| Wizzup | freemangordon: do you think it'll be a lot of work? mafw-gst-renderer? | 12:25 |
|---|---|---|
| freemangordon | no idea | 13:20 |
| arno11 | Wizzup: combining your n900-pm script (with off-mode blocked) + cpufreq conservative + overclock 250-850MHz + modem ON + hsdpa ON and connected, I get 44mA/165mW idle ! (battery 1337mA / 30 hours idle) | 13:30 |
| arno11 | and for calls, it wasn't working anymore after Sunday's upgrade...now it's repaired and i have a simple process to test on a fresh install before we send stuff to Pavel | 13:51 |
| arno11 | it still works only in one way | 13:51 |
| arno11 | but no crash or weird things | 13:52 |
| arno11 | and sound is good | 13:52 |
| Wizzup | ok, please mail any instructions :D | 13:53 |
| arno11 | ok | 13:54 |
| Wizzup | and great @ pm | 13:55 |
| Wizzup | that's probably the serial idle? | 13:55 |
| arno11 | the serial idle ? | 13:55 |
| Wizzup | the effect of the pm script is just to idle the serials | 13:59 |
| arno11 | ah yes ok | 13:59 |
| Wizzup | n900 has a serial port | 13:59 |
| dsc_ | I looked a bit at porting `mafw-gst-renderer` and its quite a bit of work I think | 19:31 |
| dsc_ | for one it depends on "libhal" https://gitlab.freedesktop.org/archived-projects/hal/hal | 19:31 |
| dsc_ | which in turn depends on an older version of glib | 19:32 |
| freemangordon | libhal is replaced with udev | 19:32 |
| freemangordon | afaik | 19:32 |
| dsc_ | exactly | 19:32 |
| freemangordon | but it udev should provide the same functionlaity | 19:32 |
| freemangordon | or similar | 19:32 |
| dsc_ | sure, its not the end of the world in terms of implementation, but I was seeing if I could port `mafw-gst-renderer` myself | 19:33 |
| freemangordon | ah | 19:33 |
| dsc_ | but this libhal->udev might be too much for me | 19:33 |
| dsc_ | ;') | 19:33 |
| freemangordon | what exactly does it need from libhal? | 19:33 |
| dsc_ | well | 19:33 |
| freemangordon | anyway | 19:34 |
| dsc_ | I think libhal is used to signal if the "tv output" is connected/disconnected | 19:34 |
| freemangordon | oh | 19:34 |
| dsc_ | so, the screen basically | 19:34 |
| freemangordon | well... that's something we shaould reconsider | 19:34 |
| freemangordon | *should | 19:34 |
| freemangordon | like, what behaviour do we want | 19:35 |
| freemangordon | ok, so... | 19:35 |
| dsc_ | brb supermarket | 19:35 |
| freemangordon | lemme do mce input-ctrl module firs (whatever it is) | 19:35 |
| freemangordon | and then I'll try to port that | 19:35 |
| freemangordon | brb | 19:35 |
| Wizzup | I think we can stub that signal for now, we just need it in some working state to test OMP | 19:40 |
| Wizzup | dsc is working on OMP atm | 19:40 |
| freemangordon | right | 19:46 |
| Wizzup | even maybe just copying a db from fremantle could work | 19:47 |
| Wizzup | well, I guess not, for the renderer | 19:47 |
| dsc_ | freemangordon: some info I guess: https://github.com/maemo-leste/mafw-gst-renderer/issues/1 | 20:12 |
| Wizzup | gst 0.10 -> 1.0 is a little annoying, depending on what is used | 20:29 |
| Wizzup | both fmg, uvos and I have done the ports before | 20:29 |
| Wizzup | there is also gnome vfs -> gio porting | 20:33 |
| Wizzup | but as I understand it, this is just one piece, right? | 20:41 |
| Wizzup | this isn't actually the db indexing | 20:41 |
| dsc_ | i got the impression that db stuff is done by libmafw | 20:41 |
| dsc_ | in OMP, the `player/mafw` directory provides a Qt wrapper around libmafw | 20:43 |
