**** BEGIN LOGGING AT Thu Jan 04 03:00:02 2007 Jan 04 07:47:01 03hrw 07org.oe.dev * r4d3df47b... 10/ (1 packages/syslinux/syslinux-native_3.11.bb): syslinux: fix SRC_URI Jan 04 07:47:07 03hrw 07org.oe.dev * rd880ea61... 10/ (4 files in 2 dirs): syslinux: unify Jan 04 07:47:12 03hrw 07org.oe.dev * r195a518c... 10/ (6 files in 2 dirs): syslinux: added 3.31, fixed previous commit, unified Jan 04 07:54:01 03hrw 07org.oe.dev * r4330d1ea... 10/ (7 files in 4 dirs): ifupdown: updated to 0.6.8 (current Debian version) Jan 04 07:59:34 ~logs Jan 04 07:59:40 apt/ibot/jbot/purl all log to http://ibot.rikers.org// where channelname is html encoded ie: %23debian | lines that start with a space are not shown | some channels have stats at http://ibot.rikers.org/stats/.html.gz, or updated "nightly" Jan 04 08:00:09 03hrw 07org.oe.dev * r2a00a5d3... 10/ (1 packages/cvs/cvs_1.12.8.bb): cvs 1.12.8: fix SRC_URI Jan 04 08:00:43 hrw: morning, saw you submitted a patch to palm TX Jan 04 08:00:49 can it do OE? Jan 04 08:15:01 03hrw 07org.oe.dev * rfab3408b... 10/ (1 packages/lftp/lftp_3.5.1.bb): lftp: fix SRC_URI for 3.5.1 Jan 04 08:15:15 morning fooks Jan 04 08:21:23 03hrw 07org.oe.dev * r09305c82... 10/ (1 packages/uboot/uboot_cvs.bb): u-boot: dropped CVS version - upstream moved to git tree Jan 04 08:21:29 morning Jan 04 08:21:30 03hrw 07org.oe.dev * r624e4cbb... 10/ (1 packages/uboot/u-boot_git.bb): u-boot: added GIT version with DEF_PREF = -1 to not select it automatically Jan 04 08:30:25 morning Jan 04 08:44:06 morning Jan 04 08:44:17 niv_one_three: palmtx can run OE derived distros Jan 04 08:44:27 hrw: thnkas Jan 04 08:44:45 ok, here is the deal, I got my dad a TX Jan 04 08:44:58 can I flash it and be confident ? Jan 04 08:45:11 dont know Jan 04 08:46:10 my dad is tech- chlanged Jan 04 08:46:32 how uses the palm TX distro/. Jan 04 08:46:34 ? Jan 04 08:47:27 ask hack&dev people Jan 04 08:47:43 seems they still sleep Jan 04 08:47:48 thanks Jan 04 09:06:46 fsck. Jan 04 09:12:44 morning all Jan 04 09:12:55 hi RP koen Jan 04 09:13:22 hey RP Jan 04 09:14:51 03hrw 07org.oe.dev * r625edac0... 10/ (1 packages/linux/linux-hackndev-2.6_svn.bb): linux-hackndev: added recipe (not ideal one) Jan 04 09:15:53 good morning all Jan 04 09:15:57 heh Jan 04 09:16:03 "not ideal one" Jan 04 09:16:33 * koen suspects H&D is a cover for pavel and sash working together to create to most crappy software ever Jan 04 09:25:56 ~change 2000 usd to pln Jan 04 09:26:03 2,000.00 United States Dollar (USD) makes 5,867.41 Zloty (PLN) (from http://www.xe.com/) Jan 04 09:30:42 * koen still wonders where to put module_autoload_usbcore = "usbcore" Jan 04 09:31:15 linux-rp.inc Jan 04 09:31:31 that doesn't help my netbook Jan 04 09:31:56 * koen has http://rafb.net/p/8sdddz82.html Jan 04 09:32:26 that'll do as well Jan 04 09:33:03 I thinking of writing a bb that has an initscript that does "modprobe usbcore ; mount /proc/bus/usb" Jan 04 09:33:23 and make task-base include that for MACHINE_FEATURES=usbhost Jan 04 09:34:20 but I'm puzzled why usbfs 'works in OZ' in .dev Jan 04 09:35:03 it does? Jan 04 09:35:10 according to justinp Jan 04 09:35:23 I wonder too Jan 04 09:35:36 usbcore isnt loaded here Jan 04 09:35:46 or wasnt, I have no oz installs anymore Jan 04 09:36:10 usbfs /proc/bus/usb usbfs noauto 0 0 Jan 04 09:36:15 thats spitz/oz Jan 04 09:36:44 right 'noauto' Jan 04 09:37:19 I changed that to 'defaults' to get it mounted for people that build it into the kernel Jan 04 09:37:23 bbbl Jan 04 09:41:24 03hrw 07org.oe.dev * rab77b999... 10/ (1 conf/machine/palmld.conf conf/machine/palmtx.conf): palmld/palmtx: tune for XScale Jan 04 09:41:44 hi folks Jan 04 09:42:25 ~lart me and give me coffee Jan 04 09:42:26 * ibot decapitates me and give me coffee conan the destroyer style Jan 04 09:43:03 RP: can we extact config handling from cooker? Jan 04 09:43:09 RP: in bitbake Jan 04 09:47:01 hi need some help. Jan 04 09:48:17 anushsh: dont ask to ask but ask Jan 04 09:48:41 ok, i have been compiling for the kernel2.6 Jan 04 09:49:14 i have been encountering some do_compile error in the qemu Jan 04 09:50:30 hrw|work : what could be that Jan 04 09:50:36 you do not have gcc 3.3 or 3.4 on host Jan 04 09:51:10 hrw|work : i installed gcc 3.4 now Jan 04 09:51:16 but still the same error Jan 04 09:51:31 bitbake -crebuild qemu-native fails still? Jan 04 09:51:52 do i have to specify anywhere gcc-3.4 anywhere ? Jan 04 09:52:05 anushsh: first do Jan 04 09:52:06 bitbake -crebuild qemu-native Jan 04 09:52:18 03hrw 07org.oe.dev * rca7eaf62... 10/ (1 packages/zip/zip-native_2.31.bb packages/zip/zip_2.31.bb): zip: drop 2.31 (not fetchable) Jan 04 09:52:35 hrw|work : ok Jan 04 09:53:31 hrw|work: Which bit of the config handling. I left the option parsing in bin/bitbake Jan 04 09:53:46 RP: I will show you my bittest version Jan 04 09:54:14 ~lart qemu Jan 04 09:54:15 * ibot strangles qemu with a 9-pole serial cable Jan 04 09:54:51 RP: http://www.openembedded.org/~hrw/bittest.txt Jan 04 09:56:21 hrw|work: So you mean add BBConfiguration somewhere accessible? Jan 04 09:56:49 hrw|work : now again do_configure failed Jan 04 09:57:06 RP: exactly Jan 04 09:57:11 anushsh: What is your gcc 3.4 compiler called? Jan 04 09:57:46 hrw|work: The problem is we need to add configuration handling to cooker, not the other way around! :) Jan 04 09:58:18 RP: ok Jan 04 09:58:47 03hrw 07org.oe.dev * rde62ae92... 10/ (1 packages/libxml/libxml2-native_2.6.9.bb): libxml2-native: fixed SRC_URI for 2.6.9 Jan 04 09:59:13 RP: I will try to clean bittest/source_checker or write bitfetch from them Jan 04 09:59:41 hrw|work: I guess we need to move more of bin/bitbake into a lib/bb/bbconfig.py Jan 04 09:59:45 03hrw * r723 10bitbake/setup.py: setup.py: use FHS compatible directories Jan 04 09:59:59 RP: yep Jan 04 10:00:41 03hrw 07bitbake-1.6 * r724 10/setup.py: setup.py: use FHS compatible directories (from trunk) Jan 04 10:03:25 RP : gcc-3.4 Jan 04 10:04:07 RP : the log says SDL needs to be installed for the graphical output Jan 04 10:05:20 03hrw * r725 10bitbake/lib/bb/ (17 files in 2 dirs): Jan 04 10:05:20 drop shebangs from lib/bb scripts Jan 04 10:05:20 - those scripts are not called by hand by bitbake users Jan 04 10:05:20 and most of package linting tools (lintian/Debian, rpmlint/Fedora) Jan 04 10:05:20 complain about not executable scripts. Jan 04 10:05:20 - if someone want to run them by hand then he will be able to do it Jan 04 10:05:22 with calling 'python SCRIPT' Jan 04 10:05:38 anushsh: so you know the answer Jan 04 10:06:01 RP: now I will apply it to bitbake-1.6 Jan 04 10:06:16 hrw|work : yep, trying to find the ubuntu packeages for libsdl Jan 04 10:06:20 *packages Jan 04 10:06:34 anushsh: apt-get install libsdl12-dev iirc Jan 04 10:06:56 hrw|work : thanks a lot Jan 04 10:08:19 RP: http://bugs.openembedded.org/show_bug.cgi?id=97 - can you look? Jan 04 10:09:14 03hrw 07bitbake-1.6 * r726 10/lib/bb/ (15 files in 2 dirs): Jan 04 10:09:14 drop shebangs from lib/bb scripts (from trunk) Jan 04 10:09:14 - those scripts are not called by hand by bitbake users Jan 04 10:09:14 and most of package linting tools (lintian/Debian, rpmlint/Fedora) Jan 04 10:09:14 complain about not executable scripts. Jan 04 10:09:14 - if someone want to run them by hand then he will be able to do it Jan 04 10:09:14 with calling 'python SCRIPT' Jan 04 10:17:54 hrw|work: go for it Jan 04 10:18:23 hrw|work: It will need rewriting once we have the xmlrpc event interface implmented but such is life Jan 04 10:25:34 Packaged contents of task-base into /a/home/hrw/devel/build/angstrom/tmp/deploy/glibc/ipk/task-base_1.0-r15_palmld.ipk Jan 04 10:26:02 now time for palmtx Jan 04 10:26:45 hrw|work: OE builds libsdl-native Jan 04 10:26:54 no need for sdl on the host Jan 04 10:27:40 koen|away: good to know Jan 04 10:34:24 03hrw * r727 10bitbake/lib/bb/build.py: Jan 04 10:34:24 build.by: Added option to limit amount of log lines displayed - close #97 Jan 04 10:34:24 - option is named BBINCLUDELOGS_LINES and is integer Jan 04 10:34:24 - if option is set then logs are passed via 'tail' command Jan 04 10:38:37 good morning Jan 04 10:39:00 03hrw 07org.oe.dev * r871ce362... 