/[suikacvs]/markup/html/whatpm/Whatpm/HTML.pm.src

Diff of /markup/html/whatpm/Whatpm/HTML.pm.src

Parent Directory | Revision Log | View Patch Patch

-revision 1.129 by wakaba,
Sat Apr 12 15:37:33 2008 UTC
+revision 1.171 by wakaba,
Sun Sep 14 01:51:08 2008 UTC
 Line 8 
 use Error qw(:try);
  ## doc.write ('');
  ## alert (doc.compatMode);
- ## TODO: Control charcters and noncharacters are not allowed (HTML5 revision 1263)
+ require IO::Handle;
- ## TODO: 1252 parse error (revision 1264)
- ## TODO: 8859-11 = 874 (revision 1271)
  my $HTML_NS = q<http://www.w3.org/1999/xhtml>;
  my $MML_NS = q<http://www.w3.org/1998/Math/MathML>;
-Line 47 
 sub MISC_SPECIAL_EL () { 0b1000000000000
+Line 45 
 sub MISC_SPECIAL_EL () { 0b1000000000000
  sub FOREIGN_EL () { 0b10000000000000000000000000 }
  sub FOREIGN_FLOW_CONTENT_EL () { 0b100000000000000000000000000 }
  sub MML_AXML_EL () { 0b1000000000000000000000000000 }
+ sub RUBY_EL () { 0b10000000000000000000000000000 }
+ sub RUBY_COMPONENT_EL () { 0b100000000000000000000000000000 }
  sub TABLE_ROWS_EL () {
    TABLE_EL |
 Line 54 
 sub TABLE_ROWS_EL () {
    TABLE_ROW_GROUP_EL
  }
+ ## NOTE: Used in "generate implied end tags" algorithm.
+ ## NOTE: There is a code where a modified version of END_TAG_OPTIONAL_EL
+ ## is used in "generate implied end tags" implementation (search for the
+ ## function mae).
  sub END_TAG_OPTIONAL_EL () {
    DD_EL |
    DT_EL |
    LI_EL |
-   P_EL
+   P_EL |
+   RUBY_COMPONENT_EL
  }
+ ## NOTE: Used in </body> and EOF algorithms.
  sub ALL_END_TAG_OPTIONAL_EL () {
-   END_TAG_OPTIONAL_EL |
+   DD_EL |
+   DT_EL |
+   LI_EL |
+   P_EL |
    BODY_EL |
    HTML_EL |
    TABLE_CELL_EL |
-Line 98 
 sub SPECIAL_EL () {
+Line 108 
 sub SPECIAL_EL () {
    ADDRESS_EL |
    BODY_EL |
    DIV_EL |
-   END_TAG_OPTIONAL_EL |
+   DD_EL |
+   DT_EL |
+   LI_EL |
+   P_EL |
    FORM_EL |
    FRAMESET_EL |
    HEADING_EL |
-Line 172 
 my $el_category = {
+Line 187 
 my $el_category = {
    param => MISC_SPECIAL_EL,
    plaintext => MISC_SPECIAL_EL,
    pre => MISC_SPECIAL_EL,
+   rp => RUBY_COMPONENT_EL,
+   rt => RUBY_COMPONENT_EL,
+   ruby => RUBY_EL,
    s => FORMATTING_EL,
    script => MISC_SPECIAL_EL,
    select => SELECT_EL,
-Line 205 
 my $el_category_f = {
+Line 223 
 my $el_category_f = {
      mtext => FOREIGN_FLOW_CONTENT_EL,
    },
    $SVG_NS => {
-     foreignobject => FOREIGN_FLOW_CONTENT_EL, ## TODO: case
+     foreignObject => FOREIGN_FLOW_CONTENT_EL,
      desc => FOREIGN_FLOW_CONTENT_EL,
      title => FOREIGN_FLOW_CONTENT_EL,
    },
    ## NOTE: In addition, FOREIGN_EL is set to non-HTML elements.
  };
+ my $svg_attr_name = {
+   attributename => 'attributeName',
+   attributetype => 'attributeType',
+   basefrequency => 'baseFrequency',
+   baseprofile => 'baseProfile',
+   calcmode => 'calcMode',
+   clippathunits => 'clipPathUnits',
+   contentscripttype => 'contentScriptType',
+   contentstyletype => 'contentStyleType',
+   diffuseconstant => 'diffuseConstant',
+   edgemode => 'edgeMode',
+   externalresourcesrequired => 'externalResourcesRequired',
+   filterres => 'filterRes',
+   filterunits => 'filterUnits',
+   glyphref => 'glyphRef',
+   gradienttransform => 'gradientTransform',
+   gradientunits => 'gradientUnits',
+   kernelmatrix => 'kernelMatrix',
+   kernelunitlength => 'kernelUnitLength',
+   keypoints => 'keyPoints',
+   keysplines => 'keySplines',
+   keytimes => 'keyTimes',
+   lengthadjust => 'lengthAdjust',
+   limitingconeangle => 'limitingConeAngle',
+   markerheight => 'markerHeight',
+   markerunits => 'markerUnits',
+   markerwidth => 'markerWidth',
+   maskcontentunits => 'maskContentUnits',
+   maskunits => 'maskUnits',
+   numoctaves => 'numOctaves',
+   pathlength => 'pathLength',
+   patterncontentunits => 'patternContentUnits',
+   patterntransform => 'patternTransform',
+   patternunits => 'patternUnits',
+   pointsatx => 'pointsAtX',
+   pointsaty => 'pointsAtY',
+   pointsatz => 'pointsAtZ',
+   preservealpha => 'preserveAlpha',
+   preserveaspectratio => 'preserveAspectRatio',
+   primitiveunits => 'primitiveUnits',
+   refx => 'refX',
+   refy => 'refY',
+   repeatcount => 'repeatCount',
+   repeatdur => 'repeatDur',
+   requiredextensions => 'requiredExtensions',
+   requiredfeatures => 'requiredFeatures',
+   specularconstant => 'specularConstant',
+   specularexponent => 'specularExponent',
+   spreadmethod => 'spreadMethod',
+   startoffset => 'startOffset',
+   stddeviation => 'stdDeviation',
+   stitchtiles => 'stitchTiles',
+   surfacescale => 'surfaceScale',
+   systemlanguage => 'systemLanguage',
+   tablevalues => 'tableValues',
+   targetx => 'targetX',
+   targety => 'targetY',
+   textlength => 'textLength',
+   viewbox => 'viewBox',
+   viewtarget => 'viewTarget',
+   xchannelselector => 'xChannelSelector',
+   ychannelselector => 'yChannelSelector',
+   zoomandpan => 'zoomAndPan',
+ };
+ my $foreign_attr_xname = {
+   'xlink:actuate' => [$XLINK_NS, ['xlink', 'actuate']],
+   'xlink:arcrole' => [$XLINK_NS, ['xlink', 'arcrole']],
+   'xlink:href' => [$XLINK_NS, ['xlink', 'href']],
+   'xlink:role' => [$XLINK_NS, ['xlink', 'role']],
+   'xlink:show' => [$XLINK_NS, ['xlink', 'show']],
+   'xlink:title' => [$XLINK_NS, ['xlink', 'title']],
+   'xlink:type' => [$XLINK_NS, ['xlink', 'type']],
+   'xml:base' => [$XML_NS, ['xml', 'base']],
+   'xml:lang' => [$XML_NS, ['xml', 'lang']],
+   'xml:space' => [$XML_NS, ['xml', 'space']],
+   'xmlns' => [$XMLNS_NS, [undef, 'xmlns']],
+   'xmlns:xlink' => [$XMLNS_NS, ['xmlns', 'xlink']],
+ };
+ ## ISSUE: xmlns:xlink="non-xlink-ns" is not an error.
  my $c1_entity_char = {
 x80 => 0x20AC,
 x81 => 0xFFFD,
-Line 248 
 my $c1_entity_char = {
+Line 348 
 my $c1_entity_char = {
  }; # $c1_entity_char
  sub parse_byte_string ($$$$;$) {
+   my $self = shift;
+   my $charset_name = shift;
+   open my $input, '<', ref $_[0] ? $_[0] : \($_[0]);
+   return $self->parse_byte_stream ($charset_name, $input, @_[1..$#_]);
+ } # parse_byte_string
+ sub parse_byte_stream ($$$$;$$) {
+   # my ($self, $charset_name, $byte_stream, $doc, $onerror, $get_wrapper) = @_;
    my $self = ref $_[0] ? shift : shift->new;
-   my $charset = shift;
+   my $charset_name = shift;
-   my $bytes_s = ref $_[0] ? $_[0] : \($_[0]);
+   my $byte_stream = $_[0];
-   my $s;
+   my $onerror = $_[2] || sub {
-   if (defined $charset) {
+     my (%opt) = @_;
-     require Encode; ## TODO: decode(utf8) don't delete BOM
+     warn "Parse error ($opt{type})\n";
-     $s = \ (Encode::decode ($charset, $$bytes_s));
+   };
-     $self->{input_encoding} = lc $charset; ## TODO: normalize name
+   $self->{parse_error} = $onerror; # updated later by parse_char_string
-     $self->{confident} = 1;
-   } else {
+   my $get_wrapper = $_[3] || sub ($) {
-     ## TODO: Implement HTML5 detection algorithm
+     return $_[0]; # $_[0] = byte stream handle, returned = arg to char handle
+   };
+   ## HTML5 encoding sniffing algorithm
+   require Message::Charset::Info;
+   my $charset;
+   my $buffer;
+   my ($char_stream, $e_status);
+   SNIFFING: {
+     ## NOTE: By setting |allow_fallback| option true when the
+     ## |get_decode_handle| method is invoked, we ignore what the HTML5
+     ## spec requires, i.e. unsupported encoding should be ignored.
+       ## TODO: We should not do this unless the parser is invoked
+       ## in the conformance checking mode, in which this behavior
+       ## would be useful.
+     ## Step 1
+     if (defined $charset_name) {
+       $charset = Message::Charset::Info->get_by_html_name ($charset_name);
+           ## TODO: Is this ok?  Transfer protocol's parameter should be
+           ## interpreted in its semantics?
+       ## ISSUE: Unsupported encoding is not ignored according to the spec.
+       ($char_stream, $e_status) = $charset->get_decode_handle
+           ($byte_stream, allow_error_reporting => 1,
+            allow_fallback => 1);
+       if ($char_stream) {
+         $self->{confident} = 1;
+         last SNIFFING;
+       } else {
+         ## TODO: unsupported error
+       }
+     }
+     ## Step 2
+     my $byte_buffer = '';
+     for (1..1024) {
+       my $char = $byte_stream->getc;
+       last unless defined $char;
+       $byte_buffer .= $char;
+     } ## TODO: timeout
+     ## Step 3
+     if ($byte_buffer =~ /^\xFE\xFF/) {
+       $charset = Message::Charset::Info->get_by_html_name ('utf-16be');
+       ($char_stream, $e_status) = $charset->get_decode_handle
+           ($byte_stream, allow_error_reporting => 1,
+            allow_fallback => 1, byte_buffer => \$byte_buffer);
+       $self->{confident} = 1;
+       last SNIFFING;
+     } elsif ($byte_buffer =~ /^\xFF\xFE/) {
+       $charset = Message::Charset::Info->get_by_html_name ('utf-16le');
+       ($char_stream, $e_status) = $charset->get_decode_handle
+           ($byte_stream, allow_error_reporting => 1,
+            allow_fallback => 1, byte_buffer => \$byte_buffer);
+       $self->{confident} = 1;
+       last SNIFFING;
+     } elsif ($byte_buffer =~ /^\xEF\xBB\xBF/) {
+       $charset = Message::Charset::Info->get_by_html_name ('utf-8');
+       ($char_stream, $e_status) = $charset->get_decode_handle
+           ($byte_stream, allow_error_reporting => 1,
+            allow_fallback => 1, byte_buffer => \$byte_buffer);
+       $self->{confident} = 1;
+       last SNIFFING;
+     }
+     ## Step 4
+     ## TODO: <meta charset>
+     ## Step 5
+     ## TODO: from history
+     ## Step 6
      require Whatpm::Charset::UniversalCharDet;
-     $charset = Whatpm::Charset::UniversalCharDet->detect_byte_string
+     $charset_name = Whatpm::Charset::UniversalCharDet->detect_byte_string
-         (substr ($$bytes_s, 0, 1024));
+         ($byte_buffer);
-     $charset ||= 'windows-1252';
+     if (defined $charset_name) {
-     $s = \ (Encode::decode ($charset, $$bytes_s));
+       $charset = Message::Charset::Info->get_by_html_name ($charset_name);
-     $self->{input_encoding} = $charset;
+       ## ISSUE: Unsupported encoding is not ignored according to the spec.
