**** BEGIN LOGGING AT Sun Jan 29 03:00:00 2017 Jan 29 05:12:34 morning europe Jan 29 05:14:30 bencoh: so this sounds to me that it could be set as changable option in OMP. Or im guessing too much? Jan 29 05:22:54 Was changing battery right now. Had surprising GUI glitch before that, even Ctrl+Backspace didn't work, while powerkey and systemui did work. Wish Ctrl+Shift+P was working... Jan 29 05:31:01 bencoh: i've always thought of that short playback pause caused by cpu resources being used to bring up unlockslider Jan 29 06:24:12 Vajb: you'd have to patch mafw-gst-renderer and expose some option through its API (or through dbus) Jan 29 06:25:03 Vajb: the playback pause doesn't happen when playing with mafw-based players Jan 29 06:30:11 freemangordon: i guess my system is ready... what do i need to download and compile exactly? Jan 29 06:30:22 well i could argue that running anything else, but mafw-based is that more taxing, but honestly i have no idea Jan 29 06:31:43 it's actually roughly equivalent when it comes to audio playback (unlike video) Jan 29 06:34:20 hmm maybe different route through sw layers Jan 29 06:37:56 it's a completely different code but audio playback through dsp still takes cpu anyway Jan 29 06:41:22 beats me Jan 29 07:23:56 Juesto: do git clone of hildon-desktop, it will tell you all the needed dependencies Jan 29 07:24:30 Juesto: hildon-desktop from https://github.com/fremantle-gtk2/ Jan 29 07:24:45 almost all the needed stuff is there as well Jan 29 07:25:28 what extra libs are needed to try launching apps from SDK chroot? Jan 29 07:26:56 freemangordon: differences ? Jan 29 07:27:30 between the gtk versions of hildon Jan 29 07:28:16 NeKit: SDK chroot == ScratchBox? Jan 29 07:28:21 yes Jan 29 07:28:31 Juesto: gtk3 hildon is not ready at all Jan 29 07:28:53 NeKit: I am not aware of any, it should work OOB Jan 29 07:29:17 but, last time I tried was a couple of years ago, so... :) Jan 29 07:39:20 could it be pulse audio related? (audio pause when ui activity does some system work) Jan 29 07:39:49 another thing to try would be renicing player to -20 Jan 29 07:40:23 im only renicing oscp playback to -1, and it still skips Jan 29 07:41:28 bencoh: audio playback through dsp takes a lot less code than with sw decoding Jan 29 07:41:46 and a lot less mem bandwidth Jan 29 07:41:52 KotCzarny: try to move the player to multimedia(or somesuch) cgroup Jan 29 07:42:06 fmg: got a handy c func? Jan 29 07:42:30 KotCzarny: no C code, just look in /system Jan 29 07:42:37 uhum Jan 29 07:42:59 also, there was something in /etc, ohmd related Jan 29 07:43:13 can't remember the details Jan 29 07:43:49 https://bugs.maemo.org/show_bug.cgi?id=7132 Jan 29 07:43:51 04Bug 7132: audio hiccups in fm radio when sleeping Jan 29 07:43:52 he he. Jan 29 07:44:09 surprisingly its still 'NEW' Jan 29 07:45:14 huhu Jan 29 07:45:29 KotCzarny: it could be pulse related, although I doubt it, since it should occur with mafw as well then Jan 29 07:45:45 and my players already are in the same xpolicy group as mafw Jan 29 07:46:06 freemangordon: why NEWS and README is empty? Jan 29 07:46:41 aren't they always? :D Jan 29 07:46:44 (j/k) Jan 29 07:46:53 he is a dev, not doc person Jan 29 07:47:00 ;) Jan 29 07:47:50 configure doesnt even exist lol Jan 29 07:50:26 um Jan 29 07:50:42 im a bit stuck here... Jan 29 07:50:44 Juesto: which pakcge is that? Jan 29 07:50:48 *package Jan 29 07:50:59 downloaded fremantle-gtk2/hildon-desktop Jan 29 07:51:06 and i dont know to ./configure Jan 29 07:51:06 downloaded? Jan 29 07:51:13 yeah, git clone Jan 29 07:51:16 ah Jan 29 07:51:28 you should dbpkg-buildpackage -rfakeroot -b Jan 29 07:51:35 *dpkg-buildpackage Jan 29 07:51:56 Juesto: make sure to not do ^^^ as root Jan 29 07:52:00 well the instructions are wrong lol Jan 29 07:52:03 :/ Jan 29 07:52:09 im rooted right now Jan 29 07:52:14 and its in a root dir Jan 29 07:52:19 Juesto: which instructions?!? Jan 29 07:52:25 this is debian source Jan 29 07:52:26 INSTALL Jan 29 07:52:31 Right in the repo Jan 29 07:52:32 lol Jan 29 07:52:57 * Juesto shrugs Jan 29 07:52:58 well, if you want ./configure, you should do autofoo stuff by hand Jan 29 07:53:10 but you'd better use debian scripts Jan 29 07:53:23 autogen.sh and ./configure-ac Jan 29 07:53:26 there is Jan 29 07:53:35 yes, but better nto do it Jan 29 07:53:49 as debian/rules does more things Jan 29 07:53:52 whats the problem with rooted enviorment for the package building Jan 29 07:54:16 you should not pass -rfakeroot to dpkg-buildpackage in that case Jan 29 07:54:28 as you're already root Jan 29 07:54:34 just that Jan 29 07:55:24 KotCzarny: look in /usr/share/ploicy/ Jan 29 07:55:30 grrrr Jan 29 07:55:42 "/usr/share/policy" Jan 29 07:56:39 fmg, will do, but i suspect its cpu/mem bandwidth starvation/stall Jan 29 07:56:55 does stutter happen all the time? Jan 29 07:57:06 most noticable on screen lock/unlock Jan 29 07:57:06 what is the cpu load? Jan 29 07:57:19 and when notifications show Jan 29 07:57:29 KotCzarny: the it is a priority thing Jan 29 07:57:32 *then Jan 29 07:57:41 aiui Jan 29 07:57:47 will try -20 first then cgroups, then both Jan 29 07:58:04 (as soon as i find the charger, poor thing is hungry) Jan 29 07:58:19 if you do cgroups, you should not change the priority afaik Jan 29 07:59:50 freemangordon: got like... 10 unobtainable package dependencies Jan 29 08:00:07 5am in here Jan 29 08:00:10 KotCzarny: you should move to either "media" or "mediasrc" group Jan 29 08:00:18 Juesto: that's normal :) Jan 29 08:00:43 Juesto: just git-clone the relevent packages, build and install them Jan 29 08:01:06 you'll have more missing stuff, but the code is on the repo Jan 29 08:01:21 installing what's standard/exists Jan 29 08:01:39 full xserver dev stuff apparently Jan 29 08:02:17 if there is a dependency that is not on github, then you should the standard package by using apt-get Jan 29 08:02:30 *should install Jan 29 08:02:50 I already did that lol Jan 29 08:02:57 good Jan 29 08:02:57 except the github ones Jan 29 08:03:23 how did you found what -dev packages are needed? Jan 29 08:03:26 *find Jan 29 08:03:37 * freemangordon needs more coffee Jan 29 08:03:43 fmg, you can copypaste your converstion with juesto into BUILD file ;) Jan 29 08:03:52 will also serve as FAQ Jan 29 08:03:59 KotCzarny: /me is developer, not doc writer :p Jan 29 08:04:22 developers do copypaste a lot sometimes :P Jan 29 08:04:29 KotCzarny: Even if I can take over the everything, I refuse to Jan 29 08:04:38 so you should be proficient with that skill Jan 29 08:04:50 freemangordon: should i just download the entire fremantle-gtk2 ? Jan 29 08:05:07 download = git clone Jan 29 08:05:13 Juesto: actually yes, but I have no idea if it is possible Jan 29 08:05:22 to do it with one command only Jan 29 08:05:25 it is possible Jan 29 08:05:31 well not Jan 29 08:05:36 you have to git clone package-by-package Jan 29 08:05:51 but yes, everything there is needed Jan 29 08:06:24 those are about maybe 7-% of fremantle core packages Jan 29 08:06:27 daaamn Jan 29 08:06:31 70% Jan 29 08:07:10 KotCzarny: do you think it is "fair" /me to take all the job? Jan 29 08:07:21 not at all Jan 29 08:07:32 and do you think it makes sense if nobody else is interested? Jan 29 08:07:37 but right now you have to repeat all