Report a bug
If you spot a problem with this page, click here to create a Bugzilla issue.
Improve this page
Quickly fork, edit online, and submit a pull request for this page. Requires a signed-in GitHub account. This works well for small changes. If you'd like to make larger changes you may want to consider using a local clone.

std.encoding

Classes and functions for handling and transcoding between various encodings.
For cases where the encoding is known at compile-time, functions are provided for arbitrary encoding and decoding of characters, arbitrary transcoding between strings of different type, as well as validation and sanitization.
Encodings currently supported are UTF-8, UTF-16, UTF-32, ASCII, ISO-8859-1 (also known as LATIN-1), ISO-8859-2 (LATIN-2), WINDOWS-1250 and WINDOWS-1252.
Category Functions
Decode codePoints decode decodeReverse safeDecode
Conversion codeUnits sanitize transcode
Classification canEncode isValid isValidCodePoint isValidCodeUnit
BOM BOM BOMSeq getBOM utfBOM
Length & Index firstSequence encodedLength index lastSequence validLength
Encoding schemes encodingName EncodingScheme EncodingSchemeASCII EncodingSchemeLatin1 EncodingSchemeLatin2 EncodingSchemeUtf16Native EncodingSchemeUtf32Native EncodingSchemeUtf8 EncodingSchemeWindows1250 EncodingSchemeWindows1252
Representation AsciiChar AsciiString Latin1Char Latin1String Latin2Char Latin2String Windows1250Char Windows1250String Windows1252Char Windows1252String
Exceptions INVALID_SEQUENCE EncodingException
For cases where the encoding is not known at compile-time, but is known at run-time, the abstract class EncodingScheme and its subclasses is provided. To construct a run-time encoder/decoder, one does e.g.
auto e = EncodingScheme.create("utf-8");
This library supplies EncodingScheme subclasses for ASCII, ISO-8859-1 (also known as LATIN-1), ISO-8859-2 (LATIN-2), WINDOWS-1250, WINDOWS-1252, UTF-8, and (on little-endian architectures) UTF-16LE and UTF-32LE; or (on big-endian architectures) UTF-16BE and UTF-32BE.
This library provides a mechanism whereby other modules may add EncodingScheme subclasses for any other encoding.
Authors:
Janice Caron
enum dchar INVALID_SEQUENCE;
Special value returned by safeDecode
enum AsciiChar: ubyte;

alias AsciiString = immutable(AsciiChar)[];
Defines various character sets.
enum Latin1Char: ubyte;
Defines an Latin1-encoded character.
alias Latin1String = immutable(Latin1Char)[];
Defines an Latin1-encoded string (as an array of immutable(Latin1Char)).
enum Latin2Char: ubyte;
Defines a Latin2-encoded character.
alias Latin2String = immutable(Latin2Char)[];
Defines an Latin2-encoded string (as an array of immutable(Latin2Char)).
enum Windows1250Char: ubyte;
Defines a Windows1250-encoded character.
alias Windows1250String = immutable(Windows1250Char)[];
Defines an Windows1250-encoded string (as an array of immutable(Windows1250Char)).
enum Windows1252Char: ubyte;
Defines a Windows1252-encoded character.
alias Windows1252String = immutable(Windows1252Char)[];
Defines an Windows1252-encoded string (as an array of immutable(Windows1252Char)).
pure nothrow @nogc @safe bool isValidCodePoint(dchar c);
Returns true if c is a valid code point
Note that this includes the non-character code points U+FFFE and U+FFFF, since these are valid code points (even though they are not valid characters).

Supersedes: This function supersedes std.utf.startsValidDchar().

