**** BEGIN LOGGING AT Fri Oct 12 03:00:02 2012 Oct 12 05:25:23 http://www.edis.at/en/server/colocation/austria/raspberrypi/ Oct 12 08:36:02 kerio, hm that wizard mounter segfaults here Oct 12 09:12:03 wtf... why do we even vote? Oct 12 10:16:44 chem|st_: hey! because voting is democracy Oct 12 10:18:32 chem|st_: also becasue (I think) the council election rules are kinda flawed to step down to a 3 heads gremium when there are 5 candidates standing Oct 12 10:19:14 this seems kinda contra-productive and paradox Oct 12 10:45:56 DocScrutinizer51: having only 3 is BS already Oct 12 10:45:56 but who am I to complain... I dont care Oct 12 10:45:56 just in general Oct 12 11:06:22 freemangordon: ping Oct 12 11:39:15 vi_: pong Oct 12 13:02:41 ping freemangordon Oct 12 13:03:27 freemangordon, do we really need special code for IBE bit in uboot rx51.c? Oct 12 13:03:36 see http://git.denx.de/?p=u-boot.git;a=blob;f=arch/arm/cpu/armv7/omap3/board.c;h=9cee1d9b490692d02129d98470ba7b57b8e61ce2;hb=HEAD#l238 Oct 12 13:03:43 and http://git.denx.de/?p=u-boot.git;a=blob;f=arch/arm/cpu/armv7/omap3/board.c;h=9cee1d9b490692d02129d98470ba7b57b8e61ce2;hb=HEAD#l426 Oct 12 13:04:12 it already calling omap3_update_aux_cr_secure(0xE0, 0); Oct 12 13:39:33 Pali: ROM call parameters differ Oct 12 13:40:13 n900 wants 1st parameter to be the number of parapeters+1 Oct 12 13:40:30 bb want 1st to be the number of params Oct 12 13:40:34 *want Oct 12 13:40:39 *wants Oct 12 13:40:59 Pali: btw why bme should be disabled with 3.5.4? Oct 12 13:41:22 freemangordon, because it using maemo specific interface Oct 12 13:41:32 ok Oct 12 13:41:40 fast charger detection, ... Oct 12 13:41:49 Pali: BTW I got stock PVR driver running with 3.5.4 Oct 12 13:41:51 /dev/twl-madc... Oct 12 13:42:13 freemangordon, nice :-) with hildon opengles? Oct 12 13:42:15 Pali: but I did that too, otherwise mce does not start Oct 12 13:42:27 Pali: yes, with everything sock in userland Oct 12 13:42:31 *stock Oct 12 13:42:46 getbootstate wants it too, etc.. Oct 12 13:43:00 "/dev/twl4030-adc that is" Oct 12 13:43:17 we need to 1) remove/fix maemo specified apps or 2) port missing kernel interfaces... Oct 12 13:43:30 1 should not be needed Oct 12 13:43:51 freemangordon, can you write somewhere how you changed maemo to run upstream kernel? Oct 12 13:43:53 Pali: even cellular modes userland starts :P Oct 12 13:44:04 do you have omap ssi driver? Oct 12 13:44:14 (that is not upstreamed...) Oct 12 13:44:25 Pali: still not ready for that. And i've just realised I've loast my debian/patches directory :(:(:( Oct 12 13:44:30 Pali: yes Oct 12 13:45:03 but having PVR is the best thing Oct 12 13:45:28 freemangordon, use extundelete for ext2/3 fs Oct 12 13:45:32 i just need to tweak a bit led driver and a couple of others Oct 12 13:45:51 I reverted back more deleted files with extundelete Oct 12 13:46:06 but i have no clue how that happened Oct 12 13:46:26 add everything to git Oct 12 13:46:53 the only thing that comes to my mind is my damn fucking stupid windoze freezed so I had to reset the machine Oct 12 13:46:59 Pali: already did that Oct 12 13:47:11 now going to do git diff to split the patches Oct 12 13:47:25 now I'm playing with uboot... I rewritten rx51 relocation code Oct 12 13:47:45 Pali: so if we have /dev/twl4030-adc bme will run? Oct 12 13:47:50 but uboot not working if starting from other address (after relocation) Oct 12 13:48:16 freemangordon, somebody wrote that bme using some kernel interface which was deleted in 2.6.32 Oct 12 13:48:29 so maemo bme not running on >= 2.6.32 Oct 12 13:48:38 but I do not know if it is true or not Oct 12 13:49:22 freemangordon, it will really not run - bme depends on some sysfs interfaces on patched maemo musb driver Oct 12 13:49:42 Pali: /dev/twl.. is deffinitely missing in upstream, but anyway i added it. Oct 12 13:49:48 aah, ok Oct 12 13:49:50 makes sense Oct 12 13:49:54 is there a reason why bme should be kept around anyways? remove the old crud! Oct 12 13:50:05 Skry: no reaon Oct 12 13:50:09 *reason Oct 12 13:50:23 freemangordon, driver which export /dev/twlmadc on maemo was upstreamed, but without that /dev device Oct 12 13:50:39 anyway, kernel tree backup is finished, going to resore my patches Oct 12 13:50:42 *restore Oct 12 13:50:44 so you need to only add that missing /dev device into driver Oct 12 13:50:55 for twl?" Oct 12 13:51:03 twlmadcsomething Oct 12 13:51:07 it takes 5 lines of code Oct 12 13:51:16 ok, 15 :D Oct 12 13:51:31 * freemangordon is busy restoring the patches Oct 12 13:51:40 but really, I do not want to see bme :-) Oct 12 13:52:19 I RE getbootstate and it is possible to remove twlmadc code (and use rx51_battery driver) Oct 12 13:52:55 mce will be problem, but some open source version exists (+ there is also source code of diablo version) Oct 12 13:52:58 Pali: in upstream there is hwmon twl-madc, where all of the channels are exposed Oct 12 13:53:15 so we won;t need rx51_battery Oct 12 13:53:17 freemangordon, I forgot about hwmon driver :-) Oct 12 13:53:25 hehe Oct 12 13:54:14 freemangordon, we need to grep 'twl4030-adc' on full n900 rootfs image Oct 12 13:54:49 Pali: no need, I told you, add ing twl4030-adc takes 15 lines and no hacks Oct 12 13:54:58 ok Oct 12 13:55:26 will show you the patch later (when I have everything restored) Oct 12 13:56:23 freemangordon, what about create some fremantle kernel driver which export all needed interfaces? it is better to maintain separated driver (and not a lot of patches which needs to be fixed for all new kernel version...) Oct 12 13:56:50 hmm, why not Oct 12 13:57:02 makes sense Oct 12 13:57:07 bootreason, component_version, ... Oct 12 13:57:17 twl4030-adc Oct 12 13:57:17 