build steps for new people Jan 29 08:07:51 and having this thing copypasted would save you some time Jan 29 08:07:52 KotCzarny: hopefully that would change soon Jan 29 08:08:05 I'm going to write down instructions perhaps Jan 29 08:08:35 as I expect to have a repo on either devuan or maemo servers someday Jan 29 08:08:46 so no manual build will be needed Jan 29 08:09:19 I am wating to see what Wizzup will do with autobuilder and repo with devuan guys Jan 29 08:09:32 if he fails, we'll setup a deb repo on maemo servers Jan 29 08:10:00 hopefully, if the community is fine with that Jan 29 08:11:25 Juesto: good Jan 29 08:19:06 10:08 < Juesto> I'm going to write down instructions perhaps Jan 29 08:19:10 that's the spirit :) Jan 29 08:19:22 :nod: Jan 29 08:19:23 KotCzarny: did you try the /usr/share/policy thing? Jan 29 08:19:50 I guess he is still looking for a charger :p Jan 29 08:20:23 done, i checked out everything Jan 29 08:20:47 freemangordon: anywhere i can find the dependency chain? Jan 29 08:20:51 bencoh: so its an entry in some system text file? Jan 29 08:21:32 am i the first into this? Jan 29 08:21:38 Juesto: debian/control? Jan 29 08:21:43 i have the feeling im not Jan 29 08:21:46 Juesto: dunno, every package dependencies are in debian/control Jan 29 08:21:57 KotCzarny: no idea yet Jan 29 08:22:01 hm Jan 29 08:22:17 Juesto: there is a command to install those dependencies, gimme a minute Jan 29 08:22:53 freemangordon: what i am looking for is the entire dependency tree Jan 29 08:22:55 interestingly enough there is a rawcam entry there Jan 29 08:23:15 KotCzarny: just move the pid to the appropriate dir in /system Jan 29 08:23:19 not what each repository comes with x packages Jan 29 08:23:46 Juesto: i don;t think you can get the whole dependency chain Jan 29 08:24:01 but i am not *thet* good in debian packaging, so... Jan 29 08:24:06 *that* Jan 29 08:25:45 freemangordon: okay i tried.... apparently i have to hack up something where it automatically cds and builds Jan 29 08:25:55 Should i just skip deps or they are required? Jan 29 08:26:00 hmm, fmg, which /system are you talking about? Jan 29 08:26:07 Juesto: re if you're the first - no, you're not Jan 29 08:26:10 apparently the default is build both src and bin Jan 29 08:26:37 [classify system] ? Jan 29 08:26:48 [classify mediarend] ? Jan 29 08:26:56 /sys I suppose Jan 29 08:27:03 err maybe not Jan 29 08:27:06 /usr/share/policy/etc/rx51/syspart.conf Jan 29 08:27:08 Juesto: it might be faster is you just cd to the needed package directory and do "dpkg-buildpackage -b" and then install the produced .debs Jan 29 08:27:10 this one probably Jan 29 08:27:27 * freemangordon checks Jan 29 08:28:14 yeah, /syspart Jan 29 08:28:33 freemangordon: im trying a faster way Jan 29 08:28:42 automation Jan 29 08:28:57 ok Jan 29 08:29:07 KotCzarny: hmm, seems there are no relevant dirs there Jan 29 08:29:20 so you should add your binary in the policy file Jan 29 08:29:27 and probably restart ohmd Jan 29 08:29:50 will try classify mediasrc Jan 29 08:29:56 I see mafw-gst-renderer there Jan 29 08:30:00 so you might be right :) Jan 29 08:30:13 if yes that'd be really cool Jan 29 08:30:43 KotCzarny: /usr/share/policy/etc/current/syspart.conf Jan 29 08:31:13 how do i restart ohmd? Jan 29 08:31:17 stop/start? Jan 29 08:31:26 * KotCzarny is services n00b Jan 29 08:31:31 * freemangordon wonders if ohmd supports something like run-parts Jan 29 08:31:43 yep, stop/start shoud do the job Jan 29 08:32:52 syspart didnt change a thing Jan 29 08:33:10 renicing oscp audio renderer thread to -20 worked Jan 29 08:33:19 so i was right about cpu starvation Jan 29 08:33:33 KotCzarny: try to reboot Jan 29 08:34:06 it could be something else besides ohmd that needs to be restarted so syspart change to take effect Jan 29 08:34:39 KotCzarny: also, -20 is too much imo Jan 29 08:34:42 i know Jan 29 08:34:48 ok Jan 29 08:34:50 i was just checking if it changes anything Jan 29 08:34:53 yeah Jan 29 08:35:17 unlock takes a tiny bit longer, but no audio skip Jan 29 08:35:35 i might add it as a config option Jan 29 08:35:51 btw. there are some binary things in /usr/share/policy/rules/rx51 Jan 29 08:35:59 maybe policies need to be 'recompiled' ? Jan 29 08:36:01 yes, some prolog stuff :) Jan 29 08:36:10 could be Jan 29 08:36:32 lets see if they change after reboot Jan 29 08:37:10 KotCzarny: https://github.com/community-ssu/policy-settings-rx51 Jan 29 08:37:29 btw. can i check current cgroup in /sys somewhere? Jan 29 08:37:47 dunno Jan 29 08:37:54 it should be possible, but... Jan 29 08:38:04 im trying quick and dirty Jan 29 08:38:11 repeating a mass build package Jan 29 08:38:20 and then dpkg -i *.deb Jan 29 08:38:23 /proc//cgroup Jan 29 08:38:36 currently 7:freezer,memory,cpu:/ Jan 29 08:38:39 reoot time. Jan 29 08:39:43 hmm, fun fact, on reboot when x dies i saw last pic i was viewing in pic viewer (or maybe captured by camera) Jan 29 08:41:01 Juesto: the first to try was NeKit, on android phone, in arch linux chroot iirc :) Jan 29 08:41:16 NeKit: right ^^^ Jan 29 08:42:00 reboot didnt change a thing Jan 29 08:42:02 what a weird mix, lol Jan 29 08:42:05 still skips Jan 29 08:42:08 what are you trying btw? gtk3 on maemo? Jan 29 08:42:15 at least those 8 cores can be put to some use Jan 29 08:42:33 bencoh: fremantle-gtk2 on devuan\ Jan 29 08:42:45 ah Jan 29 08:43:04 KotCzarny: could you check in which cgroup it is now? Jan 29 08:43:14 yes, the one i've pasted Jan 29 08:43:18 09:38 #maemo |> currently 7:freezer,memory,cpu:/ Jan 29 08:43:23 same as mafw-renderer Jan 29 08:43:38 could you check where is PA? Jan 29 08:43:45 juesto.pw:9600 Jan 29 08:43:47 VNC ^ Jan 29 08:44:08 fmg: same Jan 29 08:44:34 yeah Jan 29 08:45:16 checked some other process for comparison: /usr/libexec/gst-video-thumbnailerd, 7:freezer,memory,cpu:/applications/standby Jan 29 08:45:29 ... Jan 29 08:46:08 so either policies arent working or something else is wrong Jan 29 08:46:44 hmm, yeah, there is somethign weird Jan 29 08:47:01 is SELinux enabled? Jan 29 08:47:04 anyway, me gtg Jan 29 08:47:12 * Juesto sighs Jan 29 08:47:13 bye, bbl Jan 29 08:47:24 i guess i'll sleep Jan 29 08:47:39 Juesto: I don;t think you'll need my help anytime soon Jan 29 08:47:47 you have lots of packages to build Jan 29 08:47:56 :) Jan 29 08:47:59 actually yes, still missing stuff freemangordon Jan 29 08:48:40 you have to build and install ~30-40 packages before hildon-desktop is buildable Jan 29 08:49:05 I really hope we'll have a functional debian repo soon Jan 29 08:49:18 but until then... Jan 29 08:49:22 anyway, bye Jan 29 08:49:32 fun fact, oscp pauses playback during phone call, didnt knew it does that :> Jan 29 08:49:53 most likely audiodevice stops accepting data and it waits Jan 29 08:50:15 welp, do i need to use upstart? Jan 29 08:50:24 should i switch to ubuntu? Jan 29 08:50:25 juesto: recommended Jan 29 08:50:30 devuan Jan 29 08:50:55 we hate systemd here Jan 29 08:51:11 KotCzarny: It seems to remove sysv and put upstart Jan 29 08:51:22 :/ Jan 29 08:51:27 i am neutral Jan 29 08:51:27 as long the system boots.. Jan 29 08:51:29 :) Jan 29 08:51:41 i think upstart honors sysv scripts Jan 29 08:51:59 and upstart itself is very simple to use Jan 29 08:52:15 (i know, i was using sysv before and had to change something in