**** BEGIN LOGGING AT Fri May 11 03:00:04 2018 May 11 09:18:35 MonkeyofDoom: for that other discussion according to https://wiki.debian.org/Debate/initsystem/systemd - 'Systemd’s upstream is very accommodating to distributors. They are taking a lot of Debian’s needs into account...' May 11 12:05:08 adam_b (IRC): but it's still bloody difficult to get it to mount a drive at boot! May 11 12:07:57 depscribe[m]: I've not tried it yet, but full debian install gives you an fstab which is respected as you say (TheKit used multistrap to create the debian image we're using) May 11 12:25:06 seems if you create an fstab it must contain an entry for the main root, it appears to replace not compiment the automagic (or it wont boot) May 11 13:29:57 adam_b: is there a key combination to enlarge windows, the way alt-drag moves them? May 11 14:26:17 I missed the conversation about touch and linux yesterday but would like to add that the work Purism is doing on Gnome may be helpful in the next year or so. May 11 14:26:36 https://puri.sm/posts/librem5-progress-report-11/ May 11 15:15:49 adam_b: btw, an observation from yesterday -- when copying kwinrc from a different user, it needs to be chowned or else there might be problems later. (as with pretty much anything copied from one user to another by brute force!) May 11 15:36:23 does anyone here actually use android on their gemini? May 11 15:37:21 I do, it's okay. May 11 15:38:08 i've had mine for a while, but i haven't had the chance to really do much with it May 11 15:38:31 I'll run linux on it if/when it seems like a viable everyday option, but I'm not too concerned May 11 15:38:38 plus i'm assuming there is more work needed to get it working daily May 11 15:38:41 yes May 11 15:38:55 i know, it's very early in its lifespan May 11 15:40:02 for now I'll just use the reasonably usable smartphone OS that's already there. I want phone calls, sms, notifications all to work or I won't be able to rely on it as my phone. May 11 15:41:34 apart from a few gimpy apps that don't like landscape mode, I haven't run into anything that makes me really resent using Android on it. May 11 15:42:48 but I'm also not too fussy, not really that interested in phones or operating systems, so me; I have a keyboard and that's nice. May 11 15:43:01 *meh May 11 15:47:06 indeed May 11 15:47:42 i'm endeavoring to use linux. haven't even set up googledroid except for a password, and when i flashed it, i used adam_b's scatter file that gives maximum space to linux. but you have to really want linux at this point for it to be something you would use all the time. i really want linux. May 11 15:48:18 understandable May 11 15:49:30 i might give lineage os a shot on the android side of things, but they don't have phone working yet, either. May 11 15:49:31 i've been trying for a very long time for a phone keyboard solutinon May 11 15:50:00 right now i'm using a twiddler with a oneplus May 11 15:50:20 which, honestly, has its advantages May 11 15:50:40 but it's not the same as a fully integrated keyboard May 11 15:52:56 the keyboard does take some getting used to. there are some things in unexpected places, and those who have mentioned that the keyboard is dark against the bright screen are right. and very good eyesight is a plus -- i'm hoping to get some of that, if i can find where it's offered. May 11 15:57:38 i've the key action to be a little wanting May 11 15:58:11 like it doesn't do so wellif you press a key a bit offset May 11 18:38:00 I use android on my Gemini. I'm actually using it right now to type this message, and I get about the same level of speed and accuracy on typing on the Gemini as I do on my laptop keyboards. I'd say I average about 40 WPM on this thing. The problem I keep running into is the weird way keyboard keys interact with apps. May 11 18:38:42 I'd use Linux, but for the life of me I can't get it to sleep. May 11 18:39:36 Give it time sure it will all work out.. the more devices land in hackers hands. May 11 18:40:27 systemctl sleep and echo "mem" |sudo tee /sys/power/state just mostly suspends for a few seconds. May 11 18:40:47 mondoshawan[m]: you've tried out the sleep section on the wiki? May 11 18:40:48 Yeah. I'm one of those hackers -- kernel stuff is my thing. I just wonder where the history went because the kernel sources up on github arent enough to make an image. May 11 18:41:13 Yeah. repowerd is installed, etc. May 11 18:41:22 well it should sleep on close May 11 18:41:52 The problem is the kernel. For some reason it sleeps and quiesces the cores, but something fires the wakeup IRQ/timer too soon and reawakens the kernel. May 11 18:42:00 hmm May 11 18:42:07 So the display turns off for a second and then comes back on. May 11 18:42:08 in dmesg you see battery stuff every few seconds May 11 18:42:10 sleep could be improved, you need to turn off wifi too May 11 18:42:15 Yeah prolly needs an input event to sleep properly May 11 18:42:29 mondoshawan[m]: odd, I get a better sleep than that May 11 18:42:46 Oho. Didn't know that. May 11 18:42:47 Have you tried running powertop on it.. might reveal what is being naughty May 11 18:42:55 with wifi off over a 12 hour period the power indicator goes down 3% or so May 11 18:42:56 Do you have wifi on? May 11 18:43:11 (with closed) May 11 18:43:28 mondoshawan[m], no proper kernel history, unfortunately May 11 18:44:17 but the sources themselves can be compiled, see this channel history or maybe https://raw.githubusercontent.com/postmarketOS/pmbootstrap/device-planet-geminipda/aports/device/linux-planet-geminipda/APKBUILD May 11 18:44:44 I'll try again later today. If it works, maybe I'll stick with