Commit | Line | Data |
---|---|---|
920dae64 AT |
1 | package bytes; |
2 | ||
3 | our $VERSION = '1.02'; | |
4 | ||
5 | $bytes::hint_bits = 0x00000008; | |
6 | ||
7 | sub import { | |
8 | $^H |= $bytes::hint_bits; | |
9 | } | |
10 | ||
11 | sub unimport { | |
12 | $^H &= ~$bytes::hint_bits; | |
13 | } | |
14 | ||
15 | sub AUTOLOAD { | |
16 | require "bytes_heavy.pl"; | |
17 | goto &$AUTOLOAD if defined &$AUTOLOAD; | |
18 | require Carp; | |
19 | Carp::croak("Undefined subroutine $AUTOLOAD called"); | |
20 | } | |
21 | ||
22 | sub length ($); | |
23 | sub chr ($); | |
24 | sub ord ($); | |
25 | sub substr ($$;$$); | |
26 | sub index ($$;$); | |
27 | sub rindex ($$;$); | |
28 | ||
29 | 1; | |
30 | __END__ | |
31 | ||
32 | =head1 NAME | |
33 | ||
34 | bytes - Perl pragma to force byte semantics rather than character semantics | |
35 | ||
36 | =head1 SYNOPSIS | |
37 | ||
38 | use bytes; | |
39 | ... chr(...); # or bytes::chr | |
40 | ... index(...); # or bytes::index | |
41 | ... length(...); # or bytes::length | |
42 | ... ord(...); # or bytes::ord | |
43 | ... rindex(...); # or bytes::rindex | |
44 | ... substr(...); # or bytes::substr | |
45 | no bytes; | |
46 | ||
47 | ||
48 | =head1 DESCRIPTION | |
49 | ||
50 | The C<use bytes> pragma disables character semantics for the rest of the | |
51 | lexical scope in which it appears. C<no bytes> can be used to reverse | |
52 | the effect of C<use bytes> within the current lexical scope. | |
53 | ||
54 | Perl normally assumes character semantics in the presence of character | |
55 | data (i.e. data that has come from a source that has been marked as | |
56 | being of a particular character encoding). When C<use bytes> is in | |
57 | effect, the encoding is temporarily ignored, and each string is treated | |
58 | as a series of bytes. | |
59 | ||
60 | As an example, when Perl sees C<$x = chr(400)>, it encodes the character | |
61 | in UTF-8 and stores it in $x. Then it is marked as character data, so, | |
62 | for instance, C<length $x> returns C<1>. However, in the scope of the | |
63 | C<bytes> pragma, $x is treated as a series of bytes - the bytes that make | |
64 | up the UTF8 encoding - and C<length $x> returns C<2>: | |
65 | ||
66 | $x = chr(400); | |
67 | print "Length is ", length $x, "\n"; # "Length is 1" | |
68 | printf "Contents are %vd\n", $x; # "Contents are 400" | |
69 | { | |
70 | use bytes; # or "require bytes; bytes::length()" | |
71 | print "Length is ", length $x, "\n"; # "Length is 2" | |
72 | printf "Contents are %vd\n", $x; # "Contents are 198.144" | |
73 | } | |
74 | ||
75 | chr(), ord(), substr(), index() and rindex() behave similarly. | |
76 | ||
77 | For more on the implications and differences between character | |
78 | semantics and byte semantics, see L<perluniintro> and L<perlunicode>. | |
79 | ||
80 | =head1 LIMITATIONS | |
81 | ||
82 | bytes::substr() does not work as an lvalue(). | |
83 | ||
84 | =head1 SEE ALSO | |
85 | ||
86 | L<perluniintro>, L<perlunicode>, L<utf8> | |
87 | ||
88 | =cut |