note
	description: "Eiffel implementations for Unicode encoding conversion."
	legal: "See notice at end of class."
	status: "See notice at end of class."
	date: "$Date: 2020-05-20 11:14:50 +0000 (Wed, 20 May 2020) $"
	revision: "$Revision: 104299 $"

class interface
	UNICODE_CONVERSION

create {ENCODING}
	default_create

feature -- Query

	is_code_page_valid (a_code_page: READABLE_STRING_8): BOOLEAN
			-- Is a_code_page valid?

	is_code_page_convertible (a_from_code_page, a_to_code_page: READABLE_STRING_8): BOOLEAN
			-- Is a_from_code_page convertible to a_to_code_page.

	Last_conversion_lost_data: BOOLEAN = False
			-- Has last conversion between two encodings lost data?

	is_valid_utf8 (a_string: READABLE_STRING_8): BOOLEAN
			-- Is a_string valid UTF-8 string?
		require
			a_string_not_void: a_string /= Void

	is_valid_as_string_16 (a_string: READABLE_STRING_GENERAL): BOOLEAN
			-- Check high 16 bit of any char in a_string is zero.
	
feature -- Conversion

	convert_to (a_from_code_page: READABLE_STRING_8; a_from_string: READABLE_STRING_GENERAL; a_to_code_page: READABLE_STRING_8)
			-- Convert between Unicode encodings.
	
feature -- Explicit Conversion

	utf8_to_utf32 (a_string: READABLE_STRING_8): STRING_32
			-- UTF8 to UTF32 conversion, Eiffel implementation.
		require
			a_string_not_void: a_string /= Void
		ensure
			result_not_void: Result /= Void

	utf32_to_utf8 (a_string: READABLE_STRING_32): STRING_8
			-- Convert UTF32 to UTF8.
		require
			a_string_not_void: a_string /= Void
		ensure
			result_not_void: Result /= Void

	utf32_to_utf16 (a_str: READABLE_STRING_32): STRING_32
			-- Convert utf32 to utf16 without data lose.
		require
			a_str_not_void: a_str /= Void
		ensure
			instance_free: class
			result_not_void: Result /= Void

	utf16_to_utf32 (a_str: READABLE_STRING_32): STRING_32
			-- Convert utf16 to utf32.
		require
			a_str_not_void: a_str /= Void
		ensure
			instance_free: class
			result_not_void: Result /= Void

	append_code_point_to_utf8 (a_code: NATURAL_32; a_string: STRING_8)
			-- Append a Unicode code point a_code to a UTF-8 stream.
		require
			a_string_not_void: a_string /= Void
			a_code_is_valid: a_code >= 0 and then a_code <= 1114111
		ensure
			a_string_appended: (a_code <= 127 implies a_string.count = old a_string.count + 1) and ((a_code > 127 and a_code <= 2047) implies a_string.count = old a_string.count + 2) and ((a_code > 2047 and a_code <= 65535) implies a_string.count = old a_string.count + 3) and ((a_code > 65535 and a_code <= 1114111) implies a_string.count = old a_string.count + 4)

	read_character_from_utf8 (a_position: INTEGER_32; a_read_bytes: detachable INTEGER_32_REF; a_string: READABLE_STRING_8): CHARACTER_32
			-- Read a Unicode character from UTF-8 string.
			-- a_string is in UTF-8.
			-- a_position is the starting byte point of a character.
			-- a_read_bytes is the number of bytes read.
		require
			a_string_not_void: a_string /= Void
			a_position_in_range: a_position > 0 and a_position <= a_string.count
			a_position_valid: a_string.code (a_position).to_natural_8 <= 127 or (a_string.code (a_position).to_natural_8 & 224) = 192 or (a_string.code (a_position).to_natural_8 & 240) = 224 or (a_string.code (a_position).to_natural_8 & 248) = 240 or (a_string.code (a_position).to_natural_8 & 252) = 248 or (a_string.code (a_position).to_natural_8 & 254) = 252
	
note
	library: "Encoding: Library of reusable components for Eiffel."
	copyright: "Copyright (c) 1984-2020, Eiffel Software and others"
	license: "Eiffel Forum License v2 (see http://www.eiffel.com/licensing/forum.txt)"
	source: "[
		Eiffel Software
		5949 Hollister Ave., Goleta, CA 93117 USA
		Telephone 805-685-1006, Fax 805-685-6869
		Website http://www.eiffel.com
		Customer support http://support.eiffel.com
	]"

end -- class UNICODE_CONVERSION

Generated by ISE EiffelStudio