Changes

Prior to this PR, WithResponseUnmarshal was unmarshalling a response body to a bytes.Buffer by:

1. Reading all the response body in a slice of bytes; and
2. Writing that slice in the receiving bytes.Buffer.

Step 1 unnecessarily increases the memory footprint as the response body is essentially stored twice in memory during unmarshalling. This PR fixes that problem by loading the body directly in the buffer without passing by an intermediate container.

Tests

Added a few unit test to cover all unmarshalling types.

Memory benchmark:

var longString = strings.Repeat("a", 1<<30) 

func BenchmarkWithResponseUnmarshal_byteBuffer(b *testing.B) {
	b.ReportAllocs()

	client := NewApiClient(ClientConfig{
		Transport: hc(func(r *http.Request) (*http.Response, error) {
			return &http.Response{
				StatusCode: 200,
				Body:       io.NopCloser(strings.NewReader(longString)),
				Request:    r,
			}, nil
		}),
	})

	for i := 0; i < b.N; i++ {
		var got bytes.Buffer
		client.Do(context.Background(), "GET", "/a", WithResponseUnmarshal(&got))
	}
}

Results:

Before:	6810268826 B/op	140 allocs/op
After:	4294972868 B/op	101 allocs/op

• make test passing
• make fmt applied
• relevant integration tests applied