Compatibility instead of comparability 😁

Thanks for your advise. But I think use Semantic Versioning to handle this situation is a better idea.

@sturfee-petrl @im7mortal So sorry for this issue. I will use Semantic Versioning later.

I want to reassure you one more time.

There are only 2 options LittleEndian or BigEndian.
It's much more clear to set it in New then do redundant SetByteOrder.
There are no more important parameters for the binpacker .

Like I told before 👉 go way is when code is ridiculously simple 👈. There no any clearance why it have LittleEndian by default. It's black box. 💥

There a lot of examples when New function have important explicit parameters in standard library.
For example encoding/base32

    func NewDecoder(enc *Encoding, r io.Reader) io.Reader
    func NewEncoder(enc *Encoding, w io.Writer) io.WriteCloser

⚠️ You can notice that Encoding even first parameter! ⚠️ Because it's important to tell people what they are going to do!! ❗

LittleEndian and BigEndian are also encoding
People should first choose encoding. Nobody forgot about Reader or Writer. But everyone forgot about SetByteOrder. 💯

@im7mortal I'm think about this again and I think your idea is the correct solution. I'm a little busy these days, can you send a PR for this please?