**** BEGIN LOGGING AT Wed Feb 08 03:00:03 2017 Feb 08 04:19:02 I was wondering if anyone has information related to installing Arch Linux on N810s? Feb 08 04:19:12 I saw the Debian/Gentoo tutorials Feb 08 04:23:41 n810moron: likely similar to gentoo tutorial Feb 08 04:34:51 Juseto: it seems like the gentoo tutorial uses a lot of very gentoo-specific scripts. There was a tutorial on archmobile.org but that seems dead now. Feb 08 05:13:19 hmm, n810moron needs to learn lingering on in IRC Feb 08 05:13:25 https://web.archive.org/web/*/archmobile.org Feb 08 05:24:24 also "seems dead now"? Feb 08 06:07:23 Arch Linux ARM is alive though Feb 08 06:07:57 nowadays it doesnt really matter, all you need is working bootloader which initializes hw and kernel Feb 08 06:08:08 userland can be transpplanted Feb 08 06:09:04 xes: i need some UI design, i'm code guy, not UI and it shows. if you can outline how it should look/work i would gladly implement it Feb 08 06:41:26 KotCzarny: u didn't ask me, but OMP would be nice baseline. Feb 08 06:41:46 i'm asking anyone interested Feb 08 06:42:05 so you want as simple as it gets? Feb 08 06:42:11 did you try winamp skin mode? Feb 08 06:42:34 well, what i said and maybe get it integrated well Feb 08 06:42:43 i didn't :/ Feb 08 06:43:22 though i should probably add auto core spawn to gui to make things working for new users Feb 08 06:43:49 can't recall the version i tried, but felt like "ok, but too much everything at once" Maybe menu's and submenu's Feb 08 06:45:02 you can switch to wa mode in options (style menu) Feb 08 06:45:47 hmm maybe u can connect efforts with OMP guys? User selectable which one is default player. Feb 08 06:46:10 their skin and ur engine so to speak Feb 08 06:46:50 omp has skins? or are you talking about default ui? Feb 08 06:47:54 default ui Feb 08 06:48:15 i meant that if one wants to he could replace omp with oscp Feb 08 06:53:13 anyway, try winamp skin mode if you have a minute or two Feb 08 06:53:52 ok will do, but don't hold ur breath for it ;) Feb 08 06:54:07 btw do we replacement for defauly gallery? Feb 08 06:54:15 do we have* Feb 08 06:54:20 haven't seen Feb 08 07:06:39 i think there is .. never tried it though Feb 08 07:09:05 btw thought of what xes said and thought that maybe it could be tested via mplayer launching shell script too? Feb 08 07:09:59 tested what? Feb 08 07:18:11 about that tmo post. Where he said that sometimes when he clicks hi saved shortcut for a radiostream it starts to play something else. Feb 08 08:06:15 DocScrutinizer05: i have no experience with supl. Or would you mean some kind of supl proxy like this: https://www.tajuma.com/supl/supl-proxy-ex.html ? Then, we should evaluate also user's concerns about collected information about imsi,mnc,mcc.. (you know... my supl server setting is disabled) Feb 08 08:12:55 KotCzarny: thanks. I will try to use it with the will to find something to change :) Feb 08 08:13:49 xes, simple outline could be a good start, or quick thoughts Feb 08 08:16:30 yep, it's easy. It's enough to use it after a night of 3 hours to understand what you can't find or use at first try :) Feb 08 08:17:48 its easier on pc because mouse hover actually gives a hint ;) Feb 08 08:18:17 reading first time help also helps instead of instantly clicking 'close' Feb 08 08:24:37 you are right. But while you forget help informations after some month, if the app is user friendly/instinctive you can use it easily also when you mind is still dreaming or doing something else. Then, this will produce a better integration with the 'human side' and also with the rest of the OS (as secondary effect while trying to reach a good usability) Feb 08 08:25:42 uhum, though part of the problem is my usability profile is different from 'default user profile' Feb 08 08:26:18 i like 'less clicking/scrolling to do things' Feb 08 08:26:24 ...but don't feel this as complaint. Every mind has his shape and paths Feb 08 08:26:32 which results in plenty of controls, which isnt user friendly Feb 08 08:26:54 still, i would like it to be usable for others too, that's why i ask for hints Feb 08 08:28:04 