IRC log of #maemo for Monday, 2017-04-17

*** dafox has joined #maemo00:04
*** florian_ has joined #maemo00:20
*** florian_ is now known as florian00:34
*** florian has joined #maemo00:34
*** Vajb has joined #maemo00:46
DocScrutinizer05apropos twinkle: funny what stuff I wrote a nearly 10 years ago
DocScrutinizer05prolly even older since that file is already migrated from a yet older box/system00:49
DocScrutinizer05meanwhile both bash and I learned how to do such multithreaded stuff a tad smarter00:50
DocScrutinizer05and SIP registrars learned to not listen to the IP announced in SIP messages and rather use STUN/ICE and server assisted SIP to establish a call even to clients behind NAT00:52
DocScrutinizer05and nobody really does direct IP2IP VoIP (without registrar) anymore00:54
DocScrutinizer05if anybody except me ever did00:55
kerioDocScrutinizer05: yeah because SIP is a shit protocol00:56
DocScrutinizer05not really. Just complex00:57
DocScrutinizer05and not made from beginning with stuff like roaming and all sorts of nasty NATs in mind00:57
DocScrutinizer05SMTP is a real shit protocol ;-D01:00
kerioat least SMTP uses one connection :<01:03
*** l_bratch has joined #maemo01:06
*** luke-jr has quit IRC01:06
DocScrutinizer05ftp doesn't01:11
DocScrutinizer05but yeah, multiple ports are actually a nightmare. at least in SIP you can fix the ports to use01:11
*** luke-jr has joined #maemo01:12
DocScrutinizer05for ftp you have *lots* of fun with NATs01:12
DocScrutinizer05also for IRC DCC btw01:13
DocScrutinizer05which is in many aspects similar to SIP VoIP01:13
*** Smily has joined #maemo01:15
DocScrutinizer05A initiating a ftp file transfer from B to C is very similar to SIP A connecting B and C in a call, by A either forwarding a inbound call or steeping out of a 3way-conference01:17
DocScrutinizer05the problem always is NATs don't open receive ports, that's why they are considered a good safeguard against active attacks from 'outside' but that implies they also are a problem for stuff like ftp or SIP01:21
DocScrutinizer05VPN is just a glorified distributed NAT01:22
DocScrutinizer05so you have exactly same issues like with any other (local) NAT01:23
*** Vajb has quit IRC01:23
*** Alexxxlrus has joined #maemo01:24
*** Alexxxlrus has quit IRC01:30
Enrico_MenottiHello guys, hope you had a nice Easter day.01:38
Enrico_MenottiI'm trying to learn the last couple of things from Debian before switching to Devuan.01:38
Enrico_MenottiNow I have a strange issue about key mapping. I have been able to map some symbols to the fourth level. But with my surprise, the fourth level is not working for the keys a, z and x. And just for them - the rest is working.01:40
Enrico_MenottiAnybody know about a similar issue?01:40
*** Vajb has joined #maemo01:49
Enrico_MenottiI read it (without thinking too much about the details...). If I got it, Sh-Fn-x and Sh-Fn-z produce the same hardware event. So it is not possible to map them to different symbols. Is this right?01:49
DocScrutinizer05sorry, I just closed it so can't tell if z/x is right. The rest is correct01:56
DocScrutinizer05note that this stuff is merely deduced from schematics, I didn't verify against device behavior when I wrote it01:58
DocScrutinizer05anyway this is generally one of the reasons to use "sticky qualifier keys", like shift-down shift up a-down a-up -> uppercase-A01:59
DocScrutinizer05this works in maemo for shift and blue, not though for ctrl. And I think it doesn't work in the generic debian kbd driver either02:01
DocScrutinizer05the canonical pattern is (not found in maemo): press qualifier once: sticky for next keypress (like explained above). Press qualifier twice: lock (like capslock for pressing shift twice). With lock engaged any further press of the qualifier key will unlock and not create sticky state02:07
DocScrutinizer05so for repeated pressing a qualifier, the sequence is sticky -> locked -> off -> sticky -> locked -> off ->...02:08
DocScrutinizer05ooh sorry, maemo DOES that too02:09
DocScrutinizer05it kist lacks to show recent state, though that had been very very easy via indicator LED02:11
