev * rf3f1577d... 10/ (1 packages/angstrom/angstrom-version.bb): angstrom-version: bump PE Oct 18 16:59:15 03hrw 07org.oe.dev * r40d52bf1... 10/ (1 contrib/source-checker/oe-source-checker.py): source-checker: added progressbar (from BitBake) and status info at end Oct 18 16:59:19 03hrw 07org.oe.dev * r972c9f9c... 10/ (1 contrib/source-checker/oe-source-checker.py): source-checker: move bad news to end of program (list of errors when program end) Oct 18 16:59:27 03hrw 07org.oe.dev * r6495b1bc... 10/ (1 conf/bitbake.conf conf/checksums.ini): checksums: dropped geda snapshots (unfetchable) Oct 18 16:59:33 03hrw 07org.oe.dev * racb1b3bb... 10/ (1 conf/checksums.ini): checksums: few more fixes from Koen's run Oct 18 16:59:40 03hrw 07org.oe.dev * r9ad4a408... 10/ (1 conf/distro/generic.conf): generic distro: prefer X.Org 7.2 instead of 7.1 (which is mostly not-existant) Oct 18 16:59:48 03koen 07org.oe.dev * r686e9c3c... 10/ (1 packages/ezx/ezxd/ezxd.init packages/ezx/ezxd_svn.bb): ezxd: remove runparts from init script Oct 18 17:00:12 good evening Oct 18 17:00:54 likewise, 8313, mmmmm. thanks for that. Oct 18 17:30:10 HopsNBarley: Any e300c3 core, I'll add e300c2 as well (no FPU). Oct 18 17:31:37 hi steliosk Oct 18 17:31:45 hmm, can I rely on PACKAGES = "${PN}-dbg ${PN} ${PN}-doc ${PN}-dev ${PN}-locale" in bitbake.conf? e.g. most every PACKAGES += "${PN}-dev" is a bug? Oct 18 17:32:01 Henryk: yes Oct 18 17:32:06 ok Oct 18 17:32:32 package.bbclass will throw a waring about PN-dev being in PACKAGES twice Oct 18 17:33:26 koen: i know. i 'found' some of these during my current build. I fixed one, or at least thought so. python-pygtk had += ...-dev in there twice (making for 3 -dev in PACKAGES) Oct 18 17:33:52 I'll grep around a bit and see whether i can find more of those Oct 18 17:34:15 the only valid reason to frob PACKAGES is when adding packages when splitting something up Oct 18 17:39:42 vivijim : Hi ! Oct 18 17:40:37 koen : is there a feed for x86 on angstrom repo ? doing an ipkg update gets 404 errors Oct 18 17:41:02 steliosk: I've configured a lot of machines with icecc here but it is really not working.... The packages are being compiled local only :( Oct 18 17:42:15 and I'm sure that I've configured everything like we did last week... and following the tutorial on wiki... Oct 18 17:42:20 do you have any idea? Oct 18 17:42:27 vivijim : Did you check the logs from the machine that you have assigned to do the scheduling ? Oct 18 17:42:50 actually every icecc logs are empty Oct 18 17:43:03 steliosk: i[456]86 are there Oct 18 17:43:11 steliosk: everbody here is using ubuntu feisty/gutsy Oct 18 17:43:23 steliosk: you can check with http://www.angstrom-distribution.org/repo/ Oct 18 17:43:31 (the dropdown box) Oct 18 17:44:00 * koen heads out Oct 18 17:44:03 later all Oct 18 17:45:19 koen: packages/flumotion/flumotion_0.4.1.bb cd15f8b1.. by koen 2007-06-17: #Yes, we know: ERROR: flumotion-dbg is listed in PACKAGES mutliple times, this leads to packaging errors. Oct 18 17:45:45 vivijim : weird. here are the logs from my scheduler -> http://rafb.net/p/cT0tS075.html Oct 18 17:47:14 koen : -> http://rafb.net/p/8HP4Zl99.html Oct 18 17:48:32 vivijim : here is my icecc init script- > http://rafb.net/p/fzTKBi12.html Oct 18 17:49:11 vivijim : notice about the logs with the parameters passed to the icecc daemon when it starts Oct 18 17:50:55 vivijim : once you get the logs, you can see if the machines actually log in the scheduling que or there is some protocol error Oct 18 17:52:14 koen : -> the feeds are there but empty.... Oct 18 17:53:36 steliosk: we've just noticed that there is an error on init.d/icecc that isn't using correct path to logs... Oct 18 18:07:54 hi Oct 18 18:10:53 vivijim : have to go. Ping me tomorrow if you still have troubles to get it going Oct 18 18:11:41 eee-pc Oct 18 18:12:02 steliosk: thanks Oct 18 18:13:32 03henryk 07org.oe.dev * re2424a77... 