humanbeta | bug report: With Pinephone volume settings get lost in restart, after restart volume always in maximum. | 13:01 |
---|---|---|
buZz | used to be the same on droid4 , reboot always reset the volumes to whatever was default | 13:07 |
humanbeta | so it does not read saved volume setting..or does it even save? | 13:10 |
buZz | i believe it wasnt saving them at the time, not sure if thats implemented currently | 13:10 |
buZz | really should dig up my d4 and retry :) | 13:11 |
humanbeta | Maemo 5 with N900 saves and it have even different volume setting for phone calls. | 13:15 |
humanbeta | so that make me think feature request: add volume slider to sphone for save earpiece volume. | 13:17 |
humanbeta | with Maemo 5 I used volume buttons during call to set earpiece volume and it saved it automatically. | 13:20 |
buZz | i dont think it's sphone's role to do so, but just normal OS | 13:22 |
humanbeta | if OS saves volume settings for each apps then it is OS. | 13:26 |
buZz | its not a app volume, voice volume is a seperate mixer setting, see 'alsamixer' or 'pavucontrol' output | 13:27 |
buZz | the audio path of calls is a seperate hw path for audio | 13:27 |
Wizzup | I think the status area saves and should restore | 13:28 |
Wizzup | or maybe it's pulse | 13:29 |
humanbeta | but I think phone call volume in phones are special situation. Maybe best would be save call volume settings in addressbook so you can set it for best for every caller. | 13:29 |
buZz | humanbeta: either way, the mixer is in alsamixer/pavucontrol for you to change ;) | 13:29 |
buZz | you could likely force it to a setting in /etc/rc.local with some command i guess | 13:30 |
buZz | i havent seen any phone saving 'call volume per number' :) but that would be a fun feature yeah | 13:31 |
humanbeta | Maemo Leste should have something that any another phone OS does not have. | 13:32 |
humanbeta | ok..it has already something, but.. | 13:33 |
humanbeta | something that every user understand. | 13:33 |
humanbeta | but first things first, volume settings have to work during call. | 13:37 |
buZz | they do, open alsamixer ;) | 13:38 |
buZz | or pavucontrl | 13:38 |
buZz | control* | 13:38 |
humanbeta | yes I have pavu control, but how to make shortcut to desktop, now I have to go debian-folder and start it..it is not easy when you speak in phone. | 13:41 |
humanbeta | if volume buttons are not possible to do that how about some keyboard buttons? | 13:44 |
humanbeta | btw..it would be nice write SIM–pin code (in start up) with keyboard too (you can do that with PostMarketOS). | 13:47 |
humanbeta | btw..does Maemo use both GPU cores and acceleration of Pinephone in programs in Debian folder?...text rendering has someting badly wrong when testing with Hardinfo or even simple with UXTerm. | 14:02 |
humanbeta | badly wrong when speaking speed. | 14:02 |
Wizzup | what do you mean 'badly wrong' ? | 14:03 |
humanbeta | very slow speed in text rendering. | 14:03 |
Wizzup | is it better in portrait mode? | 14:04 |
Wizzup | I think there is still some slowdown in landscape mode due to everything needing to be rotated | 14:04 |
humanbeta | is hardware rotate only in closed source Mali drivers? | 14:05 |
Wizzup | I don't think it has something to do with the closed source drivers, but rather glamor | 14:06 |
Wizzup | I don't have more info/details, others might | 14:07 |
buZz | imho, pinephone's GPU speed was always a bit 'meh' | 14:09 |
buZz | compared to droid4 or n900 | 14:09 |
humanbeta | but anyway X Terminal what comes with Maemo Leste has much faster text rendering than UXTerm or Xterm in Debian folder. | 14:10 |
humanbeta | they had come from default repository. | 14:12 |
humanbeta | buZZ: I afraid that open source Mali drivers, Lima drivers, does not use full potentional of GPU. And if Maemo does not use full potentional of those drivers..it would be problem. I remember that problem with OQO handhelds..they had quite good Silicon Motion Lynx3D+ GPU, but nobody ever made good accelerated Linux drivers although Silicon Motion | 14:24 |
humanbeta | gave all info. | 14:24 |
buZz | humanbeta: from all i saw, the pinephone just has a A64 soc with a pretty old/basic mali version | 14:24 |
