That’s the intent, at least:

Are you intending to ship a (close to) mainline kernel, or a Board Support Package (BSP)/vendor kernel and make it work with a libhybris/Halium approach?

We’ll go with bare-metal Linux—no Halium, no libhybris. We want to stay as close to mainline as possible and actively contribute upstream.

(I haven’t really used them a lot in the heat yet) Last enclosure was ASA, but AFAIK, black ABS is OK too because black pigment absorbs most of the light/UV, preventing plastic from degrading as fast

For offline navigation on Linux, have you looked at osmin? It was pretty decent on a PinePhone.

How do you handle power-off? Does Raspberry Pi just shut down? My thoughts were to use Alpine or some RAM-based OS that would not corrupt SD card or the hard drive.

I have been messing around with building an in-car navigation from e-waste for a while now. Right now, I settled on an old smartphone with OsmAnd and wrote my own app to view the reverse camera.

Slighty altered:

when prosecutors reportedly alleged the murderer used Google Maps to locate and visit places where handguns were known to be sold.

Sounds kinda dumb, doesn’t it? Once again, a lesser known app gets a portion of the blame because of shitty people.

P.S. I use iNaturalist and greatly appreciate all the people who correct my identifications.

.NET applications using .NET Core or later are intended to be cross-platform, so technically, Linux can run .NET apps. (The use-case I know is running .NET sites on Linux servers)

I don’t find it difficult, and have enjoyed sxmo on the PinePhone. I understand the suckless approach, but I do have to admit that many people that I know, even tech-savvy ones, probably would not want to rebuild to configure something.

To configure most suckless tools you need to… recompile them. The readme says:

Because dwm is customized through editing its source code, it’s pointless to make binary packages of it. This keeps its userbase small and elitist. No novices asking stupid questions.

But if you are trying to compile suckless tools, you are already in too deep.

I just installed Miniflux on my server as well.

Advantages (in my opinion) are: Package is in Debian repos (safe and no compilation needed), software is a static binary (thus does not require docker and only needs postgreSQL), documentation is good.

This tool looks fantastic, thank you!

And Snikket for super-easy setup and management

Synapse has seemingly improved since 2020. A word of warning though: if you join large rooms from your server, Synapse will eventually grow the DB to a huge size due to a “lookup” table state_groups_state, and will require manual cleanup. See https://www.sequentialread.com/matrix-synapse-out-of-disk-space-state_groups_state/

Xed editor comes close in terms of handling files-as-files, but I found it more cumbersome and buggy than Markor.

When I ran prosody a few years ago, I did so without docker.

I did try snikket in docker though, and it looks like it is still actively maintained.

IIRC the Windows version of Midori was the only browser that was light enough to watch Netflix on my ~2005 laptop.

Reminds me of a “minimalist text editor” that my coworker showed me circa 2015. It was an Electron app that consumed more RAM to display a empty file than Firefox with 5 active tabs.

Mailbox.org

Mailbox Standard compared to ProtonMail Plus:

• Cheaper (€30/yr vs ~€50/yr; if you don’t need custom domains, €1/mo)
• More aliases (25 on mailbox, 50 on own domain. Proton has 10 TOTAL - why custom domain aliases are counted against Proton ones does not make sense to me.)
• Support for any number of custom domains AFAICT (Proton Plus supports only one)
• Trial account is not allowed to send emails, so fewer issues with services blacklisting proton.me and protonmail.com for spam (hasn’t happened to me, but I have heard of some cases)
• Can use a regular email client (security tradeoff for E2EE messages - but there already were plenty of discussions on whether E2EE has benefits, especially sending mail to other services)

TBH I don’t even understand why Android Auto needs to exist in the first place.

The same (or even better) functionality can be achieved by using a standard video output (DisplayPort, HDMI) from the device to the in-car screen, while the touches on the in-car screen can be translated into USB mouse position and clicks on the device (unless there is a better touch protocol).

I know there are regulations about live video on in-car screens, but 1. That does not stop people from watching videos on their phones while driving and 2. Somehow that does not apply to maps?

Tried it on an older Pixel with GrapheneOS.

Can’t zoom, can’t switch to wide-angle lens. Camera does not balance brightness by the focus point. Otherwise pictures look pretty much the same.

I suppose this was made for specific devices in mind?

Proxies started getting blocked (by some auth-account methods) last year, but libreddit/redlib dev was able to outsmart them multiple times. Now it seems like reddit is blocking IPs (and/or IP ranges). Running redlib from a residential IP still works, but I would not expect it to work from a VPS.

Tried it out a while ago, and found that I prefer GNOME’s UX and configurable shortcuts better, and that two side-by-side applications on my laptop is the most “tiles” I would realistically want.