**** BEGIN LOGGING AT Fri Feb 29 10:59:56 2008 Feb 29 11:02:50 hi mickeyl Feb 29 11:02:59 hi hrw Feb 29 11:05:27 morning mickeyl Feb 29 11:05:45 hi RP Feb 29 11:06:39 hey RP Feb 29 11:06:46 yeah, my lost+found size depens on my image-size...ist alway approx. 5% Feb 29 11:12:27 hmm, who do I talk to about gpsd? Feb 29 11:12:56 Its including c++ bindings which mean everything using it pulls in libstdc++ and needs to link using g++ Feb 29 11:13:37 ugly Feb 29 11:13:44 so we should increase package granularity? Feb 29 11:13:49 gpsd is dead anyways Feb 29 11:13:54 but yes, lets improve it Feb 29 11:14:14 mickeyl: Its not just package granularity, the c++ bindings are in the base lib Feb 29 11:14:42 ooh Feb 29 11:14:44 i see Feb 29 11:14:46 nasty nasty Feb 29 11:14:56 mickeyl: Its just broken my sysroot enabled build since when linking with gcc, it no longer can find libstdc++ Feb 29 11:15:03 heh Feb 29 11:15:14 which is correct and hightlights nastiness like this :/ Feb 29 11:15:14 ~praise sysroot Feb 29 11:15:15 All hail sysroot! Feb 29 11:15:32 if you have time to make a patch, feel free to rip the C++ binding out Feb 29 11:15:36 i don't think anyone is using this atm. Feb 29 11:15:46 otherwise i could take a look next week Feb 29 11:15:47 I'll try a quick patch... Feb 29 11:16:09 * mickeyl is much less anal about C++ after having seen Fefe's presentation "A case against C++" Feb 29 11:16:22 *g* Feb 29 11:16:29 mickeyl: url? Feb 29 11:16:29 it's horror Feb 29 11:16:32 but it's a fact Feb 29 11:16:37 wait, i'll dig it out Feb 29 11:17:04 http://www.fefe.de/c++/c%2b%2b-talk.pdf Feb 29 11:17:06 this is the pdf Feb 29 11:17:11 but you need to see the video of the presentation Feb 29 11:17:15 otherwise you're missing good stuff Feb 29 11:17:19 hmm... where was that... Feb 29 11:17:50 ftp://dewy.fem.tu-ilmenau.de/CCC/CCCamp07/video/m4v/cccamp07-en-1951-A_Case_Against_C++.m4v Feb 29 11:17:53 here we are Feb 29 11:18:36 thx Feb 29 11:18:43 np Feb 29 11:18:43 mickeyl: lol Feb 29 11:18:57 i thought it was FUD Feb 29 11:19:01 but it doesn't look like it... Feb 29 11:19:35 mickeyl: sure, templates allow recursion and finding the fixpoint for the semantic of a recursive function is tough :) Feb 29 11:20:51 108MB... Feb 29 11:21:08 ~curse perl in OE Feb 29 11:21:09 May you be reincarnated as a Windows XP administrator, perl in OE ! Feb 29 11:21:15 ~curse perl in general too Feb 29 11:21:16 May the fleas of a thousand camels infest your most sensitive regions, perl in general too ! Feb 29 11:25:07 ~curse imagemagick Feb 29 11:25:08 May you be reincarnated as a Windows XP administrator, imagemagick ! Feb 29 11:25:20 ~curse cursing Feb 29 11:25:21 May the fleas of a thousand camels infest your most sensitive regions, cursing ! Feb 29 11:41:42 morning! Feb 29 11:43:03 RP: have you looked at the patches from OE bug 3879? Feb 29 11:50:01 lumag: sorry, I haven't :( Feb 29 11:50:30 mickeyl, you can shoot yourself in the foot in many languages :) Feb 29 11:52:40 Crofton: I think the main sounds of shooting yourself in the foot remain the same in any language Feb 29 11:53:01 yeah Feb 29 11:53:32 I think the fact you can shoot yourself in the foot keeps the guys that write coding style books in business :) Feb 29 11:54:14 * XorA wonders if there is regional variations on screams of pain Feb 29 11:54:25 I was feeling so optimistic this morning, now everything is exploding in this build :( Feb 29 11:55:18 RP: that just makes it time for pub Feb 29 11:55:19 ~curse perl in Oe Feb 29 11:55:19 May you be reincarnated as a Windows XP administrator, perl in Oe ! Feb 29 11:55:24 hrw: having fun? Feb 29 11:55:28 o yes Feb 29 11:55:53 with private recipe Feb 29 11:55:55 * RP glares at tcl/tk Feb 29 11:56:03 ~hail quilt Feb 29 11:56:04 * ibot bows down to quilt and chants, "I'M NOT WORTHY!!" Feb 29 11:56:37 Does packages from OE works fine in Chinook? Feb 29 11:56:45 the problem with all perl stuff is that we need to run native perl to config and build for target Feb 29 11:56:57 ynezz: with DISTRO="chinook-compat" they do Feb 29 11:57:00 mostly Feb 29 11:57:31 how do you build custom kernels in OE? Feb 29 11:57:31 what do you mean by mostly? :) Feb 29 11:57:38 re Feb 29 11:57:48 ynezz: those which I tried work - except gvim Feb 29 11:58:28 ok fine Feb 29 11:58:36 althought gvim would be handy also Feb 29 11:59:43 RP: can we have a 'stacked' FILESPATH so uclibc/mymachine/uclinuxeabi/defconfig and uclibc/mymachine/uclinux-uclibc/defconfig both work? Koen hinted to ask you about it :> Feb 29 12:00:32 03xora 07org.oe.dev * r20657a67... 