=head1 NAME
perlunicode - Unicode support in Perl
=head1 DESCRIPTION
If you haven't already, before reading this document, you should become
familiar with both L and L.
Unicode aims to B-fy the en-B-ings of all the world's
character sets into a single Standard. For quite a few of the various
coding standards that existed when Unicode was first created, converting
from each to Unicode essentially meant adding a constant to each code
point in the original standard, and converting back meant just
subtracting that same constant. For ASCII and ISO-8859-1, the constant
is 0. For ISO-8859-5, (Cyrillic) the constant is 864; for Hebrew
(ISO-8859-8), it's 1488; Thai (ISO-8859-11), 3424; and so forth. This
made it easy to do the conversions, and facilitated the adoption of
Unicode.
And it worked; nowadays, those legacy standards are rarely used. Most
everyone uses Unicode.
Unicode is a comprehensive standard. It specifies many things outside
the scope of Perl, such as how to display sequences of characters. For
a full discussion of all aspects of Unicode, see
L.
=head2 Important Caveats
Even though some of this section may not be understandable to you on
first reading, we think it's important enough to highlight some of the
gotchas before delving further, so here goes:
Unicode support is an extensive requirement. While Perl does not
implement the Unicode standard or the accompanying technical reports
from cover to cover, Perl does support many Unicode features.
Also, the use of Unicode may present security issues that aren't
obvious, see L below.
=over 4
=item Safest if you C
In order to preserve backward compatibility, Perl does not turn
on full internal Unicode support unless the pragma
L>|feature/The 'unicode_strings' feature>
is specified. (This is automatically
selected if you S> or higher.) Failure to do this can
trigger unexpected surprises. See L below.
This pragma doesn't affect I/O. Nor does it change the internal
representation of strings, only their interpretation. There are still
several places where Unicode isn't fully supported, such as in
filenames.
=item Input and Output Layers
Use the C<:encoding> layer to read from and write to
filehandles using the specified encoding. (See L.)
=item You must convert your non-ASCII, non-UTF-8 Perl scripts to be
UTF-8.
The L module has been deprecated since perl 5.18 and the
perl internals it requires have been removed with perl 5.26.
=item C still needed to enable L in scripts
If your Perl script is itself encoded in L,
the S> pragma must be explicitly included to enable
recognition of that (in string or regular expression literals, or in
identifier names). B> is needed.> (See L).
If a Perl script begins with the bytes that form the UTF-8 encoding of
the Unicode BYTE ORDER MARK (C, see L), those
bytes are completely ignored.
=item L scripts autodetected
If a Perl script begins with the Unicode C (UTF-16LE,
UTF16-BE), or if the script looks like non-C-marked
UTF-16 of either endianness, Perl will correctly read in the script as
the appropriate Unicode encoding.
=back
=head2 Byte and Character Semantics
Before Unicode, most encodings used 8 bits (a single byte) to encode
each character. Thus a character was a byte, and a byte was a
character, and there could be only 256 or fewer possible characters.
"Byte Semantics" in the title of this section refers to
this behavior. There was no need to distinguish between "Byte" and
"Character".
Then along comes Unicode which has room for over a million characters
(and Perl allows for even more). This means that a character may
require more than a single byte to represent it, and so the two terms
are no longer equivalent. What matter are the characters as whole
entities, and not usually the bytes that comprise them. That's what the
term "Character Semantics" in the title of this section refers to.
Perl had to change internally to decouple "bytes" from "characters".
It is important that you too change your ideas, if you haven't already,
so that "byte" and "character" no longer mean the same thing in your
mind.
The basic building block of Perl strings has always been a "character".
The changes basically come down to that the implementation no longer
thinks that a character is always just a single byte.
There are various things to note:
=over 4
=item *
String handling functions, for the most part, continue to operate in
terms of characters. C, for example, returns the number of
characters in a string, just as before. But that number no longer is
necessarily the same as the number of bytes in the string (there may be
more bytes than characters). The other such functions include
C, C, C, C, C, C,
C, C, and C.
The exceptions are:
=over 4
=item *
the bit-oriented C
E
=item *
the byte-oriented C/C C format
However, the C specifier does operate on whole characters, as does the
C specifier.
=item *
some operators that interact with the platform's operating system
Operators dealing with filenames are examples.
=item *
when the functions are called from within the scope of the
S>> pragma
Likely, you should use this only for debugging anyway.
=back
=item *
Strings--including hash keys--and regular expression patterns may
contain characters that have ordinal values larger than 255.
If you use a Unicode editor to edit your program, Unicode characters may
occur directly within the literal strings in UTF-8 encoding, or UTF-16.
