unicode

Documentation

Overview

Package unicode provides Unicode encodings such as UTF-16.

Index

• Variables
• func BOMOverride(fallback transform.Transformer) transform.Transformer
• func UTF16(e Endianness, b BOMPolicy) encoding.Encoding
• type BOMPolicy
• type Endianness

Variables

var All = []encoding.Encoding{
	UTF16(BigEndian, UseBOM),
	UTF16(BigEndian, IgnoreBOM),
	UTF16(LittleEndian, IgnoreBOM),
}

All lists a configuration for each IANA-defined UTF-16 variant.

var ErrMissingBOM = errors.New("encoding: missing byte order mark")

ErrMissingBOM means that decoding UTF-16 input with ExpectBOM did not find a starting byte order mark.

Functions

func BOMOverride

func BOMOverride(fallback transform.Transformer) transform.Transformer

BOMOverride returns a new decoder transformer that is identical to fallback, except that the presence of a Byte Order Mark at the start of the input causes it to switch to the corresponding Unicode decoding. It will only consider BOMs for UTF-8, UTF-16BE, and UTF-16LE.

This differs from using ExpectBOM by allowing a BOM to switch to UTF-8, not just UTF-16 variants, and allowing falling back to any encoding scheme.

This technique is recommended by the W3C for use in HTML 5: "For compatibility with deployed content, the byte order mark (also known as BOM) is considered more authoritative than anything else." http://www.w3.org/TR/encoding/#specification-hooks

Using BOMOverride is mostly intended for use cases where the first characters of a fallback encoding are known to not be a BOM, for example, for valid HTML and most encodings.

func UTF16

func UTF16(e Endianness, b BOMPolicy) encoding.Encoding

UTF16 returns a UTF-16 Encoding for the given default endianness and byte order mark (BOM) policy.

When decoding from UTF-16 to UTF-8, if the BOMPolicy is IgnoreBOM then neither BOMs U+FEFF nor noncharacters U+FFFE in the input stream will affect the endianness used for decoding, and will instead be output as their standard UTF-8 encodings: "\xef\xbb\xbf" and "\xef\xbf\xbe". If the BOMPolicy is UseBOM or ExpectBOM a staring BOM is not written to the UTF-8 output. Instead, it overrides the default endianness e for the remainder of the transformation. Any subsequent BOMs U+FEFF or noncharacters U+FFFE will not affect the endianness used, and will instead be output as their standard UTF-8 encodings. For UseBOM, if there is no starting BOM, it will proceed with the default Endianness. For ExpectBOM, in that case, the transformation will return early with an ErrMissingBOM error.

When encoding from UTF-8 to UTF-16, a BOM will be inserted at the start of the output if the BOMPolicy is UseBOM or ExpectBOM. Otherwise, a BOM will not be inserted. The UTF-8 input does not need to contain a BOM.

There is no concept of a 'native' endianness. If the UTF-16 data is produced and consumed in a greater context that implies a certain endianness, use IgnoreBOM. Otherwise, use ExpectBOM and always produce and consume a BOM.

In the language of http://www.unicode.org/faq/utf_bom.html#bom10, IgnoreBOM corresponds to "Where the precise type of the data stream is known... the BOM should not be used" and ExpectBOM corresponds to "A particular protocol... may require use of the BOM".

Types

type BOMPolicy

type BOMPolicy uint8

BOMPolicy is a UTF-16 encoding's byte order mark policy.

const (

	// IgnoreBOM means to ignore any byte order marks.
	IgnoreBOM BOMPolicy = 0

	// UseBOM means that the UTF-16 form may start with a byte order mark, which
	// will be used to override the default encoding.
	UseBOM BOMPolicy = writeBOM | acceptBOM

	// ExpectBOM means that the UTF-16 form must start with a byte order mark,
	// which will be used to override the default encoding.
	ExpectBOM BOMPolicy = writeBOM | acceptBOM | requireBOM
)

type Endianness

type Endianness bool

Endianness is a UTF-16 encoding's default endianness.

const (
	// BigEndian is UTF-16BE.
	BigEndian Endianness = false
	// LittleEndian is UTF-16LE.
	LittleEndian Endianness = true
)