| freemangordon | n900, gtkperf, with traces removed, 2D slowpath rendering: | 08:00 |
|---|---|---|
| freemangordon | Total time: 62.33 | 08:00 |
| Wizzup | freemangordon: sweet | 09:48 |
| lel | MerlijnWajer closed a pull request: https://github.com/maemo-leste/leste-config/pull/25 (Various leste-config changes) | 10:41 |
| Wizzup | uvos: when we go for drmfb on n900 we might need to revert the changemode directfb maybe | 10:44 |
| _inky | https://forum.pine64.org/showthread.php?tid=9444 hope leste will work after flushing like that from the internal emmc. | 11:10 |
| _inky | you cannot imagine how much i hope pinephone will become usable under maemo. i believe it can, if the gfx issues get fixed. because droid4 is very usable for me. | 11:12 |
| _inky | and i have two friends that have pinephones, and pinephone is slow under plasma and phosh, but will be very fast under maemo. | 11:12 |
| uvos | thb i think even if the gfx issues are fixed by disabeling glamor (which is likely) i would not call leste very usable on pp until we improve the quality and functionality of him | 11:16 |
| uvos | i honestly cant use the bionic in ernest | 11:16 |
| uvos | even though it works slightly better than the d4 generally | 11:16 |
| uvos | Wizzup: what changemode for driectfb? | 11:19 |
| uvos | Wizzup: i dident add anything | 11:19 |
| uvos | besides once ddk1.17 works i intend to drop driectfb... | 11:19 |
| Wizzup | aha | 11:19 |
| uvos | because charging sdl works fine on kms/drm on ddk1.17 | 11:20 |
| uvos | what else would you want directfb for? | 11:20 |
| Wizzup | fair | 11:20 |
| uvos | btw we should activate charge-mode on the deivces | 11:21 |
| uvos | atm its not even in devel | 11:21 |
| Wizzup | ok | 11:21 |
| uvos | it hasend caused me any trouble and mce-battery guard has the unforitonate side effect that if you do manage to get well below the theshold voltage mce will shut the device down just as it boots with usb plugged in | 11:23 |
| uvos | without charging mode | 11:23 |
| uvos | with charging mode the device will hang around there ofc | 11:23 |
| uvos | probubly should add something to charging-sdl that prevents the device from booting to the default runlevel until the battery is at least 3.5V | 11:27 |
| Wizzup | do we have a pr to enable it? | 11:28 |
| uvos | no | 11:28 |
| uvos | it needs to be added to https://github.com/maemo-leste/mapphone-kexecboot-config | 11:28 |
| uvos | and some runlevel | 11:28 |
| uvos | and the cmdline on n900 | 11:28 |
| uvos | (somehow) | 11:28 |
| uvos | *and some metapackage | 11:29 |
| uvos | cmdline needs softlevel=charge-mode | 11:30 |
| uvos | i dont think cmdline on n900 is controled by a package? | 11:30 |
| uvos | same problem with pp | 11:30 |
| uvos | but pp dosent have battery-guard either | 11:30 |
| uvos | both would need testing first | 11:30 |
| Wizzup | _inky: pp with keyboard would be nice to support, but fixing lima really isn't our prio atm I think | 11:31 |
| Wizzup | also the battery life is kind of a problem for me personally | 11:31 |
| uvos | do wie sleep with crust at all atm? | 11:31 |
| Wizzup | only if you invoke it manuall | 11:33 |
| Wizzup | y | 11:33 |
| Wizzup | in any case I don't think hibernate-ram is a way to use a phone | 11:33 |
| Wizzup | but yeah | 11:33 |
| uvos | yeah its not ideal | 11:34 |
| uvos | you could make it suspend to ram for 50 of 60s of a minute or soemthing and have the deamons download new ip messages/emails etc during the 10s | 11:35 |
| uvos | wich is how android works on platforms that cant do runtime pm | 11:35 |
| uvos | but yeah | 11:35 |
| Wizzup | I don't know how you will listen to music that way | 11:36 |
| uvos | yeah you would need a concept of wakelocks like android | 11:36 |
| Wizzup | right, but then the battery would drain like razy | 11:38 |
| Wizzup | my point is mostly I'm not happy with the big power usage | 11:38 |
| Wizzup | but yeah it's still a cool device. | 11:38 |
| Wizzup | I'm sure we'll get to support it more when we get our other devices usable for daily usage | 11:38 |
| uvos | well d4 sleeps exactly never when you play music | 11:38 |
| uvos | on leste | 11:38 |
| Wizzup | It does for me I'm pretty sure | 11:38 |
| uvos | on android it seams to use the m3 cores | 11:39 |
| uvos | or something | 11:39 |
| Wizzup | I am pretty sure the ret count increases | 11:39 |
| uvos | what do you play? | 11:39 |
| uvos | 44khz flac? | 11:39 |
| Wizzup | opus files | 11:40 |
| uvos | ok | 11:40 |
| uvos | maybe its my files/codec | 11:41 |