+       require Whatpm::Charset::DecodeHandle;
+       $buffer = Whatpm::Charset::DecodeHandle::ByteBuffer->new
+           ($byte_stream);
+       ($char_stream, $e_status) = $charset->get_decode_handle
+           ($buffer, allow_error_reporting => 1,
+            allow_fallback => 1, byte_buffer => \$byte_buffer);
+       if ($char_stream) {
+         $buffer->{buffer} = $byte_buffer;
+         !!!parse-error (type => 'sniffing:chardet',
+                         text => $charset_name,
+                         level => $self->{level}->{info},
+                         layer => 'encode',
+                         line => 1, column => 1);
+         $self->{confident} = 0;
+         last SNIFFING;
+       }
+     }
+     ## Step 7: default
+     ## TODO: Make this configurable.
+     $charset = Message::Charset::Info->get_by_html_name ('windows-1252');
+         ## NOTE: We choose |windows-1252| here, since |utf-8| should be
+         ## detectable in the step 6.
+     require Whatpm::Charset::DecodeHandle;
+     $buffer = Whatpm::Charset::DecodeHandle::ByteBuffer->new
+         ($byte_stream);
+     ($char_stream, $e_status)
+         = $charset->get_decode_handle ($buffer,
+                                        allow_error_reporting => 1,
+                                        allow_fallback => 1,
+                                        byte_buffer => \$byte_buffer);
+     $buffer->{buffer} = $byte_buffer;
+     !!!parse-error (type => 'sniffing:default',
+                     text => 'windows-1252',
+                     level => $self->{level}->{info},
+                     line => 1, column => 1,
+                     layer => 'encode');
      $self->{confident} = 0;
+   } # SNIFFING
+   if ($e_status & Message::Charset::Info::FALLBACK_ENCODING_IMPL ()) {
+     $self->{input_encoding} = $charset->get_iana_name; ## TODO: Should we set actual charset decoder's encoding name?
+     !!!parse-error (type => 'chardecode:fallback',
+                     #text => $self->{input_encoding},
+                     level => $self->{level}->{uncertain},
+                     line => 1, column => 1,
+                     layer => 'encode');
+   } elsif (not ($e_status &
+                 Message::Charset::Info::ERROR_REPORTING_ENCODING_IMPL())) {
+     $self->{input_encoding} = $charset->get_iana_name;
+     !!!parse-error (type => 'chardecode:no error',
+                     text => $self->{input_encoding},
+                     level => $self->{level}->{uncertain},
+                     line => 1, column => 1,
+                     layer => 'encode');
+   } else {
+     $self->{input_encoding} = $charset->get_iana_name;
    }
    $self->{change_encoding} = sub {
      my $self = shift;
-     my $charset = lc shift;
+     $charset_name = shift;
      my $token = shift;
-     ## TODO: if $charset is supported
-     ## TODO: normalize charset name
-     ## "Change the encoding" algorithm:
+     $charset = Message::Charset::Info->get_by_html_name ($charset_name);
+     ($char_stream, $e_status) = $charset->get_decode_handle
+         ($byte_stream, allow_error_reporting => 1, allow_fallback => 1,
+          byte_buffer => \ $buffer->{buffer});
+     if ($char_stream) { # if supported
+       ## "Change the encoding" algorithm:
-     ## Step 1
+       ## Step 1
-     if ($charset eq 'utf-16') { ## ISSUE: UTF-16BE -> UTF-8? UTF-16LE -> UTF-8?
+       if ($charset->{category} &
-       $charset = 'utf-8';
+           Message::Charset::Info::CHARSET_CATEGORY_UTF16 ()) {
-     }
+         $charset = Message::Charset::Info->get_by_html_name ('utf-8');
+         ($char_stream, $e_status) = $charset->get_decode_handle
+             ($byte_stream,
+              byte_buffer => \ $buffer->{buffer});
+       }
+       $charset_name = $charset->get_iana_name;
+       ## Step 2
+       if (defined $self->{input_encoding} and
+           $self->{input_encoding} eq $charset_name) {
+         !!!parse-error (type => 'charset label:matching',
+                         text => $charset_name,
+                         level => $self->{level}->{info});
+         $self->{confident} = 1;
+         return;
+       }
-     ## Step 2
+       !!!parse-error (type => 'charset label detected',
-     if (defined $self->{input_encoding} and
+                       text => $self->{input_encoding},
-         $self->{input_encoding} eq $charset) {
+                       value => $charset_name,
-       $self->{confident} = 1;
+                       level => $self->{level}->{warn},
-       return;
+                       token => $token);
+       ## Step 3
+       # if (can) {
+         ## change the encoding on the fly.
+         #$self->{confident} = 1;
+         #return;
+       # }
+       ## Step 4
+       throw Whatpm::HTML::RestartParser ();
      }
+   }; # $self->{change_encoding}
-     !!!parse-error (type => 'charset label detected:'.$self->{input_encoding}.
+   my $char_onerror = sub {
-         ':'.$charset, level => 'w', token => $token);
+     my (undef, $type, %opt) = @_;
+     !!!parse-error (layer => 'encode',
-     ## Step 3
+                     %opt, type => $type,
-     # if (can) {
+                     line => $self->{line}, column => $self->{column} + 1);
-       ## change the encoding on the fly.
+     if ($opt{octets}) {
-       #$self->{confident} = 1;
+       ${$opt{octets}} = "\x{FFFD}"; # relacement character
-       #return;
+     }
-     # }
+   };
-     ## Step 4
+   my $wrapped_char_stream = $get_wrapper->($char_stream);
-     throw Whatpm::HTML::RestartParser (charset => $charset);
+   $wrapped_char_stream->onerror ($char_onerror);
-   }; # $self->{change_encoding}
    my @args = @_; shift @args; # $s
    my $return;
    try {
-     $return = $self->parse_char_string ($s, @args);
+     $return = $self->parse_char_stream ($wrapped_char_stream, @args);
    } catch Whatpm::HTML::RestartParser with {
-     my $charset = shift->{charset};
+     ## NOTE: Invoked after {change_encoding}.
-     $s = \ (Encode::decode ($charset, $$bytes_s));
-     $self->{input_encoding} = $charset; ## TODO: normalize
+     if ($e_status & Message::Charset::Info::FALLBACK_ENCODING_IMPL ()) {
+       $self->{input_encoding} = $charset->get_iana_name; ## TODO: Should we set actual charset decoder's encoding name?
+       !!!parse-error (type => 'chardecode:fallback',
+                       level => $self->{level}->{uncertain},
+                       #text => $self->{input_encoding},
+                       line => 1, column => 1,
+                       layer => 'encode');
+     } elsif (not ($e_status &
+                   Message::Charset::Info::ERROR_REPORTING_ENCODING_IMPL())) {
+       $self->{input_encoding} = $charset->get_iana_name;
+       !!!parse-error (type => 'chardecode:no error',
+                       text => $self->{input_encoding},
+                       level => $self->{level}->{uncertain},
+                       line => 1, column => 1,
+                       layer => 'encode');
+     } else {
+       $self->{input_encoding} = $charset->get_iana_name;
+     }
      $self->{confident} = 1;
-     $return = $self->parse_char_string ($s, @args);
+     $wrapped_char_stream = $get_wrapper->($char_stream);
+     $wrapped_char_stream->onerror ($char_onerror);
+     $return = $self->parse_char_stream ($wrapped_char_stream, @args);
    };
    return $return;
- } # parse_byte_string
+ } # parse_byte_stream
  ## NOTE: HTML5 spec says that the encoding layer MUST NOT strip BOM
  ## and the HTML layer MUST ignore it.  However, we does strip BOM in
-Line 327 
 sub parse_byte_string ($$$$;$) {
+Line 615 
 sub parse_byte_string ($$$$;$) {
  ## such as |parse_byte_string| in this module, must ensure that it does
  ## strip the BOM and never strip any ZWNBSP.
- *parse_char_string = \&parse_string;
+ sub parse_char_string ($$$;$$) {
+   #my ($self, $s, $doc, $onerror, $get_wrapper) = @_;
+   my $self = shift;
+   my $s = ref $_[0] ? $_[0] : \($_[0]);
+   require Whatpm::Charset::DecodeHandle;
+   my $input = Whatpm::Charset::DecodeHandle::CharString->new ($s);
+   if ($_[3]) {
+     $input = $_[3]->($input);
+   }
+   return $self->parse_char_stream ($input, @_[1..$#_]);
+ } # parse_char_string
+ *parse_string = \&parse_char_string; ## NOTE: Alias for backward compatibility.
