**** BEGIN LOGGING AT Mon May 01 03:00:01 2017
May 01 11:37:29 <Enrico_Menotti> I have Devuan running on my n900 fine. Boot fast. No udev. Power button working. Screen going to sleep. Keyboard set up. Wifi. Ssh.
May 01 11:37:53 <Enrico_Menotti> I'd need some help with the following steps:
May 01 11:38:14 <Enrico_Menotti> 1. How to check whether the battery is charging? If not, what is needed to do that?
May 01 11:38:43 <Enrico_Menotti> 2. How to have the right switch working to wake up the screen? (I have scpi-support-base installed.)
May 01 11:38:59 <Enrico_Menotti> 3. What is the present state of the art about desktop environments?
May 01 11:46:21 <Maxdamantus> My plan is to just modify dwm or i3 or something.
May 01 11:46:40 <Maxdamantus> (for the desktop environment)
May 01 11:46:54 <Maxdamantus> well, for a window manager at least, which should be sufficient.
May 01 11:48:39 <bencoh> Enrico_Menotti: 1. see the bq27xx driver (probably located under /sys/class/power_supply), or the bq27*.sh script
May 01 11:50:54 <bencoh> oh err, which kernel are you using btw?
May 01 11:51:06 <bencoh> (and why no udev?)
May 01 12:03:39 <Enrico_Menotti> bencoh As discussed with parazyd , udev is not needed for hotplug. And it was creating problems at boot - a 30 seconds delay, with net.agent timing out due to the loopback lo no being mounted, which in turn is a consequence of the missing initrd (I boot directly to rootfs). Hope I don't miss here any step - many details, I don't remember exactly everything.
May 01 12:04:16 <Enrico_Menotti> I'm using mainline kernel with Pali's rx51_defconfig, modified to include the two watchdogs in the kernel and not as modules.
May 01 12:05:30 <Enrico_Menotti> Maxdamantus You talk about a "plan". I'd deduce there is no DE running yet. Am I right?
May 01 12:05:54 <Maxdamantus> I don't know. It's just a personal plan.
May 01 12:07:59 <Maxdamantus> Just because I'm not really interested in most of the things hildon seems to do, and I suspect I can just modify an existing tiling WM to do the things I would actually be interested in.
May 01 12:09:22 <Maxdamantus> (and the ones I mentioned would probably use a lot less memory than hildon)
May 01 12:12:58 <Enrico_Menotti> Maxdamantus Ok. Thanks.
May 01 12:27:24 <Wizzup> Enrico_Menotti: the kernel should manage the charging stuff by itself, but you can read the currnet state from /sys
May 01 12:27:41 <Wizzup> /sys/devices/platform/68000000.ocp/48072000.i2c/i2c-2/2-0055/power_supply/bq27200-0/
May 01 12:28:22 <Wizzup> as for making specific keys wake up the screen/phone, do you suspend/hibernate-ram he phone, or just want X turned off?
May 01 12:28:39 <Wizzup> I am pretty sure the switch on the side is just an input device, so it should not be too hard to make it wake up
May 01 12:28:54 <Wizzup> I don't have my gentoo & devuan n900s next to me right now, but it should not be too hard to make that happen
May 01 12:29:05 <Wizzup> As for 3, hildon? :) 
May 01 12:31:05 <Maxdamantus> in the normal maemo system it's a gpio called "slide"
May 01 12:31:14 <Maxdamantus> dunno how the names are assigned.
May 01 12:32:34 <Wizzup> I just usuall do tail -f /dev/input/event* and see which one activates
May 01 12:32:52 <Enrico_Menotti> Wizzup Thanks. I'm trying to read the battery state with the help of a couple of scripts suggested by DocScrutinizer05 . (If I'm able to get them work; they have a few dependences).
May 01 12:33:00 <Wizzup> also, for led fun - 
May 01 12:33:01 <Wizzup> for i in $(ls /sys/devices/platform/68000000.ocp/48072000.i2c/i2c-2/2-0032/leds/lp5523:*/trigger /sys/devices/platform/68000000.ocp/48072000.i2c/i2c-2/2-0032/leds/lp5523:kb*/trigger) ; do echo timer > $i; done
May 01 12:33:14 <Maxdamantus> It's not handled by /dev/input/*
May 01 12:33:20 <Maxdamantus> it's just gpio.
May 01 12:33:20 <Wizzup> Enrico_Menotti: with the current setup, even normal battery monitors will pick it up
May 01 12:33:24 <Wizzup> Maxdamantus: mainline
May 01 12:33:50 <Wizzup> Enrico_Menotti: turned off wiht for i in $(ls /sys/devices/platform/68000000.ocp/48072000.i2c/i2c-2/2-0032/leds/lp5523:*/trigger /sys/devices/platform/68000000.ocp/48072000.i2c/i2c-2/2-0032/leds/lp5523:kb*/trigger) ; do echo none > $i; done
May 01 12:33:56 <Maxdamantus> Eh, so mainline has an input driver for the GPIO switches?