10/ (1 conf/documentation.conf): documentation.conf: describe BBINCLUDELOGS/BBINCLUDELOGS_LINES Jan 04 10:41:10 03hrw * r728 10bitbake/lib/bb/build.py: moved 'here comes the log' message back to proper place Jan 04 10:42:22 Packaged contents of task-base into /a/home/hrw/devel/build/angstrom/tmp/deploy/glibc/ipk/task-base_1.0-r15_palmtx.ipk Jan 04 10:46:26 hrw|work: s/include/require/g in conf/machine Jan 04 10:46:43 koen: thats the plan in todo Jan 04 10:46:55 mtn: 8 heads on branch 'org.openembedded.dev' Jan 04 10:47:33 8!??! Jan 04 10:48:21 62 files changed, 80 insertions(+), 80 deletions(-) Jan 04 10:48:33 koen: fault here Jan 04 10:53:50 now I have only 3 heads Jan 04 10:55:11 03hrw 07org.oe.dev * r500bd26a... 10/ (1 packages/uboot/uboot-devkitidp-pxa255_cvs.bb): uboot-devkitidp-pxa255: merged in old uboot_cvs recipe, marked as BROKEN Jan 04 10:55:18 03hrw 07org.oe.dev * re0d9912a... 10/ (40 files in 2 dirs): conf/machine/*.conf: require everything which needs to be included Jan 04 11:03:03 hmmm Jan 04 11:03:22 I forgot to bill some work Jan 04 11:04:02 hrw|work: can you share the name of the company that contracted haerwu yet? Jan 04 11:07:59 morning all Jan 04 11:11:25 koen: in 2 weeks will do it Jan 04 11:11:27 hi Dirk Jan 04 11:11:35 hey Dirk Jan 04 11:11:42 hey Marcin, koen Jan 04 11:12:14 hrw: I found www.hrw-consulting.de :) Jan 04 11:16:02 do13_: thats small when you compare them to hrw.org ;) Jan 04 11:17:14 ok. I see Jan 04 11:18:10 hey do13_ Jan 04 11:18:23 hey XorA Jan 04 11:18:52 hey XorA Jan 04 11:19:17 do13_: was it my imagination, or did I catch something from you said you started a collie SD driver? Jan 04 11:19:55 XorA: Over the chrismas days I started work: Jan 04 11:20:38 XorA: I have a spi_locomo.ko, but it isn't complete Jan 04 11:21:33 do13_: wicked Jan 04 11:21:49 do13_: you had card-insert detection working a while ago, right? Jan 04 11:22:17 koen: No, card-insertion won't work, because my connector is broken Jan 04 11:22:27 do13_: want another collie? Jan 04 11:22:37 koen: I have some communication Jan 04 11:22:45 do13_: or need? Jan 04 11:22:58 hrw|work: maybe later Jan 04 11:23:40 hrw|work: If I have another collie then the pressure is too high :) Jan 04 11:23:45 do13_: ;) Jan 04 11:24:42 do13_: Is what you've written so far available anywhere? Jan 04 11:25:15 RP: not yet Jan 04 11:26:23 RP: let me clean up some parts. Jan 04 11:26:34 * XorA cant wait until collie Angstrom Jan 04 11:27:21 collie needs more work. the other subsystems are ugly Jan 04 11:27:35 * RP agrees Jan 04 11:27:42 beautifull compared to 2.4.crappix :) Jan 04 11:28:51 then we can scoure ebay for old collies, clean them, upgrade, sell them for a profit as useful :-) Jan 04 11:29:23 XorA: software upgrades... Jan 04 11:29:48 amsdelta - who maintain it? Jan 04 11:29:55 hrw|work: noodles Jan 04 11:30:19 I will tweak config a bit Jan 04 11:30:28 RP: basically I can read / write on the spi, also I know how to switch the freq (25Mhz, 6.25Mhz, 3.12Mhz, 0.385Mhz), assert CS, apply power Jan 04 11:30:49 RP: now I working on registering this on the spi bus Jan 04 11:31:23 do13_: if i would be able to get my collie bootable with my kernel i could do some testing for ya Jan 04 11:31:52 rw: thx Jan 04 11:32:06 i'm not really in love with bitbake Jan 04 11:32:11 +yet Jan 04 11:32:32 hi Crofton Jan 04 11:33:48 do13_: _maybe_ i could write some stuff...but i have some exams this month Jan 04 11:35:23 rw: cool. setup bitbake and have a look at the kernel :) Jan 04 11:36:03 ,) Jan 04 11:36:44 btw. - is 2.6.19 known to work on collie - cause my kernel is unable to find partitions on the MTD-Device *shrug* Jan 04 11:38:22 rw: yep. do you changed the config? Jan 04 11:38:43 re Jan 04 11:40:58 do13_: That sounds like good progress :) Jan 04 11:41:03 do13_: i used oz-config for 2.6.17 and it didn't work Jan 04 11:49:54 WTF, trolltech webinars on multithreading??? Jan 04 11:52:03 where? Jan 04 11:52:16 hrw|work: adverts in theregister Jan 04 11:55:39 ah. Jan 04 11:55:45 03hrw 07org.oe.dev * r00d78946... 10/ (1 conf/machine/amsdelta.conf): amsdelta: do not hand tune for arm920t but include proper tune config Jan 04 11:55:48 I stopped reading the register some time ago Jan 04 11:55:56 too many posts per day Jan 04 12:04:38 rw: http://www.do13.de/oz/temp/defconfig-collie Jan 04 12:05:56 do13_: i will have a look at it Jan 04 12:05:58 bbl Jan 04 12:06:01 thsdx Jan 04 12:06:04 -sd Jan 04 12:14:36 RP: Ok, so I feel like an idiot! The issue causing my kernel to crash was that when I switched to linux-rp_2.6.19.bb I never set up mem=68M for the kernel config and although I was using the same config file the linux-rp.inc writes over the config line with its on settings. Once I added in settings for the lpd270 with mem=64M my crashes went away. Jan 04 12:21:27 http://pastebin.ca/304552 Jan 04 12:22:55 Crofton: fscking ppc != powerpc problem Jan 04 12:24:07 yeh Jan 04 12:24:32 not sure how it is deatl with though Jan 04 12:25:40 I thought koen had built angstrom for the efika Jan 04 12:25:50 maybe he excluded certain bb files :) Jan 04 12:26:45 or added more 'if ppc then powerpc' tweaks Jan 04 12:35:16 hrw|work : its workign well. Thanks a lot once again :) Jan 04 12:37:13 I modified something to not include madwifi Jan 04 12:41:21 :) Jan 04 12:41:42 MACHINE_FEATURES -= "pci" Jan 04 12:42:55 ~seen _law_ Jan 04 12:42:59 _law_ was last seen on IRC in channel #gpe, 44d 2h 57m 29s ago, saying: 'hi florian_kc '. Jan 04 12:45:16 * koen is tempted to blaim the dbus problems on the arm cache issue Jan 04 12:50:13 koen: angstrom have one arm-angstrom-linux-gnueabi-gcc and it then generate arm v4/v5/v6 code? Jan 04 12:50:49 hrw|work: in practice, yes Jan 04 13:05:33 koen: where is this weeks super gtk+ optimisation?? Jan 04 13:06:10 XorA: I think the gtk dudes got drunk and spent their sober time on figuring out svn Jan 04 13:06:45 ~hail Debian Jan 04 13:06:59 * ibot bows down to Debian and chants, "I'M NOT WORTHY!!" Jan 04 13:07:13 NOTE: package libnet-1.0.2a-r1: task do_rebuild: completed Jan 04 13:11:06 koen: how often ewi syncs with main? Jan 04 13:11:37 every 10 minutes Jan 04 13:11:52 if it can connect to oe.org Jan 04 13:12:31 ok Jan 04 13:14:09 03hrw 07org.oe.dev * ref1a4a23... 10/ (3 files in 2 dirs): libnet: fix SRC_URI by pointing to Debian mirrors Jan 04 13:25:15 ~help Jan 04 13:26:50 hi Cliff Jan 04 13:29:00 mickeyl: screenshot! Jan 04 13:29:08 mickeyl: good morning :) Jan 04 13:29:11 heh Jan 04 13:29:11 morning Jan 04 13:29:14 screenshot from what? Jan 04 13:29:22 hrw|work: morning Jan 04 13:29:27 the bitbake curses thing Jan 04 13:29:29 heh Jan 04 13:29:35 not much you can see, frankly Jan 04 13:30:29 mickeyl: did you get your devboard working? Jan 04 13:30:33 RP: you still want me to make this more generic way of coupling the main loop plus additional stuff or can you go from now? Jan 04 13:31:43 koen: I took Harald's LCM with me and if it didn't get broken during transport then I'm finally able to work tonite with the production LCM on the devel board. *crossing fingers* first thing i'm doing is some benchmarks & dpi experiments Jan 04 13:31:47 mickeyl: what time zone are you on? Jan 04 13:32:30 niv_one_three: CET. Jan 04 13:32:47 the beautiful city of Frankfurt/Main Jan 04 13:33:13 oh , ok Jan 04 13:33:28 just that you sho up here round this time Jan 04 13:33:40 while koen and hrw arrive before Jan 04 13:33:43 my presence is totally unrelated to time Jan 04 13:33:47 sometimes i work during day Jan 04 13:33:50 sometimes i work during night Jan 04 13:34:03 that's the "free" in "freelancer" :D Jan 04 13:34:34 so, I want to add bluetooth by using albertr hack Jan 04 13:34:47 I see familier 0.8.4 got bluetooth Jan 04 13:34:58 what is my best bet to track this? Jan 04 13:35:32 _if_ you want to use OE, then you could diff the git repository against OE .dev Jan 04 13:35:34 I am seeking a file to show what code is used Jan 04 13:35:56 your best bet may be dropping over to #handhelds.org or #familiar and just ask the guys what they did to add BT to it Jan 04 13:38:35 mickeyl: "familiar got bluetooth" == "bluez is installed" Jan 04 13:38:48 possibly Jan 04 13:38:57 certainly Jan 04 13:45:08 ok Jan 04 13:48:50 If I where to ask about where bluetooth code for the module is in OE Jan 04 13:48:54 now someone make linux bluetooth work nice :-) Jan 04 13:49:00 what file should I look in to? Jan 04 13:49:13 XorA: how? Jan 04 13:49:50 niv_one_three: not being insanely difficult to make two devices talk would be a start :-) Jan 04 13:51:24 .oO(under gnome this works pretty well) Jan 04 13:53:12 rw: gnome learned SDP while I wasnt looking? Jan 04 13:53:50 maybe ,) Jan 04 13:56:45 re Jan 04 14:01:22 where is the bluetooth code for the module