also, it's meant to be usable with cursor keys (even in gtk ui mode) Feb 08 08:28:37 left is 'go back/go to parent dir' right is 'play this file/enter dir or archive' Feb 08 08:29:50 plenty of controls is good if inside an easy to guess tree Feb 08 08:31:56 anyway, thanks for your efforts and availabilty, i will come back with some idea :) Feb 08 08:32:03 great! Feb 08 08:58:38 connect fan or led Feb 08 08:58:48 and test different values Feb 08 08:58:56 eh, echan. **** BEGIN LOGGING AT Wed Feb 08 09:36:31 2017 Feb 08 13:41:30 xes: I have had a look in my alarmed for the script I run to start the radio, it's the same as your post but I am discarding the stdout have you tried that to see if that helps? Feb 08 14:16:37 sixwheeledbeast: the problem is not related only to that script. If you read my post, you can see that the same happens while vieweing a video in the camera-ui2 "preview" Feb 08 14:21:09 sixwheeledbeast: then, maybe i'm wrong but i think that in the past that video preview was different, maybe embedded inside camera-ui.. But as said maybe i'm wrong Feb 08 16:59:59 I don't have camera-ui2 on my daily device but I will try my old radio script for a bit and see what happens. Feb 08 17:29:35 what's the issue? Feb 08 18:00:02 http://talk.maemo.org/showpost.php?p=1522786&postcount=1717 Feb 08 18:15:43 hmm, sounds a tad like race with dbus_listen() during app startup. Possibly also related: gstreamer and mafw Feb 08 18:19:50 the question would be ... does it happen with stock MediaPlayer, then. Feb 08 18:33:51 yes, first of all. Not only stock mediaplayer but also (just) stock alsaped and policyfile, or just stock dbus (proxy) settings, or just stock pulseaudio... Feb 08 18:34:40 this is a damn complex problem Feb 08 18:36:08 aiui there might be differences in stock mediaplayer vs the mediaplayer exhibiting this problem, regarding if and how mafw is used Feb 08 18:37:51 huh, that's why i hate dbus and overcomplicated things Feb 08 18:38:01 but aiui omp is replacing stock mediaplayer and listens to the very same dbus messages Feb 08 18:38:34 and it uses mafw as well Feb 08 18:40:02 but any app acting as a default player would have to implement such interface Feb 08 18:52:33 yeah, so it's either what I originally suspected: race condition when dbus proxy stores the signal and replays it to mediaplayer after process startup, which might happen at a point in time where that process isn't ready to deal with the dbus message yet Feb 08 18:53:10 or it's totally unrelated to dbus and mediaplayer and rather caused by one of the other domains I listed above Feb 08 18:55:31 I suggest to start the mediaplayer with a debug wrapper (or even start the debugger internally inside the mediaplayer) to see exactly what it does, when and why Feb 08 18:57:40 also don't miss to look into the mediaplayer desktop widget Feb 08 18:58:22 which might change behavior of mediaplayer by patches applied to hildon desktop or whatever hildon-blabla Feb 08 18:58:57 (obviously only when that widget is installed) Feb 08 19:00:13 a lot of preconditions to check and rule out Feb 08 19:01:50 the intriguing question is: where is that info (last song played yesterday) stored persistently? Feb 08 19:02:22 and why doesn' Feb 08 19:02:42 "current song" maybe Feb 08 19:03:14 and why doesn't the new instruction (dbus cmd 'play http://foo.bar) override that old info Feb 08 19:04:34 it's either stored in some persisten file with 'settings' like playlists etc, or in some running process work memory aka variables/datastructures Feb 08 19:05:25 I suspect it's the latter, namely OMP Feb 08 19:05:37 or the desktop widget Feb 08 19:07:04 but mafw itself is also a possible culprit, since you could control mafw from multiple frontends, means one frontend instructing a new song playing makes all frontends know about and display that fact Feb 08 19:07:42 s/from multiple frontends/from multiple frontends concurrently/ Feb 08 19:07:42 DocScrutinizer05 meant: but mafw itself is also a possible culprit, since you could control mafw from multiple frontends concurrently, means one frontend instructing a new