Linux. Possibly with a running stumpwm if I can manage it. :-p May 11 18:45:38 yea we need to figure out why wifi keeps it awake so much May 11 18:45:46 Android is nice and all in that its a more refined experience, but I bleed Linux, so... :-) May 11 18:46:25 Probably a wakelock is my guess. May 11 18:46:27 not related, i just plugged in the hub/mouse to have access to a right mouse button in lxqt. i made the icons much bigger -- 48px vs. 22 in the taskbar/kicker thing, and it's now usable. there is an autohide checkbox which i did not check, because i feared losing it never to return. but it made me wonder: could it be tied to a press -release of the alt key (or if not, something else; just thought alt would be good May 11 18:46:28 because that performs the same function on the android side) so it would life the taskbar/app bar thing for 5 seconds or so? it would free up screen space but still then be useful. May 11 18:46:29 Are the wifi drivers OSS? May 11 18:46:54 (I'm assuming the firmware isnt which is why we only get 2.4GHz in Linux instead of 5GHz) May 11 18:48:41 it's the mt-wifi driver May 11 18:49:07 makes me wonder what is different on Linux though, since kernel part is the same between Linux/Android May 11 18:49:13 ^ May 11 18:49:57 (maybe they have some custom commands to wpa_supplicant? see https://gitlab.com/mt6797/vendor/tree/ALPS-MP-M0.MP9-V1.78.1_NB6797_6C_M/mediatek/proprietary/hardware/connectivity/wlan/wpa_supplicant_8_lib) May 11 18:52:50 it has things like set_band_5g inside May 11 18:54:06 mondoshawan[m]: but with wifi on, I still get the screen to stay off when I close it, just the power goes down much quicker than I'd expected May 11 18:55:20 @adam_b I'd be curious to see what your dmesg says. It mentions how long the device was actually asleep for. Mine reads as 0 every time. May 11 18:56:27 If I let mine sit long enough, the screen times out, but its not actually asleep. I can still manipulate things via ssh. May 11 18:57:01 yea when sleeping you can still ssh in, but thats sensible? May 11 18:57:02 Maybe that was the issue, actually. Incoming packets via wifi might have triggered a resume? May 11 18:57:22 I did notice lag on the SSH connection after I did the suspend at the CLI. May 11 18:57:50 Really? That means the SoC is still running in high power. I May 11 18:58:05 well it goes slower May 11 18:58:23 I'd expect it to be unresponsive. May 11 18:58:34 Screen off / blank != suspended May 11 18:58:38 yea its probably more responsive than it should be May 11 18:59:04 > Screen off / blank != suspended May 11 18:59:05 Yup.. that is the truth May 11 18:59:12 note that due to the lack of activity monitoring we only activate sleep when the gemian-lock is up May 11 18:59:37 So basically, what we're all saying is that suspend doesn't actually work currently. May 11 18:59:54 How does gemian-lock trigger kernel suspends? May 11 19:00:13 custom signals to say 'lock active', which then lets repowerd enable sleep May 11 19:00:37 it was a simple hack to get sleeping to work when closed May 11 19:01:02 mondoshawan[m], you can SSH when device is suspended, but incoming packets will wake the device up May 11 19:01:02 using xss-lock to launch the lock means it triggers on screen sleep too May 11 19:01:06 *can't May 11 19:02:14 (I don't strictly know if it's Wi-Fi packets or something else that does cause device to wake up, but on Android/Sailfish you can also SSH when device is "sleeping", but it will be unresponsive due to it constantly going to sleep/waking up) May 11 19:02:38 (here I meant other devices, not only Gemini PDA) May 11 19:03:11 this means that the setting of the screen lock is an essential part of the sleep story, also stops the stray half esc key press that comes through on resume May 11 19:04:02 ahh. Okay, so then repowerd is the key bit providing the dbus power interface, then. May 11 19:05:03 yea check the diffs on gemian repo if you want full details May 11 19:07:48 Yeah, will do! May 11 19:16:54 interesting: the little power-level app on the lxqt bottom tray provides no information, but the little circle on the tray seems to show how much battery is left. May 11 20:00:24 yep May 11 20:01:35 just found about Orange Pi 4G-IoT, which has low-tier MediaTek hardware, different SoC from Gemini, but might be still interesting to try some ideas May 11 20:02:17 the battery interface provided by mediatek is bunk, /sys/class/power_supply/battery is full of weirdness May 11 21:54:15 has anyone gotten the GPS hardware to work under linux? A long that line we should have a page on the wiki that lists what hardware is and is not working May 11 21:59:26 that would include the adaptive brightness sensor and the softwarw brightness control. screen brightness so close to the keyboard in part of why the keyboard is so hard to see. May 11 21:59:57 softwarw=software May 11 22:03:41 depscribe[m]: you can echo 20 > /sys/class/leds/lcd-backlight/brightness for now May 11 22:04:04 eventually xbacklight and such should work May 11 23:05:04 Re the battery indicator: it seems that the battery info is exposed properly in /sys/class/power, but not really sure where that is getting the detailed info. May 11 23:15:38 I don't know about *properly* May 11 23:22:46 https://www.kernel.org/doc/Documentation/power/power_supply_class.txt defines what devices should be doing May 11 23:26:32 it seems the "/sys/class/power_supply/battery/InstatVolt" file is the one to read to determine current battery voltage May 12 01:52:12 Anyone manage to get root working with android? May 12 02:41:24 apparently some people on the oesf forum did. **** ENDING LOGGING AT Sat May 12 03:00:03 2018