Lit is the best Web Components library out there. It's simple and web-friendly. I spent a few days arguing that it should provide a CDN-friendly static build. I failed. The project devs firmly believe that they are saving the ecosystem by preventing people potentially using the wrong version even if it means less people are using the library. Pointing out that this is not an issue at all and that every single library out there does this wasn't enough.
I wasn't sure if I should roll my own pre-compiled version. It sounded like an obvious thing to do. It's not a fork, it's basically making a project that simply exports Lit. The reason is not done more often is because every other library out there provide a
dist/. That said, I didn't want to come across as being aggressive towards the core developers of Lit.
In then end, I had something already working for my projects, so I took the opportunity to learn more about Github Actions. I built lit-dist, a project that abuses Github Action to track the official Lit repo and automatically generate pre-compiled versions of the library. Reception so far has been good.
This week I also kinda unhappy with my choice of self-hosted RSS Reader. My criteria are simple, and yet very hard to achieve: I want a self-hosted reader that is well designed, has reasonable key bindings, and that uses the only reasonable database to use for a single-user project: SQLite.
Last time I looked for it I ended up with only one option. Now there are 3. I picked the nicest one: Yarr. The main issue with Yarr is that the developer decided that it's feature complete and it won't change a thing. I can respect that in a project, but this meant that its weird corners won't be fixed. So I forked it and improved it the way it made sense to me. So far, I'm very happy with it.