(The former requires a C, the latter may require a C.)
L gives other ways to place non-ASCII
characters in your strings.
=item *
The C and C functions work on whole characters.
=item *
Regular expressions match whole characters. For example, C matches
a whole character instead of only a single byte.
=item *
The C functionality has been removed. For similar functionality to
that, see C and C).
=item *
C reverses by character rather than by byte.
=item *
The bit string operators, C and (starting in v5.22)
C can operate on bit strings encoded in UTF-8, but this
can give unexpected results if any of the strings contain code points
above 0xFF. Starting in v5.28, it is a fatal error to have such an
operand. Otherwise, the operation is performed on a non-UTF-8 copy of
the operand. If you're not sure about the encoding of a string,
downgrade it before using any of these operators; you can use
L|utf8/Utility functions>.
=back
The bottom line is that Perl has always practiced "Character Semantics",
but with the advent of Unicode, that is now different than "Byte
Semantics".
=head2 ASCII Rules versus Unicode Rules
Before Unicode, when a character was a byte was a character,
Perl knew only about the 128 characters defined by ASCII, code points 0
through 127 (except for under L>|perllocale>). That
left the code
points 128 to 255 as unassigned, and available for whatever use a
program might want. The only semantics they have is their ordinal
numbers, and that they are members of none of the non-negative character
classes. None are considered to match C for example, but all match
C.
Unicode, of course, assigns each of those code points a particular
meaning (along with ones above 255). To preserve backward
compatibility, Perl only uses the Unicode meanings when there is some
indication that Unicode is what is intended; otherwise the non-ASCII
code points remain treated as if they are unassigned.
Here are the ways that Perl knows that a string should be treated as
Unicode:
=over
=item *
Within the scope of S>
If the whole program is Unicode (signified by using 8-bit Bnicode
Bransformation Bormat), then all literal strings within it must be
Unicode.
=item *
Within the scope of
L>|feature/The 'unicode_strings' feature>
This pragma was created so you can explicitly tell Perl that operations
executed within its scope are to use Unicode rules. More operations are
affected with newer perls. See L.
=item *
Within the scope of S> or higher
This implicitly turns on S>.
=item *
Within the scope of
L>|perllocale/Unicode and UTF-8>,
or L>|perllocale> and the current
locale is a UTF-8 locale.
The former is defined to imply Unicode handling; and the latter
indicates a Unicode locale, hence a Unicode interpretation of all
strings within it.
=item *
When the string contains a Unicode-only code point
Perl has never accepted code points above 255 without them being
Unicode, so their use implies Unicode for the whole string.
=item *
When the string contains a Unicode named code point C
The C construct explicitly refers to a Unicode code point,
even if it is one that is also in ASCII. Therefore the string
containing it must be Unicode.
=item *
When the string has come from an external source marked as
Unicode
The L|perlrun/-C [numberElist]> command line option can
specify that certain inputs to the program are Unicode, and the values
of this can be read by your Perl code, see L.
=item * When the string has been upgraded to UTF-8
The function L|utf8/Utility functions>
can be explicitly used to permanently (unless a subsequent
C<:utf8_downgrade> is called) cause a string to be treated as
Unicode.
=item * There are additional methods for regular expression patterns
A pattern that is compiled with the C> or C> modifiers is
treated as Unicode (though there are some restrictions with C>).
Under the C> and C> modifiers, there are several other
indications for Unicode; see L.
=back
Note that all of the above are overridden within the scope of
C>; but you should be using this pragma only for
debugging.
Note also that some interactions with the platform's operating system
never use Unicode rules.
When Unicode rules are in effect:
=over 4
=item *
Case translation operators use the Unicode case translation tables.
Note that C, or C in interpolated strings, translates to
uppercase, while C, or C in interpolated strings,
translates to titlecase in languages that make the distinction (which is
equivalent to uppercase in languages without the distinction).
There is a CPAN module, C>, which allows you to
define your own mappings to be used in C, C, C,
C, and C (or their double-quoted string inlined versions
such as C). (Prior to Perl 5.16, this functionality was partially
provided in the Perl core, but suffered from a number of insurmountable
drawbacks, so the CPAN module was written instead.)
=item *
Character classes in regular expressions match based on the character
properties specified in the Unicode properties database.
C can be used to match a Japanese ideograph, for instance; and
C a Bengali number.
=item *
Named Unicode properties, scripts, and block ranges may be used (like
bracketed character classes) by using the C "matches property"
construct and the C negation, "doesn't match property".
See L"Unicode Character Properties"> for more details.
You can define your own character properties and use them
in the regular expression with the C or C construct.
See L"User-Defined Character Properties"> for more details.
