| * inky got mz617 in summer in hope to use it with maemo. it lives in a drawer waiting for better times. | 02:07 | |
| Wizzup | ~ | 07:38 |
|---|---|---|
| freemangordon | hmm? | 07:38 |
| Wizzup | inky: OBlag :) | 07:38 |
| Wizzup | freemangordon: lag | 07:39 |
| Wizzup | lol | 07:39 |
| freemangordon | heh | 07:39 |
| Wizzup | inky: soon enough | 07:39 |
| Wizzup | I am using mz617 now with qalendar and apart from qalendar sometimes crashing it's quite nice! | 07:39 |
| Wizzup | fun to type notes and tasks using the vkb only in qt | 07:39 |
| freemangordon | Wizzup: hmm, not sure what to do | 08:22 |
| freemangordon | seems xorg has issues with drm devices order | 08:22 |
| Wizzup | freemangordon: this being multiple 'video cards' in drm or something? | 08:40 |
| freemangordon | yes | 09:14 |
| freemangordon | like we have on omap | 09:14 |
| freemangordon | for example modesetting will use card0 by default, despite it is for SGX | 09:15 |
| freemangordon | I am just moaning, not that we can do anything about that | 09:15 |
| freemangordon | I'll just fix omap driver to return false from Probe function if there is platform probe support in xorg | 09:16 |
| freemangordon | not sure that's the proper way, but meh | 09:16 |
| Wizzup | freemangordon: we might be able to set te card in xorg.conf | 11:22 |
| Wizzup | if they are predictable (usually no0t | 11:22 |
| freemangordon | Wizzup: no, I just disable driver probe if platform probe is supported | 13:37 |
| freemangordon | (whatever those are :) ) | 13:37 |
| freemangordon | Wizzup: ok, got it working with daedalus Xorg, will take a rest and will push the changes | 14:15 |
| freemangordon | btw, do we have fulscreen QT application to test with, if it still crashes SGX? | 14:15 |
| Wizzup | freemangordon: I think we can add that to jib | 14:29 |
| Wizzup | (again) | 14:29 |
| Wizzup | freemangordon: great! | 14:29 |
| freemangordon | Wizzup: well, but I want to test before we put anythong in the repo | 14:32 |
| freemangordon | *anything | 14:32 |
| freemangordon | maybe smplayer | 14:32 |
| Wizzup | mpv? | 14:32 |
| Wizzup | freemangordon: ok, but nobody uses repo yet | 14:33 |
| Wizzup | :) | 14:33 |
| Wizzup | it is just us | 14:33 |
| Wizzup | and only you have a d4 img | 14:33 |
| freemangordon | ok, if you can do it | 14:39 |
| freemangordon | fort jib that is | 14:39 |
| freemangordon | *for | 14:39 |
| freemangordon | anyway, bbl | 14:39 |
| dsc_ | yes jib crashed previously | 14:46 |
| dsc_ | @fullscreen | 14:46 |
| Wizzup | freemangordon: sure, when you push the changes I will test :D | 14:48 |
| arno11 | note that it is not affecting only qt apparently, pcsx rearmed crashes in fullscreen with opengles rendering too | 16:04 |
| arno11 | works perfectly fine in window mode even with scale x2 | 16:05 |
| freemangordon | Wizzup: BTW, we have to change xorg service because of the changed log fule location | 18:12 |
| freemangordon | *file | 18:12 |
| freemangordon | Wizzup: ok, with xorg/ddx from repos d4 boots to h-d | 18:20 |
| freemangordon | hmm, and it crashed :( | 18:22 |
| freemangordon | hmm, maybe low battery | 18:23 |
| freemangordon | no, it really crashes | 18:26 |
| freemangordon | what the? | 18:26 |
| freemangordon | hmm, maybe mce shuts-down because battery is not calibarated | 18:36 |
| freemangordon | yeah, ok, does not reset with charger connected | 18:39 |
| sicelo | :-) | 18:40 |
| sicelo | stock upower? | 18:40 |
| freemangordon | yes | 18:40 |
| freemangordon | hmm, wait | 18:41 |
| freemangordon | I remember we changed the config | 18:41 |
| freemangordon | sicelo: do you know where config is? | 18:41 |
| sicelo | while you debug, i'm interested in your `/sys/class/power_supply/...battery/uevent` | 18:41 |
| sicelo | freemangordon: config for? | 18:41 |
| freemangordon | upower | 18:41 |
| freemangordon | or, was it for logind? | 18:42 |
