| Wizzup | freemangordon: need anything from me wrt above patch? | 07:15 |
|---|---|---|
| freemangordon | no | 07:15 |
| freemangordon | unless if you want to fix CI bug | 07:18 |
| Wizzup | what ci bug? | 07:23 |
| freemangordon | see #leste-ci | 07:24 |
| Wizzup | arno11 I'll use my leste n900 more regularly now, but my other sim in it | 08:43 |
| Wizzup | arno11: I do have to say that on my end I still quite haven't gotten the n900 to not be quite laggy, maybe I just never found the right microsd card | 09:15 |
| Wizzup | I'll try 6.6 this week | 09:15 |
| arno11 | Wizzup: ok cool | 09:33 |
| arno11 | oops didn't see last msgs | 09:34 |
| arno11 | should be an easy stuff to boost it | 09:36 |
| Wizzup | I don't think it is boost | 09:40 |
| arno11 | with boost, i mean speed up | 09:42 |
| arno11 | cpufreq boost is not really needed | 09:42 |
| Wizzup | right | 09:44 |
| arno11 | when the sd is quite slow, the main problem is trackers and transitions | 09:44 |
| Wizzup | yeah but I think I bought the right sd card | 09:44 |
| Wizzup | also transitions are off I think | 09:44 |
| Wizzup | this one says U3 A2 (sandisk extreme 128gb) | 09:45 |
| arno11 | a bit overkilled lol | 09:46 |
| arno11 | but should work fine | 09:47 |
| arno11 | do you still have trackers in $HOME ? | 09:47 |
| Wizzup | overkill? | 09:48 |
| Wizzup | so it should be good / fast? | 09:48 |
| Wizzup | I don't know, maybe my emmc is worn out or something :) | 09:48 |
| Wizzup | (for swap) | 09:48 |
| arno11 | yes ofc this card should be ok, yes indeed emmc is maybe the problem | 09:49 |
| arno11 | i recommend you to remove gnome tracker from $HOME, remove mafw tracker from xsession (for the moment), and use a swapfile | 09:51 |
| Wizzup | swapoff -a takes a (very) long time so indeed maybe emmc is busted | 09:52 |
| arno11 | (for swapfile, commands from the wiki work fine) | 09:52 |
| arno11 | that was the same for me | 09:53 |
| Wizzup | on 6.1 I think | 09:53 |
| Wizzup | btw | 09:53 |
| Wizzup | also I used this for swap | 09:53 |
| Wizzup | fallocate -l 1g /swap | 09:53 |
| Wizzup | mkswap /swap | 09:53 |
| Wizzup | chmod 0600 /swap | 09:53 |
| Wizzup | swapon /swap | 09:53 |
| Wizzup | ooh | 09:53 |
| Wizzup | I was already on 6.6 | 09:53 |
| Wizzup | lol | 09:53 |
| arno11 | lol | 09:53 |
| arno11 | ok so makes sense for emmc now | 09:54 |
| Wizzup | it's such a difference lol | 09:54 |
| arno11 | indeed | 09:57 |
| arno11 | btw once your install is ok you should try the last custom transitions.ini: H-D is amazingly smooth :) | 10:00 |
| Wizzup | rught | 10:01 |
| Wizzup | right* | 10:01 |
| Wizzup | first upgrading to latest 6.6 that we have | 10:01 |
| arno11 | yep | 10:01 |
| Wizzup | arno11: what is your /etc/fstab line for swap? | 10:35 |
| arno11 | currently /swapfile swap swap defaults 0 0 | 10:40 |
| Wizzup | ok, I think my line didn't work so then it didn't manage to boot all the wya, added emmc back now and will see what's up | 10:40 |
| Wizzup | will be slow of course | 10:41 |
| arno11 | ok | 10:41 |
| arno11 | btw for ucm2, switching between speaker and headset works fine on the fly but i wonder how to keep the headset active between calls (in a car for example) | 13:52 |
| arno11 | because sphone switches back to hifi profile at the end of each call (which is fine) so we actually depend on jackcontrol at the beginnig of each call | 13:54 |
| arno11 | uvos: ^^^ any tips ? i suppose route-pulseaudio module could help | 13:55 |
| arno11 | the use of the handset button could do the trick btw | 13:57 |
| arno11 | really not sure what could be the right approach | 14:01 |
| * sicelo doesn't understand the question | 14:03 | |