Standards:
Unicode 5.0, ASCII, ISO-8859-1, ISO-8859-2, WINDOWS-1250, WINDOWS-1252
Parameters:
dchar c the code point to be tested
@property string encodingName(T)();
Returns the name of an encoding.
The type of encoding cannot be deduced. Therefore, it is necessary to explicitly specify the encoding type.
Standards:
Unicode 5.0, ASCII, ISO-8859-1, ISO-8859-2, WINDOWS-1250, WINDOWS-1252
Examples:
writeln(encodingName!(char)); // "UTF-8"
writeln(encodingName!(wchar)); // "UTF-16"
writeln(encodingName!(dchar)); // "UTF-32"
writeln(encodingName!(AsciiChar)); // "ASCII"
writeln(encodingName!(Latin1Char)); // "ISO-8859-1"
writeln(encodingName!(Latin2Char)); // "ISO-8859-2"
writeln(encodingName!(Windows1250Char)); // "windows-1250"
writeln(encodingName!(Windows1252Char)); // "windows-1252"
bool canEncode(E)(dchar c);
Returns true iff it is possible to represent the specified codepoint in the encoding.
The type of encoding cannot be deduced. Therefore, it is necessary to explicitly specify the encoding type.
Standards:
Unicode 5.0, ASCII, ISO-8859-1, ISO-8859-2, WINDOWS-1250, WINDOWS-1252
Examples:
assert( canEncode!(Latin1Char)('A'));
assert( canEncode!(Latin2Char)('A'));
assert(!canEncode!(AsciiChar)('\u00A0'));
assert( canEncode!(Latin1Char)('\u00A0'));
assert( canEncode!(Latin2Char)('\u00A0'));
assert( canEncode!(Windows1250Char)('\u20AC'));
assert(!canEncode!(Windows1250Char)('\u20AD'));
assert(!canEncode!(Windows1250Char)('\uFFFD'));
assert( canEncode!(Windows1252Char)('\u20AC'));
assert(!canEncode!(Windows1252Char)('\u20AD'));
assert(!canEncode!(Windows1252Char)('\uFFFD'));
assert(!canEncode!(char)(cast(dchar) 0x110000));
Examples:
How to check an entire string
import std.algorithm.searching : find;
import std.utf : byDchar;

assert("The quick brown fox"
    .byDchar
    .find!(x => !canEncode!AsciiChar(x))
    .empty);
bool isValidCodeUnit(E)(E c);
Returns true if the code unit is legal. For example, the byte 0x80 would not be legal in ASCII, because ASCII code units must always be in the range 0x00 to 0x7F.
Standards:
Unicode 5.0, ASCII, ISO-8859-1, ISO-8859-2, WINDOWS-1250, WINDOWS-1252
Parameters:
E c the code unit to be tested
Examples:
assert(!isValidCodeUnit(cast(char) 0xC0));
assert(!isValidCodeUnit(cast(char) 0xFF));
assert( isValidCodeUnit(cast(wchar) 0xD800));
assert(!isValidCodeUnit(cast(dchar) 0xD800));
assert(!isValidCodeUnit(cast(AsciiChar) 0xA0));
assert( isValidCodeUnit(cast(Windows1250Char) 0x80));
assert(!isValidCodeUnit(cast(Windows1250Char) 0x81));
assert( isValidCodeUnit(cast(Windows1252Char) 0x80));
assert(!isValidCodeUnit(cast(Windows1252Char) 0x81));
bool isValid(E)(const(E)[] s);
Returns true if the string is encoded correctly

Supersedes: This function supersedes std.utf.validate(), however note that this function returns a bool indicating whether the input was valid or not, whereas the older function would throw an exception.

