kutil VS goja

For reference, here's my implementation. Note that my wrapper supports both structured and traditional logging APIs.

For JS in Go use goja, which offers excellent integration with Go. Also I maintain an extension that adds Common-JS compatibility to it.

I think your instinct is right. For all my projects I use these wrappers. If you have any feedback or contributions they would be much appreciated. I know it's a losing battle, but there's no choice but to fight. Semper fi.

This is mostly what I did with my kutil logging library (sorry, no good documentation on how to use it yet). So I can use the same API in all my code and then pick various backends per project in order to better unify the log. For example, in Kubernetes I can use klog as the backend. I also wrote my own "simple" backend that does everything that I personally want. I'm sure there are other similar libraries out there.

I really like the idea of an Extended JSON, so I created my own version called "CJSON". The code is pretty straightforward.

dop251/goja

Your comment was posted 4 minutes ago. That means you still have enough time to edit your comment to change it so it contains real URLs:

<https://github.com/PuerkitoBio/goquery>

<https://github.com/dop251/goja>

(Please do not reply to this comment—I won't be able to delete it once the previous post is fixed if it contains replies.)

Apparently "not depending on Node or V8" means depending on some random Go JS engine instead.

Goja https://github.com/dop251/goja

Either call a JavaScript interpreter like node with exec.Command and read its stdout, or use a pure Go JavaScript interpreter like goja or otto.

Just to also say this is implemented in pure Go we aren't including V8 or any external dependencies we instead use https://github.com/dop251/goja which is a JS VM written completely in Go.

If you want to serialize a function to JSON one idea may be to embed a scripting language like JavaScript into your program. The goja package is a very good solution: a native ES5 JavaScript (with some ES6 syntax support as well) natively implemented in Go so you can get tight data bindings to your Go types and funcs. For your JSON marshalling you could serialize a JavaScript function source (text) and when reloading that, parse that text with goja to be able to run it dynamically in your Go program. Basically you'd need to get away from pure Go for this and towards something that is JSON compatible to (de)serialize to text.

Not really. It was built in-house and I don't know of anything about it that went public. I recall it using Goja for the JS runtime. Code was embedded into the binary (think embed package). There was some kind of sorcery to convert what would be HTTP network calls in the browser into local function calls during SSR, but I'm hazy on how it worked I'm afraid.