n900) Jan 29 08:53:09 hm Jan 29 08:53:11 interesting Jan 29 08:53:46 http://pastebin.com/raw/dB6x2Wc9 Jan 29 08:53:52 cgroups are b0rken? Jan 29 08:54:11 maybe that policy thing doesnt use cgroups but something else? Jan 29 08:54:19 (stock n900 here btw.) Jan 29 08:54:31 bencoh, can you paste /proc/cgroups if you have cssu? Jan 29 08:55:39 cpu 7 6 1 Jan 29 08:55:42 same for memory and freezer Jan 29 08:57:34 In what order i should install the debs Jan 29 08:58:57 ... Jan 29 08:59:08 either i install upstart-dev or maemo-system-sevices-dev Jan 29 08:59:10 :| Jan 29 09:01:28 Juesto: oh, don't install upstart-dev, sorry, it was deprecated some time ago Jan 29 09:01:49 I have to remove it from the repo Jan 29 09:01:52 :/ Jan 29 09:01:56 i'll remove that then Jan 29 09:02:18 is is superseded by maemo-system-sevices-dev Jan 29 09:02:18 freemangordon: also https://defuse.ca/b/MXjB2GJk Jan 29 09:02:44 "Sorry, the paste you were looking for could not be found. " Jan 29 09:03:02 lowercase g sorry Jan 29 09:03:57 https://defuse.ca/b/MXjB2gJk Jan 29 09:04:01 Juesto: hmm, make sure you're on the correct branch Jan 29 09:04:14 im on fremantle-gtk2 Jan 29 09:04:18 you need branches with gtk2 or clutter1 in the name Jan 29 09:04:28 no, for individual packages that is Jan 29 09:04:35 ah Jan 29 09:04:42 git branch -a will list all the branches Jan 29 09:04:47 i have a mixed system Jan 29 09:04:49 ah Jan 29 09:04:52 * freemangordon is afk Jan 29 09:09:16 hm, btw. freaking old, but ... http://wiki.maemo.org/Maemo_on_Debian Jan 29 09:09:27 i have a headache Jan 29 09:09:54 I dont think they got anywhere, but ... Jan 29 09:53:26 how do i switch branches Jan 29 09:53:32 im used to tortoisegit Jan 29 09:54:06 git checkout Jan 29 09:55:15 doesnt match it says Jan 29 09:55:51 ? Jan 29 09:56:33 git checkout gives pathspec didnt match any files known to git Jan 29 09:56:34 :/ Jan 29 09:59:30 i had to use -B Jan 29 10:01:16 [rule /opt/oscp/bin/oscp] Jan 29 10:01:16 group mediasrc Jan 29 10:01:22 this works, weird Jan 29 10:02:12 cant make it skip anymore (lock/unlock, running telephony app etc) Jan 29 10:03:08 bencoh, freemangordon: ^^^ Jan 29 10:03:39 now, how do i sensibly modify syspart.conf on install? Jan 29 10:04:02 KotCzarny: why didn't it work earlier then? Jan 29 10:04:20 (because that's exactly what fmg was referring too ;p) Jan 29 10:04:38 (and what I was referring to when I said I found mafw-gst-renderer there) Jan 29 10:04:43 bencoh, i've tried to add it in [classify mediasrc] Jan 29 10:04:46 that didnt work Jan 29 10:04:51 ah, well Jan 29 10:05:07 but amazingly, cant make it skip anymore :) Jan 29 10:05:23 Finally im managing someting!! Jan 29 10:05:25 even running cpu intensive things (as starting oscp-remote.py Jan 29 10:05:26 something* Jan 29 10:05:42 KotCzarny: I don't think it was cpu-bound anyway :) Jan 29 10:05:43 i feel horrible Jan 29 10:05:51 :) Jan 29 10:06:07 but great, thanks a lot! :) Jan 29 10:06:10 i cant imaging running .py script affecting audio Jan 29 10:06:14 *imagine Jan 29 10:06:30 KotCzarny: starting oscp-remote made it stutter? Jan 29 10:06:55 freemangordon: Finally! compiled libmatchbox2 in upstream-gtk2 Jan 29 10:07:02 yeah, because running py scripts require quite a lot of cpu at start Jan 29 10:07:20 and even sometimes opening new windows (info, help etc) Jan 29 10:09:16 now im actually getting somewhere Jan 29 10:10:05 KotCzarny: well, good to know anyway Jan 29 10:10:12 I wonder if this file can be fragmented Jan 29 10:10:25 because you don't want to alter it on every player package installation Jan 29 10:10:34 i dont see any include or similar entries Jan 29 10:10:44 I saw a similar one with rawcam Jan 29 10:10:51 (similar file) Jan 29 10:11:02 maybe it manages its own cgroup? Jan 29 10:11:09 but I haven't dived further Jan 29 10:11:37 i will have to invent some sed magic for oscp now (with checking if modification is needed) Jan 29 10:11:40 I'll have a go at it this evening (in case there's anything left to do, that is ;p) Jan 29 10:11:51 I'd rather not, but ... Jan 29 10:12:08 actually same goes for xpolicy (pulse policy) btw Jan 29 10:12:27 i think pa doesnt have to be modified Jan 29 10:12:32 I had to manually add games/emulators to [games] and players/radio/whatever to relevant groups Jan 29 10:12:39 unless you do some voice related work Jan 29 10:12:40 not regarding that issue no Jan 29 10:12:58 but if you want to be able to listen to music while still in silent mode ... Jan 29 10:13:05 o.O Jan 29 10:13:10 * KotCzarny checks Jan 29 10:13:31 oh wow Jan 29 10:13:34 yeah ;p Jan 29 10:13:37 they really messed that one Jan 29 10:13:45 you are right :> Jan 29 10:13:57 otoh, silent is silent, right? :P Jan 29 10:14:10 the thing is ... connecting headphones doesn't change that behavior Jan 29 10:14:44 and default mediaplayer on the other hand can always play music (be it in silent or general, with or without headsets) Jan 29 10:15:17 so you'd expect either 1. a consistent experience, or 2. hearing something when connecting headphones, even in silent mode Jan 29 10:15:21 anyway :) Jan 29 10:16:18 (another solution would be to patch responsible daemon to switch to general when pluggin headphones and revert back when unplugging - and even that is not a proper solution) Jan 29 10:20:20 hmm, which group should i add to? Jan 29 10:20:57 player, iirc Jan 29 10:21:06 just look for mafw in xpolicy ;> Jan 29 10:21:06 starting to write notes Jan 29 10:21:10 Juesto: yay Jan 29 10:21:42 and just covering what i needed to do specific to the thing Jan 29 10:22:09 i am omitting all the deps that the user can install from the distro Jan 29 10:29:51 bencoh, what service shall i restart for xpolicy.conf ? Jan 29 10:30:10 pulse Jan 29 10:32:07 stop: Unknown job: pulse Jan 29 10:33:09 isnt pulseaudio? Jan 29 10:33:15 oh nvm Jan 29 10:34:16 yeha, pulseaudio Jan 29 10:35:05 fixed silence in silent mode, but it skips again Jan 29 10:36:07 hrrm Jan 29 10:37:06 check resources usage? Jan 29 10:37:17 are you sure the media isnt corrupted? Jan 29 10:38:07 after restarting ohm after pulseaudio fixed it Jan 29 10:38:34 2 essential fixes in one day, yay Jan 29 10:38:47 now to add them to installer. Jan 29 10:39:16 freemangordon: i use the gtk2 branch for libhildondesktop anyway ? Jan 29 10:39:21 it appears older Jan 29 10:39:36 9 commits ahead and 6 behind master Jan 29 10:40:15 freemangordon: (re ping) I have time to work on it in the eve. So I expect first results tonight Jan 29 10:40:34 oh hey there Jan 29 10:40:48 hi Jan 29 10:40:58 im going through the package building as you saw... :) Jan 29 10:41:54 great. I need to read the backlog in full still Jan 29 10:43:05 KotCzarny: wait, restarting pulse after ohm broke the stuttering thing? Jan 29 10:43:16 and restarting ohm after pulse fixes it back? Jan 29 10:46:42 Juesto: just curious, for what device are you building Jan 29 10:47:01 Wizzup: none Jan 29 10:47:09 ok, just your laptop/desktop? Jan 29 10:47:24 yep, rebasing for devuan... Jan 29 10:47:37 see the conversation i had yesterday Jan 29 10:51:02 Juesto: by rebasing, you mean building? Jan 29 10:51:08 Or do you mean you're modding the debian build scripts? Jan 29 10:51:25 Wizzup: just building Jan 29 10:52:16 again, see the conversation yesterday, started on -ssu then moved