10/ (3 files in 2 dirs): python-pygtk: Remove more duplicate ${PN}-dev Oct 18 18:22:00 hu the n810 with mozilla and flash 9? Oct 18 18:22:11 flashplugin under arm? Oct 18 18:30:47 woglinde: the 770 already had flash (but not 9) Oct 18 18:31:05 SFF SDR is booting with root over nfs Oct 18 18:31:07 yay Oct 18 18:31:53 crofton yay Oct 18 18:32:04 keesj yes with opera Oct 18 18:32:10 but now with mozilla Oct 18 18:34:23 right Oct 18 18:57:50 does one know how to use the .map file generated by gcc Oct 18 19:02:18 no, what is that? class hierachy? Oct 18 19:22:43 03likewise 07org.oe.dev * r947287d0... 10/ (1 MAINTAINERS): MAINTAINERS: Updated my info. Oct 18 19:23:25 zecke: the .map file is just like the system.map file in the kernel it gives funtion/address translations (something like that) http://unixhelp.ed.ac.uk/CGI/man-cgi?ld (the -M option). given that information it should be possible that when given a address (from a backtrace) find the code /method in the source Oct 18 19:25:26 I want to use a program like addr2line Oct 18 19:27:10 anyone know what package contains ulimit? Oct 18 19:27:41 jsheldon hm Oct 18 19:27:43 shell Oct 18 19:27:45 nusynox Oct 18 19:27:47 busybox Oct 18 19:28:29 woglinde: really? a search in busybox's "command help" page doesn't list ulimit :( Oct 18 19:29:02 hm Oct 18 19:29:28 bash has ulimit Oct 18 19:29:54 actually, a more general question would be "has anyone dealt with grabbing coredumps on an arm system?" Oct 18 19:30:04 on most systems ulimit is a bash builtin, yes Oct 18 19:30:07 woglinde: oh sorry, right, I'll try the full bash Oct 18 19:37:32 hey guys...I created a bb file for i2ctools and can build them just fine with bitbake i2ctools. however, when I include them in my image.bb file, bitbake says it can't find it. Anyone have a clue of what I am doing wrong? Oct 18 19:43:47 it suggests running ipkg update. I haven't had to do that before...where would I do that? Oct 18 19:46:29 christopher: hi, is the .bb file in your BBFILES? Oct 18 19:46:39 christopher: what do you add to the image? Oct 18 19:47:48 I added i2c-tools to the RDEPENDS (where i've put other packages) Oct 18 19:47:54 the .bb file is in it's own directory Oct 18 19:48:12 bitbake i2c-tools builds it? Oct 18 19:48:27 yes. Oct 18 19:48:41 I've added other packages without an issue. Oct 18 19:48:45 so, this is a first Oct 18 19:48:53 -crebuild your-image? Oct 18 19:49:01 I tried changing the name thinking that could do something and it worked every time no problem. Oct 18 19:49:16 do you have i2c-tools packages in tmp/deploy/glibc/ipk/*? Oct 18 19:49:25 do they have a special PACKAGE_ARCH? Oct 18 19:49:27 if I -crebuild, will that clean out all of the prebuilt projects or just the rootfs Oct 18 19:49:32 I'll check if they are there Oct 18 19:49:39 i used autotools so I would think they were Oct 18 19:49:42 bitbake -b yourimage -crebuild Oct 18 19:49:53 it won't rebuild everything Oct 18 19:50:56 What's the recommended way of staging a native u-boot mkimage tool? Oct 18 19:51:11 