Wizzup | ^Tcache: welcome :) | 00:38 |
---|---|---|
buZz | freemangordon: measured with http://space.nurdspace.nl/~buzz/maemo/batmon.py | 00:57 |
^Tcache | <Wizzup> "𝔗𝔠𝔞𝔠𝔥𝔢: welcome :)..." <- Thanks Wizzup | 03:15 |
norayr | hello, folks. | 03:33 |
norayr | i was thinking to build 'tablet mode' applet, but i am afraid it uses some nokia dbus stuff which doesn't work anymore. | 03:33 |
norayr | can someone take a look? we need something like that - to disable cellular but still be able to use wifi. | 03:33 |
buZz | freemangordon: ah, after ~10 minutes of screenoff, it drops to 70-90mA | 04:22 |
freemangordon | buZz: that's still too high, please confirm TS is inhibited | 08:28 |
freemangordon | buZz: cd /sys/class/power_supply/battery/ | 09:28 |
freemangordon | while [ 1 ]; do cat power_avg; sleep 60; done | 09:28 |
freemangordon | Wizzup: what the?!? https://github.com/maemo-leste/dsme/blob/master/Makefile#L64 | 09:48 |
freemangordon | also, why do we have this https://github.com/maemo-leste/dsme/blob/master/debian/control#L14 ? | 09:49 |
Wizzup | freemangordon: I think that's work from spinal that we merged | 09:49 |
freemangordon | yes, but you changes systemd to logind | 09:49 |
freemangordon | neither ow which are used on devuen | 09:49 |
freemangordon | *of | 09:49 |
Wizzup | hm? | 09:50 |
freemangordon | do you mind if I enable systemd/elogind support almost unconditionally (based on pkgconfig package presence)? Instead of checking the build machine OS current Makefile does | 09:50 |
freemangordon | see the check in the Makefile | 09:51 |
Wizzup | makes sense | 09:51 |
freemangordon | "ifeq (,$(wildcard /etc/devuan_version))" | 09:51 |
freemangordon | this is .... | 09:51 |
Wizzup | a hack ;) | 09:51 |
freemangordon | weird to say at least :) | 09:51 |
freemangordon | so, will enable systemd/elogind support based on pkgconfig packages presence | 09:52 |
arno11 | sicelo: Wizzup: guys...n900 calls are working both ways !!! | 12:34 |
arno11 | with pulseaudio and cmt_pulse | 12:34 |
Wizzup | arno11: amazing | 12:34 |
Wizzup | info pls :D | 12:35 |
arno11 | well, sample rate was the issue: disabling resampling and forcing pulseaudio to accept 4000 hz rate authorize cmtspeech to record and send data. | 12:38 |
arno11 | as simple as that | 12:39 |
Wizzup | super... | 12:39 |
arno11 | the solution was in source code | 12:39 |
arno11 | yes really cool | 12:39 |
Wizzup | so you changed pa src? | 12:40 |
arno11 | i just changed daemon.conf resampling option and alternate default samplerate | 12:41 |
Wizzup | ok, care to share the changes? :D | 12:41 |
arno11 | yes of course. | 12:41 |
arno11 | i will share stuff this evening | 12:44 |
Wizzup | great | 12:45 |
arno11 | i just need time to test few other things in alsamixer and sphone | 12:45 |
arno11 | but it works | 12:45 |
arno11 | able to take a call from my wife | 12:46 |
Wizzup | this is really great news man :) | 12:46 |
arno11 | :) | 12:46 |
Wizzup | probably warrants a new post together with the chimaera releas when we integrate it | 12:47 |
arno11 | :) cool | 12:48 |
freemangordon | arno11: great work man! | 12:55 |
freemangordon | I guess we'll be able to divert .conf through leste-conig | 12:55 |
freemangordon | in the meanwhile I taught dsme on elogind sessions :) | 12:56 |
arno11 | freemangordon: thx :) and great work from you too as usual | 12:58 |
freemangordon | heh | 12:59 |
sicelo | arno11: you're THE man | 12:59 |
sicelo | how's quality? | 12:59 |
arno11 | :) quality is acceptable but few cracking noise | 13:00 |
arno11 | normal at 4000hz | 13:00 |
arno11 | i think its possible to use 8000hz | 13:00 |
arno11 | i will try | 13:01 |
arno11 | changing few things in mixers | 13:01 |
sicelo | this news will be very welcome on #maemo and tmo | 13:01 |
sicelo | and might generate a bit more interest in Leste | 13:02 |
arno11 | yes probably | 13:02 |
Wizzup | and latency? | 13:04 |