yeah, you are right Oct 12 13:57:31 hmm, i'll do that Oct 12 13:58:31 and what did you do with omap gpio switch? Oct 12 13:58:40 NFC :D:D:D Oct 12 13:58:48 in upstream kernel is some other non compatible interface with that Oct 12 13:59:11 Pali: please, lemme restore my patches, i'll publish them Oct 12 13:59:17 ok :-) Oct 12 14:00:01 Skry: where were hsi and cmt patches? Oct 12 14:01:17 freemangordon, in my meego patch I disabled platform_device_register(&rx51_gpio_keys_device); (in board-rx51-peripherals.c) & enabled maemo gpio-switch.o driver Oct 12 14:01:32 Pali: I didn't do that Oct 12 14:01:44 ok, good Oct 12 14:02:06 i used Skry's kernel, ask him what has he did for gpios Oct 12 14:02:36 I will look at patches... Oct 12 14:04:43 now building new pre release kernel-power Oct 12 14:07:05 :D Oct 12 14:07:05 freemangordon, wait, does 3.5.4 boot Fremantle now? Oct 12 14:08:45 Hurrian: very -ish Oct 12 14:08:56 O.M.G. Oct 12 14:09:04 ...and that's all I can say about that. Oct 12 14:09:21 guessing frrom the chat logs, bme is the only major component that breaks (?) Oct 12 14:11:09 Hurrian: not really, a lot of stuff is nonworking now Oct 12 14:16:34 ...speaking of Maemo though, a massive part of preinit and rcS-late are unnecessary Oct 12 14:17:00 as in, completely replacable with reconfiguration of Fremantle. Oct 12 14:17:43 I'm not sure preinit is even necessary, it just very backwardly remounts / and chroots and boots. Oct 12 14:18:24 rcS-late may be excusable for entering into ACT_DEAD mode, and nothing else. Oct 12 14:19:03 Hurrian: sure, but I am doing the kernel stuff :P Oct 12 14:19:23 and TBH I am not that good in shell scripts ;) Oct 12 14:19:34 freemangordon, I know. Just thinking of some tweaks I'm making to Maemo to make it easier to boot on non-UBIFS root Oct 12 14:20:14 because with them, they really complicate everything, by a lot. Oct 12 14:20:38 I'd say ~80% of code in the "if" statements are unused on the N900 Oct 12 14:20:50 Hurrian: yeah, apart for bootmenu and ACT_DEAD, /sbin/preinit basically boils down to exec /sbin/init Oct 12 14:20:55 unfortunately, they carried over to harmattan too :( Oct 12 14:21:33 Hurrian, some months ago I started rewriting preinit Oct 12 14:21:38 but never finished... Oct 12 14:22:19 see https://code.launchpad.net/~pali/+junk/maemo_getbootstate Oct 12 14:23:10 Pali, wow, that's a lot more complicated than need be Oct 12 14:23:38 preinit ideally shouldn't exist - userspace component in upstart should catch bootreason and set runlevel accordingly Oct 12 14:24:06 u-boot bootmenu can provide bootmenu support, and with the right modules, /sbin/init can be executed directly Oct 12 14:24:54 Hurrian, then you need to move all preinit code into some upstart task Oct 12 14:26:18 Pali, that's the only drawback Oct 12 14:26:41 are you going to do it? :-) Oct 12 14:26:47 with a little fixup, default k-p cmdline can have init=/sbin/init ;) Oct 12 14:27:07 does NOLO pass a cmdline? Oct 12 14:27:07 I plan to write a minimal preinit that just mounts /mnt/new_root, and then check if maemo even needs that Oct 12 14:27:20 kerio, nope, it's built into kernel at compile time Oct 12 14:27:56 it /used/ to, on pre-N900 devices, or so I heard from the old-timers on maemo.org Oct 12 14:29:30 actually, if you read thru preinit, you used to be able to select between internal flash and sd boot, using the built-in bootmenu Oct 12 14:29:58 this feature is largely neutered on Fremantle and up, and it's vestiges only add complexity to the system Oct 12 14:31:25 kerio, nolo does not pass any cmdline Oct 12 14:32:17 Hurrian, in preinit you need to activate some cmt_bsi gpio Oct 12 14:32:45 and set runlevel Oct 12 14:32:45 Pali, AIUI though wrt getbootstate, preinit + getbootstate = fremantle getbootstate, right? Oct 12 14:33:14 has your harmattan-derived open getbootstate been pulled into CSSU yet? Oct 12 14:33:15 what? Oct 12 14:33:23 the getbootstate binary. Oct 12 14:33:32 Hurrian, not my getbootstate is not in cssu Oct 12 14:33:42 ah Oct 12 14:33:53 but my getbootstate (patched harmattan) should be compatible with maemo Oct 12 14:34:02 oh well, time to take a look at that too Oct 12 14:34:13 it should act same as fremantle binary Oct 12 14:35:13 btw, in my getbootstate there is some bsi number... now we know that it is RAW value of battery capacity Oct 12 14:35:31 and in kernel driver rx51_battery is conversation formula to mAh Oct 12 14:36:18 Pali, small steps to replace BME system? Oct 12 14:37:02 Hurrian, it should be possible now Oct 12 14:37:07 kernel drivers working Oct 12 14:37:10 hald-addon-bme too Oct 12 14:37:25 and libbmeipc should work (it was not tested yet) Oct 12 14:37:33 All that's needed is charge script to start charging, right? Oct 12 14:37:46 Hurrian, no Oct 12 14:37:47 Or does it charge automatically like stock system? Oct 12 14:37:50 only load kernel modules Oct 12 14:37:59 ah Oct 12 14:38:07 and kernel modules will automatically start charging when charger is plugged Oct 12 14:38:27 kernel-power v52 will be needed Oct 12 14:38:51 now I'm going to test my new build and then relase it as pre release Oct 12 14:39:06 Pali: i'm going to test it too Oct 12 14:39:26 I'll be waiting :). Goodnight, guys. Oct 12 14:40:42 bye Oct 12 14:57:47 kerio, you can: http://talk.maemo.org/showthread.php?p=1279757#post1279757 Oct 12 14:58:28 Pali: same URIs? Oct 12 14:58:33 yes Oct 12 15:00:18 Pali: btw, are the flashable kernel images in the kernel-power debs? Oct 12 15:00:25 the ones flashable with flasher-3.5 i mean Oct 12 15:01:01 kerio, fiasco image is in kernel-power and classic zImage is in kernel-power-bootimg Oct 12 15:01:32 you can flash both images by flasher-3.5 (with -F fiasco and with -k zImage) Oct 12 15:01:54 but before flashing you need to install modules package Oct 12 15:01:58 alright, installed Oct 12 15:02:20 Pali: i know, i wanted to know the way to recover from a botched kernel update :) Oct 12 15:02:42 Pali: it boots Oct 12 15:09:44 wait, when was 52 up? Oct 12 15:12:36 yoh Oct 12 15:22:24 sup DocScrutinizer Oct 12 15:24:10 now, i wonder Oct 12 15:24:18 am i a bad enough dude to install Pali's bme replacement? Oct 12 15:25:10 kerio, just stop bme & load modules bq2415x_charger and bq27x00_battery and rx51_battery Oct 12 15:25:25 it will charge battery Oct 12 15:25:43 and when you want to switch back to bme, just rmmod above modules and start bme Oct 12 15:27:37 Pali: i don't get the battery measurement though :) Oct 12 15:28:05 Pali: when does your replacement system shut down due to low battery? Oct 12 15:28:21 edv1? Oct 12 15:28:22 when you install hald-addon-bme Oct 12 15:28:28 like, edv1 set Oct 12 15:28:32 I tested it and working Oct 12 15:28:52 but I did not tested libbmeipc (used by pulseaudio & dsme for battery temperature) Oct 12 15:29:03 hm Oct 12 15:29:40 Pali: wait, rx51_battery was already loaded Oct 12 15:29:43 i think Oct 12 15:29:54 yes, rx51_battery is autoloading Oct 12 15:30:03 it read values via twlmadc driver Oct 12 15:30:17 and bme has no problem with it Oct 12 15:30:40 so you do not have to rmmod it Oct 12 15:30:43 i see Oct 12 15:30:57 Pali: solid led as soon as i attached the wallcharger Oct 12 15:31:00 how do i check what it's doing? Oct 12 15:31:20 cat /sys/class/power_supply/bq Oct 12 15:31:29 for example file uevent Oct 12 15:31:45 and check capacity or charge now Oct 12 15:32:19 Pali: wait, is uevent supposed to be sourceable from a shellscript? Oct 12 15:32:21 that's REALLY neat Oct 12 15:32:45 yes, you can cat uevent from shell script Oct 12 15:32:52 no, not catting it :) Oct 12 15:32:54 or you can cat your property file too Oct 12 15:32:56 i meant sourcing Oct 12 15:33:13 as in, . /sys/class/power_supply/bq24150a-0/uevent Oct 12 15:33:29 :-) Oct 12 15:33:30 to get POWER_SUPPLY_STATUS as a variable in your script Oct 12 15:33:44 kerio, ls /sys/class/power_supply/bq24150a-0/ Oct 12 15:33:59 aww, doesn't properly work because there are spaces in the last line :( Oct 12 15:35:07 POWER_SUPPLY_TEMP=326 Oct 12 15:35:11 is that in kelvin? :o Oct 12 15:35:45 Pali: POWER_SUPPLY_CHARGE_FULL_DESIGN=2056320 Oct 12 15:35:47 lolnope Oct 12 15:36:07 kerio, bq27200 battery report bad full design Oct 12 15:36:12 hm Oct 12 15:36:29 because this value is in eeprom (and not calibrated in factory) Oct 12 15:36:32 alright, rx51-battery is more sensible Oct 12 15:36:36 1308238 Oct 12 15:36:41 for design capacity is needed rx51-battery Oct 12 15:37:08 Pali: will h-e-n work with the new modules? Oct 12 15:37:52 kerio, h-e-n not Oct 12 15:38:00 you need to rmmod charger module Oct 12 15:38:30 or fix booston ;-) Oct 12 15:38:32 h-e-n using its own userspace "driver" for bq2415x Oct 12 15:39:48 ok screw this, i'm installing the bme replacement Oct 12 15:40:06 I'm tempted to do exactly that: augment booston so it works regardless of kernel bq24150.ko or bme is used Oct 12 15:40:39 ~closed_packages Oct 12 15:40:39 it has been said that closed_packages is https://wiki.maemo.org/Fremantle_closed_packages Oct 12 15:42:05 Pali: hrmpf, do you have some built packages? Oct 12 15:42:28 deb packages not Oct 12 15:42:36 ??? Oct 12 15:47:21 Pali: hmm, shouldn't it be charger_mode host if you are doing usb host with charging? Oct 12 15:47:39 unless you're providing 1A to the n900 *and* powering the usb device Oct 12 15:48:12 kerio, yes Oct 12 15:48:30 http://gitorious.org/rx51-bme-replacement/usbmode/blobs/master/usbmode.sh#line354 Oct 12 15:51:59 kerio, auto is too Oct 12 15:52:25 because fast charger detection is in host mode disabled (in kernel-power by patch) Oct 12 15:52:32 i see Oct 12 15:52:32 neat Oct 12 15:56:07 Pali: what does it mean for current_now to be -518k? Oct 12 15:56:17 charging at 518mA? Oct 12 15:56:26 yes Oct 12 15:56:39 is that normal? it feels... low :s Oct 12 15:56:52 it's at 70something percent Oct 12 16:12:01 Pali: usb_attached() is buggy Oct 12 16:12:43 DocScrutinizer05, do you have better implementation? Oct 12 16:13:23 well, for a beginning, delete /var/tmp/lsusb as first action in usb_attached() Oct 12 16:13:48 though I guess I could eventually come up with a better implementation, yes Oct 12 16:14:19 DocScrutinizer05, ok, if you can write better code, then I will not touch it yet Oct 12 16:16:06 can't you just do lsusb | grep -qv "1d6b:0002" Oct 12 16:16:31 he wants to kill it when it hangs Oct 12 16:16:53 does it do that often? Oct 12 16:17:03 :shrug: Oct 12 16:19:43 line 200 is ... weird Oct 12 16:37:00 Pali: looks decent, generally Oct 12 16:37:25 not even the usb_attached() is _really_ buggy Oct 12 16:40:30 DocScrutinizer's seal of "meh" Oct 12 16:41:06 indeed Oct 12 16:41:42 I'd write some stuff a lil different, but that only means I have differing preferences how shellscripts should look like Oct 12 16:42:03 phew, I managed to restore everything :) Oct 12 16:43:01 now uboot can turn off led :-) Oct 12 16:43:07 tested and working Oct 12 16:44:33 what I don't like is the fact that there's zero feedback about bq24150 boostmode errors to user Oct 12 16:45:21 yay, h-d is here :D:D:D Oct 12 16:45:57 DocScrutinizer05, I think that info about errors can go to hald-addon-bme Oct 12 16:46:23 because hald-addon-bme now parsing /sys/class/power_supply/ Oct 12 16:46:41 freemangordon :-) Oct 12 16:47:02 Pali: well, there's still nothing to inform user about the errors, no? Oct 12 16:47:12 hmm, it even tries to attach to wireless ;) Oct 12 16:47:19 DocScrutinizer05, dmesg Oct 12 16:47:27 Pali: there's even be the mce boostpattern happily going on Oct 12 16:47:33 and also some sysfs entry Oct 12 16:47:55 hm, I will look at mce pattern Oct 12 16:48:24 now hald-addon-bme handle charging pattern, so adding another should not be problem Oct 12 16:48:48 somebody needs to tear down pattern_boost and maybe rise a notifier when e.g. boostmode stumbles and dies on too high load Oct 12 16:49:06 :nod: Oct 12 16:49:25 Pali: brilliant idea Oct 12 16:50:27 DocScrutinizer05, in bq2415 sysfs there are *_status entires Oct 12 16:51:01 thought as much - just nobody reading less watching them Oct 12 16:51:56 kernel aka bq2415x.ko should emit a kevent to let everybody know there occured an exception Oct 12 17:08:42 DocScrutinizer05, now sysfs entry of bq2415x show auto or off (if autoreseting timer is enabled) Oct 12 17:09:26 and if error occur, then it print error message (instead auto/off) Oct 12 17:09:36 what do you think? Oct 12 17:10:18 if you change timer mode, then error will be cleared Oct 12 17:10:20 Pali: i thought there was a _fault entry already Oct 12 17:10:26 hehe, keyboard leds work :P Oct 12 17:10:38 kerio, yes there are _status and fault entries Oct 12 17:10:54 but bq2415x driver must every ?s reset timer Oct 12 17:11:20 and when resetting timer it reading error status and report error message to dmesg Oct 12 17:11:26 well, that's an internal thing **** ENDING LOGGING AT Fri Oct 12 17:18:10 2012 **** BEGIN LOGGING AT Fri Oct 12 17:20:22 2012 Oct 12 17:22:12 Pali: and while errors in auto-charge mode might recover with a little help from kernel driver (or even without any, in some cases), for boostmode it definitely needs userland interaction to clear an error Oct 12 17:23:42 Pali: ideally the kernel driver emits a signal (or does other event which a process could wait() for) when an error occurs Oct 12 17:24:59 DocScrutinizer05, now I patched kernel driver. it will call sysfs_notify on error and cat timer (from userspace will show error message) Oct 12 17:25:19 auto reseting timer then is disabled Oct 12 17:25:24 cool Oct 12 17:25:41 and userspace need to echo off/auto > timer to clear error message Oct 12 17:25:59 :-) Oct 12 17:29:44 DocScrutinizer05, see: https://gitorious.org/rx51-bme-replacement/bq2415x_charger/commit/b81b6e499d1e283853eaf756577bee1d74272f19 Oct 12 17:30:40 ok, mce works, enough for today :D Oct 12 17:31:12 freemangordon, can you push git repo? Oct 12 17:32:03 Pali, not now. And I'll upload the pathches(against vanilla) to merlins server Oct 12 17:32:12 ok Oct 12 17:34:51 Pali: hmm, your charging module seems a bit eager to start charging again, or something like that Oct 12 17:35:05 or maybe hald-addon-bme cheats Oct 12 17:35:10 (the stock one) Oct 12 18:04:33 Pali: ok, now i'll stop pestering you for a new kp52 build and i'll start pestering you about a bme-rx51-replacement build Oct 12 18:04:35 is that ok? Oct 12 18:08:13 kerio, wizard mounter works ok for you? Here it segfaults Oct 12 18:08:24 grammoboy: idk, never tried it Oct 12 18:08:31 kerio, k Oct 12 18:09:04 Pali: https://gitorious.org/rx51-bme-replacement/bq2415x_charger/blobs/b81b6e499d1e283853eaf756577bee1d74272f19/bq2415x_charger.c#line607 ERRRRR Oct 12 18:09:29 Pali: https://gitorious.org/rx51-bme-replacement/bq2415x_charger/blobs/b81b6e499d1e283853eaf756577bee1d74272f19/bq2415x_charger.c#line603 is odd, I think Oct 12 18:10:21 >>V(chg)=Voffset_3.5 + bit2_20mV + bit3_40mV ..... + bit7_640mV<< or maybe not Oct 12 18:12:11 Pali: anyway - http://maemo.cloud-7.de/bq24150-sysnode.spec.txt ->"CV r/w 2[7:2]"... Oct 12 18:13:19 Pali: >>... That is why the driver does refuse settings to >4240mV without special sequence. See sourcecode.<< Oct 12 18:14:01 Pali: I thought of something like writing value twice in sequence, but second value ++ (aka +1) Oct 12 18:14:51 I.E. echo 4380 > CV; echo 4381 > CV Oct 12 18:16:08 any other write access to any of the bq2415x sysfs nodes in between the two voids the "escalation" sequence Oct 12 18:17:12 now lemme try my math on 94 Oct 12 18:17:32 what about an internal "leet mode" where you can do weird stuff like that? Oct 12 18:18:03 like, a safe_mode that you can echo 0 to to disable those sanity checks Oct 12 18:19:02 Pali: how about testing mV around line 602,603, rather than val in that location. Way more legible Oct 12 18:20:45 Pali: if (mV > (4440)) return -EINVAL; Oct 12 18:22:16 Pali: if (mV > (4240)) {wait_for_special_escalation_sequence = true;} Oct 12 18:23:09 Pali: if (mV > (4240)) {wait_for_special_escalation_sequence = true; first_escalation_mV = mV} Oct 12 18:26:08 why 4240? Oct 12 18:27:15 Pali: if (mV > 4240) { if (!wait_for_special_escalation_sequence) {wait_for_special_escalation_sequence = true; first_escalation_mV = mV} else {wait_for_special_escalation_sequence = false; if (mV <> --first_escalation_mV) return -EINVAL}} Oct 12 18:27:49 ShadowJK: I thought it's a nice value and still safe, just slightly above 4200 Oct 12 18:29:36 who's finding any syntax or typo errors may keep it Oct 12 18:30:47 on a second thought, define first_escalation_mV==0 as !wait_for_special_escalation_sequence Oct 12 18:33:19 if(mv > 4240 && lastmv-1 != mv) { lastmv=mv; mv=4240; } Oct 12 18:33:20 NO second thoughts for you! ;) Oct 12 18:33:39 Pali: if (mV > 4240) { if (first_escalation_mV==0) {first_escalation_mV = mV - 1} else { if (mV <> first_escalation_mV) first_escalation_mV=0; return -EINVAL}} Oct 12 18:33:49 Woody14619: :-D Oct 12 18:34:03 LOL... Couldn't resist. Oct 12 18:34:28 Woody14619: in THAT case this never been any second thought Oct 12 18:35:34 I actually never changed my mind regarding standing for