Consider using regex groups rather than FM_BOUNDARY

[NGenetzky]

Regex is a powerful tool and it could be used to more precisely split the data. The performance impact appears to be negligible.

• yaml: (^---$\s*[\S\s]*?\s*^---$\s)([\S\s]*)
• toml: (^\+\+\+$\s*[\S\s]*?\s*^\+\+\+$\s)([\S\s]*)
• json: (^{$\s*[\S\s]*?\s*^}$\s)([\S\s]*)

I believe this would eliminate the need for START_DELIMITER and END_DELIMITER; however, these are exposed all the way up to the public API (for some reason?) and so that may be an issue.

If someone wanted to get fancy I am sure all three of the above regex could be generated from a function that accepts start and end delimiters.

[eyeseast]

I bounced back and forth on this in earlier versions and found that for my own sanity, using split() just made everything easier.

The reason for exposing START_DELIMITER and END_DELIMITER is so people can change the output.

I'm going to close this, but I appreciate the input.