Wizzup | ok :) | 00:12 |
---|---|---|
arno11 | thx | 00:14 |
uvos | how is alsa involved in tv out on n900 | 00:31 |
uvos | is it just the audio | 00:31 |
uvos | or is the jack routing of the video signal done via alsa? (ugh) | 00:31 |
sicelo | the tv out also provides audio to the tv | 00:32 |
Wizzup | uvos: it's not @ alsa | 00:51 |
Wizzup | >Enable in alsamixer (Jack Function in alsamixer to TV OUT), enable with xrandr --output TV --mode 800x480 --same-as LCD. Set PAL/NTSC with xrandr --output TV --set SignalProperties PAL, and scale with xrandr --output TV --set TVScale 90. | 00:52 |
Wizzup | maybe it is :D | 00:52 |
sicelo | :-P | 01:00 |
sicelo | even HDMI features in alsa, for the same obvious reason | 01:00 |
freemangordon | Wizzup: uvos: actually I don;t think we can use either gconf or gsettings for fmtx | 08:08 |
freemangordon | https://github.com/maemo-leste/fmtx-middleware/blob/master/files/etc/init.d/fmtx#L3 | 08:09 |
freemangordon | we don;t want to do that ^^^ | 08:09 |
freemangordon | so, I am thinking of configfs maybe? | 08:09 |
freemangordon | umm... not configfs | 08:10 |
freemangordon | what was the name? | 08:10 |
freemangordon | ok, seems I got it wrong and there is no kernel fs that can be used as persistent parameters storage | 08:28 |
sicelo | besides frequency (and maybe region?), what else needs persistent storage? | 09:54 |
uvos__ | freemangordon: not sure why this should prevent gsettings, you can check the key like that with dconf too | 10:11 |
uvos__ | besides, this seams not so great, would it not be better if the deamon just checked the key on startup and exited sucessfully if its not set? | 10:12 |
uvos__ | am i missing something? | 10:12 |
freemangordon | uvos__: maybe you are right, I will dig further into it | 10:32 |
arno11 | Wizzup: sicelo: new UCM files sent | 11:30 |
Wizzup | ty, will deploy momentarily | 11:30 |
arno11 | ok thx | 11:31 |
Wizzup | (I'm excited for this atm) | 11:31 |
Wizzup | s/atm/btw/ | 11:31 |
Wizzup | also want to try the IR a bit later this week :) | 11:31 |
arno11 | ok cool | 11:31 |
Wizzup | arno11_: building for -devel | 12:57 |
arno11_ | cool | 12:59 |
Wizzup | it's probably already built | 13:05 |
arno11_ | ok thx | 13:06 |
arno11_ | yes i see it | 13:07 |
sicelo | cool, thanks. quick question though ... why create relationship between fmtx and headphone jack? | 13:40 |
sicelo | or that was to get autoswitching, since pulse already triggers the ucm verb/device according to jack events? | 13:41 |
arno11_ | fmtx only works if jack function is set on 'headphone' iirc | 14:11 |
sicelo | yes, that's true (the fm transmitter chip inputs connected to the headphone lines). | 14:15 |
sicelo | but my question is - you mention that one should connect the headphone jack. it should work without inserting it normally | 14:16 |
arno11_ | yes of course, just in case you insert something, fmtx must be started after | 14:18 |
arno11_ | otherwise ucm switch to headphone sequence | 14:18 |
arno11_ | and stop fmtx | 14:19 |
arno11_ | *switches *stops | 14:26 |
arno11_ | brb | 14:26 |
sicelo | i see. i guess that's why something like fmtxd is needed ... so the fmtx session is 'protected' from that kind of thing | 14:30 |
Wizzup | uvos: the thumb2 support, is that for building kernel to make it use thumb2? | 15:30 |
uvos | yes | 15:33 |
Wizzup | cool | 15:41 |
sicelo | link to the branch? i don't see it for now | 15:55 |
Wizzup | maemo/chimaera-devel | 15:57 |
Wizzup | bbl | 15:57 |
sicelo | yes, but i mean the actual kernel config changes ... maemo/chimaera-devel only contains packaging | 15:58 |
Wizzup | ah, ok, I just looked at the changelog | 15:59 |
sicelo | otherwise thanks uvos for thumb2, and i see charge-mode changes related to the booting at <10%. thanks | 16:00 |
uvos | sicelo: also the charging not being detected on boot on d4 might be fixed, at least i cant repo it right now anymore | 16:05 |
uvos | Wizzup: do we have bash installed on leste for sure? | 16:18 |
uvos | why is /bin/sh dash? | 16:18 |
sicelo | isn't that the debian/devuan default? | 16:25 |
bencoh | it is | 16:26 |
uvos | anyhow Wizzup please merge and tag this out: https://github.com/maemo-leste/osso-af-startup/pull/2 | 16:27 |
uvos | also the use of runuser here is ugh | 16:27 |