arno11 | just let me ask my wife lol | 13:05 |
arno11 | according to her (and me) latency seems not bad | 13:08 |
arno11 | anyway that's easy to deal using alsamixer | 13:08 |
arno11 | i think we can improve that a bit | 13:09 |
arno11 | let me try pa_test | 13:12 |
arno11 | back in a bit | 13:12 |
arno11 | one second latency using pa_test | 13:14 |
arno11 | with no change in mixers | 13:15 |
arno11 | i'll try playing with that to improve that | 13:16 |
arno11 | latency | 13:16 |
arno11 | sound is clear | 13:16 |
Wizzup | cool | 13:17 |
arno11 | i have to go. ttyl | 13:18 |
sicelo | calls been the main reason people state for disinterest in Leste. of course this may be a little late for some, with 2G gone (but that's the case with everything we support except pp) | 13:18 |
freemangordon | Wizzup: I'll push dsme changes to -devel without PR, if you or uvos or anybody has any concerns about them, LMK | 13:26 |
Wizzup | go for it | 13:27 |
Wizzup | sicelo: I don't think both 2g and 3g are gone in places? | 13:27 |
Wizzup | that would be truly insane | 13:27 |
freemangordon | for sure 2g is here to stay | 13:27 |
sicelo | someone was saying it is gone/going somewhere in USA. i didn't check if what they were saying was supprted by facts | 13:28 |
freemangordon | well, USA... :) | 13:29 |
freemangordon | that's another universe anyways | 13:29 |
sicelo | :-D | 13:31 |
sicelo | with voice calls working on N900, the issue of sphone preferring portrait needs to be tackled ... | 13:33 |
sicelo | or it wasn't an issue for arno11? | 13:33 |
Wizzup | yes, that is a problem still | 13:35 |
Wizzup | I imagine he maybe used ofono scripts to test | 13:35 |
freemangordon | Wizzup: can't we somehow divert reboot and shutdown commands? | 14:53 |
freemangordon | to point to dsme that is | 14:53 |
Wizzup | freemangordon: we could I think | 14:56 |
Wizzup | not sure if we should | 14:56 |
freemangordon | I think we shall do something about | 14:56 |
freemangordon | as now dsme restarts processes on reboot/shutdown from shell | 14:56 |
Wizzup | like what if dsme does not run | 14:56 |
Wizzup | during some debugging | 14:56 |
freemangordon | hmm | 14:57 |
freemangordon | iirc there are some hooks that can be made | 14:57 |
* freemangordon checks | 14:57 | |
arno11 | Wizzup: sicelo: indeed i used ofono script but surprisingly sphone works great (keyboard open. no weird crash or resolution) | 15:26 |
arno11 | sound is good. latency acceptable | 15:27 |
arno11 | latency is better in real condition than using pa_test | 15:27 |
freemangordon | Wizzup: will you have some spare time to revert the revert of xorg? | 15:27 |
Wizzup | I think I do right now actually | 15:28 |
freemangordon | thanks! | 15:28 |
Wizzup | arno11: yes keyboard open is a hack to make sphone not crash n900 driver atm | 15:28 |
freemangordon | I created policykit-1-hildon package which for now will contain only xsession script and dependency to gnome auth agent | 15:29 |
freemangordon | with those 2 we are done | 15:29 |
freemangordon | everything else is in place, unless I am missing something | 15:29 |
Wizzup | freemangordon: shall I also add the conflicts with xserver-xorg-legacy again then | 15:30 |
Wizzup | to hildon-base | 15:30 |
freemangordon | yes | 15:30 |
Wizzup | ok | 15:30 |
Wizzup | btw this is to -devel first | 15:31 |
Wizzup | then I will try it, and then move to non-devel | 15:32 |
Wizzup | freemangordon: sounds good @ policykit-1-hildon | 15:32 |
freemangordon | we shall move all @ once devel->stable | 15:33 |
freemangordon | well... | 15:33 |
freemangordon | we better do | 15:33 |
arno11 | Wizzup: freemangordon: are alsamixer default settings similar to fremantle or are they generic ? | 15:37 |
Wizzup | that's a complicated question | 15:37 |
arno11 | ah ok | 15:37 |
Wizzup | they're probably not like fremantle by default | 15:37 |
Wizzup | fremantle has very complicated audio mgmt | 15:37 |