- sub parse_string ($$$;$) {
+ sub parse_char_stream ($$$;$) {
    my $self = ref $_[0] ? shift : shift->new;
-   my $s = ref $_[0] ? $_[0] : \($_[0]);
+   my $input = $_[0];
    $self->{document} = $_[1];
    @{$self->{document}->child_nodes} = ();
-Line 350 
 sub parse_string ($$$;$) {
+Line 649 
 sub parse_string ($$$;$) {
      pop @{$self->{prev_char}};
      unshift @{$self->{prev_char}}, $self->{next_char};
-     $self->{next_char} = -1 and return if $i >= length $$s;
+     my $char;
-     $self->{next_char} = ord substr $$s, $i++, 1;
+     if (defined $self->{next_next_char}) {
+       $char = $self->{next_next_char};
+       delete $self->{next_next_char};
+     } else {
+       $char = $input->getc;
+     }
+     $self->{next_char} = -1 and return unless defined $char;
+     $self->{next_char} = ord $char;
      ($self->{line_prev}, $self->{column_prev})
          = ($self->{line}, $self->{column});
      $self->{column}++;
      if ($self->{next_char} == 0x000A) { # LF
+       !!!cp ('j1');
        $self->{line}++;
        $self->{column} = 0;
      } elsif ($self->{next_char} == 0x000D) { # CR
-       $i++ if substr ($$s, $i, 1) eq "\x0A";
+       !!!cp ('j2');
+ ## TODO: support for abort/streaming
+       my $next = $input->getc;
+       if (defined $next and $next ne "\x0A") {
+         $self->{next_next_char} = $next;
+       }
        $self->{next_char} = 0x000A; # LF # MUST
        $self->{line}++;
        $self->{column} = 0;
      } elsif ($self->{next_char} > 0x10FFFF) {
+       !!!cp ('j3');
        $self->{next_char} = 0xFFFD; # REPLACEMENT CHARACTER # MUST
      } elsif ($self->{next_char} == 0x0000) { # NULL
+       !!!cp ('j4');
        !!!parse-error (type => 'NULL');
        $self->{next_char} = 0xFFFD; # REPLACEMENT CHARACTER # MUST
+     } elsif ($self->{next_char} <= 0x0008 or
+              (0x000E <= $self->{next_char} and $self->{next_char} <= 0x001F) or
+              (0x007F <= $self->{next_char} and $self->{next_char} <= 0x009F) or
+              (0xD800 <= $self->{next_char} and $self->{next_char} <= 0xDFFF) or
+              (0xFDD0 <= $self->{next_char} and $self->{next_char} <= 0xFDDF) or
+ ## ISSUE: U+FDE0-U+FDEF are not excluded
+              {
+xFFFE => 1, 0xFFFF => 1, 0x1FFFE => 1, 0x1FFFF => 1,
+x2FFFE => 1, 0x2FFFF => 1, 0x3FFFE => 1, 0x3FFFF => 1,
+x4FFFE => 1, 0x4FFFF => 1, 0x5FFFE => 1, 0x5FFFF => 1,
+x6FFFE => 1, 0x6FFFF => 1, 0x7FFFE => 1, 0x7FFFF => 1,
+x8FFFE => 1, 0x8FFFF => 1, 0x9FFFE => 1, 0x9FFFF => 1,
+xAFFFE => 1, 0xAFFFF => 1, 0xBFFFE => 1, 0xBFFFF => 1,
+xCFFFE => 1, 0xCFFFF => 1, 0xDFFFE => 1, 0xDFFFF => 1,
+xEFFFE => 1, 0xEFFFF => 1, 0xFFFFE => 1, 0xFFFFF => 1,
+x10FFFE => 1, 0x10FFFF => 1,
+              }->{$self->{next_char}}) {
+       !!!cp ('j5');
+       if ($self->{next_char} < 0x10000) {
+         !!!parse-error (type => 'control char',
+                         text => (sprintf 'U+%04X', $self->{next_char}));
+       } else {
+         !!!parse-error (type => 'control char',
+                         text => (sprintf 'U-%08X', $self->{next_char}));
+       }
      }
    };
    $self->{prev_char} = [-1, -1, -1];
    $self->{next_char} = -1;
+   $self->{getc_until} = sub { return undef };
+ #  if ($input->can ('manakai_getc_until')) {
+     $self->{getc_until} = sub {
+       my $special_range = shift;
+       return undef if defined $self->{next_next_char};
+       my $s = $input->manakai_getc_until
+           (qr/(?![$special_range\x{FDD0}-\x{FDDF}\x{FFFE}\x{FFFF}\x{1FFFE}\x{1FFFF}\x{2FFFE}\x{2FFFF}\x{3FFFE}\x{3FFFF}\x{4FFFE}\x{4FFFF}\x{5FFFE}\x{5FFFF}\x{6FFFE}\x{6FFFF}\x{7FFFE}\x{7FFFF}\x{8FFFE}\x{8FFFF}\x{9FFFE}\x{9FFFF}\x{AFFFE}\x{AFFFF}\x{BFFFE}\x{BFFFF}\x{CFFFE}\x{CFFFF}\x{DFFFE}\x{DFFFF}\x{EFFFE}\x{EFFFF}\x{FFFFE}\x{FFFFF}])[\x20-\x7E\xA0-\x{D7FF}\x{E000}-\x{10FFFD}]/);
+       if ($s) {
+         $self->{column} += length $$s;
+         $self->{column_prev} += length $$s;
+         $self->{prev_char} = [-1, -1, -1];
+         $self->{next_char} = -1;
+       }
+       return $s;
+     }; # $self->{getc_until}
+ #  } else {
+ #    $self->{getc_until} = sub {
+ #      my $special_range = shift;
+ #      return undef if defined $self->{next_next_char};
+ #      my $c = $input->getc;
+ #      if ($c =~ /^(?![$special_range\x{FDD0}-\x{FDDF}\x{FFFE}\x{FFFF}\x{1FFFE}\x{1FFFF}\x{2FFFE}\x{2FFFF}\x{3FFFE}\x{3FFFF}\x{4FFFE}\x{4FFFF}\x{5FFFE}\x{5FFFF}\x{6FFFE}\x{6FFFF}\x{7FFFE}\x{7FFFF}\x{8FFFE}\x{8FFFF}\x{9FFFE}\x{9FFFF}\x{AFFFE}\x{AFFFF}\x{BFFFE}\x{BFFFF}\x{CFFFE}\x{CFFFF}\x{DFFFE}\x{DFFFF}\x{EFFFE}\x{EFFFF}\x{FFFFE}\x{FFFFF}])[\x20-\x7E\xA0-\x{D7FF}\x{E000}-\x{10FFFD}]/) {
+ #        $self->{column}++;
+ #        $self->{column_prev}++;
+ #        $self->{prev_char} = [-1, -1, -1];
+ #        $self->{next_char} = -1;
+ #        return \$c;
+ #      } elsif (defined $c) {
+ #        #$input->ungetc (ord $c);
+ #        $self->{next_next_char} = $c;
+ #        return undef;
+ #      } else {
+ #        return undef;
+ #      }
+ #    }; # $self->{getc_until}
+ #  }
    my $onerror = $_[2] || sub {
      my (%opt) = @_;
      my $line = $opt{token} ? $opt{token}->{line} : $opt{line};
-Line 393 
 sub parse_string ($$$;$) {
+Line 768 
 sub parse_string ($$$;$) {
    delete $self->{parse_error}; # remove loop
    return $self->{document};
- } # parse_string
+ } # parse_char_stream
  sub new ($) {
    my $class = shift;
-   my $self = bless {}, $class;
+   my $self = bless {
+     level => {must => 'm',
+               should => 's',
+               warn => 'w',
+               info => 'i',
+               uncertain => 'u'},
+   }, $class;
    $self->{set_next_char} = sub {
      $self->{next_char} = -1;
    };
-Line 426 
 sub RCDATA_CONTENT_MODEL () { CM_ENTITY
+Line 807 
 sub RCDATA_CONTENT_MODEL () { CM_ENTITY
  sub PCDATA_CONTENT_MODEL () { CM_ENTITY | CM_FULL_MARKUP }
  sub DATA_STATE () { 0 }
- sub ENTITY_DATA_STATE () { 1 }
+ #sub ENTITY_DATA_STATE () { 1 }
  sub TAG_OPEN_STATE () { 2 }
  sub CLOSE_TAG_OPEN_STATE () { 3 }
  sub TAG_NAME_STATE () { 4 }
-Line 437 
 sub BEFORE_ATTRIBUTE_VALUE_STATE () { 8
+Line 818 
 sub BEFORE_ATTRIBUTE_VALUE_STATE () { 8
  sub ATTRIBUTE_VALUE_DOUBLE_QUOTED_STATE () { 9 }
  sub ATTRIBUTE_VALUE_SINGLE_QUOTED_STATE () { 10 }
  sub ATTRIBUTE_VALUE_UNQUOTED_STATE () { 11 }
- sub ENTITY_IN_ATTRIBUTE_VALUE_STATE () { 12 }
+ #sub ENTITY_IN_ATTRIBUTE_VALUE_STATE () { 12 }
  sub MARKUP_DECLARATION_OPEN_STATE () { 13 }
  sub COMMENT_START_STATE () { 14 }
  sub COMMENT_START_DASH_STATE () { 15 }
-Line 460 
 sub AFTER_DOCTYPE_SYSTEM_IDENTIFIER_STAT
+Line 841 
 sub AFTER_DOCTYPE_SYSTEM_IDENTIFIER_STAT
  sub BOGUS_DOCTYPE_STATE () { 32 }
  sub AFTER_ATTRIBUTE_VALUE_QUOTED_STATE () { 33 }
  sub SELF_CLOSING_START_TAG_STATE () { 34 }
- sub CDATA_BLOCK_STATE () { 35 }
+ sub CDATA_SECTION_STATE () { 35 }
+ sub MD_HYPHEN_STATE () { 36 } # "markup declaration open state" in the spec
+ sub MD_DOCTYPE_STATE () { 37 } # "markup declaration open state" in the spec
+ sub MD_CDATA_STATE () { 38 } # "markup declaration open state" in the spec
+ sub CDATA_PCDATA_CLOSE_TAG_STATE () { 39 } # "close tag open state" in the spec
+ sub CDATA_SECTION_MSE1_STATE () { 40 } # "CDATA section state" in the spec
+ sub CDATA_SECTION_MSE2_STATE () { 41 } # "CDATA section state" in the spec
+ sub PUBLIC_STATE () { 42 } # "after DOCTYPE name state" in the spec
+ sub SYSTEM_STATE () { 43 } # "after DOCTYPE name state" in the spec
+ ## NOTE: "Entity data state", "entity in attribute value state", and
+ ## "consume a character reference" algorithm are jointly implemented
+ ## using the following six states:
+ sub ENTITY_STATE () { 44 }
+ sub ENTITY_HASH_STATE () { 45 }
+ sub NCR_NUM_STATE () { 46 }
+ sub HEXREF_X_STATE () { 47 }
+ sub HEXREF_HEX_STATE () { 48 }
+ sub ENTITY_NAME_STATE () { 49 }
  sub DOCTYPE_TOKEN () { 1 }
  sub COMMENT_TOKEN () { 2 }
-Line 513 
 sub IN_COLUMN_GROUP_IM () { 0b10 }
+Line 911 
 sub IN_COLUMN_GROUP_IM () { 0b10 }
  sub _initialize_tokenizer ($) {
    my $self = shift;
    $self->{state} = DATA_STATE; # MUST
+   #$self->{state_keyword}; # initialized when used
+   #$self->{entity__value}; # initialized when used
+   #$self->{entity__match}; # initialized when used
    $self->{content_model} = PCDATA_CONTENT_MODEL; # be
-   undef $self->{current_token}; # start tag, end tag, comment, or DOCTYPE
+   undef $self->{current_token};
    undef $self->{current_attribute};
    undef $self->{last_emitted_start_tag_name};
-   undef $self->{last_attribute_value_state};
+   #$self->{prev_state}; # initialized when used
    delete $self->{self_closing};
-   $self->{char} = [];
    # $self->{next_char}
    !!!next-input-character;
    $self->{token} = [];
-Line 543 
 sub _initialize_tokenizer ($) {
+Line 943 
 sub _initialize_tokenizer ($) {
  ##     |->{self_closing}| is used to save the value of |$self->{self_closing}|
  ##     while the token is pushed back to the stack.
- ## ISSUE: "When a DOCTYPE token is created, its
- ## <i>self-closing flag</i> must be unset (its other state is that it
- ## be set), and its attributes list must be empty.": Wrong subject?
  ## Emitted token MUST immediately be handled by the tree construction state.
  ## Before each step, UA MAY check to see if either one of the scripts in
-Line 555 
 sub _initialize_tokenizer ($) {
+Line 951 
 sub _initialize_tokenizer ($) {
  ## has completed loading.  If one has, then it MUST be executed
  ## and removed from the list.
- ## NOTE: HTML5 "Writing HTML documents" section, applied to
+ ## TODO: Polytheistic slash SHOULD NOT be used. (Applied only to atheists.)
- ## documents and not to user agents and conformance checkers,
+ ## (This requirement was dropped from HTML5 spec, unfortunately.)
- ## contains some requirements that are not detected by the
- ## parsing algorithm:
- ## - Some requirements on character encoding declarations. ## TODO
- ## - "Elements MUST NOT contain content that their content model disallows."
- ##   ... Some are parse error, some are not (will be reported by c.c.).
- ## - Polytheistic slash SHOULD NOT be used. (Applied only to atheists.) ## TODO
- ## - Text (in elements, attributes, and comments) SHOULD NOT contain
- ##   control characters other than space characters. ## TODO: (what is control character? C0, C1 and DEL?  Unicode control character?)
- ## TODO: HTML5 poses authors two SHOULD-level requirements that cannot
- ## be detected by the HTML5 parsing algorithm:
- ## - Text,
  sub _get_next_token ($) {
    my $self = shift;
-Line 592 
 sub _get_next_token ($) {
+Line 976 
 sub _get_next_token ($) {
          if ($self->{content_model} & CM_ENTITY and # PCDATA | RCDATA
              not $self->{escape}) {
            !!!cp (1);
-           $self->{state} = ENTITY_DATA_STATE;
+           ## NOTE: In the spec, the tokenizer is switched to the
+           ## "entity data state".  In this implementation, the tokenizer
+           ## is switched to the |ENTITY_STATE|, which is an implementation
+           ## of the "consume a character reference" algorithm.
