**** BEGIN LOGGING AT Sun Apr 29 02:59:58 2012
Apr 29 07:34:45 <Itschue> good morning
Apr 29 09:13:14 <Itschue> someone here ?
Apr 29 09:14:22 <spark666> yes
Apr 29 09:15:30 <Itschue> my n900 cant acces to nokia repository
Apr 29 09:15:46 <Itschue> can you access nokia repository?
Apr 29 09:16:12 <spark666> yes
Apr 29 09:16:23 <Itschue> you can access ?
Apr 29 09:16:33 <spark666> your internet connection works for sure?
Apr 29 09:16:39 <spark666> on phone i mean
Apr 29 09:16:58 <Itschue> yes on phone i have acces to maemo.org repository
Apr 29 09:17:11 <Itschue> can surf to google and other inet sites
Apr 29 09:18:00 <Itschue> an on pc can surf but cant acces to nokia repository
Apr 29 09:19:57 <Itschue> its curious or?
Apr 29 09:20:14 <spark666> yeah...
Apr 29 09:20:32 <spark666> brb ill need to reboot my laptop
Apr 29 09:21:01 <Itschue> ok im back later have to go to bathromm for something
Apr 29 09:22:36 <spark666> okey
Apr 29 09:38:09 <freemangordon> Pali, ping
Apr 29 09:38:17 <Pali> freemangordon, pong
Apr 29 09:39:05 <freemangordon> Pali, hald-addon-bme is to the point that I have to get the real data, Where am I supposed go get that?
Apr 29 09:39:11 <freemangordon> from
Apr 29 09:39:34 <freemangordon> which sysfs entry is that?
Apr 29 09:39:48 <Pali> /sys/class/power...
Apr 29 09:40:05 <Pali> use entry "uevent"
Apr 29 09:40:08 <freemangordon> so, i should stop bme and insmod bqxxx module?
Apr 29 09:40:13 <Pali> yes
Apr 29 09:40:20 <freemangordon> uevent?
Apr 29 09:40:26 <Pali> also for monitoring changes use inotify calls
Apr 29 09:40:41 <Pali> standard linux uevent file
Apr 29 09:40:46 <freemangordon> :D
Apr 29 09:40:55 <freemangordon> good to hear. 
Apr 29 09:41:11 <freemangordon> sorry, do you have an example or I have to google 
Apr 29 09:41:18 <Pali> after loading bq27x00_battery module, you will get /sys/class/power_supply/bq27200-*/
Apr 29 09:41:26 <freemangordon> I know that
Apr 29 09:41:27 <Pali> and here will be uevent file
Apr 29 09:41:33 <freemangordon> know that too
Apr 29 09:41:40 <Pali> format is PROPERTY=VALUE
Apr 29 09:41:44 <freemangordon> but have NFC what to do with it
Apr 29 09:41:47 <freemangordon> aaah
Apr 29 09:41:50 <freemangordon> ok
Apr 29 09:41:58 <Pali> use inotify for monitoring if that file was changed
Apr 29 09:42:05 <Pali> and when changed, read it
Apr 29 09:42:20 <freemangordon> ok
Apr 29 09:42:45 <Pali> when you load bq2415x_charger module you will get entry /sys/class/power_supplt/bq2415x*/
Apr 29 09:43:12 <freemangordon> Pali, do I need it?
Apr 29 09:43:13 <Pali> but I think for monitoring battery is only needed bq27200 chip
Apr 29 09:43:18 <freemangordon> yeah
Apr 29 09:43:45 <Pali> bq24150 chip tell you when you really charging battery
Apr 29 09:44:03 <Pali> but that can tell also bq27200
Apr 29 09:44:18 <freemangordon> ok, for now I will use 27200
Apr 29 10:19:16 <DocScrutinizer> Pali: inotify doesn't work on sysfs
Apr 29 10:19:35 <Pali> and what is working?
Apr 29 10:20:03 <freemangordon> g_file_monitor?
Apr 29 10:20:05 <DocScrutinizer> kevents, if maemo powerkernel supports that
Apr 29 10:21:03 <Pali> for example musb driver call sysfs_notify
Apr 29 10:21:14 <Pali> and then hal is monitoring status of file
Apr 29 10:22:26 <DocScrutinizer> Pali: you loked into ke-recv
Apr 29 10:22:33 <DocScrutinizer> look again!
Apr 29 10:22:50 <Pali> ke-recv receiving events from hal
Apr 29 10:22:58 <DocScrutinizer> ooh
Apr 29 10:23:06 <Pali> it connect to hal via dbus
Apr 29 10:23:14 <Pali> and listen for dbus signals
Apr 29 10:24:02 <Pali> ok, I looked into power_supply code and it call kobject_uevent KOBJ_CHANGE
Apr 29 10:24:13 <DocScrutinizer> that's kevents
Apr 29 10:24:31 <DocScrutinizer> aka kobjects
Apr 29 10:25:45 <Pali> I looked into hal year ago and I remember it monitor uevent file of power_supply
Apr 29 10:25:54 <Pali> I download sources again
Apr 29 10:26:21 <DocScrutinizer> maybe it's polling uevent
Apr 29 10:26:27 <DocScrutinizer> which is a BAD thing
Apr 29 10:28:36 <DocScrutinizer> I always thought ke-recv is KernelEvents-RECeiVe
Apr 29 10:28:56 <Pali> when I patched kernel-power to call sysfs_notify when musb driver change mode to host HAL has correct state of usb host mode
Apr 29 10:29:34 <Pali> without that patch HAL still show old informations, so hal do not polling for musb
Apr 29 10:29:47 <DocScrutinizer> sysfs-notify is using kevents afaik
Apr 29 10:30:30 <Pali> shit, now I looked into hal code for power_supply
Apr 29 10:30:34 <Pali> and it doing polling
Apr 29 10:30:53 <DocScrutinizer> I knew that
Apr 29 10:32:19 <DocScrutinizer> sysfs-notify / kevents wasn't available or very commonly used when maemo kernel version was developed
Apr 29 10:32:32 <Pali> I wil try to look what is doing with musb
Apr 29 10:32:41 <DocScrutinizer> kevents concept is relatively "new"
Apr 29 10:33:52 <DocScrutinizer> use MXR, search for kobject_uevent
Apr 29 10:34:05 <DocScrutinizer> a pity you won't get a single hit
Apr 29 10:34:18 <DocScrutinizer> we at h-e-n frowned at that a lot
Apr 29 10:34:19 <freemangordon> what about libudev?
Apr 29 10:34:54 <DocScrutinizer> freemangordon: udev is similar to kobjects, just a tiny bit different
Apr 29 10:35:13 <freemangordon> but could it be used?
Apr 29 10:35:21 <DocScrutinizer> I don't think so
Apr 29 10:35:36 <DocScrutinizer> it could be ABused
Apr 29 10:35:46 <freemangordon> so, what am I supposed to use :)
Apr 29 10:36:02 <DocScrutinizer> but why would we wanna do that when there's nice shiny kobject_uevent concept
Apr 29 10:36:24 <DocScrutinizer> freemangordon: for now you need polling
Apr 29 10:36:28 <DocScrutinizer> :-/
Apr 29 10:36:45 <DocScrutinizer> until pali fixed the kernel side and implemented kobject_uevent
Apr 29 10:36:51 <DocScrutinizer> aka sysfs-notify
Apr 29 10:37:04 <freemangordon> so, you are sure i cannot use udev_monitor?
Apr 29 10:37:33 <DocScrutinizer> yes. been there, tried that
Apr 29 10:37:47 <freemangordon> :(
Apr 29 10:39:01 <DocScrutinizer> udev-adm monitor -> nuttin
Apr 29 10:40:57 <DocScrutinizer> actually all you both said has been said by me some 2 years ago, and IIRC it was moh who educated me about maemo not using kevents
Apr 29 10:42:24 <DocScrutinizer> for sure it was moh who tought me inotify doesn't work on sysfs
Apr 29 10:42:58 <Pali> hal using special maemo addon for musb
Apr 29 10:43:04 <Pali> source code is in hald/linux/addons/addon-usb-cable.c
Apr 29 10:43:17 <freemangordon> damn, so I have to poll. well, actually stock hald-addon-bme polls bme too :)
Apr 29 10:43:25 <Pali> addon open filedesc on sysfs entry
Apr 29 10:43:31 <Pali> and call syscall poll
Apr 29 10:44:01 <Pali> after syscall exit read sysfs file
Apr 29 10:44:22 <freemangordon> Pali, where is that code?
Apr 29 10:44:54 <Pali> http://repository.maemo.org/pool/fremantle/free/h/hal/
Apr 29 10:45:54 <Pali> and if kernel (musb driver) do not call notify, poll syscall do not return
Apr 29 10:46:41 <freemangordon> so, can we use that for bq
Apr 29 10:46:51 <Pali> freemangordon, also code online here: http://mxr.maemo.org/fremantle/source/hal/hald/linux/addons/addon-usb-cable.c
Apr 29 10:47:04 <freemangordon> ok, i did apt-get source
Apr 29 10:47:36 * SpeedEvil idly wonders if you can swap emmc for a SD transparently.
Apr 29 10:47:40 <Pali> freemangordon, no idea if it will work, in power_supply is kobject_event call
Apr 29 10:48:08 <freemangordon> aah, ok, si i will have to poll it :(
Apr 29 10:48:13 <freemangordon> *so
Apr 29 10:48:48 <ShadowJK> iirc poll()/select() works in sysfs, and wakes up when driver calls notify_all() or something like that
Apr 29 10:49:16 <ShadowJK> oh you discovered already
Apr 29 10:49:22 <ShadowJK> I was scrolled up a bit :)
Apr 29 10:49:55 <ShadowJK> SpeedEvil: transparently, on the fly, online?
Apr 29 10:50:51 <ShadowJK> I think you'd edit the script which mounts emmc, and disable the scripts that mount sd? :)
Apr 29 10:52:24 <SpeedEvil> ShadowJK: Well - no
Apr 29 10:52:36 <SpeedEvil> ShadowJK: offline - with hot-air and shizzle.
Apr 29 10:52:50 <ShadowJK> oh, lol
Apr 29 10:56:50 <spark666> its okey to give user a password?
Apr 29 10:57:12 <spark666> like gpasswd user xxxxyyyxx
Apr 29 10:57:34 <DocScrutinizer> ShadowJK: Pali: freemangordon: aiui poll() and read() only wait for the function in kernel delivering new value for the sysfs node. However this is supposed to happen immediately after the read() *triggered* the function. So the poll()/read() doesn't wait for *new* state of e.g. VBUS, it simply *polls* the state each time addon-usb-cable.c main() gets called
Apr 29 10:57:37 <DocScrutinizer> AIUI
Apr 29 10:57:50 <DocScrutinizer> spark666: yes
Apr 29 10:57:56 <spark666> couse when i try to ssh my n900 (not as root) it ask me the passw
Apr 29 10:58:06 <spark666> but user doesnt have one by default afaik
Apr 29 10:58:13 <freemangordon> su user?
Apr 29 10:58:21 <freemangordon> after logged in as root
Apr 29 10:58:42 <spark666> freemangordon: and what that makes?:P
Apr 29 10:58:50 <DocScrutinizer> spark666: yes, that's how ssh works
Apr 29 10:59:13 <DocScrutinizer> ssh demands a password being set
Apr 29 10:59:14 <spark666> DocScrutinizer: humm user its allready in sudoers
Apr 29 10:59:25 <spark666> but its whit nopasswd
Apr 29 10:59:35 <DocScrutinizer> unrelated
Apr 29 10:59:40 <spark666> ok:P
Apr 29 11:00:05 <DocScrutinizer> though you might want to fix this nonsense as well
Apr 29 11:00:10 <DocScrutinizer> ~jrtools
Apr 29 11:00:11 <infobot> from memory, jrtools is http://wiki.maemo.org/User:Joerg_rw/tools
Apr 29 11:01:44 <spark666> it works
Apr 29 11:01:45 <spark666> :D
Apr 29 11:02:58 <freemangordon> so, i should setup a timer and read the state of what?
Apr 29 11:03:15 <DocScrutinizer> that's the very polling HAL does
Apr 29 11:03:39 <spark666> ohh and btw  its safe to switch to bash?
Apr 29 11:03:41 <DocScrutinizer> which I toldya you'll have to use until Pali fixed sysfs-notify cals in kernel
Apr 29 11:03:57 <DocScrutinizer> spark666: not exactly, for user
Apr 29 11:04:18 <spark666> DocScrutinizer: then ill stich to ash,even that im not a big fan :P
Apr 29 11:04:50 <DocScrutinizer> well it's safe to switch to bash for login shells, unless you also replace busybox by proper unix tools, *and* make bash a valid shell in /etc/shells
Apr 29 11:05:04 <DocScrutinizer> again, see
Apr 29 11:05:06 <DocScrutinizer> ~jrtools
Apr 29 11:05:07 <infobot> from memory, jrtools is http://wiki.maemo.org/User:Joerg_rw/tools
Apr 29 11:05:33 <spark666> im looking right now:P
Apr 29 11:07:07 <DocScrutinizer> zhe crux is: boot resp X/hildon init scripts are using some busyboxisms, and when your system user is using bash and unix tools instead busybox during boot, it will bootloop
Apr 29 11:08:25 <DocScrutinizer> root however is dafe to completely replace messybox and switch to bash, I've been told
Apr 29 11:08:30 <DocScrutinizer> safe*
Apr 29 11:10:06 <spark666> hummm
Apr 29 11:10:34 <DocScrutinizer> so if you are going to add /usr/bin/gnu etc to PATH in .bashrc, make DAMN SURE you remove this bullshit line >> . .bashrc<< from your .profile
Apr 29 11:11:17 <spark666> i wont switch to bash better..more safw
Apr 29 11:11:20 <spark666> safe*
Apr 29 11:11:40 <DocScrutinizer> as this will make busybox also source .bashrc, which in turn causes PATH make usr/bin/gnu primary executable location, which in turn causes bootloop
Apr 29 11:12:15 <spark666> i understand what you say
Apr 29 11:12:22 <DocScrutinizer> spark666: I have the setup like I shown in ~jrtools, and first cmd I enter after login is >>bash<<
Apr 29 11:12:31 <DocScrutinizer> safe bet
Apr 29 11:19:06 <freemangordon> DocScrutinizer, any idea how to force hal to load the addon?