10/ (4 files in 2 dirs): alsa-state/fic-gta02/* : merge state files with openmoko versions Feb 29 12:00:59 Jin^eLD: I think that might already work. Have a look at the FILESPATH stuff in base.bbclass to see what combinations it supports Feb 29 12:01:19 thanks Feb 29 12:01:41 * RP has just found that an extremely old overlay with default prority is causing half his problems Feb 29 12:01:52 ;) Feb 29 12:04:44 epdfview sort of worked on n810 Feb 29 12:06:26 * XorA needs to see about angstrom or poky on n810 Feb 29 12:06:41 * tom__ just wonders how to prevent that a huge amount of kernel modules is included in my image Feb 29 12:07:13 tom__: dont RRECOMMEND kernel-modules Feb 29 12:07:23 i did that Feb 29 12:07:33 i think Feb 29 12:07:55 #MACHINE_EXTRA_RRECOMMENDS = " kernel-modules" Feb 29 12:08:03 in my x86.conf Feb 29 12:08:08 I wonder how hard it would be to make complete Angstrom image for N810 also Feb 29 12:08:24 tom__: rebuild task-base now Feb 29 12:08:33 ah ok Feb 29 12:08:51 ynezz: Poky do such ones - n800/n810 image - lack keyboard map probably Feb 29 12:08:52 ...some days later... ;) Feb 29 12:08:56 ynezz: we need to pull in the OH uber work for n8x0 machines I suspect Feb 29 12:09:27 XorA: kernel is in OE already Feb 29 12:09:44 hrw: btw did you guys experiment with open source network driver? Feb 29 12:10:07 hm from debug output i see its not better Feb 29 12:10:14 Well I want to use Hildon, I like it, so Poky isn't my choice I think Feb 29 12:10:16 do i have to do some -crebuild? Feb 29 12:10:23 XorA: cx3110 opensource driver? Feb 29 12:10:47 tom: bitbake task-base -crebuild Feb 29 12:11:09 k thx Feb 29 12:11:42 hrw: dont know its title, but I know there is one Feb 29 12:12:39 o fsck.. Feb 29 12:13:09 due to 29th Feb my favorite radio station play music which they do not play normally Feb 29 12:13:20 now it is Sabrina - Boys Feb 29 12:15:09 hrw: the pain the pain Feb 29 12:15:22 XorA: nope - pure fun Feb 29 12:16:17 * tom__ thanks for the tip..works now :) Feb 29 12:16:59 one more short question...: what's best-practice for building a custom kernel? Feb 29 12:17:18 because the bzImage is quite big with 2.6mb Feb 29 12:17:45 i'd like to have it at 1mb Feb 29 12:18:52 tom__: Define your own target device and add a custom kernel config to the kernel of your choice. Feb 29 12:19:19 so, TARGET? Feb 29 12:19:56 tom__: if i read this correct you have some x86 based machine, right? Feb 29 12:20:04 yes Feb 29 12:20:13 i'll build for x86 Feb 29 12:20:23 i want to use my builds in a VM Feb 29 12:20:31 it works very nice in qemu yet Feb 29 12:20:59 when i use qemux86 as MACHINE, the kernel can't find the rootfs, but with x86 it works Feb 29 12:21:43 i have a working kernel with 1mb, but in future i'd like to build everything from within OE Feb 29 12:23:04 tom__: then copy the x86.conf to .conf and copy your kernel config file to packages/linux/linux-x86-//defconfig Feb 29 12:23:06 tom__: add the defconfig of you 1mb kernel to OE Feb 29 12:23:25 XorA: now it is pain... O-Zone Feb 29 12:23:31 thanks a lot! Feb 29 12:24:02 XorA: what would you like to know about cx3110x network driver? Feb 29 12:24:57 hmm whats PREFERRED_PROVIDER_virtual/kernel in x86.conf? Feb 29 12:25:11 should i let it point at "linux"? Feb 29 12:25:12 ssvb: does it work better/the same as the closed source one? Feb 29 12:26:16 XorA: well, does a completely open source driver for cx3110x exist? Feb 29 12:26:42 Happy leap day everyone Feb 29 12:26:48 ssvb: yes, https://garage.maemo.org/projects/cx3110x/ Feb 29 12:27:01 but IIRC it relies on a closed source module or firmware or something Feb 29 12:27:22 how can i pull the OE.db from a date ? Feb 29 12:28:29 trickie: well, the emphasis was on the word *completely* :) I know about this garage project and I'm even trying to fix/optimize the driver for nokia 770 Feb 29 12:29:39 ssvb: ok :) well i guess wireless operation does not *competely* open software Feb 29 12:29:45 but i may be wrong Feb 29 12:30:25 * XorA knows nothing apart from rumours about needed nokias .ko file which would mean Angstrom could never work right Feb 29 12:30:42 the problem is that connexant does not plan to release anything so nokia tablets wifi will stay closed Feb 29 12:31:46 hrw: what wlan driver did XorA have in mind then? Feb 29 12:32:20 no idea Feb 29 12:32:26 * XorA just saw mention in internettablettalk.com of an open source driver Feb 29 12:32:36 user forums... Feb 29 12:33:13 rob_w|laptop: see monotone.ca Feb 29 12:33:52 XorA: that was probably a discussion about my patch for 770 Feb 29 12:33:57 so status is even with garage driver nokia .ko file still needed? Feb 29 12:35:05 XorA: yes, cx3110x.ko driver is open source, but it depends on umac.ko which is closed source, also wlan chip is operated by closed source firmware which gets uploaded to it on initialization Feb 29 12:35:23 ssvb: the firmware is normal situation for any wlan Feb 29 12:36:13 rob_w|laptop: http://www.monotone.ca/docs/Selectors.html#Selectors Feb 29 12:36:52 thanks likewise .. its time to read the pdf Feb 29 12:37:28 ssvb: would it be possible to have driver which can do injecting? Feb 29 12:37:51 or it's that umac.ko what's the problem? Feb 29 12:38:35 ynezz: I have no idea, cx3110x.ko is just a glue which provides communication between the firmware in wlan chip and umac.ko Feb 29 12:39:57 why won't my card reader create a valid ext2 file system on my micro-sd card ???? Feb 