in OE ? where should I run the search for bluez? google aint helping me Jan 04 14:01:37 (sorry about asking about familier here) Jan 04 14:02:21 userland part is in bluez/* Jan 04 14:02:25 kernel part is in kernel Jan 04 14:02:40 and for simpad-bt hack nothing more is needed Jan 04 14:02:47 it is simple BT-over-serial Jan 04 14:04:28 hrw: but I need to tell th kernel what pins to use - right? Jan 04 14:04:45 that is general perpess IO Jan 04 14:04:54 niv_one_three: iirc DECT port in simpad was normal serial... Jan 04 14:05:21 hrw: I soldered the bt as well as a mmc Jan 04 14:05:47 niv_one_three: ah.. then kernel part is needed rather Jan 04 14:05:49 so I need to tell the kernel to use GPIO xx for mmc and GPIO xx for BT Jan 04 14:06:20 as I am new, pls just point me in to where to seek the code Jan 04 14:06:45 plz , if you can Jan 04 14:08:21 how does the kernel get his info about the mtd-partitions on collie? Jan 04 14:08:53 niv_one_three: I cant Jan 04 14:09:02 hrw: ok thanks Jan 04 14:09:11 I will keep lookin Jan 04 14:10:48 hrw: one thing , is your new webpad running? do you actually took it to class? Jan 04 14:10:58 class? Jan 04 14:11:07 Uni Jan 04 14:11:18 niv_one_three: I'm 30 years old. Jan 04 14:11:20 so? Jan 04 14:11:21 ok Jan 04 14:11:32 did you try hand recognition? Jan 04 14:11:39 handwrite Jan 04 14:11:45 I finished studies few years ago Jan 04 14:11:46 niv_one_three: not yet Jan 04 14:11:50 oh ok Jan 04 14:19:37 mickeyl: i've got a patch so that bitbake correctly handles svn://user:pass@bla urls, interested? Jan 04 14:20:18 giel: submit it to bugtracker please Jan 04 14:20:38 sure Jan 04 14:20:40 giel: it would be nice improvment Jan 04 14:20:47 hrw|work: the oe bugtracker? because i couldn't really find a bitbake bugtracker Jan 04 14:20:59 oe bugtracker contains a severity for bitbake bugs Jan 04 14:21:03 okay Jan 04 14:21:34 mickeyl: #97 finally closed Jan 04 14:21:41 !oebugs 97 Jan 04 14:21:47 !oebug 97 Jan 04 14:21:49 * * Bug 97, Status: VERIFIED, Created: 2005-06-15 03:57 Jan 04 14:21:50 * * openembedded(AT)hrw.one.pl: BBINCLUDELOGS should have option to display only XX last lines Jan 04 14:21:51 * * http://bugs.openembedded.org/show_bug.cgi?id=97 Jan 04 14:21:54 aah, very cool Jan 04 14:22:07 RP, I'm now using usb_pxa27x_udc-r3.patch and I still get the DMA errors. Jan 04 14:31:13 hi mickeyl Jan 04 14:31:49 hey pb_. happy new year! hope you're fine Jan 04 14:32:07 btw, its Jan ----> where can one buy the fic cellphone? Jan 04 14:32:42 wouldn't be the first time something comes a bit later ... Jan 04 14:33:26 as long as it aint my girl - ok Jan 04 14:34:00 btw, what is the status of the open system for the fic? Jan 04 14:34:20 open system? Jan 04 14:35:11 mickeyl: how do you think - will you get something to show on fosdem or not? Jan 04 14:35:42 openmoko? Jan 04 14:35:47 4got the name.. Jan 04 14:35:54 hrw|work: given yesterday night's development, 100% sure on that Jan 04 14:36:28 hrw|work: a HTC universal gaffa taped into a smaller case :-) Jan 04 14:36:30 ah. as is the hardware, the openmoko system is under rapid development and will be released together with the hardware. Jan 04 14:36:42 FIC doesn't want to release any code earlier than hw Jan 04 14:37:18 XorA: who needs a case when you have gaffa tape? Jan 04 14:39:51 ~gaffa tape? Jan 04 14:40:15 Crofton: its like the force, dark on one side, light on the other and holds the universe together Jan 04 14:40:34 duck tape, or duct tape then :) Jan 04 14:40:57 Crofton: thats the stuff Jan 04 14:41:20 in germany it's called "tank tape (panzer tape)" Jan 04 14:41:25 You learn new "foreign" words every day :) Jan 04 14:41:55 well at least in an emergency in .de I can find the tape to rule them all :-) Jan 04 14:47:28 * koen reads http://www.iceni.org/cgi-bin/blosxom.cgi/Dude__Where_s_my_NetworkManager_. Jan 04 14:47:40 Good evening. Jan 04 14:48:02 .. and happy new year. Jan 04 14:48:02 hey sirfred Jan 04 14:49:04 hi sirfred Jan 04 14:49:08 HNY Jan 04 14:49:14 sirfred: HNY Jan 04 14:49:37 XorA: the bit about talking to wpa-supplicant over dbus is interesting Jan 04 14:50:22 I'm happy to announce you that I was working into a new kdrive imageon xserver. Jan 04 14:50:39 libw100 based? Jan 04 14:50:48 koen: No, standalone code. Jan 04 14:50:56 hey koen Jan 04 14:51:01 sirfred: that's great to hear Jan 04 14:51:05 hwy mreimer Jan 04 14:51:28 At this stage, it's supporting internal and external offscreen surfaces, copy and solid hardware operations, and an XV port. Jan 04 14:51:34 And hardware cursors. Jan 04 14:51:36 guys: what is needed for bitbake to REALLY start shell on failed patching or for devshell? Jan 04 14:51:36 s/hwy/hey/ Jan 04 14:51:51 koen|away: I was trying to decipher that. :-) Jan 04 14:51:53 on rhel machine it does not spawn shell.. Jan 04 14:51:58 sirfred: XV?? wow Jan 04 14:52:08 koen|away: was thinking you were saying "hny" as short for "happy new year" Jan 04 14:52:21 mreimer: no, just fat fingers Jan 04 14:52:22 hrw|work: I've only tested it with mplayer, but it seems to work , at least for common planar surfaces. Jan 04 14:53:04 Once I finish fighting against randr, I will release something. Jan 04 14:53:06 and thanks to ssvb he have some faster idct routines for armv5 as well Jan 04 14:53:16 that's great news Jan 04 14:53:23 s/he/we/ Jan 04 14:53:27 I would like to have randr working without the shadow framebuffer. Jan 04 14:53:41 sirfred: that would rock Jan 04 14:54:05 sirfred: have you done any disassembly of the wince drivers? Jan 04 14:54:28 I'd like to figure out how to get scaling working on w3220 Jan 04 14:55:03 mreimer: The problem with what I've seen in the wince drivers is that everything is done with the CP and packets. Jan 04 14:55:24 sirfred: sorry, I don't understand what that means Jan 04 14:55:39 sirfred: some kind of microcode? Jan 04 14:55:48 mreimer: The imageon series have a component they call CP, or PM4, or something so. Jan 04 14:55:50 mreimer: Yes. Jan 04 14:55:57 sirfred: ah Jan 04 14:56:22 mreimer: It's commonly used in ATI cards. You just feed the commands into a ringbuffer, and setup some pointer to feed them to the card. Jan 04 14:56:43 sirfred: sounds clever. does anyone have contacts with ATI that might be able to give clues? Jan 04 14:57:07 mreimer: I've tried sometimes to get specs without luck. Jan 04 14:57:23 mreimer: So, I'm using only direct register programming. Perhaps it's slower, but it seems to work. Jan 04 14:57:44 sirfred: it will still be way faster than no accel :-) Jan 04 14:57:45 * XorA seriously suspects ATI never designed the imageon chips Jan 04 14:58:19 sirfred: do you have any suggestions for how I might figure out how to get scaling working on w3220? Jan 04 14:58:54 mreimer: I suppose that the greater advantage of the ringbuffer approach is that you can fifo a lot of commands into it, while the internal cmdfifo has only 16 entries. Jan 04 14:59:19 sirfred: and it might allow for a sort of compression, sort of like using macros Jan 04 14:59:29 mreimer: I suspect that there are two registers that should be different: mmENG_CNTL (where we can setup rotated trajectories) and mmSRC_INC Jan 04 15:00:32 mreimer: Probably SRC_INC, where we can setup the scaling are different. The w100 one provides only 5 bits IIRC. Jan 04 15:00:59 sirfred: does mmENG_CNTL control rotation too? I'm seeing failures with hostcopyrot, ovrotate, ovstretch, ovyuvrotscale, rotblt, scaling, stretchblt, stretchblt2, and videotest Jan 04 15:01:49 mreimer: With any rotation configuration? Jan 04 15:01:58 sirfred: let me check Jan 04 15:02:45 mreimer: All the demos using overlay need to play with ENG_CNTL, because the overlay surfaces don't honour the screen orientation. Jan 04 15:03:05 mreimer: So, we need to scale + rotate to go from the source image to the destination surface. Jan 04 15:03:06 sirfred: AC_ROT0. Is that what you mean? Jan 04 15:03:11 firewalls suxx Jan 04 15:03:34 buildmachine -> firewall/router -> internet -> firewall/router -> me Jan 04 15:03:36 mreimer: Yes. With AC_ROT8 we shouldn't need to setup mmENG_CNTL. Jan 04 15:03:42 sirfred: this mean I need to enable x11 and xv support in mplayer :-) Jan 04 15:03:43 s/AC_ROT8/AC_ROT0/ Jan 04 15:03:51 XorA: Yes. :) Jan 04 15:04:09 XorA: Its' easy , as I was able to do it. Jan 04 15:04:20 sirfred: it'll upset the opieists Jan 04 15:04:25 XorA: thats why I do not like mplayer - no pluggable outputs Jan 04 15:04:45 XorA: mplayer is in