NAME

String::Multibyte - manipulation of multibyte character strings

SYNOPSIS

    use String::Multibyte;

    $utf8 = String::Multibyte->new('UTF8');
    $utf8_len = $utf8->length($utf8_str);

DESCRIPTION

This module provides some functions which emulate the corresponding CORE functions for locale-independent manipulation of multiple-byte character strings.

Why this module is locale-independent? Well, because this module only consider the byte sequence structure of charsets and is not aware of any Locale stuff! Locale-dependent methods like uc(), lc(), etc., will not be supported at all.

Definition of Multibyte Charsets

The definition files are sited under the directory where String::Multibyte is sited. E.g. if String::Multibyte is perl/site/lib/String/Multibyte.pm, copy String::Multibyte::Foo as perl/site/lib/String/Multibyte/Foo.pm.

The definition file must return a hashref, having key(s) named as following.

charset

The value for the key 'charset' stands for a string of the charset name. In almost case, omission of the 'charset' matters very little, but keep them not conflict among another charset.

regexp

The value for the key 'regexp', REQUIRED, is a regular expression that matchs a single character of charset in question. (You may use qr// if available.)

If the 'regexp' is omitted, calling any method is croaked.

nextchar

The value for the key 'nextchar' must be a coderef that returns the next character to the specified character. If the 'nextchar' coderef is omitted, mkrange() and strtr() methods don't understand hyphen metacharacter for character ranges.

cmpchar

The value for the key 'cmpchar' must be a coderef that compares the specified two characters. If the 'cmpchar' coderef is omitted, mkrange and strtr functions don't understand reverse character ranges.

hyphen

The value for the key 'hyphen' is a character to stand for a character range. The default is '-'.

escape

The value for the key 'escape' is an escape character for a hyphen character. The default is '\\'. The 'escape' character is valid only before a hyphen or another 'escape' (e.g. '\\\\-]' means '\\' to ']'; '\\\\\-]' means '\\', '-', and ']'). If an 'escape' character is followed by any character other than 'escape' or 'hyphen', it is parsed literally.

Constructor

$mbcs = String::Multibyte->new(CHARSET)
$mbcs = String::Multibyte->new(CHARSET, VERBOSE)

CHARSET is the charset name; exactly speaking, the file name of the definition file (without the suffix .pm). It returns the instance to tell methods in which charset the specified strings should be handled.

CHARSET may be a hashref; this is how to define a charset without any .pm file.

    # see perlfaq6  :-)
    my $martian  = String::Multibyte->new({
        charset => "martian",
        regexp => '[A-Z][A-Z]|[^A-Z]',
    });

If true value is specified as VERBOSE, the called method (excepting islegal) will check its arguments and carps if any of them is not legally encoded.

Otherwise such a check won't be carried out (saves a bit of time, but unsafe, though you can use the islegal method if necessary).

$mbcs->islegal(LIST)

Returns a boolean indicating whether all the strings in arguments are legally encoded in the concerned charset. Returns false even if one element is illegal in LIST.

Length

$mbcs->length(STRING)

Returns the length in characters of the specified string.

Reverse

$mbcs->strrev(STRING)

Returns a reversed string in characters.

$mbcs->index(STRING, SUBSTR)
$mbcs->index(STRING, SUBSTR, POSITION)

Returns the position of the first occurrence of SUBSTR in STRING at or after POSITION. If POSITION is omitted, starts searching from the beginning of the string.

If the substring is not found, returns -1.

$mbcs->rindex(STRING, SUBSTR)
$mbcs->rindex(STRING, SUBSTR, POSITION)

Returns the position of the last occurrence of SUBSTR in STRING at or after POSITION. If POSITION is specified, returns the last occurrence at or before that position.

If the substring is not found, returns -1.

$mbcs->strspn(STRING, SEARCHLIST)

Returns returns the position of the first occurrence of any character not contained in the search list.

  $mbcs->strspn("+0.12345*12", "+-.0123456789");
  # returns 8.

If the specified string does not contain any character in the search list, returns 0.

The string consists of characters in the search list, the returned value equals the length of the string.