29 12:40:26 ynezz: I'm only trying to optimize DMA transfers done over SPI bus which is this driver responsible for, the goal is to get faster transfers and lower cpu load Feb 29 12:41:34 Jop, I've seen the thread with that Nokia WiFi guy, good luck then :) Feb 29 12:43:23 RP: whom should I ping then? you? or koen? Feb 29 12:44:31 lumag: Probably me Feb 29 12:45:07 RP :) Feb 29 12:45:27 RP: the FILESPATH thing.. in order for it to support the stuff I was asking for it should have a ${TARGET_OS} in there, right? Feb 29 12:45:44 Jin^eLD: Probably, yes Feb 29 12:45:50 then it's not there yet Feb 29 12:46:02 have to check what ${PF} is, but probably not what I am looking for Feb 29 12:46:13 its not Feb 29 12:54:40 NPR's business news starts with a story on the dollar's descent against the euro Feb 29 12:54:44 * Crofton cries Feb 29 13:02:21 * tom__ asks: anybody know what to do about: ERROR: Information not available for target 'vhe_x86-linux' Feb 29 13:02:48 Crofton: dollar? it is that currency which is more worth as paper/metal then value on it? Feb 29 13:02:58 :)) Feb 29 13:04:43 not yet :) Feb 29 13:04:52 maybe the penny, not sure on the copper percent though Feb 29 13:05:13 hrw, so you say i could sell the coins at a higher price ,as of its metal ? huhuhu Feb 29 13:05:38 you can sell the metal at a higher price than the coin ;) Feb 29 13:05:39 well, hope it doesn't start WWIII Feb 29 13:05:45 but it's forbidden... Feb 29 13:06:03 rob_w|laptop: in Poland ~30 years ago we got 10 & 20 pln notes as metal coins were more expensive then value on them Feb 29 13:15:24 hi, all! do anybody remembers, if I enable mouse in kdrive at compile time, is it possible to prevent it from using it and make it use tslib? tslib is compiled-in, but Xfbdev doesn't try to open touchscreen device and garbles console :( Feb 29 13:15:37 hi all! Feb 29 13:16:23 slapin: 1.3 or 1.4? Feb 29 13:16:38 lumag, 0.6.3 :( Feb 29 13:19:05 * * OE Bug 3895 has been created by  Feb 29 13:19:07 * * glibc-2.6.1-r2-do_package Feb 29 13:19:09 * * http://bugs.openembedded.net/show_bug.cgi?id=3895 Feb 29 13:19:53 gm gremlin[it] Feb 29 13:20:20 the penny could be worth more in metal, but that is largely driven by the rise in Copper prices Feb 29 13:22:56 right Feb 29 13:23:54 Crofton, are you back ? ... i heard you had sick after the fly ... Feb 29 13:24:00 yeah Feb 29 13:24:04 I had a cold Feb 29 13:24:10 started as I was leaving Feb 29 13:24:11 hmm my custom kernel builds now and gets deployed, but bitbake breaks because it wants to copy the modules (which aren't there) Feb 29 13:24:39 basically, buying a block of Copper a few years back would have been as good or better investment than a suitcase full of euros Feb 29 13:25:49 it breaks at tar -cvzf /stuff/build/tmp/deploy/images/modules-2.6.21-r12-vhe.tgz -C /stuff/build/tmp/work/vhe-linux/linux-2.6.21-r12/image lib Feb 29 13:26:14 image seems to be empty Feb 29 13:30:26 no, image/lib is nonexistent Feb 29 13:37:02 tom: your kernel lack modules Feb 29 14:03:47 hrw: yeah, because i don't want modules :) Feb 29 14:04:08 i commented out MACHINE_EXTRA_RRECOMMENDS Feb 29 14:04:28 but it still wants to build a module image Feb 29 14:12:25 03john_lee 07org.oe.dev * ra8f89599... 10/ (1 packages/asterisk/asterisk_1.4.17.bb): asterisk: add unset RANLIB to make the build success. Feb 29 14:12:33 03john_lee 07org.oe.dev * r61e5777f... 10/ (3 files in 3 dirs): busybox: 1.9.1 add cmp into openmoko distribution Feb 29 14:12:42 03john_lee 07org.oe.dev * re8061e0e... 10/ (4 files in 2 dirs): Feb 29 14:12:42 gnash: create gnash.inc Feb 29 14:12:42 * add gnash 0.8.1 Feb 29 14:12:52 03john_lee 07org.oe.dev * r65ad0338... 10/ (7 files in 3 dirs): Feb 29 14:12:52 gpsdrive: add 2.10pre4 Feb 29 14:12:52 * http://bugs.openembedded.org/show_bug.cgi?id=3670 Feb 29 14:12:52 * plus some polishment Feb 29 14:13:00 03john_lee 07org.oe.dev * r64b051ac... 10/ (4 files in 3 dirs): Feb 29 14:13:00 gsm: update gsmd initscript for gta01 and gta02 Feb 29 14:13:00 * use the same gsmd initscript for gta01 and gta02 Feb 29 14:13:00 * this only works with kernel 2.6.24 Feb 29 14:13:08 03john_lee 07org.oe.dev * r88eb006e... 10/ (5 files in 4 dirs): multitap-pad: multi-tap input method for GTK Feb 29 14:13:14 03john_lee 07org.oe.dev * r0ce45085... 10/ (6 files in 5 dirs): Feb 29 14:13:14 openmoko-dialer2: add machine specific patch for backlight and vibrator Feb 29 14:13:14 * this should be done in external config but for now we use patch Feb 29 14:13:14 * gta01 and gta02 use different names for backlight and vibrator under /sys Feb 29 14:13:14 * this only works with kernel 2.6.24 Feb 29 14:13:23 03john_lee 07org.oe.dev * r8bf6f11e... 10/ (2 files in 2 dirs): openmoko-panel-memory: shows an out-of-memory warning in the OpenMoko panel Feb 29 14:13:30 03john_lee 07org.oe.dev * r6f0b6659... 10/ (3 files in 3 dirs): openmoko-session2: add openmoko-panel-memory Feb 29 14:13:36 03john_lee 07org.oe.dev * r3b510b98... 