Apr 29 11:19:14 <DocScrutinizer> nope
Apr 29 11:19:19 <DocScrutinizer> HAL is EVIL
Apr 29 11:19:25 <DocScrutinizer> I tend not to touch it
Apr 29 11:19:31 <DocScrutinizer> :-)
Apr 29 11:20:23 <vdv> following dbus call has no effect my n900 dbus-send --system --type=method_call --dest=com.nokia.mce /com/nokia/mce/request com.nokia.mce.request.req_led_pattern_activate string:"PatternCommunicationIM"
Apr 29 11:20:32 <vdv> what can be the reason for that?
Apr 29 11:20:34 <DocScrutinizer> esp since I heard it's already dead, and what we see in maemo is a zombie
Apr 29 11:22:04 <DocScrutinizer> vdv: either you got a typo, or your etc/mce/mce.ini has no pattern of that name, or that pattern has lower prio than the one currently active, or it's not defined to be active for the state your device is in
Apr 29 11:22:34 <vdv> hmm
Apr 29 11:22:49 <Pali> freemangordon, forgot about kobject, notify, ...
Apr 29 11:23:00 <vdv> doubt, that it typo is, because copy-pasted from here http://wiki.maemo.org/Phone_control#LED
Apr 29 11:23:06 <DocScrutinizer> iirc per default that pattern is defined as being active on locked screen only
Apr 29 11:23:20 <freemangordon> Pali, yeah, going to poll 
Apr 29 11:24:25 <DocScrutinizer> vdv: lemme check - a minute please
Apr 29 11:24:49 <Pali> bq27200 kernel driver doing polling internally
Apr 29 11:24:59 <vdv> there's a line in mce.ini, PatternCommunicationIM=30;1;0;0000;0000;40007f00017f017f050001ff01ff0000
Apr 29 11:25:11 <DocScrutinizer> vdv: works great on my locked device here
Apr 29 11:25:30 <Pali> so there cannot be proper implementation. you need to poll in kernel or poll in user space
Apr 29 11:25:37 <DocScrutinizer> also works on unlocked device3
Apr 29 11:25:52 <DocScrutinizer> I c&p your post above
Apr 29 11:26:09 <freemangordon> Pali, can I poll this uevent file?
Apr 29 11:26:23 <vdv> very interesting..
Apr 29 11:26:31 <vdv> something is wrong with my n900
Apr 29 11:26:35 <DocScrutinizer> Pali: I explained bq27200.ko is CRAP, just yesterday
Apr 29 11:26:35 <Pali> I tried syscall select, but it not worked correctly
Apr 29 11:27:30 <DocScrutinizer> Pali: OTOH this chip has no infra to do proper event driven (IRQ-based) signalling
Apr 29 11:28:13 <vdv> DocScrutinizer, PatternBatteryFull worked
Apr 29 11:28:15 <DocScrutinizer> vdv: check your settings-notificationlight
Apr 29 11:28:31 <Pali> bq driver read properties every 360s
Apr 29 11:28:39 <DocScrutinizer> LOL
Apr 29 11:28:47 <vdv> DocScrutinizer, device is connected via usb at the moment
Apr 29 11:28:49 <Pali> so uevent file is updated every 360s
Apr 29 11:29:02 <Pali> but if you read it, kernel driver update it immediately
Apr 29 11:29:06 <vdv> maybe that is the reason?
Apr 29 11:29:25 <Pali> so select/poll syscalls is not good idea
Apr 29 11:29:47 <freemangordon> no, i meant "read" every 5 secs or so
Apr 29 11:29:59 <Pali> freemangordon, use some timer and every 10s read uevent file
Apr 29 11:30:04 <freemangordon> yep
Apr 29 11:30:04 <DocScrutinizer> freemangordon: which is exactly the same
Apr 29 11:30:16 <freemangordon> the same as?
Apr 29 11:30:21 <Pali> no better solution
Apr 29 11:30:28 <DocScrutinizer> indeed
Apr 29 11:30:34 <freemangordon> :nod:
Apr 29 11:30:35 <Pali> bq chip has no signal...
Apr 29 11:30:42 <freemangordon> well, going to do it
Apr 29 11:30:54 <DocScrutinizer> as fsckng HAL doesn'T signal hald-addon-bme/whatever when to *poll* the sysfs file
Apr 29 11:31:14 <DocScrutinizer> and bq27200 doesn't tell kernel when values have changed
Apr 29 11:31:25 <vdv> DocScrutinizer, settings-notificationlight is a file?
Apr 29 11:31:35 <DocScrutinizer> SETTINGS
Apr 29 11:31:48 <vdv> ah, menu
Apr 29 11:32:03 <vdv> but there's no notificationlight submenu
Apr 29 11:32:08 <DocScrutinizer> meh
Apr 29 11:32:11 <DocScrutinizer> there is
Apr 29 11:32:18 <vdv> ah, right
Apr 29 11:32:20 <vdv> sorry )
Apr 29 11:32:59 <vdv> works now
Apr 29 11:33:02 <vdv> thanks
Apr 29 11:33:03 <DocScrutinizer> :-P
Apr 29 11:33:07 <DocScrutinizer> yw
Apr 29 11:35:01 <DocScrutinizer> freemangordon: however please ponder how to make the 10s sleep configurable
Apr 29 11:35:39 <freemangordon> yeah, sure
Apr 29 11:35:48 <DocScrutinizer> so you (at least on boottime, via parameter, $env, whatever) could set it to 1s for rapid reporting, or to 300s for better battery saving
Apr 29 11:35:48 <freemangordon> gconf or something
Apr 29 11:36:15 <DocScrutinizer> gconf is too way up?
Apr 29 11:36:27 <DocScrutinizer> dunno when HAL gets instantiated
Apr 29 11:36:47 <DocScrutinizer> if gconf is available at that time
Apr 29 11:37:22 <DocScrutinizer> I gather hal has other settings to read during startup, just add something there
Apr 29 11:37:35 <freemangordon> ok, will find a way, lets first have the functionality in place
Apr 29 11:37:54 <DocScrutinizer> if that's gconf - fine. If it's some /etc/* file, add it there
Apr 29 11:38:07 <freemangordon> ok, will remember it
Apr 29 11:38:11 <DocScrutinizer> :-
Apr 29 11:38:12 <DocScrutinizer> )
Apr 29 11:39:45 <DocScrutinizer> freemangordon:   sysfs_path = getenv ("HAL_PROP_LINUX_SYSFS_PATH");  (/addon-usb-cable.c 271)
Apr 29 11:41:01 <DocScrutinizer>   poll_period = getenv ("HAL_PROP_POLL_PERIOD_SECONDS");
Apr 29 11:41:23 <DocScrutinizer> yw
Apr 29 11:42:01 <DocScrutinizer> err
Apr 29 11:42:26 <DocScrutinizer>   poll_period = getenv ("HAL_PROP_VBUS_POLL_PERIOD_SECONDS");
Apr 29 11:42:41 <DocScrutinizer> analog for other stuff
Apr 29 11:43:28 <DocScrutinizer> while VBUS actually *has* IRQ so can get implemented event-driven/IRQ-based
Apr 29 11:43:42 <DocScrutinizer> bq27200 hasn't
Apr 29 11:43:46 <DocScrutinizer> so maybe
Apr 29 11:43:55 <DocScrutinizer>   poll_period = getenv ("HAL_PROP_BQ27200_POLL_PERIOD_SECONDS");
Apr 29 11:44:07 <DocScrutinizer> I guess you got it :-)
Apr 29 11:45:17 <DocScrutinizer> for the chanlog:
Apr 29 11:45:22 <DocScrutinizer> poll_period = getenv ("HAL_PROP_BQ27200_POLL_PERIOD_SECONDS");
Apr 29 11:50:36 <DocScrutinizer> on a sidenote: this is exactly the crap sensorfw of meego did - for *all* sensors
Apr 29 11:52:04 <DocScrutinizer> in any A(B(C(sysfs))) config, system C polls sysfs, B polls C, A polls B. Each on random periods, each causing wakeup of CPU.
Apr 29 11:52:51 <DocScrutinizer> Not like a sane event driven setup where sysfs notifies C notifies B notifies A, *only* if there's actually something *new* to notify
Apr 29 11:54:56 <DocScrutinizer> form above you might have guessed this is inevitable for some rare cases where idiotic peripherals like bq27200 have no IRQ to signal events (and even then B and C could wait for A starting a poll, and then hand it down to next lower layer, instead of each subsystem implementig its own timer based polling pace)
Apr 29 11:55:51 <DocScrutinizer> on a general design rule you never want to do things like that anyway (unless there's absolutely no other way) - sensorfw does it *all the time* like that
Apr 29 11:56:49 <DocScrutinizer> that's why I always said meego has been designed by obtuse system architects and will never reach standby time of maemo
Apr 29 11:58:02 <DocScrutinizer> well, I have to admit I didn'T bother to proofread the complete sensorfw code, to check if they really use this sensor template everywhere
Apr 29 11:58:39 <DocScrutinizer> and even less I checked whole meego if actually the architects had similar mindset when designing other subsystems than sensorfw
Apr 29 11:59:47 <DocScrutinizer> but the ignorance to even understand what I was telling them made me think whole system been designed with no clue about IRQ and event-driven design
Apr 29 12:00:26 <DocScrutinizer> standard answer: "we're using upstream"
Apr 29 12:01:20 <DocScrutinizer> that'S exactly what makes maemo different: Nokia didn't "use upstream" no matter what. Each system been carefully checked for optimal design, and completely rewritten if "upstream" was crap
Apr 29 12:01:52 <DocScrutinizer> almost each system, obviously :-D
Apr 29 12:01:57 <DocScrutinizer> not HAL though
Apr 29 12:02:30 <ShadowJK> the right way is in between, fix upstream and use it
Apr 29 12:02:31 <DocScrutinizer> HAL never been meant for this usecase
Apr 29 12:02:48 <ShadowJK> instead of rewriting from scratch, or using a broken upstream
Apr 29 12:02:56 <DocScrutinizer> HAL basic design is incompatible with even driven concept
Apr 29 12:03:23 <DocScrutinizer> HAL been invented for desktop PCs
Apr 29 12:03:39 <ShadowJK> hal is dead now
Apr 29 12:03:51 <DocScrutinizer> indeed, guess why ;-D
Apr 29 12:05:16 <DocScrutinizer> evenT driven concept*
Apr 29 12:10:58 <villager> I don't think that's why HAL is dead... HAL does handle events, it can e.g. send dbus notifications when properties change (though the notification unfortunately does not say what *to*, it does say that it changed), and apps can listen, so there's obviously mechanisms for that
Apr 29 12:12:16 <DocScrutinizer> villager: that's output side. input side - like hald-addon-foobar - seems unaware of any such event concept
Apr 29 12:13:29 <DocScrutinizer> villager: but yes, you're right, it could get fixed without breaking basic HAL design
Apr 29 12:14:21 <DocScrutinizer> and I agree pretty wholeheartedly about nonsensical lack of info regarding details of a change in HAL dbus msgs
Apr 29 12:14:30 <villager> DocScrutinizer: yeah, any lack of event seemed to me to be more like an implementation issue than a design issue
Apr 29 12:15:49 <villager> DocScrutinizer: from what I hear, HAL is dead mostly because it started to become a monolithic and unmaintainable beast (as well as a piece of crap)
Apr 29 12:16:02 <DocScrutinizer> bitten me when I wrote that 4 liner gimmick for playswoosh when slider opened
Apr 29 12:19:34 <DocScrutinizer> villager: http://talk.maemo.org/showpost.php?p=897443&postcount=14 , for nasty effects of HAL's dbus msgs missing state of element that changed
Apr 29 12:19:56 <DocScrutinizer> I need to keep a local var to store old state, and compare that to sysfs
Apr 29 12:20:28 <DocScrutinizer> I *could* also use HAL to query actual state
Apr 29 12:20:51 <DocScrutinizer> but, as mentioned above, I prefer to not touch HAL whenever avoidable
Apr 29 12:23:09 <DocScrutinizer> idiot HAL sending a dbus msg "buttonpressed state has changed" - but it does NOT send HOW it changed
Apr 29 12:26:33 <freemangordon> DocScrutinizer, care to write those HAL vars somewhere for further references
Apr 29 12:26:43 <freemangordon> wiki, or wherever
Apr 29 12:26:48 <DocScrutinizer> umm
Apr 29 12:27:02 <DocScrutinizer> Chanlog: http://mg.pov.lt/maemo-irclog
Apr 29 12:27:18 <freemangordon> no, I was talking about some documentation :)
Apr 29 12:27:51 <DocScrutinizer> I just looked up the MXR page Pali tossed, and found this one line 271
Apr 29 12:28:05 <DocScrutinizer> what shall I document about that fact?
Apr 29 12:30:52 <Pali> freemangordon, documentation is in file /usr/share/doc/hald-addon-bme/README.properties.gz
Apr 29 12:31:04 <Pali> in deb package hald-addon-bme_0.0.17+0m5_armel.deb
Apr 29 12:31:04 <DocScrutinizer> wiki page "souce line in HAL", content: "http://mxr.maemo.org/fremantle/source/hal/hald/linux/addons/addon-usb-cable.c:271" ?
