IRC log for #neo900 on 20160825

00:03.50*** join/#neo900 goiken_ (~goiken@dslb-188-102-028-129.188.102.pools.vodafone-ip.de)
00:39.20*** join/#neo900 goiken_ (~goiken@dslb-188-102-028-129.188.102.pools.vodafone-ip.de)
01:01.45*** join/#neo900 chomwitt (~chomwitt@ppp-94-67-221-221.home.otenet.gr)
01:15.58*** join/#neo900 goiken_ (~goiken@dslb-188-102-028-129.188.102.pools.vodafone-ip.de)
01:52.59*** join/#neo900 goiken_ (~goiken@dslb-188-102-028-129.188.102.pools.vodafone-ip.de)
02:39.11*** join/#neo900 ravelo (~ravelo@178.165.129.23.wireless.dyn.drei.com)
02:40.56*** join/#neo900 liteIRC_ (~ravelo@93-82-103-217.adsl.highway.telekom.at)
02:50.47*** join/#neo900 herpderphurr (~afwang@c-98-234-221-193.hsd1.ca.comcast.net)
03:07.07*** join/#neo900 goiken_ (~goiken@dslb-188-102-028-129.188.102.pools.vodafone-ip.de)
03:42.14*** join/#neo900 Oksana (~chatzilla@Maemo/community/ex-council/Wikiwide)
03:48.12*** join/#neo900 goiken_ (~goiken@dslb-188-102-028-129.188.102.pools.vodafone-ip.de)
03:58.28*** join/#neo900 galiven__ (~Andrew@50-205-116-131-static.hfc.comcastbusiness.net)
04:01.40*** join/#neo900 DocScrutinizer05 (~saturn@openmoko/engineers/joerg)
04:01.40*** mode/#neo900 [+v DocScrutinizer05] by ChanServ
04:26.43*** join/#neo900 goiken_ (~goiken@dslb-188-102-028-129.188.102.pools.vodafone-ip.de)
04:42.12*** join/#neo900 pagurus` (~user@pD950E430.dip0.t-ipconnect.de)
05:00.48*** join/#neo900 goiken_ (~goiken@dslb-188-102-028-129.188.102.pools.vodafone-ip.de)
05:15.08*** join/#neo900 liteIRC_ (~ravelo@77.119.131.53.wireless.dyn.drei.com)
05:41.24*** join/#neo900 goiken_ (~goiken@dslb-188-102-028-129.188.102.pools.vodafone-ip.de)
06:16.55*** join/#neo900 goiken_ (~goiken@dslb-188-102-028-129.188.102.pools.vodafone-ip.de)
06:44.32*** join/#neo900 ecloud (quassel@nat/qt/x-dpskqyuhxpmtwing)
06:47.53*** join/#neo900 brolin_empey (~brolin@brolin.be)
06:55.04*** join/#neo900 goiken_ (~goiken@dslb-188-102-028-129.188.102.pools.vodafone-ip.de)
07:29.06*** join/#neo900 goiken_ (~goiken@dslb-188-102-028-129.188.102.pools.vodafone-ip.de)
08:06.06*** join/#neo900 goiken_ (~goiken@dslb-188-102-028-129.188.102.pools.vodafone-ip.de)
08:24.58*** join/#neo900 chomwitt (~chomwitt@ppp-94-67-221-221.home.otenet.gr)
08:49.13*** join/#neo900 radekp (~radek@82.113.39.213)
09:07.25*** join/#neo900 SylvieLorxu (~TheLastPr@541B7AAC.cm-5-4b.dynamic.ziggo.nl)
09:19.22*** join/#neo900 paulk-collins (~paulk@gagarine.paulk.fr)
09:30.40*** join/#neo900 galiven__ (~Andrew@50-205-116-131-static.hfc.comcastbusiness.net)
09:55.00atkDocScrutinizer05: You know the proximity sensor on the N900? Will the Neo900 use the same sensor?
09:55.17atkHas someone worked out how to "fix" it so that it doesn't get a false positive proximity response when in sunlight / bright light?