10/ (1 packages/tasks/task-openmoko-phone.bb): task-openmoko-phone: include openmoko-panel-memory Feb 29 14:13:43 03john_lee 07org.oe.dev * r4512622b... 10/ (3 files in 2 dirs): perl: add libwww-perl and native Feb 29 14:13:48 03john_lee 07org.oe.dev * rbd952e49... 10/ (1 packages/tasks/task-openmoko-base.bb): task-openmoko-base: replace matchbox-keyboard-inputmethod with multitap-pad Feb 29 14:13:54 03john_lee 07org.oe.dev * r63ed025a... 10/ (1 packages/u-boot/u-boot-openmoko_1.3.1+git+svn.bb): Feb 29 14:13:54 u-boot-openmoko: remove do_configure_prepend because it's no longer Feb 29 14:13:54 necessary Feb 29 14:14:01 03john_lee 07org.oe.dev * rf0bf816c... 10/ (1 conf/checksums.ini): Feb 29 14:14:01 checksums.ini: add checksums for libwww-perl-5.508.tar.gz, Feb 29 14:14:01 gpsdrive-2.10pre4 and update the location of Feb 29 14:14:01 gpsdrive-2.10pre2 Feb 29 14:14:08 03john_lee 07org.oe.dev * r1aa47190... 10/ (1 conf/distro/include/moko-autorev.inc): Feb 29 14:14:08 moko-autorev: add opkg, opkg-native, multitap-pad and Feb 29 14:14:08 openmoko-panel-memory Feb 29 14:14:15 03john_lee 07org.oe.dev * rae20897a... 10/ (1 conf/distro/include/preferred-om-2008-versions.inc): Feb 29 14:14:15 preferred-om-2008-versions: add preferred version of asterisk, gnash Feb 29 14:14:15 and gpsdrive Feb 29 14:14:22 03john_lee 07org.oe.dev * rd12cd249... 10/ (1 conf/distro/include/sane-srcrevs.inc): sane-srcrevs: add multitap-pad and openmoko-panel-memory Feb 29 14:14:30 03john_lee 07org.oe.dev * r05e26e1a... 10/ (1 conf/distro/openmoko.conf packages/tasks/task-base.bb): Feb 29 14:14:30 openmoko: add DISTRO_EXTRA_RDEPENDS += "opkg-ipkg-compat" Feb 29 14:14:30 * increase task-base PR to reflect the change into task-distro-base Feb 29 14:14:37 03john_lee 07org.oe.dev * r3d4d513f... 10/ (1 conf/distro/openmoko.conf): Feb 29 14:14:37 openmoko.conf: add stuffs for jalimo Feb 29 14:14:37 * classpath, cacao-initial and swt3.4-gtk Feb 29 14:26:50 why is it not listing uclibc? DEBUG: providers for virtual/libc are: ['external-toolchain', 'eglibc', 'glibc'] Feb 29 14:27:05 I even defined PREFERRED_PROVIDER_virtual/libc = "uclibc" Feb 29 14:27:17 not sure where I screwed up but it's trying to build glibc Feb 29 14:32:22 Jin^eLD: it doesn't think uclibc is buildable, for whatever reason Feb 29 14:35:38 morning Feb 29 14:36:48 doh... mhm Feb 29 14:36:54 which part "decides" this? Feb 29 14:37:24 I did adjust the siteinfo.bbclass for my uclinux stuff Feb 29 14:37:27 so that should be ok Feb 29 14:39:12 Jin^eLD: the runqueue (actually something related to it) Feb 29 14:39:47 phu.. where do I find that? Feb 29 14:40:00 well, maybe you can give me a hint... what I did was - I defined a custom TARGET_OS Feb 29 14:40:15 I adapted siteinfo so it returns valid stuff Feb 29 14:40:38 is there anything else that needs to be done for a new TARGET_OS ? Feb 29 14:42:36 Jin^eLD: did the number of skipped files increase? ;) Feb 29 14:43:24 NOTE: Parsing finished. 4786 cached, 0 parsed, 247 skipped, 0 masked. Feb 29 14:43:24 NOTE: Parsing finished. 4786 cached, 0 parsed, 247 skipped, 0 masked. Feb 29 14:43:26 oops Feb 29 14:43:27 ssorry Feb 29 14:43:30 hehe, shaky hands ;) Feb 29 14:43:34 Jin^eLD: check uclibc.inc Feb 29 14:44:00 what I am looking for there? Feb 29 14:44:02 Jin^eLD: why do you have shaky hands? wasn't you trained by some secret organisation (starting with K and ending with GB) to not have shaky hands? Feb 29 14:44:15 Jin^eLD: for the python thingie at the top Feb 29 14:44:25 zecke: probably because I am hungry again ;) Feb 29 14:44:31 uc_os = (re.match('.*uclibc*', bb.data.getVar('TARGET_OS', d, 1)) != None) Feb 29 14:44:52 oh yes.. Feb 29 14:45:10 thanks, that's it Feb 29 14:45:10 Jin^eLD: is your new TARGET_OS including uclibc? ;) Feb 29 14:45:14 my TARGET_OS is uclinuxeabi Feb 29 14:45:22 so it's not including that string.. Feb 29 14:45:41 btw, those strings - are they somehwere "defined" in the sense - they MUST be like this and that Feb 29 14:46:03 Jin^eLD: my knowledge is ending there, if you think uclinuxeabi is sane and supported by uclibc Feb 29 14:46:06 or did someone just come up with something and its just a string? Feb 29 14:46:07 Jin^eLD: change that line :) Feb 29 14:46:32 zecke: codesourcery toolchain build script uses uclinuxeabi, Koen used uclinux-uclibc for bfin stuff Feb 29 14:46:48 so I am not sure either, but I'll just change the line and see what will happen :> Feb 29 14:47:03 I don't know either Feb 29 14:47:12 this whole uclinux stuff is really messy Feb 29 14:48:49 thanks for the uclibc pointer :> so moving one step further... Feb 29 14:51:38 zecke: will you be at the Linux Tag? Feb 29 14:53:10 woglinde: hi, do you have some time to help me? Feb 29 14:53:57 Jin^eLD: sure, unless I'm in Taipei or Oslo or wherever... Feb 29 14:54:38 I'll see if I can get there :> LT that is Feb 29 14:59:05 hi mr_nice sorry no Feb 29 14:59:28 mr_nice you could sent me an email Feb 29 15:07:18 woglinde: no problem. it's not that important Feb 29 15:13:01 03mickeyl 07org.oe.dev * ra658066f... 10/ (3 files in 2 dirs): add enter, an experimental inputmethod using efl Feb 29 15:13:09 03mickeyl 07org.oe.dev * r08976209... 