Apr 29 12:31:16 <freemangordon> Pali, ok
Apr 29 12:31:41 <Pali> also some info is on: https://gitorious.org/mer-toggles/hald-addon-bme/blobs/master/hald-addon-bme.c and https://gitorious.org/mer-toggles/hald-addon-smartq/blobs/master/hald-addon-smartq.c
Apr 29 12:36:25 <DocScrutinizer> MEH https://gitorious.org/mer-toggles/hald-addon-smartq/blobs/master/hald-addon-smartq.c#line89
Apr 29 12:37:13 <freemangordon> yeah, the same for bme in mer
Apr 29 12:37:24 <freemangordon> 1000ms poll interval
Apr 29 12:37:50 <DocScrutinizer> yeah, https://gitorious.org/mer-toggles/hald-addon-bme/blobs/master/hald-addon-bme.c#line91
Apr 29 12:39:13 <DocScrutinizer> anyway I'm MEHing on the missing getenv() line assigning _poll_interval a new value when we want sth different from " ~1000ms (default)"
Apr 29 12:39:50 <freemangordon> well, I don't want to comment on that sourec ;)
Apr 29 12:39:52 <DocScrutinizer> since "(default)" suggests there has to be such line
Apr 29 12:39:53 <freemangordon> *source
Apr 29 12:42:26 <DocScrutinizer> btw I guess getenv() returns empty string (or nullpointer) when env undef. You need a) catch that, and b) convert string to int in a segv safe way, for poll_int
Apr 29 12:44:15 <freemangordon> DocScrutinizer, do not forget I am a dveloper after all :P
Apr 29 12:44:22 <freemangordon> *developer
Apr 29 12:44:26 <DocScrutinizer> :-)
Apr 29 12:44:31 <DocScrutinizer> sorry, no offense
Apr 29 12:44:57 <DocScrutinizer> just correcting a omission I did
Apr 29 12:45:50 <freemangordon> I saw it, code alredy written
Apr 29 12:45:59 <DocScrutinizer> :-D
Apr 29 12:46:48 <freemangordon> aah, should be in seconds
Apr 29 13:00:40 <Pali> hacking HAM
Apr 29 13:00:59 <Pali> I belive that current version in my git repo will be working now...
Apr 29 13:03:18 <DocScrutinizer> Pali: \o/
Apr 29 13:03:45 <Pali> needs testing: https://gitorious.org/~pali/community-ssu/pali-hildon-application-manager
Apr 29 13:03:57 <DocScrutinizer> I once looked into this unbearably obscure code that decides when to update, based on age of last update cache
Apr 29 13:04:58 <Pali> stupid scratchbox, it adding -mthumb
Apr 29 13:05:05 <vdv> how to know which hints does libnotify support?
Apr 29 13:05:16 <freemangordon> Pali, sorry for asking again the same stupid question, but what is supposed to be in uevent file?
Apr 29 13:05:21 <freemangordon> for bq
Apr 29 13:05:34 <DocScrutinizer> Pali: maybe you find a bug in that code, that makes HAM update after each single random user interaction
Apr 29 13:05:54 <freemangordon> yeah, that bug is driving me nuts
Apr 29 13:06:01 <Pali> PHYSDEVPATH=/class/i2c-adapter/i2c-2/2-0055
Apr 29 13:06:01 <Pali> PHYSDEVBUS=i2c
Apr 29 13:06:01 <Pali> PHYSDEVDRIVER=bq27x00-battery
Apr 29 13:06:01 <Pali> POWER_SUPPLY_NAME=bq27200-0
Apr 29 13:06:01 <Pali> POWER_SUPPLY_TYPE=Battery
Apr 29 13:06:01 <Pali> POWER_SUPPLY_STATUS=Discharging
Apr 29 13:06:04 <Pali> POWER_SUPPLY_PRESENT=1
Apr 29 13:06:06 <Pali> POWER_SUPPLY_VOLTAGE_NOW=3636000
Apr 29 13:06:08 <Pali> POWER_SUPPLY_CURRENT_NOW=436789
Apr 29 13:06:10 <Pali> POWER_SUPPLY_CAPACITY=48
Apr 29 13:06:12 <Pali> POWER_SUPPLY_TEMP=406
Apr 29 13:06:15 <Pali> POWER_SUPPLY_TIME_TO_EMPTY_NOW=4560
Apr 29 13:06:17 <Pali> POWER_SUPPLY_TIME_TO_EMPTY_AVG=4140
Apr 29 13:06:19 <Pali> POWER_SUPPLY_TECHNOLOGY=Li-ion
Apr 29 13:06:19 <DocScrutinizer> PALI!!!!
Apr 29 13:06:21 <Pali> POWER_SUPPLY_CHARGE_FULL=1147398
Apr 29 13:06:23 <Pali> POWER_SUPPLY_CHARGE_NOW=551208
Apr 29 13:06:26 <Pali> POWER_SUPPLY_CHARGE_FULL_DESIGN=2056320
Apr 29 13:06:26 <vdv> Pali, why don't you pastebin? :)
Apr 29 13:06:47 <Pali> yes, I will do
Apr 29 13:06:56 <DocScrutinizer> yay, one second from a kick
Apr 29 13:07:08 <BCMM> i don't think yelling at him will help when his IRC client is still finishing a paste
Apr 29 13:07:17 <DocScrutinizer> yes
Apr 29 13:07:21 <DocScrutinizer> thus the kick
Apr 29 13:10:02 <DocScrutinizer> POWER_SUPPLY_CHARGE_FULL_DESIGN=2056320  WTF??
Apr 29 13:11:16 <DocScrutinizer> a completely random selection of bq27200 values, some of which look even completely off resp in useless units
Apr 29 13:12:19 <Pali> ILMD register
Apr 29 13:12:48 <Pali> that which is not configured correctly on n900
Apr 29 13:12:49 <DocScrutinizer> raw
Apr 29 13:13:04 <DocScrutinizer> hmm, possible
Apr 29 13:13:14 <DocScrutinizer> anyway useless
Apr 29 13:13:24 <Pali> yes, this value is useless
Apr 29 13:13:42 <DocScrutinizer> while e.g. LMD is completely missing
Apr 29 13:13:45 <DocScrutinizer> as is CI
Apr 29 13:14:37 <DocScrutinizer> as is the highly useful set of values based on IDLE
Apr 29 13:15:42 <Pali> LMD is POWER_SUPPLY_CHARGE_FULL
Apr 29 13:17:46 <DocScrutinizer> s/IDLE/STANDBY/
Apr 29 13:17:47 <infobot> DocScrutinizer meant: as is the highly useful set of values based on STANDBY
Apr 29 13:17:47 <Pali> if CI is invalid driver do not report any other values
Apr 29 13:20:22 <DocScrutinizer> as is cycle-count
Apr 29 13:21:00 <Pali> cycle-count is missing due 2.6.28
Apr 29 13:21:09 <Pali> but is in upstream kernel
Apr 29 13:22:58 <ShadowJK> cycle count and cycle count since learning is useful for guessing if caacity has been corrected
Apr 29 13:24:00 <DocScrutinizer> basically a driver shall export a comprehensive set of registers, "expanded" and corrected for proper units
Apr 29 13:25:18 <DocScrutinizer> I suggest looking at output of http://maemo.cloud-7.de/maemo5/usr/local/sbin/bq27k-detail2
Apr 29 13:26:46 <DocScrutinizer> even that one missing the writable register access, for obvious reasons
Apr 29 13:27:53 <DocScrutinizer> we don't need the nifty calculator in bq27200, as we got a better one in CPU, but anyway you might want to write to the registers for doing other stuff - see bq27200 datasheet
Apr 29 13:29:21 <DocScrutinizer> if you don't support all registers in expanded normalized form, you at very least should add a dump sysnode, to get a complete hex register dump
Apr 29 13:30:55 <DocScrutinizer> so everybody is free to use his own userland tool like http://maemo.cloud-7.de/maemo5/usr/local/sbin/bq27k-detail2 on that dump sysnode
Apr 29 13:32:26 <DocScrutinizer> bq27200.ko is not only blocking bme access via I2C, but also i2cget tool. So bq27k-detail won't work when bq27200.ko got loaded, unless bq27200.ko exports a dump sysnode
Apr 29 13:32:38 <Pali> DocScrutinizer, in kernel-power there is interface for that
Apr 29 13:33:22 <Pali> userspace library for LD_PRELOAD is here: https://code.launchpad.net/~pali/+junk/maemo_libbqioctl
Apr 29 13:35:50 <DocScrutinizer> errrwut?
Apr 29 13:40:06 <Pali> DocScrutinizer, do not remember? search in your history :-) http://mg.pov.lt/maemo-irclog/%23maemo.2011-09-30.log.html#t2011-09-30T13:55:01
Apr 29 13:41:04 <DocScrutinizer> I know we pondered how to make bme work with bq27200.ko
Apr 29 13:41:57 <Pali> I tested it with i2cget tool and worked
Apr 29 13:42:41 <DocScrutinizer> but AIUI: bq27200.ko is locking I2Cbus 2 chipaddr 0x55. This libbqioctl will intercept I2C calls to taht addr and redirect them to /sys/*/*/bq27200/<foobar>
Apr 29 13:44:30 <DocScrutinizer> when bq27200.ko doesn't export cycle-count or STTE to sysfs, how is >libbqioctl i2cget -y 2 0x55 0x1c< going to work?
Apr 29 13:44:32 <Pali> it redirect i2c ioctl calls to /dev/bq*
Apr 29 13:44:48 <Pali> and /dev/bq* is character device created by bq kernel driver
Apr 29 13:45:04 <Pali> and it handle any call
Apr 29 13:45:19 <Pali> so make sure you do not ask for unknown register like BME :D
Apr 29 13:47:47 <DocScrutinizer> so what you say is: >find /sys/devices/ -path '*/hdq/*' -name dump<  is now >find /dev  -name bq*<?
Apr 29 13:50:11 <DocScrutinizer> as I'm not going to install a LD_PRELOAD lib for i2cget, to make a script like http://maemo.cloud-7.de/maemo5/usr/local/sbin/bq27k-detail2 work
Apr 29 13:53:49 <DocScrutinizer> if that'S what you say, I'd ask you to give me details so I can fix the script so it works on PK as well - *without* maemo_libbqioctl. And I'd ask you to test it on a system that already has bq27200.ko modprobed
Apr 29 13:56:18 <Pali> you need to build libbqioctl.so library
Apr 29 13:56:27 <DocScrutinizer> WTF?
Apr 29 13:56:34 <Pali> and then call i2cget like: LD_PRELOAD=./libbqioctl.so i2cget ....
Apr 29 13:56:42 <DocScrutinizer> no way dude
Apr 29 13:57:22 <DocScrutinizer> I want to do a pretty normal thing: read out a chip's complete register set. I'm not going to use a special library for that
Apr 29 13:57:53 <Pali> first check if bq module is loaded. if not use standard way
Apr 29 13:58:16 <DocScrutinizer> if bq27200.ko doesn't alow access to all registers of bq27200 chip, it's buggy
Apr 29 13:58:17 <Pali> if is loaded set export LD_PRELOAD=/path/to/libbqioctl.so
Apr 29 13:58:28 <Pali> it allow access to all reigsters
Apr 29 13:58:44 <DocScrutinizer> that's BS
Apr 29 13:58:50 <DocScrutinizer> sorry
Apr 29 13:59:20 <DocScrutinizer> cat /dev/bq27200 is the supposedly right way
Apr 29 13:59:42 <DocScrutinizer> if that's your direct access to chip registers now
Apr 29 14:00:20 <Pali> cat not working
Apr 29 14:00:25 <Pali> it use ioctl
Apr 29 14:00:31 <Pali> like /dev/i2c-*
Apr 29 14:00:33 <DocScrutinizer> then it's broken
Apr 29 14:00:46 <Pali> what broken?
Apr 29 14:00:52 <Pali> /dev/i2c-*?
Apr 29 14:01:05 <vdv> anybody knows, whether sound utilities are accessable from qt? http://maemo.org/api_refs/5.0/5.0-final/hildon/hildon-Sound-Utilities.html
Apr 29 14:02:29 <DocScrutinizer> listen: a kernel driver is meant to allow *simple* *comprehensive* access to hardware
Apr 29 14:02:52 <DocScrutinizer> *sigh*
Apr 29 14:03:01 <DocScrutinizer> toldya bq27200.ko is crap
Apr 29 14:04:34 <Pali> tested, bq27200.sh working fine with library
Apr 29 14:04:43 <Pali> sudo sh -c "LD_PRELOAD=./libbqioctl.so ./bq27200.sh"
Apr 29 14:04:44 <DocScrutinizer> if /dev/bq27200 is implementing a API ...
Apr 29 14:05:41 <DocScrutinizer> DANG I don't give a flying shit about LD_PRELOADing a lib I ave to build and install on device, to do a *pretty* *normal* thing that worked before bq27200.ko messed up shit
Apr 29 14:06:59 <DocScrutinizer> if /dev/bq27200 is implementing a API like /dev/I2C-2 then maybe we can symlink /dev/I2C-99 to /dev/I2C-2 and use i2cget -y 99 0x55...
Apr 29 14:09:28 <Pali> I added this to top of file bq27200.sh: if [ -e /sys/class/power_supply/bq27200-0 ]; then export LD_PRELOAD=/home/user/libbqioctl.so; fi
Apr 29 14:09:51 <Pali> it is not same api
Apr 29 14:10:08 <DocScrutinizer> IOW a /dev/foobar device is udeless when there are no tools to access this device in a standard manner. LD_PRELOAD is *no* standard manner/tool
Apr 29 14:10:46 <Pali> it is only for i2cget
Apr 29 14:10:58 <Pali> easy way how to fix scripts using i2cget
Apr 29 14:11:17 <DocScrutinizer> this will never fly, nobody sane will accept a LD_PRELOAD to make a script work where you got the source of to fix it the rpoper way
Apr 29 14:11:34 <Pali> at least see that code old (2011-08-02)
Apr 29 14:11:56 <Pali> I do not want to rewrite it now, no time for it
Apr 29 14:12:01 <Pali> and is working
Apr 29 14:12:12 <DocScrutinizer> if you define a new device in /dev, you as well need to provide the *proper* tools to access it. LD_PRELOAD=foobar i2cget is NO proper tool
Apr 29 14:12:42 <Pali> in bzr is tool bqread.c
Apr 29 14:12:48 <Pali> for that interface
Apr 29 14:12:51 <DocScrutinizer> so I declare bq27200.kp buggy, same as it always been
Apr 29 14:13:31 <vdv> found, have to include <hildon-1/hildon/hildon-sound.h>
Apr 29 14:13:55 <DocScrutinizer> since it's neither exporting all registers to sys, nor offers any standard alternative method to access those registers
Apr 29 14:16:03 <DocScrutinizer> honestly what's the common sense in inventing a /dev/bq27200 device that is neither accessable via standard >cat< nor compatible to /dev/I2C* ?