Wizzup | and the n900 alsamixer settings if you ask me are also complicated... :D | 15:37 |
arno11 | ok because i use fremantle settings to get audio calls | 15:37 |
Wizzup | ah, I see, do you have a way to diff them? | 15:38 |
Wizzup | because I think our default alsamixer settings are in the n900 image | 15:38 |
freemangordon | arno11: sorry, I am not the one to ask about audio routing, Wizzup/uvos know more than me. Most probably you know more than me as well :) | 15:39 |
Wizzup | leste-config-n900 contains leste-config-n900/var/lib/alsa/asound.state | 15:39 |
Wizzup | arno11: but really what we need in the very least is UCM2 for n900 | 15:39 |
arno11 | Wizzup: yes definitly | 15:40 |
Wizzup | in general fremantle does a lot of funky stuff audio wise, which we currently definitely don't do, but we don't need it immediately either | 15:40 |
arno11 | freemangordon: ok | 15:40 |
arno11 | ok i just asked to be sure my changes in alsamixer had an inpact | 15:41 |
Wizzup | yeah, they should be saved and persistent across reboots maybe even | 15:44 |
Wizzup | depending on what PA does | 15:44 |
arno11 | yes | 15:44 |
arno11 | Wizzup: what is the best way to share voice call stuff ? | 15:52 |
Wizzup | email and patches I guess | 15:53 |
Wizzup | or github pull requests | 15:53 |
Wizzup | maybe just an email with description first | 15:54 |
arno11 | ok | 15:54 |
arno11 | Wizzup: btw there are NO error messages during calls: no buffer or float point errors | 16:07 |
Wizzup | cool | 16:08 |
arno11 | that's incredible how a simple thing (samplerate) could generate so much errors are troubles not directly related | 16:08 |
arno11 | anyway that's usual stuff with pulseaudio... | 16:11 |
arno11 | *or | 16:12 |
freemangordon | Wizzup: we shall ad dependency to policykit-1-hildon in one of the metapackages | 16:13 |
freemangordon | *add | 16:13 |
freemangordon | any idea which one? | 16:13 |
Wizzup | hildon-meta ? | 16:16 |
freemangordon | ok, please add when changing for xorg-legacy | 16:16 |
Wizzup | hildon-base and hildon-meta are not the same | 16:17 |
Wizzup | but I can add it to hildon-meta | 16:17 |
freemangordon | ok | 16:17 |
freemangordon | Wizzup: what is hildon-session package? | 16:19 |
freemangordon | do re need it still? | 16:19 |
freemangordon | *do we | 16:19 |
Wizzup | seems like not I think | 16:19 |
freemangordon | shell I remove it from repos_core.py ? | 16:20 |
freemangordon | while I am at it | 16:20 |
freemangordon | removed it | 16:21 |
Wizzup | yes | 16:22 |
Wizzup | please | 16:22 |
arno11 | Wizzup: i compared stock alsamixer settings | 16:27 |
arno11 | there are really different | 16:27 |
arno11 | i think we should use fremantle ones | 16:28 |
arno11 | but there are more than 200 parameters... | 16:28 |
arno11 | i think i'll do a copy of my custom asound.state (calls working. music and video as well) and send it by email ? | 16:31 |
arno11 | i did mine by hand using fremantle as a model | 16:31 |
arno11 | because i didn't find a clean way to load another asound profile | 16:34 |
freemangordon | Wizzup: did I get it right that you will add the dependency? | 16:37 |
Wizzup | arno11: keep in mind that fremantle settings will be different depending on what the phone is doing | 16:37 |
Wizzup | arno11: there is no 'one' good state | 16:37 |
Wizzup | freemangordon: sure | 16:37 |
freemangordon | ok | 16:37 |
Wizzup | freemangordon: did you build it for maemo/chimaera? | 16:38 |
freemangordon | maemo/chimaera-devel | 16:38 |
freemangordon | it needs dsme from -devel | 16:38 |
Wizzup | ok | 16:38 |
arno11 | Wizzup: i know. that's why i tried to capture alsamixer settings during call | 16:38 |
arno11 | and it works on leste | 16:39 |
Wizzup | right, well, keep the files around, they will help with making UCM2 | 16:40 |
Wizzup | UCM2 is just some weird syntax around setting amixer states mostly | 16:40 |
arno11 | yes ok | 16:40 |