10/ (3 files in 3 dirs): moko-underground-image: add some fonts to make matchbox happy, bump SRCREV for pyneod Feb 29 15:18:02 03mickeyl 07org.oe.dev * r91382b46... 10/ (6 files in 3 dirs): contrib: move svgs into dedicated directory Feb 29 15:18:06 03mickeyl 07org.oe.dev * rae9980c1... 10/ (1 packages/python/python_2.5.2.bb): python 2.5.2 python-crypt needs to RRECOMMEND openssl, otherwise import md5 (_hashlib actually) fails Feb 29 15:21:56 re robert Feb 29 15:28:55 heh http://www.youtube.com/watch?v=uTta0W5sXt0 Feb 29 15:39:16 hello steliosk , long time no see Feb 29 15:39:21 hi HopsNBarley ! Feb 29 15:40:47 likewise : hi ! Feb 29 15:41:13 likewise : My machine logs in but i am running arround a lot these days Feb 29 15:47:54 steliosk: hi Feb 29 15:48:03 steliosk: I got newsletter mail(s) - thx Feb 29 15:48:26 * steliosk is an official spammer now according to aol :) Feb 29 15:49:00 hi likewise - how's it going? Feb 29 15:50:21 HopsNBarley: pretty well, swamped with work tho Feb 29 15:50:46 steliosk: you ever worked with AMCC PowerPC parts? Feb 29 15:50:59 bbl Feb 29 15:51:20 likewise : yep. mostly work with ppc405 these days and ppc440 Feb 29 15:57:43 steliosk: any PCIe stuff you worked with on the AMCC parts? Feb 29 15:58:10 hey likewise, how's it going? Feb 29 15:58:28 Jin^eLD: quoting myself: "pretty well, swamped with work tho" :-) Feb 29 15:58:34 :) Feb 29 15:58:37 Jin^eLD: How's things in .at? Feb 29 15:58:47 bye Feb 29 15:58:52 rainy :> Feb 29 15:58:58 and messy in regard to my uclinux attempts Feb 29 15:59:05 likewise : no PCIe yet, only pci and now optimizing Oslec echo canceller to use 405's dsp instructions Feb 29 15:59:31 and hopefully cpu based ram as buffer Feb 29 16:00:35 steliosk: nice project Feb 29 16:01:41 likewise : heh gdb screen is imprinted on my retina now ;) Feb 29 16:01:55 steliosk: LOL :-) Feb 29 16:02:17 steliosk: Are you using OE to target the boards? Feb 29 16:02:25 likewise : yes Feb 29 16:03:00 03likewise 07org.oe.dev * rcc62ac9d... 10/ (5 files in 4 dirs): linux-rt-2.6.24: Added leds-cpu-activity. Feb 29 16:03:11 03mickeyl 07org.oe.dev * r9d51ade6... 10/ (5 files in 4 dirs): busybox 1.9.1 enable cmp utility, remove openmoko-specific defconfig (not necessary) Feb 29 16:03:12 steliosk: Because I see lots of patches on the linuxppc mailing list for the amcc dev boards Feb 29 16:03:49 likewise : what is happening is that the ppc tree that the amcc boards used is "moving" to the powerpc tree Feb 29 16:04:25 but in my case 2.6.20 ppc tree is fine Feb 29 16:04:53 the pcc tree is scheduled for removal in summer Feb 29 16:05:57 steliosk: ah ok, I'm now using 2.6.24.3-rt3 against the Freescale MPC831x. Feb 29 16:06:09 steliosk: that one is powerpc already, mostly fully working Feb 29 16:07:49 likewise : i think that machines that move from ppc->powerpc will have a few months till all new "features" get figured out Feb 29 16:08:09 so stayed with the old-and-working 2.6.20 Feb 29 16:25:28 Hmm Feb 29 16:25:38 Are there rules for when to move a bug you files from resolved/fixed to closed? Feb 29 16:25:59 s/files/filed/ Feb 29 16:28:09 rwhitby: ping? Feb 29 16:38:53 bye! Feb 29 16:54:03 is anyone using the gcc-csl* compilers? Feb 29 16:54:07 I see there are recipes available Feb 29 16:54:33 hm arnt they old ones? Feb 29 16:54:42 where csl did a lot of arm work Feb 29 16:54:54 well, csl also has some newer stuff Feb 29 16:55:09 there is a gcc 4.2.1 one with uclinux eabi support and some other stuff Feb 29 16:55:38 I hit a problem when trying to add uclinux to OE, gcc did not quite want to compile and after further digging I found out that it is missing some stuff that I need Feb 29 16:55:47 and csl had this stuff Feb 29 16:55:57 the csl*.bb in OE are indeed older Feb 29 17:01:28 Jin^eLD: They came from poky Feb 29 17:04:05 RP: I see.. well I'll try to get in the 2007q3 version as I seem to need it Feb 29 17:04:57 03rpurdie 07org.oe.dev * rbca6af36... 10/ (8 files in 3 dirs): libmatchbox: Remove old versions, add patch from poky which fixes the .pc file Feb 29 17:05:06 03rpurdie 07org.oe.dev * r8407454b... 10/ (9 files in 3 dirs): matchbox-wm: Drop old versions, add patch from poky to fix expat configure option Feb 29 17:05:13 03rpurdie 07org.oe.dev * rf0423a79... 10/ (3 files in 3 dirs): dtnrg: Add missing patch Feb 29 17:05:21 03rpurdie 07org.oe.dev * r64013d98... 10/ (3 files in 3 dirs): pango: Add patch to fix pkgconfig breakage (from poky) Feb 29 17:05:28 03rpurdie 07org.oe.dev * rf8805ae6... 10/ (3 files in 3 dirs): xorg-xserver-1.4: Fix pkgconfig file Feb 29 17:05:35 03rpurdie 07org.oe.dev * r12f1d8b8... 10/ (3 files in 3 dirs): matchbox-desktop-0.9.1: Add fix for pkgconfig file Feb 29 17:05:43 03rpurdie 07org.oe.dev * r2a94e224... 10/ (3 files in 3 dirs): gnutls-1.6.3: Update configure patch to address pkgconfig issues Feb 29 17:05:50 03rpurdie 07org.oe.dev * r359194a9... 