zecke: ok. I see the ipk under armv4 Oct 18 19:52:08 christopher: now, go to tmp/deploy/rootfs/etc/ipkg/arch.conf (IIRC) and check if armv4 is in the list... Oct 18 19:52:19 | Cannot find package i2c-tools. Oct 18 19:52:19 | Check the spelling or perhaps run 'ipkg update' Oct 18 19:52:35 christopher: also check if if i2c-tools is in the Packages file of the ipkg directories Oct 18 19:54:16 zecke : in the packages list, there are -dev and -dbg versions...but no i2c-tools version Oct 18 19:54:51 christopher: a) you lied to me because you didn't find a package in the directory b) i2c-tools installed somewhere else and OE didn't create an empty package Oct 18 19:55:22 zecke: a) I wouldn't lie. :) I did find a package. Oct 18 19:55:59 zecke: but those packages, now that I look closer, are dev and dbg pacakges Oct 18 19:56:00 christopher: there was a output on the console like "Not creating empty package" Oct 18 19:57:07 zecke: yeah, I see that now Oct 18 19:57:11 :) Oct 18 19:57:16 so what does that mean? how do I get that created? Oct 18 19:57:23 that's the first time it hasn't built one Oct 18 19:57:35 ALLOW_EMPTY = "1" but that is not what you want! Oct 18 19:57:45 what do I want? Oct 18 19:58:01 christopher: OE also warns about installed but unpackaged files, do you have any? Oct 18 19:58:50 NOTE: Not creating empty archive for i2c-tools-3.0.0-r0 Oct 18 19:58:50 NOTE: Not creating empty archive for i2c-tools-doc-3.0.0-r0 Oct 18 19:58:50 Packaged contents of i2c-tools-dev into /oe/build/tmp/deploy/glibc/ipk/armv4t/i2c-tools-dev_3.0.0-r0_armv4t.ipk Oct 18 19:58:50 NOTE: Not creating empty archive for i2c-tools-locale-3.0.0-r0 Oct 18 19:58:54 that's all I see Oct 18 19:59:24 *g* Oct 18 19:59:29 no files Oct 18 19:59:32 no package Oct 18 19:59:32 christopher: what binary do you expect to see packaged? Oct 18 19:59:58 christopher: I would go to the ${WORKDIR} of i2c-tools and search for the binary Oct 18 20:00:01 there are some like i2cdetect etc Oct 18 20:00:12 christopher: also there is a directory where stuff got installed to Oct 18 20:00:39 ok, I see that... Oct 18 20:00:42 it is under 'tools' Oct 18 20:00:55 christopher: a) find out if it was build b) check if it was installed c1) fix the installation dir c2) adjust the FILES_${PN} += to add the path/binary in the .bb files Oct 18 20:01:04 christopher: it is not installed? Oct 18 20:01:28 zecke: not installed where? I see that they are built. Oct 18 20:01:52 basically, there are i2cget and lots of utilities in the tools directory under i2ctools/i2ctools Oct 18 20:02:52 so basically do a install -d from the tools directory? Oct 18 20:02:54 in the bb file? Oct 18 20:03:36 christopher: OE is executing make install (and setting a destination dir) so they should be installed into a special directory Oct 18 20:04:07 christopher: if you want to have a fast solution you could do do_install_prepend in your bb and install the files manually do ${D}${bindir} (there are a lot of examples...) Oct 18 20:04:35 woglinde: I just tried the "Web Collage" screensaver and it took up to 3 minutes to see the first porn... Oct 18 20:04:36 zecke: ok. Oct 18 20:04:53 zecke what? Oct 18 20:12:21 zecke: what is a good example? i would have thought it was a post install and not a pre-install? Oct 18 20:12:47 install -m 0755 ${WORKDIR}/tools/i2cdetect ${D}${bindir}/ Oct 18 20:12:49 that is what I tried Oct 18 20:18:24 christopher: pre-install, post-install are for scripts to be executed during