council (well modulo the short time when I felt extremely pissed by all those who flamed me) Oct 12 18:35:41 Oh, it's ok... apparently now I'm a power-grabbing person, trying to take over everything, don't-cha-know... Oct 12 18:35:54 suuuuure Oct 12 18:36:01 shame on you ;-P Oct 12 18:36:25 You know, for following election rules... and "interpreting" them the same way most sane people would... Oct 12 18:36:40 Woody14619: again many thanks for the cochones Oct 12 18:37:34 yeah, I've seen it as live as possible, I *felt* it Oct 12 18:38:19 there's actually a flaw in rules though, which now caused a board of three election when actually 5 are running Oct 12 18:38:28 lol. :) Well, all is in motion now, so... Oct 12 18:38:34 I don't see that as a flaw really. Oct 12 18:38:44 The intent is to have a selection of people vs appointment. Oct 12 18:38:56 debatable Oct 12 18:39:29 I'd think on exactly 5 candidates the nomination should get extended, but still trying to get 5 Oct 12 18:39:37 or more than 5 actually Oct 12 18:40:08 then *after* extended nomination you go with 5 w/o voting Oct 12 18:40:29 my take on spirit of rules Oct 12 18:40:35 maybe that's just me Oct 12 18:41:10 the rules are flawed anyway, as they are not clear regarding those "niche" cases Oct 12 18:41:26 nobody anticipated we'd *ever* run into that situation Oct 12 18:42:10 no, it's actually quite clear.... If there are 5 or fewer, the election is for 3. Oct 12 18:42:19 If there are more than that, the election is for 5. Oct 12 18:42:40 whatever, would you occasionally add me in on http://wiki.maemo.org/Community_Council/Council_election_September_2012/Candidate_declarations nagain? Oct 12 18:43:08 you might be right, but it feels paradox Oct 12 18:43:57 It's also very clear that if there are fewer than 3 after nomination period closes that nomination extend. Oct 12 18:44:54 The problem is that the wording should be either "before elections begin", or spell out clearly what happens if a nomination is pulled at any time, but especially in the window between nomination close and election start. Oct 12 18:45:31 Woody14619: and would you send me a mail from your "private" addr, so I could contact you after doom hits me (which God forbid), as I think there's probably little to no help from any other of the current council members for me Oct 12 18:46:13 We'll be adding such to the bylaws in the next revision, as requested in the tmo thread. Oct 12 18:46:29 ooh, missed that Oct 12 18:46:41 I will pm you on tmo. No worries... I'm not going anywhere. Oct 12 18:47:10 * DocScrutinizer05 afk for dinner Oct 12 18:47:17 WEEEEEEEEEEEEEKEEEEEEEEEEEEEND Oct 12 18:48:46 damn, 3 thirds of https://gitorious.org/rx51-bme-replacement/bq2415x_charger/blobs/b81b6e499d1e283853eaf756577bee1d74272f19/bq2415x_charger.c#line603 left over for auditing/review Oct 12 18:48:54 later.... ;-P Oct 12 18:50:27 ShadowJK: please review https://gitorious.org/rx51-bme-replacement/bq2415x_charger/blobs/b81b6e499d1e283853eaf756577bee1d74272f19/bq2415x_charger.c#line525 (only 5 or 6 lines!) Oct 12 18:52:20 https://gitorious.org/rx51-bme-replacement/bq2415x_charger/blobs/b81b6e499d1e283853eaf756577bee1d74272f19/bq2415x_charger.c#line632 \o/ kudos Pali Oct 12 18:54:11 >> I(term)= (offset_3.4 + bit0_3.4 + bit1_6.8 + bit2_13.6) / R(SNS) [mA] << Oct 12 18:54:58 >> Default for R(SNS)=0.068, other can be set values by module parameter. I(chg)= (offset_37.4 + bit4_6.8 + bit5_13.6 + bit6_27.2) / R(SNS) [mA] << Oct 12 18:57:12 * DocScrutinizer05 now idly wonders if bq2415x.ko will take parameter like Rsns=0.08 Oct 12 18:57:19 for bq->init_data.resistor_sense; Oct 12 18:57:47 modprobe bq2415x.ko Rsns=0.08 Oct 12 19:02:55 Pali: ^^^ Oct 12 19:03:56 Pali: don't miss previous highlights, I suggested a better solution for https://gitorious.org/rx51-bme-replacement/bq2415x_charger/blobs/b81b6e499d1e283853eaf756577bee1d74272f19/bq2415x_charger.c#line607 Oct 12 19:04:38 Wow really? Oct 12 19:05:03 hmm? Oct 12 19:05:05 My next stop is the "Bob Hope Internaltional Airport". They named an airport after Bob Hope? How random. Oct 12 19:05:48 what? alice? WTF is... bob hope? Oct 12 19:06:43 He was an actor... many many years ago. Oct 12 19:07:33 http://en.wikipedia.org/wiki/Bob_Hope Oct 12 19:10:28 Ah... Burbank... Bleh.. I been there before. It was just Burbank airport back then... Wonder when it changed names. Oct 12 19:10:43 Anyway.. off I go. Have a nice weekend all. Oct 12 19:10:46 What sounds better, Burbank or Bob Hope Oct 12 19:10:56 point... ;) Oct 12 19:11:09 ;) have a nice weekend Oct 12 19:41:26 Pali: by the way, do you reckon the kernel can be flashed with nandwrite, or do we still need some kind of specific flasher, for that? Oct 12 19:41:30 (i mean on-device flashing) Oct 12 19:42:34 kerio: depends on format of image you wanna flash Oct 12 19:43:56 if nanddump produces a binary identical to your fiasco excerpt, then why should nandwrite fail? Oct 12 19:44:13 idk, error correction on blocks that have become bad Oct 12 19:44:42 nandwrite knows that too Oct 12 19:45:02 DocScrutinizer51: then why does kernel-flasher not use nandwrite? Oct 12 19:46:18 I suspect there's some fancy header on fiasco stff Oct 12 19:46:24 stuff* Oct 12 19:46:41 maybe even checksums Oct 12 19:46:56 ask Pali Oct 12 19:47:11 that's what i did! Oct 12 19:47:24 he's the 0xffff savvy Oct 12 19:47:57 kerio, flashing on n900 is done by softupd Oct 12 19:48:18 and it fake raw zImage, it append some header and flash it Oct 12 19:48:31 we need to RE that header & it should work Oct 12 19:48:37 thought as much Oct 12 19:48:57 Pali: what about nandwriting an image that was nanddumped from the device itself? Oct 12 19:49:08 I