Apr 29 14:16:52 <DocScrutinizer> is it compatible to *anything*
Apr 29 14:17:03 <DocScrutinizer> maybe V4L2 interface?
Apr 29 14:17:06 <DocScrutinizer> :-P
Apr 29 14:20:44 <spark666> DocScrutinizer: does it worth to give megoo a shot?
Apr 29 14:21:11 <DocScrutinizer> suggested fix: export ALL registers to /sys/*/bq*/*
Apr 29 14:21:13 <spark666> meego*
Apr 29 14:21:36 <DocScrutinizer> spark666: I don't think so, but that's your beer
Apr 29 14:21:57 <spark666> DocScrutinizer: i mean its more "linux" than maemo/
Apr 29 14:22:00 <spark666> ?
Apr 29 14:22:17 <DocScrutinizer> it's less mature than maemo
Apr 29 14:22:33 <DocScrutinizer> maemo is pretty much linux as much as it gets
Apr 29 14:22:56 <DocScrutinizer> I mean, maemo is not android
Apr 29 14:23:01 <spark666> yeah!
Apr 29 14:23:17 <spark666> meego looks a bit like android from what i see :P
Apr 29 14:24:16 <spark666> but does n950 run meego?or something else?
Apr 29 14:24:40 <DocScrutinizer> maemo has some userland apps that rely on some /sys et al idiosyncrasies of this particular maemo kernel, which otherwise is a normal linux kernel also regarding those idiosyncrasies - at least 99.99%
Apr 29 14:25:12 <DocScrutinizer> it's just you will break those APIs when you update to newer linux kernel versions
Apr 29 14:26:14 <DocScrutinizer> meego otoh doesn't have those userland apps that require a special kernel land. Mostly because meego doesn't have *any* userland apps ;-P
Apr 29 14:26:30 <spark666> hahahah :P
Apr 29 14:26:46 <spark666> at least im happy that we have CSSU community
Apr 29 14:27:02 <spark666> couse i see that nokia droped us :(
Apr 29 14:28:10 <DocScrutinizer> well, btw meego != meego
Apr 29 14:28:20 <spark666> huh?
Apr 29 14:28:53 <DocScrutinizer> there's "true" meego, as in meego CE, and there's Harmattan "meego" which basically is Nokia's maemo rechristinaized
Apr 29 14:29:11 <spark666> lol
Apr 29 14:29:11 <DocScrutinizer> the latter running on N9 and N950
Apr 29 14:29:30 <DocScrutinizer> my comments above were about true meego
Apr 29 14:29:33 <spark666> so the second meego its kinda maemo6?:P
Apr 29 14:29:38 <DocScrutinizer> yes
Apr 29 14:29:58 <spark666> so your opinion about the second its good,right?
Apr 29 14:30:14 <DocScrutinizer> not really
Apr 29 14:30:44 <DocScrutinizer> it's again loaded with blob libs that are impossible to port to any other platform
Apr 29 14:30:55 <DocScrutinizer> and it's almost more deadend than maemo5
Apr 29 14:31:03 <auenf> meego is more open than meego
Apr 29 14:31:03 <spark666> lol
Apr 29 14:31:24 <spark666> lets name them meego and meego 2 :P
Apr 29 14:31:40 <auenf> meego is more open than harmattan
Apr 29 14:31:48 <DocScrutinizer> nah, they are called meegoCE and HARM(attan)
Apr 29 14:31:58 <spark666> :D
Apr 29 14:32:05 <auenf> all maemos had quite a lot of closed source components
Apr 29 14:32:10 <auenf> meego would have fixed most of that
Apr 29 14:32:19 <DocScrutinizer> HARM doesn't
Apr 29 14:32:31 <auenf> harmattan is only meego by name
Apr 29 14:32:35 <spark666> so meego CE its full opensource?
Apr 29 14:32:38 <DocScrutinizer> indeed
Apr 29 14:32:48 <DocScrutinizer> though it's no proper maemo either
Apr 29 14:32:53 <spark666> and its not made by nokia/
Apr 29 14:32:54 <spark666> ?
Apr 29 14:32:57 <DocScrutinizer> at least in my book
Apr 29 14:33:16 <DocScrutinizer> spark666: (meegoCE) yes
Apr 29 14:33:16 <auenf> meego was nokia & intel with a whole bunch of other minor input company
Apr 29 14:33:18 <auenf> s
Apr 29 14:33:30 <DocScrutinizer> now "continued" as nemo/mer
Apr 29 14:33:35 <spark666> lol
Apr 29 14:33:55 <spark666> i saw something about tizen to (even that i have no idea whats that) :P
Apr 29 14:34:05 <DocScrutinizer> see #mer #nemomobile
Apr 29 14:34:28 <DocScrutinizer> tizen is intel's idea of meego's future
Apr 29 14:34:36 <DocScrutinizer> AIUI
Apr 29 14:34:45 <spark666> maemo its the best,and it would be even better if we hade all parts opensource,so ppl could develop more stuf for it
Apr 29 14:35:03 <DocScrutinizer> yep
Apr 29 14:35:13 <auenf> nokia wanted meego for arm
Apr 29 14:35:25 <auenf> intel wanted to fight in the mobile arm war with x86
Apr 29 14:35:28 <DocScrutinizer> that's where CSSU and our arguments here about bq27200.ko come in
Apr 29 14:35:53 <spark666> whats bq27200.ko is?
Apr 29 14:36:07 <DocScrutinizer> a kernel driver for the battery gauge chip
Apr 29 14:36:41 <spark666> humm
Apr 29 14:36:52 <DocScrutinizer> that didn't exist in original maemo5 and that resulted in quite some troubles when we tried to augment the system at that end
Apr 29 14:37:41 <DocScrutinizer> as this driver (again) conflicts with closed userland stuff, namely bme
Apr 29 14:38:00 <spark666> that s**ks
Apr 29 14:38:01 <spark666> :(
Apr 29 14:38:37 <DocScrutinizer> well, we found our way around that, we're just "fighting" for the best implementation still :-D
Apr 29 14:38:50 <spark666> :>
Apr 29 14:39:36 <spark666> now im running the defaulf stuffs + cssu testing version
Apr 29 14:39:43 <DocScrutinizer> better "fight" now, rather than moan later
Apr 29 14:39:55 <spark666> its been more than a year since i had the last n900
Apr 29 14:40:05 <spark666> and i kinda forgot some stuffs
Apr 29 14:40:22 <spark666> hahah nice words :D
Apr 29 14:40:40 <DocScrutinizer> once stuff is deployed, it's hard to change some details in some parts
Apr 29 14:41:11 <villager> noo, don't say "the last n900"
Apr 29 14:41:18 <DocScrutinizer> so my part always is the bitching for the better
Apr 29 14:41:28 <spark666> villager: the last that i had
Apr 29 14:41:35 <spark666> now i buyed a new one :D:D:D
Apr 29 14:41:44 <DocScrutinizer> *new*?
Apr 29 14:41:51 <spark666> second hand :(
Apr 29 14:41:55 <DocScrutinizer> :-/
Apr 29 14:42:08 <DocScrutinizer> I'm looking for really new ones
Apr 29 14:42:11 <spark666> the first that i had whas new more than 2 years ago
Apr 29 14:42:31 <spark666> i payed 500 euros on that one
Apr 29 14:42:33 <villager> spark666: phew... will be a sad day when there's only one n900 left in the world
Apr 29 14:42:40 <spark666> but now i dont have much money anymore :(
Apr 29 14:42:45 <DocScrutinizer> I only have three devices, makes me kinda nervous ;-D
Apr 29 14:42:55 <spark666> villager: we will fight for it :P 
Apr 29 14:43:18 <spark666> DocScrutinizer: wow :P your rich :P
Apr 29 14:43:27 <DocScrutinizer> not at all :-)
Apr 29 14:43:32 <DocScrutinizer> just crazy
Apr 29 14:43:51 <spark666> im crazy to
Apr 29 14:43:54 <spark666> :P
Apr 29 14:44:44 <DocScrutinizer> I'm less poor now than I've been half a year ago, but there's no more really unused new N900 available now :-/
Apr 29 14:45:00 <spark666> im not a big fan of ubuntu,but if they will make ubuntu mobile version will be nice
Apr 29 14:45:04 <ThreeM> sure there is
Apr 29 14:45:10 <ThreeM> but at a inhuman price
Apr 29 14:45:24 <DocScrutinizer> where?
Apr 29 14:45:24 <ThreeM> amazon what 599eur (in germany) for a new N900
Apr 29 14:45:27 <villager> hope they continue to sell replacement parts for a while, at least... I still haven't fixed my other n900, the one with the cracked lcd
Apr 29 14:45:34 <DocScrutinizer> o.O
Apr 29 14:45:37 <spark666> in my country its still 500 euros in stores
Apr 29 14:45:56 <DocScrutinizer> villager: spare parts getting rare as well
Apr 29 14:46:03 <ThreeM> http://www.amazon.de/Nokia-Smartphone-Maemo-QWERTZ-Tastatur-black/dp/B002N2Z0MQ/ref=sr_1_1?ie=UTF8&qid=1335710754&sr=8-1
Apr 29 14:47:24 <DocScrutinizer> hmmm
Apr 29 14:47:50 <DocScrutinizer> westcom, maybe this one isn't a sucker cheating his customers this time
Apr 29 15:02:35 <DocScrutinizer> did ebay nuke the "ask seller a question" function?
Apr 29 15:03:09 <ThreeM> i hate ebay. some of the worst sites on the internet
Apr 29 15:03:16 <DocScrutinizer> sure
Apr 29 15:03:21 <DocScrutinizer> same here
Apr 29 15:03:51 <ThreeM> periot im searching for functions... they are there but you have to serach for it
Apr 29 15:04:23 <DocScrutinizer> I'd rather do bunjee jumping or play russian roulette, if that would get me a new N900
Apr 29 15:05:25 <ThreeM> my n900 is in pretty good condition
Apr 29 15:05:51 <ThreeM> im lookging forward for the N9 but im still not sure to buy
Apr 29 15:06:29 <DocScrutinizer> you can have mine ;-) nah, not really
Apr 29 15:07:23 <DocScrutinizer> it's a lend-out anyway, so I mustn't give it away
Apr 29 15:07:43 <ThreeM> is the n) a good device?
Apr 29 15:07:49 <DocScrutinizer> and actually last few weeks I started to look into hostmode for N9 again
Apr 29 15:07:55 <ThreeM> a "like N900 good" device?
Apr 29 15:08:08 <DocScrutinizer> depends on what you expect
Apr 29 15:08:15 <DocScrutinizer> HARM is no maemo anymore
Apr 29 15:08:40 <DocScrutinizer> and the device never been an option for me, with crappy capacitive touchscreen and no hw-kbd
Apr 29 15:09:05 <ThreeM> when this means i get no HAM... yea! ;)
Apr 29 15:09:07 <ThreeM> hmm
Apr 29 15:09:27 <ThreeM> damn, im looking for a replacement for the N900 but still.... there is none
Apr 29 15:10:01 <DocScrutinizer> there definitely is none, otherwise I'd not ponder to buy new N900 for even astronomic price tags
Apr 29 15:10:09 <villager> imho, capacitive touchscreens should be outlawed
Apr 29 15:10:15 <DocScrutinizer> indeed
Apr 29 15:11:09 <DocScrutinizer> maybe one day Apple will do us the favour and claim they got all the patents on c-ts so nobody else may use it X-P
Apr 29 15:11:38 <ThreeM> :)
Apr 29 15:11:45 <ThreeM> i like them
Apr 29 15:11:52 <ThreeM> on tablets
Apr 29 15:12:44 <DocScrutinizer> so the sane 90% of industry (duh, did I write that?) could switch to really nice stuff, like Stantum http://www.youtube.com/watch?v=Bv3X5y-ajtc
Apr 29 15:13:04 <villager> I'm not going to buy any tablet or anything with a capacitive screen... I even hate the laptop's touchpad
Apr 29 15:14:39 <villager> I bought a n9 for the principle of it, but as a smartphone I use it as little as possible, the n900 has a much more comfortable resistive touchscreen
Apr 29 15:15:25 <ThreeM> hmm what is more comfortable on a resistive ts?
Apr 29 15:16:11 <villager> everything, in my opinion... too bad the n900 doesn't have multitouch, but I think that's also possible on more advanced resistive screens
Apr 29 15:16:29 <DocScrutinizer> it's WAAAAY more precide, and works with arbitrary objects you touch it
Apr 29 15:16:37 <DocScrutinizer> precise*
Apr 29 15:16:57 <DocScrutinizer> villager: watch the youtube video!
Apr 29 15:17:13 <DocScrutinizer> you'll drool on your PC screen
Apr 29 15:17:16 <DocScrutinizer> :-D
Apr 29 15:18:26 <DocScrutinizer> (or n900 screen if you're watching it there :-D)
Apr 29 15:18:35 <villager> with resistive, I can hit anything I want, also actually see what I'm hitting since my finger doesn't have to be in the way, I can use gloves or a stylus, and I don't get aches from hours of electric currents through my fingers
Apr 29 15:20:49 <DocScrutinizer> c-ts has exactly one immanent advantage: it's more ruggedized by design. R-TS needs a somewhat flexible foil as cover, while C-TS can use even glass
Apr 29 15:21:53 <DocScrutinizer> on absolutely every other discipline C-TS sucks competing to R-TS
Apr 29 15:22:53 <DocScrutinizer> raindrops: R-TS check, C-TS BZZZZ
Apr 29 15:23:16 <DocScrutinizer> gloves, stylii: R-TS check, C-TS BZZZ
Apr 29 15:23:40 <DocScrutinizer> accuracy: R-TS check, C-TS BZZZZZZZZ
Apr 29 15:24:21 <DocScrutinizer> multitouch, R-TS check, C-TS check
Apr 29 15:27:05 <DocScrutinizer> even sensibility and well defined activation: R-TS check (when done properly), C-TS BZZZZ (either acts on not-really-rouches since too sensible, or it decides to not really accept a touch at all)
Apr 29 15:28:54 <DocScrutinizer> another nice one of Stantum: http://www.youtube.com/watch?v=HnQHoCDxiRw&NR=1, 1:43 *awesome*
Apr 29 15:32:44 <freemangordon> Pali, ping
Apr 29 15:32:52 <Pali> freemangordon, pong
Apr 29 15:33:17 <Pali> now Integrating autodetection of charger to kernel-power for bq2415x_charger driver
Apr 29 15:33:20 <freemangordon> Pali, all hal entries are filled by someone when bq module is insmoded
Apr 29 15:33:48 <DocScrutinizer> Pali: you checked it against hostmode and general USB?