09:55.38DocScrutinizer05well, as long as we can source the slider mech with flex cable, yes
09:55.38atkBecause personally I think the solution would be quite simple, but I have no idea how much control the Neo900 will have over the sensor.
09:56.14atkI presume the sensor is just an IR led and a IR receiver?
09:56.47DocScrutinizer05it's an integrated module consisting of IR LED, sensor, and controller, yes
09:57.00atkcontroller?
09:57.04DocScrutinizer05sure
09:57.10atkWhat does this controller do?
09:57.11DocScrutinizer05it pulses
09:57.35atkah, is there any way to get the output of the sensor without pulsing the led?
09:57.44DocScrutinizer05hm?
09:58.13DocScrutinizer05the sensor has a VCC, GND, and a logical output line
09:58.40DocScrutinizer05btw the cam door uses basically same sensor
09:58.41atk(this would be the obvious solution, if the sensor detects light without the led being on, it means that the sensor is obviously not in close proximity to anything, if the sensor doesn't detect light then the secondary test (blink + detect) would check if the device was simply in a dark area and not in proximity to anything.
09:59.24DocScrutinizer05the controller is supposed to detect the signal generated by the LED pulse
09:59.29atkah, so the whole sensor is one integrated thing with just VCC, GND, and output?
09:59.32DocScrutinizer05ambient light shouldn't matter
09:59.39DocScrutinizer05yes
09:59.56atkbut the N900 does apparently have an issue (I've noticed this) with the device receiving calls in bright light / sunlight.
10:00.00atkThe device will lock the screen
10:00.25DocScrutinizer05don't you think the screen simply disables backlight?
10:00.46atkThat's the ambient light sensor.
10:01.03atkI mean the sensor to detect when someone has placed the device against their ear (to prevent their face touching buttons)
10:02.06atkHmm. I guess this does mean that it might be possible to fix this problem another way - combine the ambient light sensor and the proximity sensor, if the ambient light sensor can sense a lot of light but the prox sensor senses proximity, then the device is just in bright light (overwhelming the prox sensor)
10:02.32*** join/#neo900 galiven (~Andrew@50-205-116-131-static.hfc.comcastbusiness.net)
10:02.50DocScrutinizer05the ambient light sensor switches off screen backlight with sufficient ambient brightness
10:03.17DocScrutinizer05you easily could see this a s locked screen
10:03.31atkDocScrutinizer05: http://talk.maemo.org/showthread.php?p=1013440
10:03.40DocScrutinizer05sorry
10:04.09atk(the lights at my workplace trigger this bug)
10:04.37DocScrutinizer05http://neo900.org/stuff/block-diagrams/neo900/neo900.html  "proximity sensor"
10:05.30DocScrutinizer05I only consider something a bug when I verified it by commandline means
10:06.14atkAs in, it is well known that this happened with N900 for some people in some scenarios (outdoors / odd lighting)
10:07.25atkthe prox sensor does have "high ambient light suppression" which works perfectly fine 99% of the time for me, but it has happened that I could not pick up a call on the device without using the slider when in specific lighting conditions / outside on a really sunny day.
10:08.21DocScrutinizer05hmm
10:10.25atkIf you know how I can get the output of the prox sensor via command line, I could try it out for you at various places.
10:11.24DocScrutinizer05watch cat /sys/devices/platform/gpio-switch/proximity/state
10:12.31DocScrutinizer05watch -n 1 cat /sys/devices/platform/gpio-switch/proximity/state
10:12.36atk(using the soft keyboard for the terminal is a flash back to days of android :( )
10:14.27DocScrutinizer05you found the SFH7741 datasheet link in our block diagram?
10:15.07atkYeah
10:15.46atkIt appears that I can get it to happen very randomly if I get relatively close to a energy saving (fluorescent) lightbulb.
10:16.04atkThe ones at work are fluoros so I will check there when I get back on Monday.
10:16.12atkI'll also try outside on a sunny day.
10:16.22atkBut I can't get it to give me a false positive with my halogen lightbulb.
