Upgrade to Encode 2.00.

[p5sagit/p5-mst-13.2.git] / ext / Encode / lib / Encode / JP / JIS7.pm

diff --git a/ext/Encode/lib/Encode/JP/JIS7.pm b/ext/Encode/lib/Encode/JP/JIS7.pm
index 8cc40ca..28503ec 100644 (file)
--- a/ext/Encode/lib/Encode/JP/JIS7.pm
+++ b/ext/Encode/lib/Encode/JP/JIS7.pm
@@ -1,14 +1,15 @@
 package Encode::JP::JIS7;
 use strict;
 
-our $VERSION = do { my @r = (q$Revision: 1.1 $ =~ /\d+/g); sprintf "%d."."%02d" x $#r, @r };
+our $VERSION = do { my @r = (q$Revision: 2.0 $ =~ /\d+/g); sprintf "%d."."%02d" x $#r, @r };
+
+use Encode qw(:fallbacks);
 
-require Encode;
 for my $name ('7bit-jis', 'iso-2022-jp', 'iso-2022-jp-1'){
     my $h2z     = ($name eq '7bit-jis')    ? 0 : 1;
     my $jis0212 = ($name eq 'iso-2022-jp') ? 0 : 1;
-    
-    $Encode::Encoding{$name} =  
+
+    $Encode::Encoding{$name} =
         bless {
                Name      =>   $name,
                h2z       =>   $h2z,
@@ -16,8 +17,10 @@ for my $name ('7bit-jis', 'iso-2022-jp', 'iso-2022-jp-1'){
               } => __PACKAGE__;
 }
 
-sub name { shift->{'Name'} }
-sub new_sequence { $_[0] };
+use base qw(Encode::Encoding);
+
+# we override this to 1 so PerlIO works
+sub needs_lines { 1 }
 
 use Encode::CJKConstants qw(:all);
 
@@ -25,71 +28,125 @@ use Encode::CJKConstants qw(:all);
 # decode is identical for all 2022 variants
 #
 
-sub decode
+sub decode($$;$)
 {
-    my ($obj,$str,$chk) = @_;
-    jis_euc(\$str);
-    return Encode::decode('euc-jp', $str, $chk);
+    my ($obj, $str, $chk) = @_;
+    my $residue = '';
+    if ($chk){
+       $str =~ s/([^\x00-\x7f].*)$//so and $residue = $1;
+    }
+    $residue .= jis_euc(\$str);
+    $_[1] = $residue if $chk;
+    return Encode::decode('euc-jp', $str, FB_PERLQQ);
 }
 
 #
 # encode is different
 #
 
-sub encode
+sub encode($$;$)
 {
     require Encode::JP::H2Z;
-    my ($obj,$str,$chk) = @_;
+    my ($obj, $utf8, $chk) = @_;
+    # empty the input string in the stack so perlio is ok
+    $_[1] = '' if $chk;
     my ($h2z, $jis0212) = @$obj{qw(h2z jis0212)};
-    my $result = Encode::encode('euc-jp', $str, $chk);
-    $h2z and &Encode::JP::H2Z::h2z(\$result);
-    euc_jis(\$result, $jis0212);
-    return $result;
+    my $octet = Encode::encode('euc-jp', $utf8, FB_PERLQQ) ;
+    $h2z and &Encode::JP::H2Z::h2z(\$octet);
+    euc_jis(\$octet, $jis0212);
+    return $octet;
 }
 