install ;) Oct 18 20:18:41 christopher: search for do_compile_prepend but your line above looks sane Oct 18 20:22:28 ok, so use do_compile_prepend and not do_install_prepend? Oct 18 20:23:09 sorry.... Oct 18 20:23:28 just do_install_prepend ( I was compiling when writing this...) Oct 18 20:32:49 zecke: ok. and do you generally do each file or can you do directories and it knows which files are executable? Oct 18 20:36:21 for file in files; do install done ;) Oct 18 20:36:26 christopher: or fixing the Makefile etc... Oct 18 20:36:54 * hrw for a moment Oct 18 20:37:16 i2ctools? how does it differ from lmsensors-apps? Oct 18 20:37:33 lmsensors-apps also gives i2c detect/set/get tools Oct 18 20:39:02 hrw: lmsensors-apps didn't build for me. so, I was going with i2ctools since that is all I needed Oct 18 20:39:22 christopher: how it does not build? which way? Oct 18 20:39:29 christopher: glibc or uclibc build? Oct 18 20:39:41 hrw: this is a glibc build. Oct 18 20:39:46 hrw: first, it fails on a patch. Oct 18 20:39:56 # file://prefix-fix.patch;patch=1 \ Oct 18 20:40:00 so, I took that one out Oct 18 20:40:01 zecke: did you saw base.bbclass generator of checksums.ini and contrib/source-checker script? Oct 18 20:40:13 christopher: how old metadata you use? Oct 18 20:40:16 and now it fails with | make: *** No rule to make target `install'. Stop. Oct 18 20:40:27 I fixed lmsensors during oedem Oct 18 20:40:30 hrw: I did a pull and update a couple of weeks ago I suppose Oct 18 20:40:35 hrw: your mail is marked as to be read... Oct 18 20:40:47 christopher: so you use obsolete tree Oct 18 20:41:01 i use org.openembedded.dev if that is what you mean? Oct 18 20:41:05 christopher: lmsensors-apps in OE.dev builds Oct 18 20:41:15 christopher: few weeks old == obsolete Oct 18 20:41:24 ok. is there a way to update just that bb file? Oct 18 20:41:41 it isn't a revert, but I am sure mtn has a command for that Oct 18 20:42:07 task-base takes a long time to build :) Oct 18 20:42:07 christopher: use two dirs: one with current .dev and second with your-old one Oct 18 20:42:31 hrw: hmmm. yeah, I can do that, it is just that it takes a day to build it all Oct 18 20:42:46 christopher: I do not mean build tree but metadata tree Oct 18 20:43:09 hrw: ok. so, just make a copy of the org.oe.dev directory Oct 18 20:43:10 ? Oct 18 20:43:17 and then do an update? Oct 18 20:43:21 christopher: yes Oct 18 20:43:39 christopher: and then copy lmsensors from current dev to your old copy Oct 18 20:44:10 hi hrw Oct 18 20:44:21 hrw: ok. so, to get that, all I should have to do is 'mtn pull && mtn update' Oct 18 20:44:23 hi woglinde Oct 18 20:44:26 hrw: and I should have the updated one Oct 18 20:44:29 christopher: right Oct 18 20:44:49 hrw: do you remember what you had to change? Oct 18 20:45:14 christopher: I basically rewrote whole recipe, regenerated missing patch etc Oct 18 20:46:03 * hrw goes back to wife - bye Oct 18 20:47:54 nite hrw Oct 18 20:58:04 * * OE Bug 3174 has been created by inv2002(AT)yandex.ru Oct 18 20:58:06 * * sound problem over CE-RH2. Oct 18 20:58:08 * * http://bugs.openembedded.org/show_bug.cgi?id=3174 Oct 18 20:59:26 * cbrake is pretty impressed how much Linux you can fit in a 4MB flash :-) Oct 18 21:00:51 hehe Oct 18 21:00:56 cbrake: like a whole kernel? Oct 18 21:02:27 zecke: kernel + alsa + tslib + libpcre + mtd-utils + dropbear + ... Oct 18 21:02:50 zecke: can do quite a bit if you are willing to write apps in C :-\ Oct 18 21:04:16 