May 01 12:34:02 <Wizzup> Enrico_Menotti: e.g. 'batti' will work
May 01 12:34:12 <Wizzup> Maxdamantus: I would not be surprised, is this not the point of dts and such
May 01 12:34:20 <Maxdamantus> I don't think so.
May 01 12:34:24 <Wizzup> otherwise, if it is just gpio, it's also easy to open+select() the right gpio
May 01 12:34:38 <Wizzup> well, I am not in a position to check atm.
May 01 12:34:38 <Maxdamantus> Right, I think that's the standard way to interact with GPIO.
May 01 12:35:10 <Wizzup> many keyboards are also gpio, but they have drivers and are exposed not via gpio in sys
May 01 12:35:21 <Maxdamantus> I have a funny program I wrote that polls the status of the shutter GPIO.
May 01 12:35:22 <Wizzup> same for simple buttons on many arm dev boards
May 01 12:35:48 <Wizzup> the input layer is a perfect place for these things
May 01 12:36:00 <Wizzup> e.g. closing/opening a laptop will typically also trigger something in /dev/input/
May 01 12:36:29 <Maxdamantus> That seems weird to me.
May 01 12:36:48 * Maxdamantus might try it sometime with his laptop.
May 01 12:37:11 <Maxdamantus> maybe there'd be something on the trackpad device to indicate that it's active/inactive, dunno.
May 01 12:37:12 <Wizzup> you can run evdev on all of htem
May 01 12:37:16 <Wizzup> it will print all the keys they expose
May 01 12:37:24 <Wizzup> evtest*
May 01 12:37:59 <Wizzup> Input device name: "Lid Switch"
May 01 12:38:13 <Wizzup> Event: time 1493642283.711271, type 5 (EV_SW), code 0 (SW_LID), value 1
May 01 12:38:13 <Wizzup> Event: time 1493642283.711271, -------------- SYN_REPORT ------------
May 01 12:38:13 <Wizzup> Event: time 1493642286.611207, type 5 (EV_SW), code 0 (SW_LID), value 0
May 01 12:38:13 <Wizzup> Event: time 1493642286.611207, -------------- SYN_REPORT ------------
May 01 12:38:51 <Wizzup> similarly, for mincrophone inserts:     Event code 4 (SW_MICROPHONE_INSERT) state 0
May 01 12:42:03 <DocScrutinizer05> would depend on DT I guess
May 01 12:42:56 <DocScrutinizer05> yes, all sensors should show up as /dev/*, in maemo however they usually are only in /sys
May 01 12:43:26 <Maxdamantus> anyway, https://gist.github.com/Maxdamantus/23783a33d9ec43e8be42632566e8d1f9
May 01 12:43:30 <Maxdamantus> Who needs malloc?
May 01 12:45:05 <DocScrutinizer05> any unlike Vol+/- the cam trigger, lockswitch, slider, proxy and camdoor are not input devices
May 01 12:45:35 <DocScrutinizer05> no idea how that got inplemented in devuan aka mainline kernel
May 01 14:43:58 <Enrico_Menotti> About the slide key, there is a module called gpio_keys. I modprobed it. Now the slide key is alive, but it only produces a ^@ symbol on the screen.
May 01 14:47:52 <bencoh> which might be okay actually
May 01 14:51:43 <Enrico_Menotti> Ok, so now how to handle that?
May 01 14:52:50 <bencoh> gpiokeys are supposed to output events to /dev/input/event*
May 01 14:53:37 <bencoh> a glance at /sys/class/input might tell you which one
May 01 14:54:10 <DocScrutinizer05> but X11 isn't supposed to read input form that device, I guess
May 01 14:54:22 <bencoh> /dev/input/event* follow a simple read api
May 01 14:54:31 <bencoh> (read() syscall I mean)
May 01 14:54:36 <bencoh> DocScrutinizer05: sure it does
May 01 14:54:41 <bencoh> it isa
May 01 14:54:49 <bencoh> is* damnit
May 01 14:55:23 <DocScrutinizer05> err, well
May 01 14:56:09 <bencoh> dunno how the switch button was implemented .... I suppose it's a simple press/release key
May 01 14:56:33 <bencoh> (gpiokeys are basically that anyway ... press and release events with associated keycode)
May 01 14:56:47 <DocScrutinizer05> sure
May 01 14:57:05 <bencoh> on maemo mce is responsible for it iirc
May 01 14:57:07 <DocScrutinizer05> anyway maemo doesn't show an event for lockswitch
May 01 14:57:14 <DocScrutinizer05> yes
May 01 14:57:20 <DocScrutinizer05> that's what I said
May 01 14:58:19 <DocScrutinizer05> and mce opens a file handle to whatever-that-gpioswitch directly instead of listening to X11 events
May 01 14:58:50 <Enrico_Menotti> The power button works just fine without modules, just installed acpi-support-base (Devuan).