=back
=head2 Extended Grapheme Clusters (Logical characters)
Consider a character, say C. It could appear with various marks around it,
such as an acute accent, or a circumflex, or various hooks, circles, arrows,
I, above, below, to one side or the other, I. There are many
possibilities among the world's languages. The number of combinations is
astronomical, and if there were a character for each combination, it would
soon exhaust Unicode's more than a million possible characters. So Unicode
took a different approach: there is a character for the base C, and a
character for each of the possible marks, and these can be variously combined
to get a final logical character. So a logical character--what appears to be a
single character--can be a sequence of more than one individual characters.
The Unicode standard calls these "extended grapheme clusters" (which
is an improved version of the no-longer much used "grapheme cluster");
Perl furnishes the C regular expression construct to match such
sequences in their entirety.
But Unicode's intent is to unify the existing character set standards and
practices, and several pre-existing standards have single characters that
mean the same thing as some of these combinations, like ISO-8859-1,
which has quite a few of them. For example, C was already in this standard when Unicode came along.
Unicode therefore added it to its repertoire as that single character.
But this character is considered by Unicode to be equivalent to the
sequence consisting of the character C
followed by the character C.
C is called a "pre-composed"
character, and its equivalence with the "E" and the "COMBINING ACCENT"
sequence is called canonical equivalence. All pre-composed characters
are said to have a decomposition (into the equivalent sequence), and the
decomposition type is also called canonical. A string may be comprised
as much as possible of precomposed characters, or it may be comprised of
entirely decomposed characters. Unicode calls these respectively,
"Normalization Form Composed" (NFC) and "Normalization Form Decomposed".
The C> module contains functions that convert
between the two. A string may also have both composed characters and
decomposed characters; this module can be used to make it all one or the
other.
You may be presented with strings in any of these equivalent forms.
There is currently nothing in Perl 5 that ignores the differences. So
you'll have to specially handle it. The usual advice is to convert your
inputs to C before processing further.
For more detailed information, see L of C (letter) property (see
L below). Brackets are not
required for single letter property names, so C is equivalent to C.
More formally, C matches any single character whose Unicode
C property value is C, and C matches any character
whose C property value is C, and they could have been written as
C and C, respectively.
This formality is needed when properties are not binary; that is, if they can
take on more values than just C and C. For example, the
C property (see L"Bidirectional Character Types"> below),
can take on several different
values, such as C, C, C, and others. To match these, one needs
to specify both the property name (C), AND the value being
matched against
(C, C, I). This is done, as in the examples above, by having the
two components separated by an equal sign (or interchangeably, a colon), like
C.
All Unicode-defined character properties may be written in these compound forms
of C=I}> or C:I}>, but Perl provides some
additional properties that are written only in the single form, as well as
single-form short-cuts for all binary properties and certain others described
below, in which you may omit the property name and the equals or colon
separator.
Most Unicode character properties have at least two synonyms (or aliases if you
prefer): a short one that is easier to type and a longer one that is more
descriptive and hence easier to understand. Thus the C and
C properties above are equivalent and can be used
interchangeably. Likewise, C is a synonym for C,
and we could have written C equivalently as C.
Also, there are typically various synonyms for the values the property
can be. For binary properties, C has 3 synonyms: C,
C, and C; and C has correspondingly C,
C, and C. But be careful. A short form of a value for one
property may not mean the same thing as the short form spelled the same
for another.
Thus, for the C> property, C means
C, but for the L|/Bidirectional Character Types>
property, C means C. A complete list of properties and
synonyms is in L.
Upper/lower case differences in property names and values are irrelevant;
thus C means the same thing as C or even C.
Similarly, you can add or subtract underscores anywhere in the middle of a
word, so that these are also equivalent to C. And white space
is generally irrelevant adjacent to non-word characters, such as the
braces and the equals or colon separators, so C and
C are equivalent to these as well. In fact, white
space and even hyphens can usually be added or deleted anywhere. So
even C is equivalent. All this is called
"loose-matching" by Unicode. The "name" property has some restrictions
on this due to a few outlier names. Full details are given in
L.
The few places where stricter matching is
used is in the middle of numbers, the "name" property, and in the Perl
extension properties that begin or end with an underscore. Stricter
matching cares about white space (except adjacent to non-word
characters), hyphens, and non-interior underscores.
You can also use negation in both C and C by introducing a caret
(C) between the first brace and the property name: C is
equal to C.
Almost all properties are immune to case-insensitive matching. That is,
adding a C regular expression modifier does not change what they
match. There are two sets that are affected.
The first set is
C,
C,
and C,
all of which match C under C matching.