cbrake much more when you use klibc ort diet-libc Oct 18 21:05:01 woglinde: yeah, I'm still using uclibc Oct 18 21:05:41 Yay. 2.6.23 boots on powerpc mpc8313. Looks like a toolchain and kernel in <5 hours of work. Hugs to OE! Oct 18 21:05:58 woglinde: the nice thing it "just works" in OE without any extra effort :-) Oct 18 21:06:07 likewise: nice Oct 18 21:06:36 cbrake me too but with diet-libc things are much smaller Oct 18 21:06:47 * likewise brings on a ramdisk beer, uh, build. Oct 18 21:06:52 does anyone know what package generates /etc/version? ipkg search gives no provider Oct 18 21:06:54 likewise gratz Oct 18 21:07:52 < 5 hours of work? or compiling :) Oct 18 21:08:18 compiling Oct 18 21:10:18 sakoman: I would assume base-files Oct 18 21:25:18 steliosk: I'm asked to drink your ouzo now :} Oct 18 21:27:04 steliosk: ah, right, the autobuilder doesn't build for i486 machines Oct 18 21:28:37 zecke: I don't see it in the base-files recipe, have to keep searching Oct 18 21:28:47 sakoman: search for base* Oct 18 21:28:57 sakoman: your work history is quite interesting Oct 18 21:29:43 koen:I've been lucky to be involved in lost of interesting stuff Oct 18 21:30:22 lots, though I suppose lost might be a freudian slip :-) Oct 18 21:31:31 :) Oct 18 21:32:00 nite Oct 18 21:32:00 bleh, my hands smell like fixer Oct 18 21:32:20 koen: do you know what package generates /etc/version? Oct 18 21:32:44 koen: I thought you used the odorless stuff :-) Oct 18 21:32:56 the darkroom tutorial used a different one Oct 18 21:33:10 I would guess that base-files generates /etc/version Oct 18 21:34:07 oooh, n810, zaurus replacement finally Oct 18 21:34:22 XorA: register for the dev discount Oct 18 21:34:33 koen: where is that? Oct 18 21:34:36 xora mozilla based browser with flash Oct 18 21:34:38 maemo.org Oct 18 21:34:51 XorA: and in the footnote of my latest mail to oe-dev :) Oct 18 21:34:53 xora where comes the flash9 for arm from? Oct 18 21:35:10 glash, swfdec, adobe? Oct 18 21:36:28 koen from adobe might be intressting Oct 18 21:36:36 didnt know the other can flah9 now Oct 18 21:36:44 the previous ones were from adobe as well Oct 18 21:37:07 hm I thought the were from opera Oct 18 21:39:01 I do not think /etc/version comes from base-files Oct 18 21:39:06 from looking at the bb file Oct 18 21:40:14 angstrom-version comes from angstrom-version.bb .... Oct 18 21:40:39 rootfs_deb.bbclass: echo ${BUILDNAME} > ${IMAGE_ROOTFS}/${sysconfdir}/version Oct 18 21:40:40 rootfs_ipk.bbclass: echo ${BUILDNAME} > ${IMAGE_ROOTFS}/${sysconfdir}/version Oct 18 21:40:48 * koen thinks that should go Oct 18 21:41:09 btw, I am very pleased with SFF SDR progress atm Oct 18 21:41:23 Crofton: congrats Oct 18 21:41:29 thanks Oct 18 21:41:34 Crofton: did you hand the angstrom toolchain to your student yet? Oct 18 21:41:41 koen: bah, so the login page doesnt actually work for applying for discount Oct 18 21:41:43 no Oct 18 21:41:51 we do not have one yet for this purpose Oct 18 21:42:01 but I will likely try the tool chain soon Oct 18 21:42:02 Crofton: I agree -- it isn't from base-files Oct 18 21:42:22 sakoman: scroll up 2 minutes :) Oct 18 21:42:25 sakoman, see koen's comment, it is in rootfs*.bbclass Oct 18 21:42:27 :-) Oct 18 21:42:47 lesson: don't respond till you've read everything :-) Oct 18 21:43:03 I need to run a quick test on the tool chain and prod my advisor to fins a guy in his class to try some wmmx stuff Oct 18 21:43:25 you can always hook the guy up