Apr 29 15:34:08 <Pali> DocScrutinizer, I only export visible interface to bq module
Apr 29 15:34:22 <freemangordon> so I am not sure what this is supposed to do (besides sending dbus signals)
Apr 29 15:34:31 <DocScrutinizer> well, that's enough to already break hostmode I guess
Apr 29 15:34:31 <freemangordon> i mean hal-addon-bme
Apr 29 15:34:32 <Pali> like dmesg messages: twl4030_usb twl4030_usb: HW_CONDITIONS 0xd0/208; link 2
Apr 29 15:35:27 <Pali> freemangordon, what? someone is rewrting hal entries in hald-addnon-bme?
Apr 29 15:35:39 <freemangordon> yep
Apr 29 15:35:52 <Pali> but it is not possible
Apr 29 15:36:13 <Pali> did you replaced original hald-addon-bme?
Apr 29 15:36:22 <freemangordon> yep
Apr 29 15:36:26 <DocScrutinizer> I toldya >cat /sys/devices/platform/musb_hdrc/vbus< invokes a function in kernel musb-hdrc that breaks USB data session
Apr 29 15:36:43 <freemangordon> and the only thing I set so far is  battery.present 
Apr 29 15:36:47 <Pali> DocScrutinizer, I do not call ANY functions
Apr 29 15:37:08 <Pali> I only store informations from dmesg
Apr 29 15:37:09 <DocScrutinizer> actually not vbus
Apr 29 15:37:30 <DocScrutinizer> cat /sys/devices/platform/musb_hdrc/charge does
Apr 29 15:38:32 <DocScrutinizer> well, then you'll never learn about fastcharger, as the test in hw is only tiggered by reading this sysnode
Apr 29 15:38:59 <DocScrutinizer> and that mustn't be done after you enabled hostmode
Apr 29 15:39:41 * DocScrutinizer sighs and wishes some guys would talk to him a bit more
Apr 29 15:40:09 <Pali> DocScrutinizer, when you enable hostmode bq2415x driver going to manual mode
Apr 29 15:40:13 <freemangordon> Apli, http://pastebin.com/GPXR1KJZ
Apr 29 15:40:19 <freemangordon> Pali ^^^
Apr 29 15:40:33 <Pali> and then autodetection not working...
Apr 29 15:40:55 <DocScrutinizer> uhuh
Apr 29 15:41:06 <DocScrutinizer> and how do you detect "hostmode"?
Apr 29 15:41:42 <Pali> that user enable hostmode via sysfs entry...
Apr 29 15:42:12 <DocScrutinizer> yay, so I *really* wonder how you're going to solve those layering problems
Apr 29 15:42:53 <DocScrutinizer> you'll have to compile bq27200 as monolitic
Apr 29 15:43:00 <freemangordon> err, and what about hal addon?
Apr 29 15:43:24 <DocScrutinizer> otherwise I fail to see how you hook into the syfs node of musb_hdrc to detect hostmode
Apr 29 15:48:08 <DocScrutinizer> Pali: and honestly I fail to see how you'll make THAT kernel with bq24150 builtin to kernel compatible to stock rootfs that tries to start bme and will barf up when bme finds it can't conect to bq24150 anymore
Apr 29 15:49:00 <DocScrutinizer> you're definitely leaving maemo compatibility
Apr 29 15:49:46 <DocScrutinizer> you'll need userland patches to make KP work
Apr 29 15:50:03 <DocScrutinizer> or rather, make userland work with KP
Apr 29 15:50:04 <Pali> no, bq* drivers will be always modules
Apr 29 15:50:09 <Pali> not linked to kernel
Apr 29 15:50:29 <Pali> all will be working...
Apr 29 15:50:37 <DocScrutinizer> but HOW?
Apr 29 15:50:39 <Pali> without patching userspace
Apr 29 15:50:43 <Pali> hooks in board code
Apr 29 15:51:08 <Pali> from musb code is calling function in board code
Apr 29 15:51:28 <DocScrutinizer> hmmm
Apr 29 15:51:35 <Pali> and bq module register hook function from board code
Apr 29 15:51:55 <DocScrutinizer> sound story, though I never heard of it
Apr 29 15:52:01 <DocScrutinizer> might actually work
Apr 29 15:55:58 <DocScrutinizer> if that all pans out, it will be a nifty cute piece of software, and we'll dance on bme's grave :-D
Apr 29 15:58:30 <DocScrutinizer> actually incredible - you can run on "stock" config (original BME and no bq27200.ko), you can load bq27200 and still get bme running on LD_PRELOAD maemo_libbqioctl, plus you can nuke bme completely and run system on bq27200.ko plus bq27150.ko
Apr 29 15:59:31 <DocScrutinizer> I'd just really like to augment bq27200.ko to support *all* register the chip offers
Apr 29 15:59:40 <DocScrutinizer> in /sys
Apr 29 16:00:26 <freemangordon> and waht about battery calibration?
Apr 29 16:00:34 <freemangordon> is it auto-calibrated?
Apr 29 16:00:42 <DocScrutinizer> it never been
Apr 29 16:00:51 <DocScrutinizer> you need to run a learning cycle
Apr 29 16:01:02 <freemangordon> so, how is that going to work?
Apr 29 16:01:27 <DocScrutinizer> you charge completely, then discharge completely
Apr 29 16:01:55 <DocScrutinizer> when CI goes "0" your learning cycle got completed
Apr 29 16:03:04 <DocScrutinizer> at least that's what bq27x00 does
Apr 29 16:03:49 <freemangordon> the question was - how (and where ) is that going to be integrated?
Apr 29 16:03:51 <DocScrutinizer> alas it seems bq27200.ko tries to be overly smart and doesn't report anything when CI=1
Apr 29 16:04:29 <DocScrutinizer> what? I don't want my device to randomly decide to completely deplete battery to complete a learning cycle
Apr 29 16:04:44 <DocScrutinizer> it's user's very discretion to do that
Apr 29 16:05:11 <DocScrutinizer> so probably no "integration"
Apr 29 16:05:27 <freemangordon> but without at least one calibration, bq module does not work properly
Apr 29 16:05:51 <DocScrutinizer> you're free to write a calibrator app that stops charging until CI=0
Apr 29 16:06:05 <freemangordon> yeah
Apr 29 16:06:18 <DocScrutinizer> yeah, that'S another point why bq27200.ko is broken
Apr 29 16:06:28 <DocScrutinizer> alas it seems bq27200.ko tries to be overly smart and doesn't report anything when CI=1
Apr 29 16:06:39 <DocScrutinizer> [/quote]
Apr 29 16:07:08 <freemangordon> Pali, back to my question, what is hald-addon-bme supposed to do besides sending dbus signals
Apr 29 16:07:37 <DocScrutinizer> this is not only unfortunate, it's utterly wrong, as CI=1 doesn't mean the values are bogus, it just means they are not *guaranteed* anymore
Apr 29 16:08:06 <freemangordon> and one more thing, bme reports charge_level.current in unit bars, while bq module in mAh
Apr 29 16:08:42 <DocScrutinizer> for THAT you invent a nice function 100%->8bars :-D
Apr 29 16:08:56 <Pali> DocScrutinizer, bq27200.ko do not report values if battery is not calibrated
Apr 29 16:09:04 <freemangordon> but that is already set in hal 
Apr 29 16:09:11 <DocScrutinizer> it's basically up to you and your taste what this function actually does
Apr 29 16:09:31 <DocScrutinizer> Pali: yes, and that's WRONG
Apr 29 16:09:37 <freemangordon> DocScrutinizer, http://pastebin.com/GPXR1KJZ
Apr 29 16:09:46 <Pali> and what should be correct?
Apr 29 16:10:01 <Pali> paste is empty?
Apr 29 16:10:09 <DocScrutinizer> always report the values as seen in chip, only convert/normalize them
Apr 29 16:10:29 <freemangordon> is not
Apr 29 16:11:06 <DocScrutinizer> get a /sys/*/bq27200/CI_VALID that's either true or false, depending on CI
Apr 29 16:11:09 <Pali> ah, rekonq default kde borser is stupid...
Apr 29 16:11:35 <Pali> freemangordon, that hal property is filled by HAL itself
Apr 29 16:11:44 <freemangordon> yeah
Apr 29 16:11:56 <freemangordon> what am I supposed to do?
Apr 29 16:12:04 <Pali> ignore it
Apr 29 16:12:13 <Pali> or copy values to BME udi
Apr 29 16:12:24 <freemangordon> what?
Apr 29 16:12:37 <freemangordon> you mean to create a ned device?
Apr 29 16:12:41 <freemangordon> *new
Apr 29 16:12:57 <Pali> you need to have hal device with BME name
Apr 29 16:13:12 <freemangordon> hmm, let me check
Apr 29 16:13:18 <Pali> and hald-addon-bme exports more properties
Apr 29 16:13:47 <freemangordon> aah, ok, now i got it
Apr 29 16:17:03 <Pali> freemangordon, I think it can be better to copy values from /org/freedesktop/Hal/devices/computer_power_supply_battery_bq27200_0 to hald-addon-bme
Apr 29 16:17:07 <DocScrutinizer> Pali: (and what should be correct?) see output of http://maemo.cloud-7.de/maemo5/usr/local/sbin/bq27k-detail2. If all those values are showing up /sys/class/power_supply/bq27200-0 then the driver is somewhat complete, and it doesn't even need to bother about special treatment for CI=1
Apr 29 16:17:39 <freemangordon> Pali, does not kake sense, and I already made the parses :)
Apr 29 16:17:39 <DocScrutinizer> it's raher HAL/whoever that may want to report about CI status
Apr 29 16:17:56 <Pali> freemangordon, ok
Apr 29 16:18:19 <DocScrutinizer> s/report/care/
Apr 29 16:18:19 <infobot> DocScrutinizer meant: it's raher HAL/whoever that may want to care about CI status
Apr 29 16:18:37 <Pali> but power_supply interface does not support all those properties...
Apr 29 16:19:18 <DocScrutinizer> well, maybe that'S because bq27200 is not any kind of power-supply
Apr 29 16:19:39 <DocScrutinizer> it's a battery GAUGE
Apr 29 16:21:29 <DocScrutinizer> whatever, there's no reason I'm aware of why you can't get a /sys/class/power_supply/bq27200-0/CI node
Apr 29 16:22:31 <DocScrutinizer> as well as /sys/class/power_supply/bq27200-0/CYCL and  /sys/class/power_supply/bq27200-0/CYCT
Apr 29 16:22:42 <DocScrutinizer> and a few others that are missing
Apr 29 16:23:34 <DocScrutinizer> if HAL doesn't care about state of CI, what is most sensible thing to do? report capacity=0% always? or 100%?
Apr 29 16:24:04 <DocScrutinizer> or simply report the probably still somewhat sane values from chip, even when CI=1
Apr 29 16:26:12 <Pali> DocScrutinizer, ask that maintainers of battery-2.6.git tree
Apr 29 16:26:30 <DocScrutinizer> no, I'm not going to do that
Apr 29 16:27:01 <DocScrutinizer> as I'm not a meego guy that always answers "we're just using upstream"
Apr 29 16:28:20 <vdv> anybody knows, how modest merges several subsequent notifications into one? it increases number of notifications instead of showing multiple popus http://imagebin.org/210268 ( i've marked with red number of notifications)
Apr 29 16:28:56 <vdv> it's done here i think http://gitorious.org/community-ssu/modest/blobs/master/src/hildon2/modest-platform.c
Apr 29 16:29:03 <DocScrutinizer> I made my point - get the missing registers/flags into  /sys/class/power_supply/bq27200-0/*, don't fake reported values based on (flase) assumtions what e.g. CI means
Apr 29 16:29:20 <vdv> but i can't find exact piece of code which does that
Apr 29 16:31:01 <DocScrutinizer> If I had to guess, I'd say modest cancels the old notification and sets up a new one instead
Apr 29 16:31:47 <vdv> DocScrutinizer, most likely, but how to set that number?
Apr 29 16:31:58 <vdv> i.e. for new notification it's always 1
Apr 29 16:32:12 <DocScrutinizer> err, just concatenate it with your message string?
Apr 29 16:32:42 <DocScrutinizer> ooh, the notifier already has a dedicated integer parameter for that?
Apr 29 16:32:55 <vdv> which one?
Apr 29 16:32:59 <vdv> amount?
Apr 29 16:33:31 <DocScrutinizer> or no parameter at all, while still showing a number that's *not* in notifier call's msg string?
Apr 29 16:34:21 <vdv> yep
Apr 29 16:34:48 <vdv> i.e. that number 2 in screenshot is not part of msg string
Apr 29 16:34:57 <DocScrutinizer> I.E when you call the notification like notify("foobar") it actually displays "1 foobar" ?
Apr 29 16:35:06 <DocScrutinizer> mhm
Apr 29 16:35:10 <vdv> yes
Apr 29 16:35:35 <DocScrutinizer> so I guess it will increase that number for each subsequent call 
Apr 29 16:35:36 <vdv> and i can't find the way to set number of notifications
Apr 29 16:35:43 <villager> vfv: line 1709 maybe?
Apr 29 16:36:29 <vdv> villager, 1709? modest_platform_remove_new_mail_notifications?
Apr 29 16:36:47 <villager> vdv: yes looks to me like it cancels the old notifications
Apr 29 16:37:15 <vdv> villager, that is clear for me
Apr 29 16:37:38 <vdv> i understand that modest cancels previous notification before displaying a new one
Apr 29 16:38:04 <vdv> but how it sets "number of notifications" for new notification?
Apr 29 16:38:25 <vdv> villager, look at screenshot
Apr 29 16:38:39 <vdv> number marked with red
Apr 29 16:39:05 <vdv> how can i set that number for new notificaitons?
