Long posts

I love having all these parts in my head.

I'm working on a major overhaul of the API and front end.

• file improvements: processing new derivatives, Rust backend, and higher throughput/bulk processing. Faster int'l time-to-upload with an AWS Lambda on the front, then processing later, localized.
• break out work that should be on background jobs: post, user, file work, fan-out processes, when big things happen.
• separate pnut.io frontend from backend: server client with internal API.
• modernize code and data design: better async/await handling, atomize what should be guaranteed, error handling, vastly improved cache handling.

Some of the work is in separate projects.

• "Cabinet" is the AWS Lambda + local file processing system.
• "Inferno" is the server-side web client and internal API for it.
• "Hare" is the job processing system, which has been in use for a long time now handling websocket messages (not the websockets themselves, but figuring out filtering and rendering data). But I added a whole wing of operations to it for background jobs. Used to just be nofitication-related.

Other pieces that are not being worked on until these big changes are stabilized:

• "Socks", new websocket system.
• "Farm", RSS system.
• "Heylisten", audio file processor, is being integrated into "Cabinet".