perlunicode - Unicode support in Perl
WARNING: While the implementation of Unicode support in Perl is now fairly complete it is still evolving to some extent.
In particular the way Unicode is handled on EBCDIC platforms is still rather experimental. On such a platform references to UTF-8 encoding in this document and elsewhere should be read as meaning UTF-EBCDIC as specified in Unicode Technical Report 16 unless ASCII vs EBCDIC issues are specifically discussed. There is no utfebcdic pragma or ":utfebcdic" layer, rather "utf8" and ":utf8" are re-used to mean platform's "natural" 8-bit encoding of Unicode. See perlebcdic for more discussion of the issues.
The following areas are still under development.
A filehandle can be marked as containing perl's internal Unicode encoding (UTF-8 or UTF-EBCDIC) by opening it with the ":utf8" layer. Other encodings can be converted to perl's encoding on input, or from perl's encoding on output by use of the ":encoding()" layer. There is not yet a clean way to mark the perl source itself as being in an particular encoding.
The regular expression compiler does now attempt to produce polymorphic opcodes. That is the pattern should now adapt to the data and automatically switch to the Unicode character scheme when presented with Unicode data, or a traditional byte scheme when presented with byte data. The implementation is still new and (particularly on EBCDIC platforms) may need further work.
use utf8 still needed to enable a few featuresThe utf8 pragma implements the tables used for Unicode support. These tables are automatically loaded on demand, so the utf8 pragma need not normally be used.
However, as a compatibility measure, this pragma must be explicitly used to enable recognition of UTF-8 encoded literals and identifiers in the source text on ASCII based machines or recognize UTF-EBCDIC encoded literals and identifiers on EBCDIC based machines.
Beginning with version 5.6, Perl uses logically wide characters to represent strings internally. This internal representation of strings uses either the UTF-8 or the UTF-EBCDIC encoding.
In future, Perl-level operations can be expected to work with characters rather than bytes, in general.
However, as strictly an interim compatibility measure, Perl v5.6 aims to provide a safe migration path from byte semantics to character semantics for programs. For operations where Perl can unambiguously decide that the input data is characters, Perl now switches to character semantics. For operations where this determination cannot be made without additional information from the user, Perl decides in favor of compatibility, and chooses to use byte semantics.
This behavior preserves compatibility with earlier versions of Perl, which allowed byte semantics in Perl operations, but only as long as none of the program's inputs are marked as being as source of Unicode character data. Such data may come from filehandles, from calls to external programs, from information provided by the system (such as %ENV), or from literals and constants in the source text.
If the -C command line switch is used, (or the ${^WIDE_SYSTEM_CALLS} global flag is set to 1), all system calls will use the corresponding wide character APIs. This is currently only implemented on Windows since UNIXes lack API standard on this area.
Regardless of the above, the bytes pragma can always be used to force byte semantics in a particular lexical scope. See bytes.
The utf8 pragma is primarily a compatibility device that enables recognition of UTF-(8|EBCDIC) in literals encountered by the parser. It may also be used for enabling some of the more experimental Unicode support features. Note that this pragma is only required until a future version of Perl in which character semantics will become the default. This pragma may then become a no-op. See utf8.
Unless mentioned otherwise, Perl operators will use character semantics when they are dealing with Unicode data, and byte semantics otherwise. Thus, character semantics for these operations apply transparently; if the input data came from a Unicode source (for example, by adding a character encoding discipline to the filehandle whence it came, or a literal UTF-8 string constant in the program), character semantics apply; otherwise, byte semantics are in effect. To force byte semantics on Unicode data, the bytes pragma should be used.
Under character semantics, many operations that formerly operated on bytes change to operating on characters. For ASCII data this makes no difference, because UTF-8 stores ASCII in single bytes, but for any character greater than chr(127), the character may be stored in a sequence of two or more bytes, all of which have the high bit set. For C1 controls or Latin 1 characters on an EBCDIC platform the character may be stored in a UTF-EBCDIC multi byte sequence. But by and large, the user need not worry about this, because Perl hides it from the user. A character in Perl is logically just a number ranging from 0 to 2**32 or so. Larger characters encode to longer sequences of bytes internally, but again, this is just an internal detail which is hidden at the Perl level.