May 01 14:59:20 <DocScrutinizer05> mce         824       root   10r   REG        0,0     4096       4101 /sys/devices/platform/gpio-switch/kb_lock/state
May 01 15:00:10 <DocScrutinizer05> http://paste.ubuntu.com/24493079
May 01 15:01:42 <DocScrutinizer05> it's questionable if you want a ^@ input event in X11 for lockswitch
May 01 15:02:43 <DocScrutinizer05> I guess it *could* get implemented that way, but feels somewhat... wrong
May 01 15:04:06 <DocScrutinizer05> at very least it should provide the correct keysym
May 01 15:06:47 <DocScrutinizer05> though don't ask me what's the name
May 01 15:08:34 <DocScrutinizer05> can't find something reasonable in http://wiki.linuxquestions.org/wiki/XF86_keyboard_symbols
May 01 15:09:08 <DocScrutinizer05> maybe XF86ScreenSaver
May 01 15:13:25 <DocScrutinizer05> anyway that's keysym, not scancode. First of all you'd need to make sure the lockswitch generates a unique scancode
May 01 15:16:29 <DocScrutinizer05> Enrico_Menotti: please run xev and pastebin the lockswitch event
May 01 15:17:19 <Enrico_Menotti> xev: command not found. Let me investigate.
May 01 15:19:16 <Enrico_Menotti> It's package x11-utils. A moment.
May 01 15:22:21 <DocScrutinizer05> freaking... WTF?! xev doesn't report events from N900 hw kbd (except for vol+/- aka F6/7(?)) when started via ssh
May 01 15:23:58 <DocScrutinizer05> whatever, busy. AFK, bbl
May 01 15:24:11 <Enrico_Menotti> xev:  unable to open display ''
May 01 15:24:24 <Enrico_Menotti> ?
May 01 15:24:50 <DocScrutinizer05> no X11 environment and/or no $DISPLAY defined
May 01 15:25:34 <Enrico_Menotti> Yes, no X11 environment. So how to use that?
May 01 15:25:39 <DocScrutinizer05> IroN900:~# echo $DISPLAY 
May 01 15:25:41 <DocScrutinizer05> :0.0
May 01 15:26:05 <Enrico_Menotti> Ok, I'll try to define it.
May 01 15:26:19 <DocScrutinizer05> doesn't help if it doesn't exist
May 01 15:26:45 <Enrico_Menotti> In fact.
May 01 15:27:37 <Enrico_Menotti> bbl (maybe)
May 01 15:28:25 <DocScrutinizer05> http://paste.ubuntu.com/24493190
May 01 15:29:58 <Enrico_Menotti> I just have event0 - 3.
May 01 15:30:52 <DocScrutinizer05> http://paste.ubuntu.com/24493197
May 01 16:09:06 <halftux> sicelo: I guess I found the solution for irssi which you were looking for. https://talk.maemo.org/showpost.php?p=1527540&postcount=10
May 01 17:25:48 <vahe> hi all :) I'm looking for bitcoin wallet for n900
May 01 17:55:55 <DocScrutinizer05> https://talk.maemo.org/showthread.php?t=66756
May 01 18:56:30 <sicelo> merlin1991: halfux's solution worked, yay! https://garage.maemo.org/pipermail/extras-cauldron-builds/2017-May/156638.html
May 01 18:57:27 <sicelo> Vajb: looks like perl irssi is coming to repos soon :)
May 01 19:00:56 <buZz> they ported irssi to 100% perl?
May 01 19:01:50 <parazyd> lol
May 01 21:08:49 <bencoh> DocScrutinizer05: ^@ is probably just some undisplayable character
May 01 21:12:26 <DocScrutinizer05> sounds like 0x00
May 01 21:17:02 <DocScrutinizer05> compare less `which date`
May 01 21:30:07 <DocScrutinizer05> as such it's indeed "undisplayable" and gets represented in ^notation
May 01 21:31:41 <DocScrutinizer05> :o>
May 01 21:32:26 <DocScrutinizer05> shouldn't try to type a ctrl+q to get a ctrl+@ ;-P
May 01 21:33:46 <bencoh> :D
May 02 02:06:28 <Vajb> sicelo: cool. I bet then there'll be regular updates ;)
**** ENDING LOGGING AT Tue May 02 03:00:01 2017