sparseMatrixI'm going to redo the machine anyway, it needs other work pretty badly. It wasn't really put together for this, so much as hijacked for it ;)
sknebelI found the docs for caddy v2 quite confusing last I tried to use it and went back to trusty nginx. but it is a viable choice with some interesting things
sknebelthe docs pushed the "we have an API! we are an API gateway!" angle too much for me at that point, whereas with nginx I knew where to copy stuff from :D
ShadowKyogre joined the channel; ShadowKyogre left the channel
petermolnaryou mean something like "a streaming solution for internet-scale webmentions support using microservices on a container based public cloud infrastructure backed by data warehouses" ?
jackyalcinemainly things like indieauth and incoming webmentions - things that I'd want to have a very strict structure around. elixir's way better for things like microsub
jackyalcinelemme know if there's things you'd want to see in that crate and I'll try to get to it - I'm optimizing it right now to have some shared functionality in a few projects
[chrisaldrich], [jacky], [fluffy] and [aciccarello] joined the channel
vikanezrimayajackyalcine: if you have an MF2 parser there that would be as HTTP-client-agnostic as possible, I'd totally add this crate to my Cargo.toml right away! I'm using surf as my HTTP client right now and I hope to possibly reuse connections by owning a handle to surf::Client instead of creating a new one for every single request (i'm afraid it might
vikanezrimayaoverhead too, and I need to keep the Micropub parts of my site blazing fast so I wouldn't feel a delay when posting! even though I'm currently sending webmentions from a background task)
[jacky]there was a bit of discussion re: a Rust MF2 parser (namely between someone else and I) - the issue was picking a flexible enough HTML parser tbh
vikanezrimayaI kinda hate JF2 but I understand that other people use it, so if you have any ideas or concepts, open an issue or a PR! I'm also considering disassembling one of the major MF2 parsers in other language to see how it works and possibly port it to Rust
vikanezrimayaThen probably keep an eye on that repo, it might start moving in the future when I run out of low-hanging fruit to implement in Kittybox-rs
vikanezrimayafor now it's mostly onboarding (which I'm trying to code probably for the third time already!) and indexing posts for ?q=source, ?q=geo, ?q=food and ?q=contacts
vikanezrimayais there a way to make, say, a fieldset <legend> only visible to screen-readers? it's in the way of my visual style but I'd like to leave it where it is for semantic purposes (since I use something like a progress-bar or breadcrumbs to track form progress and it's weird to have a section title be duplicated twice)
vikanezrimayaexcept the no JS part isn't gonna be relevant anyway! The endpoint for which I'm building the form requires JSON and it'll be painful to use form-encoded for it... form-encoded is generally a pain for me
vikanezrimayabasically I have something like Vec<{ slug: String, name: String }> in my form for creating feeds (mapped to Micropub channels - see extension) when onboarding Kittybox
vikanezrimayaand then my Micropub client (including the one that will ship with the software - for now it also serves as a debugging instrument) will read those and allow me to either auto-detect channels based on post type (h-entry, h-card etc.) or post to custom channels
[jacky]I think the most important thing is that I could use whichever Web server I want despite this being the business bits of it - if I can figure out how to extract some bits, I might even make the Micropub server logic into a library so anyone using Rust can just _get_ a Micropub server with a single import
reed, astrojl_matrix, forest[m], Paul[m]2, ytqb[m], smacko[m], VictoriaDrake[m], Ethan[m], drmr and edsu joined the channel
vikanezrimayaFor example, separating the normalize_mf2 function that makes MF2 posts have consistent properties, type-checks some of them (especially datetimes!) and inserts default properties when they aren't defined - namely, `uid`, `published`, `url`, `author` and `channel` - was the best decision in the entire development cycle
[girrodocus]I'm experimenting with a VPS for the first time (via digitalocean) and am wondering where to insert sorting out SSL certificates into this process: 1. Set up SSH keys. 2. Create the droplet (Ubuntu 20.04(LTS)x64). 3. Initial Ubuntu updates and set-up. 4. ufw firewall. 5. Install nginx
[girrodocus]I know I need to do these at some point: A: delegate domain from namecheap to digitalocean (done). B: Set up A and AAAA records pointing example.com and www.example.com to the relevant IP addresses (done but not working yet). C: Do something that I don't understand yet with CAA records, LetsEncrypt and cerbot (the mysterious aspect).
[girrodocus]I think I've accidentally registered for the LetsEncrypt thing about 4 times in the last couple days while I've tried to set up digitalocean droplets. I'm hoping this attempt will survive the getting-deleted-in-frustration test.
ShadowKyogre, Caleb[m]1, astralbijection[, jamietanna[m], JackyAlcin[m], Salt[m], dianoetic[m], smacko[m], VictoriaDrake[m], ytqb[m], astrojl_matrix, forest[m], Paul[m]2, drmr, edsu, reed, Ethan[m], KartikPrabhu and [dianoetic_net] joined the channel; ShadowKyogre left the channel
[dianoetic_net]girrodocus, if you want a self-signed cert with LetsEncrypt it should come down to running some shell commands. Have you seen this? https://letsencrypt.org/getting-started/