And the second set is
C,
C,
and C,
all of which match C under C matching.
This set also includes its subsets C and C both
of which under C match C.
(The difference between these sets is that some things, such as Roman
numerals, come in both upper and lower case so they are C, but
aren't considered letters, so they aren't C's.)
See L for special considerations when
matching Unicode properties against non-Unicode code points.
=head3 B
Every Unicode character is assigned a general category, which is the "most
usual categorization of a character" (from
L).
The compound way of writing these is like C
(short: C). But Perl furnishes shortcuts in which everything up
through the equal or colon separator is omitted. So you can instead just write
C.
Here are the short and long forms of the values the C property
can have:
Short Long
L Letter
LC, L& Cased_Letter (that is: [\p{Ll}\p{Lu}\p{Lt}])
Lu Uppercase_Letter
Ll Lowercase_Letter
Lt Titlecase_Letter
Lm Modifier_Letter
Lo Other_Letter
M Mark
Mn Nonspacing_Mark
Mc Spacing_Mark
Me Enclosing_Mark
N Number
Nd Decimal_Number (also Digit)
Nl Letter_Number
No Other_Number
P Punctuation (also Punct)
Pc Connector_Punctuation
Pd Dash_Punctuation
Ps Open_Punctuation
Pe Close_Punctuation
Pi Initial_Punctuation
(may behave like Ps or Pe depending on usage)
Pf Final_Punctuation
(may behave like Ps or Pe depending on usage)
Po Other_Punctuation
S Symbol
Sm Math_Symbol
Sc Currency_Symbol
Sk Modifier_Symbol
So Other_Symbol
Z Separator
Zs Space_Separator
Zl Line_Separator
Zp Paragraph_Separator
C Other
Cc Control (also Cntrl)
Cf Format
Cs Surrogate
Co Private_Use
Cn Unassigned
Single-letter properties match all characters in any of the
two-letter sub-properties starting with the same letter.
C and C are special: both are aliases for the set consisting of everything matched by C, C, and C.
=head3 B
Because scripts differ in their directionality (Hebrew and Arabic are
written right to left, for example) Unicode supplies a C property.
Some of the values this property can have are:
Value Meaning
L Left-to-Right
LRE Left-to-Right Embedding
LRO Left-to-Right Override
R Right-to-Left
AL Arabic Letter
RLE Right-to-Left Embedding
RLO Right-to-Left Override
PDF Pop Directional Format
EN European Number
ES European Separator
ET European Terminator
AN Arabic Number
CS Common Separator
NSM Non-Spacing Mark
BN Boundary Neutral
B Paragraph Separator
S Segment Separator
WS Whitespace
ON Other Neutrals
This property is always written in the compound form.
For example, C matches characters that are normally
written right to left. Unlike the
C> property, this
property can have more values added in a future Unicode release. Those
listed above comprised the complete set for many Unicode releases, but
others were added in Unicode 6.3; you can always find what the
current ones are in L. And
L
The world's languages are written in many different scripts. This sentence
(unless you're reading it in translation) is written in Latin, while Russian is
written in Cyrillic, and Greek is written in, well, Greek; Japanese mainly in
Hiragana or Katakana. There are many more.
The Unicode C > is C. But because their use is somewhat dangerous,
Perl will warn (using the warning category C, which is a
sub-category of C) if an attempt is made
to do things like take the lower case of one, or match
case-insensitively, or to output them. (But don't try this on Perls
before 5.14.)
=item *
UTF-32, UTF-32BE, UTF-32LE
The UTF-32 family is pretty much like the UTF-16 family, except that
the units are 32-bit, and therefore the surrogate scheme is not
needed. UTF-32 is a fixed-width encoding. The C signatures are
C for BE and C for LE.
=item *
UCS-2, UCS-4
Legacy, fixed-width encodings defined by the ISO 10646 standard. UCS-2 is a 16-bit
encoding. Unlike UTF-16, UCS-2 is not extensible beyond C,
because it does not use surrogates. UCS-4 is a 32-bit encoding,
functionally identical to UTF-32 (the difference being that
UCS-4 forbids neither surrogates nor code points larger than C).
=item *
UTF-7
A seven-bit safe (non-eight-bit) encoding, which is useful if the
transport or storage is not eight-bit safe. Defined by RFC 2152.
=back
=head2 Noncharacter code points
66 code points are set aside in Unicode as "noncharacter code points".
These all have the C (C) C>, and
no character will ever be assigned to any of them. They are the 32 code
points between C and C inclusive, and the 34 code
points:
U+FFFE U+FFFF
U+1FFFE U+1FFFF
U+2FFFE U+2FFFF
...
