mojo VS JSON-to-Go

Initial thoughts

If you end up doing web development, check out Mojolicious:

https://mojolicious.org/

If you choose to go down the Mojolicious road, there's lots of deployment information and guides in the Mojolicious Cookbook.

I guess this will make it harder to search for Mojo(licious)-related stuff. 😩

But if I want the return value of 'test_p' or the error message 'This is an error', I can't seem to figure that out. I tried looking at the promise tests (https://github.com/mojolicious/mojo/blob/main/t/mojo/promise.t) but that didn't seem to work either.

Several! The 3 big players in order of release are Catalyst, (released in 2005), Dancer2 (Dancer was first released in 2009, but went through a complete re-write as Dancer2 around 2013), and Mojolicious (released in 2010).

This project sounds to me like the perfect excuse to learn Mojolicious if you're interested in converting your scripts into a web application using Perl.

Awesome! I still use Perl on a pretty regular basis both for work and fun. I really enjoy it. Definitely take a look at Mojolicious if you haven't already. It's primarily focused on being a web framework (both server and client), but it's nicely modular so you can use bits and pieces of the stack. In i3mojo, I used the Mojo::IOLoop event loop, Mojo::Base as a base class system, and Mojo::UserAgent as a web client for some plugins.

Last time I used Perl for anything web it was via https://mojolicious.org/

It even does event-based and websockets

That site likely uses Matt's lib for that functionality: https://github.com/mholt/json-to-go

https://mholt.github.io/json-to-go/ is a nice tool for generating a struct definition to fit some sample JSON.

json-to-go is your friend.

JSON to struct. I use this about 100 times a day, no joke.

https://mholt.github.io/json-to-go/

Do your best to get static types on everything. Play a game with yourself where interface{} is banned and try to do your work. One of the most common sources of such loosely-typed values in JSON, so I'll point out this automatic struct generator for instance. If you have questions about a specific place you don't see how to avoid interface, feel free to ask. Every once in a while I do have to reach for something like this, but it should be rare.

But given we know ChatGPT can't accurately generate structs from even moderately complex json objects why bother doing all of that legwork of verifying and chiding chatgpt, when you can just drop it into a tool like https://mholt.github.io/json-to-go/ and get a correct struct in a fraction of the time?

I use https://marketplace.visualstudio.com/items?itemName=maracko.json-to-go, which is an extension version of https://mholt.github.io/json-to-go/

I think using struct in such cases is almost always the right approach. There are tools which generates go structs from json this.