Character semantics have the following effects:
Strings and patterns may contain characters that have an ordinal value larger than 255.
Presuming you use a Unicode editor to edit your program, such characters will typically occur directly within the literal strings as UTF-(8|EBCDIC) characters, but you can also specify a particular character with an extension of the \x notation. UTF-X characters are specified by putting the hexadecimal code within curlies after the \x. For instance, a Unicode smiley face is \x{263A}.
Identifiers within the Perl script may contain Unicode alphanumeric characters, including ideographs. (You are currently on your own when it comes to using the canonical forms of characters--Perl doesn't (yet) attempt to canonicalize variable names for you.)
Regular expressions match characters instead of bytes. For instance, "." matches a character instead of a byte. (However, the \C pattern is provided to force a match a single byte ("char" in C, hence \C).)
Character classes in regular expressions match characters instead of bytes, and match against the character properties specified in the Unicode properties database. So \w can be used to match an ideograph, for instance.
Named Unicode properties and block ranges make be used as character classes via the new \p{} (matches property) and \P{} (doesn't match property) constructs. For instance, \p{Lu} matches any character with the Unicode uppercase property, while \p{M} matches any mark character. Single letter properties may omit the brackets, so that can be written \pM also. Many predefined character classes are available, such as \p{IsMirrored} and \p{InTibetan}.
The special pattern \X match matches any extended Unicode sequence (a "combining character sequence" in Standardese), where the first character is a base character and subsequent characters are mark characters that apply to the base character. It is equivalent to (?:\PM\pM*).
The tr/// operator translates characters instead of bytes. Note that the tr///CU functionality has been removed, as the interface was a mistake. For similar functionality see pack('U0', ...) and pack('C0', ...).
Case translation operators use the Unicode case translation tables when provided character input. Note that uc() translates to uppercase, while ucfirst translates to titlecase (for languages that make the distinction). Naturally the corresponding backslash sequences have the same semantics.
Most operators that deal with positions or lengths in the string will automatically switch to using character positions, including chop(), substr(), pos(), index(), rindex(), sprintf(), write(), and length(). Operators that specifically don't switch include vec(), pack(), and unpack(). Operators that really don't care include chomp(), as well as any other operator that treats a string as a bucket of bits, such as sort(), and the operators dealing with filenames.
The pack()/unpack() letters "c" and "C" do not change, since they're often used for byte-oriented formats. (Again, think "char" in the C language.) However, there is a new "U" specifier that will convert between UTF-8 characters and integers. (It works outside of the utf8 pragma too.)
The chr() and ord() functions work on characters. This is like pack("U") and unpack("U"), not like pack("C") and unpack("C"). In fact, the latter are how you now emulate byte-oriented chr() and ord() under utf8.
The bit string operators & | ^ ~ can operate on character data. However, for backward compatibility reasons (bit string operations when the characters all are less than 256 in ordinal value) one cannot mix ~ (the bit complement) and characters both less than 256 and equal or greater than 256. Most importantly, the DeMorgan's laws (~($x|$y) eq ~$x&~$y, ~($x&$y) eq ~$x|~$y) won't hold. Another way to look at this is that the complement cannot return both the 8-bit (byte) wide bit complement, and the full character wide bit complement.
And finally, scalar reverse() reverses by character rather than by byte.
See Encode.
As of yet, there is no method for automatically coercing input and output to some encoding other than UTF-8 or UTF-EBCDIC. This is planned in the near future, however.
Whether an arbitrary piece of data will be treated as "characters" or "bytes" by internal operations cannot be divined at the current time.
Use of locales with utf8 may lead to odd results. Currently there is some attempt to apply 8-bit locale info to characters in the range 0..255, but this is demonstrably incorrect for locales that use characters above that range (when mapped into Unicode). It will also tend to run slower. Avoidance of locales is strongly encouraged.