the base opie install? Jan 04 15:05:00 mplayer will now depend on fontconfig, freetype, sdl, xv, x11, alsa Jan 04 15:05:03 How about totem? Jan 04 15:05:06 koen: no, but you know some of them just revert changes with x11 in the name :-) Jan 04 15:05:18 Or some of the xine players? Jan 04 15:05:40 Perhaps mplayer is not the best matching gpe. Jan 04 15:05:41 sirfred: xine + arm == small disaster Jan 04 15:05:49 gstreamer Jan 04 15:06:04 hrw|work: :( Jan 04 15:06:24 koen: Is gstreamer working on arm? Jan 04 15:06:41 sirfred: it was designed to run on arm :) Jan 04 15:06:42 so, yes Jan 04 15:06:49 the nokia770 is using it Jan 04 15:07:00 Great. Perhaps that's the way to go. Jan 04 15:07:03 we used to have some gst dudes in here Jan 04 15:07:19 I have a doubt about this work. Jan 04 15:07:55 gst support would certainly fit gpe well Jan 04 15:07:59 The think is that I had to hack kdrive itself, to be able to use the two video memory zones of the imageon. Jan 04 15:08:08 but I shall always love mplayer Jan 04 15:08:20 So, I have modified kdrive in a way that is not compatible with the existing drivers. Jan 04 15:08:20 s/gpe/x11 based images/ Jan 04 15:08:22 sirfred: s/hack/fix/ Jan 04 15:08:36 XorA: Well, I don't know if the kdrive people will love my changes. Jan 04 15:08:58 sirfred: then they should be kicked in the stones until they realise not everyone has a PCI card Jan 04 15:09:25 I've just changed KdScreenInfo , replacing the videomem_base and offscreen_base or something so with a pointer to a new VideoMemArea array. Jan 04 15:09:49 you can always send the patch to the xorg (xarch?) list for feedback Jan 04 15:09:54 The think is that I don't want to fix all the existing drivers. I've just changed fbdev and my imageon one. Jan 04 15:10:28 sirfred: tf@o-hand.com has done work on kdrive, so he should know who to harass for feedback Jan 04 15:11:16 koen: And if they don't like it, we could always live with the patches. Jan 04 15:11:24 indeed Jan 04 15:11:47 in normal linux style it might take a bit of hammering, but I think the concept is good Jan 04 15:12:10 its much more likely in embedded systems that non contiguose areas are used for screen Jan 04 15:12:41 RP: did chris write all this himself, or did he copy from OE without saying so? (http://svn.o-hand.com/view/poky?rev=1113&view=rev) Jan 04 15:16:26 koen, actually current optimized armv5 idct was developed by Mans Rullgard (mru from #ffmpeg channel), I only tried to improve it further Jan 04 15:21:08 XorA: An strange thing I'm suffering with my mplayer tests. Jan 04 15:21:40 XorA: It's giving me more than 200% time into audio_out Jan 04 15:21:49 XorA: So, I have to make the tests with audio disabled Jan 04 15:22:04 sirfred: on a c7x0? Jan 04 15:22:10 XorA: Yes. Jan 04 15:22:17 sirfred: Ill try and check that tonight Jan 04 15:22:21 sirfred: which kernel Jan 04 15:22:26 XorA: Let me see... Jan 04 15:22:39 2.6.18 Jan 04 15:22:44 openzaurus Jan 04 15:23:06 I see a lot of soc errors before starting, and most of times, no sound. Jan 04 15:23:24 sirfred: which asoc version? Jan 04 15:23:28 sirfred: what kind of audio decoder is used? Jan 04 15:23:46 XorA: How could I know that? Jan 04 15:23:54 ssvb: You mean, the audio-out mplayer module? Jan 04 15:23:59 sirfred: its in the dmesg when snd-soc-corgi is modprobed Jan 04 15:24:30 <6>ASoC version 0.12 Jan 04 15:24:30 <6>wm8731: WM8731 Audio Codec 0.12 Jan 04 15:24:30 <6>asoc: WM8731 <-> pxa2xx-i2s mapping ok Jan 04 15:24:34 XorA: You mean that? Jan 04 15:24:50 sirfred: earlier than that Jan 04 15:25:06 <6>ASoC version 0.13.0pre1 Jan 04 15:25:09 style Jan 04 15:25:17 sirfred: no, just decoder, if it uses floating point math, it will be extremely slow Jan 04 15:25:44 XorA: Isn't that the first line I pasted? Jan 04 15:25:57 sirfred: oh yeah, I cant read Jan 04 15:26:16 ssvb: How could I know that? Jan 04 15:26:18 sirfred: we are upto 0.12.6 Jan 04 15:26:37 XorA: So, I should upgrade my kernel? Jan 04 15:26:50 sirfred: default for 2.6.18 now is 12.4 Jan 04 15:26:55 sirfred: which should work Jan 04 15:27:16 XorA: OK, I will try to upgrade my kernel. Thanks. Jan 04 15:27:29 * XorA knows there was some breakage in earlier 12.X series Jan 04 15:27:48 sirfred: mplayer sgould display information anbout the decoder, if it is [ffmp3] or [mad], it is ok, but if it is [mp3], you are in a trouble Jan 04 15:28:02 ssvb: OK. let's see.. Jan 04 15:28:05 mp3lib should be disabled in the oe build Jan 04 15:28:55 XorA: I've used a modified oe bb to get the XV+X11 support. Jan 04 15:29:01 XorA: So, I suppose it's disabled. Jan 04 15:29:32 ~lart bloody audio Jan 04 15:29:32 * ibot executes killall -TERM bloody audio Jan 04 15:30:27 Selected audio codec: [ffvorbis] afm: ffmpeg (FFmpeg Vorbis decoder) Jan 04 15:30:32 another package dropped from OE Jan 04 15:30:35 Something like that? Jan 04 15:30:55 sirfred, ffvorbis uses floating point Jan 04 15:31:01 I'm using an ogm file to make the tests, with xvid video and vorbis audio. Jan 04 15:31:04 use tremor instead Jan 04 15:31:24 gah, when did they sneak in ffvorbis Jan 04 15:31:35 tremor works nice Jan 04 15:32:10 That means I need to recompile my mplayer to include tremor support, or is there any hope it's already included? Jan 04 15:32:10 they changed default to ffvorbis as it the best choice for x86 in 1.0rc1 Jan 04 15:32:56 sirfred, you will have to edit codecs configuration file and change ffvorbis and tremor priorities Jan 04 15:33:06 sirfred: tremor should be in there, might be called ivorbis Jan 04 15:33:21 I'm too mplayer newbie. Sorry. :( Jan 04 15:33:47 otherwise if both of them are compiled in, ffvorbis will be still used by default Jan 04 15:34:04 Any command line switch to change that behaviour? Jan 04 15:34:14 mplayer -ac tremor Jan 04 15:34:19 or whatever its called Jan 04 15:34:22 try '-afm tremor', but I'm not sure Jan 04 15:34:43 mplayer -ac help should give a list Jan 04 15:35:47 XorA: thanks, I was just looking at the list Jan 04 15:36:03 root@c7x0:/media/card/libw100/bin$ ./mplayer.sh -vo xv -ac help | grep vorbis Jan 04 15:36:04 ffvorbis ffmpeg working FFmpeg Vorbis decoder [vorbis] Jan 04 15:36:04 vorbis libvorbis working OggVorbis Audio Decoder [libvorbis] Jan 04 15:36:20 It seems it's not included Jan 04 15:36:55 sirfred: try tremor Jan 04 15:37:08 XorA: It doesn't work. Jan 04 15:37:16 XorA: Perhaps I've messed something while building mplayer. No problem. Jan 04 15:37:30 I don't need sound for these tests. I will go back into this later. Jan 04 15:37:35 sirfred: probably the configure script changed in random ways like it does Jan 04 15:37:52 And anyway, when mplayer starts, dmesg gets full of: Jan 04 15:38:02 <4>asoc: DAI[15:2] failed to match rate Jan 04 15:38:04 sirfred, tremor sources are included in mplayer source tree, it should be easy to configure mplayer with it, also you may check the latest revision of codecs.conf with ffvorbis entry removed: https://garage.maemo.org/plugins/scmsvn/viewcvs.php/trunk/etc/codecs.conf?root=mplayer&view=log Jan 04 15:38:26 sirfred: thats normal, its not a failure Jan 04 15:38:39 I also had this problem when building mplayer for nokia 770 :) Jan 04 15:39:01 XorA: OK. The word "failure" and the output into the kernel message fifo always makes me suspect. ;-) Jan 04 15:40:48 sirfred: its ok, 0.13 gets rid of that message/codfe section Jan 04 15:40:59 Nice Jan 04 15:41:51 03mwester 07org.oe.dev * r2c8930a5... 10/ (6 files in 3 dirs): madwifi-ng: add slugos support and bump to latest snapshot Jan 04 15:41:55 03hrw 07org.oe.dev * rbb994d04... 10/ (1 packages/tin/tin_1.7.3.bb packages/tin/tin_1.9.1.bb): tin: fix SRC_URI, unify Jan 04 15:41:59 03hrw 07org.oe.dev * r4cc808d3... 