Standards:
Unicode 5.0, ASCII, ISO-8859-1, ISO-8859-2, WINDOWS-1250, WINDOWS-1252
Parameters:
const(E)[] s the string to be tested
Examples:
assert( isValid("\u20AC100"));
assert(!isValid(cast(char[3])[167, 133, 175]));
size_t validLength(E)(const(E)[] s);
Returns the length of the longest possible substring, starting from the first code unit, which is validly encoded.
Standards:
Unicode 5.0, ASCII, ISO-8859-1, ISO-8859-2, WINDOWS-1250, WINDOWS-1252
Parameters:
const(E)[] s the string to be tested
immutable(E)[] sanitize(E)(immutable(E)[] s);
Sanitizes a string by replacing malformed code unit sequences with valid code unit sequences. The result is guaranteed to be valid for this encoding.
If the input string is already valid, this function returns the original, otherwise it constructs a new string by replacing all illegal code unit sequences with the encoding's replacement character, Invalid sequences will be replaced with the Unicode replacement character (U+FFFD) if the character repertoire contains it, otherwise invalid sequences will be replaced with '?'.
Standards:
Unicode 5.0, ASCII, ISO-8859-1, ISO-8859-2, WINDOWS-1250, WINDOWS-1252
Parameters:
immutable(E)[] s the string to be sanitized
Examples:
writeln(sanitize("hello \xF0\x80world")); // "hello \xEF\xBF\xBDworld"
size_t firstSequence(E)(const(E)[] s);
Returns the length of the first encoded sequence.
The input to this function MUST be validly encoded. This is enforced by the function's in-contract.
Standards:
Unicode 5.0, ASCII, ISO-8859-1, ISO-8859-2, WINDOWS-1250, WINDOWS-1252
Parameters:
const(E)[] s the string to be sliced
Examples:
writeln(firstSequence("\u20AC1000")); // "\u20AC".length
writeln(firstSequence("hel")); // "h".length
size_t lastSequence(E)(const(E)[] s);
Returns the length of the last encoded sequence.
The input to this function MUST be validly encoded. This is enforced by the function's in-contract.
Standards:
Unicode 5.0, ASCII, ISO-8859-1, ISO-8859-2, WINDOWS-1250, WINDOWS-1252
Parameters:
const(E)[] s the string to be sliced
Examples:
writeln(lastSequence("1000\u20AC")); // "\u20AC".length
writeln(lastSequence("hellö")); // "ö".length
ptrdiff_t index(E)(const(E)[] s, int n);
Returns the array index at which the (n+1)th code point begins.
The input to this function MUST be validly encoded. This is enforced by the function's in-contract.

Supersedes: This function supersedes std.utf.toUTFindex().

Standards:
Unicode 5.0, ASCII, ISO-8859-1, ISO-8859-2, WINDOWS-1250, WINDOWS-1252
Parameters:
const(E)[] s the string to be counted
int n the current code point index
Examples:
writeln(index("\u20AC100", 1)); // 3
writeln(index("hällo", 2)); // 3
dchar decode(S)(ref S s);
Decodes a single code point.
This function removes one or more code units from the start of a string, and returns the decoded code point which those code units represent.
The input to this function MUST be validly encoded. This is enforced by the function's in-contract.

Supersedes: This function supersedes std.utf.decode(), however, note that the function codePoints() supersedes it more conveniently.

Standards:
Unicode 5.0, ASCII, ISO-8859-1, ISO-8859-2, WINDOWS-1250, WINDOWS-1252
Parameters:
S s the string whose first code point is to be decoded
dchar decodeReverse(E)(ref const(E)[] s);
Decodes a single code point from the end of a string.
This function removes one or more code units from the end of a string, and returns the decoded code point which those code units represent.
The input to this function MUST be validly encoded. This is enforced by the function's in-contract.
Standards:
Unicode 5.0, ASCII, ISO-8859-1, ISO-8859-2, WINDOWS-1250, WINDOWS-1252
Parameters:
const(E)[] s the string whose first code point is to be decoded
dchar safeDecode(S)(ref S s);
Decodes a single code point. The input does not have to be valid.
This function removes one or more code units from the start of a string, and returns the decoded code point which those code units represent.
This function will accept an invalidly encoded string as input. If an invalid sequence is found at the start of the string, this function will remove it, and return the value INVALID_SEQUENCE.
Standards:
Unicode 5.0, ASCII, ISO-8859-1, ISO-8859-2, WINDOWS-1250, WINDOWS-1252
Parameters:
S s the string whose first code point is to be decoded
size_t encodedLength(E)(dchar c);
Returns the number of code units required to encode a single code point.
The input to this function MUST be a valid code point. This is enforced by the function's in-contract.
The type of the output cannot be deduced. Therefore, it is necessary to explicitly specify the encoding as a template parameter.
Standards:
Unicode 5.0, ASCII, ISO-8859-1, ISO-8859-2, WINDOWS-1250, WINDOWS-1252
Parameters:
dchar c the code point to be encoded
E[] encode(E)(dchar c);
Encodes a single code point.
This function encodes a single code point into one or more code units. It returns a string containing those code units.
The input to this function MUST be a valid code point. This is enforced by the function's in-contract.
The type of the output cannot be deduced. Therefore, it is necessary to explicitly specify the encoding as a template parameter.

Supersedes: This function supersedes std.utf.encode(), however, note that the function codeUnits() supersedes it more conveniently.

