note
	description: "UTF-16 encoding routines"
	library: "Gobo Eiffel Kernel Library"
	copyright: "Copyright (c) 2002-2018, Eric Bezault and others"
	license: "MIT License"
	date: "$Date: 2019-02-07 22:54:15 +0000 (Thu, 07 Feb 2019) $"
	revision: "$Revision: 102807 $"

class interface
	UC_UTF16_ROUTINES

create 
	default_create

feature -- Status report

	valid_utf16 (a_string: STRING_8): BOOLEAN
			-- Are the bytes in a_string a valid UTF-16 encoding?
			-- 'a_string' has one byte per character.
			-- Default to big endian when no BOM.
		require
			a_string_not_void: a_string /= Void
			a_string_is_string: Any_.same_types (a_string, "")
		ensure
			instance_free: class
			empty_is_true: a_string.count = 0 implies Result
			utf16_even_count: Result implies ((a_string.count \\ 2) = 0)

	valid_utf16be (a_string: STRING_8): BOOLEAN
			-- Are the bytes in a_string valid UTF-16BE?
			-- 'a_string' has one byte per character.
		require
			a_string_not_void: a_string /= Void
			a_string_is_string: Any_.same_types (a_string, "")
		ensure
			instance_free: class
			empty_is_true: a_string.count = 0 implies Result
			utf16_even_count: Result implies ((a_string.count \\ 2) = 0)

	valid_utf16le (a_string: STRING_8): BOOLEAN
			-- Are the bytes in a_string valid UTF-16LE?
			-- 'a_string' has one byte per character.
		require
			a_string_not_void: a_string /= Void
			a_string_is_string: Any_.same_types (a_string, "")
		ensure
			instance_free: class
			empty_is_true: a_string.count = 0 implies Result
			utf16_even_count: Result implies ((a_string.count \\ 2) = 0)
	
feature -- Endian-ness detection

	Bom_be: STRING_8
			-- BOM in big-endian format
		ensure
			instance_free: class
			bom_be_not_void: Result /= Void
			two_bytes: Result.count = 2
			first_byte: Result.item_code (1) = Hex_fe
			second_byte: Result.item_code (2) = Hex_ff

	Bom_le: STRING_8
			-- BOM in little-endian format
		ensure
			instance_free: class
			bom_le_not_void: Result /= Void
			two_bytes: Result.count = 2
			first_byte: Result.item_code (1) = Hex_ff
			second_byte: Result.item_code (2) = Hex_fe

	is_endian_detection_character_most_first (first, second: INTEGER_32): BOOLEAN
			-- Do the two bytes first and second represent the character
			-- 0xFEFF with first being the most significant byte?
		require
			a_byte_is_byte: is_byte (first)
			other_byte_is_byte: is_byte (second)
		ensure
			instance_free: class
			definition: Result = (is_endian_detection_character (first, second) and (first = Hex_fe))

	is_endian_detection_character_least_first (first, second: INTEGER_32): BOOLEAN
			-- Do the two bytes first and second represent the character
			-- 0xFEFF with first being the least significant byte?
		require
			a_byte_is_byte: is_byte (first)
			other_byte_is_byte: is_byte (second)
		ensure
			instance_free: class
			definition: Result = (is_endian_detection_character (first, second) and (first = Hex_ff))

	is_endian_detection_character (a_byte, other_byte: INTEGER_32): BOOLEAN
			-- Can these two bytes represent ZERO WIDTH NON-BREAKING SPACE?
			-- (It has to be unicode character 0xFEFF, because 0xFFFE is not a valid character.)
		require
			a_byte_is_byte: is_byte (a_byte)
			other_byte_is_byte: is_byte (other_byte)
		ensure
			instance_free: class
			definition: Result = (a_byte.min (other_byte) = Hex_fe and a_byte.max (other_byte) = Hex_ff)
	
feature -- Surrogate

	is_surrogate (a_most: INTEGER_32): BOOLEAN
			-- Is this a high surrogate byte?
		require
			byte: is_byte (a_most)
		ensure
			instance_free: class

	is_high_surrogate (a_most: INTEGER_32): BOOLEAN
			-- Is this a high surrogate byte?
		require
			byte: is_byte (a_most)
		ensure
			instance_free: class

	is_low_surrogate (a_most: INTEGER_32): BOOLEAN
			-- Is this a low surrogate byte?
		require
			byte: is_byte (a_most)
		ensure
			instance_free: class

	least_10_bits (msb, lsb: INTEGER_32): INTEGER_32
			-- UTF16 least 10 bytes of a byte pair
		require
			msb_byte: is_byte (msb)
			lsb_byte: is_byte (lsb)
			surrogate: is_surrogate (msb)
		ensure
			instance_free: class
			ten_bits: Result >= 0 and Result < Hex_400

	surrogate (a_high_10: INTEGER_32; a_low_10: INTEGER_32): INTEGER_32
			-- Supplementary code point from high and low values
		require
			high_10: a_high_10 >= 0 and a_high_10 < 1024
			low_10: a_low_10 >= 0 and a_low_10 < 1024
		ensure
			instance_free: class
			more_than_16bits: Result >= Hex_10000

	surrogate_from_bytes (a_high_most, a_high_least, a_low_most, a_low_least: INTEGER_32): INTEGER_32
			-- Supplementary code point from bytes
		require
			surrogate_high: is_high_surrogate (a_high_most)
			high_least_byte: is_byte (a_high_least)
			surrogate_low: is_low_surrogate (a_low_most)
			low_least_byte: is_byte (a_low_least)
		ensure
			instance_free: class
			more_than_16bits: Result >= Hex_10000

	is_byte (a: INTEGER_32): BOOLEAN
			-- Is a a byte?
		ensure
			instance_free: class
			definition: Result = (a >= 0 and a < Hex_100)

	supplementary_to_high_surrogate (a_code: INTEGER_32): INTEGER_32
			-- High surrogate for a_code
		require
			code_high_enough: a_code > Maximum_bmp_character_code
			code_low_enough: a_code <= Maximum_unicode_character_code
		ensure
			instance_free: class
			high_surrogate: Result >= 256 * Hex_d8
			not_too_big: Result < 256 * Hex_dc

	supplementary_to_low_surrogate (a_code: INTEGER_32): INTEGER_32
			-- Low surrogate for a_code
		require
			code_high_enough: a_code > Maximum_bmp_character_code
			code_low_enough: a_code <= Maximum_unicode_character_code
		ensure
			instance_free: class
			low_surrogate: Result >= 256 * Hex_dc
			not_too_big: Result < 256 * Hex_e0
	
end -- class UC_UTF16_ROUTINES

Generated by ISE EiffelStudio