Apr 29 16:39:22 <vdv> i can change title, icon, text
Apr 29 16:39:34 <vdv> but that number is always 1 for me
Apr 29 16:40:04 <villager> vdv: is it line 1660?
Apr 29 16:40:13 <DocScrutinizer> nevertheless it seems to me there's no parameter for number of msgs to display in notifier
Apr 29 16:40:31 <DocScrutinizer> g_object_get(G_OBJECT (notification), "id", &notif_id, NULL);  notifications_list = g_slist_prepend (notifications_list, GINT_TO_POINTER(notif_id));
Apr 29 16:40:39 <DocScrutinizer> is probably doing the trick
Apr 29 16:41:14 <vdv> villager, i've tried that, but it has no effect
Apr 29 16:41:34 <vdv> i.e. amount hint does something else
Apr 29 16:43:23 <vdv> DocScrutinizer, that code just adds a notification to list, not?
Apr 29 16:43:45 <DocScrutinizer> /* Set the number of new notifications */  notify_notification_set_hint_int32 (NOTIFY_NOTIFICATION (notification) "amount", g_list_length (URI_list));
Apr 29 16:43:58 <villager> vdv: ok, what happens if you do open more than one hildon notification object?
Apr 29 16:44:18 <DocScrutinizer> (1659)
Apr 29 16:44:48 <vdv> DocScrutinizer, i also thought that it is exactly what i need
Apr 29 16:44:52 <vdv> but it's not
Apr 29 16:45:11 <DocScrutinizer> I'd bet it is
Apr 29 16:45:13 <vdv> villager, i get two popus
Apr 29 16:45:33 <vdv> *popups
Apr 29 16:45:49 <DocScrutinizer> vdv: probably only if you use a new  "id", &notif_id,
Apr 29 16:45:56 <villager> vdv: according to the code, modest doesn't destroy the old popups, so maybe some of the parameters somehow convince hildon that it's from the same app
Apr 29 16:46:28 <DocScrutinizer> id does that, in my world
Apr 29 16:46:56 <vdv> DocScrutinizer, do you see also, whether new notification is created with old id?
Apr 29 16:47:26 <DocScrutinizer> not yet, I only looked cursory at the code
Apr 29 16:47:27 <villager> DocScrutinizer: that code gets the id and stores it internally, it doesn't set it, I think
Apr 29 16:47:50 <vdv> villager, yes
Apr 29 16:48:01 <vdv> villager, with amout hint i do see that amount
Apr 29 16:48:04 <vdv> but
Apr 29 16:48:20 <DocScrutinizer> I think if it doesn't reuse it, either for nuking the notification, or for sending an "update", it's a rather silly thing to store the id at all. no?
Apr 29 16:48:22 <vdv> not on the place which i've pointed on screenshot
Apr 29 16:48:45 <villager> DocScrutinizer: it nukes them from another function, but not sure when exactly
Apr 29 16:49:39 <villager> I don't use modest so not sure when it usually does that
Apr 29 16:49:43 <vdv> villager, i see the setted number of notifications in small popup which appears at the top near the clock
Apr 29 16:49:52 <DocScrutinizer> simple approach: run modest under gdb, see what it actually does
Apr 29 16:50:22 <DocScrutinizer> when increasing the number of msgs in notification
Apr 29 16:51:13 <vdv> ok, thanks anyway for hints
Apr 29 16:52:49 <DocScrutinizer> or, even simpler, use dbus-monitir
Apr 29 16:53:14 <DocScrutinizer> as I gaher the notifications are handled via dbus interface by hildon-*
Apr 29 16:53:18 <DocScrutinizer> gather*
Apr 29 17:02:04 <DocScrutinizer> anyway, Set the number of new notifications */  notify_notification_set_hint_int32 (NOTIFY_NOTIFICATION (notification) "amount", g_list_length (URI_list));  is just clear enough
Apr 29 17:04:29 <vdv> yes, but it doesn't work, i've tried
Apr 29 17:08:08 <vdv> DocScrutinizer, do you also understand comment on line 1644?
Apr 29 17:08:33 <vdv> it says it sets the led pattern, but hint "time" is setted
Apr 29 17:09:39 <DocScrutinizer> 1651
Apr 29 17:09:52 <vdv> DocScrutinizer, is that normal, that i have two processes named modest in output of ps ax?
Apr 29 17:10:21 <vdv> DocScrutinizer, yes, but comment is above another line
Apr 29 17:10:44 <vdv> not 1651, which actually sets led pattern
Apr 29 17:15:57 <villager> it's not that mysterious... programmers shuffle code around without remembering to update the comments all the time
Apr 29 17:17:53 <villager> many programmers seem to have developed comment-blindness
Apr 29 17:18:14 <vdv> villager, correct, that's why i'm not trusting to other comments too
Apr 29 17:45:44 <Pali> patched kernel with autocharger mode for bq2415x not working...
Apr 29 17:45:53 <Pali> it boots to bootmenu
Apr 29 17:46:08 <Pali> and after booting real /sbin/init black screen and reboot
Apr 29 17:46:24 <Pali> giving up today
Apr 29 17:46:37 <Pali> patch on top of kernel-power is here: http://atrey.karlin.mff.cuni.cz/~pali/bq2415x/0001-Added-bq2415x-driver.patch
Apr 29 17:47:32 <Pali> (driver is really blacklisted)
Apr 29 17:53:36 <Pali> in /dev/mtd2 is no error output (all dumps are for other kernel version strings)
Apr 29 17:56:41 <Itschue> hello
Apr 29 17:57:02 <freemangordon> Pali, what is POWER_SUPPLY_CAPACITY?
Apr 29 17:57:14 <Pali> capacity in %
Apr 29 17:57:17 <Itschue> after flashing emmc and than flashing rootsfs device is booting
Apr 29 17:57:26 <freemangordon> you mean the charge in %?
Apr 29 17:57:40 <Pali> yes
Apr 29 17:57:43 <Itschue> i think all is ok but i will be 100%sure if all is really correct   how can i check this
Apr 29 17:57:46 <freemangordon> ok, thanks
Apr 29 17:59:14 <Itschue> ca someone please help me
Apr 29 18:02:17 <DocScrutinizer> vdv: the first uint32 you're passing to the call is 0 in this example: http://wiki.maemo.org/Phone_control#Make_an_.22Email_Style.22_notification_dialog
Apr 29 18:03:04 <DocScrutinizer> the point is it's NOT 0 on system's reply to the call, instead it holds a handle to the notification
Apr 29 18:03:43 <DocScrutinizer> if you call same dbus call again, with this very ID / handle, it does NOT open a new notifier
Apr 29 18:03:57 <Itschue> docscutinizer is this for me?
Apr 29 18:04:05 <DocScrutinizer> alas I was not able to set the amount value in dbus-send
Apr 29 18:04:20 <DocScrutinizer> Itschue: to vdv
Apr 29 18:04:37 <Itschue> ok did you read my question ?
Apr 29 18:04:55 <vdv> DocScrutinizer, hmm
Apr 29 18:05:09 <vdv> i've tried that before, as i was in python
Apr 29 18:05:23 <vdv> don't remember whether i worked or not
Apr 29 18:05:34 <vdv> now i'm on c++/qt
Apr 29 18:05:49 <DocScrutinizer> vdv: so I'd bet when you set BOTH the "amount" AND the "id" it shall just add the number of msgs
Apr 29 18:06:08 <vdv> and seems c/c++ api doesn't allow to set id
Apr 29 18:07:26 <DocScrutinizer> meh, obviously there's a way to do it, and modest does
Apr 29 18:09:04 <DocScrutinizer> Itschue: test the device, use it, see if there are problems
Apr 29 18:09:15 <DocScrutinizer> what do you expect to go awry?
Apr 29 18:10:12 <Itschue> to go awry? 
Apr 29 18:10:32 <DocScrutinizer> why are you not 100% sure ?
Apr 29 18:10:48 <DocScrutinizer> are you searching for problems?
Apr 29 18:11:06 <Itschue> the flasher doesnt say "device successfully flashed" after flashing emmc
Apr 29 18:11:15 <DocScrutinizer> hmmm
Apr 29 18:11:23 <DocScrutinizer> does it return?
Apr 29 18:11:42 <DocScrutinizer> does it print an error message?
Apr 29 18:11:47 <Itschue> no
Apr 29 18:11:55 <Itschue> no error message
Apr 29 18:12:04 <DocScrutinizer> then I bet everything is fine
Apr 29 18:12:43 <DocScrutinizer> just check content of MyDocs, if there is something left over from before you flashed
Apr 29 18:12:43 <Itschue> hmm but i have the same problem as before flashing
Apr 29 18:13:05 <DocScrutinizer> *sigh*
Apr 29 18:13:20 <DocScrutinizer> will you ever tell us what that problem is?
Apr 29 18:15:25 <Itschue> yes first mydocs looks fine there are no files from me
Apr 29 18:15:27 <DocScrutinizer> I guess your original question wasn't " i will be 100%sure if all is really correct   how can i check this" anyway
Apr 29 18:16:13 <DocScrutinizer> so why don't you simply state your problem, tell us you just flashed and that didn't help?
Apr 29 18:16:25 <Itschue> problem is i start application mannager and when he check for updates it takes very long time and he cant connect to all repos
Apr 29 18:16:40 <DocScrutinizer> that's normal
Apr 29 18:17:26 <DocScrutinizer> sometimes some of the repos are simply not online, or messed up. And HAM err Hildon App Manager is notorious for being *very* slow
Apr 29 18:17:43 <DocScrutinizer> you shouldn't keep extras-devel enabled by al means
Apr 29 18:18:03 <Itschue> i had problems to connecting to the nokia repos
Apr 29 18:18:18 <Pali> aaaaaa, I forgot to run depmod - reason why kernel does not boot
Apr 29 18:18:26 <DocScrutinizer> haha
Apr 29 18:18:27 <Pali> stupid
Apr 29 18:18:33 <DocScrutinizer> ;-P
Apr 29 18:18:43 <Itschue> i know tha ham is very slow
Apr 29 18:18:49 * DocScrutinizer prods Pali
Apr 29 18:18:50 <Itschue> but this problem i never had
Apr 29 18:19:12 <DocScrutinizer> Itschue: see my explanation above
Apr 29 18:19:43 <DocScrutinizer> sometimes the repos are messed up, try again tomorrow, if problem persists, come back for getting help
Apr 29 18:19:55 <Itschue> when i decativate extras devel ham connect to nokia repos?
Apr 29 18:20:09 <Itschue> problem persists since yesterday
Apr 29 18:20:10 <DocScrutinizer> reposa are known for acting up occasionally, especially on weekends
Apr 29 18:20:54 <DocScrutinizer> have you changed anything yesterday?
Apr 29 18:21:30 <DocScrutinizer> if not, then wait for problem to vanish tomorrow, just like it apeared yesterday
Apr 29 18:21:48 <Itschue> i also did apt-get update in terminal. there its very slow to. on some repo links was an "ign" 
Apr 29 18:22:13 <DocScrutinizer> yes, see above
Apr 29 18:23:39 <DocScrutinizer> "ign" probably means "ignored" -> problems with the repo. If you haven't changed anything in HAM catalog settings, it's a problem on internet side. Nothing you can do
Apr 29 18:24:30 <Itschue> can it be a problem with the router or the internet ip address
Apr 29 18:24:55 <DocScrutinizer> nope, as long as you can get any connection to internet from device
Apr 29 18:24:56 <Itschue> because im by my aunt and not @home. @home i never had problems like this
Apr 29 18:25:15 <DocScrutinizer> can you use browser?
Apr 29 18:25:42 <SpeedEvil> I've not been using my n900 to share wifi. What's the recommended 3G-wifi bridge? I want to connect my cheap android tablet to it.
Apr 29 18:25:56 <Itschue> can start browse and browser loading inet site too
Apr 29 18:26:12 <DocScrutinizer> SpeedEvil: mameo hotspot?
Apr 29 18:26:30 <DocScrutinizer> Itschue: repo defect, see above
Apr 29 18:26:42 <DocScrutinizer> wait for Nokia fixing it
Apr 29 18:27:19 <SpeedEvil> k
Apr 29 18:27:45 <DocScrutinizer> (if it's related to your aunt's WLAN you probably can't fix it anyway, so better just wait for monday)
Apr 29 18:28:30 <Itschue> tommorrow im at home problem is here by my aunt the mbile network isnt good
Apr 29 18:28:52 <DocScrutinizer> yes, we all understood quite well
Apr 29 18:29:01 <Itschue> or ist it an unfortunate coincidence
Apr 29 18:29:06 <DocScrutinizer> the advice won't change though
Apr 29 18:34:14 <Itschue> :(
Apr 29 18:36:56 <DocScrutinizer> what does HAM show when you select "catalogs" in menu?
Apr 29 18:37:56 <DocScrutinizer> some of the repos are probably marked as having problems, you can check the details of the problem when you click on it 
Apr 29 18:41:51 <Itschue> mom docscrutnizer i must wait until i can see it on the n900
Apr 29 18:43:30 <DocScrutinizer> anyway since you reflashed and thus the settings for the repos are quite probably ok, there's basically nothing we can do for you
Apr 29 18:44:31 <DocScrutinizer> no matter if Nokia's repo server or your auntie's WLAN is the culprit
Apr 29 18:44:38 <Itschue> its curious now there isnt any error
Apr 29 18:45:15 <Itschue> i think its my aunties wlan its an speedport router from german telekom :(
Apr 29 18:45:20 <Itschue> i hate these routers
Apr 29 18:45:40 <Itschue> at home i have a good linksys router
Apr 29 18:45:55 <DocScrutinizer> so how can we help? send some punching ball to treat?
Apr 29 18:46:31 <Itschue> now there were no errors shon in ham
Apr 29 18:46:33 <DocScrutinizer> I'm not inclined to discuss pros and cons of telekom branded speedport routers
Apr 29 18:46:43 <Itschue> yesterday i had seen an error text in red
Apr 29 18:47:01 <DocScrutinizer> yes, we all have seen this, dozens of times
Apr 29 18:47:09 <Itschue> at the moment ham is checking updates 
Apr 29 18:47:41 <Itschue> i see it only along time ago because i typed the correct link
Apr 29 18:47:44 <DocScrutinizer> I'm getting tired
Apr 29 18:48:37 <Itschue> from my problem?