buZz | plus a -way- too high resolution display for that gpu | 14:25 |
buZz | also; mali is just the -3D- gpu | 14:25 |
buZz | not the 2D one | 14:25 |
Wizzup | humanbeta: the lima drivers are very unlikely the problem | 14:26 |
Wizzup | did you try it in portrait like I asked? | 14:26 |
humanbeta | Wizzup: some how I did not get UXterm and Xterm to start to portrait...X Terminal of Maemo rotate well..it was not slow in first place. | 14:33 |
humanbeta | buZz: Actually AllWinner has dual core Mali version with nice dual alpha layer 2D rotate and zoom mixing add on, but it might be hard to use in normal X11 use. | 14:35 |
buZz | humanbeta: i know the mali400, its in many socs | 14:36 |
humanbeta | do you know can it speed up with 16 bit per pixel mode? Actually how many real colors Pinephone LCD can show? | 14:45 |
freemangordon | humanbeta: bad performance in landscape mode is because modesetting/glamor rotate via SW | 14:54 |
freemangordon | this has nothing to do with maemo | 14:55 |
freemangordon | or GPU being slow | 14:55 |
freemangordon | you can compare dragging desktop views landscape vs portrait | 14:56 |
freemangordon | maybe open a bug against modesetting driver | 14:56 |
freemangordon | also, unfortunately 16bpp is a nogo | 14:57 |
uvos__ | true but i mean the pinephones gpu is still pretty slow, its fillrate is bearly faster than the sgx in d4 and it pushes way more pixels | 14:57 |
freemangordon | lots of 'modern' programs/frameworks assume 24bpp | 14:57 |
freemangordon | uvos__: right, but it is fast enough to do 60fps in portrait vs 15fps in landscape, because of SW rotation | 14:58 |
uvos__ | resotring volumes is the resposablility of pulse | 14:58 |
uvos__ | and it works fine on d4 | 14:58 |
uvos__ | should work on pp too | 14:59 |
uvos__ | the issue on d4 previously was that one of the mixers had an invalid name | 14:59 |
humanbeta | and nobody tried to use of hardware rotate in start up to change coordinates (of course portrait would be then slower) | 14:59 |
uvos__ | causeing the saveing to error out | 14:59 |
uvos__ | humanbeta: ? | 15:00 |
humanbeta | uvos: has anybody tried use hardware rotate of AllWinner? | 15:00 |
uvos__ | the way rotation should work in glamor is that the randr extentions sets a transformation matrix on every x surface and then glamor renders every surface on its side via ogl, dosent work on gles however | 15:00 |
Wizzup | weren't a lot of patches submitted for glamor gles a few months ago? | 15:08 |
Wizzup | lot of glamor/xv stuff too recently https://gitlab.freedesktop.org/xorg/xserver/-/commits/master | 15:09 |
freemangordon | humanbeta: it does not seem that modesetting supports HW rotate :( | 15:13 |
freemangordon | unless that has been fixed recently | 15:13 |
freemangordon | Wizzup: yeah, I gess we shall upgrade xorg at some point | 15:13 |
freemangordon | *guess | 15:13 |
freemangordon | uvos__: glamor on lime should use GL, not GLES | 15:13 |
freemangordon | *lima | 15:14 |
maxwelld | can someone who built qwebbrowser patch it also in a way that it would open the link it gets as a commandline argument? when clicking a link in somewhere, lets say this chat, it opens qtwebbrowser for me (dont remember if i edited something for that by hand or just made sure this is the last installed browser) but when opened it just opens blank page. i guess it has no code to open a link it gets fro | 15:16 |
maxwelld | m commandline. | 15:16 |
buZz | maxwelld: alas qtwebbrowser is just a sdk example, not really a full browser :( | 20:32 |
maxwelld | yeah i know... i guess it is possible to modify it to use argc/argv and pass it to the rendering engine by default. | 20:40 |
buZz | sure, have at it :D | 20:40 |
buZz | quite sure there's nothing changed to the upstream sdk example | 20:41 |
dsc_ | i am working on improving jib a bit, ill make sure to handle that use-case | 20:42 |
maxwelld | jib? what is that? | 20:45 |
dsc_ | https://github.com/maemo-leste-extras/jib | 20:48 |
dsc_ | its considered alpha still, needs UX improvements | 20:48 |
dsc_ | but you can try it out | 20:48 |
maxwelld | isn't it on repos? | 20:50 |
maxwelld | is there a repo i can add? how does it work on d4? | 20:50 |
