Skip to content

gomatic/go-kit-phases

00-init
Switch branches/tags

Name already in use

A tag already exists with the provided branch name. Many Git commands accept both tag and branch names, so creating this branch may cause unexpected behavior. Are you sure you want to create this branch?
Code

Latest commit

 

Git stats

Files

Permalink
Failed to load latest commit information.
Type
Name
Latest commit message
Commit time
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Introduction

“Namespaces are one honking great idea–let’s do more of those!”

A good package starts with a good name

This project began as just an attempt to refresh my understanding of go-kit while thinking through the organization and how to iteratively add features to a go-kit project. It evolved into a specialized layout designed for iterative development of transports and endpoints, isolation of implementation details, and to be a clear reflection of protobuf service definitions to make navigating and implementing the code more obvious and straightforward based mostly on the folder hierarchy.

A primary design principle of this project is centered around exposing the specifics of the service, its methods, and its types while retaining the great concepts and separation of concerns introduced by go-kit.

Additionally, within the package layout, import-aliases are utilized to standardize packages, making the code for the packages highly similar even though the types being referenced are incompatible. This technique was chosen over using interface types or reflection to keep the code more easily navigable and explicit.

import (
  in "github.com/gomatic/go-kit-phases/internal/api/moody/transform/feeling"
  out "github.com/gomatic/go-kit-phases/internal/api/moody/transform/overall"
)

func Request(_ context.Context, q interface{}) (interface{}, error)  { return in.To(q) }
func Response(_ context.Context, p interface{}) (interface{}, error) { return out.To(p) }

This design allows each folder to be highly standardized and allowing patterns to standout, making this layout suitable for code generation based entirely on the protobuf and the transport requirements of the service. And allowing for the endpoint package to be almost the only code requiring domain-specific customizations. That is, a generation command can take the protobuf and transport type(s) as input and fully generate everything except the implementations of the endpoints.

Getting started

Each branch documentation links to comparisons with the prior branch and explains the changes made.

Start here and follow along using the branches.