| sicelo | /etc/UPower.conf ... but UPower doesn't really have much options there | 18:42 |
| freemangordon | sicelo: uevent is from 6.6, what do you expect to see? | 18:42 |
| freemangordon | Linux devuan-droid4 6.6.53 #1 SMP PREEMPT Sat Nov 30 17:40:52 UTC 2024 armv7l GNU/Linux | 18:43 |
| sicelo | just what an uncalibrated uevent looks like. want to keep it for my notes, and perhaps future upower work | 18:43 |
| freemangordon | ah, ok | 18:43 |
| freemangordon | I guess I will have to stop upower, roght? | 18:43 |
| sicelo | i can get it from my D4 as well though, later on | 18:43 |
| freemangordon | hmm, seems I cannot stop it | 18:44 |
| freemangordon | sicelo: https://pastebin.com/xzUSnxiJ | 18:45 |
| freemangordon | charger/no charger | 18:45 |
| freemangordon | I wonder why upower ignores POWER_SUPPLY_CAPACITY_LEVEL=High | 18:45 |
| sicelo | when upower decides that shutdown should happen, it sends a request to logind for that. so a quick hack would be to request a shutdown inhibit lock through logind | 18:45 |
| freemangordon | but we alredy do | 18:45 |
| freemangordon | *already | 18:45 |
| freemangordon | or, I thought so | 18:45 |
| sicelo | freemangordon: they say CAPACITY_LEVEL only matters for peripherals (they're wrong, and wouldn't budge) | 18:46 |
| freemangordon | bbiab | 18:46 |
| sicelo | but, upower maintainership seems to have changed since then, so we could have a go at this conversation with them | 18:46 |
| Wizzup | we shouldhave logind ignore all of this | 18:54 |
| Wizzup | the key names changed perhaps | 18:55 |
| freemangordon | Wizzup: seems not | 18:55 |
| freemangordon | bbl, dinner | 18:55 |
| Wizzup | they changed on my gentoo laptop recently | 18:55 |
| Wizzup | ttyl | 18:55 |
| sicelo | i'm also not in a position to check atm, but under normal circumstances (upower aside), i think we do want logind to respond to shutdown requests. e.g. users should be able to `loginctl shutdown` (which is basically the same request upower makes). leste currently works fine with this | 18:58 |
| sicelo | what i would need to confirm is - who actually makes the request to shut down the system (Leste) atm - is it MCE, after receiving appropriate signal from upower, or the shutdown request goes directly from upower to logind | 19:00 |
| sicelo | most stacks simply let logind (and upower) handle this stuff on their own. in our case, afaiui, we want mce to control system shutdown, etc., so it would seem mce needs to start holding inhibitor lock against shutdown operation | 19:05 |
| sicelo | https://systemd.io/INHIBITOR_LOCKS/ | 19:05 |
| sicelo | i think mce qualifies for the 'blocking' lock type in that documentation | 19:08 |
| freemangordon | sicelo: mce asks dsme which asks logind :D | 19:08 |
| freemangordon | it is not that simple | 19:08 |
| freemangordon | it is not only mce that can request shutdwon | 19:08 |
| freemangordon | (iirc) | 19:08 |
| sicelo | besides the user, who else can request it? | 19:08 |
| sicelo | anyhow, since upower 1.90.5 (https://gitlab.freedesktop.org/upower/upower/-/releases/v1.90.5), upower can be told to do nothing when it thinks shutdown is needed. maybe this would suit our needs | 19:14 |
| sicelo | i use that config under pmos with an udev rule that initiates shutdown when kernel sends uevent to say CAPACITY_LEVEL = Low/Critical, thus completely bypassing upower for making shutdown decisions | 19:17 |
| freemangordon | cool | 19:18 |
| freemangordon | hmm | 19:18 |
| freemangordon | upower 0.99.20-2 | 19:18 |
| freemangordon | won't fly | 19:19 |
| freemangordon | maybe we shal jump to upstream upoerw anyways | 19:19 |
| sicelo | what glib2 is in daedalus? this upower needs at least 2.66 | 19:20 |
| sicelo | https://gitlab.freedesktop.org/upower/upower/-/releases/v1.90.3 | 19:20 |
| freemangordon | libglib2.0-0:armhf 2.74.6-2+deb12u5 | 19:22 |