Apr 29 18:49:09 <DocScrutinizer> http://talk.maemo.org/archive/index.php/t-74220.html
Apr 29 18:49:16 <DocScrutinizer> yes
Apr 29 18:49:36 <Itschue> we stop it now
Apr 29 18:49:43 <Itschue> i check it tomorrow at home
Apr 29 18:50:15 <Itschue> an when i have the problem tomorrow at home than i have a big problem
Apr 29 18:50:26 <Itschue> i think at home will be all fine
Apr 29 18:59:09 <DocScrutinizer> I thought you just found it works even where you are right now?
Apr 29 18:59:41 <DocScrutinizer> [2012-04-29 20:47:09] <Itschue> at the moment ham is checking updates 
Apr 29 18:59:59 <DocScrutinizer> so why the heck should there any problems tomorrow?
Apr 29 19:00:08 <DocScrutinizer> should there be*
Apr 29 19:00:53 <Itschue|2> back
Apr 29 19:01:05 <DocScrutinizer> repo (or telekom DNS, or whatever) fixed itself, problem identified as transient non-issue
Apr 29 19:02:30 <DocScrutinizer> Pali: depmod helped?
Apr 29 19:02:41 <Pali> yes
Apr 29 19:02:52 <Pali> pluged, unpluged detection working
Apr 29 19:02:56 <DocScrutinizer> so it's "working"??? :-o
Apr 29 19:02:59 <Pali> and battery is charging :-)
Apr 29 19:03:01 <DocScrutinizer> WOW!
Apr 29 19:03:18 * DocScrutinizer hands Pali a lollipop
Apr 29 19:03:32 <Pali> now I must thinging about wall charger and usb charger correct detection
Apr 29 19:04:05 <DocScrutinizer> USB host charger is absolutely nontrivial, and clearly a userland thing
Apr 29 19:04:25 <DocScrutinizer> and probably so is fastcharger detection
Apr 29 19:04:48 <DocScrutinizer> since you want to check for fastcharger only if ENUM fails
Apr 29 19:05:15 <freemangordon> Pali, congrats
Apr 29 19:05:26 <freemangordon> :)
Apr 29 19:07:02 * DocScrutinizer just wonders how to swap hald-addon-bme during running system
Apr 29 19:07:18 <DocScrutinizer> to switch from bme to bq24150.ko
Apr 29 19:07:41 <DocScrutinizer> not exactly trivial
Apr 29 19:08:26 <freemangordon> if you find a way I will owe you a beer
Apr 29 19:08:49 <freemangordon> i am restarting the device everytime a need to test a new version :)
Apr 29 19:09:04 <DocScrutinizer> freemangordon: maybe your hald-addon-bme replacement needs a fallback to original implementation on certain signals that bq24150.ko doesn't provide the needed info
Apr 29 19:11:08 <DocScrutinizer> freemangordon: well you could try initctl restart hal
Apr 29 19:11:20 <freemangordon> hal restart does not help
Apr 29 19:11:34 <DocScrutinizer> but I'm afraid hal is so central a conponent that this will make whole device reboot
Apr 29 19:11:39 <freemangordon> well, i am dount /etc/init.d/hal restart, but it should be the same
Apr 29 19:11:50 <freemangordon> *doing
Apr 29 19:11:50 <DocScrutinizer> nope
Apr 29 19:12:02 <DocScrutinizer> maybe not
Apr 29 19:12:05 <freemangordon> sure? ok, let me try
Apr 29 19:13:25 <freemangordon> instant reboot :)
Apr 29 19:13:44 <DocScrutinizer> toldya
Apr 29 19:14:31 <DocScrutinizer> you'll have to implement this fallback all across your code
Apr 29 19:14:48 <DocScrutinizer> foo(){
Apr 29 19:15:24 <DocScrutinizer> if bq24150-works a;
Apr 29 19:15:38 <DocScrutinizer> else fallback-to-original;
Apr 29 19:15:57 <DocScrutinizer> }
Apr 29 19:28:56 <Itschue|2> docscrutinizer  "resolving host timed out"
Apr 29 19:29:18 <Macer> kind of sucks no longer being able to fully use my n900
Apr 29 19:29:25 <Macer> i'm using it like a damn nit lol
Apr 29 19:29:34 <Macer> what a waste
Apr 29 19:32:39 <SpeedEvil> :/
Apr 29 19:46:29 <vdv> DocScrutinizer, found the reason
Apr 29 19:46:49 <vdv> DocScrutinizer, title must be same
Apr 29 19:47:00 <vdv> title of the notification
Apr 29 19:47:12 <vdv> and same id also
Apr 29 19:48:36 <vdv> and "amount" hint
Apr 29 19:50:12 <DocScrutinizer> vdv: toldya
Apr 29 19:50:50 <DocScrutinizer> you send exactly same message again, just with updated ID and amount
Apr 29 19:52:14 <vdv> hmm.. just tested with different id's
Apr 29 19:52:18 <vdv> and it also works
Apr 29 19:52:33 <vdv> so, message title must be same
Apr 29 19:52:39 <vdv> and amount must be set to one
Apr 29 19:52:53 <DocScrutinizer> err, definitely not
Apr 29 19:53:03 <vdv> just tested
Apr 29 19:53:08 <vdv> i print out id's
Apr 29 19:53:14 <DocScrutinizer> amount is set as number of new msgs to add to the already signalled ones
Apr 29 19:53:20 <vdv> for two subsequent notifications
Apr 29 19:53:45 <vdv> DocScrutinizer, yes, amount is a number to add
Apr 29 19:54:03 <DocScrutinizer> if this doesn't work then there's a bug in this API and another one in modest
Apr 29 19:54:12 <vdv> if there was one notification, and i create a new notification with "amount = 2", then i get three
Apr 29 19:54:23 <DocScrutinizer> yep
Apr 29 19:54:42 <vdv> so, it works as you said
Apr 29 19:54:48 <vdv> just id's must not be same
Apr 29 19:54:56 <DocScrutinizer> strange
Apr 29 19:55:05 <vdv> i.e. they can be same or not
Apr 29 19:55:12 <vdv> no effect on result
Apr 29 19:55:13 <DocScrutinizer> I dunno what's that first integer in dbus msg
Apr 29 19:55:21 <DocScrutinizer> I just see it's an integer
Apr 29 19:55:27 <vdv> it's really id of the notification
Apr 29 19:55:36 <vdv> and maybe it's used in some way
Apr 29 19:55:58 <vdv> for example, to reference later to the specific notification
Apr 29 19:56:05 <vdv> hide it, or show it again
Apr 29 19:59:20 <DocScrutinizer> >>> interface.Notify('Notification', 555, 'control_bluetooth_paired', 'Testing 123', 'Hello World', [], {}, 0)
Apr 29 19:59:21 <DocScrutinizer> ##method call sender=:1.310 -> dest=:1.28 serial=19 path=/org/freedesktop/Notifications; interface=org.freedesktop.Notifications; member=Notify
Apr 29 19:59:23 <DocScrutinizer>    string "Notification" 
Apr 29 19:59:24 <DocScrutinizer>    uint32 555 
Apr 29 19:59:41 <DocScrutinizer> returns:
Apr 29 20:00:23 <DocScrutinizer> method return sender=:1.28 -> dest=:1.310 reply_serial=19
Apr 29 20:00:25 <DocScrutinizer>    uint32 18 
Apr 29 20:00:26 <DocScrutinizer> dbus.UInt32(18L)....
Apr 29 20:00:54 <DocScrutinizer> now I can send as many  interface.Notify('Notification', 18... as I want
Apr 29 20:01:03 <DocScrutinizer> it won't open new requester
Apr 29 20:01:23 <DocScrutinizer> so "18" is what I called handle or ID
Apr 29 20:07:51 <DocScrutinizer> vdv: http://paste.debian.net/166597/
Apr 29 20:10:05 <villager> sure it's the id, but you don't need to set it to the id of the previous notification, keep in mind that modest doesn't do that either
Apr 29 20:13:36 <freemangordon> Pali, after battery change bq module is reporting total nonsense POWER_SUPPLY_CHARGE_NOW=0
Apr 29 20:14:27 <Pali> freemangordon, try to cat /sys/class/..../bq2.../charge_now
Apr 29 20:15:00 <freemangordon> still 0
Apr 29 20:15:31 <freemangordon> WTF?
Apr 29 20:15:43 <DocScrutinizer> vdv: see my pastebin, it's wxactly what happened there - when I don't set the id to the returned value, it opens a new notifier
Apr 29 20:16:47 <DocScrutinizer> It opens a new notifier for any integre it put there, except when that integer is a handle to an existing notifier
Apr 29 20:17:30 <DocScrutinizer> ooops, villager ^^^
Apr 29 20:18:38 <DocScrutinizer> freemangordon: CI=1
Apr 29 20:18:58 <DocScrutinizer> I already elaborated on this being a bug in bq27200.ko
Apr 29 20:19:08 <freemangordon> so? no charge?
Apr 29 20:19:11 <freemangordon> meh
Apr 29 20:19:52 <DocScrutinizer> I actually elaborated about it in epic length
Apr 29 20:21:02 <DocScrutinizer> [2012-04-29 18:23:33] <DocScrutinizer> if HAL doesn't care about state of CI, what is most sensible thing to do? report capacity=0% always? or 100%?  [2012-04-29 18:24:03] <DocScrutinizer> or simply report the probably still somewhat sane values from chip, even when CI=1 
Apr 29 20:21:56 <DocScrutinizer> o/
Apr 29 20:22:01 <DocScrutinizer> off for dinner
Apr 29 20:24:21 <villager> DocScrutinizer: perhaps the way you're calling it, but that's almost certainly not the way vdv or modest is doing it... for example, the id would be set to 0 for new notifications, not some random number, and hints such as "amount" etc are set (not sure which hints vdv found were needed for this to work)
Apr 29 20:27:54 <vdv> DocScrutinizer, villager: in order to combine several notification objects into one (with different or same id's) i had to 1) set "amount" hint and 2) set same title for both
Apr 29 20:28:49 <villager> and the "amount" hint is not set in that pastebin thing
Apr 29 20:30:03 <villager> I think you would replace [], {} with [], {"amount": 1} or something
Apr 29 20:32:30 <villager> this is probably some dirty hack that nokia made anyway, not really clean design
Apr 29 20:33:42 <villager> ideally it probably should have worked like Doc says, but...
Apr 29 20:34:44 <DocScrutinizer51> but you're missing the point of what I said
Apr 29 20:35:04 <DocScrutinizer51> damn roaming
Apr 29 20:35:29 <DocScrutinizer51> villager: I'm not a fool, and I for sure wouldn't decide to instantiate a new dialog 'with any random number for ID'
Apr 29 20:35:34 <DocScrutinizer51> I'm well aware that 0 is the only valid 'id' value to request a new dialog
Apr 29 20:35:36 <DocScrutinizer51> but you're missing the point of what I said
Apr 29 20:35:39 <DocScrutinizer51> damn roaming
Apr 29 20:37:16 <DocScrutinizer51> and yes, probably we're supposed to set amount hint to a non-0 value
Apr 29 20:38:02 <vdv> DocScrutinizer, and maybe also use "update" feature
Apr 29 20:38:15 <DocScrutinizer51> just I was too lazy to find out about 'or something'
Apr 29 20:38:21 <vdv> it's possible to update text and icon of execsting notification
Apr 29 20:38:46 <DocScrutinizer51> sounds reasonable, yes
Apr 29 20:38:49 <villager> DocScrutinizer51: well, try that "amount" thing then
Apr 29 20:39:05 <DocScrutinizer51> why?
Apr 29 20:39:17 <DocScrutinizer51> I don't need the whole cruft
Apr 29 20:39:27 <DocScrutinizer51> I just checked it for vdv 
Apr 29 20:39:33 <vdv> :)
Apr 29 20:39:58 <vdv> villager, and i've already get what i wanted, actually
Apr 29 20:40:01 <villager> DocScrutinizer51: as otherwise your pastebin shows you've missed the point?
Apr 29 20:40:04 <vdv> got
Apr 29 20:40:14 <villager> well, doesn't matter, vdv is happy
Apr 29 20:40:16 <DocScrutinizer51> pfff
Apr 29 20:40:34 <villager> I'm not really in the arguing mood
Apr 29 20:40:34 <vdv> yep :)
Apr 29 20:40:45 <DocScrutinizer51> me neither
Apr 29 20:41:02 <vdv> thanks a lot
Apr 29 20:41:12 <DocScrutinizer51> yw
Apr 29 20:41:35 <DocScrutinizer51> was intriguing, I admit
Apr 29 20:42:14 <DocScrutinizer51> so I just did it to make sure my advice wasn't utter BS
Apr 29 20:53:58 <DocScrutinizer51> villager: I even could figure a sane developer would check for other details of msg when and only when id=0, to allow some error tolerance
Apr 29 20:55:32 <DocScrutinizer51> and of course such an API would handle id!=0 as 0 if no such id is actually in use
Apr 29 20:56:13 <DocScrutinizer51> that's why my 'random id' worked like I wouldve used 0
Apr 29 21:01:30 <freemangordon> DocScrutinizer, Pali, https://gitorious.org/hald-addon-bme/hald-addon-bme
Apr 29 21:01:40 <Pali> nice
Apr 29 21:01:53 <freemangordon> not complete of course
Apr 29 21:02:16 <DocScrutinizer51> sorry, on n900
Apr 29 21:02:31 <freemangordon> but as a start should be enough
Apr 29 21:05:18 <freemangordon> Pali, is there anything to to to force bq module to report non-zero charge (besides calibration cycle)?
Apr 29 21:05:31 <Pali> patch it
Apr 29 21:05:34 <freemangordon> s/to to to/to do to/
Apr 29 21:05:35 <infobot> freemangordon meant: Pali, is there anything to do to force bq module to report non-zero charge (besides calibration cycle)?
Apr 29 21:05:40 <freemangordon> yeah, nice
Apr 29 21:06:17 <freemangordon> But it seems this is a readout from the chip, ain't?