+           $self->{entity_additional} = -1;
+           $self->{prev_state} = DATA_STATE;
+           $self->{state} = ENTITY_STATE;
            !!!next-input-character;
            redo A;
          } else {
-Line 656 
 sub _get_next_token ($) {
+Line 1046 
 sub _get_next_token ($) {
                     data => chr $self->{next_char},
                     line => $self->{line}, column => $self->{column},
                    };
+       my $s = $self->{getc_until}->(q[-!<>&]);
+       if ($s) {
+         $token->{data} .= $$s;
+       }
        ## Stay in the data state
        !!!next-input-character;
        !!!emit ($token);
        redo A;
-     } elsif ($self->{state} == ENTITY_DATA_STATE) {
-       ## (cannot happen in CDATA state)
-       my ($l, $c) = ($self->{line_prev}, $self->{column_prev});
-       my $token = $self->_tokenize_attempt_to_consume_an_entity (0, -1);
-       $self->{state} = DATA_STATE;
-       # next-input-character is already done
-       unless (defined $token) {
-         !!!cp (13);
-         !!!emit ({type => CHARACTER_TOKEN, data => '&',
-                   line => $l, column => $c,
-                  });
-       } else {
-         !!!cp (14);
-         !!!emit ($token);
-       }
-       redo A;
      } elsif ($self->{state} == TAG_OPEN_STATE) {
        if ($self->{content_model} & CM_LIMITED_MARKUP) { # RCDATA | CDATA
          if ($self->{next_char} == 0x002F) { # /
-Line 762 
 sub _get_next_token ($) {
+Line 1137 
 sub _get_next_token ($) {
            redo A;
          } else {
            !!!cp (23);
-           !!!parse-error (type => 'bare stago');
+           !!!parse-error (type => 'bare stago',
+                           line => $self->{line_prev},
+                           column => $self->{column_prev});
            $self->{state} = DATA_STATE;
            ## reconsume
-Line 777 
 sub _get_next_token ($) {
+Line 1154 
 sub _get_next_token ($) {
          die "$0: $self->{content_model} in tag open";
        }
      } elsif ($self->{state} == CLOSE_TAG_OPEN_STATE) {
+       ## NOTE: The "close tag open state" in the spec is implemented as
+       ## |CLOSE_TAG_OPEN_STATE| and |CDATA_PCDATA_CLOSE_TAG_STATE|.
        my ($l, $c) = ($self->{line_prev}, $self->{column_prev} - 1); # "<"of"</"
        if ($self->{content_model} & CM_LIMITED_MARKUP) { # RCDATA | CDATA
          if (defined $self->{last_emitted_start_tag_name}) {
+           $self->{state} = CDATA_PCDATA_CLOSE_TAG_STATE;
-           ## NOTE: <http://krijnhoetmer.nl/irc-logs/whatwg/20070626#l-564>
+           $self->{state_keyword} = '';
-           my @next_char;
+           ## Reconsume.
-           TAGNAME: for (my $i = 0; $i < length $self->{last_emitted_start_tag_name}; $i++) {
+           redo A;
-             push @next_char, $self->{next_char};
-             my $c = ord substr ($self->{last_emitted_start_tag_name}, $i, 1);
-             my $C = 0x0061 <= $c && $c <= 0x007A ? $c - 0x0020 : $c;
-             if ($self->{next_char} == $c or $self->{next_char} == $C) {
-               !!!cp (24);
-               !!!next-input-character;
-               next TAGNAME;
-             } else {
-               !!!cp (25);
-               $self->{next_char} = shift @next_char; # reconsume
-               !!!back-next-input-character (@next_char);
-               $self->{state} = DATA_STATE;
-               !!!emit ({type => CHARACTER_TOKEN, data => '</',
-                         line => $l, column => $c,
-                        });
-               redo A;
-             }
-           }
-           push @next_char, $self->{next_char};
-           unless ($self->{next_char} == 0x0009 or # HT
-                   $self->{next_char} == 0x000A or # LF
-                   $self->{next_char} == 0x000B or # VT
-                   $self->{next_char} == 0x000C or # FF
-                   $self->{next_char} == 0x0020 or # SP
-                   $self->{next_char} == 0x003E or # >
-                   $self->{next_char} == 0x002F or # /
-                   $self->{next_char} == -1) {
-             !!!cp (26);
-             $self->{next_char} = shift @next_char; # reconsume
-             !!!back-next-input-character (@next_char);
-             $self->{state} = DATA_STATE;
-             !!!emit ({type => CHARACTER_TOKEN, data => '</',
-                       line => $l, column => $c,
-                      });
-             redo A;
-           } else {
-             !!!cp (27);
-             $self->{next_char} = shift @next_char;
-             !!!back-next-input-character (@next_char);
-             # and consume...
-           }
          } else {
            ## No start tag token has ever been emitted
+           ## NOTE: See <http://krijnhoetmer.nl/irc-logs/whatwg/20070626#l-564>.
            !!!cp (28);
-           # next-input-character is already done
            $self->{state} = DATA_STATE;
+           ## Reconsume.
            !!!emit ({type => CHARACTER_TOKEN, data => '</',
                      line => $l, column => $c,
                     });
            redo A;
          }
        }
        if (0x0041 <= $self->{next_char} and
            $self->{next_char} <= 0x005A) { # A..Z
          !!!cp (29);
-Line 886 
 sub _get_next_token ($) {
+Line 1223 
 sub _get_next_token ($) {
                                    line => $self->{line_prev}, # "<" of "</"
                                    column => $self->{column_prev} - 1,
                                   };
-         ## $self->{next_char} is intentionally left as is
+         ## NOTE: $self->{next_char} is intentionally left as is.
-         redo A;
+         ## Although the "anything else" case of the spec not explicitly
+         ## states that the next input character is to be reconsumed,
+         ## it will be included to the |data| of the comment token
+         ## generated from the bogus end tag, as defined in the
+         ## "bogus comment state" entry.
+         redo A;
+       }
+     } elsif ($self->{state} == CDATA_PCDATA_CLOSE_TAG_STATE) {
+       my $ch = substr $self->{last_emitted_start_tag_name}, length $self->{state_keyword}, 1;
+       if (length $ch) {
+         my $CH = $ch;
+         $ch =~ tr/a-z/A-Z/;
+         my $nch = chr $self->{next_char};
+         if ($nch eq $ch or $nch eq $CH) {
+           !!!cp (24);
+           ## Stay in the state.
+           $self->{state_keyword} .= $nch;
+           !!!next-input-character;
+           redo A;
+         } else {
+           !!!cp (25);
+           $self->{state} = DATA_STATE;
+           ## Reconsume.
+           !!!emit ({type => CHARACTER_TOKEN,
+                     data => '</' . $self->{state_keyword},
+                     line => $self->{line_prev},
+                     column => $self->{column_prev} - 1 - length $self->{state_keyword},
+                    });
+           redo A;
+         }
+       } else { # after "<{tag-name}"
+         unless ({
+x0009 => 1, # HT
+x000A => 1, # LF
+x000B => 1, # VT
+x000C => 1, # FF
+x0020 => 1, # SP
+x003E => 1, # >
+x002F => 1, # /
+                  -1 => 1, # EOF
+                 }->{$self->{next_char}}) {
+           !!!cp (26);
+           ## Reconsume.
+           $self->{state} = DATA_STATE;
+           !!!emit ({type => CHARACTER_TOKEN,
+                     data => '</' . $self->{state_keyword},
+                     line => $self->{line_prev},
+                     column => $self->{column_prev} - 1 - length $self->{state_keyword},
+                    });
+           redo A;
+         } else {
+           !!!cp (27);
+           $self->{current_token}
+               = {type => END_TAG_TOKEN,
+                  tag_name => $self->{last_emitted_start_tag_name},
+                  line => $self->{line_prev},
+                  column => $self->{column_prev} - 1 - length $self->{state_keyword}};
+           $self->{state} = TAG_NAME_STATE;
+           ## Reconsume.
+           redo A;
+         }
        }
      } elsif ($self->{state} == TAG_NAME_STATE) {
        if ($self->{next_char} == 0x0009 or # HT
-Line 1057 
 sub _get_next_token ($) {
+Line 1454 
 sub _get_next_token ($) {
          if (exists $self->{current_token}->{attributes} # start tag or end tag
              ->{$self->{current_attribute}->{name}}) { # MUST
            !!!cp (57);
-           !!!parse-error (type => 'duplicate attribute:'.$self->{current_attribute}->{name}, line => $self->{current_attribute}->{line}, column => $self->{current_attribute}->{column});
+           !!!parse-error (type => 'duplicate attribute', text => $self->{current_attribute}->{name}, line => $self->{current_attribute}->{line}, column => $self->{current_attribute}->{column});
            ## Discard $self->{current_attribute} # MUST
          } else {
            !!!cp (58);
-Line 1228 
 sub _get_next_token ($) {
+Line 1625 
 sub _get_next_token ($) {
          redo A;
        } else {
-         !!!cp (82);
+         if ($self->{next_char} == 0x0022 or # "
+             $self->{next_char} == 0x0027) { # '
+           !!!cp (78);
+           !!!parse-error (type => 'bad attribute name');
+         } else {
+           !!!cp (82);
+         }
          $self->{current_attribute}
              = {name => chr ($self->{next_char}),
                 value => '',
-Line 1263 
 sub _get_next_token ($) {
+Line 1666 
 sub _get_next_token ($) {
          !!!next-input-character;
          redo A;
        } elsif ($self->{next_char} == 0x003E) { # >
+         !!!parse-error (type => 'empty unquoted attribute value');
          if ($self->{current_token}->{type} == START_TAG_TOKEN) {
            !!!cp (87);
            $self->{last_emitted_start_tag_name} = $self->{current_token}->{tag_name};
-Line 1327 
 sub _get_next_token ($) {
+Line 1731 
 sub _get_next_token ($) {
          redo A;
        } elsif ($self->{next_char} == 0x0026) { # &
          !!!cp (96);
-         $self->{last_attribute_value_state} = $self->{state};
+         ## NOTE: In the spec, the tokenizer is switched to the
-         $self->{state} = ENTITY_IN_ATTRIBUTE_VALUE_STATE;
+         ## "entity in attribute value state".  In this implementation, the
+         ## tokenizer is switched to the |ENTITY_STATE|, which is an
+         ## implementation of the "consume a character reference" algorithm.
+         $self->{prev_state} = $self->{state};
+         $self->{entity_additional} = 0x0022; # "
+         $self->{state} = ENTITY_STATE;
          !!!next-input-character;
          redo A;
        } elsif ($self->{next_char} == -1) {
-Line 1369 
 sub _get_next_token ($) {
+Line 1778 
 sub _get_next_token ($) {
          redo A;
        } elsif ($self->{next_char} == 0x0026) { # &
          !!!cp (102);
-         $self->{last_attribute_value_state} = $self->{state};
+         ## NOTE: In the spec, the tokenizer is switched to the
-         $self->{state} = ENTITY_IN_ATTRIBUTE_VALUE_STATE;
+         ## "entity in attribute value state".  In this implementation, the
+         ## tokenizer is switched to the |ENTITY_STATE|, which is an
+         ## implementation of the "consume a character reference" algorithm.
+         $self->{entity_additional} = 0x0027; # '
+         $self->{prev_state} = $self->{state};
+         $self->{state} = ENTITY_STATE;
          !!!next-input-character;
          redo A;
        } elsif ($self->{next_char} == -1) {
-Line 1415 
 sub _get_next_token ($) {
+Line 1829 
 sub _get_next_token ($) {
          redo A;
        } elsif ($self->{next_char} == 0x0026) { # &
          !!!cp (108);
-         $self->{last_attribute_value_state} = $self->{state};
+         ## NOTE: In the spec, the tokenizer is switched to the
-         $self->{state} = ENTITY_IN_ATTRIBUTE_VALUE_STATE;
+         ## "entity in attribute value state".  In this implementation, the
+         ## tokenizer is switched to the |ENTITY_STATE|, which is an
+         ## implementation of the "consume a character reference" algorithm.
