improve byteorder docs for ReadBytesExt/WriteBytesExt

I believe I'm doing this correctly. I'm trying to create an f64 from 4 u16s.
```rust
extern crate byteorder;

use byteorder::{LittleEndian, ReadBytesExt, WriteBytesExt};

// hello
fn main() {
let mut cursor = std::io::Cursor::new(vec![]);
cursor.write_u16::(0u16).unwrap();
cursor.write_u16::(1u16).unwrap();
cursor.write_u16::(2u16).unwrap();
cursor.write_u16::(3u16).unwrap();

println!("{:?}", cursor);

let f = cursor.read_f64::().unwrap();
println!("{}", f);
}
```

My `println` outputs `Cursor { inner: [0, 0, 1, 0, 2, 0, 3, 0], pos: 8 }`, but the `read_f64` line panics with:

```
thread 'main' panicked at 'called `Result::unwrap()` on an `Err` value: Custom { kind: UnexpectedEof, error: StringError
("failed to fill whole buffer") }', libcore/result.rs:945:5
```

Am I doing something wrong? 8 x 8 = 64, so I believe the cursor has the correct number of bytes in it. I see no reason why this would be erroring in this way.

The issue names the ReadBytesExt and WriteBytesExt traits and includes a Rust reproduction; start by reading those trait docs and rerunning the example. Improve the documentation around the reported byte-count and cursor behavior so a reader can understand why read_f64 returns EOF and how to verify the intended read.

Written by the indexing model from the issue text.