U+EFFFE U+EFFFF
U+FFFFE U+FFFFF
U+10FFFE U+10FFFF
Until Unicode 7.0, the noncharacters were "B for use in open
interchange of Unicode text data", so that code that processed those
streams could use these code points as sentinels that could be mixed in
with character data, and would always be distinguishable from that data.
(Emphasis above and in the next paragraph are added in this document.)
Unicode 7.0 changed the wording so that they are "B for
use in open interchange of Unicode text data". The 7.0 Standard goes on
to say:
=over 4
"If a noncharacter is received in open interchange, an application is
not required to interpret it in any way. It is good practice, however,
to recognize it as a noncharacter and to take appropriate action, such
as replacing it with C replacement character, to indicate the
problem in the text. It is not recommended to simply delete
noncharacter code points from such text, because of the potential
security issues caused by deleting uninterpreted characters. (See
conformance clause C7 in Section 3.2, Conformance Requirements, and
L)."
=back
This change was made because it was found that various commercial tools
like editors, or for things like source code control, had been written
so that they would not handle program files that used these code points,
effectively precluding their use almost entirely! And that was never
the intent. They've always been meant to be usable within an
application, or cooperating set of applications, at will.
If you're writing code, such as an editor, that is supposed to be able
to handle any Unicode text data, then you shouldn't be using these code
points yourself, and instead allow them in the input. If you need
sentinels, they should instead be something that isn't legal Unicode.
For UTF-8 data, you can use the bytes 0xC1 and 0xC2 as sentinels, as
they never appear in well-formed UTF-8. (There are equivalents for
UTF-EBCDIC). You can also store your Unicode code points in integer
variables and use negative values as sentinels.
If you're not writing such a tool, then whether you accept noncharacters
as input is up to you (though the Standard recommends that you not). If
you do strict input stream checking with Perl, these code points
continue to be forbidden. This is to maintain backward compatibility
(otherwise potential security holes could open up, as an unsuspecting
application that was written assuming the noncharacters would be
filtered out before getting to it, could now, without warning, start
getting them). To do strict checking, you can use the layer
C<:encoding>.
Perl continues to warn (using the warning category C, which
is a sub-category of C) if an attempt is made to output
noncharacters.
=head2 Beyond Unicode code points
The maximum Unicode code point is C, and Unicode only defines
operations on code points up through that. But Perl works on code
points up to the maximum permissible signed number available on the
platform. However, Perl will not accept these from input streams unless
lax rules are being used, and will warn (using the warning category
C, which is a sub-category of C) if any are output.
Since Unicode rules are not defined on these code points, if a
Unicode-defined operation is done on them, Perl uses what we believe are
sensible rules, while generally warning, using the C
category. For example, C will generate such a
warning, returning the input parameter as its result, since Perl defines
the uppercase of every non-Unicode code point to be the code point
itself. (All the case changing operations, not just uppercasing, work
this way.)
The situation with matching Unicode properties in regular expressions,
the C and C constructs, against these code points is not as
clear cut, and how these are handled has changed as we've gained
experience.
One possibility is to treat any match against these code points as
undefined. But since Perl doesn't have the concept of a match being
undefined, it converts this to failing or C. This is almost, but
not quite, what Perl did from v5.14 (when use of these code points
became generally reliable) through v5.18. The difference is that Perl
treated all C matches as failing, but all C matches as
succeeding.
One problem with this is that it leads to unexpected, and confusing
results in some cases:
chr(0x110000) =~ \p{ASCII_Hex_Digit=True} # Failed on for
C, and C for C. For example:
chr(0x110000) =~ \p{ASCII_Hex_Digit=True} # Fails, no warning
chr(0x110000) =~ \p{ASCII_Hex_Digit=False} # Succeeds, with warning
In both these examples, the character being matched is non-Unicode, so
Unicode doesn't define how it should match. It clearly isn't an ASCII
hex digit, so the first example clearly should fail, and so it does,
with no warning. But it is arguable that the second example should have
an undefined, hence C, result. So a warning is raised for it.
Thus the warning is raised for many fewer cases than in earlier Perls,
and only when what the result is could be arguable. It turns out that
none of the optimizations made by Perl (or are ever likely to be made)
cause the warning to be skipped, so it solves both problems of Perl's
earlier approach. The most commonly used property that is affected by
this change is C which is a short form for
C. Starting in v5.20, all non-Unicode
code points are considered C. In earlier releases the
matches failed because the result was considered undefined.
The only place where the warning is not raised when it might ought to
have been is if optimizations cause the whole pattern match to not even
be attempted. For example, Perl may figure out that for a string to
match a certain regular expression pattern, the string has to contain
the substring C. Before attempting the match, Perl may look
for that substring, and if not found, immediately fail the match without
actually trying it; so no warning gets generated even if the string
contains an above-Unicode code point.