Apr 29 21:08:07 <merlin1991> hm silly question, how do I write a regex to check for a-z except for a few?
Apr 29 21:08:27 <DocScrutinizer51> freemangordon: nope
Apr 29 21:08:49 <merlin1991> is there a syntax for this caracter class minus xyz?
Apr 29 21:09:12 <DocScrutinizer51> it's an atrificial intentional bug introduced by oversmart bq27200.ko
Apr 29 21:09:22 <DocScrutinizer51> aiui
Apr 29 21:09:37 <DocScrutinizer51> chip doesn't do that nonsense
Apr 29 21:11:01 <DocScrutinizer51> merlin1991: either [^xyz]
Apr 29 21:11:15 <merlin1991> well that allows anything but xyz
Apr 29 21:11:27 <merlin1991> but I'd like todo a-z but not b e and o
Apr 29 21:11:49 <DocScrutinizer51> or [A-Wa-w]
Apr 29 21:11:49 <merlin1991> ofc I can type it out :D
Apr 29 21:12:25 <DocScrutinizer51> yes, you'll have to
Apr 29 21:12:48 <DocScrutinizer51> there's no set subtractor afaik
Apr 29 21:13:53 <merlin1991> :/
Apr 29 21:14:33 <DocScrutinizer51> [acdf-np-z]
Apr 29 21:15:23 <DocScrutinizer51> i dunno of any nicer term
Apr 29 21:19:18 <DocScrutinizer51> maybe inverted logic: [^_-Zbeo*-]
Apr 29 21:21:48 <DocScrutinizer51> I'm not sure sth like [^-k] would actually catch and forbid all from very beginning up to k
Apr 29 21:22:18 <DocScrutinizer51> I rather guess that forbids just - and k
Apr 29 21:24:00 <DocScrutinizer51> also I don't think there's any & analog to | in regex
Apr 29 21:27:02 <DocScrutinizer51> kde had a nice regex wizard, I think in konqueror multi-rename or sth
Apr 29 21:27:35 <DocScrutinizer51> merlin1991: ^
Apr 29 21:59:22 <Pali> freemangordon, DocScrutinizer, here bq2415x patch for kp50: http://atrey.karlin.mff.cuni.cz/~pali/bq2415x/0001-Added-bq2415x-driver.patch
Apr 29 21:59:36 <Pali> wallcharger detection does not working correctly yet
Apr 29 22:00:05 <Pali> need to call: cat /sys/devices/platform/musb_hdrc/charger
Apr 29 22:05:30 <DocScrutinizer> toldya
Apr 29 22:06:05 <DocScrutinizer> the detection function in musb and hw is triggered only when you cat that sysnode
Apr 29 22:06:33 <DocScrutinizer> anyway, nice work, worth even another lollipop :-D
Apr 29 22:07:42 <DocScrutinizer> (cat sysnode) a thing usually done by ke-recv aiui
Apr 29 22:08:51 <DocScrutinizer> or by bme itself, or knows-the-flying-spaghettimonszer-whom
Apr 29 22:08:58 <Pali> DocScrutinizer, it will work, but I need to wait 1-2s
Apr 29 22:09:11 <Pali> so some delay is needed
Apr 29 22:09:52 <DocScrutinizer> yeah, the delay to allow ke-recv do its job
Apr 29 22:10:11 <Pali> going offline now, bye all
Apr 29 22:10:16 <DocScrutinizer> or hal, or (see above)
Apr 29 22:10:25 <DocScrutinizer> bye Pali 
Apr 29 22:11:05 <vdv> what does category mean for a notification in maemo? http://developer-next.gnome.org/libnotify/0.7/NotifyNotification.html#notify-notification-set-category
Apr 29 22:11:21 <DocScrutinizer> prolly the icon used
Apr 29 22:11:40 <vdv> if i set "email-message" category for new notification i get behaviour of modest automatically
Apr 29 22:12:08 <vdv> behavior = led pattern, sound, notification counter
Apr 29 22:12:10 <DocScrutinizer> dang, don't ask me
Apr 29 22:12:12 <DocScrutinizer> :-D
Apr 29 22:12:48 <DocScrutinizer> maybe some sorta meta-macro
Apr 29 22:12:50 <vdv> it's very interesting where is that defined
Apr 29 22:13:03 <DocScrutinizer> God knows
Apr 29 22:13:12 <vdv> :)
Apr 29 22:13:33 <vdv> and whether i'm allowed to use it
Apr 29 22:13:47 <DocScrutinizer> "when I wrote that code, only God and me knew what I was doing - now only God knows"
Apr 29 22:14:34 <DocScrutinizer> o/
Apr 29 22:14:52 <vdv> i've killed modest, thought, that category behaviour is defined by modest at runtime
Apr 29 22:15:30 <vdv> DocScrutinizer, http://gitorious.org/community-ssu/modest/blobs/master/src/hildon2/modest-platform.c#line1630
Apr 29 22:16:21 <vdv> it's all like a magic
Apr 29 22:17:18 <vdv> meaning of "category" is not documented
Apr 29 22:22:27 <DocScrutinizer51> vdv: that's defined somewhere in the abyss of hildon-notification-daemon or sth
Apr 29 22:24:43 <vdv> it's interesting, whether it's standard, or added by modest
Apr 29 22:25:05 <vdv> and if uninstalling modest will remove that category too
Apr 29 22:57:16 <vdv> /etc/hildon-desktop/notification-groups.conf
Apr 29 23:19:29 <topobot> Are there any companies developing smartphones with maemo?
Apr 29 23:21:37 <GeneralAntilles> topobot, no. Maemo's not being developed commercially any longer.
Apr 29 23:21:58 <topobot> which is the last smartphone with maemo?
Apr 29 23:24:51 <SpeedEvil> The n900
Apr 29 23:25:05 <topobot> oh
Apr 29 23:25:07 <SpeedEvil> Well - arguably the n9
Apr 29 23:25:12 <topobot> is it good the n900? 
Apr 29 23:25:17 <SpeedEvil> Yes.
Apr 29 23:25:21 <SpeedEvil> But it's two years old.
Apr 29 23:25:21 <topobot> or is it better iphone?
Apr 29 23:25:27 <SpeedEvil> Closing on three.
Apr 29 23:25:54 <topobot> is there any benefit in comparation with iphone or galaxy?
Apr 29 23:26:18 <SpeedEvil> At the moment - for 99.9% of users, no.
Apr 29 23:26:44 <topobot> why?
Apr 29 23:28:39 <SpeedEvil> Because nokia has switched for the mmomet entirely away from linux, and is trying to push their windows handsets.
Apr 29 23:29:31 <topobot> and what anout that os based on maemo and intel?
Apr 29 23:30:17 <SpeedEvil> It fell apart.
Apr 29 23:30:19 <SpeedEvil> Meego.
Apr 29 23:30:22 <SpeedEvil> Gone.
Apr 29 23:30:28 <topobot> oh yes meego
Apr 29 23:30:33 <topobot> gone? why?
Apr 29 23:31:26 <topobot> if im developer and i want to develop a tool for smarphone, which is the best os should i use?
Apr 29 23:31:30 <SpeedEvil> Nokia became uninterested in the project.
Apr 29 23:32:00 <topobot> i want to sell my app
Apr 29 23:32:13 <SpeedEvil> If you do not have a spare ten or twenty million to invest in attempting to get a new plaform on the market, android, ios, or blackberry.
Apr 29 23:32:25 <SpeedEvil> You may also consider windows phone, of course.
Apr 29 23:33:33 <topobot> android or ios?
Apr 29 23:36:36 <topobot> hi?
Apr 29 23:37:40 <SpeedEvil> yes.
Apr 29 23:37:42 <SpeedEvil> one of them.
Apr 29 23:38:03 <SpeedEvil> The problem with either appstore is you are competing against tens of thousands of apps.
Apr 29 23:38:11 <SpeedEvil> Windows may be somewhat better in that respect.
Apr 29 23:38:28 <topobot> ummm
Apr 29 23:40:21 <topobot> maemo is linux>
Apr 29 23:40:22 <topobot> ?
Apr 29 23:40:34 <topobot> no android
Apr 29 23:40:42 <topobot> is android linux?
Apr 29 23:41:28 <nox-> their kernel is linux
Apr 29 23:41:44 <nox-> (patched...)
Apr 29 23:44:14 <SpeedEvil> Android is based on a linux kernel.
Apr 29 23:44:29 <SpeedEvil> You might consider developing for N9.
Apr 29 23:44:45 <SpeedEvil> Which isn't - quite - a dead platform, though nokia are trying hard to kill it.
Apr 29 23:47:05 <topobot> why nokia wants to kill n900?
Apr 29 23:48:38 <nox-> elop wants to push win phone... :/
Apr 29 23:48:52 <nox-> he killed n950 too
Apr 29 23:48:58 <topobot> elop?
Apr 29 23:49:05 <nox-> new nokia ceo
Apr 29 23:49:25 <topobot> no business with linux based stuff?
Apr 29 23:56:08 <SpeedEvil> There are no new linux based phones, other than android, which is sort-of-arguable.
Apr 30 00:03:39 <topobot> and which are the differences between maemo and android?
Apr 30 00:15:02 <SpeedEvil> Mameo is - compared to android - a 'normal' desktop linux
Apr 30 00:31:36 <topobot> so maemo is better than android?
Apr 30 01:24:29 <luke-jr> SpeedEvil: not really, it just has X11
Apr 30 01:24:53 <SpeedEvil> And libc, and runs binary apps, ...
Apr 30 01:25:06 <luke-jr> topobot: Maemo is better, because it's designed for normal computing, not just a super-phone
Apr 30 01:25:06 <SpeedEvil> there isn't for example - a JVM in the way of all apps
Apr 30 01:25:55 <topobot> what does it means "designed for normal computing"?
Apr 30 01:25:57 <robbiethe1st> I want to see a Kubuntu Mobile phone.
Apr 30 01:26:42 <robbiethe1st> topobot: you can do anything you can do with a normal Linux install with Maemo
Apr 30 01:27:10 <topobot> you mean, you can run linux software on maemo?
Apr 30 01:27:15 <robbiethe1st> Which includes custom kernel modules, USB host(including odd stuff like USB soundcards and serial adaptors), etc.
Apr 30 01:27:54 <robbiethe1st> Yes, though unless there's a version *built* for maemo, you have to run a chroot(like easy debian) to give the program the supporting files it needs.
Apr 30 01:27:59 <topobot> maemo is more hackeable than android?
Apr 30 01:28:03 <robbiethe1st> Far more.
Apr 30 01:28:19 <robbiethe1st> N900 is unlocked, both bootloader wise and carrier wise
Apr 30 01:28:36 <topobot> but n900 is slow ?
Apr 30 01:28:50 <Hurrian> not really.
Apr 30 01:28:51 <robbiethe1st> Somewhat, yeah. Stock 600mhz, single core.
Apr 30 01:29:12 <Hurrian> js-heavy webpages suck on microb, work fine in firefox with easy-debian
Apr 30 01:29:24 <topobot> what can you do with maemo that you cannot do with android?
Apr 30 01:29:34 <Hurrian> chrooting doesn't suck
Apr 30 01:29:36 <robbiethe1st> Uh... run Open Office?
Apr 30 01:29:47 <robbiethe1st> Run X-based apps natively
Apr 30 01:29:59 <Hurrian> robbiethe1st, you can also chroot under android, it's just that performance sucks.
Apr 30 01:30:09 <robbiethe1st> Because of no X server, just VNC
Apr 30 01:30:13 <robbiethe1st> right?
Apr 30 01:30:24 <Hurrian> to put it into perspective, launching OOo is as fast as desktop on N900
Apr 30 01:30:34 <Hurrian> it will take ~2 minutes on my Tegra 2 tablet
Apr 30 01:31:05 <topobot> i was wondering if its possible to connect a secondary camera with a n900?
Apr 30 01:31:24 <joga> does android do usb host?
Apr 30 01:31:47 <robbiethe1st> topo, yeah, I bet it is possible to use a USB webcam with it.
Apr 30 01:31:55 <SpeedEvil> joga: In some circumstances, yes.
Apr 30 01:32:10 <SpeedEvil> joga: It can suport various devices. For example, mass storage is common
Apr 30 01:32:12 <robbiethe1st> You might have to use EasyDebian to find software that'll /use/ the usb camera, though.
Apr 30 01:33:03 <robbiethe1st> Personally, I've used USB mode to run a flash drive, USB soundcard(giving me SPDIF out), and USB->Serial adaptor for interfacing with a CNC machine)
Apr 30 01:33:10 <joga> I used a pololu servo controller with n900 (+ usb adapter)
Apr 30 01:33:38 <robbiethe1st> Cool
Apr 30 01:34:02 <joga> also you can use an arduino
Apr 30 01:34:20 <joga> or anything that you can talk serial to via usb
Apr 30 01:35:20 <joga> also plugging keyb+mouse is a nice feat :) maybe with tv-out you could do word processing in easy-debian oowriter hehe
Apr 30 01:44:00 <Hurrian> joga, iirc the usb serial driver is in the maemo kernel
Apr 30 01:44:03 <Hurrian> so yeah, it should work
Apr 30 02:30:46 <LaoLang_cool> Any tool to allow me navigate quickly to top/bottom?
Apr 30 02:32:32 <robbiethe1st> Uh, LaoLang_cool ?
Apr 30 02:34:30 <LaoLang_cool> robbiethe1st: sorry for my poor English, say I'm viewing a mms in fmms, and I'm in the middle of the long page, it's hard to go to the top/bottom of that page
Apr 30 02:34:47 <robbiethe1st> try shift + up/down?
Apr 30 02:34:53 <robbiethe1st> perhaps alt+up/down
Apr 30 02:35:17 <LaoLang_cool> tried it, nothing happened..
Apr 30 02:37:47 <robbiethe1st> darn
Apr 30 02:37:53 <robbiethe1st> Well, I don't use it myself, so..
Apr 30 02:38:33 <LaoLang_cool> robbiethe1st: actually, I'm needing a universe way to go to top/bottom of a page on maemo :)
**** ENDING LOGGING AT Mon Apr 30 02:59:58 2012