10:17.06*** join/#neo900 arcean (~arcean@62.159.77.166)
10:19.08DocScrutinizer05which is weird since it's (supposed to be) sensitive to IR only, and incandescent has more IR than any flourescent lamp. But what just occurs to me: FL have fast pulses that might occasionally fall in sync freq (by a multiple) with the sensor pulses, so sensor thinks the pulse it sees is an echo of own LED
10:19.30atkI just thought of that too.
10:20.32atkMaybe I've just been unlucky :P
10:21.59atkI bet the sensor uses the technique I described earlier too (check for light without pulse, check for light during pulse)
10:22.35DocScrutinizer05yes, that's what I think it does
10:23.18*** join/#neo900 chomwitt (~chomwitt@ppp-94-67-221-221.home.otenet.gr)
10:23.50atkI'll find a spare photo transistor and hook it up to my oscilloscope and do some tests to see what the light of the fluorescent lightbulb would look like to the device
10:23.55atkas well as seeing the pulse frequency of the prox sensor
10:24.28DocScrutinizer05:-)
10:24.54DocScrutinizer05you can actually see the prox IR LED with several cameras
10:25.21atkYes, but it won't give you a good idea of the frequency
10:25.59DocScrutinizer05possible fix: make ALS read out a current value whenever the prox goes open->closed
10:26.17DocScrutinizer05only allow true "closed" when ALS doesn't say "bright"
10:27.01DocScrutinizer05needs to check: how fast is ALS via /sys/ interface
10:27.07atkYeah, that's what I was saying earlier, it would avoid the 1/100 scenario where you can't pick up a call because the prox sensor thinks there's something close by.
10:27.22DocScrutinizer05yep
10:27.46atkIn any case, I need to acquire some cow extract for my coffee.
10:28.42DocScrutinizer05watch -n 1 cat /sys/class/i2c-adapter/i2c-2/2-0029/lux
10:29.43DocScrutinizer05I think the lagginess of ALS is only by MCE poll interval
10:30.21DocScrutinizer05since mce got RE'ed, you can patch it to fix the issue
10:31.07DocScrutinizer05if that's even MCE that does screen lock on prox close
10:32.26DocScrutinizer05IroN900:~# lsof |grep /sys/devices/platform/gpio-switch/proximity/state
10:32.27DocScrutinizer05mce         815       root   18r   REG        0,0     4096       4115 /sys/devices/platform/gpio-switch/proximity/state
10:32.29DocScrutinizer05hald-addo   874  haldaemon    6r   REG        0,0     4096       4115 /sys/devices/platform/gpio-switch/proximity/state
10:38.11DocScrutinizer05what I'd love: pick up call by gesture - like: press cam button and shake
10:40.08DocScrutinizer05or rather: hold device screen down, preass cam trigger, turn screen up (or into portrait orientation that's natural for listening to the phone)
10:40.59DocScrutinizer05device already mutes ringtone when turning it screen down
10:41.54DocScrutinizer05so screen-down, cam-trigger, !screen-down is a natural gesture for picking up a call
10:42.05atkI like that these features were already being done in 2008 by nokia for the N900
10:42.17atkbut it's only recently that android phones get similar kinds of useful features
10:42.33DocScrutinizer05meh andriod
10:42.37atkit appears no android phone has yet mastered the ability of the N900 to be off but still be able to wake you up in the morning
10:42.54DocScrutinizer05hahaha
10:43.48atkSomething I miss from android which the N900 doesn't do is the ability to use the volume rocker to fast forward / reverse songs and music when the device screen is locked.
10:43.58atkIf you hold down the volume up it skips to the next song
10:44.14atkhold down and it goes to the beginning of the song, hold down again and it goes to the previous song
10:44.23atkI wonder how hard it would be to get that working on the N900
10:44.30DocScrutinizer05err N900 does this
10:44.39atkIt does?
10:44.41atk:P
10:44.52DocScrutinizer05needs "an app"
10:45.49DocScrutinizer05easy to write yourself whatever you like, with dbus scripting
10:45.57atkHmm.