+         $self->{entity_additional} = -1;
+         $self->{prev_state} = $self->{state};
+         $self->{state} = ENTITY_STATE;
          !!!next-input-character;
          redo A;
        } elsif ($self->{next_char} == 0x003E) { # >
-Line 1480 
 sub _get_next_token ($) {
+Line 1899 
 sub _get_next_token ($) {
          !!!next-input-character;
          redo A;
        }
-     } elsif ($self->{state} == ENTITY_IN_ATTRIBUTE_VALUE_STATE) {
-       my $token = $self->_tokenize_attempt_to_consume_an_entity
-           (1,
-            $self->{last_attribute_value_state}
-              == ATTRIBUTE_VALUE_DOUBLE_QUOTED_STATE ? 0x0022 : # "
-            $self->{last_attribute_value_state}
-              == ATTRIBUTE_VALUE_SINGLE_QUOTED_STATE ? 0x0027 : # '
-            -1);
-       unless (defined $token) {
-         !!!cp (117);
-         $self->{current_attribute}->{value} .= '&';
-       } else {
-         !!!cp (118);
-         $self->{current_attribute}->{value} .= $token->{data};
-         $self->{current_attribute}->{has_reference} = $token->{has_reference};
-         ## ISSUE: spec says "append the returned character token to the current attribute's value"
-       }
-       $self->{state} = $self->{last_attribute_value_state};
-       # next-input-character is already done
-       redo A;
      } elsif ($self->{state} == AFTER_ATTRIBUTE_VALUE_QUOTED_STATE) {
        if ($self->{next_char} == 0x0009 or # HT
            $self->{next_char} == 0x000A or # LF
-Line 1539 
 sub _get_next_token ($) {
+Line 1936 
 sub _get_next_token ($) {
          $self->{state} = SELF_CLOSING_START_TAG_STATE;
          !!!next-input-character;
          redo A;
+       } elsif ($self->{next_char} == -1) {
+         !!!parse-error (type => 'unclosed tag');
+         if ($self->{current_token}->{type} == START_TAG_TOKEN) {
+           !!!cp (122.3);
+           $self->{last_emitted_start_tag_name} = $self->{current_token}->{tag_name};
+         } elsif ($self->{current_token}->{type} == END_TAG_TOKEN) {
+           if ($self->{current_token}->{attributes}) {
+             !!!cp (122.1);
+             !!!parse-error (type => 'end tag attribute');
+           } else {
+             ## NOTE: This state should never be reached.
+             !!!cp (122.2);
+           }
+         } else {
+           die "$0: $self->{current_token}->{type}: Unknown token type";
+         }
+         $self->{state} = DATA_STATE;
+         ## Reconsume.
+         !!!emit ($self->{current_token}); # start tag or end tag
+         redo A;
        } else {
          !!!cp ('124.1');
          !!!parse-error (type => 'no space between attributes');
-Line 1571 
 sub _get_next_token ($) {
+Line 1988 
 sub _get_next_token ($) {
          !!!emit ($self->{current_token}); # start tag or end tag
          redo A;
+       } elsif ($self->{next_char} == -1) {
+         !!!parse-error (type => 'unclosed tag');
+         if ($self->{current_token}->{type} == START_TAG_TOKEN) {
+           !!!cp (124.7);
+           $self->{last_emitted_start_tag_name} = $self->{current_token}->{tag_name};
+         } elsif ($self->{current_token}->{type} == END_TAG_TOKEN) {
+           if ($self->{current_token}->{attributes}) {
+             !!!cp (124.5);
+             !!!parse-error (type => 'end tag attribute');
+           } else {
+             ## NOTE: This state should never be reached.
+             !!!cp (124.6);
+           }
+         } else {
+           die "$0: $self->{current_token}->{type}: Unknown token type";
+         }
+         $self->{state} = DATA_STATE;
+         ## Reconsume.
+         !!!emit ($self->{current_token}); # start tag or end tag
+         redo A;
        } else {
          !!!cp ('124.4');
          !!!parse-error (type => 'nestc');
-Line 1581 
 sub _get_next_token ($) {
+Line 2018 
 sub _get_next_token ($) {
        }
      } elsif ($self->{state} == BOGUS_COMMENT_STATE) {
        ## (only happen if PCDATA state)
-       ## NOTE: Set by the previous state
-       #my $token = {type => COMMENT_TOKEN, data => ''};
-       BC: {
-         if ($self->{next_char} == 0x003E) { # >
-           !!!cp (124);
-           $self->{state} = DATA_STATE;
-           !!!next-input-character;
-           !!!emit ($self->{current_token}); # comment
-           redo A;
-         } elsif ($self->{next_char} == -1) {
-           !!!cp (125);
-           $self->{state} = DATA_STATE;
-           ## reconsume
-           !!!emit ($self->{current_token}); # comment
+       ## NOTE: Unlike spec's "bogus comment state", this implementation
+       ## consumes characters one-by-one basis.
+       if ($self->{next_char} == 0x003E) { # >
+         !!!cp (124);
+         $self->{state} = DATA_STATE;
+         !!!next-input-character;
-           redo A;
+         !!!emit ($self->{current_token}); # comment
-         } else {
+         redo A;
-           !!!cp (126);
+       } elsif ($self->{next_char} == -1) {
-           $self->{current_token}->{data} .= chr ($self->{next_char}); # comment
+         !!!cp (125);
-           !!!next-input-character;
+         $self->{state} = DATA_STATE;
-           redo BC;
+         ## reconsume
-         }
-       } # BC
-       die "$0: _get_next_token: unexpected case [BC]";
+         !!!emit ($self->{current_token}); # comment
+         redo A;
+       } else {
+         !!!cp (126);
+         $self->{current_token}->{data} .= chr ($self->{next_char}); # comment
+         ## Stay in the state.
+         !!!next-input-character;
+         redo A;
+       }
      } elsif ($self->{state} == MARKUP_DECLARATION_OPEN_STATE) {
        ## (only happen if PCDATA state)
-       my ($l, $c) = ($self->{line_prev}, $self->{column_prev} - 1);
-       my @next_char;
-       push @next_char, $self->{next_char};
        if ($self->{next_char} == 0x002D) { # -
+         !!!cp (133);
+         $self->{state} = MD_HYPHEN_STATE;
          !!!next-input-character;
-         push @next_char, $self->{next_char};
+         redo A;
-         if ($self->{next_char} == 0x002D) { # -
-           !!!cp (127);
-           $self->{current_token} = {type => COMMENT_TOKEN, data => '',
-                                     line => $l, column => $c,
-                                    };
-           $self->{state} = COMMENT_START_STATE;
-           !!!next-input-character;
-           redo A;
-         } else {
-           !!!cp (128);
-         }
        } elsif ($self->{next_char} == 0x0044 or # D
                 $self->{next_char} == 0x0064) { # d
+         ## ASCII case-insensitive.
+         !!!cp (130);
+         $self->{state} = MD_DOCTYPE_STATE;
+         $self->{state_keyword} = chr $self->{next_char};
          !!!next-input-character;
-         push @next_char, $self->{next_char};
+         redo A;
-         if ($self->{next_char} == 0x004F or # O
-             $self->{next_char} == 0x006F) { # o
-           !!!next-input-character;
-           push @next_char, $self->{next_char};
-           if ($self->{next_char} == 0x0043 or # C
-               $self->{next_char} == 0x0063) { # c
-             !!!next-input-character;
-             push @next_char, $self->{next_char};
-             if ($self->{next_char} == 0x0054 or # T
-                 $self->{next_char} == 0x0074) { # t
-               !!!next-input-character;
-               push @next_char, $self->{next_char};
-               if ($self->{next_char} == 0x0059 or # Y
-                   $self->{next_char} == 0x0079) { # y
-                 !!!next-input-character;
-                 push @next_char, $self->{next_char};
-                 if ($self->{next_char} == 0x0050 or # P
-                     $self->{next_char} == 0x0070) { # p
-                   !!!next-input-character;
-                   push @next_char, $self->{next_char};
-                   if ($self->{next_char} == 0x0045 or # E
-                       $self->{next_char} == 0x0065) { # e
-                     !!!cp (129);
-                     ## TODO: What a stupid code this is!
-                     $self->{state} = DOCTYPE_STATE;
-                     $self->{current_token} = {type => DOCTYPE_TOKEN,
-                                               quirks => 1,
-                                               line => $l, column => $c,
-                                              };
-                     !!!next-input-character;
-                     redo A;
-                   } else {
-                     !!!cp (130);
-                   }
-                 } else {
-                   !!!cp (131);
-                 }
-               } else {
-                 !!!cp (132);
-               }
-             } else {
-               !!!cp (133);
-             }
-           } else {
-             !!!cp (134);
-           }
-         } else {
-           !!!cp (135);
-         }
        } elsif ($self->{insertion_mode} & IN_FOREIGN_CONTENT_IM and
                 $self->{open_elements}->[-1]->[1] & FOREIGN_EL and
                 $self->{next_char} == 0x005B) { # [
+         !!!cp (135.4);
+         $self->{state} = MD_CDATA_STATE;
+         $self->{state_keyword} = '[';
          !!!next-input-character;
-         push @next_char, $self->{next_char};
+         redo A;
-         if ($self->{next_char} == 0x0043) { # C
-           !!!next-input-character;
-           push @next_char, $self->{next_char};
-           if ($self->{next_char} == 0x0044) { # D
-             !!!next-input-character;
-             push @next_char, $self->{next_char};
-             if ($self->{next_char} == 0x0041) { # A
-               !!!next-input-character;
-               push @next_char, $self->{next_char};
-               if ($self->{next_char} == 0x0054) { # T
-                 !!!next-input-character;
-                 push @next_char, $self->{next_char};
-                 if ($self->{next_char} == 0x0041) { # A
-                   !!!next-input-character;
-                   push @next_char, $self->{next_char};
-                   if ($self->{next_char} == 0x005B) { # [
-                     !!!cp (135.1);
-                     $self->{state} = CDATA_BLOCK_STATE;
-                     !!!next-input-character;
-                     redo A;
-                   } else {
-                     !!!cp (135.2);
-                   }
-                 } else {
-                   !!!cp (135.3);
-                 }
-               } else {
-                 !!!cp (135.4);
-               }
-             } else {
-               !!!cp (135.5);
-             }
-           } else {
-             !!!cp (135.6);
-           }
-         } else {
-           !!!cp (135.7);
-         }
        } else {
          !!!cp (136);
        }
-       !!!parse-error (type => 'bogus comment');
+       !!!parse-error (type => 'bogus comment',
-       $self->{next_char} = shift @next_char;
+                       line => $self->{line_prev},
-       !!!back-next-input-character (@next_char);
+                       column => $self->{column_prev} - 1);
+       ## Reconsume.
        $self->{state} = BOGUS_COMMENT_STATE;
        $self->{current_token} = {type => COMMENT_TOKEN, data => '',
-                                 line => $l, column => $c,
+                                 line => $self->{line_prev},
+                                 column => $self->{column_prev} - 1,
                                 };
        redo A;
+     } elsif ($self->{state} == MD_HYPHEN_STATE) {
-       ## ISSUE: typos in spec: chacacters, is is a parse error
+       if ($self->{next_char} == 0x002D) { # -
-       ## ISSUE: spec is somewhat unclear on "is the first character that will be in the comment"; what is "that will be in the comment" is what the algorithm defines, isn't it?
