cgscript VS JSON-to-Go

The thing that annoys me about golang is startup time. Sure, it's fast, but when I write my scripts `go run` compiles the script for half a second just for it to run for literally 7ms. It annoyed me, and I've created cgscript which caches the compiled executable that `go build` produces so that you won't have to recompile your scripts every time you run them.

I love using go for scripting, but `go run` builds my script for like half a second just for it to run for literally 7ms every time. So I've created cgscript which caches go executables so that go doesn't have to recompile them again and again. Sorry for the shameless plug.

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.