10:46.13atkdbus in C is a pain
10:46.22atk(If you avoid sd-bus and gdbus)
10:46.38atkI plan on one day writing some simple generator for the silly amounts of boilerplate
10:46.39DocScrutinizer05well, moderately easy. You still need the hildon-desktop hotkey feature
10:46.53atkDocScrutinizer05: http://sprunge.us/RFSP
10:48.25atk(This is calling a org.freedesktop.Notifications.Notify method)
10:48.55atk(and there's no error checking on all the dbus_message_iter* calls (which return false if they run out of memory))
10:49.09DocScrutinizer05http://maemo.org/packages/view/dbus-scripts/
10:49.37atkis this like the udev of dbus?
10:49.55atk(sans the device node creation, I mean the scripting side :P)
10:50.07atkCool
10:50.13atkI think I'll try hacking this together one day.
10:50.19atkAmazing flexibility
10:52.50DocScrutinizer05http://talk.maemo.org/showthread.php?p=464473&highlight=volume#post464473
10:57.06atkHmm, using the camera button would also be a nice feature for play pause.
10:57.41DocScrutinizer05you see it's all there already, usually at least 80%
10:57.42atkI also had this interesting idea of using espeak and some kind of audio ducking to say the current time out loud via the headphones when the camera button / some other button was pressed in a specific way
10:59.33DocScrutinizer05hmm, 80% already there in "say who's calling"
10:59.59DocScrutinizer05and again dbus-scripting
11:00.03atkhmm
11:00.15atkWell, I know what I'm doing during some future weekend.
11:00.21DocScrutinizer05:-)
11:02.11DocScrutinizer05actually for "say the time" I'd use a dbus-script that simply picks the right audio snippets to play back via aplay
11:02.18DocScrutinizer05or playsound
11:03.03atkI think for that I would go with python.
11:06.08DocScrutinizer05case `date +hh:mm` in 00:*) playsound midnigt_and.wav;; 01:*) playsound one.wav;; ... esac;
11:08.38DocScrutinizer05you can run python scripts too, as debus-script
11:11.14DocScrutinizer05my payswoosh: http://paste.opensuse.org/44081634
11:11.22DocScrutinizer05playswoosh even
11:20.23DocScrutinizer05finding the parameters for /etc/dbus-scripts.d/dbus-scripts-settings took me half a day though, it's somewhat twisted
11:21.31DocScrutinizer05and NEVER place the script into  /etc/dbus-scripts.d/ ! it made the whole thing forkbomb ;-P
11:26.16DocScrutinizer05iirc what took the most time though in finding the right parameters etfc, was to learn and to accept that freedesktop.org in their *eternal* wisdom specified the switches dbus signals in a way that doesn't tell whether the switch got closed or opened
11:27.47DocScrutinizer05which makes me feel like grabbing the rifle and load it with owl shit
11:34.35atkThat EOMA68 device on crowd supply just got 10% funded in a day
11:34.43atkI think they might make it: https://www.crowdsupply.com/eoma68/micro-desktop
11:34.57atkIt was at 80% yesterday afternoon.
11:45.58atkMaybe it would be good PR for the Neo900 project if it publicly endorses the EOMA68 micro desktop
11:54.44*** join/#neo900 herpderphurr (~afwang@c-98-234-221-193.hsd1.ca.comcast.net)
11:56.31*** join/#neo900 galiven_ (~Andrew@50-205-116-131-static.hfc.comcastbusiness.net)
11:59.32*** join/#neo900 galiven__ (~Andrew@50-205-116-131-static.hfc.comcastbusiness.net)
12:26.30*** join/#neo900 ploop (~ploop@unaffiliated/ruukasu)
13:43.48wpwrakthey have a fun observation here: https://www.crowdsupply.com/eoma68/micro-desktop/updates/speedy-devuan
13:43.53wpwrakgood PR for devuan, for sure :)
14:00.19*** join/#neo900 drrrz (~drrrrz@172.98.67.95)
14:36.26*** join/#neo900 MonkeyofDoom (~one@2601:240:c400:5d33:dd62:5953:ca08:b49c)
14:41.22*** join/#neo900 chomwitt (~chomwitt@ppp-94-67-221-221.home.otenet.gr)
14:48.53atkWell, I got one, why not.