SEARCHLIST can be an ARRAYREF. e.g. if a charset treats CRLF as a single character, "\r\n" is a one-element list of only "\r\n". A two-element list of "\r" and "\n" can be given as ["\r", "\n"] (of course "\n\r" is also ok since the character order of SEARCHLIST doesn't matter in strspn).

$mbcs->strcspn(STRING, SEARCHLIST)

Returns returns the position of the first occurrence of any character contained in the search list.

If the specified string does not contain any character in the search list, the returned value equals the length of the string.

SEARCHLIST can be an ARRAYREF. e.g. if a charset treats CRLF as a single character, "\r\n" is a one-element list of only "\r\n". A two-element list of "\r" and "\n" can be given as ["\r", "\n"] (of course "\n\r" is also ok since the character order of SEARCHLIST doesn't matter in strcspn).

Substring

$mbcs->substr(STRING or SCALAR REF, OFFSET)
$mbcs->substr(STRING or SCALAR REF, OFFSET, LENGTH)
$mbcs->substr(SCALAR, OFFSET, LENGTH, REPLACEMENT)

It works like CORE::substr, but using character semantics of multibyte charset encoding.

If the REPLACEMENT as the fourth argument is specified, replaces parts of the SCALAR and returns what was there before.

You can utilize the lvalue reference, returned if a reference of scalar variable is used as the first argument.

    ${ $mbcs->substr(\$str,$off,$len) } = $replace;

        works like

    CORE::substr($str,$off,$len) = $replace;

The returned lvalue is not multibyte-aware, then successive assignment may lead to odd results.

Split

$mbcs->strsplit(SEPARATOR, STRING)
$mbcs->strsplit(SEPARATOR, STRING, LIMIT)

This function emulates CORE::split, but splits on the SEPARATOR string, not by a pattern.

If not in list context, only return the number of fields found, but does not split into the @_ array.

If empty string is specified as SEPARATOR, splits the specified string into characters.

  $bytes->strsplit('', 'This is perl.', 7);
  # ('T', 'h', 'i', 's', ' ', 'i',  's perl.')

Character Range

$mbcs->mkrange(CHARLIST, ALLOW_REVERSE)

Returns the character list (not in list context, as a concatenated string) gained by parsing the specified character range.

The result depends on the the character order for the concerned charset. About the character order for each charset, see its definition file.

If the character order is undefined in the definition file, returns an identical string with the specified string.

A character range is specified with a hyphen ('-', but exactly speaking, $obj->{hyphen}).

The backslashed combinations '\-' and '\\' (exactly speaking, "$obj->{escape}$obj->{hyphen}" and "$obj->{escape}$obj->{escape}") are used instead of the characters '-' and '\', respectively. The hyphen at the beginning or the end of the range is also evaluated as the hyphen itself.

For example, $mbcs->mkrange('+\-0-9A-F') returns ('+', '-', '0', '1', '2', '3', '4', '5', '6', '7', '8', '9', 'A', 'B', 'C', 'D', 'E', 'F') and scalar $mbcs->mkrange('A-P') returns 'ABCDEFGHIJKLMNOP'.

If true value is specified as the second argument, reverse character ranges such as '9-0', 'Z-A' are allowed.

  $bytes = String::Multibyte->new('Bytes');
  $bytes->mkrange('p-e-r-l', 1); # ponmlkjihgfefghijklmnopqrqponml

Transliteration

$mbcs->strtr(STRING or SCALAR REF, SEARCHLIST, REPLACEMENTLIST)
$mbcs->strtr(STRING or SCALAR REF, SEARCHLIST, REPLACEMENTLIST, MODIFIER)

Transliterates all occurrences of the characters found in the search list with the corresponding character in the replacement list.

If a reference of scalar variable is specified as the first argument, returns the number of characters replaced or deleted; otherwise, returns the transliterated string and the specified string is unaffected.

If 'h' modifier is specified, returns a hash of histogram in list context; a reference to hash of histogram in scalar context;

SEARCHLIST and REPLACEMENTLIST

Character ranges (internally utilizing mkrange()) are supported.

If the REPLACEMENTLIST is empty (specified as '', not undef, because the use of uninitialized value causes warning under -w option), the SEARCHLIST is replicated.

If the replacement list is shorter than the search list, the final character in the replacement list is replicated till it is long enough (but differently works when the 'd' modifier is used).