here. Wizzup Jan 29 10:54:57 ok, I read the log. I just wanted to check. Jan 29 10:55:06 Since you said 'rebasing' :) Jan 29 11:01:46 Juesto: hi! I'm looking into your login problem Jan 29 11:16:20 bencoh: yeah Jan 29 14:03:48 bencoh, oscp 0.9.67 in repos, no skipping, no muting ;) Jan 29 14:04:06 KotCzarny: in free? Jan 29 14:04:19 soon, if you want source you can grab my arch package Jan 29 14:04:33 i have yet to repackage it for maemo's autobuilder Jan 29 14:04:39 nice Jan 29 14:05:07 which will be adventure in itself, right now i dont have to care about ancient autobuilder's gcc Jan 29 14:05:54 that shouldn't be that of a problem Jan 29 14:06:02 it is for libav Jan 29 14:06:02 ancient ffmpeg, on the other hand... Jan 29 14:06:14 do you seriously intend to ship with a builtin ffmpeg? Jan 29 14:06:20 (and seriously drop libav ...) Jan 29 14:06:31 nope, i will probably make oscp-libav package Jan 29 14:06:41 *sigh* Jan 29 14:06:48 iirc there is already a not-to-old ffmpeg in maemo Jan 29 14:06:51 libav is cleaner and saner than ffmpeg Jan 29 14:07:32 ffmpeg is 'oooh, a patch!, lets add it' Jan 29 14:07:56 okay, I wont feed it ;p Jan 29 14:07:58 anyway Jan 29 14:10:41 http://maemo.org/packages/view/libavcodec53/ Jan 29 14:10:43 hehe Jan 29 14:10:52 so eventfull Jan 29 14:12:16 can i see package voting status somewhere ? Jan 29 14:15:41 it seems that ffmpeg/libav packages were created by felipec (nokia employee) and apparently promoted without any voting Jan 29 14:20:37 libav-0.7 in repos is.. well, ancient Jan 29 14:21:04 and ffmpeg is even more ancient Jan 29 14:25:56 I think android_808 built a new one Jan 29 14:26:00 but maybe he didn't uploadi t Jan 29 14:29:22 did he manage to build neon optimized one? Jan 29 14:29:35 afair old gcc fails to compile optimizations Jan 29 14:36:29 KotCzarny: I still build ffmpeg with old gcc, but it works fine enough, and you can always remove intrinsics Jan 29 14:36:53 since most interesting ffmpeg optims are built using yasm Jan 29 14:37:06 (they're neither intrinsics nor gcc inline afaict) Jan 29 14:53:12 oo cool https://www.youtube.com/watch?v=JgdNwT1NOzo Jan 29 14:54:10 Juesto: yes, gtk2 branch for libhildondesktop Jan 29 14:54:20 KotCzarny: so, what was the final fix? Jan 29 14:54:39 fmg: adding as a [rule /some/path/oscp] Jan 29 14:54:57 KotCzarny: also, it is not that good idea to sed syspart.conf Jan 29 14:55:11 fmg, got any better way? Jan 29 14:55:25 and i've added few safeguards Jan 29 14:55:49 not sure, maybe check ohmd if it supports $dir.d run-parts type of configuration Jan 29 14:56:17 Pali: ^^^ is anything besides syspart.conf read by ohmd? Jan 29 14:56:37 if it doesn't, maybe patching ohmd would be an option? Jan 29 14:56:43 KotCzarny: if not, maybe propose a patch to CSSU :p Jan 29 14:56:52 ah, bencoh beat me to it :) Jan 29 14:56:55 fmg: what about stock fremantle? :P Jan 29 14:56:56 ;) Jan 29 14:57:01 who cares Jan 29 14:57:02 my main device is stock ;) Jan 29 14:57:11 time to move to CSSU then Jan 29 14:57:32 it works, why should i? ;) Jan 29 14:57:33 yeah, still dont understand why you still carry a stock device Jan 29 14:57:41 because it ""works"" Jan 29 14:57:43 it's full of bugs Jan 29 14:57:50 which bugs affect me? Jan 29 14:57:55 memleaks for sure Jan 29 14:58:09 didnt notice, and sometimes battery runs dry anyway Jan 29 14:58:13 KotCzarny: no, really, I will downvote such package if it enters extras-testing Jan 29 14:58:22 battery runs dry because of bugs as well :D Jan 29 14:58:33 KotCzarny: wgat about openssl? Jan 29 14:58:36 *what Jan 29 14:58:51 certificates? Jan 29 14:58:51 freemangordon: reminds me I still havent try your foss devlock replacement Jan 29 14:58:57 tried* Jan 29 14:59:10 iisn;t it in CSSU? Jan 29 14:59:15 i dont use browser much, and when i do it mostly works Jan 29 14:59:43 KotCzarny: it is not about the browser, but anything internet related Jan 29 14:59:52 but yeah, it is up to you Jan 29 15:00:00 internet, but only https etc Jan 29 15:00:08 ssl, tls, etc Jan 29 15:00:21 for plain nonencrypted connections ssl isnt used Jan 29 15:00:25 sure Jan 29 15:00:41 freemangordon: hmm it might be in -testing, but I have a -stable device with cherry-picked changes Jan 29 15:00:55 bencoh: ah, right Jan 29 15:01:01 could be like that Jan 29 15:01:09 (that's quite stupid as well, tbh) Jan 29 15:01:20 (running a -stable device these days) Jan 29 15:01:21 anyway, I expect new stable soon Jan 29 15:01:24 yay :) Jan 29 15:01:33 has it been tested enough though? Jan 29 15:01:51 bencoh: yeah, we have a new maintainer for -testing and hopefully merlin1991 will issue a new stable Jan 29 15:02:04 :) Jan 29 15:02:27 it has been, but there is(was) noone to make the new release Jan 29 15:02:33 fmg: check syspart.conf, quite a lot of things are added to that file (mahjong, lmarbles, chess, gnuchess) Jan 29 15:02:47 KotCzarny: all those are stock nokia applications Jan 29 15:03:04 so they were added by Nokia Jan 29 15:03:06 regarding cssu packaging, I know we've already discussed it, but ... how feasible would building cssu packages using autobuilder be? Jan 29 15:03:21 then its very important to add include.d functionality to syspart.conf Jan 29 15:03:36 KotCzarny: yeah Jan 29 15:03:40 I guess fully feasible, just somebody has to do it Jan 29 15:03:41 same goes for xpolicy.conf Jan 29 15:03:43 otherwise ANY application not using mafw for audio will glitch Jan 29 15:03:45 KotCzarny: :nod: Jan 29 15:04:12 KotCzarny: I suspect nokia's policy back then was "thou shalt not play audio without mafw" anyway Jan 29 15:04:45 (which would be quite a good idea if it really helped saving power ... but I doubt it does) Jan 29 15:04:47 which is not very smart, as afaik QT apps does not play through it Jan 29 15:04:56 yeah they don't Jan 29 15:05:02 phonon uses gstreamer directly Jan 29 15:05:03 bencoh: it is not about the power, but about the priorities Jan 29 15:05:14 that too, but still Jan 29 15:06:53 hmm, btw Jan 29 15:07:05 freemangordon: have you ever worked with notifications on hildon? Jan 29 15:07:11 is there any way to quickly grep sources repo for possible syspart.conf modifications? Jan 29 15:07:17 it looks like notify_update_notification() does not work properly Jan 29 15:07:25 which is a pity Jan 29 15:07:28 bencoh: could be Jan 29 15:07:41 I used them, but it was a while ago Jan 29 15:07:59 it means you have to close previous notifications and open a new one from your app, which is racy unless you open the new one in the closed cb of the old one Jan 29 15:08:05 which is ... spaghetti Jan 29 15:08:19 KotCzarny: even if there was, I doubt you'll find anything Jan 29 15:08:24 bencoh: :nod: Jan 29 15:08:47 KotCzarny: modifying others config files is not the correct approach Jan 29 15:08:58 well, maybe I could try and fix notification support in our libnotify or in hildon itself instead of kludging app Jan 29 15:09:15 fmg, right now there is no other way to fix this bug (because it is a bug in os) Jan 29 15:09:26 notifications are in hildon iirc Jan 29 15:09:46 what about opt-in with user requester? would that be passable? Jan 29 15:10:11 KotCzarny: toldya, the correct way is to raise a bug against CSSU and ideally to provide a patch Jan 29 15:10:29 got cssu bugtracker link handy? Jan 29 15:10:31 freemangordon: we use