1 package Encode::CN::HZ;
6 #$VERSION = do { my @r = (q$Revision: 1.5 $ =~ /\d+/g); sprintf "%d."."%02d" x $#r, @r };
9 use Encode qw(:fallbacks);
11 use base qw(Encode::Encoding);
12 __PACKAGE__->Define('hz');
14 # HZ is a combination of ASCII and escaped GB, so we implement it
15 # with the GB2312(raw) encoding here. Cf. RFCs 1842 & 1843.
17 # not ported for EBCDIC. Which should be used, "~" or "\x7E"?
23 my ($obj,$str,$chk) = @_;
25 my $GB = Encode::find_encoding('gb2312-raw');
27 my $in_ascii = 1; # default mode is ASCII.
30 if ($in_ascii) { # ASCII mode
31 if ($str =~ s/^([\x00-\x7D\x7F]+)//) { # no '~' => ASCII
33 # EBCDIC should need ascii2native, but not ported.
35 elsif ($str =~ s/^\x7E\x7E//) { # escaped tilde
38 elsif ($str =~ s/^\x7E\cJ//) { # '\cJ' == LF in ASCII
41 elsif ($str =~ s/^\x7E\x7B//) { # '~{'
42 $in_ascii = 0; # to GB
44 else { # encounters an invalid escape, \x80 or greater
48 else { # GB mode; the byte ranges are as in RFC 1843.
49 if ($str =~ s/^((?:[\x21-\x77][\x21-\x7E])+)//) {
50 $ret .= $GB->decode($1, $chk);
52 elsif ($str =~ s/^\x7E\x7D//) { # '~}'
60 $_[1] = '' if $chk; # needs_lines guarantees no partial character
65 my ($obj, undef, $src, $pos, $trm, $chk) = @_;
66 my ($rdst, $rsrc, $rpos) = \@_[1..3];
68 my $GB = Encode::find_encoding('gb2312-raw');
70 my $in_ascii = 1; # default mode is ASCII.
72 my $ini_pos = pos($$rsrc);
74 substr($src, 0, $pos) = '';
76 my $ini_len = bytes::length($src);
78 # $trm is the first of the pair '~~', then 2nd tilde is to be removed.
79 # XXX: Is better C<$src =~ s/^\x7E// or die if ...>?
80 $src =~ s/^\x7E// if $trm eq "\x7E";
84 if ($in_ascii) { # ASCII mode
85 if ($src =~ s/^([\x00-\x7D\x7F])//) { # no '~' => ASCII
88 elsif ($src =~ s/^\x7E\x7E//) { # escaped tilde
91 elsif ($src =~ s/^\x7E\cJ//) { # '\cJ' == LF in ASCII
94 elsif ($src =~ s/^\x7E\x7B//) { # '~{'
95 $in_ascii = 0; # to GB
98 else { # encounters an invalid escape, \x80 or greater
102 else { # GB mode; the byte ranges are as in RFC 1843.
103 if ($src =~ s/^((?:[\x21-\x77][\x21-\x7F])+)//) {
104 $now = $GB->decode($1, $chk);
106 elsif ($src =~ s/^\x7E\x7D//) { # '~}'
115 next if ! defined $now;
121 $$rpos = $ini_pos + $pos + $ini_len - bytes::length($src);
122 pos($$rsrc) = $ini_pos;
128 $$rpos = $ini_pos + $pos + $ini_len - bytes::length($src);
129 pos($$rsrc) = $ini_pos;
130 return ''; # terminator not found
136 my ($obj,$str,$chk) = @_;
138 my $GB = Encode::find_encoding('gb2312-raw');
140 my $in_ascii = 1; # default mode is ASCII.
142 no warnings 'utf8'; # $str may be malformed UTF8 at the end of a chunk.
144 while (length $str) {
145 if ($str =~ s/^([[:ascii:]]+)//) {
147 $tmp =~ s/~/~~/g; # escapes tildes
149 $ret .= "\x7E\x7D"; # '~}'
152 $ret .= pack 'a*', $tmp; # remove UTF8 flag.
154 elsif ($str =~ s/(.)//) {
155 my $tmp = $GB->encode($1, $chk);
156 last if !defined $tmp;
157 if (length $tmp == 2) { # maybe a valid GB char (XXX)
159 $ret .= "\x7E\x7B"; # '~{'
164 elsif (length $tmp) { # maybe FALLBACK in ASCII (XXX)
166 $ret .= "\x7E\x7D"; # '~}'
172 else { # if $str is malformed UTF8 *and* if length $str != 0.
176 $_[1] = $str if $chk;
178 # The state at the end of the chunk is discarded, even if in GB mode.
179 # That results in the combination of GB-OUT and GB-IN, i.e. "~}~{".
180 # Parhaps it is harmless, but further investigations may be required...
183 $ret .= "\x7E\x7D"; # '~}'
194 Encode::CN::HZ -- internally used by Encode::CN