10/ (4 files in 3 dirs): curl: Fix pkgconfig file Feb 29 17:05:57 03rpurdie 07org.oe.dev * r6d9979f4... 10/ (7 files in 4 dirs): gtk-webcore: Tweak pkgconfig usage so it works within a sysroot Feb 29 17:52:24 Jin^eLD: hmm, what are you missing from CSL in our compilers? Feb 29 17:52:47 Jin^eLD: I mean, buildroot builds for uclibc, we could glance there. Feb 29 17:53:09 Jin^eLD: Or use the stuff Greg did for snapgear's / uclinux toolchain Feb 29 17:54:37 one sec phone Feb 29 17:55:32 re Feb 29 17:55:41 likewise: which one is the snapgears/uclinux toolchain? Feb 29 17:55:49 likewise: basically, I am trying to build uclinux stuff with OE Feb 29 17:55:54 I had to add another TARGET_OS Feb 29 17:56:13 and I noticed that gcc 4.x.x does not have a config/arm/uclinux-eabi.h Feb 29 17:56:19 but the CSL one has this Feb 29 17:56:51 I do not know anything about Greg's snapgear stuff so if you have any pointers I'd be very happy Feb 29 18:01:28 Jin^eLD: hold on. I know of a few build projects (buildroot, ptxdist, ltib, uclinux) that *might* have uclibc support as of today. Feb 29 18:02:07 likewise: well, I want to have uclinux support in OE :) that's the goal Feb 29 18:02:13 I did look at the uclinux build system and did not really like it Feb 29 18:02:27 Jin^eLD: it does not build a toolchain Feb 29 18:02:40 right... Feb 29 18:02:46 Jin^eLD: they provide seperate toolchains+sources. However, looking at the web site, they are too old. Feb 29 18:02:52 I did look at the CSL toolchain build script Feb 29 18:03:01 Jin^eLD: buildroot might be interesting, let me check that one and ptxdist Feb 29 18:03:22 is anything speaking against csl? Feb 29 18:03:45 I also picked it because the customer wanted to use CSL for no idea what reasons Feb 29 18:04:25 Jin^eLD: CSL is commercially supported, but usually lagggs behind Feb 29 18:04:51 I'ld opt for ptxdist. I exchanged some gcc patches with them and they are 4.1.x and 4.2.x users Feb 29 18:05:26 Jin^eLD: you just go ahead with CSL, I'll look at ptxdist, ok ? :-) Feb 29 18:06:10 ok :) Feb 29 18:06:31 well, in the end we should be able to use anything supporting uclinux Feb 29 18:06:39 I guess that will be easier if I manage to put the rest of the things in place Feb 29 18:27:55 hmm, I have a gcc-cross_csl-arm-2007q3.bb and I defined PREFERRED_PROVIDER_virtual/arm-angstrom-uclinuxeabi-gcc = "gcc-cross" and also PREFERRED_VERSION_gcc-cross_uclinuxeabi = "csl-arm-2007q3" but still it says: Feb 29 18:27:56 preferred version csl-arm-2007q3 of gcc-cross not available (for item virtual/arm-angstrom-uclinuxeabi-gcc) Feb 29 18:28:04 any ideas what it might be disliking? Feb 29 18:30:28 Jin^eLD: It might be possible that it assumes that it is impossible to build your preferred one. Feb 29 18:30:36 why would it assume that? Feb 29 18:30:55 it does go through the .bb file without errors Feb 29 18:33:13 Jin^eLD: a missing dependency e.g. you could try to build it directly Feb 29 18:33:30 thanks, let me se... Feb 29 18:33:33 s/se/see/ Feb 29 18:33:48 you are right, I forgot to update binutils Feb 29 18:33:53 thanks Feb 29 18:34:18 Jin^eLD: yw :) Feb 29 18:34:21 :) Feb 29 18:34:27 * florian -> home, bbl Feb 29 19:21:43 l8r Feb 29 19:22:06 03jeremy_laine 07org.oe.dev * r1c40da9a... 10/ (1 packages/madwifi/madwifi-ng_r.inc): madwifi-ng_r.inc: fix the list of packages built from madwifi-ng Feb 29 19:22:11 03mickeyl 07org.oe.dev * r2450408c... 10/ (1 conf/distro/include/sane-srcrevs.inc): sane-srcrevs.inc: bump opkg[-native] to a revision that compiles Feb 29 19:22:16 03cbrake 07org.oe.dev * rac63b4f4... 10/ (12 files in 5 dirs): linux 2.6.24: update cm-x270 to use 2.6.24 kernel Feb 29 19:22:22 03rpurdie 07org.oe.dev * r1543d455... 10/ (18 files in 6 dirs): pango: Remove unstable testing versions and stable versions with later bugfix point releases Feb 29 19:32:55 RP: did you touch the git branch code lately? Feb 29 19:33:02 it looks like it no longer gets the correct rev here Feb 29 19:33:04 instead I get 1 Feb 29 19:34:28 cbrake: ping. There is also 2.6.24-rt (real-time) which you might find interesting. Feb 29 19:36:26 likewise: good deal. I'm using the rt patch on one project, but did not realize you had set up a recipe un dev Feb 29 19:37:21 cbrake: I'm understanding how to deal with proper interrupt thread prioritization etc, from userland. Feb 29 19:39:37 likewise: yeah, there is a lot to all that! Feb 29 20:06:04 any adzs-bf548 users? I'd like to play with this platform, but it looks like binutils doesn't support it's blackfin processor and the build dies. Feb 29 20:08:12 hi? Feb 29 20:08:15 anybody using miniinit here? Feb 29 20:09:14 apaulsen : If i remember correctly Koen gave bf a try but had issues Feb 29 20:09:32 apaulsen : not sure at what stage its now Feb 29 20:09:47 steliosk: looks to be at the same stage :) Feb 29 20:10:02 I will ask koen Feb 29 20:18:30 re Feb 29 20:34:28 "-c rebuild" seems not to do anything. I have to manually do a "-c clean" step. has something changed recently with this? Feb 29 20:34:34 03mickeyl 07org.oe.dev * r013a96e5... 