a generic lib (libnotify) between applications and hildon Jan 29 15:10:36 so bug could sit there as well Jan 29 15:10:51 yeah, could be Jan 29 15:11:01 KotCzarny: bugs.maemo.org Jan 29 15:11:03 I'll try and have a look Jan 29 15:11:07 cool Jan 29 15:11:26 KotCzarny: I guess there are already a couple of similar bugs Jan 29 15:11:34 KotCzarny: wait... Jan 29 15:12:05 maybe it is a better idea /me to put ohmd on cssu github repo and you to raise an issue there Jan 29 15:12:21 k, let me know when to post report Jan 29 15:12:40 I know it will take a while until it hits -stable, but this is the correct way Jan 29 15:12:42 ok Jan 29 15:13:19 are we still using stock ohmd? Jan 29 15:13:29 https://bugs.maemo.org/show_bug.cgi?id=8699 Jan 29 15:13:30 04Bug 8699: Sound pauses when screen is manually blanked or unbanked with the slider switch Jan 29 15:13:45 bencoh: yes Jan 29 15:13:49 wow :) Jan 29 15:13:50 this is one of the reports Jan 29 15:13:52 2010 ;) Jan 29 15:13:57 yeah Jan 29 15:14:16 hitting lock when music plays is easiest to trigger this bug Jan 29 15:14:56 fmg, there is similar bug in pulseaudio's xpolicy.conf, with similar fix Jan 29 15:15:47 freemangordon: no idea what is used by ohmd Jan 29 15:16:41 anyway, ohmd is open source and maemo5 source code can be found in meego ohmd git history Jan 29 15:18:35 https://bugs.maemo.org/show_bug.cgi?id=11099 Jan 29 15:18:36 04Bug 11099: Audio interruptions when locking screen or receiving notifications Jan 29 15:19:04 maybe i shall grab all those bugs and move to some container-bug-report? Jan 29 15:20:48 KotCzarny: it is said to be fixed in libplayback, lemme check Jan 29 15:22:13 libplayback? Jan 29 15:23:31 yes Jan 29 15:23:40 KotCzarny: I guess you should use it Jan 29 15:23:57 KotCzarny: see http://maemo.org/api_refs/5.0/5.0-final/libplayback-1/playback_8h.html Jan 29 15:24:10 that thing is ... funny Jan 29 15:24:29 if its better than pulseaudio api, why not, i've already added specific n900's pa interface Jan 29 15:24:51 wth, it plays via dbus? Jan 29 15:24:58 look at https://talk.maemo.org/showpost.php?p=899694&postcount=1 Jan 29 15:25:25 * libplayback offers the following features: Jan 29 15:25:25 * - Audio output when silent mode is active Jan 29 15:25:25 * - Notifications that we should stop playing (f.e. incoming call) Jan 29 15:25:25 * - Override of privacy mode / silent mode for our own audio stream Jan 29 15:25:25 * if the playback class is set to "PB_CLASS_CALL". Jan 29 15:25:25 * - Simultaneous mute/unmute of all mics (seen on dbus: TV out, headset, headmike, microphone) Jan 29 15:26:06 I wonder if it allows proper handling of volume keys when "locked" Jan 29 15:26:28 the fact that you have to modify app to run is pretty annoying though Jan 29 15:26:35 +1 Jan 29 15:27:14 KotCzarny: iiuc it doesn;t play through dbus Jan 29 15:27:17 hmm, nokia-binaries Jan 29 15:27:27 means it's closed source ... shit Jan 29 15:27:31 yes it is Jan 29 15:28:22 it uses /org/maemo/Playback/Manager Jan 29 15:29:49 http://maemo.org/api_refs/5.0/5.0-final/libplayback-1/ Jan 29 15:29:52 har har Jan 29 15:32:26 KotCzarny: http://maemo.org/api_refs/5.0/5.0-final/libplayback-1/files.html Jan 29 15:32:37 and http://maemo.org/api_refs/5.0/5.0-final/libplayback-1/playback_8h.html Jan 29 15:33:39 KotCzarny: also, look at the example code attached to the post Jan 29 15:33:49 yeah, saw that Jan 29 15:33:58 looks like what you need Jan 29 15:34:08 but c++/object oriented code makes me dizzy Jan 29 15:34:40 this is pure C code, where did you find c++? Jan 29 15:35:06 anything with g_loop and callbacks Jan 29 15:35:19 this is C :) Jan 29 15:35:25 glib Jan 29 15:35:26 not c++ specifically, but i meant this model Jan 29 15:35:49 well, the whole maemo is like that Jan 29 15:36:46 looks like Playback.Manager actually belongs to ohmd (or rather ohm-session-agent) Jan 29 15:37:02 brings us back to read ohmd code to understand what it does :) Jan 29 15:37:08 :> Jan 29 15:37:20 http://wiki.maemo.org/Documentation/Maemo_5_Developer_Guide/Architecture/Multimedia_Domain Jan 29 15:37:59 freemangordon: but ... that's cheating! :* Jan 29 15:38:18 i like the simple model of 'here's the file, here's the device, i decode file and write to device' ;) Jan 29 15:39:04 KotCzarny: that might work if it is only you, your file and the kernel but no other applications around :) Jan 29 15:39:45 fmg, or if the policies are implanted into device drivers ;) Jan 29 15:39:58 KotCzarny: however, you still might want to propose a patch for ohmd Jan 29 15:40:05 as it looks way simpler Jan 29 15:40:38 i feel that main source of slowness nowadays is shuffling data around needlessly Jan 29 15:42:26 fmg, i think oscp isnt the only app that could benefit from such option Jan 29 15:42:40 sure Jan 29 15:42:59 interestingly enough mplayer links against libplayback Jan 29 15:43:14 I wonder why and what they did there ... because afaict it still stutters Jan 29 15:43:29 bencoh: I guess to report playback state and to receive cb events (phone call etc) Jan 29 15:43:30 maybe they didnt fix libplayback well enough? Jan 29 15:43:46 try that example from the thread? Jan 29 15:43:51 mediaplayer stutters? this is new to me Jan 29 15:43:59 mplayer != mediaplayer Jan 29 15:44:05 ah, mplayer Jan 29 15:44:10 yeah, mplayer Jan 29 15:44:21 which is why I find it odd Jan 29 15:44:34 maybe libplayback is used but not in the correct way Jan 29 16:02:38 dunno Jan 29 16:04:41 what the hell is this Jan 29 16:04:48 sslayer@portatil:~/Proyectos/neo900$ host radio Jan 29 16:04:48 radio has address 127.0.53.53 Jan 29 16:04:48 radio mail is handled by 10 your-dns-needs-immediate-attention.radio. Jan 29 16:04:56 what? Jan 29 16:05:10 ceene: .lan, right? Jan 29 16:05:28 no search or domain entry on /etc/resolv.conf Jan 29 16:05:43 :> Jan 29 16:05:44 sslayer@portatil:~/Proyectos/neo900$ host test Jan 29 16:05:44 Host test not found: 3(NXDOMAIN) Jan 29 16:05:49 https://www.icann.org/en/.../name-collision-mitigation-final-redline-28oct15-en.pdf Jan 29 16:05:52 what is your resolver? Jan 29 16:05:57 sslayer@portatil:~/Proyectos/neo900$ cat /etc/resolv.conf Jan 29 16:05:58 # Generated by NetworkManager Jan 29 16:05:58 nameserver 8.8.8.8 Jan 29 16:05:58 nameserver 8.8.4.4 Jan 29 16:06:22 then you shouldnt send names without . to public resolver Jan 29 16:06:31 well I guess .radio is part of the new gTLDs Jan 29 16:06:31 i guess, ie. forgot domain in that file Jan 29 16:06:35 or dhcp Jan 29 16:06:56 meaning radio now points to ... the tld itself Jan 29 16:07:10 yeah, i forgot i don't have defined any domain search on resolv.conf Jan 29 16:07:15 bencoh: that makes sense Jan 29 16:07:15 http://serverfault.com/questions/626612/dns-just-started-resolving-my-server-prod-addresses-to-127-0-53-53 Jan 29 16:08:12 it makes sense, a little at least Jan 29 16:08:29 thanks Jan 29 16:08:46 i don't like this everything-can-be-a-tld thing Jan 29 16:08:58 i can end up masquerading existing dns hosts, can't i? Jan 29 16:09:11 what if had named my internal lan as .radio? Jan 29 16:09:25 dont ask me, its money talking Jan 29 16:09:26 now i wouldn't be able to see mtv.radio Jan 29 16:10:42 then run local lan resolver instead of asking google every time? Jan 29 16:11:38 that wouldn't fix that i can be masquerading a suddenly new tld Jan 29 