song playing makes all frontends know about and display that fact Feb 08 19:08:42 at least that's what I think it works like, you would have to ask mohammadAG about details, I never looked into it Feb 08 19:09:45 anyway mafw is maybe a smart idea, but alas a flawed implementation: it lacks for example gapless playback capabilities Feb 08 19:12:36 that's why back when the discussion resulted in the conclusion that really high quality media players will want to ignore mafw and coexist with stock mediaplayer instead trying to replace it Feb 08 19:12:51 lol Feb 08 19:13:14 yeah, gapless (along with x-fade) is on my todo too Feb 08 19:42:48 re mafw, I don't know how exactly the API is defined, but it's quite possible it's a mere handling error, e.g. like not stopping but only pausing "the last song of yesterday", and then mafw interpreting a new PLAY command as UNPAUSE and ignoring any parameter that asks for a new media file to playback Feb 08 19:45:24 in such situation the "persistent storage" would be in mafw itself, but the true culprit would be the controlling process that handles mafw incorrectly Feb 08 20:08:25 bencoh: I think the "does it happen with stock media player too" is pretty much to the point. What gives me headache is the "it works ... 80%", which sounds prety much like a race condition. The nasty part about race conditions is to reproduce and/or evaluate them since literally everything you change on the test platform can have an impact to the result Feb 08 20:09:10 odds are even running OMP under strace will change timing enough so the race doesn't hit anymore Feb 08 20:09:25 xes: ^^^ Feb 08 20:54:56 that link should go into paragrap Feb 08 20:54:57 h Feb 08 20:57:02 did it for you Feb 08 21:06:12 magic! Feb 08 21:07:54 uh, ECHAN again. sorry. Feb 08 21:52:18 Maemo Auto-Complete bug: Auto-Complete dictionary for Russian language is contaminated strongly with English words. My impression is that if virtual keyboard is set to Russian, and human is meanwhile typing in English on hardware keyboard, the Auto-Complete catches the words as if they were in Russian. Feb 08 22:29:07 of course Feb 08 22:33:24 iirc there's an app to edit the user dictionary Feb 08 22:58:55 oh wait, completion is 'wider' than spellcheck, it also uses history on top of the dictionary. I guess simply deleting history will do the trick Feb 08 22:59:08 disclaimer: afaik Feb 08 22:59:48 not using spellcheck nor autocompletion. Disabled both during 48h after unboxing my first N900 Feb 08 22:59:58 (also N810) Feb 08 23:02:12 I found it pretty pointless and annoying particularly because it remembers also typos and 'atypical words' in history, and it sucks big time to only use the built-in dictionary and update it manually by allowing only those new words in that are really relevant Feb 08 23:03:18 in tests my words per minute *dropped* *significantly* when using autocompletion Feb 08 23:07:55 No idea about spellcheck aka built-in dictionary. My auto-complete uses only history, which I occasionally clean from misprints. How do wpm drop, if typer can always ignore auto-complete? It's not an aggressive auto-correct. Feb 08 23:10:22 I would appreciate if auto-complete would recognise words which deviate from sentence case, like PuTTY, or names like DocScrutinizer05, and keep them this way, exempt from sentence case. Feb 08 23:12:54 One useful and under advertised feature of auto-complete is ability to finish phrases, and not just words. Good morning, for example. However, such phrases have to be manually added to dictionary. And it's unfortunately refusing to remember case in, say, Dunkin' Donuts. Feb 08 23:23:14 And that's taking all the usefulness out of adding such specially cased phrases to the dictionary. Either user auto-completes to Krispy kreme, or user types Krispy Kreme most of the way just to be able to specify this upper case K. Feb 08 23:26:39 Normally cased phrases work magnificently, though. I haven't used Auto-complete Editor for a long time, but it is still a useful ace in the sleeve sometimes. Feb 09 00:12:47 If autocorrect doesnt work well i just disable it Feb 