+         !!!cp (127);
+         $self->{current_token} = {type => COMMENT_TOKEN, data => '',
+                                   line => $self->{line_prev},
+                                   column => $self->{column_prev} - 2,
+                                  };
+         $self->{state} = COMMENT_START_STATE;
+         !!!next-input-character;
+         redo A;
+       } else {
+         !!!cp (128);
+         !!!parse-error (type => 'bogus comment',
+                         line => $self->{line_prev},
+                         column => $self->{column_prev} - 2);
+         $self->{state} = BOGUS_COMMENT_STATE;
+         ## Reconsume.
+         $self->{current_token} = {type => COMMENT_TOKEN,
+                                   data => '-',
+                                   line => $self->{line_prev},
+                                   column => $self->{column_prev} - 2,
+                                  };
+         redo A;
+       }
+     } elsif ($self->{state} == MD_DOCTYPE_STATE) {
+       ## ASCII case-insensitive.
+       if ($self->{next_char} == [
+             undef,
+x004F, # O
+x0043, # C
+x0054, # T
+x0059, # Y
+x0050, # P
+           ]->[length $self->{state_keyword}] or
+           $self->{next_char} == [
+             undef,
+x006F, # o
+x0063, # c
+x0074, # t
+x0079, # y
+x0070, # p
+           ]->[length $self->{state_keyword}]) {
+         !!!cp (131);
+         ## Stay in the state.
+         $self->{state_keyword} .= chr $self->{next_char};
+         !!!next-input-character;
+         redo A;
+       } elsif ((length $self->{state_keyword}) == 6 and
+                ($self->{next_char} == 0x0045 or # E
+                 $self->{next_char} == 0x0065)) { # e
+         !!!cp (129);
+         $self->{state} = DOCTYPE_STATE;
+         $self->{current_token} = {type => DOCTYPE_TOKEN,
+                                   quirks => 1,
+                                   line => $self->{line_prev},
+                                   column => $self->{column_prev} - 7,
+                                  };
+         !!!next-input-character;
+         redo A;
+       } else {
+         !!!cp (132);
+         !!!parse-error (type => 'bogus comment',
+                         line => $self->{line_prev},
+                         column => $self->{column_prev} - 1 - length $self->{state_keyword});
+         $self->{state} = BOGUS_COMMENT_STATE;
+         ## Reconsume.
+         $self->{current_token} = {type => COMMENT_TOKEN,
+                                   data => $self->{state_keyword},
+                                   line => $self->{line_prev},
+                                   column => $self->{column_prev} - 1 - length $self->{state_keyword},
+                                  };
+         redo A;
+       }
+     } elsif ($self->{state} == MD_CDATA_STATE) {
+       if ($self->{next_char} == {
+             '[' => 0x0043, # C
+             '[C' => 0x0044, # D
+             '[CD' => 0x0041, # A
+             '[CDA' => 0x0054, # T
+             '[CDAT' => 0x0041, # A
+           }->{$self->{state_keyword}}) {
+         !!!cp (135.1);
+         ## Stay in the state.
+         $self->{state_keyword} .= chr $self->{next_char};
+         !!!next-input-character;
+         redo A;
+       } elsif ($self->{state_keyword} eq '[CDATA' and
+                $self->{next_char} == 0x005B) { # [
+         !!!cp (135.2);
+         $self->{current_token} = {type => CHARACTER_TOKEN,
+                                   data => '',
+                                   line => $self->{line_prev},
+                                   column => $self->{column_prev} - 7};
+         $self->{state} = CDATA_SECTION_STATE;
+         !!!next-input-character;
+         redo A;
+       } else {
+         !!!cp (135.3);
+         !!!parse-error (type => 'bogus comment',
+                         line => $self->{line_prev},
+                         column => $self->{column_prev} - 1 - length $self->{state_keyword});
+         $self->{state} = BOGUS_COMMENT_STATE;
+         ## Reconsume.
+         $self->{current_token} = {type => COMMENT_TOKEN,
+                                   data => $self->{state_keyword},
+                                   line => $self->{line_prev},
+                                   column => $self->{column_prev} - 1 - length $self->{state_keyword},
+                                  };
+         redo A;
+       }
      } elsif ($self->{state} == COMMENT_START_STATE) {
        if ($self->{next_char} == 0x002D) { # -
          !!!cp (137);
-Line 2010 
 sub _get_next_token ($) {
+Line 2457 
 sub _get_next_token ($) {
          redo A;
        } elsif ($self->{next_char} == 0x0050 or # P
                 $self->{next_char} == 0x0070) { # p
+         $self->{state} = PUBLIC_STATE;
+         $self->{state_keyword} = chr $self->{next_char};
          !!!next-input-character;
-         if ($self->{next_char} == 0x0055 or # U
+         redo A;
-             $self->{next_char} == 0x0075) { # u
-           !!!next-input-character;
-           if ($self->{next_char} == 0x0042 or # B
-               $self->{next_char} == 0x0062) { # b
-             !!!next-input-character;
-             if ($self->{next_char} == 0x004C or # L
-                 $self->{next_char} == 0x006C) { # l
-               !!!next-input-character;
-               if ($self->{next_char} == 0x0049 or # I
-                   $self->{next_char} == 0x0069) { # i
-                 !!!next-input-character;
-                 if ($self->{next_char} == 0x0043 or # C
-                     $self->{next_char} == 0x0063) { # c
-                   !!!cp (168);
-                   $self->{state} = BEFORE_DOCTYPE_PUBLIC_IDENTIFIER_STATE;
-                   !!!next-input-character;
-                   redo A;
-                 } else {
-                   !!!cp (169);
-                 }
-               } else {
-                 !!!cp (170);
-               }
-             } else {
-               !!!cp (171);
-             }
-           } else {
-             !!!cp (172);
-           }
-         } else {
-           !!!cp (173);
-         }
-         #
        } elsif ($self->{next_char} == 0x0053 or # S
                 $self->{next_char} == 0x0073) { # s
+         $self->{state} = SYSTEM_STATE;
+         $self->{state_keyword} = chr $self->{next_char};
          !!!next-input-character;
-         if ($self->{next_char} == 0x0059 or # Y
+         redo A;
-             $self->{next_char} == 0x0079) { # y
-           !!!next-input-character;
-           if ($self->{next_char} == 0x0053 or # S
-               $self->{next_char} == 0x0073) { # s
-             !!!next-input-character;
-             if ($self->{next_char} == 0x0054 or # T
-                 $self->{next_char} == 0x0074) { # t
-               !!!next-input-character;
-               if ($self->{next_char} == 0x0045 or # E
-                   $self->{next_char} == 0x0065) { # e
-                 !!!next-input-character;
-                 if ($self->{next_char} == 0x004D or # M
-                     $self->{next_char} == 0x006D) { # m
-                   !!!cp (174);
-                   $self->{state} = BEFORE_DOCTYPE_SYSTEM_IDENTIFIER_STATE;
-                   !!!next-input-character;
-                   redo A;
-                 } else {
-                   !!!cp (175);
-                 }
-               } else {
-                 !!!cp (176);
-               }
-             } else {
-               !!!cp (177);
-             }
-           } else {
-             !!!cp (178);
-           }
-         } else {
-           !!!cp (179);
-         }
-         #
        } else {
          !!!cp (180);
+         !!!parse-error (type => 'string after DOCTYPE name');
+         $self->{current_token}->{quirks} = 1;
+         $self->{state} = BOGUS_DOCTYPE_STATE;
          !!!next-input-character;
-         #
+         redo A;
        }
+     } elsif ($self->{state} == PUBLIC_STATE) {
+       ## ASCII case-insensitive
+       if ($self->{next_char} == [
+             undef,
+x0055, # U
+x0042, # B
+x004C, # L
+x0049, # I
+           ]->[length $self->{state_keyword}] or
+           $self->{next_char} == [
+             undef,
+x0075, # u
+x0062, # b
+x006C, # l
+x0069, # i
+           ]->[length $self->{state_keyword}]) {
+         !!!cp (175);
+         ## Stay in the state.
+         $self->{state_keyword} .= chr $self->{next_char};
+         !!!next-input-character;
+         redo A;
+       } elsif ((length $self->{state_keyword}) == 5 and
+                ($self->{next_char} == 0x0043 or # C
+                 $self->{next_char} == 0x0063)) { # c
+         !!!cp (168);
+         $self->{state} = BEFORE_DOCTYPE_PUBLIC_IDENTIFIER_STATE;
+         !!!next-input-character;
+         redo A;
+       } else {
+         !!!cp (169);
+         !!!parse-error (type => 'string after DOCTYPE name',
+                         line => $self->{line_prev},
+                         column => $self->{column_prev} + 1 - length $self->{state_keyword});
+         $self->{current_token}->{quirks} = 1;
-       !!!parse-error (type => 'string after DOCTYPE name');
+         $self->{state} = BOGUS_DOCTYPE_STATE;
-       $self->{current_token}->{quirks} = 1;
+         ## Reconsume.
+         redo A;
+       }
+     } elsif ($self->{state} == SYSTEM_STATE) {
+       ## ASCII case-insensitive
+       if ($self->{next_char} == [
+             undef,
+x0059, # Y
+x0053, # S
+x0054, # T
+x0045, # E
+           ]->[length $self->{state_keyword}] or
+           $self->{next_char} == [
+             undef,
+x0079, # y
+x0073, # s
+x0074, # t
+x0065, # e
+           ]->[length $self->{state_keyword}]) {
+         !!!cp (170);
+         ## Stay in the state.
+         $self->{state_keyword} .= chr $self->{next_char};
+         !!!next-input-character;
+         redo A;
+       } elsif ((length $self->{state_keyword}) == 5 and
+                ($self->{next_char} == 0x004D or # M
+                 $self->{next_char} == 0x006D)) { # m
+         !!!cp (171);
+         $self->{state} = BEFORE_DOCTYPE_SYSTEM_IDENTIFIER_STATE;
+         !!!next-input-character;
+         redo A;
+       } else {
+         !!!cp (172);
+         !!!parse-error (type => 'string after DOCTYPE name',
+                         line => $self->{line_prev},
+                         column => $self->{column_prev} + 1 - length $self->{state_keyword});
+         $self->{current_token}->{quirks} = 1;
-       $self->{state} = BOGUS_DOCTYPE_STATE;
+         $self->{state} = BOGUS_DOCTYPE_STATE;
-       # next-input-character is already done
+         ## Reconsume.
-       redo A;
+         redo A;
+       }
      } elsif ($self->{state} == BEFORE_DOCTYPE_PUBLIC_IDENTIFIER_STATE) {
        if ({
 x0009 => 1, 0x000A => 1, 0x000B => 1, 0x000C => 1, 0x0020 => 1,
-Line 2328 
 sub _get_next_token ($) {
+Line 2786 
 sub _get_next_token ($) {
          redo A;
        } elsif ($self->{next_char} == 0x003E) { # >
          !!!cp (208);
-         !!!parse-error (type => 'unclosed PUBLIC literal');
+         !!!parse-error (type => 'unclosed SYSTEM literal');
          $self->{state} = DATA_STATE;
          !!!next-input-character;
-Line 2364 
 sub _get_next_token ($) {
+Line 2822 
 sub _get_next_token ($) {
          redo A;
        } elsif ($self->{next_char} == 0x003E) { # >
          !!!cp (212);
-         !!!parse-error (type => 'unclosed PUBLIC literal');
+         !!!parse-error (type => 'unclosed SYSTEM literal');
          $self->{state} = DATA_STATE;
          !!!next-input-character;
-Line 2412 
 sub _get_next_token ($) {
+Line 2870 
 sub _get_next_token ($) {
        } elsif ($self->{next_char} == -1) {
          !!!cp (217);
          !!!parse-error (type => 'unclosed DOCTYPE');
          $self->{state} = DATA_STATE;
          ## reconsume
-Line 2453 
 sub _get_next_token ($) {
+Line 2910 
 sub _get_next_token ($) {
          !!!next-input-character;
          redo A;
        }
-     } elsif ($self->{state} == CDATA_BLOCK_STATE) {
+     } elsif ($self->{state} == CDATA_SECTION_STATE) {
-       my $s = '';
+       ## NOTE: "CDATA section state" in the state is jointly implemented
+       ## by three states, |CDATA_SECTION_STATE|, |CDATA_SECTION_MSE1_STATE|,
+       ## and |CDATA_SECTION_MSE2_STATE|.