| sicelo | seems like this would be best option then :-) | 19:22 |
| freemangordon | hmm, latest is Released: 2025-01-09 | 19:22 |
| freemangordon | lemme pull it | 19:23 |
| halftux_ | Wizzup: My rpi3 was broken got in early stage ovserial console PMIC: timeout reading reg 00 (error -1) | 19:25 |
| freemangordon | sicelo: ok, upstream upower is entangled with systemd :( | 19:31 |
| freemangordon | I wonder if we shouldn't move to systemd | 19:32 |
| sicelo | no, you can run it just fine without systemd. just need the correct ./configure options | 19:34 |
| freemangordon | ok, but won;t' do it now | 19:35 |
| sicelo | haven't compiled it in a while, but iirc something like `-D systemdunitdir=/no` | 19:35 |
| sicelo | it | 19:35 |
| sicelo | it's "-Dsystemdsystemunitdir=no" | 19:36 |
| freemangordon | ok, lemme check | 19:37 |
| freemangordon | in the meanwhile chromium crashes :( | 19:40 |
| freemangordon | https://pastebin.com/Hr6A3Mrs | 19:40 |
| freemangordon | sicelo: ok, upgraded upower, lets see if system wil boot at all :) | 19:43 |
| freemangordon | ok, still boots | 19:46 |
| sicelo | so you'll need to mod /etc/UPower.conf as per https://gitlab.postmarketos.org/postmarketOS/pmaports/-/blob/5f059f959e53395c7c9f3d15e5caa98f8f8f56cc/device/community/device-nokia-n900/upower.conf#L84-L104 | 19:49 |
| freemangordon | not now | 19:49 |
| freemangordon | :) | 19:49 |
| sicelo | then upower will never try to shutdown the system. obviously, that needs the shutting down decision to move somewhere else. i went with udev, but we can do it in mce perhaps | 19:50 |
| sicelo | sure, no rush :-) | 19:50 |
| freemangordon | ugh, will have to build mesa on device :( | 20:48 |
| sicelo | i wonder how long that'd take on N900 ... a week? :-D | 20:53 |
| Wizzup | freemangordon: why not on the armhf vm? | 20:54 |
| freemangordon | I'll directly build in ci | 20:55 |
| freemangordon | the bug is abvious | 20:55 |
| freemangordon | *obvious | 20:55 |
| Wizzup | ok | 20:55 |
| freemangordon | Wizzup: BTW, we will have to set DPI | 21:50 |
| freemangordon | RN everything is tiny on d4 :) | 21:50 |
| Wizzup | that's new | 22:02 |
| freemangordon | I think we were doing something about DPI? | 22:03 |
| Wizzup | don't remember, now in leste-config I don't think | 22:03 |
| Wizzup | X just reads dpi from screen | 22:03 |
| Wizzup | leste-config-common/etc/profile.d/gtk-maemo.sh.leste:export GDK_DPI_SCALE=1.7 | 22:03 |
| Wizzup | there is this but I think this is gtk3 | 22:03 |
| Wizzup | freemangordon: should I build a new img, or do you want me to wait? | 22:07 |
| freemangordon | I think you can | 22:07 |
| freemangordon | ugh: | 22:08 |
| freemangordon | dimensions: 960x540 pixels (203x114 millimeters) | 22:08 |
| freemangordon | hmm, wait | 22:08 |
| freemangordon | what should be the dpi? | 22:09 |
| freemangordon | dimensions: 960x540 pixels (254x142 millimeters) | 22:09 |
| freemangordon | 254x142, eh? | 22:10 |
| Wizzup | that is the screen size | 22:12 |
| freemangordon | the real dpi is 270, IIUC | 22:12 |
| freemangordon | not 96 | 22:12 |
| Wizzup | 270 sounds about right | 22:13 |
| Wizzup | on chimaera X says DPI set to (96, 96) | 22:14 |
| Wizzup | nothing in xrdb | 22:14 |
| freemangordon | yes, but here it seems dpi is taken into account | 22:15 |
| Wizzup | by who? gtk? h-d? | 22:16 |
| freemangordon | xorg | 22:17 |
| freemangordon | *everything* is tiny | 22:17 |
| Wizzup | looks like xorg changes more than we think :) | 22:19 |
| Wizzup | 254mm = 2.5cm - that seems wrong, no? | 22:19 |
| Wizzup | wait, 25cm | 22:19 |
| Wizzup | that's also wron | 22:20 |
| freemangordon | yes, it is | 22:20 |
| freemangordon | but that's reported by xdpyinfo | 22:20 |
| Wizzup | it this in dts? | 22:20 |
| Wizzup | is this* | 22:20 |
| freemangordon | that means dpi 96 | 22:20 |
| freemangordon | I think it is nowhere | 22:20 |