09 00:13:20 otherwise i just make it so it suggests only instead of autocorrect, additionally i disable autocapitalize Feb 09 00:13:24 oksanaa hi :) Feb 09 00:15:27 Auto-capitalise is mostly useful, it knows dots, exclamations and questions. Feb 09 00:16:23 Additionally i disable auto space after dot Feb 09 00:16:34 sometimes it is annoying Feb 09 00:24:23 Well, I would appreciate recognition of triple dot... With which space after dot occasionally interferes. But not always, keyboard behaviour is not necessarily consistent. Feb 09 00:25:09 can i pm you oksanaa ? Feb 09 00:26:44 ok Feb 09 00:39:16 Oksanaa: when you're so obviously experienced in how autocompletion works, I don't understand how you can call "BUG!" the normal and correct behavior of remembering *all* words user types (irrespective if they are in a particular language-specific dictionary or not) Feb 09 00:40:20 DocScrutinizer05: Because the corrector should know what words corresponds to X language and should NOT mix them Feb 09 00:40:21 also why calling it "auto-complete *dictionary*" then? Feb 09 00:40:25 It sorts them into language dictionaries incorrectly. Feb 09 00:40:40 Juesto: and how should it do that? Feb 09 00:40:55 Separate each language Feb 09 00:41:00 Juesto: and how should it do that? Feb 09 00:41:00 Android does this very well Feb 09 00:41:29 I call them dictionaries because they are separate for different languages. If it was all in one mixture, then it wouldn't be a dictionary, but it would be difficult to use. Feb 09 00:41:39 then maybe android has a larger more comprehensive dictionary Feb 09 00:42:01 meh Feb 09 00:42:16 Android has a user dictionary, and a per-language dictionary Feb 09 00:42:27 Juesto: not good Feb 09 00:42:42 and its like saying, a plain database vs a database with folders for each language Feb 09 00:42:44 There should be user's per-language dictionaries Feb 09 00:43:09 Oksanaa: the user dictionary itself has "folders" for each language, so each user dictionary is different between languages Feb 09 00:43:20 Okay, good Feb 09 00:44:01 And GBoard (previously called Google Keyboard, internal name: LatinIME) has implemented a two language mixture of suggestions/corrections Feb 09 00:44:13 But built-in predefined per-language dictionaries aren't helpful if different users have different vocabularies. Unless Google aims for uniformity of users' vocabularies. Feb 09 00:44:32 it does the job well because it detects what language you're writing to provide the right suggestions Feb 09 00:44:37 you said: 1) the keyboard was set to Russian but you typed english words. 2) the history list used for autocompletion stores those english words 3) you imply that those english words are not in the spelling dictionary Feb 09 00:44:43 I do not know for sure but yes they're aiming that Feb 09 00:44:45 so what the heck is the bug? Feb 09 00:45:06 Oksanaa: That's not a bug, instead it is a user mistake. Feb 09 00:45:48 3) do not have a spelling dictionary afaik ; 1) it's possible when virtual keyboard is set to Russian and hardware keyboard to English? Feb 09 00:45:59 you're storing english words in the russian user directionary Feb 09 00:46:15 that's still not a bug, it's you doing things wrong! Feb 09 00:55:40 How to prevent it? Switch virtual keyboard to English just because I am using hardware keyboard? Feb 09 00:58:08 Yes or do not use virtual keyboard, Oksanaa Feb 09 00:58:22 I think it's actually a bug when not noth the hw and virtual kbd are switched insync Feb 09 00:58:35 both* Feb 09 00:58:40 Perhaps Feb 09 00:58:56 but thing is: the hardware keyboard does not have for russian for example Feb 09 01:08:16 There is a way to switch languages on hardware keyboard, akin to Ctrl+Shift, I just don't remember the shortcuts. Feb 09 01:08:45 why do you seem to forget me oksanaa ? Feb 09 01:09:03 and nice, control+shift shortcuts can be hard to learn Feb 09 01:09:08 just like PC shortcuts Feb 09 01:12:49 Смешно. I found that Ctrl+Space switches language... Of virtual keyboard. Well, that's easier than touching the on-screen buttons for switching languages. Feb 09 01:13:48 wasn't