-       my ($l, $c) = ($self->{line}, $self->{column});
+       if ($self->{next_char} == 0x005D) { # ]
+         !!!cp (221.1);
+         $self->{state} = CDATA_SECTION_MSE1_STATE;
+         !!!next-input-character;
+         redo A;
+       } elsif ($self->{next_char} == -1) {
+         $self->{state} = DATA_STATE;
+         !!!next-input-character;
+         if (length $self->{current_token}->{data}) { # character
+           !!!cp (221.2);
+           !!!emit ($self->{current_token}); # character
+         } else {
+           !!!cp (221.3);
+           ## No token to emit. $self->{current_token} is discarded.
+         }
+         redo A;
+       } else {
+         !!!cp (221.4);
+         $self->{current_token}->{data} .= chr $self->{next_char};
+         ## Stay in the state.
+         !!!next-input-character;
+         redo A;
+       }
-       CS: while ($self->{next_char} != -1) {
+       ## ISSUE: "text tokens" in spec.
-         if ($self->{next_char} == 0x005D) { # ]
+     } elsif ($self->{state} == CDATA_SECTION_MSE1_STATE) {
-           !!!next-input-character;
+       if ($self->{next_char} == 0x005D) { # ]
-           if ($self->{next_char} == 0x005D) { # ]
+         !!!cp (221.5);
-             !!!next-input-character;
+         $self->{state} = CDATA_SECTION_MSE2_STATE;
-             MDC: {
+         !!!next-input-character;
-               if ($self->{next_char} == 0x003E) { # >
+         redo A;
-                 !!!cp (221.1);
+       } else {
-                 !!!next-input-character;
+         !!!cp (221.6);
-                 last CS;
+         $self->{current_token}->{data} .= ']';
-               } elsif ($self->{next_char} == 0x005D) { # ]
+         $self->{state} = CDATA_SECTION_STATE;
-                 !!!cp (221.2);
+         ## Reconsume.
-                 $s .= ']';
+         redo A;
-                 !!!next-input-character;
+       }
-                 redo MDC;
+     } elsif ($self->{state} == CDATA_SECTION_MSE2_STATE) {
-               } else {
+       if ($self->{next_char} == 0x003E) { # >
-                 !!!cp (221.3);
+         $self->{state} = DATA_STATE;
-                 $s .= ']]';
+         !!!next-input-character;
-                 #
+         if (length $self->{current_token}->{data}) { # character
-               }
+           !!!cp (221.7);
-             } # MDC
+           !!!emit ($self->{current_token}); # character
-           } else {
-             !!!cp (221.4);
-             $s .= ']';
-             #
-           }
          } else {
-           !!!cp (221.5);
+           !!!cp (221.8);
-           #
+           ## No token to emit. $self->{current_token} is discarded.
          }
-         $s .= chr $self->{next_char};
+         redo A;
+       } elsif ($self->{next_char} == 0x005D) { # ]
+         !!!cp (221.9); # character
+         $self->{current_token}->{data} .= ']'; ## Add first "]" of "]]]".
+         ## Stay in the state.
          !!!next-input-character;
-       } # CS
+         redo A;
-       $self->{state} = DATA_STATE;
-       ## next-input-character done or EOF, which is reconsumed.
-       if (length $s) {
-         !!!cp (221.6);
-         !!!emit ({type => CHARACTER_TOKEN, data => $s,
-                   line => $l, column => $c});
        } else {
-         !!!cp (221.7);
+         !!!cp (221.11);
+         $self->{current_token}->{data} .= ']]'; # character
+         $self->{state} = CDATA_SECTION_STATE;
+         ## Reconsume.
+         redo A;
+       }
+     } elsif ($self->{state} == ENTITY_STATE) {
+       if ({
+x0009 => 1, 0x000A => 1, 0x000B => 1, 0x000C => 1, # HT, LF, VT, FF,
+x0020 => 1, 0x003C => 1, 0x0026 => 1, -1 => 1, # SP, <, &
+         $self->{entity_additional} => 1,
+       }->{$self->{next_char}}) {
+         !!!cp (1001);
+         ## Don't consume
+         ## No error
+         ## Return nothing.
+         #
+       } elsif ($self->{next_char} == 0x0023) { # #
+         !!!cp (999);
+         $self->{state} = ENTITY_HASH_STATE;
+         $self->{state_keyword} = '#';
+         !!!next-input-character;
+         redo A;
+       } elsif ((0x0041 <= $self->{next_char} and
+                 $self->{next_char} <= 0x005A) or # A..Z
+                (0x0061 <= $self->{next_char} and
+                 $self->{next_char} <= 0x007A)) { # a..z
+         !!!cp (998);
+         require Whatpm::_NamedEntityList;
+         $self->{state} = ENTITY_NAME_STATE;
+         $self->{state_keyword} = chr $self->{next_char};
+         $self->{entity__value} = $self->{state_keyword};
+         $self->{entity__match} = 0;
+         !!!next-input-character;
+         redo A;
+       } else {
+         !!!cp (1027);
+         !!!parse-error (type => 'bare ero');
+         ## Return nothing.
+         #
        }
-       redo A;
+       ## NOTE: No character is consumed by the "consume a character
+       ## reference" algorithm.  In other word, there is an "&" character
-       ## ISSUE: "text tokens" in spec.
+       ## that does not introduce a character reference, which would be
-       ## TODO: Streaming support
+       ## appended to the parent element or the attribute value in later
-     } else {
+       ## process of the tokenizer.
-       die "$0: $self->{state}: Unknown state";
-     }
+       if ($self->{prev_state} == DATA_STATE) {
-   } # A
+         !!!cp (997);
+         $self->{state} = $self->{prev_state};
-   die "$0: _get_next_token: unexpected case";
+         ## Reconsume.
- } # _get_next_token
+         !!!emit ({type => CHARACTER_TOKEN, data => '&',
+                   line => $self->{line_prev},
- sub _tokenize_attempt_to_consume_an_entity ($$$) {
+                   column => $self->{column_prev},
-   my ($self, $in_attr, $additional) = @_;
+                  });
+         redo A;
-   my ($l, $c) = ($self->{line_prev}, $self->{column_prev});
+       } else {
+         !!!cp (996);
+         $self->{current_attribute}->{value} .= '&';
+         $self->{state} = $self->{prev_state};
+         ## Reconsume.
+         redo A;
+       }
+     } elsif ($self->{state} == ENTITY_HASH_STATE) {
+       if ($self->{next_char} == 0x0078 or # x
+           $self->{next_char} == 0x0058) { # X
+         !!!cp (995);
+         $self->{state} = HEXREF_X_STATE;
+         $self->{state_keyword} .= chr $self->{next_char};
+         !!!next-input-character;
+         redo A;
+       } elsif (0x0030 <= $self->{next_char} and
+                $self->{next_char} <= 0x0039) { # 0..9
+         !!!cp (994);
+         $self->{state} = NCR_NUM_STATE;
+         $self->{state_keyword} = $self->{next_char} - 0x0030;
+         !!!next-input-character;
+         redo A;
+       } else {
+         !!!parse-error (type => 'bare nero',
+                         line => $self->{line_prev},
+                         column => $self->{column_prev} - 1);
-   if ({
+         ## NOTE: According to the spec algorithm, nothing is returned,
-x0009 => 1, 0x000A => 1, 0x000B => 1, 0x000C => 1, # HT, LF, VT, FF,
+         ## and then "&#" is appended to the parent element or the attribute
-x0020 => 1, 0x003C => 1, 0x0026 => 1, -1 => 1, # SP, <, & # 0x000D # CR
+         ## value in the later processing.
-        $additional => 1,
-       }->{$self->{next_char}}) {
+         if ($self->{prev_state} == DATA_STATE) {
-     !!!cp (1001);
+           !!!cp (1019);
-     ## Don't consume
+           $self->{state} = $self->{prev_state};
-     ## No error
+           ## Reconsume.
-     return undef;
+           !!!emit ({type => CHARACTER_TOKEN,
-   } elsif ($self->{next_char} == 0x0023) { # #
+                     data => '&#',
-     !!!next-input-character;
+                     line => $self->{line_prev},
-     if ($self->{next_char} == 0x0078 or # x
+                     column => $self->{column_prev} - 1,
-         $self->{next_char} == 0x0058) { # X
+                    });
-       my $code;
+           redo A;
-       X: {
-         my $x_char = $self->{next_char};
-         !!!next-input-character;
-         if (0x0030 <= $self->{next_char} and
-             $self->{next_char} <= 0x0039) { # 0..9
-           !!!cp (1002);
-           $code ||= 0;
-           $code *= 0x10;
-           $code += $self->{next_char} - 0x0030;
-           redo X;
-         } elsif (0x0061 <= $self->{next_char} and
-                  $self->{next_char} <= 0x0066) { # a..f
-           !!!cp (1003);
-           $code ||= 0;
-           $code *= 0x10;
-           $code += $self->{next_char} - 0x0060 + 9;
-           redo X;
-         } elsif (0x0041 <= $self->{next_char} and
-                  $self->{next_char} <= 0x0046) { # A..F
-           !!!cp (1004);
-           $code ||= 0;
-           $code *= 0x10;
-           $code += $self->{next_char} - 0x0040 + 9;
-           redo X;
-         } elsif (not defined $code) { # no hexadecimal digit
-           !!!cp (1005);
-           !!!parse-error (type => 'bare hcro', line => $l, column => $c);
-           !!!back-next-input-character ($x_char, $self->{next_char});
-           $self->{next_char} = 0x0023; # #
-           return undef;
-         } elsif ($self->{next_char} == 0x003B) { # ;
-           !!!cp (1006);
-           !!!next-input-character;
          } else {
-           !!!cp (1007);
+           !!!cp (993);
-           !!!parse-error (type => 'no refc', line => $l, column => $c);
+           $self->{current_attribute}->{value} .= '&#';
+           $self->{state} = $self->{prev_state};
+           ## Reconsume.
+           redo A;
          }
+       }
-         if ($code == 0 or (0xD800 <= $code and $code <= 0xDFFF)) {
+     } elsif ($self->{state} == NCR_NUM_STATE) {
-           !!!cp (1008);
+       if (0x0030 <= $self->{next_char} and
-           !!!parse-error (type => (sprintf 'invalid character reference:U+%04X', $code), line => $l, column => $c);
+           $self->{next_char} <= 0x0039) { # 0..9
-           $code = 0xFFFD;
-         } elsif ($code > 0x10FFFF) {
-           !!!cp (1009);
-           !!!parse-error (type => (sprintf 'invalid character reference:U-%08X', $code), line => $l, column => $c);
-           $code = 0xFFFD;
-         } elsif ($code == 0x000D) {
-           !!!cp (1010);
-           !!!parse-error (type => 'CR character reference', line => $l, column => $c);
-           $code = 0x000A;
-         } elsif (0x80 <= $code and $code <= 0x9F) {
-           !!!cp (1011);
-           !!!parse-error (type => (sprintf 'C1 character reference:U+%04X', $code), line => $l, column => $c);
-           $code = $c1_entity_char->{$code};
-         }
-         return {type => CHARACTER_TOKEN, data => chr $code,
-                 has_reference => 1,
-                 line => $l, column => $c,
-                };
-       } # X
-     } elsif (0x0030 <= $self->{next_char} and
-              $self->{next_char} <= 0x0039) { # 0..9
-       my $code = $self->{next_char} - 0x0030;
-       !!!next-input-character;
-       while (0x0030 <= $self->{next_char} and
-                 $self->{next_char} <= 0x0039) { # 0..9
          !!!cp (1012);
-         $code *= 10;
+         $self->{state_keyword} *= 10;
-         $code += $self->{next_char} - 0x0030;
+         $self->{state_keyword} += $self->{next_char} - 0x0030;
+         ## Stay in the state.
