First time I've heard audio on the #fairphone5 running #nixos, thanks to #postmarketos research in https://gitlab.postmarketos.org/postmarketOS/pmaports/-/issues/3793#note_510714, it will soon™ Just Work™ in a mainline kernel near you!

#MobileLinux #LinuxMobile #Fairphone5 #Fairphone #weeknotes

@matthewcroughan Nice.

Comletely unrelated but I wonder why #phosh's icons leak beneath the rounded corners? Below image is how it should look on FP5.

Maybe the corner radii we have in gmobile are off? I don't have hardware to verify myself.

Because I'm running NixOS and have done nothing to supply phosh with the gmobile bits, since I'm just bringing things up and haven't gotten to the point of packaging gmobile yet

@matthewcroughan The build pulls in a subproject wrap if not found as shared lib. That is an old version though which would explain why some things are off.

What is a "subproject wrap"? Sounds tasty

@matthewcroughan Like https://gitlab.gnome.org/World/Phosh/phosh/-/blob/main/subprojects/gmobile.wrap?ref_type=heads . If you'd be working on OP6T we could make it a fajita.wrap.

I don't know what I'm supposed to do with this .wrap file, where I'm supposed to put it, etc, but I was going to figure that out some other time.

@matthewcroughan@defenestrate. It's used to fetch e.g. gmobile from git if not found on the system so nothing to do here. Just enjoy the wrap 🌯 .

Oh yeah I would never include such a thing in my system that would randomly fetch and compile or shift code/config around on the system at runtime, not for me. Nix builds configuration/code and puts them into place statically, and it's how everything in the system works top to bottom with no exceptions. If I found behavior like this in the system I would likely patch it out.

@matthewcroughan Well, you get gmobile from somewhere. Otherwise it wouldn't build.

Yeah looks like it https://github.com/NixOS/nixpkgs/blob/nixos-unstable/pkgs/applications/window-managers/phosh/default.nix#L102

So how do I configure Phosh to use the cutout for this device? Is that bit documented anywhere?

@matthewcroughan The library has documentation and then there's https://phosh.mobi/posts/notch-support/

There's nothing to configure per se as everything is picked up automatically based on DT compatible if the necessary information is available in the libs data (which is the case for FP5, that's why I asked 😃).

If I read this right, it should be doing it based on my device-tree's model right? So shouldn't this be working automatically since I'm using https://github.com/torvalds/linux/blob/master/arch/arm64/boot/dts/qcom/qcm6490-fairphone-fp5.dts#L28 ?

@matthewcroughan Yes, that's why I asked the initial question: why are the symbols leaking under the corner on your device? Maybe the border radius that we have is wrong.

@matthewcroughan That was ... mean. I hope you get copyright strike for this, or something. :-).

Anyway, good to see one more phone getting usable.

Ah must be, I can just patch phosh easily if you want me to test anything

@matthewcroughan Thanks for the offer. As I don't have a device it would be hit and miss and the value wouldn't be accurate after all., Let's wait until someone with a device bothers enough to have the resolved on FP5 too.

(Just in case someone comes across this: changed notches / corner radii can be tested without rebuilding due to `G_RESOURCE_OVERLAYS`: https://phosh.mobi/posts/notch-support/#looking-at-existing-device-information)