Standards:
Unicode 5.0, ASCII, ISO-8859-1, ISO-8859-2, WINDOWS-1250, WINDOWS-1252
Parameters:
dchar c the code point to be encoded
size_t encode(E)(dchar c, E[] array);
Encodes a single code point into an array.
This function encodes a single code point into one or more code units The code units are stored in a user-supplied fixed-size array, which must be passed by reference.
The input to this function MUST be a valid code point. This is enforced by the function's in-contract.
The type of the output cannot be deduced. Therefore, it is necessary to explicitly specify the encoding as a template parameter.

Supersedes: This function supersedes std.utf.encode(), however, note that the function codeUnits() supersedes it more conveniently.

Standards:
Unicode 5.0, ASCII, ISO-8859-1, ISO-8859-2, WINDOWS-1250, WINDOWS-1252
Parameters:
dchar c the code point to be encoded
E[] array the destination array
Returns:
the number of code units written to the array
void encode(E)(dchar c, void delegate(E) dg);
Encodes a single code point to a delegate.
This function encodes a single code point into one or more code units. The code units are passed one at a time to the supplied delegate.
The input to this function MUST be a valid code point. This is enforced by the function's in-contract.
The type of the output cannot be deduced. Therefore, it is necessary to explicitly specify the encoding as a template parameter.

Supersedes: This function supersedes std.utf.encode(), however, note that the function codeUnits() supersedes it more conveniently.

Standards:
Unicode 5.0, ASCII, ISO-8859-1, ISO-8859-2, WINDOWS-1250, WINDOWS-1252
Parameters:
dchar c the code point to be encoded
void delegate(E) dg the delegate to invoke for each code unit
size_t encode(Tgt, Src, R)(in Src[] s, R range);
Encodes the contents of s in units of type Tgt, writing the result to an output range.
Returns:
The number of Tgt elements written.
Parameters:
Tgt Element type of range.
Src[] s Input array.
R range Output range.
CodePoints!E codePoints(E)(immutable(E)[] s);
Returns a foreachable struct which can bidirectionally iterate over all code points in a string.
The input to this function MUST be validly encoded. This is enforced by the function's in-contract.
You can foreach either with or without an index. If an index is specified, it will be initialized at each iteration with the offset into the string at which the code point begins.

Supersedes: This function supersedes std.utf.decode().

Standards:
Unicode 5.0, ASCII, ISO-8859-1, ISO-8859-2, WINDOWS-1250, WINDOWS-1252
Parameters:
immutable(E)[] s the string to be decoded

Example:

string s = "hello world";
foreach (c;codePoints(s))
{
    // do something with c (which will always be a dchar)
}
Note that, currently, foreach (c:codePoints(s)) is superior to foreach (c;s) in that the latter will fall over on encountering U+FFFF.

Examples:
string s = "hello";
string t;
foreach (c;codePoints(s))
{
    t ~= cast(char) c;
}
writeln(s); // t
CodeUnits!E codeUnits(E)(dchar c);
Returns a foreachable struct which can bidirectionally iterate over all code units in a code point.
The input to this function MUST be a valid code point. This is enforced by the function's in-contract.
The type of the output cannot be deduced. Therefore, it is necessary to explicitly specify the encoding type in the template parameter.

Supersedes: This function supersedes std.utf.encode().

Standards:
Unicode 5.0, ASCII, ISO-8859-1, ISO-8859-2, WINDOWS-1250, WINDOWS-1252
Parameters:
dchar c the code point to be encoded
Examples:
char[] a;
foreach (c;codeUnits!(char)(cast(dchar)'\u20AC'))
{
    a ~= c;
}
writeln(a.length); // 3
writeln(a[0]); // 0xE2
writeln(a[1]); // 0x82
writeln(a[2]); // 0xAC
void transcode(Src, Dst)(Src[] s, out Dst[] r);
Convert a string from one encoding to another.

Supersedes: This function supersedes std.utf.toUTF8(), std.utf.toUTF16() and std.utf.toUTF32() (but note that to!() supersedes it more conveniently).

