Introduction

First of all, simple_dependency_finder(sdf) is built to figure out what test suits to run.

Big Project has a lot of unit tests. It takes long time to run all of test suits.

Some possible solution for faster test execution is..

1. Run chunk of test on multiple CPU
2. Run chunk of test on different machine
3. Make less IO by refactor test files
4. Run test suits as little as possible

sdf helps to do 4. Run test suits as little as possible

Install

download executable for your platform from Github Release.

How it works

Graph

Above graph show dependency graph.

Each Node represents main module unit of any Programing Language. For example, class for Ruby, function for JS, package for Go, etc.

Arrow Between Nodes represents Which module knows which module. For example, OrderItem knows both Order and Item directly, and knows User Indirectly through Order

Below table shows what to test or what is effected based on change on this dependency graph.

Basically, we need to test only effected modules.

diff	what to test
User	User, Order, OrderItem
Order	Order, OrderItem
Item	Item, OrderItem
OrderItem	OrderItem
CircularA	CircularA, CircularB
CircularC	CircularC, CircularD, CircularE

Function

sdf is cli tool, which is like function below.

Outputed effected modules are what we need to test.

It's important to understand that feeding these input is your responsibility.

How to parse code is very different based on context like Programing Language. So sdf does not care about it by design.

cli parameters...

option name	desc	required	default	type
mpath	path to json file which has modules info	true	NA	string
rpath	path to json file which has relation info	true	NA	string
diffs	file diffs comma separated	true	NA	string
depth	how deep you want to search	false	0(unlimited)	int

each json scheme will be found in testdata directory

example usage

sdf \
  -mpath testdata/modules.json \
  -rpath testdata/relations.json \
  -diffs order.rb

Example Usage

Case: Ruby project

sdf \
  -mpath testdata/modules.json \
  -rpath testdata/relations.json \
  -diffs $(git ls-files -omd --exclude-standard | sed "s,^,$PWD/,") \
  | jq -r '.[] | .name' > patterns.txt

grep -rl -f patterns.txt spec/ | xargs rspec
rm patterns.txt