16:29:11 ceene: yup, that's silly Jan 29 16:29:21 worst example is all those organizations that used .lan Jan 29 16:29:52 err, .local Jan 29 16:30:01 (instead of .lan) Jan 29 16:31:00 yep, anyhow, it's the same thing Jan 29 16:31:03 all screwed up :) Jan 29 17:32:51 pointing out far back, but i think it wasn't mentioned that mplayer still plays even if i have a phone call Jan 29 17:33:11 i don't hear it but file advances Jan 29 17:34:39 also panucci stops playback when unlocking, but i think it pauses when phonecall Jan 29 17:35:04 but both mplayer and panucci play in headphones in silent mode Jan 29 17:37:32 well, could be related then Jan 29 17:39:48 #ifdef CONFIG_MAEMO Jan 29 17:39:48 ossocontext = osso_initialize("mplayer-cli", "1.0 SVN", 0, NULL); Jan 29 17:39:48 playback = pb_playback_new_2 ((DBusConnection *) osso_get_dbus_connection (ossocontext), PB_CLASS_MEDIA, PB_FLAG_AUDIO | PB_FLAG_VIDEO | PB_FLAG_AUDIO_RECORDING | PB_FLAG_VIDEO_RECORDING, PB_STATE_PLAY, libplayback_state_request_handler, NULL); Jan 29 17:39:52 #endif Jan 29 17:39:52 in mplayer.c Jan 29 17:39:56 so ... indeed Jan 29 17:58:55 ROTFL http://paste.opensuse.org/497230 Jan 29 18:03:18 maemo audio handling is totally FUBAR. IIRC it decides on last process started which audio profile to 'enforce' Jan 29 18:04:14 I found a file (years ago) that lists all games (process names) that should use audio policy 'gaming' (or somesuch) Jan 29 18:04:56 DocScrutinizer05: you should probably backlog the whole discussion before going on :) Jan 29 18:08:02 could you give me a timestamp where it started? Jan 29 18:12:55 can't confirm >>11099: Audio interruptions when locking screen or receiving notifications<< Jan 29 18:13:36 actually mediaplayer mutes system sounds (a good thing) Jan 29 18:14:32 DocScrutinizer05: it happens with non-mafw players Jan 29 18:15:19 then those are "flawed", IOW they don't take care about the fubar audio handling on maemo Jan 29 18:16:38 that was our conclusion as well, yeah. actually we were pretty sure of that, just wanted to understand what should be done to "fix" those players ;) Jan 29 18:18:29 in a sane (ALSA) system environment, you'd define a few standard audio devices like "systemsound", "alarm", "media", "music", "gaming" and most apps would simply open the appropriate audio device. Apps with special needs would define their own audio devices by providing the ALSA stack, or they would construct the stack by complex audio device 'name' (incl parameters) they open Jan 29 18:19:40 *every* proper ALSA audio app has a way to configure the audio device to open. So no app would need "#ifdef CONFIG_MAEMO" crap Jan 29 18:21:13 xes: news? Jan 29 18:26:05 freemangordon: hey there, im progressing and writing down a log, fell asleep Jan 29 18:27:19 see for example http://alsa.opensrc.org/How_to_use_softvol_to_control_the_master_volume and implementation like in https://groups.yahoo.com/neo/groups/twinklephone/conversations/topics/1731 Jan 29 18:31:37 is there a distro that's not PA-based nowadays? Jan 29 18:33:30 hmm likely Jan 29 18:33:37 good question. Prolly not, but afaik you still can uninstall PA and get plain working ALSA Jan 29 18:34:07 android doesnt use PA for example Jan 29 18:34:23 is not PA-based* Jan 29 18:35:19 android is not linux (despite they tell me it was, every day) Jan 29 18:35:29 hehe @ Android Jan 29 18:37:11 re ALSA "right audio device" I even did a tiny fancy and created a ALSA audio device that executes arbitrary binaries on open() and on close() Jan 29 18:37:12 i don't consider Android when I think of LInux too .. anyway, Juesto what does Android use for audio? Jan 29 18:37:27 plain alsa afaik Jan 29 18:37:34 oh Jan 29 18:39:20 http://pastebin.com/f7015c561 Jan 29 18:39:36 i always apt-get --purge remove pulseaudio Jan 29 18:39:42 sometimes it's the *only* way to get audio to work Jan 29 18:40:40 :p Jan 29 18:40:51 i played with jackd at some point .. i liked it Jan 29 18:41:22 i read about low latency and thought: this is more than what i need Jan 29 18:43:06 http://maemo.cloud-7.de/ALSAso.tar.gz Jan 29 18:43:39 i should make time for alsa at some point Jan 29 18:43:50 my experience with alsa is pretty limited too Jan 29 18:44:07 configuring asoundrc a little bit :) Jan 29 18:44:13 never really used it as pa was default by the time i started using linux Jan 29 18:44:33 i think i created a virtual device with its own volume control so i could turn up the volume of notifications or something like that, years ago Jan 29 18:44:35 meh Jan 29 18:44:40 i feel overwhelmed here :( Jan 29 18:44:46 and for me audio (pa) always worked well Jan 29 18:48:07 not sure if just alsa or esound or stagefreight Jan 29 18:48:20 ^ about android audio, actually Jan 29 18:50:37 ceene: have a look at https://groups.yahoo.com/neo/groups/twinklephone/conversations/topics/1731 Jan 29 18:50:58 if we used systemd on N900, could we expect bad performance? :) Jan 29 18:51:37 Sicelo: you could expect bootloop since systemd needs cgroups but maemo occupies those Jan 29 18:52:18 systemd is absolutely incompatible to maemo Jan 29 18:52:57 maemo using cgroups to ensure priority of e.g. phone apps Jan 29 18:53:04 freemangordon: cant install libhildonmime because of missing l10n packages Jan 29 18:55:01 should i log out from garage before attempting to login in maemo.org? Jan 29 18:55:07 one of the total fuckups of systemd that it exclusively occupies cgroups and doesn't allow existing solutions needing cgroups to coexist Jan 29 18:55:39 or rather, systemd can't work without cgroups Jan 29 18:56:14 xes, for some reason logging out from garage logged me in maemo.org Jan 29 18:56:17 now it works fine Jan 29 18:56:27 zhere are systems that *don't HAVE* cgroup support at all Jan 29 18:56:28 strange Jan 29 19:00:12 i get you Jan 29 19:18:31 re http://maemo.cloud-7.de/ALSAso.tar.gz you could even do stuff like triggering vibra (via dbus-send) on any arbitrary app opening the ringer audio device. Or muting all system sounds incl ringtones as soon as any arbitrary media player opens "music" or "media" audio device. NO PATCHES NEEDED at all in any of the apps Jan 29 19:19:24 BUT.... Nokia couldn't have kept that a closed blob and thus ensure "User doesn't disrupt key functionality of device" Jan 29 19:19:59 Hmmm Jan 29 19:22:46 really the ONLY thing ALSA needs to learn to be on par with PA is rerouting of audio streams on hotplug of a new audio card (think pairing a BT headset). There's no support in ALSA for doing that without prior configuration knowing about the new audiocard properties Jan 29 19:24:15 you *could* handle that with existing ALSA when you accept you have to close() and re-open() the audio device of any app that's using the old and new audio card Jan 29 19:26:59 tbh I think that's the favorable concept over sinply brute-force switching to a new audio card as soon as it becomes available to the device, without even taking care if all the audio streams are really meant to get played back via that new device - but that's what PA in maemo does now, more or less Jan 29 19:27:23 Can't someone write a plugin to do that? Or would it be considered not very ALSA-like since it would probably involve multiple configurations? Jan 29 19:28:16 ie, pcm { type "connect"; connect.socket = "/run/alsamaster"; }, or