This behavior is more "Do what I mean" than in earlier Perls for most
applications. But it catches fewer issues for code that needs to be
strictly Unicode compliant. Therefore there is an additional mode of
operation available to accommodate such code. This mode is enabled if a
regular expression pattern is compiled within the lexical scope where
the C warning class has been made fatal, say by:
use warnings FATAL => "non_unicode"
(see L). In this mode of operation, Perl will raise the
warning for all matches against a non-Unicode code point (not just the
arguable ones), and it skips the optimizations that might cause the
warning to not be output. (It currently still won't warn if the match
isn't even attempted, like in the C example above.)
In summary, Perl now normally treats non-Unicode code points as typical
Unicode unassigned code points for regular expression matches, raising a
warning only when it is arguable what the result should be. However, if
this warning has been made fatal, it isn't skipped.
There is one exception to all this. C looks like a Unicode
property, but it is a Perl extension that is defined to be true for all
possible code points, Unicode or not, so no warning is ever generated
when matching this against a non-Unicode code point. (Prior to v5.20,
it was an exact synonym for C, matching code points C
through C.)
=head2 Security Implications of Unicode
First, read
L.
Also, note the following:
=over 4
=item *
Malformed UTF-8
UTF-8 is very structured, so many combinations of bytes are invalid. In
the past, Perl tried to soldier on and make some sense of invalid
combinations, but this can lead to security holes, so now, if the Perl
core needs to process an invalid combination, it will either raise a
fatal error, or will replace those bytes by the sequence that forms the
Unicode REPLACEMENT CHARACTER, for which purpose Unicode created it.
Every code point can be represented by more than one possible
syntactically valid UTF-8 sequence. Early on, both Unicode and Perl
considered any of these to be valid, but now, all sequences longer
than the shortest possible one are considered to be malformed.
Unicode considers many code points to be illegal, or to be avoided.
Perl generally accepts them, once they have passed through any input
filters that may try to exclude them. These have been discussed above
(see "Surrogates" under UTF-16 in L ,
L, and L).
=item *
Regular expression pattern matching may surprise you if you're not
accustomed to Unicode. Starting in Perl 5.14, several pattern
modifiers are available to control this, called the character set
modifiers. Details are given in L.
=back
As discussed elsewhere, Perl has one foot (two hooves?) planted in
each of two worlds: the old world of ASCII and single-byte locales, and
the new world of Unicode, upgrading when necessary.
If your legacy code does not explicitly use Unicode, no automatic
switch-over to Unicode should happen.
=head2 Unicode in Perl on EBCDIC
Unicode is supported on EBCDIC platforms. See L.
Unless ASCII vs. EBCDIC issues are specifically being discussed,
references to UTF-8 encoding in this document and elsewhere should be
read as meaning UTF-EBCDIC on EBCDIC platforms.
See L.
Because UTF-EBCDIC is so similar to UTF-8, the differences are mostly
hidden from you; S> (and NOT something like
S>) declares the script is in the platform's
"native" 8-bit encoding of Unicode. (Similarly for the C
layer.)
=head2 Locales
See L
=head2 When Unicode Does Not Happen
There are still many places where Unicode (in some encoding or
another) could be given as arguments or received as results, or both in
Perl, but it is not, in spite of Perl having extensive ways to input and
output in Unicode, and a few other "entry points" like the C
array (which can sometimes be interpreted as UTF-8).
The following are such interfaces. Also, see L.
For all of these interfaces Perl
currently (as of v5.16.0) simply assumes byte strings both as arguments
and results, or UTF-8 strings if the (deprecated) C pragma has been used.
One reason that Perl does not attempt to resolve the role of Unicode in
these situations is that the answers are highly dependent on the operating
system and the file system(s). For example, whether filenames can be
in Unicode and in exactly what kind of encoding, is not exactly a
portable concept. Similarly for C and C: how well will the
"command-line interface" (and which of them?) handle Unicode?
=over 4
=item *
C, C, C, C, C, C, C,
C, C, C, C, C, C, C, C
=item *
C
=item *
C (aka the C*E>)
=item *
C, C, C
=item *
C (aka the backtick operator), C
=item *
C, C
=back
=head2 The "Unicode Bug"
The term, "Unicode bug" has been applied to an inconsistency with the
code points in the C block, that is, between
128 and 255. Without a locale specified, unlike all other characters or
code points, these characters can have very different semantics
depending on the rules in effect. (Characters whose code points are
above 255 force Unicode rules; whereas the rules for ASCII characters
are the same under both ASCII and Unicode rules.)