Enrico_MenottiOk, thank you, got it. Would be interesting to enable these sticky qualifier keys in Debian as well. But it's not necessary, so I won't investigate further.02:16
DocScrutinizer05I *think* maemo does it on X11 level02:17
DocScrutinizer05not sure though02:18
DocScrutinizer05could as well be kernel level02:18
DocScrutinizer05pr anything intermediate. keybpard service layering is a mystery to me. I have studied it from bottom to top a dozen times already and each single time I forgot about bottom details when I reached top02:19
*** Pali has quit IRC02:20
DocScrutinizer05hmm, maybe useful (for me and my problem):
*** jon_y has quit IRC02:28
DocScrutinizer05>>The device file transmits key up/down events with a scan code, which may or may not be identical to the value transmitted by the hardware (the kernel may translate the scan code from a keyboard-dependent value to a common value, and Linux doesn't retransmit the scan codes that it doesn't know).<<02:28
*** jon_y has joined #maemo02:29
DocScrutinizer05now it gets funny only with >>There are two mechanisms to configure the mapping from keycodes to keysyms:<<02:31
DocScrutinizer05there's also and which I bookmarked for a particular reason. The latter might be interesting for you too02:39
DocScrutinizer05evtest is rather great02:53
*** florian has quit IRC02:59
*** atk has quit IRC03:00
*** atk has joined #maemo03:00
*** infobot has quit IRC03:17
*** M4rtinK has quit IRC03:18
*** infobot has joined #maemo03:18
MaxdamantusMaemo does it at the HIM level.03:39
Maxdamantus(the sticky keys mechanism)03:40
MaxdamantusHIM = hildon-input-method03:40
*** dafox has quit IRC03:52
*** NotKit has joined #maemo04:38
*** TheKit has quit IRC04:42
*** LauRoman has quit IRC04:44
*** pagurus` has joined #maemo06:41
*** pagurus` has joined #maemo06:42
*** pagurus has quit IRC06:43
*** pagurus` has quit IRC06:45
*** pagurus has joined #maemo06:45
*** pagurus has quit IRC06:48
*** pagurus has joined #maemo06:48
*** pagurus has quit IRC06:51
*** pagurus has joined #maemo06:51
*** pagurus has quit IRC06:54
*** pagurus has joined #maemo06:54
*** pagurus has quit IRC06:57
*** pagurus has joined #maemo06:57
*** pagurus has quit IRC07:00
*** pagurus has joined #maemo07:00
*** pagurus has quit IRC07:03
*** pagurus has joined #maemo07:03
*** pagurus has quit IRC07:06
*** pagurus has joined #maemo07:06
*** pagurus has quit IRC07:09
*** pagurus has joined #maemo07:09
*** pagurus has quit IRC07:12
*** pagurus has joined #maemo07:12
*** pagurus has quit IRC07:15
*** pagurus has joined #maemo07:15
*** pagurus has quit IRC07:18
*** pagurus has joined #maemo07:18
*** pagurus has quit IRC07:21
*** pagurus has joined #maemo07:21
*** pagurus has quit IRC07:24
*** DocScrutinizer05 has quit IRC07:38
*** DocScrutinizer05 has joined #maemo07:38
*** vahe has joined #maemo07:52
*** KotCzarny has quit IRC07:55
*** KotCzarny has joined #maemo08:05
*** L29Ah has joined #maemo09:02
*** LauRoman has joined #maemo09:13
siceloi thought pagurus was going to test his irc stuff elesewhere ...10:08
KotCzarnyis he even alive?10:09
siceloyou mean online now? no10:27
sicelobut i think yesterday he promised something to that effect10:27
DocScrutinizer05again spam?10:45
DocScrutinizer05holler if you want sth getting fone about it, I don't see such stuff10:45
infobotDocScrutinizer05 meant: holler if you want sth getting done about it, I don't see such stuff10:46
KotCzarnyit seems to got fixed, for now10:48
*** L29Ah has left #maemo11:01
sicelo01:11 < DocScrutinizer05> it kist lacks to show recent state, though that had been very very easy via indicator LED11:04
sicelo01:11 < DocScrutinizer05> just*11:04
sicelopossibly deemed unnecessary as Maemo resets that sticky state as son as another program comes to the foreground11:05
siceloso it's not global11:05
sicelo(as soon as ... )11:05
DocScrutinizer05well, that's both puzzling and a flawed design, and doesn't really jelp with making lock state more clear11:06
DocScrutinizer05no, Nokia thought that notifier banner was sufficient11:06