Wizzup | it's in extras, but I think dsc didn't yet add a X-Maemo... entry and a desktop ico | 20:51 |
Wizzup | n | 20:51 |
buZz | thats not needed when you dont run maemo ;) | 20:51 |
maxwelld | good it is in extras. | 20:52 |
maxwelld | Wizzup, i am trying to build pango, libadwaita, gtk4 to build latest dino, since this dino has lots of bugs. | 20:52 |
maxwelld | if i succeed i'll ask for repos, hope you are ok with it. | 20:52 |
Wizzup | I am not sure if gtk4 apps will look any good currently | 20:54 |
maxwelld | you cannot imagine how buggy this dino is. | 20:55 |
Wizzup | I can I think, I use it daily on my laptop | 20:55 |
Wizzup | I bet it's very power hungry too | 20:56 |
maxwelld | i actually use d4, so i need communication. everybody around uses omemo, otherwise i would be fine with pidgin. | 20:56 |
maxwelld | yes, it is very power hungry. both old version and new version. | 20:56 |
maxwelld | i carry powerbank with me, my battery is also not in good shape. | 20:56 |
Wizzup | we have an open funding goal to add omemo to tp | 20:56 |
Wizzup | in case anyone is interested | 20:56 |
maxwelld | i don't think i am the best person for that. | 20:57 |
Wizzup | freemangordon: I wonder if we maybe want to run tracker only when phones are connected to external power | 21:34 |
freemangordon | Wizzup: why? | 21:58 |
Wizzup | because it drains battery in these semi corrupt states | 21:58 |
freemangordon | isn't it better to just fix the issue? | 22:05 |
uvos__ | sure | 22:08 |
uvos__ | but i think at the very least it should be removed from the metapackages until it works | 22:09 |
freemangordon | it is in -devel | 22:09 |
uvos__ | even in -devel | 22:09 |
freemangordon | and -devel is meant to break your devce | 22:09 |
freemangordon | *device | 22:09 |
uvos__ | havin pm broken in devel for long lengths of time has the issue of masking subsiquent issues intrduced | 22:09 |
uvos__ | that can be quite hard to find | 22:09 |
uvos__ | so its not great | 22:10 |
uvos__ | even in devel | 22:10 |
freemangordon | tracker seems to have issue that is easily reproducible on Wizzup's device | 22:10 |
uvos__ | it happens on my device too | 22:10 |
freemangordon | ok, then please, either attach gdb or provide logs | 22:10 |
freemangordon | disabling tracker will not help to move ahead | 22:11 |
uvos__ | its not so easy to repduce sometimes it just gets into states where it uses to mutch cpu | 22:11 |
uvos__ | but its qute random | 22:11 |
freemangordon | I am sure that high cpu usage is because it crashes | 22:11 |
uvos__ | it dosent crash | 22:11 |
freemangordon | how do you know? | 22:11 |
uvos__ | pid dosent change | 22:12 |
freemangordon | I think on Wizzup's device it crashes | 22:12 |
freemangordon | otherwise it does not affect battery | 22:12 |
freemangordon | at least here | 22:12 |
uvos__ | powertop lists it as a lot of wakeups/S | 22:13 |
freemangordon | tracker wakes-up every 30 seconds | 22:13 |
uvos__ | causeing the device to use ~400mw | 22:13 |
uvos__ | but its not wedged or anything | 22:13 |
freemangordon | BTW, when we you 'tracker', what exactly do you mean | 22:13 |
freemangordon | because there are couple of processes involved | 22:14 |
freemangordon | tracker-miner, tracker-fs, etc | 22:14 |
freemangordon | *when you say | 22:14 |
uvos__ | have to check witch process it is | 22:14 |
uvos__ | its allways the same one | 22:14 |
freemangordon | ok, that might be the one starting the other, the other crashes and 'the one' restarts the other, etc | 22:15 |
freemangordon | thus high CPU usage | 22:15 |
freemangordon | do you have any idea how to repro? | 22:15 |
freemangordon | did you do tracker reset perhaps? | 22:18 |
* freemangordon copies 7GiB of music to his d4 | 22:23 | |
dsc_ | all CC0 licensed | 23:04 |
Wizzup | freemangordon: agree trying to fix it is better | 23:21 |
freemangordon | Wizzup: you can enable logs | 23:22 |
freemangordon | see TRACKER_USE_LOG_FILES | 23:23 |
freemangordon | and also "tracker daemon --set-log-verbosity" | 23:23 |
Wizzup | ok | 23:23 |
Wizzup | will do (later tonight) | 23:23 |
freemangordon | maybe there are more logs | 23:24 |
freemangordon | but loets start with those | 23:24 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!