SEARCHLIST and REPLACEMENTLIST can be an ARRAYREF. e.g. if a charset treats "\r\n" (CRLF) as a single character, "\r\n" is a one-element list of only "\r\n". A two-element list of "\r" and "\n" should be given as ["\r", "\n"]. Of course "\n\r" is also ok but the character order is different; cf. strtr($str, ["\r", "\n"], ["\n", "\r"]) that swaps "\n" and "\r".

Each elements of ARRAYREF can include character ranges (the modifiers R and r affect their evaluation as usual).

["A-C", "h-z"] is evaluated like "A-Ch-z" if charset does not include grapheme "Ch". The former prevents "C" and "h" from evaluation as "Ch" even if the charset included grapheme "Ch".

MODIFIER

    c   Complement the SEARCHLIST.
    d   Delete found but unreplaced characters.
    s   Squash duplicate replaced characters.
    h   Return a hash (or a hashref) of histogram.
    R   No use of character ranges.
    r   Allows to use reverse character ranges.
    o   Caches the conversion table internally.

If 'R' modifier is specified, '-' is not evaluated as a meta character but hyphen itself like in tr'''. Compare:

  $mbcs->strtr("90 - 32 = 58", "0-9", "A-J");
    # output: "JA - DC = FI"

  $mbcs->strtr("90 - 32 = 58", "0-9", "A-J", "R");
    # output: "JA - 32 = 58"
    # cf. ($str = "90 - 32 = 58") =~ tr'0-9'A-J';
    # '0' to 'A', '-' to '-', and '9' to 'J'.

If 'r' modifier is specified, reverse character ranges are allowed. e.g.

   $mbcs->strtr($str, "0-9", "9-0", "r")

     is equivalent to

   $mbcs->strtr($str, "0123456789", "9876543210")

Caching the conversion table

If 'o' modifier is specified, the conversion table is cached internally. e.g.

  foreach (@source_strings) {
    print $mbcs->strtr($_, $from_list, $to_list, 'o');
  }

will be almost as efficient as this:

  $trans = $mbcs->trclosure($from_list, $to_list);

  foreach (@source_strings) {
    print &$trans($_);
  }

You can use whichever you like.

Without 'o',

  foreach (@source_strings) {
    print $mbcs->strtr($_, $from_list, $to_list);
  }

will be very slow since the conversion table is made whenever the function is called.

Generation of the Closure to Transliterate

$mbcs->trclosure(SEARCHLIST, REPLACEMENTLIST)
$mbcs->trclosure(SEARCHLIST, REPLACEMENTLIST, MODIFIER)

Returns a closure to transliterate the specified string. The return value is an only code reference, not blessed object. By use of this code ref, you can save yourself time as you need not specify arguments every time.

  my $trans = $mbcs->trclosure($from_list, $to_list);
  print &$trans ($string); # ok to perl 5.003
  print $trans->($string); # perl 5.004 or better

The functionality of the closure made by trclosure() is equivalent to that of strtr(). Frankly speaking, the strtr() calls trclosure() internally and uses the returned closure.

SEARCHLIST and REPLACEMENTLIST can be an ARRAYREF same as strtr().

CAVEAT

$[

This modules supposes $[ is always equal to 0, never 1.

Grapheme manipulation

Since v. 1.01, manipulation of sequence of graphemes is to be supported.

In a grapheme-aware manipulation, notice that the beginning and the end of a string always lie on a grapheme boundary.

E.g. imagine a grapheme set where a grapheme comprises either a leading latin capital letter followed by one or more latin small letters, or a single byte. Such a set can be define as below.

   $gra = String::Multibyte->new({
         regexp => '[A-Z][a-z]*|[\x00-\xFF]',
      });

Think about $gra->index("Perl", "Pe"). As both "Perl" and "Pe" are a single grapheme, they are not equal to each other. So the result of this must be -1 (meaning no match).

AUTHOR

SADAHIRO Tomoyuki <SADAHIRO@cpan.org>

Copyright(C) 2001-2015, SADAHIRO Tomoyuki. Japan. All rights reserved.

This module is free software; you can redistribute it and/or modify it under the same terms as Perl itself.

SEE ALSO

perl(1).