with devshell Oct 18 21:43:30 yeah Oct 18 21:43:39 I just need something really easy Oct 18 21:44:08 I may have to stick the tool chain in a vmware image .... Oct 18 21:44:36 hm do I want a nokia 810 or not Oct 18 21:45:18 sakoman, how good is connex support in .dev? Oct 18 21:45:45 could we point people at the Angstrom site and have them try images and packages from there? Oct 18 21:45:56 for a sane set of hw? Oct 18 21:46:00 * XorA gives up on the busted nokia pages Oct 18 21:46:01 Crofton: I haven't tried oe.dev in quite some time Oct 18 21:46:15 xora? Oct 18 21:46:19 been using my custom hw.conf Oct 18 21:46:23 * Crofton needs to read the web page on the netwifimicrosd Oct 18 21:46:25 yeah Oct 18 21:46:35 woglinde: I enter login, no error, it just pushes me back to login page again Oct 18 21:46:39 I need to compare those :) Oct 18 21:46:54 XorA: there's a waiting time Oct 18 21:47:05 XorA: the backend needs a few minutes to update Oct 18 21:47:26 koen: bah, dont these guys know how databases work :-) Oct 18 21:47:38 "> I've registred to maemo.org creating an account. I got the Oct 18 21:47:39 > > confirmation email from maemo.org and I was able to log into the Oct 18 21:47:39 > > "garage". But if I go to www.maemo.org and I try to enter my user and Oct 18 21:47:40 > > password I cannot login :( Oct 18 21:47:40 There is some delay before accounts are replicated from Gforge to Oct 18 21:47:40 Midgard. Try again a bit later." Oct 18 21:47:55 Crofton: scruggs and I are both using it with connex. Oct 18 21:48:36 I can't remember if scruggs is using anything from your collection Oct 18 21:48:57 he told me he's using all of it Oct 18 21:49:02 ok Oct 18 21:49:16 need to see what makes sense to move into .dev Oct 18 21:49:30 Between the two of us I think we've tested everything but bluetooth Oct 18 21:49:31 anyway, I'll worry about this another day :) Oct 18 21:49:37 that is very cool Oct 18 21:49:49 I have had actual cool work to do :) Oct 18 21:50:02 You deserve to have some fun! Oct 18 21:50:15 heh Oct 18 21:50:35 The school thing will be over soon . . . Oct 18 21:50:37 I need this to work so I can convince some people to pay me to make a complete package for the board :) Oct 18 21:50:50 anyway, time to go Oct 18 21:50:50 Which board? Oct 18 21:50:55 I am using sakoman's collection and its working great Oct 18 21:51:03 Lyrtech small form factor SDR board Oct 18 21:51:13 ~sffsdr Oct 18 21:51:17 I'll have to check it out! Oct 18 21:51:21 pricey Oct 18 21:51:35 I need to convince them not to take it back Oct 18 21:52:16 l8r guys! Oct 18 21:52:24 OK, later! Oct 18 21:52:37 take it easy, Crofton Oct 18 21:53:19 sakoman: thanks for updating me with your collection changes... its sped my other devving up nicely! Oct 18 21:53:52 scruggs: I really like what you've done. That's a very neat project! Oct 18 21:54:34 how is performance on a connex? Oct 18 21:54:47 sakoman: thanks! its been a fun one Oct 18 21:56:15 the biggest performance hit is drawing a new map - I generate a map around 3X the size of the screen and then just re-center it on a new fix, but if the map to be generated has lots of contour lines generation time can be around 3sec on a connex Oct 18 21:56:49 scruggs_: floating point calculations? Oct 18 21:56:49 if I have it just render streets/water features the generate time is a fraction of a sec Oct 18 21:58:04 koen: no, more just the amount of vector points to draw all the contour lines for the map Oct 18 21:58:24 