10/ (1 packages/tin/tin.inc): tin: added tin.inc Jan 04 15:42:04 03hrw 07org.oe.dev * rcf2bb56e... 10/ (5 files in 3 dirs): tin: dropped 1.7.3, added libpcre into DEPENDS Jan 04 15:49:02 koen: angstrom use full libx11? Jan 04 15:49:39 Checking for OggVorbis support ... yes (internal low accuracy Tremor) Jan 04 15:49:57 hrw|work: angstrom can use full x11 Jan 04 15:50:00 XorA: I'm going to see my configure.log... Jan 04 15:50:11 hrw|work: it's a distribution, you can install what you want :) Jan 04 15:50:32 koen: I meant: 'does angstrom build full libx11 by default'? Jan 04 15:50:49 Checking for OggVorbis support ... yes (internal low accuracy Tremor) Jan 04 15:51:28 hrw|work: check the angstrom-2007.1.conf Jan 04 15:51:29 maybe -ac vorbis is tremor Jan 04 15:53:01 XorA: You're right. With that flag, audio time goes down to less that 30% Jan 04 15:53:10 Without it it's more than 200% Jan 04 15:53:30 koen: distro set to diet-x11.. I got full-x11 anyway Jan 04 15:54:49 hrw|work: multithreaded butbake? Jan 04 15:54:53 koen: yes Jan 04 15:55:21 hrw|work: that's why I pasted libx11 into diet-x11 and mtn mv -e libx11_* ../nonworking/foo Jan 04 15:56:23 in other words you cheated? Jan 04 15:56:40 heh Jan 04 15:56:45 yes :D Jan 04 15:57:08 koen: so your builds are not official but from private branch Jan 04 15:57:22 org.openembedded.dev.koen ;d Jan 04 15:59:39 hrw|work: amusing definition of "official" :-) Jan 04 16:00:28 XorA: there is no way to get images in same way as koen gets them Jan 04 16:00:59 XorA: he got efika/angstrom builds which Crofton and I are unable to do Jan 04 16:01:17 now we got info that his builds has diet-x11 when my build got full-x11 Jan 04 16:01:20 etc Jan 04 16:03:16 hrw|work: ssh koen@angstrombuild.org password: angstom, bitbake angstrom-image :-) Jan 04 16:04:00 XorA: ssh koen@angstrom.ko.en Jan 04 16:04:22 hrw|work: BTW when is your wedding? Jan 04 16:04:47 how do i tell bitbake which kernel to build? Jan 04 16:05:06 * greentux greentux has experience in wedding planning (full last year :) -> can share with hrw @fosdem Jan 04 16:05:14 XorA: finally we got to decision to not plan wedding time Jan 04 16:05:36 rw: set it in machine config and then do 'bitbake virtual/kernel' Jan 04 16:07:08 works for collie too? Jan 04 16:07:14 rw: yes Jan 04 16:07:30 # following MACHINE types: poodle, tosa and simpad Jan 04 16:07:30 well, okay ,) Jan 04 16:07:47 hrw|work: Vegas for you? Jan 04 16:08:18 no ;) Jan 04 16:12:39 mickeyl: python-pyqt and qt4 are your stuff? Jan 04 16:12:49 yes Jan 04 16:13:12 mickeyl: python-pyqt4 4.0 is not available - need update to 4.1.1 Jan 04 16:13:16 Hi! Jan 04 16:13:23 qt4, sip4 probably too Jan 04 16:13:50 koen: Btw, 20061217 (or ~) angstrom snapshot for h4000 was broken (appears to be a780) Jan 04 16:15:14 sip4: ? Jan 04 16:15:24 hrw|work, I'll try excluding stuff that doesn't build for angstrom/efika Jan 04 16:15:35 ekiga for anstrong ?? Jan 04 16:15:56 Crofton__: most will be kernel related I think Jan 04 16:16:05 or twinkle , which is sip using qt... Jan 04 16:17:44 morning Jan 04 16:17:48 hi chouimat Jan 04 16:18:18 * chouimat is feeling masochistic today ... building OOo 2.1 from sources Jan 04 16:19:30 So BBMASK is one regex? Jan 04 16:19:42 Guess I need to remember hwo to do or's .. Jan 04 16:19:49 hrw|work: http://ewi546.ewi.utwente.nl/tmp/diff Jan 04 16:23:13 does angstrom-bootstrap-image build x11? Jan 04 16:23:44 it could Jan 04 16:23:59 that says nothing about whether it will install it Jan 04 16:24:11 I understand that Jan 04 16:25:18 BBMASK'ing madwifi is not helpful :) Jan 04 16:25:20 cu Jan 04 16:25:28 bye Jan 04 16:25:38 unresolved dependency Jan 04 16:26:08 why does angstrom want to build madwifi? Jan 04 16:26:33 MACHINE_FEATURES = "pci" Jan 04 16:27:27 <- professional D-C & Reset presser Jan 04 16:31:35 03koen 07org.oe.dev * rb7578129... 10/ (3 files in 2 dirs): poppler: drop old stuff Jan 04 16:42:41 * CosmicPenguin has a new years wish that somebody with better python skillz then he would help get Sugar into OE Jan 04 16:46:24 in arm-linux-g++ is there a way to align doubles on 8-byte boundaries, like how -malign-double works in the x86 compiler? Jan 04 17:19:09 I'm just curious if nobody is answering because nobody is listening, nobody knows, nobody cares, or if I am asking in the wrong place... Jan 04 17:20:26 hi Jan 04 17:22:09 ~lart one core machines Jan 04 17:22:09 * ibot --purges one core machines Jan 04 17:28:14 sbaker, that is the sort of question that may get an answer here, but you would be lucky if you did get an answer here ..... Jan 04 17:49:09 mickeyl: I updated patches for qt4 4.2.1 Jan 04 17:49:22 that's much appreciated. thank you Jan 04 17:49:32 i just have 0 time for OE for a couple of more weeks :/ Jan 04 17:50:03 i still need to find some time to plan our FOSDEM appearance though *sigh* Jan 04 17:50:07 mickeyl: get openmoko done and phone v1 released Jan 04 17:50:30 Anyone thinking of making some stuff to sell at FOSDEM? Jan 04 17:51:17 Crofton: could you ask on the mailing list? I'd love to see that happening, but I just can't take the wheel in organizing it. Perhaps someone volunteers for that Jan 04 17:51:25 ok Jan 04 17:51:27 thanks Jan 04 17:51:42 I would offer to be helpful, but since I am in US, figuring out details would be hard Jan 04 17:51:51 I do not want to import a box of stuff :) Jan 04 17:51:59 yeah, it'd be better if someone from EU handles that Jan 04 17:52:08 I fear cavity searches :) Jan 04 17:52:13 hehe Jan 04 17:52:27 and lost baggage Jan 04 18:02:23 shit.. updated patches but it try to use 'c' as compiler and 'o' as linker.. Jan 04 18:03:56 time to check 4.1.2 Jan 04 18:04:02 somebody can copy/paste their /etc/xinet.d/atftpd? Ubuntu Feisty has a broken post-install... Jan 04 18:06:15 ~praise RP for http://www.rpsys.net/wp/?p=11 Jan 04 18:06:21 All hail RP for http://www.rpsys.net/wp/?p=11! Jan 04 18:07:05 yes, good sleuthing RP! Jan 04 18:09:01 hrw: i like your 2006 timeline post Jan 04 18:09:10 RP: Excellent work. How were you able to debug this one? Jan 04 18:09:12 mickeyl: then comment there ;D Jan 04 18:09:58 mickeyl: it also showed me that writing on blog is good thing - I just looked into http://hrw.one.pl/2006/ pages ;D Jan 04 18:10:10 hehe, right Jan 04 18:10:13 * mickeyl comments Jan 04 18:10:38 ping RP Jan 04 18:11:12 mickeyl: autoaccepted Jan 04 18:11:15 so OE no longer supports Familiar? Jan 04 18:11:16 :) Jan 04 18:11:46 mccarthy: Familiar does not support OE and viceversa Jan 04 18:11:50 mccarthy: i'm afraid not. we try to continue support for the ipaqs though, so there is not that much of a loss :) Jan 04 18:12:04 hmm Jan 04 18:12:15 where the heck is all the bad blood coming from :( Jan 04 18:12:23 *nod* Jan 04 18:12:37 * mickeyl refuses to point to people not in the channel Jan 04 18:12:41 anyone want to suggest a more appropriate distro for my h5550? Jan 04 18:12:49 generic or angstrom Jan 04 18:12:52 mickeyl, :) Jan 04 18:13:15 can I do a full-on build with either? Jan 04 18:13:34 i think so, but in the absence of a h5550 device i never got a chance to test that Jan 04 18:13:35 shit. I lost win98 on my progear.. Jan 04 18:13:43 does the h5550 run kernel 2.6? Jan 04 18:13:47 if not, then generic may be a better choice Jan 04 18:13:50 mickeyl: badly Jan 04 18:13:57 mine does not run it Jan 04 18:14:04 hrw, darn Jan 04 18:14:51 mickeyl: mmp got everything minus sleeves working AFAIK on the h5xxx Jan 04 18:15:05 fair enough Jan 04 18:15:18 koen, so then it should work ok for me (I have no sleeves) Jan 04 18:15:43 you'd have to ask mmp for correct version to use Jan 04 18:15:45 ~seen mmp Jan 04 18:16:14 mmp was last seen on IRC in channel #handhelds.org, 32m 4s ago, saying: 'bye:)'. Jan 04 18:16:25 03hrw 07bitbake-1.6 * r729 10/doc/bitbake.1: added manpage from trunk Jan 04 18:17:32 koen: so wouldn't it be good to support it in Angstrom officially? Jan 04 18:17:57 iirc it is added into Angstrom queue already Jan 04 18:18:18 03hrw 07bitbake-1.6 * r730 10/AUTHORS: updated list of authors (from trunk) Jan 04 18:18:21 * koen looks at http://www.angstrom-distribution.org/unstable/images/20070101/h5000/ Jan 04 18:19:00 koen, cool. maybe I will try it Jan 04 18:19:01 :-) Jan 04 18:19:07 * florian_kc is too slow Jan 04 18:19:27 hi florian_kc Jan 04 18:19:39 florian_kc: too slow is monotone 0.32 not you Jan 04 18:19:55 I'm regenerating caches... Jan 04 18:21:02 hrw: there is no monotone release fast enough ;) Jan 04 18:22:19 or I have too slow machines? :D Jan 04 18:23:33 03hrw 07bitbake-1.6 * r731 10/lib/bb/fetch/svn.py: svn.py: updated (C) to me Jan 04 18:23:59 gods... I hate git fetcher Jan 04 18:24:06 and mvista git server Jan 04 18:24:33 need to make bitfetch multithreaded one day Jan 04 18:28:21 koen: no progear images this time? Jan 04 18:29:15 I'm doing another build run Jan 04 18:29:24 lets see if progear builds now Jan 04 18:30:22 I'm coping one of my images from oregon Jan 04 18:31:12 anyone tried angstrom for the OSK? Jan 04 18:32:09 03koen 07org.oe.dev * r61602e7d... 10/ (1 packages/minipredict/minipredict.bb): minipredict: fix packaging Jan 04 18:32:13 03koen 07org.oe.dev * r26a106d4... 