Multi Byte Characters cause flat file definitions to shift

[QuickenTheDead]

In current state, if you have a file that contain multi-byte characters (for example, em dash, accented characters, etc). The unmarshal moves the characters over x positions depending on the bytes. This causes some values from other fields to not be in the correct struct field.

This PR will change the string unmarshal to convert to a rune before picking the start and end positions and select full characters instead of bytes.

[ianlopshire]

@QuickenTheDead, Thanks for taking the time to submit a PR!

I believe this case is covered by using the "use codepoint indices" option. Can you test after setting that flag and see if the problem persists?

decoder := fixedwidth.NewDecoder(strings.NewReader(data))
decoder.SetUseCodepointIndices(true)
// Decode as usual now

[QuickenTheDead]

@ianlopshire You're right, I was able to set that in a specific decoder and it does work as expected.

Question: Why isn't SetUseCodepointIndices(true) the default for the Unmarshal() method? Maybe I'm missing something, but I would thinking that would be a much more user friendly option then the number of bytes that it currently does.

With that, thanks for writing package! It's helped me save time on my current project

[ianlopshire]

@QuickenTheDead,

You are 100% right that using codepoint indices would be the ideal default behavior.

The initial releases of the library didn't support the option, so to maintain backward compatibility, it was made an opt-in feature.

When I get around to releasing v1.0.0 I'll make the behavior opt-out instead of opt-in. I've added #53 and #54 to track this.

Thanks again for the PR!