14:51.04atkI then wrote an email to the FSF thanking them for bringing the device to my attention and asking them to take a look at the Neo900.
14:51.36atkJust to see what the response will be, and to see if the FSF would be interested in helping Neo900 along in one way or another.
14:51.43*** join/#neo900 xman (~xman@user-0cdft6f.cable.mindspring.com)
14:51.46atkI think I'll send another email to the EFF at some point.
14:52.29DocScrutinizer05fun observations? http://paste.opensuse.org/64720552
14:52.48atk(I did of course make it clear that I am not part of the project but that you guys (or more specifically DocScrutinizer05) really liked the idea of me sending emails to these organisations asking for some guidance / help)
14:53.05atkDocScrutinizer05: hah
14:53.10DocScrutinizer05the oh-so-optimized journald is 10000 times slower than plain files
14:53.17atkDocScrutinizer05: you know, you can get arch working without systemd :P
14:53.32MonkeyofDoomjournald is apocalyptically slow
14:54.34atkhas his logs done by socklog for syslog and klog and all of that is done via svlogd from runit
14:54.56atkgets easy log rotation and nice text files out of it
14:56.17atkWhen I get this device, I'll probably end up putting voidlinux on it, (I'll figure out how debian got their stuff to run so fast)
14:56.27atkdevuan*
14:57.08DocScrutinizer05other fun observation: the devuan chroot nuked my suse system's cgroups so systemd-crond is not happy anymore (and probably no other login as well)
14:59.08DocScrutinizer05http://paste.opensuse.org/49841194 http://susepaste.org/3918705 (line 165 more likely than from line 3)
15:01.50DocScrutinizer05alas I have no clue how to cope with cgroups and systemd, to recover this without reboot to restart dang systemd
15:04.24*** join/#neo900 galiven (~Andrew@50-205-116-131-static.hfc.comcastbusiness.net)
15:05.06atkwhy do you need crond?
15:05.15atkor well, systemd-crond?
15:05.48DocScrutinizer05well... there are a few tasks that ... meh you're kidding me
15:07.41DocScrutinizer05you don't honestly ask why there are cronjobs on a system?
15:08.14atkWell, why you can't replace systemd-crond with normal cron
15:08.35atkAnd actually, I don't have any cronjobs on either this laptop or my server at all.
15:08.58DocScrutinizer05will that change anything in the damn session management that systemd and freinds implement?
15:09.11atkThe server doesn't have any because they're don via systemd timers though, when I replace systemd with runit on the server, it will need cron
15:09.32atkIf you don't need session management you can literally do away with it.
15:09.38atkNo idea how easy your distribution will make that.
15:09.51atkMight not work with a full DE, but works for a wm
15:10.05atkThere are replacements for systemd-logind too
15:10.19atkand I don't quite see how session management relates to cron (user cronjobs?)
15:10.32DocScrutinizer05prolly
15:11.04DocScrutinizer05in /var/log/syslog I get periodic (every 15 min) http://paste.opensuse.org/33575402 systemd cgroup error
15:12.08atkah, it creates a session for root to run the cronjob
15:12.14atkYeah, with a normal cron you can just ignore all that
15:12.33atkjust replicate all these cronjobs with a normal cron (anacron for a pc) without using cgroups
15:15.27DocScrutinizer05this all doesn't help me out, and I won't even try to start get rid of parts of systemd-madness in opensuse
15:18.47wpwrakdevuse ;-)
15:18.54wpwrakor susian ?
15:22.03DocScrutinizer05cgroups anybody?
15:22.31DocScrutinizer05alas their introduction completely bypassed my attention
15:23.26DocScrutinizer05*cough*
15:23.28wpwrakyou could design new ones, that depend on dbus, jQuery, mysql, and ASN.1. call them jgroups.
15:23.29DocScrutinizer05saturn:/var/log # apropos cgroup
15:23.31DocScrutinizer05cgroup: nothing appropriate.