10/ (1 packages/matchbox-common/matchbox_0.9.1.bb): matchbox-common: clean up metadata a bit Jan 04 18:32:19 03koen 07org.oe.dev * r0cd1f696... 10/ (1 packages/linux/linux-x86-2.6.17.9/guinness/defconfig): linux-x86: update guinness defconfig Jan 04 18:32:26 03koen 07org.oe.dev * r4a2d77dc... 10/ (1 packages/intltool/intltool_0.35.0.bb): intltool: also fix 0.35.0 RDEPENDS Jan 04 18:32:33 03koen 07org.oe.dev * re42b5f28... 10/ (1 packages/gstreamer/gst-plugins.inc): gst-plugins: add PKG_DYNAMIC for gst-plugins Jan 04 18:32:40 03koen 07org.oe.dev * r020324d3... 10/ (3 files in 2 dirs): gpe-scap: update to latest release Jan 04 18:32:46 03koen 07org.oe.dev * rfa43512e... 10/ (3 files in 3 dirs): gpe-login: update removeblue patch Jan 04 18:32:52 03koen 07org.oe.dev * red848ee6... 10/ (1 packages/gnome/gnome-keyring_0.4.4.bb): gnome-keyring: update DEPENDS Jan 04 18:32:59 03koen 07org.oe.dev * r6eaba88e... 10/ (1 packages/glibc/glibc-initial_2.5.bb): glibc-2.5 initial: hide the 2.3 inheritance a bit better Jan 04 18:33:07 03koen 07org.oe.dev * rf8334d10... 10/ (1 packages/fltk/fltk_1.1.4.bb): fltk: 1.1.4: also build gl plugins Jan 04 18:33:13 03koen 07org.oe.dev * rb60c5f93... 10/ (1 packages/gnome/gnome-menus_2.16.1.bb): gnome-menus: add 2.16.1 Jan 04 18:33:23 03koen 07org.oe.dev * rdf64b65d... 10/ (1 packages/gnome/gnome-panel_2.12.1.bb): gnome-panel: add 2.12.1 Jan 04 18:33:31 03koen 07org.oe.dev * r2f874ef2... 10/ (1 packages/gnome/gnome-desktop_2.16.1.bb): gnome-desktop: add 2.16.1 Jan 04 18:33:56 * koen whistles innocently Jan 04 18:34:20 lets boot Angstrom on progear (first time) Jan 04 18:34:45 koen: fltk was in bugtracker iirc Jan 04 18:35:05 hrw: I wouldn't know, this is a few months old :) Jan 04 18:35:50 RP: not all machine-via-env patches are in OE yet Jan 04 18:45:24 what time is it in Norway right now? Jan 04 18:46:11 somewhere around 8pm Jan 04 18:46:19 give or take an hour Jan 04 18:46:37 thanks Jan 04 18:47:05 http://www.worldtimeserver.com/current_time_in_NO.aspx Jan 04 18:47:05 some guy here is trying to help someone in Norway, I suspect they haven't develop a sense of time zones yet (the local) Jan 04 18:48:13 ~lart me Jan 04 18:48:14 * ibot crushes hrw with a full height scsi disk Jan 04 18:49:41 it is really 8:15PM there? Jan 04 18:50:06 Crofton: its 20:20 here so also in Norway Jan 04 18:50:14 yeah Jan 04 18:50:21 you guys work late :) Jan 04 18:50:37 is 8:18PM also in italy ... i think in norway or is the same or 7:xx Pm Jan 04 18:50:52 I need to reconfigure progear kernel.. Jan 04 18:52:46 thanks for the world time server link btw Jan 04 18:53:01 progear got colinux issue files... Jan 04 18:55:12 Is there an easy way to pull all kernel modules into the Angstrom (bootstrap) image? Jan 04 18:55:38 likewise: install 'kernel-modules' package Jan 04 18:55:48 hrw: thanks Jan 04 18:55:57 koen: doesnt the comment in glibc still read 2.3.x :-) Jan 04 18:56:29 XorA|gone: could be :) Jan 04 18:57:22 bbl Jan 04 19:00:44 koen: can Angstrom get task-all-needed-to-start-gpe? Jan 04 19:01:13 * mrz80 is away: Metaphysically I'm still here; physically, I'm elsewhere Jan 04 19:01:57 03Oliver 07org.oe.oz354x * r29548317... 10/ (4 files in 3 dirs): corsair: updated to 0.2.6 Jan 04 19:02:04 03Oliver 07org.oe.dev * r5e009cc3... 10/ (4 files in 3 dirs): corsair: updated to 0.2.6 Jan 04 19:02:08 03hrw 07org.oe.dev * rded62027... 10/ (1 packages/corsair/corsair_0.2.6.bb): corsair: s/DEPENDS/RDEPENDS (my fault) Jan 04 19:02:24 hrw: bitbake kernel-modules -c install gave me this: ERROR: No providers of build target kernel-modules (for []) Jan 04 19:02:49 hrw: should I (r)depend on kernel-modules instead? Jan 04 19:02:51 likewise: bitbake virtual/kernel Jan 04 19:03:09 and RDEPENDS=kernel-modules Jan 04 19:03:38 koen: why doesn't my approach work? Because its a metapkg? Jan 04 19:03:43 hrw: the plan is to move gpe-image to x11-base image which is gpe-login + gpe-conf + matchbox + rxvt Jan 04 19:04:00 likewise: yes, because it's a package instead of a recipe Jan 04 19:05:27 koen: ok, I should pay better attention, because often package=recipe. Jan 04 19:05:27 likewise: more answers after 'wie is de mol' :) Jan 04 19:05:33 :-) Jan 04 19:06:39 I need a minute away from this screen as well :-) Jan 04 19:07:30 how to tell ipkg to NOT add --passive-ftp for wget? Jan 04 19:08:04 what wireless kernel mods would I need for a h5550? Jan 04 19:12:41 kergoth: ping Jan 04 19:36:55 I wonder how good this new MIT Superyeast (http://www.machinedesign.com/ASP/viewSelectedArticle.asp?strArticleId=61770&strSite=MDSite&catId=0) whould be for beer.... Jan 04 19:42:39 hvontres|poodle, probably not so good. you don't want the yeast to go too fast - too much ester production. i may be in the minority on this next point, but you don't want it to get too alcoholic either. Jan 04 19:46:14 HopsNBarley: yeah, your probably right... but just think of the potential market for "fortified beer" :) Jan 04 19:47:05 we used to call that a "boilermaker" in college. yuck. Jan 04 19:48:30 Ångström boots on my progear in 1024x768 vesafb Jan 04 19:49:57 ~lart psplash Jan 04 19:49:57 * ibot teaches psplash the basics, including how to RTM Jan 04 19:52:00 RP: 2.6.19+git boots on poodle but I have two major problems I need to investigate some more. The sound modules complain about a bunch of unresolved symbols and the screen gets corrupted on resume. Jan 04 19:52:39 RP: I'll do some more testing tonight or tomorrow moring on the way to work to get some more detail. Jan 04 19:52:51 * mrz80 is back (gone 00:51:39) Jan 04 19:55:27 mccarthy: the atc modules, with a patch from mmp Jan 04 19:56:22 koen, thanks. now I just have to find them :) (too bad they weren't in that image) Jan 04 19:58:31 03pfalcon 07org.oe.dev * r8be6dc26... 10/ (1 conf/bitbake.conf): Jan 04 19:58:31 bitbake.conf: Add support for cpio and cpio.gz image formats. Jan 04 19:58:31 * Required to create initramfs bundles. Jan 04 19:58:31 * Single image can be used both as rootfs archive to expand to some media Jan 04 19:58:31 and as initrd (initramfs-style). In the latter case, "rdinit" kernel param Jan 04 19:58:32 should be used instead of "init". See Documentation/filesystems/ramfs-rootfs-initramfs.txt Jan 04 19:58:36 for more info. Jan 04 19:59:55 psplash suxx Jan 04 20:00:30 it scale good to XGA but works... badly Jan 04 20:01:22 koen: RDEPENDS on kernel-modules only gives me a small subset. Is this the result of the combined MACHINE/DISTRO feature set? Jan 04 20:01:44 koen: I am test driving efika sound, if it matters :-) Jan 04 20:17:12 03pfalcon 07org.oe.dev * r07e0c923... 10/ (1 conf/bitbake.conf): Jan 04 20:17:12 bitbake.conf: IMAGE_CMD_ext2.gz: Remove tmp.gz/ before creating it. Jan 04 20:17:12 * Sometimes due to errors ${DEPLOY_DIR_IMAGE}/tmp.gz is not deleted, and Jan 04 20:17:12 then next time it will error again on mkdir. Pretty confusing, especially Jan 04 20:17:12 fro newbies. So, just kill it to make sure no stale content is there. Jan 04 20:18:24 how to debug i2c? Jan 04 20:18:39 i2c_adapter i2c-0: Error: command never completed Jan 04 20:18:42 lots of them Jan 04 20:25:07 HopsNBarley: Hi Jan 04 20:25:13 hi likewise! Jan 04 20:25:22 HopsNBarley: everything cool? Jan 04 20:25:42 HopsNBarley: got some quotes for the MPC8313 processor today Jan 04 20:25:57 ooooh. as cheap as they said it would be? Jan 04 20:26:29 HopsNBarley: batch prices coming up, but it runs at 40% of the 8343 price here. Jan 04 20:27:13 that's fantastic. Jan 04 20:28:39 XorA|gone: I have randr working without using the shadow framebuffer. Suspend/resume also seems to work. Tomorrow I will try to setup the patches. Jan 04 20:28:53 sirfred: good work! Jan 04 20:29:14 mreimer: Thanks. I'm afraid that by the moment the Xv extension is not going to work for your card. :-( Jan 04 20:29:31 mreimer: We need to discover the problem with scaling & rotating. Jan 04 20:29:34 sirfred: it might work as long as no scaling or rotation are needed Jan 04 20:29:40 sirfred: yes, I want to figure that out Jan 04 20:29:50 hrw: pong Jan 04 20:30:31 mreimer: If your machine uses 0 rotation by default, it's one problem less. But scaling is used by the Xv extension if you, for example, want to play a video fullscreen. Jan 04 20:30:45 sirfred: :-( Jan 04 20:31:54 kergoth: can you change that DNS entry for OZ? Jan 04 20:31:57 so what pdas are people playing with nowadays? thinking i may get a new one and actually use it (*gasp*) Jan 04 20:32:00 mreimer: My theory is that the two