lang.pn

general:
	>	Indentation starts a block. Tabs are preferred, and are equal to two
	>	spaces. (Valid) UTF-8 is allowed inside strings and comments. ASCII is
	>	allowed everywhere, except for control characters, which aren't allowed
	>	anywhere. Whitespace means tabs or spaces.

types:
	scalar:
		null:   null
		bool:   [true, false]
		int:    [0, 20, +1, -40]
		float:  [0.0, +0.0, -0.0, +6.02e23, inf, +inf, -inf, nan]
	vector:
		data:
			short:  $0f1e2d3c
			long:
				$	0000 0000 0000 0000 1111 1111 1111 1111
				$	2222 2222 2222 2222 3333 3333 3333 3333
				$	4444 4444 4444 4444 5555 5555 5555 5555
				$	6666 6666 6666 6666 7777 7777 7777 7777
				$	8888 8888 8888 8888 9999 9999 9999 9999
				$	aaaa aaaa aaaa aaaa bbbb bbbb bbbb bbbb
				$	cccc cccc cccc cccc dddd dddd dddd dddd
				$	eeee eeee eeee eeee ffff ffff ffff ffff
		string:
			short:  "use quotes\u0000use escapes"
			long:
				>	Multi-line literals use ">", "|", and "!". One tab or space character
				>	may follow them and is ignored. With ">", non-empty input lines are
				>	joined to previous non-empty lines by a space. With "|" (always) or ">"
				>	(when the current or previous line is empty), each input line begins a
				>	new line.
				>
				>	A trailing newline is added unless the last line uses "!". A "!" ends
				>	the string: no text may occur after the "!" (except an optional space or
				>	tab) and no ">" or "!" lines may follow it.
				!
		array:
			short:  [1, 2, 3]
			long:
				*	"multi-line arrays"
				*	"use stars"
		dict:
			short:  {use-braces: true, use-colons: true, "and sometimes quotes": true}
			long:
				multi-line:     "maps"
				use:            "colons"
				"quoted keys":  "OK too"

	#	There are comments too, but they aren't a type.
	#	They have to follow the same indentation rules as everything else.

notes:
	>	Short vector literals must always fit on one line. They are essentially
	>	the same as JSON, except that dict keys do not have to be quoted if they
	>	contain only [0-9A-Za-z./+-].