think it should work Oct 12 19:49:12 but never tested Oct 12 19:49:23 how about simply stripping fancy header? Oct 12 19:50:36 what became of nandtester? Oct 12 19:51:12 he ever dared to flash my CAL image? Oct 12 19:51:16 he got pretty much everything back working, afaik Oct 12 19:51:32 but idk if he used your CAL image Oct 12 19:51:35 i've got mine, now :3 Oct 12 19:53:19 i seriously wonder why nokia decided to use such an half-assed format to store configurations Oct 12 20:09:00 like CAL? Oct 12 20:09:01 legacy Oct 12 20:09:37 same reason, why we have 2MB of nand unused Oct 12 20:09:49 all the el-weirdo hw calib like for ALS is inCAL, so phenix knows how to handle it Oct 12 20:09:56 we could have +2MB on rootfs... Oct 12 20:10:08 yup Oct 12 20:10:36 or find a hackery use for the initfs partiton Oct 12 20:10:58 we can store here uboot :-) Oct 12 20:11:03 or kernel Oct 12 20:15:59 is there any other phone is awesome as the n900? Oct 12 20:16:05 dafaq, my N900 stole my posts Oct 12 20:16:37 I suggested that like 2 years ago Oct 12 20:17:39 are there proprietary blobs apart for BME that require cal? Oct 12 20:18:41 I even asked like 23 times if N900 nolo still tries to load initrd partition Oct 12 20:18:46 a lot Oct 12 20:18:59 there's bluetooth-sysinfo but iirc it has an open replacement Oct 12 20:20:07 there's quite a number of settings around NOLO, bme, dsme(?), cmt, BT, WLAN, in CAL Oct 12 20:20:34 BT and WLAN are open, bme is replaceable, and i thought cmt had *its own* CAL Oct 12 20:22:05 strings /dev/mtd1|less Oct 12 20:23:47 ConF Oct 12 20:23:49 lock_code Oct 12 20:23:50 12345 Oct 12 20:23:52 ;-P Oct 12 20:24:03 heh Oct 12 20:24:38 content-ver Oct 12 20:24:40 ({w!lA Oct 12 20:24:41 wRX-51_2009SE_1.2009.41-1.UK-IRELAND Oct 12 20:25:18 whaaaaaaaaaaaaat Oct 12 20:25:18 sw-release-ver Oct 12 20:25:20 RX-51_2009SE_1.2009.44-1_PR_MR0 Oct 12 20:25:28 you don't have the global one installed? D: Oct 12 20:25:46 ConF Oct 12 20:25:47 kernel-ver Oct 12 20:25:49 MD2.6.28-20093204eno3+0m5 Oct 12 20:25:55 dude, spam Oct 12 20:27:46 honestly, do strings /dev/mtd1ro|less and wonder Oct 12 20:28:21 ConF Oct 12 20:28:23 kernel-ver Oct 12 20:28:24 MD2.6.28-20093204eno3+0m5 Oct 12 20:29:22 keep in mind CAL usually doesn't delete or overwrite old records Oct 12 20:30:22 with >> strings /dev/mtd1ro|less << it's pretty hard to tell what's valid and what's obsolete Oct 12 20:32:27 IroN900:~# strings /dev/mtd1|grep RX-51_2009SE|wc -l Oct 12 20:32:28 19 Oct 12 20:33:20 IroN900:~# strings /dev/mtd1|grep ConF|wc -l Oct 12 20:33:21 63 Oct 12 20:34:59 ~spam kerio Oct 12 20:34:59 * infobot subscribes kerio to every possible mailing list he can find on the 'net AND arranges for kerio to get at least 50 snail mail spams a day Oct 12 20:34:59 you have a metal n900? :p Oct 12 20:35:57 yes, pure IroN Oct 12 20:36:49 the other one is way more exciting though: t900 Oct 12 20:37:07 he probably got one of estel's aluminium body replacements >:D Oct 12 20:37:15 I'm still working on the red glowing eyes Oct 12 20:38:50 ut hey, I'm known as skynet, so what else I could name my coprocessors Oct 12 20:39:48 funny sidenote: I'm yawning Oct 12 20:40:11 ~botsnack Oct 12 20:40:11 DocScrutinizer05: aw, gee Oct 12 20:40:49 thanks purl Oct 12 20:40:49 DocScrutinizer05: gern geschehen Oct 12 20:41:22 meh! Oct 12 20:41:24 ~die Oct 12 20:41:44 WAT Oct 12 20:41:51 no you didn't Oct 12 20:42:02 oh no you didn't Oct 12 20:42:25 DocScrutinizer: infobot joined! Oct 12 20:42:50 ~+uptime Oct 12 20:42:50 - Uptime for purl - Oct 12 20:42:50 Now: 1m 5s running infobot 1.5.4 (SVN) -- linux Oct 12 20:42:51 1: 59d 8h 41m 19s running infobot 1.5.4 (SVN) -- linux, ended Sun Nov 14 18:39:57 2010 Oct 12 20:42:51 2: 57d 3h 9m 23s running infobot 1.5.4 (SVN) -- linux, ended Fri Jun 26 20:39:27 2009 Oct 12 20:42:52 3: 36d 20h 47m 14s running infobot 1.5.4 (SVN) -- linux, ended Tue Aug 4 17:38:59 2009 Oct 12 20:42:59 dang, still purl Oct 12 20:43:13 ~sorry Oct 12 20:43:13 * infobot appologizes profusely to everyone, points at the spammers, and says: "IT WAS THEM, THEY MADE ME DO IT!!!" Oct 12 20:48:25 I'm really missing ~die like usual mortals enjoy it Oct 12 20:48:33 ~literal die Oct 12 20:48:33 "die" is " takes two shots to the head and crumples to the ground, lifeless." Oct 12 20:48:53 DocScrutinizer05, CAL is linked list structure Oct 12 20:49:01 yeah I know Oct 12 20:49:12 hard to read for humans Oct 12 20:49:29 DocScrutinizer05, now I have open source libcal compatible library for reading CAL Oct 12 20:49:38 #I know Oct 12 20:49:46 awesome achievement Oct 12 20:50:02 make that CALedit Oct 12 20:50:24 DocScrutinizer05, for edit somebody with hexrays should use IDA Oct 12 20:50:32 in SDK is x86 binary too Oct 12 20:50:37 also give humaity a proper kevent-mon Oct 12 20:50:41 cmdline tool Oct 12 20:51:05 and when I looked into objdump of arm and x86 binary, seems that both are same Oct 12 20:51:19 nokia for unknown reason compiled also x86 binary of libcal... Oct 12 20:51:22 to SDK Oct 12 20:51:36 autobuilder does that Oct 12 20:52:26 ping merlin1991 Oct 12 20:52:28 you bet Nokia has a cloded autobuilder, they can feed src blobs Oct 12 20:52:34 ping freemangordon Oct 12 20:52:36 closed* Oct 12 20:53:03 ping Sicelo Oct 12 20:53:32 I'd be more interested in a proper kevent-adm monitor, alike udevadm monitor Oct 12 20:53:43 onl ymarginally twekaed Oct 12 20:53:57 only marginally tweked Oct 12 20:54:03 dafaq Oct 12 20:54:10 is there anybody who can test new uboot version? Oct 12 20:54:14 U know what I meant Oct 12 20:54:26 2101 Oct 12 20:54:32 :-/ Oct 12 20:54:38 4 devices Oct 12 20:54:41 all 2101 Oct 12 20:55:03 bad, on my 2101 is uboot working Oct 12 20:55:36 