cards are not completely register-compatible. We would need some register spec, but I don't know where to find it. Jan 04 20:32:37 kergoth: I just shutdowned 6 years old webpad Jan 04 20:33:25 Son of N770? Jan 04 20:33:27 hrw: okay, done. just need to wait for propogation Jan 04 20:33:43 how is the 770? do people actually use it for doing real things? Jan 04 20:34:02 not sure, but I am certain if you had one, it would become useful :) Jan 04 20:34:07 hehe Jan 04 20:34:09 I like the hw Jan 04 20:34:18 nice screen Jan 04 20:34:37 i mostly just want a portable device to manage my tasks. gtd like outline of things to do, reminders, etc Jan 04 20:34:46 if its perty and capable beyond that, that much the better :P Jan 04 20:34:53 yeah Jan 04 20:35:10 I'm not sure what SW is availble for the 770 to do that Jan 04 20:35:18 * kergoth hmms Jan 04 20:35:46 hey kergoth, did you do the php package? Jan 04 20:36:41 mreimer: For example, looking at the w100fb.h file, I see that the mmGRAPHIC_CTRL is different for the w100 and the w32xx, but there are no more differences registered here. Perhaps Ian Molton, the one that added support for the w32xx, has more info. Jan 04 20:36:51 wouldnt surprise me. been a while since i maintained any oe packages though, i'd have to poke at it Jan 04 20:36:54 why do you ask? Jan 04 20:37:20 i'm no expert with php, but there was something in the build I considered broken, and wanted to talk about it. Jan 04 20:37:30 no matter. Jan 04 20:38:10 kergoth: thx. Jan 04 20:38:23 hi Jan 04 20:38:36 mreimer: Also, I will give a look at the "Core Pocket Media Player" source code, as it contains some code specific for some imageon chips, and also a hardware IDCT I didn't had time to look at. Jan 04 20:38:49 HopsNBarley: speaking of php.. 4.3.6/4.3.10 which we have should be updated to 4.4.x or dropped Jan 04 20:39:21 i've got a 5.x building, but had to change the recipie a bit. was hoping to bounce it off somebody. Jan 04 20:42:41 kergoth: the 770 is the only device I haven't (re)flashed with an OE built thing Jan 04 20:42:50 Good night. Jan 04 20:42:55 'night sirfred Jan 04 20:43:00 koen, that is why he should get one :) Jan 04 20:44:23 I use mine ~weekly Jan 04 20:44:35 hmm, what do you use it for? Jan 04 20:44:35 more often when I go out photographing Jan 04 20:44:41 maemo-mapper rocks Jan 04 20:45:03 kergoth: mainly for maemo-mapper and gpe-calendar Jan 04 20:45:11 hrm Jan 04 20:45:15 how bulky is it? Jan 04 20:46:10 ah, theres the dimensions Jan 04 20:46:12 couldnt find em on the site Jan 04 20:47:53 * RP has a Poky image for his 770 :) Jan 04 20:48:19 :) Jan 04 20:48:26 http://cgi.ebay.com/Nokia-770-Internet-Tablet-w-64MB-Memory-Card_W0QQitemZ330070910259QQihZ014QQcategoryZ38331QQrdZ1QQcmdZViewItem Jan 04 20:48:30 this one may go cheap Jan 04 20:48:38 some screen issue Jan 04 20:48:45 others seem to go high Jan 04 20:49:24 the successor should be out soon according to rumours Jan 04 20:54:56 RP: I have working backlight driver for progear. It use pci_read/write config to get/set status and allow to set it by /proc/ entry. Jan 04 20:55:44 /proc? Jan 04 20:55:53 hrw: why not use a backlight class dev? Jan 04 20:55:54 RP: http://pastebin.ca/305551 - should be easy enough to port it to backlight subsystem? Jan 04 20:55:54 isn't /sys used for that nowadays? Jan 04 20:56:03 pH5: I did not wrote it Jan 04 20:57:56 hrw: ok. you are right, porting to backlight classdev shouldn't be hard. Jan 04 20:58:10 hrw: very easy Jan 04 20:58:56 so maybe I will write first part of code for kernel ;d Jan 04 20:59:24 Gerrath: I'm pleased you found the problem. Simple bugs like that can be hard to find. If you do solve the DMA problem, let me know as I'd like to get that driver fixed Jan 04 20:59:27 mreimer: did you every try using a LAB kernel in the kernel partition on the hx4700? Jan 04 20:59:42 mreimer: I'm getting a bit tired of having to search for a unused cf card :) Jan 04 20:59:51 koen: What problem did you see with machine-via-env? Jan 04 21:00:05 RP: fni-array Jan 04 21:00:12 aka binutils go boom Jan 04 21:01:12 koen: Did the script in work/binutils/temp/run* have an unset MACHINE in it? Jan 04 21:02:06 mickey|dvd: Sorry I wasn't around earlier. I can probably make what I have work but if there is a neater way to handle the main loop, I doubt I'll use it unless shown as my python isn't that good yet ;-) Jan 04 21:02:07 koen: no, I haven't tried that, but psokolovsky may have. He's been making commits to LAB lately. I've been dual-booting, using haret to boot from CF Jan 04 21:14:49 koen: My suggestion on angstrom-dev breaks bitbake btw :-/ Jan 04 21:15:13 RP: that's why I have buildscripts Jan 04 21:15:28 since package_write is a insane.bbclass thing, right? Jan 04 21:16:01 koen: no, its standard packaging Jan 04 21:18:27 * koen ponders about inheriting insane by default Jan 04 21:21:06 whats insane? Jan 04 21:21:16 kergoth: a checking bbclass Jan 04 21:21:32 kergoth: still requiring python2.4 and completely unfinished crap :} Jan 04 21:22:06 ah Jan 04 21:22:14 kergoth: http://www.openembedded.org/repo/org.openembedded.dev/classes/insane.bbclass Jan 04 21:22:14 kergoth: it will scan through .la, .pc files, check package contents for .dbg, check binaries for bogus/dangerous RPATH, and will check if the arch of the binaries match the one inside the package name Jan 04 21:22:37 kergoth: I think 75.34% of the above is varpoware Jan 04 21:22:55 zecke: Its a sound idea though :) Jan 04 21:23:23 RP: and replacing 'pass' with a true check makes it reality Jan 04 21:24:07 zecke: It'll get there Jan 04 21:24:42 RP: does backlight driver need to care about level < 0 > 100%? or just set max_brightness and ignore checking? Jan 04 21:25:11 hrw: The class does the checking for you iirc Jan 04 21:25:25 hrw: Have a look in backlight.c :) Jan 04 21:26:03 zecke: Any thoughts on stamps? I haven't had any :-/ Jan 04 21:26:20 RP: hehe, well I was swamped in assignments Jan 04 21:26:38 RP: as my sleeping schedule is totally messed up, get back to me in two hours (relaxing ATM) Jan 04 21:26:56 zecke: np, I just wondered. It can wait ;-) Jan 04 21:27:50 koen: You have me hooked on a different bitbake bug now. Looks like an off by one error in the depths of runQueue :-/ Jan 04 21:27:57 * koen wonders about bitbake-server + icecream Jan 04 21:28:12 RP: I hate off by one errors Jan 04 21:28:43 N \ {1} and off by ones are history Jan 04 21:29:00 * mwester knows of fried ice cream, but never bit-baked ice cream Jan 04 21:29:07 koen: me too :-/ Jan 04 21:30:59 koen: can you add fbset/i686 to feeds? Jan 04 21:31:27 RP: you know offhand if theres a nice outlining/gtd type app for maemo/n770? Jan 04 21:32:10 kergoth: only mallum's old figmet Jan 04 21:32:21 hehe, was afraid of that Jan 04 21:32:22 (which I use quite often...) Jan 04 21:32:35 certainly useful Jan 04 21:32:38 kergoth: See zecke :) Jan 04 21:32:49 even on OSX Jan 04 21:33:10 Its not off by one. Its invented a load of imaginary tasks and appended them! :} Jan 04 21:33:44 RP: I got first version of progear_bl built Jan 04 21:34:24 hrw: You're an official kernel hacker now? :) Jan 04 21:34:30 man, i got rehired by the company i worked at before ohand.. and the very next day i saw land warrior being canceled on slashdot.. Jan 04 21:34:37 dont suppose anyone knows of any jobs offhand.. :P Jan 04 21:35:01 :( Jan 04 21:35:08 kergoth: That is bad luck :-( Jan 04 21:35:18 getting back to digisys? Jan 04 21:35:21 kergoth: oops.. Jan 04 21:35:25 * kergoth shakes fist at hte government Jan 04 21:35:59 watching dice for jobs in phoenix, sick of moving. either need telework or local.. and no tech support, damnit Jan 04 21:36:02 :) Jan 04 21:37:04 kergoth: I can see you as a dogbert technical support type somehow :} Jan 04 21:37:25 ouch Jan 04 21:38:20 hrw: I can add fbset tomorrow after I have a x86 toolchain again Jan 04 21:38:30 * koen is building images from scratch tonight Jan 04 21:38:46 and I still need to put the images-from-feeds stuff in .dev Jan 04 21:38:50 * koen needs more time! Jan 04 21:38:59 koen: ok Jan 04 21:39:17 RP: actually, he's more like Buck in "Get Fuzzy". :-D Jan 04 21:40:22 if i didnt like money so much, i'd totally go into construction like office space Jan 04 21:40:49 office space is a nice movie Jan 04 21:41:04 http://www.comics.com/comics/getfuzzy/index.html Jan 04 21:41:07 do you have the required number of flair items? Jan 04 21:41:20 a-atwood: I was just