Under Unicode rules, these upper-Latin1 characters are interpreted as
Unicode code points, which means they have the same semantics as Latin-1
(ISO-8859-1) and C1 controls.
As explained in L, under ASCII rules,
they are considered to be unassigned characters.
This can lead to unexpected results. For example, a string's
semantics can suddenly change if a code point above 255 is appended to
it, which changes the rules from ASCII to Unicode. As an
example, consider the following program and its output:
$ perl -le'
no feature "unicode_strings";
$s1 = "\xC2";
$s2 = "\x{2660}";
for ($s1, $s2, $s1.$s2) {
print /\w/ || 0;
}
'
0
0
1
If there's no C in C nor in C, why does their concatenation
have one?
This anomaly stems from Perl's attempt to not disturb older programs that
didn't use Unicode, along with Perl's desire to add Unicode support
seamlessly. But the result turned out to not be seamless. (By the way,
you can choose to be warned when things like this happen. See
C>.)
L>|feature/The 'unicode_strings' feature>
was added, starting in Perl v5.12, to address this problem. It affects
these things:
=over 4
=item *
Changing the case of a scalar, that is, using C, C, C,
and C, or C, C, C and C in double-quotish
contexts, such as regular expression substitutions.
Under C starting in Perl 5.12.0, Unicode rules are
generally used. See L for details on how this works
in combination with various other pragmas.
=item *
Using caseless (C) regular expression matching.
Starting in Perl 5.14.0, regular expressions compiled within
the scope of C use Unicode rules
even when executed or compiled into larger
regular expressions outside the scope.
=item *
Matching any of several properties in regular expressions.
These properties are C (without braces), C (without braces),
C, C, C, C, and all the Posix character classes
I C.
Starting in Perl 5.14.0, regular expressions compiled within
the scope of C use Unicode rules
even when executed or compiled into larger
regular expressions outside the scope.
=item *
In C or its inline equivalent C.
Starting in Perl 5.16.0, consistent quoting rules are used within the
scope of C, as described in L.
Prior to that, or outside its scope, no code points above 127 are quoted
in UTF-8 encoded strings, but in byte encoded strings, code points
between 128-255 are always quoted.
=item *
In the C<..> or L operator.
Starting in Perl 5.26.0, the range operator on strings treats their lengths
consistently within the scope of C. Prior to that, or
outside its scope, it could produce strings whose length in characters
exceeded that of the right-hand side, where the right-hand side took up more
bytes than the correct range endpoint.
=item *
In L's special-case whitespace splitting|perlfunc/split >>.
Starting in Perl 5.28.0, the C function with a pattern specified as
a string containing a single space handles whitespace characters consistently
within the scope of C. Prior to that, or outside its scope,
characters that are whitespace according to Unicode rules but not according to
ASCII rules were treated as field contents rather than field separators when
they appear in byte-encoded strings.
=back
You can see from the above that the effect of C
increased over several Perl releases. (And Perl's support for Unicode
continues to improve; it's best to use the latest available release in
order to get the most complete and accurate results possible.) Note that
C is automatically chosen if you S> or
higher.
For Perls earlier than those described above, or when a string is passed
to a function outside the scope of C, see the next section.
=head2 Forcing Unicode in Perl (Or Unforcing Unicode in Perl)
Sometimes (see L"When Unicode Does Not Happen"> or L)
there are situations where you simply need to force a byte
string into UTF-8, or vice versa. The standard module L can be
used for this, or the low-level calls
L|utf8/Utility functions> and
L|utf8/Utility functions>.
Note that C<:downgrade> can fail if the string contains characters
that don't fit into a byte.
Calling either function on a string that already is in the desired state is a
no-op.
L gives all the ways that a string is
made to use Unicode rules.
=head2 Using Unicode in XS
See L for an introduction to Unicode at
the XS level, and L for the API details.
=head2 Hacking Perl to work on earlier Unicode versions (for very serious hackers only)
Perl by default comes with the latest supported Unicode version built-in, but
the goal is to allow you to change to use any earlier one. In Perls
v5.20 and v5.22, however, the earliest usable version is Unicode 5.1.
Perl v5.18 and v5.24 are able to handle all earlier versions.
Download the files in the desired version of Unicode from the Unicode web
site L). These should replace the existing files in
F in the Perl source tree. Follow the instructions in
F in that directory to change some of their names, and then build
perl (see L).
