Switch the moclojer to reitit instead of pedestal-routes

[matheusfrancisco]

Currently, I run into an issue which is this one #206 to add swagger into moclojer, but pedestal does not offer a straightforward approach, but reitit does. But reitit does not offer a straightforward multi host then this metosin/reitit#660 showed me the path and I create this interceptor to reproduce https://github.com/matheusfrancisco/multihost-reitit-pedestal/blob/main/src/clojure/playground/server.clj#L26

So I see some paths.

1 - We do stay a little long without swagger, we try to figure out with pedestal-route.
2 - We migrate to reitit and then implement swagger?
3 - Or we try to implement a swagger for pedestal by our own (it could take a while)

What do you think about it?

[matheusfrancisco]

@avelino @J0sueTM

[avelino]

2 - We migrate to reitit and then implement swagger?

@matheusfrancisco this option makes sense to me, but we need to be clear on a few points:

• what is the difference between pedestal-route and reitit?
• what do we "lose" by switching to reitit?
• what do we "gain" by switching to reitit?

[matheusfrancisco]

What is the difference between pedestal-route and reitit?
Those are my opinion, what I understood about pedestal and reitit

Pedestal
Model: Pedestal uses an interceptor chain model, where each route can have its own series of interceptors (functions) that are executed before and after the actual handling of a request. This model allows for a high degree of customization and control over the request/response lifecycle.
Flexibility: Due to its interceptor model, Pedestal can be very flexible, allowing developers to insert, reorder, or remove steps in the processing of requests.
Verbosity: Configuring routes and interceptors in Pedestal can be more verbose compared to some other routing libraries, due to the explicitness required in defining the interceptor chains.

Reitit

Model: Reitit focuses on simplicity and performance, offering a more straightforward routing table definition. It supports both Clojure and Clojurescript and aims to be fast and data-driven. Also is possible to build chain interceptors using reitit.
Ease of Use: Reitit is often praised for its ease of use and simple setup, making it a good choice for projects where simplicity and speed are a priority.
Data-driven: Reitit routes are data-driven, making them easy to manipulate, generate, and introspect. This can be particularly beneficial for applications that require dynamic route generation.
Features: Reitit includes out-of-the-box support for Swagger/OpenAPI integration, making API documentation easier.

Both are flexible with middlewares and interceptors.
The big difference is:
If simplicity, performance, and ease of use are your priorities, along with good support for API documentation via Swagger/OpenAPI, then Reitit could be more suitable.

What do we "lose" by switching to reitit?

By default, we're going to lose the multi-host, and we will have to implement using interceptors

What do we "gain" by switching to reitit?

Out of the box, open api and swagger-ui.

[avelino]

The points that make sense to migrate to Reitit are:

• data-driven: mainly because we support extensive use ("framework")
• simplicity of use
• performance
• swagger by default

[matheusfrancisco]

I already create an issue to #226 .