DocScrutinizer05, btw for loading kernel image I'm using now only 0xFFFF :-) Oct 12 20:55:45 well, maybe all are self-nominating for council elections right now Oct 12 20:56:01 day when I shred flasher-3.5 is comming... Oct 12 20:56:27 Pali: I'll give you a visit that day, bring some Semtex Oct 12 20:56:57 and a MTHEL (nice offroad car as well) Oct 12 20:58:26 if only there wasn't that nasty H2(2)F Oct 12 20:58:48 or was it H(2)2F? Oct 12 20:59:40 Di-Deuterium-Flouride Oct 12 21:00:33 already turned my neighbour's trees pink Oct 12 21:01:42 ~mthel Oct 12 21:01:48 o.O Oct 12 21:02:44 ~mthel is http://www.youtube.com/watch?v=cCBwLJjzDJQ Oct 12 21:02:44 DocScrutinizer05: okay Oct 12 21:09:17 http://www.youtube.com/watch?feature=player_detailpage&v=cCBwLJjzDJQ#t=266s look at lower left side, the exhaust fumes, pure H2F Oct 12 21:10:24 eeek. Oct 12 21:12:55 Pali: pong Oct 12 21:13:14 merlin1991, I have new uboot build Oct 12 21:13:18 can you test it? Oct 12 21:16:12 Pali: sure, gimme a link, and I'll report back when I checked it :) Oct 12 21:16:41 merlin1991: http://atrey.karlin.mff.cuni.cz/~pali/u-boot/u-boot-2012.10-rc3-1+zImage-2.6.28-omap1.fiasco Oct 12 21:16:58 it also should turn off n900 led Oct 12 21:17:04 SpeedEvil: just that it's not ordinary H2, it's Deuterium Oct 12 21:18:42 (needed to push laser wavelength into a transparent window of atmosphere) Oct 12 21:19:53 with normal hydrogen the laser beanm would die after 10m Oct 12 21:21:13 anyway except for pink countryside an MTHEL makes for a nice BBQ Oct 12 21:21:58 just Deuterium gas is a tad hard to refill Oct 12 21:24:00 watch the variolight alike reflector moving 2 degrees of freedom, and wonder how they managed to feed the laser light in there Oct 12 21:32:25 glasfibre? Oct 12 21:32:45 must be glasfibre Oct 12 21:46:52 raw megawatts of pure photonic energy, awesome Oct 12 21:51:44 CW Oct 12 21:57:27 * DocScrutinizer05 sends Mohammad- to next weekend surplus auction to get another two MTHEL Oct 12 21:58:17 just weird why IL doesn't use them to stop that junk flying over from Gaza Oct 12 22:01:09 merlin1991, is new uboot working? Oct 12 22:01:16 * DocScrutinizer05 suspect wind direction being the culprit Oct 12 22:02:10 sry, I got sidetracked, but now I'm on it Oct 12 22:02:12 if wind would take that H2F cloud direction Gaza reliably, they already would shoot down each white pigeon crossing the border Oct 12 22:04:27 Pali: whatever else I find cool or not-my-style on http://gitorious.org/rx51-bme-replacement/usbmode/blobs/master/usbmode.sh, you should make the sleeep delays configurable Oct 12 22:05:26 4s is probably rather short to discharge all the buffer capacitors on VBUS so you can drop USB session and switch speed Oct 12 22:06:05 you better check session status and delay proceeding until session dropped Oct 12 22:06:45 (you can get session status from musb-hdrc status byte, iirc) Oct 12 22:07:31 also enabling charging mode on bq24150 chip might speed up discharging Oct 12 22:08:22 and finally there's a discharge command for PHY, although flawed on hw level by a useless 1k resistor Oct 12 22:09:12 ~curse mentorgrafucks musb-hdrc controller Oct 12 22:09:13 May you be reincarnated as a Windows XP administrator, mentorgrafucks musb-hdrc controller ! Oct 12 22:11:15 we've witnessed >30s without session dropping when we tested stuff back in the days of h-e-n prototyping Oct 12 22:12:34 Pali: first approach would be to wait for /sys/class/i2c-adapter/i2c-1/1-0048/twl4030_usb/vbus to go 0 Oct 12 22:12:43 and _then_ some more seconds Oct 12 22:13:02 windows xp isn't that bad compared to other windows Oct 12 22:14:54 Pali: it boots Oct 12 22:15:27 merlin1991, nice :-) Oct 12 22:15:35 DocScrutinizer05, ok Oct 12 22:16:04 merlin1991 and turn it led off? Oct 12 22:16:36 I didn't follow it closely, when should the led go off? Oct 12 22:16:51 LED? why is LED on, first instance? Oct 12 22:17:19 emergency charger steady yellow? Oct 12 22:17:36 or some friggin LP5523 driven pattern? Oct 12 22:18:12 led runs from the boot shit Oct 12 22:18:18 when you press the power button it is white Oct 12 22:18:24 and when uboot starts it stays white Oct 12 22:18:32 but Pali the led did turn off Oct 12 22:18:41 it's funny enough that NOLO has yellow charger pattern hardcoded, you'll notice that as soon as you define something different for MCE charger pattern Oct 12 22:19:15 LP5523 driven pattern "device start up" Oct 12 22:19:27 so another NOLO crap Oct 12 22:19:33 In another channel once, I saw this guy who said he once coded a 3D fps game into nokia's bootloader, but it got removed once it was discovered :o( Oct 12 22:19:35 some white color Oct 12 22:19:41 yes, by NOLO Oct 12 22:19:53 when you press power button Oct 12 22:20:00 :nod: Oct 12 22:20:40 so pali, what's new in this setup? Oct 12 22:20:47 I'd replace that by a very unique uBoot turquoise Oct 12 22:20:54 ;-) Oct 12 22:21:21 merlin1991, disabled onenand code which enabled booting Oct 12 22:21:41 so basically what fmg did? Oct 12 22:21:43 rewritten assembler code for uboot relocation Oct 12 22:22:07 reusing bootmode from NOLO Oct 12 22:22:27 added variables setup_boot_reason_atag and setup_boot_mode_atag for changing ATAGS Oct 12 22:22:27 oh quite a few things then :D Oct 12 22:22:50 hm what's the quickest way to start a local http server for some time? Oct 12 22:22:52 no turquoise LED though :-S Oct 12 22:23:04 ignore zero size ATAG_MEM Oct 12 22:23:15 reset lp5523 led Oct 12 22:23:23 on KDE there's some ultra silly taskbar applet Oct 12 22:23:51 xfce ;) Oct 12 22:23:55 + rebased on last uboot master Oct 12 22:24:20 hmm, there's some http server called like a jet fighter plane Oct 12 22:24:35 written in perl