=head2 Porting code from perl-5.6.X
Perls starting in 5.8 have a different Unicode model from 5.6. In 5.6 the
programmer was required to use the C pragma to declare that a
given scope expected to deal with Unicode data and had to make sure that
only Unicode data were reaching that scope. If you have code that is
working with 5.6, you will need some of the following adjustments to
your code. The examples are written such that the code will continue to
work under 5.6, so you should be safe to try them out.
=over 3
=item *
A filehandle that should read or write UTF-8
if ($] > 5.008) {
binmode $fh, ":encoding(UTF-8)";
}
=item *
A scalar that is going to be passed to some extension
Be it C<:zlib>, C<:request> or any extension that has no
mention of Unicode in the manpage, you need to make sure that the
UTF8 flag is stripped off. Note that at the time of this writing
(January 2012) the mentioned modules are not UTF-8-aware. Please
check the documentation to verify if this is still true.
if ($] > 5.008) {
require Encode;
$val = Encode::encode("UTF-8", $val); # make octets
}
=item *
A scalar we got back from an extension
If you believe the scalar comes back as UTF-8, you will most likely
want the UTF8 flag restored:
if ($] > 5.008) {
require Encode;
$val = Encode::decode("UTF-8", $val);
}
=item *
Same thing, if you are really sure it is UTF-8
if ($] > 5.008) {
require Encode;
Encode::_utf8_on($val);
}
=item *
A wrapper for L C and C
When the database contains only UTF-8, a wrapper function or method is
a convenient way to replace all your C and
C calls. A wrapper function will also make it easier to
adapt to future enhancements in your database driver. Note that at the
time of this writing (January 2012), the DBI has no standardized way
to deal with UTF-8 data. Please check the L to verify if
that is still true.
sub fetchrow {
# $what is one of fetchrow_{array,hashref}
my($self, $sth, $what) = @_;
if ($] $what;
} else {
require Encode;
if (wantarray) {
my @arr = $sth->$what;
for (@arr) {
defined && /[^\000-\177]/ && Encode::_utf8_on($_);
}
return @arr;
} else {
my $ret = $sth->$what;
if (ref $ret) {
for my $k (keys %$ret) {
defined
&& /[^\000-\177]/
&& Encode::_utf8_on($_) for $ret->{$k};
}
return $ret;
} else {
defined && /[^\000-\177]/ && Encode::_utf8_on($_) for $ret;
return $ret;
}
}
}
}
=item *
A large scalar that you know can only contain ASCII
Scalars that contain only ASCII and are marked as UTF-8 are sometimes
a drag to your program. If you recognize such a situation, just remove
the UTF8 flag:
utf8::downgrade($val) if $] > 5.008;
=back
=head1 BUGS
See also L above.
=head2 Interaction with Extensions
When Perl exchanges data with an extension, the extension should be
able to understand the UTF8 flag and act accordingly. If the
extension doesn't recognize that flag, it's likely that the extension
will return incorrectly-flagged data.
So if you're working with Unicode data, consult the documentation of
every module you're using if there are any issues with Unicode data
exchange. If the documentation does not talk about Unicode at all,
suspect the worst and probably look at the source to learn how the
module is implemented. Modules written completely in Perl shouldn't
cause problems. Modules that directly or indirectly access code written
in other programming languages are at risk.
For affected functions, the simple strategy to avoid data corruption is
to always make the encoding of the exchanged data explicit. Choose an
encoding that you know the extension can handle. Convert arguments passed
to the extensions to that encoding and convert results back from that
encoding. Write wrapper functions that do the conversions for you, so
you can later change the functions when the extension catches up.
To provide an example, let's say the popular C<:bar::escape_html>
function doesn't deal with Unicode data yet. The wrapper function
would convert the argument to raw UTF-8 and convert the result back to
Perl's internal representation like so:
sub my_escape_html ($) {
my($what) = shift;
return unless defined $what;
Encode::decode("UTF-8", Foo::Bar::escape_html(
Encode::encode("UTF-8", $what)));
}
Sometimes, when the extension does not convert data but just stores
and retrieves it, you will be able to use the otherwise
dangerous L|Encode/_utf8_on> function. Let's say
the popular C<:bar> extension, written in C, provides a C, C or C, or matching
regular expressions can work B faster when the underlying data are
byte-encoded.
In Perl 5.8.0 the slowness was often quite spectacular; in Perl 5.8.1
a caching scheme was introduced which improved the situation. In general,
operations with UTF-8 encoded strings are still slower. As an example,
the Unicode properties (character classes) like C are known to
be quite a bit slower (5-20 times) than their simpler counterparts
like C (then again, there are hundreds of Unicode characters matching
C compared with the 10 ASCII characters matching C).
=head1 SEE ALSO
L, L, L, L, L, L, L,
L, L,
L).
=cut