freemangordon | Wizzup: how do we debian package our omap kernels? I am asking because it is missing postinst script and I want to fix it | 17:00 |
freemangordon | ah, got it | 17:01 |
arno11 | Wizzup: is it ok to send an email to maemo-leste@lists.dyne.org ? | 18:08 |
Wizzup | yes, please! :) | 18:12 |
arno11 | ok :) | 18:12 |
arno11 | email sent | 18:32 |
Wizzup | I moved it through the queue | 18:39 |
sicelo | pavel will be pleasantly surprised | 18:39 |
Wizzup | if you add yourself to the list I won't have to approve future msgs and you will get the replies automatically | 18:39 |
Wizzup | you can subscribe here: https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/maemo-leste | 18:40 |
Wizzup | (you don't have to, just fyi) | 18:40 |
sicelo | arno11: thanks for sending to list. helps to keep everyone in the loop and able to pick up the work at any point | 18:40 |
arno11 | Wizzup: ok thx | 18:43 |
arno11 | Sicelo: i'm happy to help FOSS community | 18:44 |
Wizzup | arno11: btw, tools like parec have a --latency-msec arg that you can set to say 16ms (--latency-msec=16), it might be worth checking what pa func this calls and if the pa cmspeech stuff uses this or not | 18:50 |
Wizzup | if it doesn't, it could easily explain the delay | 18:50 |
Wizzup | (if there even is a real delay) | 18:50 |
freemangordon | Wizzup: which machine builds omap kernel? | 18:51 |
freemangordon | arm32? | 18:51 |
arno11 | ok i'll have a try | 18:51 |
arno11 | but tbh latency is acceptable in real call conditions | 18:52 |
arno11 | for me the thing to improve is micro/HP/earphone levels | 18:53 |
arno11 | i forgot to mention there is NO echo with custom mixers settings | 18:55 |
arno11 | even with pa_test or arecord/aplay which is doing exactly the same thing | 18:56 |
Wizzup | arno11: weird, I would expect there to be echo since the modem doesn't do echo cancelling | 22:52 |
Wizzup | I guess it's more of a problem with the speaker | 22:52 |
Wizzup | freemangordon: let me check | 22:52 |
Wizzup | freemangordon: devuan-armhf - yes | 22:53 |
freemangordon | Wizzup: did we ever try to cross-build it? | 22:53 |
Wizzup | we could, but it will be slower | 22:54 |
freemangordon | as it fails here: | 22:54 |
freemangordon | dpkg-gencontrol: error: current host architecture 'amd64' does not appear in package 'linux-headers-omap' architecture list (armhf armel) | 22:54 |
Wizzup | probably with bencoh's lxc setup you can | 22:54 |
freemangordon | well... | 22:54 |
freemangordon | I'd better fix rules file to be able to build it in VM | 22:55 |
Wizzup | I am not sure, but it seems to me this problem might be more with your vm than our rules? | 22:55 |
Wizzup | in any case feel free of course to change it locally and see if it helps :) | 22:55 |
freemangordon | it is in the build scripts | 22:55 |
freemangordon | or rather in dpkg-gencontrol | 22:56 |
Wizzup | well, the error is right, though, n o? | 22:56 |
freemangordon | no, because I am doing DEB_BUILD_OPTIONS="parallel=5" dpkg-buildpackage -b --target-arch=armhf | 22:56 |
freemangordon | it should take *target* arch, not host ;) | 22:57 |
Wizzup | ok | 22:57 |
freemangordon | it takes DEB_HOST_ARCH instead of DEB_BUILD_ARCH | 22:58 |
freemangordon | I'll just hack that in rules | 22:58 |
freemangordon | why I am doing that - kernel headers package lack postinst that process dkms scripts | 23:00 |
freemangordon | that's why ibhb-dkms doe not getr rebuild on new kernel version | 23:00 |
freemangordon | *does not | 23:00 |
Wizzup | ah.. | 23:01 |
freemangordon | hmm, I guess I lack a parameter to dpkg-buildpackage | 23:06 |
freemangordon | Wizzup: my bad (DEB_HOST_ARCH) | 23:13 |
Wizzup | freemangordon: check | 23:20 |
freemangordon | Wizzup: https://pastebin.com/hEVB5Eyh ? | 23:52 |
freemangordon | going to push that and make a new release from -devel | 23:57 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!