| dsc_ | actual playlist management/logic is in libmafw | 20:43 |
| Wizzup | right, but there is also a daemon that searches the fs for music files and adds them automatically | 20:43 |
| Wizzup | and I am not sure what does this | 20:44 |
| dsc_ | am sure that after fixing the renderer we'll have other OMP portings to do, for example I have not verified that the QML stuff works (OMP has some QML bits) | 21:02 |
| dsc_ | but at least now its ported to CMake and we know that the renderer is broken | 21:03 |
| Wizzup | right | 21:14 |
| Wizzup | arno11: thanks for the mail, mind sending it to the ml too? | 21:20 |
| Wizzup | dsc_: for some reason I thought the qml bits were what was causing the white screen | 21:20 |
| Wizzup | but now I guess I realised it's just an empty list? | 21:20 |
| Wizzup | or is even the playling mafw gst, and not qml/qt? | 21:20 |
| Wizzup | playlist* | 21:20 |
| arno11 | Wizzup: sure, but maybe we should try the process on a quiet 'fresh clean install' first | 21:24 |
| arno11 | anyway it doesn't matter | 21:30 |
| Wizzup | I mean, in general having to kill PA is not ideal, but it's a starting point | 21:30 |
| Wizzup | I think I was able to also get the same call audio that you do, without killing PA | 21:30 |
| Wizzup | but not both sides | 21:30 |
| arno11 | yes i was able too but it was really unstable | 21:36 |
| arno11 | and now it doesn't work anymore on my device | 21:36 |
| arno11 | sound is greater with alsa and better in cpu usage | 21:37 |
| arno11 | but probably impossible with just alsa to get sound both sides | 21:37 |
| arno11 | dealing with pa is essential | 21:47 |
| arno11 | pulseaudio-module-cmtspeech is maybe/probably the key ? | 21:50 |
| dsc_ | Wizzup: (afaik) the screen is white because there is no mafw *renderer* | 22:20 |
| dsc_ | as in, it tries to render things but there is no renderer :P | 22:20 |
| dsc_ | I think the db/playlist comes in later, and yes playlist is part of mafw | 22:20 |
| dsc_ | well, it may not immediately try to render things per-se, but it does try to setup a "rendering context" upon constructing the video or music view | 22:24 |
| dsc_ | this is what is making it white | 22:24 |
| Wizzup | hmm I see | 22:24 |
| Wizzup | arno11: well, per the mail from pavel he had some suggestions on what to look at first | 22:24 |
| Wizzup | arno11: btw, re: n900 pm, I wonder if we uncomment the off mode line for now but otherwise ship it | 22:25 |
| Wizzup | (for serial pm) | 22:25 |
| arno11 | (for pavel) ok | 22:31 |
| arno11 | Wizzup: (for pm) there is still the weird lightning screen using off mode | 22:34 |
| arno11 | we can ship it | 22:35 |
| arno11 | i mean, the 4 remaining blockers are weird | 22:36 |
| arno11 | what are they exactly ? kernel modules ? | 22:37 |
| freemangordon | Wizzup: indexi | 22:38 |
| freemangordon | part is tracker | 22:38 |
| freemangordon | *indexing | 22:38 |
| Wizzup | arno11: well, when we don't enable off mode we can still idle serial | 22:49 |
| Wizzup | arno11: well, clearly touch screen, as it seems to still use power :P | 22:50 |
| Wizzup | (one of them) | 22:50 |
| arno11 | Wizzup: yes and yes indeed | 22:51 |
| arno11 | i just have a look on last pavel emails | 22:52 |
| arno11 | and the question is how enabling full duplex mode on n900 | 22:54 |
| arno11 | i already tried but no succes :( | 22:54 |
| arno11 | anyway it's late, i'll try to find docs about it tomorrow | 22:58 |
| arno11 | gn | 22:59 |
| Wizzup | gn | 23:10 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!