Oh I just remembered but there is like 3+ ways of getting the position depending on a bunch of stuff. You can use the fused location manager API, but you need google play services so it doesn't work on stock grapheneos (and I want it to work there).
Then finally you have battery optimization stuff that even when you tell it to not get restricted, it still sometimes offsets the callbacks by 2x, 3x the recurring time you entered, which becomes extremely painful to test since even 15 minutes is a lot to wait to have feedback.
AlarmManager is discouraged and iirc also deprecated for recurring tasks but maybe I should try it again because with WorkManager, you are limited to a minimum of 15 minutes recurring task, even if the user wants it more often.
Then there is the way to do background tasks. Tauri doesn't have any way yet to do stuff that require specific background APIs such as location, so I'm coding in kotlin. I either need to use AlarmManager or WorkManager if I don't want a persistent notification (because this will be running h24)
Then you have android as a whole which is really confusing with all the versions you need to understand to setup a project and all the files it needs for a very basic project, which I am ok with calling that a skill issue. It still adds mental burden tho.
So there are quite a few things. First is a bug with what I assume to be either tauri or vite which makes the file serving often fail, so you just have a blank page and a basic error. That bug also often makes me unable to use my wifi on my laptop (it kind of kills it it's weird)
No idea if I'll continue it, since for now at least, I don't have any fun making it. If someone with mobile/android knowledge does want to contribute let me know, maybe I'll come back earlier. For now at least, I don't want to learn an entire class worth of android stuff so I'll put this on pause.
I'm gonna honest, android has mostly killed my enthusiasm I had for privacypin.
I'm able to do the server, the UI and the protocol just fine, but anything that involves background and location stuff actually makes me stressed when working on it. Not to even mention all the vite/tauri issues with it
the engineers said I couldn't use LLMs to write a program that does nothing but waste energy. did it anyway. somehow the app has a worse utility-per-watt ratio than the model that wrote it.
Yeah that was pretty much what I was thinking with the limited knowledge I have. It could help a lot tho, although I'm not sure if you can use it in cases where you might want repudiation.
I'm still sitting on my 2026 resolutions of finishing my read of the atproto. I think this situation will make me read it earlier than I thought. I just need to put it in pdf so I can read it on my eink tablet, and I have something worth quite a few bedtime stories :)
I will try to over engineer the security tho (I must be stopped sometimes)
Started to think about a unified federated messaging app about 3 years ago. I learned quite some stuff from it but it was way out of my league so it's now on the very long term backburner, would love to be apart of a team that does this tho, communication system design is basically my passion :)
So I'm a bit at a loss currently. I don't want to give up a big part of what pp stands for, but this is still a feature that a lot of other apps implements.
Only other idea I have currently is just having apps the community makes with their own features, which would probably be pretty easy tbh
5/5
I'm thinking about having a toggle where you could either hide your positions from a specific group or just hide it from any maps that include other people. One of the potential drawbacks is being socially pressured to not use that toggle. This is the best solution I currently see...
4/5
That way, whether it is a toxic-ish relationship or a family with parents that want to know their kid's positions a bit too much, as long as they are willing to install the app, it should protect the privacy of everyone. Sometimes, it's just about limiting impulses.
3/5
So I thought about being able to create "groups". You can create a family group to limit accessing other's position, but you still have the same problem within groups.
The ultimate goal would be to provide a way where you can't really have an excuse apart from "oh I missclicked"
2/5
ok so I want to add a map where you can see everyone you have as friends (since most position sharing apps have that), but one of the goals of pp is being able to see when and who saw your position. With a global map, you can now have the excuse "oh i was just checking [other person], not you".
1/5
I have a feeling that it is the case because google maps uses apis that can't be used by normal apps. So in a way, they are kind of "cheating" :p
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
I'm working on it lmao
prod=production
So something else would be like a testing or development environment which is not exposed
no idea how long that will take with everything else I have
Things left before a PP MVP:
- send/get pings
- fix a bit of stuff and harden the app against user
nah I'm actually semi-serious. I'm that slow because I do edge cases and path explorations in my head to have somewhat good code.
I still need to use the remaining 4 APIs for PP to be considered somewhat of a MVP
I'm a 0.01x dev
It would be basically for the whole day, although I'll need to re-setup obs on bazzite with the settings streamplace needs
I'll probably stream on the 2nd some PP development. Might even throw in a state machine for the main protocol
@xan.lol I'll finish PP when you finish your website /j