+#
+# cat_decode
+#
+my $re_scan_jis_g = qr{
+   \G ( ($RE{JIS_0212}) |  $RE{JIS_0208}  |
+        ($RE{ISO_ASC})  | ($RE{JIS_KANA}) | )
+      ([^\e]*)
+}x;
+sub cat_decode { # ($obj, $dst, $src, $pos, $trm, $chk)
+    my ($obj, undef, undef, $pos, $trm) = @_; # currently ignores $chk
+    my ($rdst, $rsrc, $rpos) = \@_[1,2,3];
+    local ${^ENCODING};
+    use bytes;
+    my $opos = pos($$rsrc);
+    pos($$rsrc) = $pos;
+    while ($$rsrc =~ /$re_scan_jis_g/gc) {
+       my ($esc, $esc_0212, $esc_asc, $esc_kana, $chunk) =
+         ($1, $2, $3, $4, $5);
+
+       unless ($chunk) { $esc or last;  next; }
+
+       if ($esc && !$esc_asc) {
+           $chunk =~ tr/\x21-\x7e/\xa1-\xfe/;
+           if ($esc_kana) {
+               $chunk =~ s/([\xa1-\xdf])/\x8e$1/og;
+           } elsif ($esc_0212) {
+               $chunk =~ s/([\xa1-\xfe][\xa1-\xfe])/\x8f$1/og;
+           }
+           $chunk = Encode::decode('euc-jp', $chunk, 0);
+       }
+       elsif ((my $npos = index($chunk, $trm)) >= 0) {
+           $$rdst .= substr($chunk, 0, $npos + length($trm));
+           $$rpos += length($esc) + $npos + length($trm);
+           pos($$rsrc) = $opos;
+           return 1;
+       }
+       $$rdst .= $chunk;
+       $$rpos = pos($$rsrc);
+    }
+    $$rpos = pos($$rsrc);
+    pos($$rsrc) = $opos;
+    return '';
+}
 
 # JIS<->EUC
+my $re_scan_jis = qr{
+   (?:($RE{JIS_0212})|$RE{JIS_0208}|($RE{ISO_ASC})|($RE{JIS_KANA}))([^\e]*)
+}x;
 
 sub jis_euc {
+    local ${^ENCODING};
     my $r_str = shift;
-    $$r_str =~ s(
-                ($RE{JIS_0212}|$RE{JIS_0208}|$RE{ISO_ASC}|$RE{JIS_KANA})
-                ([^\e]*)
-                )
+    $$r_str =~ s($re_scan_jis)
     {
-       my ($esc, $str) = ($1, $2);
-       if ($esc !~ /$RE{ISO_ASC}/o) {
-           $str =~ tr/\x21-\x7e/\xa1-\xfe/;
-           if ($esc =~ /$RE{JIS_KANA}/o) {
-               $str =~ s/([\xa1-\xdf])/\x8e$1/og;
+       my ($esc_0212, $esc_asc, $esc_kana, $chunk) =
+          ($1, $2, $3, $4);
+       if (!$esc_asc) {
+           $chunk =~ tr/\x21-\x7e/\xa1-\xfe/;
+           if ($esc_kana) {
+               $chunk =~ s/([\xa1-\xdf])/\x8e$1/og;
            }
-           elsif ($esc =~ /$RE{JIS_0212}/o) {
-               $str =~ s/([\xa1-\xfe][\xa1-\xfe])/\x8f$1/og;
+           elsif ($esc_0212) {
+               $chunk =~ s/([\xa1-\xfe][\xa1-\xfe])/\x8f$1/og;
            }
        }
-       $str;
+       $chunk;
     }geox;
-    $$r_str;
+    my ($residue) = ($$r_str =~ s/(\e.*)$//so);
+    return $residue;
 }
 
 sub euc_jis{
+    no warnings qw(uninitialized);
     my $r_str = shift;
     my $jis0212 = shift;
     $$r_str =~ s{
        ((?:$RE{EUC_C})+|(?:$RE{EUC_KANA})+|(?:$RE{EUC_0212})+)
        }{
-           my $str = $1;
-           my $esc = 
-               ( $str =~ tr/\x8E//d ) ? $ESC{KANA} :
-                   ( $str =~ tr/\x8F//d ) ? $ESC{JIS_0212} :
+           my $chunk = $1;
+           my $esc =
+               ( $chunk =~ tr/\x8E//d ) ? $ESC{KANA} :
+                   ( $chunk =~ tr/\x8F//d ) ? $ESC{JIS_0212} :
                        $ESC{JIS_0208};
            if ($esc eq $ESC{JIS_0212} && !$jis0212){
                # fallback to '?'
-               $str =~ tr/\xA1-\xFE/\x3F/;
+               $chunk =~ tr/\xA1-\xFE/\x3F/;
            }else{
-               $str =~ tr/\xA1-\xFE/\x21-\x7E/;
+               $chunk =~ tr/\xA1-\xFE/\x21-\x7E/;
            }
-           $esc . $str . $ESC{ASC};
+           $esc . $chunk . $ESC{ASC};
        }geox;
     $$r_str =~
        s/\Q$ESC{ASC}\E