Standards:
Unicode 5.0, ASCII, ISO-8859-1, ISO-8859-2, WINDOWS-1250, WINDOWS-1252
Parameters:
Src[] s Source string. Must be validly encoded. This is enforced by the function's in-contract.
Dst[] r Destination string
See Also:
Examples:
wstring ws;
// transcode from UTF-8 to UTF-16
transcode("hello world",ws);
writeln(ws); // "hello world"w

Latin1String ls;
// transcode from UTF-16 to ISO-8859-1
transcode(ws, ls);
writeln(ws); // "hello world"
class EncodingException: object.Exception;
The base class for exceptions thrown by this module
abstract class EncodingScheme;
Abstract base class of all encoding schemes
void register(Klass : EncodingScheme)();
Registers a subclass of EncodingScheme.
This function allows user-defined subclasses of EncodingScheme to be declared in other modules.
Parameters:
Klass The subclass of EncodingScheme to register.

Example:

class Amiga1251 : EncodingScheme
{
    shared static this()
    {
        EncodingScheme.register!Amiga1251;
    }
}

static EncodingScheme create(string encodingName);
Obtains a subclass of EncodingScheme which is capable of encoding and decoding the named encoding scheme.
This function is only aware of EncodingSchemes which have been registered with the register() function.

Example:

auto scheme = EncodingScheme.create("Amiga-1251");