| arno11 | for example: you take a call, it switches to voicecall, you plug your headset, you take the call, hangup, back to HiFi and you keep your headset for another call | 14:06 |
| arno11 | you have to unplug/plug again your headset | 14:07 |
| arno11 | to use it again | 14:07 |
| sicelo | ok | 14:11 |
| arno11 | the tricky stuff is how ucm could know if you want to keep your headset or not | 14:12 |
| arno11 | that's why hanset button in sphone could help imo | 14:12 |
| sicelo | sounds like incorrect ucm config though. HiFi should have a headphone and speaker section, as should VoiceCall | 14:12 |
| arno11 | *handset button | 14:12 |
| arno11 | that's already like that | 14:12 |
| arno11 | that's not enough | 14:13 |
| sicelo | so sphone chooses HiFi.speaker while headphone is connected? | 14:13 |
| arno11 | nope | 14:13 |
| arno11 | oops | 14:13 |
| arno11 | yes | 14:14 |
| arno11 | *if headphone is already connected | 14:14 |
| arno11 | we depend of jackcontrol but maybe there is a way in ucm to check if something is already plugged ? | 14:15 |
| uvos__ | the probelm here is that when setting profile pa returns all controls to the state that where current when the profile was last active | 14:19 |
| uvos__ | mostly we want this, but pluging in a hp while the call is going and expecting to continue to use it in hifi after is a corner case | 14:19 |
| arno11 | ok | 14:19 |
| uvos__ | needs research in pa to see if the restoreing can be controlled on a per control basis | 14:20 |
| arno11 | for the corner case, yes indeed | 14:22 |
| arno11 | that's why i'm asking | 14:22 |
| arno11 | ah maybe i found the trick in ucm :) | 14:58 |
| arno11 | jackcontrol is looking for a specific alsactl event but JackDev seems to just look @dev/input | 14:59 |
| arno11 | i'll try a bit later, no headset around | 15:06 |
| uvos__ | afaik pa dosent supprt jackdev at all | 15:19 |
| uvos__ | its a legacy interface no one should be using | 15:19 |
| uvos__ | yes seams so https://gitlab.freedesktop.org/pulseaudio/pulseaudio/-/issues/1377 | 15:20 |
| uvos__ | note that we dont use ucm directly at all | 15:20 |
| uvos__ | this ucm is a alsa thing, pa mearly has a compatability layer that allows it to load ucm modes as pa profiles | 15:21 |
| uvos__ | which is what we use | 15:21 |
| uvos__ | we dont use that actuall ucm from alsas libararies | 15:21 |
| arno11 | indeed JackDev doesn't work :( | 15:24 |
| arno11 | ok | 15:24 |
| freemangordon | ok, mafw-tracker-source memusage was reduced to ~20MB | 17:37 |
| freemangordon | arno11: ^^^ | 17:37 |
| uvos__ | thats still pretty oof on n900 | 17:50 |
| uvos__ | 10% of device ram for the media player _indexer_ | 17:50 |
| Wizzup | not sure if it's any different on fremantle | 17:51 |
| Wizzup | and that worked well, given 768mb of swap | 17:51 |
| uvos__ | i tried fremanlte for a bit and disagree that it works "well" but ok :P | 17:51 |
| arno11 | freemangordon: oh cool | 18:53 |
| arno11 | sicelo: i think i'll try to move speaker cset stuff in ucm boot sequences: this way, when stuff is back to hifi after a call, it should keep headset activated until next jack unplug event | 19:01 |
| sicelo | yes, most of the stuff belongs there (boot sequence) | 19:02 |
| arno11 | yes ok makes sense | 19:02 |
| sicelo | actually, i was looking at the ucm a few minutes ago | 19:02 |
| arno11 | ah ok :) | 19:03 |
| sicelo | why the boot sequence currently sets volume twice ? | 19:03 |
| arno11 | that's a known trick to avoid sound issue on boot | 19:03 |
| sicelo | ok | 19:03 |
| arno11 | i'll try a bit later and LYK | 19:05 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!