Add more write_X method and some bytes conversions to buffer

[SyoujyoujiNaiki]

Add methods to write UInt64, Int64, UInt, Int, Double and Float into buffer with their conversion to Bytes.

[coveralls]

Pull Request Test Coverage Report for Build 4190

Details

• 12 of 12 (100.0%) changed or added relevant lines in 1 file are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage increased (+0.04%) to 80.891%

---

Totals
Change from base Build 4188:	0.04%
Covered Lines:	4504
Relevant Lines:	5568

---

💛 - Coveralls

[Yoorkin]

Thanks for your contribution!

The write_X methods for the buffer look good to me, but the bytes_to_x seem unrelated to the buffer package. We also need further consideration about the API convention.

Can you separate out the bytes conversions part from this PR? Let's merge the write_X part first.

[Yoorkin]

Some consideration about performance:

The buffer is used to write large data consisting of multiple basic types (int, bool, double, etc.). It will call write_x frequently. We should try to eliminate the memory copy overhead in the write_x methods. Instead of converting to bytes and writing the bytes into the buffer, consider using write_byte directly

If you don't want to duplicate the code, maybe changing the implementation of X::to_bytes to use the x |> @buffer.write_x |> @buffer.to_bytes is a better choice, because those conversions are not frequently used. What do you think?