| kiva | At last I found reason for unexpected power-offs with Pinephone Keyboard, reason is simple phone side battery can go empty in whole day use, although connected with charger. | 09:30 |
|---|---|---|
| gnarface | oh, i could have warned you about that, sorry | 09:32 |
| gnarface | even if you have a spec charger for those things, it's easy to outrun it with regular CPU load | 09:33 |
| gnarface | they can only pull in about 15W as a theoretical max, but i've only seen mine able to draw about 11 | 09:33 |
| gnarface | so yea, uh... keep the screen brightness low | 09:34 |
| gnarface | sleeping it is better | 09:34 |
| kiva | Solution 1. We need two battery indicators near clock. Solution 2. Slow down CPU that much that Keyboard battery can give enough power when 10% phone side battery left. | 09:34 |
| kiva | My screen brightness is lowest anyway, because I like when white is not any brighter than papers near me...it would be nice if there is more brightness levels between lowest and whole black. Possible? | 09:43 |
| kiva | Solution 3. Phone go hilbernate when only 5% battery left and wake up when Keyboard side battery has given enough power to 10%. | 09:47 |
| kiva | My charger takes between 24.9-25.1 Watts (measuring right now), but I think connection between keyboard side and phone is much less. | 09:52 |
| kiva | Now I miss feature from 4 core PowerPC Macs, power off cpu cores one by one. | 09:59 |
| gnarface | the screen brightness thing is theoretically possible but afaik usually not without swapping the LCD itself for a different model; i know this only because early on there were different models of LCD in different batches of pinephones, and it started causing problems where minimum brightness would actually turn the screen off completely for some people while still be too bright for others | 09:59 |
| gnarface | ...though some DTB changes in theory might also improve matters if it's a situation where the panel isn't being exactly properly powered - i don't know for sure that those fixes are all completely dialed in for every model, they may have just shipped some compromised middle ground | 10:00 |
| gnarface | the problem might largely be about just finding proper specifications through some method better than trial+error | 10:01 |
| gnarface | for some models it seems they just have a higher minimum brightness too though, so improvements aren't guaranteed | 10:02 |
| gnarface | (their deal with the manufacturer allowed the manufacturer to swap out components "of equal value" as supply issues arose and the LCD and the RAM were two things the manufacturer was especially "fluid" about) | 10:03 |
| gnarface | as for powering off CPU cores individually ... i believe this hardware actually does have that feature in fact... it's pretty common on ARM i'm told, but last person i heard talking about messing with that couldn't bring them back up without crashing the phone | 10:05 |
| gnarface | some software work no doubt needs to be done in the kernel for that still | 10:05 |
| kiva | So good old Mac OSX days 20 years ago they could do that but Linux cannot still cannot handle it? | 10:07 |
| gnarface | heh, well ARM is still very immature on Linux... which is saying something because Linux is well ahead of BSD still | 10:07 |
| gnarface | the best ARM support is still locked up in proprietary Google-kernel-only Android parts, but i'm told that the support we have in Linux is very much on par with Android from 10 or so years ago | 10:08 |
| gnarface | so yea, just give it time | 10:08 |
| gnarface | ah a PowerPC phone... now that would be something to see :) | 10:10 |
| gnarface | i kinda feel like we might have really lost something important when Apple switched to Intel chips... | 10:10 |
| gnarface | i didn't at the time, but i'm coming around | 10:10 |
| kiva | We really lost when Apple moving to Intel and away from path what was built OSX 10.4. After that all with Apple going wrong. | 10:13 |
| kiva | I think there will be huge market for small retro PowerPC-machine that "accidentally" can run Mac OS9-OSX10.4 and use old firewire hardware. | 10:24 |
| kiva | Has anybody try SheepShaver MacOS9 emulator with Pinephone? ;) | 10:27 |
| kiva | or Droid4? | 10:32 |
| kiva | if somebody test, please write there how well it works: https://talk.maemo.org/showthread.php?t=101449 | 11:00 |
| uvos__ | sphone theoreticly has support for dtmf, but no real backend (ie not commtest) implements support. Mainly because the d4 dosent support dtmf on a sane interface. if dtmf works on n900 via ofono dibus calls it should be trivial to implement in sphones ofono backend | 15:50 |
| uvos__ | vcm im not sure, but probubly would not be too hard either | 15:50 |
| uvos__ | point is that all the ui bits and sphone framework bits are there for dtmf | 15:51 |
| Wizzup | does setting the screen brightness on the pinephone ewor kfo you guys? | 17:41 |
| Wizzup | work for you guys* | 17:41 |
| sicelo | kiva: i have considered your option one in the past ... see https://github.com/maemo-leste/bugtracker/issues/726 ... so your #1 is #2 in the first comment there. anyway, maybe someday i'll have time to implement it. not making any promises :-) | 18:00 |
| sicelo | uvos__: seems n900's dtmf is now broken too. bitrot i guess | 18:01 |
| sicelo | btw, something to implement in sphone - the call history should show the direction of the call (incoming/outgoing) | 18:04 |
| Wizzup | and it should take the remote_name from the db instead of matching it to a contact I think | 18:10 |
| sicelo | kiva: by the way, out of curiousity/interest, does the pinephone keyboard's battery actually show up in leste's /sys/class/power_supply/ ? | 21:31 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!