sicelo | mmm, i'm not sure 'source' is a bashism. at least it's there in pmos and used extensively. they don't use bash, but busybox. can check what their /bin/sh is aliased to | 16:34 |
uvos | source istent part of the posix shell | 16:34 |
uvos | and i think it was first introduced by bash | 16:34 |
uvos | but thats not really important | 16:35 |
uvos | point is its not in in sh and not in dash | 16:35 |
sicelo | their /bin/sh is symlink to /bin/busybox | 16:35 |
arno11 | i know that's an old issue guys but any hope to solve the weird bug with hildon composition? disabling it on the fly speeds up apps a lot (even on pinephone iirc) | 18:11 |
uvos | what wierd bug with composition? | 18:11 |
bencoh | slower rendering, I'd say | 18:12 |
arno11 | yes | 18:12 |
uvos | composition is expensive - sure maybe it can be optimized but its not a bug | 18:12 |
bencoh | (although I don't think it's specific to hildon) | 18:12 |
uvos | but composition is just expensive | 18:13 |
arno11 | ok that's really too much for N900 | 18:13 |
bencoh | I wonder if you could easily try another compositor / compositing wm for comparison | 18:13 |
uvos | kwin should work | 18:14 |
uvos | it has gles support | 18:14 |
uvos | its reaaaaaaly slow on d4 | 18:14 |
uvos | so yeah | 18:14 |
bencoh | slower than hildon? | 18:14 |
uvos | much | 18:14 |
uvos | the only fix here would be to make hildon not rely on composition, as its not nessecary for what hildon dose really | 18:15 |
uvos | but this would require major reachitecting | 18:15 |
uvos | so not going to happen | 18:15 |
freemangordon | bencoh: *everything* is slower than hildon | 18:15 |
uvos | well no plain x isent :P | 18:15 |
uvos | or openbox | 18:15 |
bencoh | freemangordon: :D | 18:15 |
uvos | or i3 | 18:15 |
bencoh | yeah, wmii is just super fast on d4 ;) | 18:16 |
uvos | sway is also faster than hildon | 18:16 |
uvos | at composition | 18:16 |
uvos | even | 18:16 |
freemangordon | uvos: sorry, I missed the context - why is bash needed in the above PR? | 18:17 |
uvos | freemangordon: so runuser runs as the users shell by default | 18:17 |
uvos | we have a user who uses xonsh (essentally python) as thair shell, so the scripts dont work with that | 18:17 |
freemangordon | omg | 18:17 |
uvos | runuser sources the profile with source | 18:17 |
uvos | wich is not in posix and not implemented by dash (default debain shell) | 18:18 |
uvos | which is not in posix and not implemented by dash (default debain shell) | 18:18 |
uvos | so we need to use bash explicitly | 18:18 |
freemangordon | I see | 18:19 |
freemangordon | ok, lemme merge that | 18:21 |
freemangordon | uvos: shall I release it for -devel? | 18:26 |
uvos | yes, please | 18:26 |
freemangordon | ok | 18:27 |
uvos | also merge master | 18:27 |
uvos | its behind devel | 18:27 |
freemangordon | sure | 18:31 |
freemangordon | done | 18:39 |
arno11 | freemangordon: any idea why ir_rx51 is causing kernel crashes ? | 19:36 |
arno11 | *when we try to send pulse | 19:37 |
maxwelld | i am getting floating point exception when running osso-xterm | 20:56 |
maxwelld | i think it happened before as well. | 20:56 |
maxwelld | don't remember how i solved it. | 20:57 |
uvos | backtrace? | 22:25 |
freemangordon | arno11: where is oops? | 22:27 |
arno11 | oops happens immediately after sending a pulse, no way to log | 22:29 |
freemangordon | why? | 22:29 |
freemangordon | wtym "no way to log"? | 22:29 |
freemangordon | mtdoops? | 22:29 |
arno11 | ok i'll try | 22:32 |
arno11 | thx | 22:34 |
Wizzup | arno11: if you tell me how to reproduce I can hook up my serial too late rtonight | 22:37 |
arno11 | ok cool and thx. For example you can just run 'ir-ctl -S rc5:0x1e01' to produce the oops | 22:41 |
arno11 | (it happens using lirc and pierogi as well) | 22:43 |
arno11 | with any command | 22:44 |
Wizzup | lirc I can install with apt? | 22:48 |
Wizzup | and I guess I need to load bt, or? | 22:48 |
sicelo | no, bt isn't needed | 22:48 |
Wizzup | ok | 22:48 |
sicelo | and for the test, ir-ctl is enough. no lirc needed | 22:49 |
Wizzup | as in it will take 1-2 hours before I can do it, so just asking any questions I might otherwise have when you guys are asleep :D | 22:49 |
sicelo | you most likely will already have ir-ctl on the device. it's provided by v4l-utils | 22:50 |
arno11 | and the ir driver is already loaded on boot | 22:53 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!