that ctrl+space or something like that? Feb 09 01:14:14 shift+space or Fn+space, somesuch Feb 09 01:14:58 oh, you found it, sorry I sound stupid answering next line what you already wrote Feb 09 01:15:05 :) Feb 09 01:15:11 no need to sorry Feb 09 01:15:14 its fine Feb 09 01:15:41 No problem, concurrent answers happen all the time. Fn+space may be @, for example. Shift+space is just space, unless mapped otherwise. Feb 09 01:15:59 only for virtual kbd? that's a terible bug Feb 09 01:16:22 makes no sense at all to switch virtual kbd only, on a hw kbd hotkey Feb 09 01:16:36 And i still have no idea how to make sdk lanucher use xorg instead of xephyr Feb 09 01:18:53 umm what? Feb 09 01:19:56 Juesto has virtual machine with Devuan and attempts to have fremantle-gtk2 running on it, last I remember? Feb 09 01:20:12 the maemo apps are 'hardcoded' for a 800*480 display resolution afaik, won't match nicely to most any of your possible plain xorg displays Feb 09 01:20:17 Oksanaa: i have to wait since it's already on the works by someone else Feb 09 01:20:51 DocScrutinizer05: of course i'll setup a 800x480 display but i want to switch servers Feb 09 01:21:00 because xephyr blocks control+shift Feb 09 01:21:16 Yes, devuan setting up maemo repository or something Feb 09 01:21:32 Ctrl+Shift? What for? Feb 09 01:22:06 Oksanaa: control+shift+ shortcuts Feb 09 01:22:11 aka Fn Feb 09 01:22:34 ..in maemo keyboard Feb 09 01:23:00 http://unix.stackexchange.com/questions/186273/emulate-key-combination-with-xdotool Feb 09 01:23:30 ah that seems better, thank you DocScrutinizer05 Feb 09 01:23:32 will look later Feb 09 01:23:46 bbl Feb 09 01:25:09 https://fedoraproject.org/wiki/Sandboxing etc - google is your friend ;-) Feb 09 01:31:14 Продолжение, if I set Russian as language of hardware keyboard, then Ctrl+Space switches hardware keyboard between Russian and Latin; virtual keyboard isn't affected anymore, no idea where dot is in Latin Feb 09 01:32:17 Of course, I can still use virtual keyboard, with English and Russian. Feb 09 01:41:26 Latin = US layout? Feb 09 01:45:15 https://bugs.maemo.org/show_bug.cgi?id=2501 Feb 09 01:45:17 04Bug 2501: Hardware keyboard doesn't switch Input language when pressing Ctrl+Space Feb 09 01:51:41 Latin layout includes бюхъ, which is messed up for English Feb 09 02:16:00 I never heard of a 'latin' keyboard layout Feb 09 02:19:17 anyway one thing is pretty obvious: Hildon Input Method stuff needs to be FOSS to receive patches. I wonder why nobody seems to be willing to tackle it Feb 09 02:23:56 Because it's confusing. Which library.so is responsible for interpretation of Ctrl+Space? Which package does it belong to? How large the package is? Feb 09 02:29:58 search for hildon input method on https://wiki.maemo.org/Fremantle_closed_packages Feb 09 02:30:09 aka hildon im Feb 09 02:30:42 https://github.com/community-ssu/hildon-input-method-plugins Feb 09 02:32:34 hmm https://github.com/community-ssu/hildon-input-method-plugins/blob/master/hildon-im-western-plugin-common3/hildon-im-western-plugin-common.c >>Copyright (C) 2012 Ivaylo Dimitrov << Feb 09 02:33:24 something around there must do the trick Feb 09 02:33:47 not necessarily this particular plugin Feb 09 02:34:23 also something around there must handle the invocation of vkbd, the shift-lock stuff etc Feb 09 02:36:37 and I'd prolly mutate from Dr Jeckyll to Mr Hyde and start coding away to get shift(-lock) and Fn(lock) state signaled by Indicator LED (pattern) if only I knew where to start Feb 09 02:38:42 WTF is libiphb alias IP heartbeat? Isn't it IP heartbeat that had the vulnerability that allowed heartbleed attack? Feb 09 02:40:17 Oksanaa: how does this sound? https://github.com/community-ssu/hildon-input-method-plugins/blob/master/hildon-im-fkb/hildon-im-word-completer.c Feb 09 02:42:50 HAHA! imengines_wp_set_data("ezitext", (void *)0xBBC58F26); /* WTF ?!? */ Feb 09 02:51:03 Huh, it should recognise all brackets, as well as ", ?, \, ! and mysterious \xC2\xBF and \xC2\xA1 as marks. Yet I often find a word) added into auto-complete with bracket. **** ENDING LOGGING AT Thu Feb 09 03:00:02 2017