←back to thread

439 points david927 | 1 comments | | HN request time: 0.21s | source

What are you working on? Any new ideas which you're thinking about?
Show context
grep_name ◴[] No.44428127[source]
The stuff I'm working on never feels worth sharing, but I am doing a lot of computer stuff lately. It's kind of the year of moving towards declarative setups for me.

- Migrating to Niri on my laptop and re-evaluating my literate config approach, switching from xkb configs to kanata and a few other QOL changes to make my tooling more composable and expressive

- Shoring up my blog / media sharing infrastructure (migrated to a landing page on an s3 bucket, with different prefixes for several different hugo deployments for different purposes, still need to get better about actually posting content)

- Preparing to migrate a bunch of my self-hosted services to a k8s cluster which can can be fully deployed locally for testing and defined in code. All this is managed through argo and testable with localstack and crossplane for some non-local resources

- Attempting (somewhat unsuccessfully) to setup a nixos config for a bunch of services that just don't feel right to run in containerized stack that I want to live in ec2 and have as close to 100% uptime as possible (uptime kuma, soju/weechat relay/bitlbee, conduit, radicale, agate, whatever else I think of that is small and has a built-in nixOS service module. Thinking about some kind of RSS aggregating solution here as well)

- Experimenting with vibecoding by trying to get an LLM to do the legwork to build a TUI interface to ynab using rust (which I don't know how to write)

I'm hoping that by the end of this summer most of the tooling I use for most things will be way more concrete and seamless. I also want to get my workflows down and get on top of converting at least a few the ~100 draft blog posts I have laying around into something I can actually post. Ditto for my photography albums, which are not yet organized into coherent groupings or exported for web.

replies(1): >>44428226 #
andrewrn ◴[] No.44428226[source]
Building a TUI for ynab is pretty compelling, so I wouldn't say that's not worth sharing. If you could pull the data on terminal startup and get a budget snapshot each time you open a terminal-- cool.
replies(1): >>44428372 #
1. grep_name ◴[] No.44428372[source]
So far that experiment is going pretty well! I haven't worked much on it but the tooling I'm using has made a great base for the project. My goals (in order of priority) are:

- Get it so that you can categorize transactions quickly in a keyboard-driven way

- Similarly have a quick, improved option for dealing with overspending / underfunding

- Add some additional reporting that I'd like to see (as well as the ability to drill down in a more fuzzy way than currently supported in ynab)

- Finally (and most importantly but also most ambitiously) develop a view with some simple tools that helps users figure out WTF is wrong when a reconciliation isn't working out. This is much harder than the other things I'm trying to do here

Luckily YNAB's API is very open and I think I can do all the things I'm looking to achieve here. If I'm successful, I plan to spin off a sister TUI project for making handling import edgecases easier in beancount, which I also use but for different reasons

Edit: but your idea of having CLI command options for printing reports on a regular basis / on opening the shell is also neat, I do plan to have some CLI options that don't require you to open the full TUI