siceloit's only sticky while you're actually in the program you 'stickied' in .. so you know what you did :)11:07
DocScrutinizer05but that notifier only shows up for like 2 seconds11:07
KotCzarnyled backlight under sticky keys would be sweet11:07
DocScrutinizer05shit no!11:07
siceloKotCzarny: perhaps neo900 should have it ;)11:08
KotCzarnywho knows?11:08
DocScrutinizer05same rationale applied to dialer would result in "why display the number you're typing, you know what you're doing"11:08
DocScrutinizer05sicelo: have waht?11:08
sicelo< KotCzarny> led backlight under sticky keys would be sweet11:08
DocScrutinizer05Neo900 is planned with 3color backlight and 3 monocrome indicator LEDs next to the leftmost keys11:09
DocScrutinizer05I'd probably prefer backlight switching from white to orange on capslock and from white to blue on Fn11:10
DocScrutinizer05but yeah, we will try to employ one or 2 of the indicator LEDs left of keyboard for that11:11
sicelo(personally i rarely, if ever, look at the state of my Caps Lock indicator on computer/laptop ... just saying)11:11
*** Pali has joined #maemo11:11
DocScrutinizer05it particularly sucks when you accidentally locked shift. You notice one char UC, then you wonder if it was sticky and thus reset or it's locked. So you type a second char just to check. Or you press shit again, just to learn it been sticky and you made another sticky11:13
DocScrutinizer05or you press shift as long as notifier shows up, and then once more >:-(11:14
DocScrutinizer05whatever you do, it's an annoyance11:15
DocScrutinizer05and God beware you're facing this situation in stuff similar to NM, where one keypress causes an action you can't revert, and there are UC and lc keys with different meaning11:17
siceloi don't think sticky keys work there, isn't it?11:18
DocScrutinizer05in BM not11:18
DocScrutinizer05but that's because there's no system at all at that point11:18
DocScrutinizer05though I didn't test, maybe it does (when it's getting handled in kernel kbd driver)11:19
DocScrutinizer05((i rarely, if ever, look at the state of my Caps Lock)) you also rarely press shift twice to emulate capslock on your PC, right?11:23
DocScrutinizer05call my old and senile nut I happened to press shift, then release it since I started to wonder if I _really_ wanted to type that char (for whatever reason, be it wording or the door ringing), then a few seconds or minutes later I press shift again, plus a char, then on next char I start swearing11:25
DocScrutinizer05if any state in a computer is sticky, you need feedback about what's the current state. If it's a sticky state in a HID, you need feedback directly on the HID and constantly11:27
DocScrutinizer05you won't prolly find a single keyboard whit capslock, scroll lock, numlock, and no LEDs for showing the state. Unless it's one of those age old keyboards that actually have mechanical toggle pushbuttons that "stay pressed" mechanically and visibly until you push them again11:30
KotCzarnyonscreen-ontop-sticky display could work too (for some people, so definitely as an option)11:35
Maxdamantusonly sticky thing I use on a normal computer is numlock .. sometimes use mouse buttons and need to have it in some state for that.11:35
Maxdamantuscan't remember which state does which, I just press numlock until it works the way I want.11:36
Maxdamantuscapslock is instead mapped to F13, and dunno what scrolllock does.11:36
MaxdamantusInteresting. Apparently scroll lock freezes console output in Linux. Good to know.11:39
MaxdamantusPresumably at a lower level than something like ctrl-s11:39
Maxdamantusoh wait, it's just another way of writing ctrl-s11:39
*** jonwil has joined #maemo11:40
*** florian has joined #maemo11:48
KotCzarnylol, that's why it's called SCROLL LOCK11:53
*** vahe has quit IRC12:11
Maxdamantusbut ctrl-s doesn't exactly lock the scroll.12:17
Maxdamantusit blocks any output.12:17
KotCzarnyit's from the times when screens were the teletypes12:21
KotCzarnyand virtual consoles emulate it12:22