looking at that :) Jan 04 21:41:46 argh.. why I selected so many modules... Jan 04 21:42:07 RP: good working finding that caching bug. I think I was probably plagued by that bug for a few months on h2200, back in 2003 or 2004 Jan 04 21:42:24 mreimer: wasn't that the ide issues? Jan 04 21:42:40 koen: no, it was random memory corruption Jan 04 21:42:44 * mreimer shudders Jan 04 21:43:01 kergoth: I keep wondering about being something like an electrician. People doing that kind of work make a fortune as there are few people qualified. I could probably get qualified without too much hasstle too :-/. Routine and boring though... Jan 04 21:43:31 mreimer: The unique thing about this one was it went away as soon as you have a context switch - the caches were flushed Jan 04 21:44:30 RP: hmm, maybe I was seeing something different. context switches seemed to be triggering my problem Jan 04 21:46:20 This bitbake failure raises a question. If a file says do_rootfs[rrecdepends] = "do_wibble", should bitbake error if not all tasks have a do_wibble task? Jan 04 21:47:02 no :} Jan 04 21:47:14 (just for convience, I know this is a bad answer) Jan 04 21:47:46 zecke: I agree but its less than ideal Jan 04 21:48:08 zecke: At the moment, it just breaks ;-) Jan 04 21:48:36 RP: make it error, and see what do_wibble is Jan 04 21:48:50 we can always add dumy tasks to base.bbclass but this will grow the dict and your runqueue Jan 04 21:48:59 maybe not your runqeue as you make up items anyway... Jan 04 21:49:18 zecke: I don't want to force thousands of empty tasks Jan 04 21:49:33 zecke: the problem case is do_deploy btw Jan 04 21:49:53 hmm, hmm Jan 04 21:52:08 PREFERRED_VERSION_linux-rp_hx2000 = "2.6.19+git" should work, right? Jan 04 21:52:35 or should I set DEFAULT_PREFERENCE_hx2000 = "1" in the .bb ? Jan 04 21:52:44 koen: Just set it in the .bb file Jan 04 21:53:42 oopsed Jan 04 21:58:44 Progear Backlight Driver Initialized. Jan 04 22:00:28 it works! Jan 04 22:00:46 hrw: yay! Jan 04 22:01:08 hello Jan 04 22:03:45 hrw: nice :) Jan 04 22:03:48 03koen 07org.oe.dev * r26b6f2ca... 10/ (1 packages/linux/linux-rp_2.6.19+git.bb): linux-rp 2.6.19+git: up default pref for hx2000 Jan 04 22:05:42 hrw: does the progear need Xorg or can it use kdrive? Jan 04 22:06:06 koen: if kdrive can use input drivers from xorg... Jan 04 22:06:30 it should Jan 04 22:06:35 koen: I need xorg-input-mutouch iirc Jan 04 22:07:01 mutouch exactly. Jan 04 22:07:13 no mt510 driver for tslib or kernel yet Jan 04 22:07:17 if it doesn't, harass keithp at breakfast in the astrid :) Jan 04 22:08:22 RP: http://pastebin.ca/305726 Jan 04 22:08:41 http://wiki.fosdem.org/DevRoom_Information Jan 04 22:08:52 I've been collecting FOSDEM info ... Jan 04 22:09:35 bbiab Jan 04 22:12:00 hrw: What happend to HW_LEVEL_MAX/MIN ? Jan 04 22:12:40 RP: ignore them Jan 04 22:12:54 hrw: You don't need to limit the range of values more? Jan 04 22:13:08 RP: will do - but not now Jan 04 22:13:16 hrw: ok, just wondered :) Jan 04 22:17:58 RP: brightness is always 0-100 values? Jan 04 22:23:42 meh, i want a gnome outliner with the ability to attach any arbitrary key/value pair metadata to each element in the outline, of which completion (yes/no, percentage) is one, and which also supports tagging (easily implemented as one of the aforementioned metadata pairs) Jan 04 22:23:46 heh Jan 04 22:24:51 bitbake-outliner ;) Jan 04 22:25:04 i should learn gtk Jan 04 22:25:13 03rpurdie * r732 10bitbake/lib/bb/taskdata.py: taskqueue.py: Add tasks to be queried but not created Jan 04 22:25:20 kergoth: mickey|dvd can give you advice on that :) Jan 04 22:26:51 03rpurdie * r733 10bitbake/lib/bb/runqueue.py: runqueue.py: Don't create recurisve tasks if they don't exist but still follow dependencies. Also allow multiple recrdeptask entries. Jan 04 22:26:59 hrw: No, any values you like Jan 04 22:27:28 hrw: In your case you probably just want to add HW_LEVEL_MIN to the intensity provided Jan 04 22:27:44 RP: yes. current version do this Jan 04 22:27:46 intensity = ((intensity * (HW_LEVEL_MAX - HW_LEVEL_MIN)) / 100) + HW_LEVEL_MIN; Jan 04 22:28:10 works nice Jan 04 22:28:34 isn't (HW_LEVEL_MAX - HW_LEVEL_MIN)) / 100 a constant? Jan 04 22:28:55 hrw: Too complex. Try intensity = intensity + HW_LEVELMIN and a max_brightness of HW_LEVEL_MAX-HW_LEVEL_MIN Jan 04 22:28:56 * koen figures gcc will optimize that Jan 04 22:29:02 sure Jan 04 22:30:14 RP: did you see my question about poky metadata this afternoon? Jan 04 22:30:30 koen: I did. I will ask chris to mention where things come from in future Jan 04 22:30:39 thanks Jan 04 22:31:18 * koen imagines cris replying with "no, I wrote them myself from scratch, including PR=r4" Jan 04 22:31:29 koen: bitbake will run do_deploy if we update image.bbclass as mentioned in that email now. The bad bit is anyone not using bitbake trunk head will have a broken bitbake (1.6 will be fine though) Jan 04 22:32:13 time for a bitbake release? Jan 04 22:32:21 koen: (HW_LEVEL_MAX - HW_LEVEL_MIN)) / 100 = 0.4 Jan 04 22:32:41 koen: fpu in kernel? Jan 04 22:32:50 fixedpoint math? Jan 04 22:34:40 RP: hx2000 image is on its way to http://www.angstrom-distribution.org/unstable/images/20070104/hx2000/ Jan 04 22:35:02 hrw: progear (and fbset) should be there next morning Jan 04 22:35:11 * koen goes to sleep now Jan 04 22:35:15 thx Jan 04 22:37:34 gnight Jan 04 22:42:28 RP: minimal level should be equal to 'nothing can be seen' or 'something is still visible'? Jan 04 22:43:07 hrw: 0 level means power off or as low as it will go Jan 04 22:44:12 hrw: I don't see why you would want to scale the values at all though as you shouldn't need to Jan 04 22:44:55 RP: intensity + MIN can be more then MAX.. Jan 04 22:45:21 hrw: It can't as intensity will never be more than max_brightness Jan 04 22:45:38 will check another Jan 04 22:46:01 hrw: Which kernel version is this? Jan 04 22:46:11 2.6.19-rc6 Jan 04 22:46:21 That should be fine Jan 04 22:47:41 RP: so userspace app check max_brightness and use 0->MAX as values? Jan 04 22:48:12 hrw: Yes and the kernel will return EINVAL to userspace if it exceeds max_brightness Jan 04 22:49:27 hrw: Look at backlight_store_brightness in drivers/video/backlight/backlight.c Jan 04 22:51:26 if I do not scale then brightness 0-40 works if I do intensity += HW_LEVEL_MIN or 80-119 if not Jan 04 22:51:53 I know where is a bug Jan 04 22:54:28 tekkken is nice :) Jan 04 22:54:30 RP: http://pastebin.ca/305838 Jan 04 22:55:44 RP: this version is OK Jan 04 22:56:41 time to sleep. tomorrow will send it to lkml Jan 04 22:57:01 cu Jan 04 22:57:22 * XorA|gone wishes his cpu would compile faster Jan 04 23:00:26 hrw|gone: Its not quite right. Read the actual_brightness property in sysfs and it will be wrong Jan 04 23:01:29 hrw|gone: Do pci_write_config_byte(pmu_dev, PMU_LPCR, intensity + HW_LEVEL_MIN); and remove intensity += HW_LEVEL_MIN;, then the brightness value stored isn't the hardware value Jan 04 23:29:54 RP: 2.6.19 still oopses on OOM for me Jan 04 23:30:52 RP: Kernel panic - not syncing: Attempted to kill init! Jan 04 23:32:19 * XorA|gone goes to sleep Jan 04 23:52:32 03koen 07org.oe.dev * r01a972a4... 10/ (1 packages/fbset/fbset_2.1.bb): Jan 04 23:52:32 fbset: stop making my eyes bleed with '############' and use the update-alternatives bbclass to update the alternatives Jan 04 23:52:32 ~lart coredump and his templates Jan 04 23:52:32 * ibot hauls coredump and his templates up by the scruff of the neck and spanks him until he waddles Jan 05 00:05:37 koen|away: hehe nice blog entry Jan 05 00:08:16 XorA|sleep: Gah :-( Jan 05 01:16:36 hmm defconfigman may address some of the headaches I am starting to have with omap kernels Jan 05 01:16:44 I hope ... Jan 05 01:19:16 Crofton, I hope too ;-). give it a try ;-) Jan 05 01:19:51 It'll will have to wait a few weeks Jan 05 01:20:06 I need to stop procrastinating and write my thesis also :) Jan 05 01:20:26 sure ;-) Jan 05 02:35:42 good nite Jan 05 02:41:39 Has anyone been able to build mysql-native? Jan 05 02:42:06 oz3541/dev Jan 05 02:49:58 ~seen Ifaistos Jan 05 02:50:03 ~seen likewise Jan 05 02:50:18 ifaistos was last seen on IRC in channel #oe, 7d 10h 10m 4s ago, saying: 'koen happy birthday :)'. Jan 05 02:50:23 likewise was last seen on IRC in channel #oe, 6h 23m 55s ago, saying: 'HopsNBarley: batch prices coming up, but it runs at 40% of the 8343 price here.'. Jan 05 02:54:02 wsueholz: you there? **** ENDING LOGGING AT Fri Jan 05 02:59:59 2007