15:25.20DocScrutinizer05as embarrassing as it is, I seem to have to revert to aunt google and stuff like https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Linux/6/html/Resource_Management_Guide/ch01.html
15:38.55*** join/#neo900 paulk-nyan-big (~paulk@gagarine.paulk.fr)
15:53.30DocScrutinizer05anyway since I got fed up and I also learned a lot about lib/dependencies management during last 4 weeks, here it comes http://wstaw.org/m/2016/08/25/plasma-desktopvp2277.png genuine opensuse kicad. Now let's test this thing and see what we got there
15:56.28DocScrutinizer05http://wstaw.org/m/2016/08/25/plasma-desktophT2277.png  hmm _not_ good
15:56.31atkwpwrak: that sounds like a series of great design decisions combined into a successful product.
15:56.47atkwpwrak: You should pitch it to lennart.
16:05.19DocScrutinizer05dang, opengl canvas doesn't work at all
16:05.38DocScrutinizer05it shows existing tracks but doesn't draw anything new
16:06.46DocScrutinizer05oops nope, that was cairo
16:07.17*** join/#neo900 Pali (~pali@Maemo/community/contributor/Pali)
16:08.39DocScrutinizer05so _maybe_ the solution to my cgroups-problems is not to run that chroot devuan kicad
16:11.17atkhmm...
16:11.36atkOddly enough, the issue might be related.
16:11.49DocScrutinizer05it for sure is
16:12.09atksomething to do with X, sessions and cgroupws
16:12.14atkcgroups*
16:12.16DocScrutinizer05http://paste.opensuse.org/49841194 http://susepaste.org/3918705 (line 165 more likely than from line 3)
16:12.35DocScrutinizer05X?
16:12.40DocScrutinizer05don't see that
16:13.14DocScrutinizer05to me the most probably culprit is line165 in that syslog excerpt ^^^
16:13.19atkX can run rootless with logind (session management magic) which might have something to do with opengl
16:13.26DocScrutinizer05after that the next cronjob fails
16:13.47DocScrutinizer05ooh, opengl, nah opengl is fine :-)
16:14.36DocScrutinizer05it was cairo that doesn't work
16:47.48wpwrakatk: the risk of doing so would be that he'd probably pick up some of these ideas :)
17:05.22*** join/#neo900 chomwitt (~chomwitt@ppp-94-67-221-221.home.otenet.gr)
17:44.41*** join/#neo900 paulk-leonov (~paulk@gagarine.paulk.fr)
18:21.53*** join/#neo900 Airwave (~Airwave@84-52-232.185.3p.ntebredband.no)
19:02.04*** join/#neo900 goiken- (~goiken@dslb-178-012-205-163.178.012.pools.vodafone-ip.de)
19:28.47DocScrutinizer05sorry?
19:29.25DocScrutinizer05oh, he as in The Poettering
19:30.11DocScrutinizer05cgroups however are a kernel concept, just hihacked by Poettering/systemd
19:32.51DocScrutinizer05the two annoying parts are: how is a chroot going to nuke host system's cgroups? if you can do this, cgroups are flawed anyway. And how is dang systemd supposed to recover from such issues without needed a reboot?
19:34.06MonkeyofDoomI thought chroot was strictly filesystem, and cgroup separation requires a different namespacing call
19:38.43*** join/#neo900 Airwave (~Airwave@84-52-232.185.3p.ntebredband.no)
19:42.15*** join/#neo900 paulk-leonov (~paulk@gagarine.paulk.fr)
21:26.03*** join/#neo900 paulk-leonov (~paulk@gagarine.paulk.fr)
21:36.50*** join/#neo900 cg132 (~cg132@173-27-250-186.client.mchsi.com)
21:38.25cg132Looking for the Bankleitzahl (BLZ) for destination bank. Tis an 8 digit code but I cant find one that matches the SWIFT.  Anyone know?
21:39.43DocScrutinizer5176070024
21:39.55cg132Thanks Doc!
21:39.58cg132You're the best.
21:39.59cg132:D
21:40.00DocScrutinizer51yw
22:15.56*** join/#neo900 paulk-collins (~paulk@gagarine.paulk.fr)
23:53.00*** join/#neo900 tsuggs (~tim@adsl-074-171-024-060.sip.int.bellsouth.net)

Generated by irclog2html.pl Modified by Tim Riker to work with infobot.