I am currently creating unit testing for a project, i need to test a function that have dependency injection without interface from an entity. Here is the code
func (u UserUsecase) CreateUser(ctx context.Context, p request.StoreUser) (entity.User, error) {
user := &entity.User{
Username: p.Username,
Email: p.Email,
Phone: p.Phone,
Password: p.Password,
CompanyID: p.CompanyID,
}
err := user.Create(ctx, u.userRepo)
if err != nil {
return entity.User{}, err
}
return *user, nil
}
In the user.Create(ctx, u.userRepo)
there is a function that call another function so when I create a test function I've been stuck and getting invalid memory address or null pointer. Here is the code in the user.Create(ctx, u.userRepo)
func (u *User) Create(ctx context.Context, ur UserStoreRepo) error {
u.UUID = uuid.New()
u.Status = constanta.ACTIVE
pass, err := helpers.Bcrypting(u.Password)
if err != nil {
return err
}
u.Password = pass
err = ur.Store(ctx, u)
return err
}
What is the best method to override the user.Create(ctx, u.userRepo)
so i can test the CreateUser(ctx context.Context, p request.StoreUser)
.
CodePudding user response:
Use interface and use mock struct implements it to test
type User interface {
CreateUser(ctx context.Context, p request.StoreUser) (entity.User, error)
}
type user struct {
//... user struct
}
func (u *user) CreateUser(ctx context.Context, p request.StoreUser) (entity.User, error) {
//... real code goes here
}
Then create a mock
type mockUser struct {
result entity.User
err error
}
func (u *mockUser) CreateUser(ctx context.Context, p request.StoreUser) (entity.User, error) {
return u.result, u.err
}
Then create a test like
func TestError(t *testing.T) {
errMockUser := &mockUser{nil,erorrs.New("error")}
result, err := errMockUser.CreateUser(context.Background(),
request.StoreUser{})
assert.Nil(t, result)
assert.Equal(t, errors.New("error"), err)
}