scruggs: I have a somewhat older Garmin GPS that has about the same performance when contours are enabled Oct 18 21:59:31 yeah.. its really not too bad except for the initial load... the program knows its about to run out of map and generates the next one before the old one is 'out of range' Oct 18 22:00:13 how large are the data files? kept on SD/CF card? Oct 18 22:03:48 depends... there is a helper program to convert the shapefile data to the minimal amount needed for the program, so for my county in CO the shapefile data is 20MB, but once its converted the data used by the program is less then 1.5MB (all stored on cf) Oct 18 22:04:17 but the data is all vector data, not rendered maps, so it is infinately scalable Oct 18 22:04:28 to a degree :) Oct 18 22:04:56 Source files are gov't downloads? Oct 18 22:05:25 yep Oct 18 22:05:46 esri shapefiles? Oct 18 22:06:21 koen: yes - and shapefiles generated from DEM data Oct 18 22:06:30 nice Oct 18 22:06:44 Cool! I've been rummaging on the net for postal code - latitude/longitude, timezone, dst data sets Oct 18 22:06:44 * koen did a minor in geo-informatics a few years ago Oct 18 22:07:04 its more of an overland nav program than a street-map one Oct 18 22:07:21 So I can automaticly set time & calulate sunrise/sunset times Oct 18 22:07:31 I've been really active in Search & Rescue for years, so it works for my other hobbies :) Oct 18 22:08:00 sakoman: just include a gps in your design :) Oct 18 22:08:14 sakoman: very cool Oct 18 22:08:23 I'm already fighting lots of product cost issues :-) Oct 18 22:08:34 you could use a gsm as well, but I think they send you to gitmo for correlating cell towers with gps data Oct 18 22:08:42 I've got US/Canada/Mexico working Oct 18 22:09:10 sakoman: if you ship to new zealand you need to update tzdata in OE :) Oct 18 22:09:28 2007e lacks their recent DST updates Oct 18 22:09:32 * koen stabs DST Oct 18 22:09:40 I hate dst too! Oct 18 22:10:04 In the US lots of areas randomly decide to adopt it or not Oct 18 22:10:14 And it changes year to year Oct 18 22:10:25 at the whim of politicians Oct 18 22:10:37 yeah.. wtf is up with dst now? Oct 18 22:10:49 dst is for hippies Oct 18 22:11:27 I think DST was around decades before hippies ;-) Oct 18 22:11:46 see you guys, taking a nap Oct 18 22:58:51 re Oct 19 01:51:05 * * OE Bug 3175 has been created by angus(AT)handheldshell.com Oct 19 01:51:07 * * Toshiba eseries kernel recipe Oct 19 01:51:09 * * http://bugs.openembedded.org/show_bug.cgi?id=3175 Oct 19 02:16:52 argggh I am getting burned by the git fetcher and 64 bit machines! Oct 19 02:20:47 Crofton: how so? Oct 19 02:23:29 git fetches are failing on my machines w64 bit installs Oct 19 02:23:46 but git installed works fine in machine Oct 19 02:23:57 I think something is fishy in git-native Oct 19 02:28:18 ah, sure. i was just about to disagree (i just had a full build of openmoko-devel-image and openmoko-feed on x86_64) but then remembered that any git failures probably have been masked by PREMIRRORS in openmoko Oct 19 02:28:48 I was wondering why it hadn't bit openmoko Oct 19 02:28:56 need to beat up RP in the morning I think Oct 19 02:29:19 or see if I can go the assume provided route until someone can figure it out Oct 19 02:29:31 running a build on the 32 bit install at the moment Oct 19 02:30:33 ok, time for bed, hopefully this build runs through :) Oct 19 02:31:25 gn Oct 19 02:32:05 good night **** ENDING LOGGING AT Fri Oct 19 02:59:58 2007