| freemangordon | and xorg assumes dpi of 96 by default | 22:21 |
| freemangordon | and then simply multiplies the resolution | 22:21 |
| freemangordon | a wild guess | 22:21 |
| Wizzup | you can set it with xrandr --dpi I think | 22:21 |
| freemangordon | no efffect | 22:21 |
| Wizzup | or when starting X, or in config, or setting in dts I guess | 22:21 |
| freemangordon | config? | 22:21 |
| Wizzup | xorg.conf | 22:23 |
| freemangordon | ok, but how? | 22:23 |
| freemangordon | hmm | 22:25 |
| Wizzup | By default, Xorg always sets DPI to 96 since 2009-01-30. A change was made with version 21.1 to provide proper DPI auto-detection, but reverted. | 22:25 |
| Wizzup | https://wiki.archlinux.org/title/Xorg | 22:25 |
| Wizzup | https://wiki.archlinux.org/title/Xorg#Display_size_and_DPI | 22:25 |
| Wizzup | you will have to restart programs after changing dpi with xrandr | 22:26 |
| freemangordon | ok, lemme try | 22:26 |
| Wizzup | but, with a monitor section you can set real display size | 22:26 |
| Wizzup | in xorg.conf.d | 22:26 |
| freemangordon | does not help much | 22:27 |
| freemangordon | only fonts in launcher are increased | 22:28 |
| freemangordon | yeah, fonts are huge now | 22:28 |
| Wizzup | after restarting X? | 22:28 |
| freemangordon | yes | 22:28 |
| Wizzup | with monitor in xorg.conf.d? | 22:28 |
| Wizzup | what about xrdb -query | grep -i dpi | 22:28 |
| freemangordon | h-d icon sizes are the same | 22:28 |
| Wizzup | https://wiki.archlinux.org/title/Xorg#Manual_DPI_Setting_Caveat | 22:29 |
| freemangordon | I need microscope for chromium, no matter what I do :) | 22:30 |
| freemangordon | ok, lemme compare with chimaera | 22:32 |
| Wizzup | what is it in daedalus? | 22:32 |
| Wizzup | what did you write in xorg.conf.d? | 22:32 |
| freemangordon | ok, h-d is ok | 22:34 |
| freemangordon | h-d in daedalus and chimaera look the same | 22:34 |
| freemangordon | but, chromium at least is tiny | 22:34 |
| freemangordon | GDK_DPI_SCALE seems to be ignored | 22:35 |
| freemangordon | --force-device-scale-factor works | 22:37 |
| freemangordon | Wizzup: is it possible that I am just missing gtk3 themes? | 22:42 |
| freemangordon | ok, the same issue in the VM | 22:44 |
| freemangordon | so not D4 related | 22:44 |
| Wizzup | well is anything you are using gtk3? | 22:44 |
| freemangordon | chromium is | 22:45 |
| freemangordon | more or less | 22:45 |
| Wizzup | ok, so ignoring chromium, what else is | 22:45 |
| freemangordon | nothing I guess | 22:45 |
| Wizzup | I'm still not clear what qrdb prints for you, or what xorg.conf.d you set | 22:45 |
| Wizzup | xrdb | 22:45 |
| freemangordon | but, GDK_SCALE=2 works with chromium | 22:45 |
| Wizzup | well, if it's just chromium, maybe we can live with that for today | 22:46 |
| Wizzup | leste-config-common/etc/profile.d/gtk-maemo.sh.leste:export GDK_DPI_SCALE=1.7 | 22:46 |
| freemangordon | xrdb -query | 22:46 |
| freemangordon | *customization:-color | 22:46 |
| Wizzup | is this not present for you? | 22:46 |
| freemangordon | it is | 22:46 |
| freemangordon | but is ignored | 22:46 |
| Wizzup | ok | 22:46 |
| freemangordon | GDK_SCALE works | 22:46 |
| Wizzup | I see | 22:46 |
| freemangordon | https://bugs.kde.org/show_bug.cgi?id=389523 | 22:47 |
| Wizzup | and this is X app,not some wayland stuff right? | 22:47 |
| freemangordon | yes | 22:47 |
| Wizzup | ok | 22:48 |
| freemangordon | oh, wait | 22:48 |
| freemangordon | lemme test something | 22:48 |
| Wizzup | but other than this, if I build an image now, will h-d and icons and osso-xterm look ok? | 22:48 |
| freemangordon | yes | 22:48 |
| Wizzup | ok | 22:48 |
| Wizzup | I will do that tomorrow then | 22:48 |
| freemangordon | ok | 22:48 |
| Wizzup | I got up at 4am so it's bed time for me :) | 22:48 |
| freemangordon | ugh | 22:49 |
| freemangordon | night! | 22:49 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!