or sth Oct 12 22:25:02 basically a scriptie Oct 12 22:25:32 Werner Almesberger using it for instance Oct 12 22:27:21 damn I need to deploy my new pubkey to ~8 devices Oct 12 22:27:29 the joys of setting up a new laptop Oct 12 22:27:36 was it tomcat? Oct 12 22:27:53 tomcat is an apache java ee shit plugin crap thingy Oct 12 22:28:19 Pali, when I build u-boot-shr/pali from git, it displays "u-boot 2012.04.01" Oct 12 22:28:22 aaah java Oct 12 22:28:36 Hurrian, changes not pushed yet Oct 12 22:28:45 I'm still testing new build Oct 12 22:28:52 ah Oct 12 22:29:11 * merlin1991 just installs lighttpd apt-get purge is there for a reason :D Oct 12 22:29:13 even the slide detection patch was not pushed? Oct 12 22:31:12 I have to say I missed the swapping desktops like a madman on a single screen Oct 12 22:32:03 merlin1991: http://en.wikipedia.org/wiki/Comparison_of_lightweight_web_servers Oct 12 22:33:01 I think I heard of http://en.wikipedia.org/wiki/Lighttpd before Oct 12 22:34:35 http://talk.maemo.org/showthread.php?p=1279882#post1279882 fs corruption /me smells oc user Oct 12 22:35:30 welllllll Oct 12 22:35:46 I think i won't change my tmo signature Oct 12 22:35:58 Oc is 1337 Oct 12 22:38:25 merlin1991: this sounds like PEBKAC though Oct 12 22:38:52 well he has a corrupted dpkg database I bet it's not purely his fault Oct 12 22:38:53 apt-get install mp-fremantle-generic-pr ;# C'MON Oct 12 22:41:38 anyway, >>dpkg: parse error, in file `/var/lib/dpkg/available` near line 37064<< rrrright, corrupt db Oct 12 22:41:44 he's not going to get there with foobar dpkg Oct 12 22:42:26 BM FTW Oct 12 22:45:28 Hurrian, detection of closed/open keyboard is already here Oct 12 22:46:06 if you have closed keyboard u-boot will boot automatically attached kernel (or other which is configured) Oct 12 22:46:53 IroN900:~# du -h /var/lib/dpkg/ Oct 12 22:46:55 27M /var/lib/dpkg/info Oct 12 22:46:56 :-/ Oct 12 22:47:19 IroN900:~# df -h /var/lib/dpkg/info Oct 12 22:47:20 Dateisystem Size Used Avail Use% EingehÃĪngt auf Oct 12 22:47:22 ubi0:rootfs 228M 169M 56M 76% / Oct 12 22:47:29 double -- :-/ Oct 12 22:48:24 33M /var/lib/dpkg/ Oct 12 22:48:39 insane Oct 12 22:52:27 IroN900:~# du -h /usr/lib/locale/locale-archive Oct 12 22:52:29 23M /usr/lib/locale/locale-archive Oct 12 22:52:49 Pali, which branch? I tried building remotes/origin/pali and it's not there Oct 12 22:52:56 even almost as insane Oct 12 22:53:11 Hurrian, I did not pushed new changes yet Oct 12 22:53:26 but it will be in branch pali Oct 12 22:53:29 ah Oct 12 22:53:30 alright. Oct 12 22:53:52 56MB for mere cruft Oct 12 22:54:02 on precious rootfs Oct 12 22:54:58 hell, my fist linux ran on a PC with a HARDISK of smaller size than those two Oct 12 22:55:19 first* Oct 12 22:56:14 hah, the fun of setting up a new scratchbox Oct 12 22:57:56 and honestly, you could easily nuke that insane locale cache, drop shit like cn_CN (sorry fellas ^_^) Oct 12 22:58:32 if only Nokia would've considered shipping the plain locale files for each lang separately Oct 12 22:59:30 or even better, would've used sane en_EN or at least technical-english strings in their binaries, rather than that tana_fi_foobar crap Oct 12 22:59:45 Now I have patch for Maemo which ask when somebody going to flash kernel fia fiasco-image-update Oct 12 23:00:04 kernel-*flasher and u-boot-flasher packages using it Oct 12 23:00:26 so now we can fix problem with Provides lines Oct 12 23:00:28 sounds like a good idea Oct 12 23:00:59 just I wonder if it will fsckup unattended installs Oct 12 23:06:15 merlin1991: http://en.qi-hardware.com/irclogs/latest.log.html#t22:26 Oct 12 23:07:12 hehe, but I have lighttpd set up now :P Oct 12 23:12:18 merlin1991: refresh! Oct 12 23:12:52 merlin1991: friggin qibot updates logs every 30 min only Oct 12 23:12:59 hehe Oct 12 23:28:02 code is here: https://code.launchpad.net/~pali/+junk/maemo_fiasco-image-update-ask Oct 12 23:29:05 it is deb package which patch fiasco-image-update to ask for flashing kernel image if version string change Oct 12 23:44:11 phew finally finished the scratchbox targets Oct 13 00:02:15 wow. WWW.aa.net.UK has word services. for example, a mobile SIM that terminates on a SIP server of your choice Oct 13 00:02:24 weird services Oct 13 00:03:21 merlin1991, freemangordon, needs testing: https://garage.maemo.org/pm/task.php?func=detailtask&project_task_id=1979&group_id=2300&group_project_id=3864 Oct 13 00:03:38 I want that u-boot will depends on this package Oct 13 00:03:50 and kernel-power too Oct 13 00:03:55 how is that going to fly in -ex Oct 13 00:04:02 so you want to push it to extras? Oct 13 00:04:17 first it needs to be tested Oct 13 00:04:21 otherwise you make kernel-power cssu-only Oct 13 00:04:32 and then yes, to extras (because of uboot & kernel-power) Oct 13 00:04:49 I'll check it out as soon as I can Oct 13 00:05:28 it allow to install -flasher package (for deb provides) without flashing image... Oct 13 00:15:27 merlin1991: [2012-10-13 01:43:46] DocScrutinizer05: my fixes are in the regular awhttpd sources, http://www.hcsw.org/awhttpd/ http://www.hcsw.org/awhttpd/awhttpd-3.0.8.tgz Oct 13 01:04:17 Is it possible to get maps for the N900 without using the PC software? Oct 13 01:13:48 http://www.maemopeople.org/index.php/jaffa/2009/10/13/downloading_ovi_maps_without_a_network_c Oct 13 01:15:06 and there also is osm based http://monav.openstreetmap.de/ Oct 13 01:21:09 nox-: that site it links to was hacked Oct 13 01:21:23 ew Oct 13 01:21:34 so where do I get the files from? Oct 13 01:22:00 dunno guess youll have to wait for it to be cleaned Oct 13 01:22:47 or use osm... Oct 13 01:26:34 :/ Oct 13 01:29:54 bbl **** ENDING LOGGING AT Sat Oct 13 03:00:02 2012