something like that. Jan 29 19:28:21 s/= // Jan 29 19:28:22 Maxdamantus meant: ie, pcm { type "connect"; connect.socket "/run/alsamaster"; }, or something like that. Jan 29 19:28:37 I wouldn't want my smart mediaplayer with carefully crafted different equalizer settings play back to wired home stereo using the IHF-speakers EQ profile, just because I plug in the AV-cable to N900 Jan 29 19:30:08 So have a differennt socket for your mediaplayer. Jan 29 19:30:42 the thing listening on that will know what equalisation to use based on the current hw output. Jan 29 19:31:24 sidenote on nice property of http://maemo.cloud-7.de/ALSAso.tar.gz : when the binary you execute as open() (or close() ) hook returns error, your asound_open() in app will throw error instead of opening the audio device Jan 29 19:32:26 and obviously (as a consequence from the above) the asound_open() is sync, I.E. it only returns _after_ the hook binary returned Jan 29 19:37:41 so whatever audio config you need to set up, you can be sure your binary (or script) you call in "" hooks.0 { type ringerdevicehook hook_args { open "date;sleep 5;date" "" will have been finished successfully before the app doing the audio device open() continues and outputs *any* audio Jan 29 19:40:04 you can set up a audio scenario handling of the complexity and basic structure of sth like sysv-init with such hooks, while keeping same flexibility and transparency/simplicity Jan 29 19:41:44 note that the hooks run in the process context of the app calling them Jan 29 19:42:01 sicelo: i'm using slackware, no systemd, no pulseaudio Jan 29 19:42:35 hehe Jan 29 19:42:49 most likely gentoo doesnt use pa either Jan 29 19:43:23 dbus/pa/systemd/gnome are all belonging to the same 'users are stupid and need windowsish like system to take care for them' Jan 29 19:43:40 and xfce? Jan 29 19:44:13 yeah gentoo doesnt use pa by default but it can use it Jan 29 19:44:16 xfce is not bad Jan 29 19:44:38 interesting since xfce is gtk based Jan 29 19:44:41 like a gnome clone Jan 29 19:44:52 any linux system could use pa, its the 'use by default' option that is bad Jan 29 19:45:01 right Jan 29 19:45:05 gnome is much more than gtk ui Jan 29 19:45:13 Well you can always defeat the "use by default" Jan 29 19:45:43 juesto: not if software starts requiring some components Jan 29 19:45:51 ahh Jan 29 19:46:01 metapackages or software itself? Jan 29 19:46:11 as in, compiled with the requirement Jan 29 19:46:13 software Jan 29 19:46:18 ahhh right Jan 29 19:47:01 KotCzarny: just added cat /usr/share/policy/etc/syspart.conf.d/cmus.syspart.conf Jan 29 19:47:01 [rule /usr/bin/cmus] Jan 29 19:47:01 group mediasrc Jan 29 19:47:02 linux was about putting little, perfectly fine-tuned, building blocks into something bigger Jan 29 19:47:10 KotCzarny: works :) Jan 29 19:47:19 bencoh: yeah, but fmg wouldnt approve ;) Jan 29 19:47:24 freemangordon: no need to patch ohmd :) Jan 29 19:47:29 KotCzarny: reread that Jan 29 19:47:34 oh, right Jan 29 19:47:47 syspart.conf.d :) Jan 29 19:47:47 did you add anything else to include that dir? Jan 29 19:47:52 nope Jan 29 19:47:57 purrfect Jan 29 19:48:08 fixes the stuttering / interruption when pressing power key Jan 29 19:48:12 yep! Jan 29 19:48:16 after 10 years of ignoring that bug, fix is simple! Jan 29 19:48:31 lel Jan 29 19:48:50 there are two possible approaches (both equally valid and up to the user or developer to decide which to use): either you split/clone e.g. "music" device audiostream into multiple sinks aka cards (IHF-speakers, wired-HS, BT-headset), each of which has its own 'mute button' and you enable the sink you want being used for output of "music". Or your app uses a dedicated "music:IHF" audio device and thus explicitly decides itself which audio Jan 29 19:48:52 card to use Jan 29 19:49:14 bencoh: is there similar fix for pa's xpolicy.conf ? Jan 29 19:49:25 KotCzarny: not that I know of Jan 29 19:49:28 sadly Jan 29 19:52:40 HAH! Jan 29 19:52:53 Sicelo: ((systemd, cgroups)) less /usr/share/policy/etc/current/syspart.conf Jan 29 19:53:27 now I'd love to find a better fix than adding a dbus request in every player I might want to use to use vol keys when device is locked Jan 29 19:53:59 because cmus (for instance) doesn't even use glib, so adding dbus there feels ... silly. Jan 29 19:54:05 and /usr/share/policy/etc/current/pulse/xpolicy.conf is the damn audio file I mentioned I found years ago Jan 29 19:54:22 DocScrutinizer05: yeah we now about it :) Jan 29 19:54:22 yup, i feel the same about libplay or whatever its called Jan 29 19:54:46 KotCzarny: libplayback wont fix the volkeys issue Jan 29 19:54:52 at least I doubt it Jan 29 19:55:06 well I can try with mplayer since it's supposed to do that already Jan 29 19:56:14 yup, doesn't work with my current mplayer version a least. and mplayer does stop playing when pressing powerkey Jan 29 19:58:33 honestly W*T*F?! Jan 29 19:58:35 [stream] Jan 29 19:58:37 property = application.process.arg0@equals:"/usr/bin/Mahjong" Jan 29 19:58:38 group = game Jan 29 20:03:17 KotCzarny: hmm, something's wrong with ohmd and this "fix" Jan 29 20:03:25 wouldn't it be much smarter to have a str audiodev = "all:\"game,Mahjong\""; asound_open(audiodev) ? Jan 29 20:03:33 looks like it doesnt work after some time (?!) Jan 29 20:03:52 I have no freaking idea what they did there Jan 29 20:03:55 seriously Jan 29 20:04:46 ohmd is sort of PITA and maemo has a closed blob extension to it afaik Jan 29 20:06:40 where i can find l10n blobs for fremantle-gtk2 ? Jan 29 20:08:06 note you could even do sth like .alsarc: default { pcm @getenv ALSAPCM... and then `ALSAPCM=games Mahjong` Jan 29 20:09:59 Juesto: nokia repositories (and mirrors) I guess Jan 29 20:10:49 bencoh: may be related to that pa policy changes Jan 29 20:11:16 s/alsarc/asoundrc/ Jan 29 20:11:16 DocScrutinizer05 meant: note you could even do sth like .asoundrc: default { pcm @getenv ALSAPCM... and then `ALSAPCM=games Mahjong` Jan 29 20:11:23 bencoh: should i add those to apt ? Jan 29 20:11:30 so when system changes policy maybe it need to reload ohmd policy too Jan 29 20:13:13 for somee reason i cant logout from maemo.org now Jan 29 20:13:15 does ohmd support some debug logging? Jan 29 20:13:23 juesto: delete cookies? Jan 29 20:14:01 KotCzarny: cookies deletion shouldnt be necessary Jan 29 20:14:20 yes, but should log you out Jan 29 20:14:36 i know. :| Jan 29 20:15:06 bencoh: ((looks like it doesnt work after some time; I have no freaking idea what they did there)) I *guess* that policy stuff doesn't always work the way we hope it would. I don't see how switching policies can get synced to e.g. task switching, and prolly also not easy to have nested policies when starting a later process overrides a earlier policy and then on end of same process the earlier policy possibly doesn't get restored Jan 29 20:32:20 Juesto: you'd better not add armel repos to your i386 devuan Jan 29 20:32:50 Juesto: you can find all the i10n packages in nokia repo mirrors Jan 29 20:32:56 just google for the package name Jan 29 20:33:28 bencoh: su syspart.conf.d is already implemented? Jan 29 20:33:31 *so Jan 29 20:36:17 DocScrutinizer05: iiuc, they tag PA packets somehow, so when they arrive in PA, they get prioritized or somesuch Jan 29 20:37:15 DocScrutinizer05: also, ohm/ohmd is FOSS Jan 29 20:37:42 yes, but Nokia built a blob extension for it afaik Jan 29 20:38:10 oh, that prolog shit? Jan 29 20:38:17 prolly, yes Jan 29 20:38:37 yep, right, but Pali REed it :) Jan 29 20:38:44 :-)) Jan 29 20:39:08 now write a doc for it and it's almost useful ;-) Jan 29 20:39:41 I really don't even grok the concept Jan 29 20:39:48 me neither Jan 29 20:40:01 yea, prolog code was decompiled :-) Jan 29 20:45:52 we should try to contact javispedro and ask him, he seemed to have some degree of understanding in that stuff Jan 29 20:46:18 ~seen javispedro Jan 29 20:46:19 javispedro <~javier@Maemo/community/contributor/javispedro> was last seen on IRC in channel #maemo, 798d 3h 45m 34s ago, saying: 'they don't, but you will have poor signal in stock n900'. Jan 29 20:46:54 iirc he did a rage quit, so I doubt he'll be cooperative Jan 29 20:46:59 o.O Jan 29 20:47:01 why? Jan 29 20:47:09 hey wb freemangordon Jan 29 20:47:28 safe link? Jan 29 20:47:30 can't recall any ragequit by javispedro Jan 29 20:48:48 rage targeted at Nokia? Jan 29 20:48:53 is being a contributor strictly a programming contribution? Jan 29 20:49:01 no Jan 29 20:49:04 no Jan 29 20:49:17 okay, perhaps i can have it once i get this working Jan 29 20:49:42 maemo cloaks managing is a royal pita Jan 29 20:49:51 how so Jan 29 20:49:57 GC MIA Jan 29 20:50:10 bencoh: that syspart.conf fix is incompatible with xpolicy.conf fix Jan 29 20:50:11 ?? Jan 29 20:50:28 group contact missing in action Jan 29 20:50:29 removing oscp from stream group in xpolicy made it unglitchable again Jan 29 20:50:44 ~seen generalantilles Jan 29 20:50:45 generalantilles <~ryan@Maemo/community/contributor/GeneralAntilles> was last seen on IRC in channel #asterisk, 841d 4h 21m 27s ago, saying: 'OK, thanks. :)'. Jan 29 20:52:55 hrm, even stranger, it stops working after i stop oscp, but restarting pulseaudio and ohmd make it work again Jan 29 20:53:01 basically Maemo/community/contributor/ is equivalent to Maemo/EX-council/ Jan 29 20:54:06 very strange. Jan 29 20:54:26 o_o Jan 29 20:54:53 bencoh, freemangordon: any clues on ^ ? Jan 29 20:55:48 KotCzarny: no idea, sorry Jan 29 20:56:15 i suspect some policies changing Jan 29 20:56:42 KotCzarny: I guess you should look at the code that parses those config files Jan 29 20:58:43 i suspect pa playing some role in all this Jan 29 20:59:56 KotCzarny: restart ohmd alone makes it work again, afaict Jan 29 21:00:04 yeah, just tried it Jan 29 21:00:06 and it stops working after some time Jan 29 21:00:12 like, a few minutes Jan 29 21:01:02 /who Maemo/community/* Jan 29 21:01:14 ? Jan 29 21:03:26 DocScrutinizer05: doesnt tell me many people because its freenode :| Jan 29 21:03:45 err? Jan 29 21:06:11 Juesto: when I was councilor, I two times (iirc) pushed a initiative to hand out Maemo/community/contributor/ cloaks to renowned members who applied for it. It bith times took about 5 months until all council agreed and GC implemented those few cloaks. I wouldn't suggest you hold your breath for a 3rd such initiative any time soon Jan 29 21:06:22 bencoh, maybe its related to some cpu usage quota? Jan 29 21:06:42 o_O Jan 29 21:06:44 holy Jan 29 21:07:12 when i enable cpu hogging effects i get glitches more easily, no matter if i restart ohmd Jan 29 21:07:56 (mind you, its not glitching during play, only when doing new app open/notification show) Jan 29 21:09:02 welp Jan 29 21:11:10 Juesto: anyway if you really really want to get such cloak, the recommended way is: find a recent council member who's actually joining this IRC channel, ping them about a cloak, send a email to council AT maemo.org applying for one, and then hope for the best Jan 29 21:11:59 okay nevermind then Jan 29 21:12:22 IF I could get a hold on Ryan Abel (our GC MIA), I could ask him to grant co-GC to me, so we could cut a few corners there, but last mail I sent been unreplied Jan 29 21:12:23 but thanks for the info DocScrutinizer05 Jan 29 21:12:45 GC MIA? Jan 29 21:13:05 [2017-01-29 Sun 21:50:27] group contact missing in action Jan 29 21:13:09 Ah Jan 29 21:13:15 sorry Jan 29 21:13:44 the contact person between owners of namespace and freenode admins who can actually assign cloaks Jan 29 21:14:58 I'm not sure if council changed the GC for #maemo*, last thing I know is: it been Ryan Abel aka generalantilles Jan 29 21:15:36 ~seen generalantilles Jan 29 21:15:37 generalantilles <~ryan@Maemo/community/contributor/GeneralAntilles> was last seen on IRC in channel #asterisk, 841d 4h 46m 19s ago, saying: 'OK, thanks. :)'. Jan 29 21:18:15 KotCzarny: doubt it, but dunno Jan 29 21:18:40 I even tried increasing min freq from 250m to 600m, didn't help Jan 29 21:40:57 DocScrutinizer05: probably it did change since... ryan abel isnt listed in community council page Jan 29 21:42:42 he was in council in 2009 tho Jan 29 21:43:03 inagural council* Jan 29 21:54:04 Juesto: sorry? GC is not mandatory member of council Jan 29 21:54:42 council usually doesn't have any administrative roles, those are delegated Jan 29 21:55:32 council decides who's administrative person for all the different tasks maemo involves Jan 29 21:56:21 though only on recent demand by situation, or by request from community Jan 29 21:57:23 actually usually council only approves suggestions made by e.g. techstaff Jan 29 21:58:10 or, in case of cloaks, by chanops and GC Jan 29 21:59:23 council isn't even expected to have the expertise to do all that administrative tasks by themselves Jan 29 22:17:30 ahhhh Jan 30 00:18:43 * Juesto for council 2040 ! Jan 30 00:35:43 anyone around? Jan 30 00:35:45 meh Jan 30 00:36:30 *tumbleweed passes* Jan 30 00:38:07 Should i gather all the noarch maemo packages from n900 repo? Jan 30 01:00:56 Is replacing N9 'sim card reader' more difficult than 'LG / Samsumg youtube videos' ? Jan 30 01:01:37 *more difficult than what videos make it out to be? Jan 30 02:51:59 hmm, "oh ey! eeew!" is a valid replacement for "OK Google" Jan 30 02:52:38 LOL Jan 30 02:52:40 :DD Jan 30 02:52:43 * DocScrutinizer05 starts to understand why arbitrary TV sound starts that google search shit Jan 30 02:52:52 ;) Jan 30 02:53:03 funny google eh Jan 30 02:55:10 Michael_a380: define 'sim card reader' Jan 30 02:55:25 welp, my computer began having stuck processes Jan 30 02:55:34 6 days uptime Jan 30 02:55:41 my netbok is still alive Jan 30 02:55:44 stuck processes? Jan 30 02:55:48 yeah Jan 30 02:55:55 what's "stuck"? Jan 30 02:55:55 kill is useless Jan 30 02:56:19 DocScrutinizer05: please use Wiktionary Jan 30 02:56:25 kill -9 should work almost always Jan 30 02:56:31 windows also Jan 30 02:56:40 and theres no kill -9... iirc Jan 30 02:56:47 errr Jan 30 02:57:24 no idea, except "ask sysinternals" Jan 30 02:57:49 Well, no dice with pskill Jan 30 02:58:00 the kill action succeeds but the process doesnt die Jan 30 02:58:14 and for worse, i have one svchost.exe stuck Jan 30 02:58:16 Those naughty metal pins that break. https://youtu.be/a9DG_9OiRdY Jan 30 02:58:43 ost dns client, NLA and workstation Jan 30 02:58:48 aka samba client Jan 30 02:58:58 lost* Jan 30 02:59:20 funny, i can still kills the process handles Jan 30 02:59:23 but not the threads Jan 30 02:59:29 its just stuck Jan 30 02:59:43 trying to attach a debugger fails **** ENDING LOGGING AT Mon Jan 30 03:00:01 2017