//! Efficiently split lines by whitespace, while handling the backslash
//! escape sequences in Rust-like string format.
//!
//! For example, if you had a string like:
//! ```text
//!   One\ two three\ four
//! ```
//! Naïve splitting on whitespace would produce four outputs:
//! * `One\`
//! * `two`
//! * `three\`
//! * `four`
//!
//! This crate will instead produce two strings:
//! * `One two`
//! * `three four`
//!
//! This crate also handles special escape sequences like "\n", which represents a newline.
//! Specifically, the escape sequences are:
//!
//! * `\a`
//! * `\b`
//! * `\t`
//! * `\n`
//! * `\v`
//! * `\f`
//! * `\r`
//! * `\\`
//!
//! If the backslash character is found, but the successive character is not special,
//! then the backslash is disregarded and the successive character is included verbatim.

use std::borrow::Cow;

/// Splits once
///
/// Returns a tuple of the first "word" up until the first unescaped whitespace character,
/// and then every after the whitespace characters.
pub fn split_one_bytes<'a>(bytes: &'a [u8])
	-> Option<(Cow<'a, [u8]>, &'a [u8])>
{
	let mut start = 0usize;
	while let Some(b) = bytes.get(start)
	{
		if b.is_ascii_whitespace()
			{ start+=1; }
		else
			{ break; }
	}

	let mut owned: Option<Vec<u8>> = None;

	let mut position = start;

	while position < bytes.len()
	{
		if bytes[position] == b'\\'
		{
			if !owned.is_some()
			{
				owned = Some(bytes[start..position].to_owned());
			}
			let b = owned.as_mut().unwrap();
			position += 1;
			match bytes.get(position)
			{
				None => return None,
				Some(b'a') => b.push(b'\x07'),
				Some(b'b') => b.push(b'\x08'),
				Some(b't') => b.push(b'\t'),
				Some(b'n') => b.push(b'\n'),
				Some(b'v') => b.push(b'\x0b'),
				Some(b'f') => b.push(b'\x0c'),
				Some(b'r') => b.push(b'\r'),
				Some(b' ') => b.push(b' '),
				Some(b'\\') => b.push(b'\\'),
				Some(a) => b.push(*a),
			}
			position+=1;
		}
		else if bytes[position].is_ascii_whitespace()
		{
			break;
		}
		else
		{
			if let Some(o) = owned.as_mut()
				{ o.push( bytes[position] ); }
			position += 1;
		}
	}

	let mut after = position;
	while let Some(b) = bytes.get(after)
	{
		if b.is_ascii_whitespace()
			{ after+=1; }
		else
			{ break; }
	}

	let after = &bytes[after..];

	if let Some(owned) = owned
	{
		Some( (Cow::Owned(owned), after) )
	}
	else
	{
		Some( (Cow::Borrowed(&bytes[start..position]), after) )
	}
}

/// Split some text by unescaped whitespace.
///
/// find the first unescaped whitespace in `text`, return
/// a tuple of the text before the whitespace and the text after
/// the whitespace.
///
/// Ignores prefixed whitespace and discards whitespace between
/// the first portion and the text after the whitespace.
///
/// Returns None if there was an escape character and then nothing
///
/// Does not look at the following text at all.
pub fn split_one<'a>(text: &'a str)
	-> Option<(Cow<'a, str>, &'a str)>
{
	if let Some((one, remainder)) = split_one_bytes(text.as_bytes())
	{
		let one_text;
		match one
		{
			Cow::Borrowed(b) =>
				one_text = unsafe { Cow::Borrowed( std::str::from_utf8_unchecked(b) ) },
			Cow::Owned(b) =>
				one_text = unsafe { Cow::Owned( String::from_utf8_unchecked(b) ) },
		}
		let remainder = unsafe { std::str::from_utf8_unchecked(remainder) };
		Some((one_text, remainder))
	}
	else
	{
		None
	}
}

/// Produce all the words as a vector
pub fn split<'a>(mut text: &'a str)
	-> Option<Vec<Cow<'a, str>>>
{
	let mut res = vec!();

	while !text.is_empty()
	{
		let s = split_one(text);
		if s.is_none() { return None; }
		let s = s.unwrap();
		res.push( s.0 );
		text = s.1;
	}
	Some(res)
}

/// Converts text with all the special characters escape with a backslash
pub fn escape<'a>(text: &'a str)
	-> Cow<'a, str>
{
	let bytes = text.as_bytes();

	let mut owned = None;

	for pos in 0..bytes.len()
	{
		let special =
			match bytes[pos]
			{
				0x07 => Some(b'a'),
				0x08 => Some(b'b'),
				b'\t' => Some(b't'),
				b'\n' => Some(b'n'),
				0x0b => Some(b'v'),
				0x0c => Some(b'f'),
				b'\r' => Some(b'r'),
				b' ' => Some(b' '),
				b'\\' => Some(b'\\'),
				_ => None,
			};
		if let Some(s) = special
		{
			if owned.is_none()
			{
				owned = Some(bytes[0..pos].to_owned());
			}
			owned.as_mut().unwrap().push(b'\\');
			owned.as_mut().unwrap().push(s);
		}
		else if let Some(owned) = owned.as_mut()
		{
			owned.push( bytes[pos] );
		}
	}

	if let Some(owned) = owned
	{
		unsafe { Cow::Owned(String::from_utf8_unchecked(owned)) }
	}
	else
	{
		unsafe { Cow::Borrowed(std::str::from_utf8_unchecked(bytes)) }
	}
}

#[cfg(test)]
mod tests
{
	use ::split_one;
	use ::split;
	use ::escape;

	fn check(text: &str, one: &str, two: &str)
	{
		let a = split_one(text);
		let a = a.unwrap();
		assert_eq!(a.0, one);
		assert_eq!(a.1, two);
	}

	#[test]
	fn failure()
	{
		assert_eq!(split_one("abc\\"), None);
	}

	#[test]
	fn fine()
	{
		check("abc\\\\", "abc\\", "");
		check("1525824000000 520893", "1525824000000", "520893");
		check("abc\\\\ def", "abc\\", "def");
		check("abc\\\\\\\\ def", "abc\\\\", "def");
		check("abc\\\\\\\\    def", "abc\\\\", "def");
		check("abc\\ def   ghi", "abc def", "ghi");
		check("abc\\ def   ghi", "abc def", "ghi");
		check("", "", "");
		check(" ", "", "");
	}
	#[test]
	fn splitting()
	{
		assert_eq!(format!("{:?}",split("abc\\\\")), "Some([\"abc\\\\\"])");
		assert_eq!(format!("{:?}",split("abc def")), "Some([\"abc\", \"def\"])");
		assert_eq!(format!("{:?}",split("abc\\ def")), "Some([\"abc def\"])");
	}

	#[test]
	fn escaping()
	{
		assert_eq!(escape("abc\ndef"), "abc\\ndef");
		assert_eq!(escape("abc\n def"), "abc\\n\\ def");
	}

	#[test]
	fn round_trip()
	{
		check(&escape("ads\nasd"), "ads\nasd", "");
	}
}