abstract const string toString();
Returns the standard name of the encoding scheme
abstract const string[] names();
Returns an array of all known names for this encoding scheme
abstract const bool canEncode(dchar c);
Returns true if the character c can be represented in this encoding scheme.
abstract const size_t encodedLength(dchar c);
Returns the number of ubytes required to encode this code point.
The input to this function MUST be a valid code point.
Parameters:
dchar c the code point to be encoded
Returns:
the number of ubytes required.
abstract const size_t encode(dchar c, ubyte[] buffer);
Encodes a single code point into a user-supplied, fixed-size buffer.
This function encodes a single code point into one or more ubytes. The supplied buffer must be code unit aligned. (For example, UTF-16LE or UTF-16BE must be wchar-aligned, UTF-32LE or UTF-32BE must be dchar-aligned, etc.)
The input to this function MUST be a valid code point.
Parameters:
dchar c the code point to be encoded
ubyte[] buffer the destination array
Returns:
the number of ubytes written.
abstract const dchar decode(ref const(ubyte)[] s);
Decodes a single code point.
This function removes one or more ubytes from the start of an array, and returns the decoded code point which those ubytes represent.
The input to this function MUST be validly encoded.
Parameters:
const(ubyte)[] s the array whose first code point is to be decoded
abstract const dchar safeDecode(ref const(ubyte)[] s);
Decodes a single code point. The input does not have to be valid.
This function removes one or more ubytes from the start of an array, and returns the decoded code point which those ubytes represent.
This function will accept an invalidly encoded array as input. If an invalid sequence is found at the start of the string, this function will remove it, and return the value INVALID_SEQUENCE.
Parameters:
const(ubyte)[] s the array whose first code point is to be decoded
abstract const @property immutable(ubyte)[] replacementSequence();
Returns the sequence of ubytes to be used to represent any character which cannot be represented in the encoding scheme.
Normally this will be a representation of some substitution character, such as U+FFFD or '?'.
bool isValid(const(ubyte)[] s);
Returns true if the array is encoded correctly
Parameters:
const(ubyte)[] s the array to be tested
size_t validLength()(const(ubyte)[] s);
Returns the length of the longest possible substring, starting from the first element, which is validly encoded.
Parameters:
const(ubyte)[] s the array to be tested
immutable(ubyte)[] sanitize()(immutable(ubyte)[] s);
Sanitizes an array by replacing malformed ubyte sequences with valid ubyte sequences. The result is guaranteed to be valid for this encoding scheme.
If the input array is already valid, this function returns the original, otherwise it constructs a new array by replacing all illegal sequences with the encoding scheme's replacement sequence.
Parameters:
immutable(ubyte)[] s the string to be sanitized
size_t firstSequence()(const(ubyte)[] s);
Returns the length of the first encoded sequence.
The input to this function MUST be validly encoded. This is enforced by the function's in-contract.
Parameters:
const(ubyte)[] s the array to be sliced
size_t count()(const(ubyte)[] s);
Returns the total number of code points encoded in a ubyte array.
The input to this function MUST be validly encoded. This is enforced by the function's in-contract.
Parameters:
const(ubyte)[] s the string to be counted
ptrdiff_t index()(const(ubyte)[] s, size_t n);
Returns the array index at which the (n+1)th code point begins.
The input to this function MUST be validly encoded. This is enforced by the function's in-contract.
Parameters:
const(ubyte)[] s the string to be counted
size_t n the current code point index
class EncodingSchemeASCII: std.encoding.EncodingScheme;
EncodingScheme to handle ASCII
This scheme recognises the following names: "ANSI_X3.4-1968", "ANSI_X3.4-1986", "ASCII", "IBM367", "ISO646-US", "ISO_646.irv:1991", "US-ASCII", "cp367", "csASCII" "iso-ir-6", "us"
class EncodingSchemeLatin1: std.encoding.EncodingScheme;
EncodingScheme to handle Latin-1
This scheme recognises the following names: "CP819", "IBM819", "ISO-8859-1", "ISO_8859-1", "ISO_8859-1:1987", "csISOLatin1", "iso-ir-100", "l1", "latin1"
class EncodingSchemeLatin2: std.encoding.EncodingScheme;
EncodingScheme to handle Latin-2
This scheme recognises the following names: "Latin 2", "ISO-8859-2", "ISO_8859-2", "ISO_8859-2:1999", "Windows-28592"
class EncodingSchemeWindows1250: std.encoding.EncodingScheme;
EncodingScheme to handle Windows-1250
This scheme recognises the following names: "windows-1250"
class EncodingSchemeWindows1252: std.encoding.EncodingScheme;
EncodingScheme to handle Windows-1252
This scheme recognises the following names: "windows-1252"
class EncodingSchemeUtf8: std.encoding.EncodingScheme;
EncodingScheme to handle UTF-8
This scheme recognises the following names: "UTF-8"
class EncodingSchemeUtf16Native: std.encoding.EncodingScheme;
EncodingScheme to handle UTF-16 in native byte order
This scheme recognises the following names: "UTF-16LE" (little-endian architecture only) "UTF-16BE" (big-endian architecture only)
class EncodingSchemeUtf32Native: std.encoding.EncodingScheme;
EncodingScheme to handle UTF-32 in native byte order
This scheme recognises the following names: "UTF-32LE" (little-endian architecture only) "UTF-32BE" (big-endian architecture only)
enum BOM: int;
Definitions of common Byte Order Marks. The elements of the enum can used as indices into bomTable to get matching BOMSeq.
none
no BOM was found
utf32be
[0x00, 0x00, 0xFE, 0xFF]
utf32le
[0xFF, 0xFE, 0x00, 0x00]
utf1
[0xF7, 0x64, 0x4C]
utfebcdic
[0xDD, 0x73, 0x66, 0x73]
scsu
[0x0E, 0xFE, 0xFF]
bocu1
[0xFB, 0xEE, 0x28]
gb18030
[0x84, 0x31, 0x95, 0x33]
utf8
[0xEF, 0xBB, 0xBF]
utf16be
[0xFE, 0xFF]
utf16le
[0xFF, 0xFE]
alias BOMSeq = std.typecons.Tuple!(BOM, "schema", ubyte[], "sequence").Tuple;
The type stored inside bomTable.
immutable Tuple!(BOM, "schema", ubyte[], "sequence")[] bomTable;
Mapping of a byte sequence to Byte Order Mark (BOM)
immutable(BOMSeq) getBOM(Range)(Range input)
if (isForwardRange!Range && is(Unqual!(ElementType!Range) == ubyte));
Returns a BOMSeq for a given input. If no BOM is present the BOMSeq for BOM.none is returned. The BOM sequence at the beginning of the range will not be comsumed from the passed range. If you pass a reference type range make sure that save creates a deep copy.
Parameters:
Range input The sequence to check for the BOM
Returns:
the found BOMSeq corresponding to the passed input.
Examples:
import std.format : format;

auto ts = dchar(0x0000FEFF) ~ "Hello World"d;

auto entry = getBOM(cast(ubyte[]) ts);
version(BigEndian)
{
    writeln(entry.schema); // BOM.utf32be
}
else
{
    writeln(entry.schema); // BOM.utf32le
}
enum dchar utfBOM;
Constant defining a fully decoded BOM