| Wizzup | in any case yeah I don't get ~24 hours of music playback time | 11:41 |
| _inky | can i disable glabor myself or some package has to be rebuilt? | 11:58 |
| _inky | glamor* | 11:58 |
| _inky | > but fixing lima really isn't our prio atm I think | 12:30 |
| _inky | Wizzup: what i wonder is if other pinephone communities do some work or request something from lima or mesa, that we can benefit, or we live in a different world because of x11? | 12:30 |
| Wizzup | I think we live in a somewhat different world | 12:32 |
| Wizzup | I don't think disabling glamor helps with anything | 12:32 |
| Wizzup | on the pp | 12:32 |
| _inky | the pinephone pro's video is different, i guess the same as on pinebook pro. i have a pinebook, so i probably can test maemo, hildon somehow on it. | 13:26 |
| _inky | and we can know how it'll work on pinephone pro. | 13:27 |
| _inky | i guess i can just write some ARM image to sdcard and boot. | 13:27 |
| _inky | hopefully. | 13:27 |
| _inky | it's a pity, these gfx problems, because the only lightweight distribution that exists for pinephone is maemo, others are very heavy. | 13:29 |
| Wizzup | iirc that has a mali-t, so the driver might work better too | 13:29 |
| Wizzup | _inky: we'll get there, we just don't have infinte time | 13:29 |
| _inky | i didn't know that time may help, i thought gfx drivers or userspace are very specialized thing, and nobody except the writers can contribute. | 13:31 |
| Wizzup | the mali-t driver is developed more actively afaik | 13:31 |
| _inky | i guess, it works not bad on my gentoo on pbp, has some issues under my favourite windowmaker, but works much smoother under enlightenment. | 13:32 |
| _inky | on x11 without wayland i mean. | 13:32 |
| _inky | i don't compile wayland. | 13:32 |
| Wizzup | isn't panfrost the driver for mali-t? | 13:33 |
| Wizzup | https://docs.mesa3d.org/drivers/panfrost.html | 13:34 |
| Wizzup | in any case you can try to get people to help fix the things, I just don't think we (core folks) currently can | 13:34 |
| Wizzup | we really should focus on other stuff | 13:34 |
| Wizzup | there might just be a way we can use it better, and perhaps newer mesa and X will also help | 13:36 |
| Wizzup | I didn't bring my pp with me on this trip so I can't help there | 13:37 |
| Wizzup | you can contribute by making it easy for the developers to reproduce the bug(s) | 13:37 |
| sicelo | _inky: maybe see how sxmo handles x11, then check if any of what they do can help leste. Seems sxmo users are happy with what they get out of it | 14:04 |
| sicelo | Yeah i like enlightenment. I wanted to try it on n900 with wayland | 14:06 |
| _inky | yes i have sxmo in emmc | 14:12 |
| _inky | it is very useful. | 14:12 |
| _inky | sicelo: i did not know they use x11. | 14:12 |
| sicelo | Hehe, I mentioned it the other day but no one noticed. What do you the the X in sxmo means ;-) | 14:13 |
| sicelo | s/do you/did you think/ | 14:14 |
| _inky | oh. great. i didn't know that. (: | 14:16 |
| _inky | well xmo works perfectly ok with xorg then. | 14:16 |
| _inky | sxmo | 14:16 |
| _inky | never noticed any problems. may be they just use more up to date software stack, because they use alpine or manjaro. | 14:17 |
| _inky | hmmmm. | 14:17 |
| sicelo | So it might give you hints on how to improve stuff for leste support | 14:17 |
| _inky | interesting if i can get it compiled on devuan and replace hildon default start with sxmo to check. | 14:17 |
| freemangordon | ok, pvr mesa segfaults :) | 14:33 |
| freemangordon | will continue on Sunday, bb | 14:33 |
| uvos | i really dislike gtk | 18:08 |
| Wizzup | uvos: yeah, also why I'd prefer to go conversations in qt | 18:12 |
| uvos | yeah anyhow, almost finished with sphone moudlarsation now. | 18:35 |
| uvos | architecturally i borrowed the datapipe concept from mce (altho implementation differs slighty) as an exmple it works like this now when a call comes in: | 18:36 |
| uvos | some plugin that has called sphone_comm_add_backend gets a call and creates a CallProperties struct and kicks it down the call_new_pipe pipe | 18:37 |
| uvos | optionally some plugin that has registerd a filter on that pipe adds stuff to CallProperties->contact struct with name image etc | 18:38 |
| uvos | and it lands in the ui | 18:38 |
| uvos | the ui then kicks the call stuct back down a call_accept_pipe when it wants | 18:38 |
| uvos | just an example | 18:39 |
| uvos | but everything works like this now | 18:39 |
| Wizzup | check | 18:40 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!