10/ (3 files in 2 dirs): Feb 29 20:34:34 u-boot-openmoko-devel git|svn build variant for gta02v5, remove variants for hardware almost no Feb 29 20:34:34 one has Feb 29 20:35:48 and "-c clean" does not remove the workdirectory any more. what is needed to restore the old behavior? Feb 29 20:43:35 rschuster: Maybe that's related to the major changes RP is working on Feb 29 20:48:14 ok Feb 29 20:58:18 rschuster: you need to call "bitbake -c fetch -f $target" now Feb 29 20:58:23 yes Feb 29 20:58:31 that is the current -c rebuild ;-) Feb 29 20:58:51 RP had sent a mail to the list about it Feb 29 20:59:23 Laibsch: ah ok. thanks! Feb 29 21:03:07 is current OE monotone repository compatible with monotone 0.39 ? Feb 29 21:07:22 I think so .... Feb 29 21:07:43 I think my Fedora box may have jsut updated to that so I am trying a pull Feb 29 21:07:49 it appears to be working gremlin[it] Feb 29 21:09:11 re Feb 29 21:09:47 Crofton me too!!! i ask for the same reason :) Feb 29 21:11:03 it looks like it is working for me :) Feb 29 21:11:15 I haven't updated yet, but mtn gives no errors Feb 29 21:11:19 or warnings Feb 29 21:22:54 good thanks Crofton Feb 29 21:33:42 * likewise measuring a worst case scheduling latency reduction from 1.7 seconds to 81 microseconds using the real-time kernel under heavy flood ping load. Whee! Feb 29 21:34:38 1.7 seconds sounds awful though, checking... Feb 29 21:37:32 bye Feb 29 21:46:04 * * OE Bug 3897 has been created by trini(AT)embeddedalley.com Feb 29 21:46:06 * * mtd-utils-dev is empty Feb 29 21:46:08 * * http://bugs.openembedded.net/show_bug.cgi?id=3897 Feb 29 21:46:29 patch attached to that one to fix it, too Feb 29 22:01:18 Tartarus: thanks, will have a look at it right away Feb 29 22:03:55 yay thanks. Feb 29 22:05:05 likewise, is there any special checking one does before moving something from resolved/fixed to closed, if they filed the bug? Feb 29 22:05:21 like would it stay open until propogated to the stable branches or something? Feb 29 22:06:44 Tartarus: no, mostly the committer (to the .dev branch) closes the bug right away. We have no review policy, nor do we wait for the bug committer to close/resolve it. Feb 29 22:07:28 Tartarus: Only one distribution (angstrom) maintains a public stable branch AFAIK. For the rest, .dev is much like the kernel development, except we don't have merge windows. Feb 29 22:07:37 ah, k, thanks. Feb 29 22:07:45 I had only heard the stable branch mentioned in passing Feb 29 22:08:00 Tartarus: angstrom-distribution.org Feb 29 22:08:35 Tartarus: it's like .dev, with carefully selected cherry picks from .dev (which do go through careful, >2, reviews) Feb 29 22:09:17 k Feb 29 22:09:30 next time we're picking a base internally i'll see just what is/isn't there I think, thanks. Feb 29 22:19:07 Tartarus: hmm yeah, I usually do pick a revision just on .dev. Feb 29 22:19:17 and stick with it during a project Feb 29 22:19:45 BTW, I have mtd-utils-1.1.0 pending for release. It missed your fix as well. I'll commit it, with DEFAULT_PREFERENCE = "-1" Feb 29 22:20:00 k Feb 29 22:20:11 technically the 1.0.0.bb has the problem too, but that wasn't the default package Feb 29 22:20:17 and it has a comment about not working Feb 29 22:20:31 as it is i'm verifying in an older snapshot and then eyeballing newer ones, so I didn't wanna push it :) Feb 29 22:27:23 Can someone else try building ltp? Feb 29 22:27:35 It doesn't package for me due to shipping some host bins that seem to come from upstream Feb 29 22:27:44 not 100% sure if that's a problem or not Feb 29 22:32:29 Tartarus: LTP need to be updated real bad. There is a new upstream ltp btw... Feb 29 22:32:39 woglinde: hi Feb 29 22:32:47 hi mrdata Feb 29 22:33:04 likewise, yeah, I noticed it's a year old, too Feb 29 22:33:08 spotted the problem Feb 29 22:33:12 rm ${D}/usr/libexec/ltp/ballista -rf Feb 29 22:33:14 woglinde: wpa works with i-gate now (hostap_cs) ;-) Feb 29 22:33:15 in do_install Feb 29 22:33:22 mrdata hehe Feb 29 22:33:23 should be ltp/testcases/ballista Feb 29 22:33:34 I am now watching boxing Feb 29 22:33:37 till later Feb 29 22:33:42 okay Feb 29 22:34:49 mrdata: wpa with these old i-gate cards? Feb 29 22:35:32 yes, old but good quality Feb 29 22:37:25 Tartarus: in fact LTP has released a new version *today*. Care for a .bb ? :-) Feb 29 22:37:28 I still have one of these... somewhere Feb 29 22:37:52 i have three of them, so working enviroment was a must for me Feb 29 22:38:38 likewise, heh Feb 29 22:38:52 flo_lap: with sta firmware 1.7.1 Feb 29 22:45:18 mickeyl: Your package doesn't have a SRCREV set Feb 29 22:49:05 * * OE Bug 3897 has been RESOLVED (FIXED) by likewise(AT)gmx.net Feb 29 22:49:07 * * mtd-utils-dev is empty Feb 29 22:49:08 * * http://bugs.openembedded.org/show_bug.cgi?id=3897 Feb 29 22:50:19 yay Feb 29 22:53:58 03likewise 07org.oe.dev * r23aaddf4... 