*** cyteen has quit IRC12:28
*** trx has quit IRC12:29
*** cyteen has joined #maemo12:32
*** Vajb has quit IRC12:34
merlin1991 you won't prolly find a single keyboard whit capslock, scroll lock, numlock, and no LEDs for showing the state ... my acer notebook13:01
KotCzarnymerlin1991: some chinese and laptop kb ;)13:04
KotCzarnyhp/lenovo they suck at it and removed useful indicators (*locks and hdd activity often)13:04
merlin1991The acer still has the hdd led, but no lock leds :/13:05
*** CcxWrk has joined #maemo13:17
CcxWrkWhat is the current way to get stock N900 install up to date?13:19
*** xorly has joined #maemo13:50
kerioDocScrutinizer05: my SIP over vpn thing is unreliable if i don't redirect the whole internet through the VPN :(14:09
*** xorly has quit IRC14:10
KotCzarnythen do that? problem solved14:11
kerionah i should set up asterisk14:12
*** xorly has joined #maemo14:14
*** dafox has joined #maemo14:31
*** jonwil has quit IRC14:40
kerioDocScrutinizer05: my brain hurts :<15:16
KotCzarnyhave you tried turning your brain off and on again?15:17
keriothis shit is worse than sendmail15:22
*** xorly has quit IRC15:59
xeskerio: have tried to fix the account ip with mc-toll? have you enable sip/rtp conntracking into your home router?16:09
*** cpt_nemo_ has joined #maemo16:32
*** L29Ah has joined #maemo17:09
*** xy2_ has joined #maemo17:58
*** xy2_ has quit IRC18:09
*** xy2_ has joined #maemo18:09
Enrico_MenottiOk, I have Debian running on the n900, keyboard quite fine, and wifi access. Also ssh. I'd like to adjust a couple of things, and then I'll move towards Devuan.18:10
Enrico_Menotti1. I have been able to customise the keyboard by editing the symbols file (<xkbroot>/symbols/nokia_vndr/rx-51) and running setupcon. But before that I had to remove the cached files in /etc/console-setup. If I run setupcon with the --save option, these files are recreated. But the system somehow doesn't use them - setupcon keeps on creating a new tmp file and loading that.18:13
Enrico_Menotti2. The wireless driver seems to generate a new mac address in a random way at each boot. So the system keeps on creating new wlan interfaces at each reboot. I'd like to have the wlan named wlan0 at each boot. I think I have to edit /etc/udev/rules.d/70-persistent-net.rules. But so far the only way I have found to reference the wireless device is via the mac address. How may I reference it in a unique way?18:15
Enrico_Menotti3. I tried to install usbutils and run lsusb, but that fails. From some debugging, I found out that the system doesn't find usbfs. Is it possible that Pali's rx51_defconfig file does not enable usb support?18:17
*** luke-jr has quit IRC18:33
*** luke-jr has joined #maemo18:35
DocScrutinizer05that's all very much OT for maemo18:47
DocScrutinizer05IOW in maemo we either don't have the problem or we have a working solution, for all your "debian problem" points you listed so far18:48
infoboti guess hen is hostmode-easy-now, or, or see ~hostmode18:52
infobothostmode is probably, or, or see ~hostmode-powered18:52
*** L29Ah has left #maemo19:00
*** L29Ah has joined #maemo19:03
*** Venemo has joined #maemo19:14
DocScrutinizer05IOW with a stock kernel you're *always* out of luck with USBhostmode on N90019:16
DocScrutinizer05I wasted over a year into solving this issue19:17
*** MoeIcenowy has quit IRC19:27
*** MoeIcenowy has joined #maemo19:28
*** M4rtinK has joined #maemo19:29
*** geaaru has quit IRC19:34
*** shentey has joined #maemo20:13
*** dmth|intevation has joined #maemo21:20
*** sunshavi has quit IRC21:36
*** sunshavi has joined #maemo21:45
*** xy2_ has quit IRC22:28
*** Cor-Ai has joined #maemo22:56
sicelohaha that's a crazy thing23:07
MaxdamantusEnrico_Menotti: if you're just trying to use the device as a USB gadget, you'd load one of the g_ modules.23:10
MaxdamantusEnrico_Menotti: a USB device is traditionally meant to be a gadget or a host, not both.23:11
Maxdamantusthough N900's controller happens to be able to work as both.23:11
*** mp107 has joined #maemo23:23
*** mp107 has quit IRC23:23
*** mp107 has joined #maemo23:24

Generated by 2.15.1 by Marius Gedminas - find it at!