10/ (1 packages/mtd/mtd-utils_1.0.0+git.bb): mtd-utils-1.0.0+git: Install headers for -dev. Feb 29 22:54:02 re Feb 29 22:54:06 03likewise 07org.oe.dev * r5b8c5c49... 10/ (3 files in 2 dirs): mtd-utils: Added 1.1.0 with -1 preference (Scheduled for default 15-03-2008). Feb 29 22:54:14 ah Feb 29 22:58:13 time to delete my mtd-utils-1.1 bb Feb 29 23:03:36 I've just updated bitbake http://pastebin.com/m689958c0 Feb 29 23:04:15 I suspect r1021 by RP, Remove digraph and switch to new stamp checking code. Feb 29 23:04:43 ynezz: Wipe out your existing tmp/cache files Feb 29 23:05:01 ynezz: I will bump the cache version so this doesn't happen to anyone else Feb 29 23:06:03 ok, thanks Feb 29 23:06:47 03rpurdie * r1026 10/ (2 files in 2 dirs): cache.py: Bump the cache version since digraph was removed and mayb still be present in old cache files Feb 29 23:06:53 there we go :) Feb 29 23:06:58 fast :) Feb 29 23:07:05 hehe Feb 29 23:12:50 woglinde: you had a local copy as well? Feb 29 23:12:54 woglinde: (mtd) Feb 29 23:13:04 mtd yeah as mrdata and mrnice Feb 29 23:14:30 'night all Feb 29 23:14:36 nite rp Feb 29 23:14:56 night Feb 29 23:15:12 also night all from me Feb 29 23:15:23 nite mrdata Feb 29 23:17:16 night all Feb 29 23:19:44 likewise: gn Feb 29 23:20:40 flo_lap: nice catch on the XML thing, you beat me to it :-) Feb 29 23:21:01 heh Feb 29 23:21:31 XorA|gone: I ran into this situation before... with hh.o Feb 29 23:21:49 flo_lap: yeah, I suppose you did Feb 29 23:22:17 flo_lap: its nice even the atachments are there Feb 29 23:22:31 * XorA|gone hadnt expected that Feb 29 23:23:00 XorA|gone: right, the bug data shold be complete but you need the accounts to get these back into bugzilla Feb 29 23:23:29 flo_lap: you could always automate account creation with random passwords and send out emails Feb 29 23:23:38 flo_lap: get the account email addresses from the bugs Feb 29 23:23:48 flo_lap: any bounces assign to the admin Feb 29 23:25:02 XorA|gone: that sounds like a good idea Feb 29 23:25:26 flo_lap: I assume a perl guru (if we have one) could crunch the data in a few hours Feb 29 23:25:58 * flo_lap fears perl Feb 29 23:27:11 * XorA|gone dones many hours in perl, but is not a guru Feb 29 23:28:45 My first open source conribution was written in perl :) Feb 29 23:29:17 But I did never knew what I were doing ;) Feb 29 23:29:27 heh heh, Ive done numerous closed source stuff in perl, normally when the windows guys wish they had X data in Y form Feb 29 23:29:56 C code to Rational Rose, C code to XML for bluetooth stack analyser, etc etc Feb 29 23:33:07 03likewise 07org.oe.dev * r541709cb... 10/ (1 packages/util-linux-ng/util-linux-ng.inc): util-linux-ng: 2.13.1 final release, remove -rc2. Feb 29 23:36:22 good nite Feb 29 23:39:39 Tartarus: ltp 20080229 flying in for a landing. Feb 29 23:39:41 XorA|gone: perl? I would do that in php rather Feb 29 23:40:11 hrw|gone: bugzilla is perl, If rather not add another crazy language in the mix :-D Feb 29 23:40:23 perl is crazy Feb 29 23:44:58 its one single curl command to dump the whole of a bugzilla XML :-) Feb 29 23:46:03 likewise, very nice Feb 29 23:47:38 Tartarus: hitting a small bug with some tests being compiled for host - hold on. Feb 29 23:50:46 anyway, is time for sleep Feb 29 23:50:52 bye Mar 01 00:47:58 03mickeyl 07org.oe.dev * r7ebcf659... 10/ (4 files in 3 dirs): Mar 01 00:47:58 glibc 2.5|2.6.1 actually apply the local-dynamic-resolvconf.patch instead of Mar 01 00:47:58 just unpacking it into debian/patches... ~lart debian... Mar 01 00:48:11 03likewise 07org.oe.dev * r9d236d0f... 10/ (9 files in 5 dirs): ltp-20080229: Newly arrived, the leap day surprise. Mar 01 00:48:21 03likewise 07org.oe.dev * rdee683d3... 10/ (4 files in 3 dirs): ltp: 2008-02-29 release. Patched to build. ballista test still fails detecting native perl. Not tested on any target yet. Mar 01 00:48:28 03mickeyl 07org.oe.dev * re0d71aa6... 10/ (1 packages/images/moko-underground-image.bb): moko-underground-image: misc dependency fixes Mar 01 00:59:05 * * OE Bug 3898 has been created by  Mar 01 00:59:07 * * gcc-cross-4.2.2-r5-do_compile Mar 01 00:59:09 * * http://bugs.openembedded.net/show_bug.cgi?id=3898 Mar 01 01:18:42 there's a problem with install and -s strip option, it's using strip from host not for target Mar 01 01:19:00 is there some easy workaroung or just patch makefile? Mar 01 01:19:14 s/workaroung/workaround Mar 01 04:47:04 * * OE Bug 3899 has been created by  Mar 01 04:47:06 * * uclibc-initial-0.9.29-r9-do_populate_staging Mar 01 04:47:08 * * http://bugs.openembedded.net/show_bug.cgi?id=3899 Mar 01 05:51:05 * * OE Bug 3900 has been created by  Mar 01 05:51:06 * * glibc-intermediate-2.6.1-r3-do_compile Mar 01 05:51:08 * * http://bugs.openembedded.net/show_bug.cgi?id=3900 Mar 01 10:24:00 hi? short question: when i have lots of RAM, for which directories of my OE-env I could profit from a shm-device? Mar 01 10:25:27 or could i profit from an fs with smaller blocksize? Mar 01 10:54:32 lo just a fyi for whoever modified opkg - it's broken **** ENDING LOGGING AT Sat Mar 01 10:59:57 2008