          !!!next-input-character;
-       }
+         redo A;
+       } elsif ($self->{next_char} == 0x003B) { # ;
-       if ($self->{next_char} == 0x003B) { # ;
          !!!cp (1013);
          !!!next-input-character;
+         #
        } else {
          !!!cp (1014);
-         !!!parse-error (type => 'no refc', line => $l, column => $c);
+         !!!parse-error (type => 'no refc');
+         ## Reconsume.
+         #
        }
+       my $code = $self->{state_keyword};
+       my $l = $self->{line_prev};
+       my $c = $self->{column_prev};
        if ($code == 0 or (0xD800 <= $code and $code <= 0xDFFF)) {
          !!!cp (1015);
-         !!!parse-error (type => (sprintf 'invalid character reference:U+%04X', $code), line => $l, column => $c);
+         !!!parse-error (type => 'invalid character reference',
+                         text => (sprintf 'U+%04X', $code),
+                         line => $l, column => $c);
          $code = 0xFFFD;
        } elsif ($code > 0x10FFFF) {
          !!!cp (1016);
-         !!!parse-error (type => (sprintf 'invalid character reference:U-%08X', $code), line => $l, column => $c);
+         !!!parse-error (type => 'invalid character reference',
+                         text => (sprintf 'U-%08X', $code),
+                         line => $l, column => $c);
          $code = 0xFFFD;
        } elsif ($code == 0x000D) {
          !!!cp (1017);
-         !!!parse-error (type => 'CR character reference', line => $l, column => $c);
+         !!!parse-error (type => 'CR character reference',
+                         line => $l, column => $c);
          $code = 0x000A;
        } elsif (0x80 <= $code and $code <= 0x9F) {
          !!!cp (1018);
-         !!!parse-error (type => (sprintf 'C1 character reference:U+%04X', $code), line => $l, column => $c);
+         !!!parse-error (type => 'C1 character reference',
+                         text => (sprintf 'U+%04X', $code),
+                         line => $l, column => $c);
          $code = $c1_entity_char->{$code};
        }
-       return {type => CHARACTER_TOKEN, data => chr $code, has_reference => 1,
+       if ($self->{prev_state} == DATA_STATE) {
-               line => $l, column => $c,
+         !!!cp (992);
-              };
+         $self->{state} = $self->{prev_state};
-     } else {
+         ## Reconsume.
-       !!!cp (1019);
+         !!!emit ({type => CHARACTER_TOKEN, data => chr $code,
-       !!!parse-error (type => 'bare nero', line => $l, column => $c);
+                   line => $l, column => $c,
-       !!!back-next-input-character ($self->{next_char});
+                  });
-       $self->{next_char} = 0x0023; # #
+         redo A;
-       return undef;
+       } else {
-     }
+         !!!cp (991);
-   } elsif ((0x0041 <= $self->{next_char} and
+         $self->{current_attribute}->{value} .= chr $code;
-             $self->{next_char} <= 0x005A) or
+         $self->{current_attribute}->{has_reference} = 1;
-            (0x0061 <= $self->{next_char} and
+         $self->{state} = $self->{prev_state};
-             $self->{next_char} <= 0x007A)) {
+         ## Reconsume.
-     my $entity_name = chr $self->{next_char};
+         redo A;
-     !!!next-input-character;
+       }
+     } elsif ($self->{state} == HEXREF_X_STATE) {
-     my $value = $entity_name;
+       if ((0x0030 <= $self->{next_char} and $self->{next_char} <= 0x0039) or
-     my $match = 0;
+           (0x0041 <= $self->{next_char} and $self->{next_char} <= 0x0046) or
-     require Whatpm::_NamedEntityList;
+           (0x0061 <= $self->{next_char} and $self->{next_char} <= 0x0066)) {
-     our $EntityChar;
+         # 0..9, A..F, a..f
+         !!!cp (990);
-     while (length $entity_name < 30 and
+         $self->{state} = HEXREF_HEX_STATE;
-            ## NOTE: Some number greater than the maximum length of entity name
+         $self->{state_keyword} = 0;
-            ((0x0041 <= $self->{next_char} and # a
+         ## Reconsume.
-              $self->{next_char} <= 0x005A) or # x
+         redo A;
-             (0x0061 <= $self->{next_char} and # a
+       } else {
-              $self->{next_char} <= 0x007A) or # z
+         !!!parse-error (type => 'bare hcro',
-             (0x0030 <= $self->{next_char} and # 0
+                         line => $self->{line_prev},
-              $self->{next_char} <= 0x0039) or # 9
+                         column => $self->{column_prev} - 2);
-             $self->{next_char} == 0x003B)) { # ;
-       $entity_name .= chr $self->{next_char};
+         ## NOTE: According to the spec algorithm, nothing is returned,
-       if (defined $EntityChar->{$entity_name}) {
+         ## and then "&#" followed by "X" or "x" is appended to the parent
-         if ($self->{next_char} == 0x003B) { # ;
+         ## element or the attribute value in the later processing.
-           !!!cp (1020);
-           $value = $EntityChar->{$entity_name};
+         if ($self->{prev_state} == DATA_STATE) {
-           $match = 1;
+           !!!cp (1005);
-           !!!next-input-character;
+           $self->{state} = $self->{prev_state};
-           last;
+           ## Reconsume.
+           !!!emit ({type => CHARACTER_TOKEN,
+                     data => '&' . $self->{state_keyword},
+                     line => $self->{line_prev},
+                     column => $self->{column_prev} - length $self->{state_keyword},
+                    });
+           redo A;
          } else {
-           !!!cp (1021);
+           !!!cp (989);
-           $value = $EntityChar->{$entity_name};
+           $self->{current_attribute}->{value} .= '&' . $self->{state_keyword};
-           $match = -1;
+           $self->{state} = $self->{prev_state};
+           ## Reconsume.
+           redo A;
+         }
+       }
+     } elsif ($self->{state} == HEXREF_HEX_STATE) {
+       if (0x0030 <= $self->{next_char} and $self->{next_char} <= 0x0039) {
+         # 0..9
+         !!!cp (1002);
+         $self->{state_keyword} *= 0x10;
+         $self->{state_keyword} += $self->{next_char} - 0x0030;
+         ## Stay in the state.
+         !!!next-input-character;
+         redo A;
+       } elsif (0x0061 <= $self->{next_char} and
+                $self->{next_char} <= 0x0066) { # a..f
+         !!!cp (1003);
+         $self->{state_keyword} *= 0x10;
+         $self->{state_keyword} += $self->{next_char} - 0x0060 + 9;
+         ## Stay in the state.
+         !!!next-input-character;
+         redo A;
+       } elsif (0x0041 <= $self->{next_char} and
+                $self->{next_char} <= 0x0046) { # A..F
+         !!!cp (1004);
+         $self->{state_keyword} *= 0x10;
+         $self->{state_keyword} += $self->{next_char} - 0x0040 + 9;
+         ## Stay in the state.
+         !!!next-input-character;
+         redo A;
+       } elsif ($self->{next_char} == 0x003B) { # ;
+         !!!cp (1006);
+         !!!next-input-character;
+         #
+       } else {
+         !!!cp (1007);
+         !!!parse-error (type => 'no refc',
+                         line => $self->{line},
+                         column => $self->{column});
+         ## Reconsume.
+         #
+       }
+       my $code = $self->{state_keyword};
+       my $l = $self->{line_prev};
+       my $c = $self->{column_prev};
+       if ($code == 0 or (0xD800 <= $code and $code <= 0xDFFF)) {
+         !!!cp (1008);
+         !!!parse-error (type => 'invalid character reference',
+                         text => (sprintf 'U+%04X', $code),
+                         line => $l, column => $c);
+         $code = 0xFFFD;
+       } elsif ($code > 0x10FFFF) {
+         !!!cp (1009);
+         !!!parse-error (type => 'invalid character reference',
+                         text => (sprintf 'U-%08X', $code),
+                         line => $l, column => $c);
+         $code = 0xFFFD;
+       } elsif ($code == 0x000D) {
+         !!!cp (1010);
+         !!!parse-error (type => 'CR character reference', line => $l, column => $c);
+         $code = 0x000A;
+       } elsif (0x80 <= $code and $code <= 0x9F) {
+         !!!cp (1011);
+         !!!parse-error (type => 'C1 character reference', text => (sprintf 'U+%04X', $code), line => $l, column => $c);
+         $code = $c1_entity_char->{$code};
+       }
+       if ($self->{prev_state} == DATA_STATE) {
+         !!!cp (988);
+         $self->{state} = $self->{prev_state};
+         ## Reconsume.
+         !!!emit ({type => CHARACTER_TOKEN, data => chr $code,
+                   line => $l, column => $c,
+                  });
+         redo A;
+       } else {
+         !!!cp (987);
+         $self->{current_attribute}->{value} .= chr $code;
+         $self->{current_attribute}->{has_reference} = 1;
+         $self->{state} = $self->{prev_state};
+         ## Reconsume.
+         redo A;
+       }
+     } elsif ($self->{state} == ENTITY_NAME_STATE) {
+       if (length $self->{state_keyword} < 30 and
+           ## NOTE: Some number greater than the maximum length of entity name
+           ((0x0041 <= $self->{next_char} and # a
+             $self->{next_char} <= 0x005A) or # x
+            (0x0061 <= $self->{next_char} and # a
+             $self->{next_char} <= 0x007A) or # z
+            (0x0030 <= $self->{next_char} and # 0
+             $self->{next_char} <= 0x0039) or # 9
+            $self->{next_char} == 0x003B)) { # ;
+         our $EntityChar;
+         $self->{state_keyword} .= chr $self->{next_char};
+         if (defined $EntityChar->{$self->{state_keyword}}) {
+           if ($self->{next_char} == 0x003B) { # ;
+             !!!cp (1020);
+             $self->{entity__value} = $EntityChar->{$self->{state_keyword}};
+             $self->{entity__match} = 1;
+             !!!next-input-character;
+             #
+           } else {
+             !!!cp (1021);
+             $self->{entity__value} = $EntityChar->{$self->{state_keyword}};
+             $self->{entity__match} = -1;
+             ## Stay in the state.
+             !!!next-input-character;
+             redo A;
+           }
+         } else {
+           !!!cp (1022);
+           $self->{entity__value} .= chr $self->{next_char};
+           $self->{entity__match} *= 2;
+           ## Stay in the state.
            !!!next-input-character;
+           redo A;
+         }
+       }
+       my $data;
+       my $has_ref;
+       if ($self->{entity__match} > 0) {
+         !!!cp (1023);
+         $data = $self->{entity__value};
+         $has_ref = 1;
+         #
+       } elsif ($self->{entity__match} < 0) {
+         !!!parse-error (type => 'no refc');
+         if ($self->{prev_state} != DATA_STATE and # in attribute
+             $self->{entity__match} < -1) {
+           !!!cp (1024);
+           $data = '&' . $self->{state_keyword};
+           #
+         } else {
+           !!!cp (1025);
+           $data = $self->{entity__value};
+           $has_ref = 1;
+           #
          }
        } else {
-         !!!cp (1022);
+         !!!cp (1026);
-         $value .= chr $self->{next_char};
+         !!!parse-error (type => 'bare ero',
-         $match *= 2;
+                         line => $self->{line_prev},
-         !!!next-input-character;
+                         column => $self->{column_prev});
+         $data = '&' . $self->{state_keyword};
+         #
        }
-     }
+       ## NOTE: In these cases, when a character reference is found,
-     if ($match > 0) {
+       ## it is consumed and a character token is returned, or, otherwise,
-       !!!cp (1023);
+       ## nothing is consumed and returned, according to the spec algorithm.
-       return {type => CHARACTER_TOKEN, data => $value, has_reference => 1,
+       ## In this implementation, anything that has been examined by the
-               line => $l, column => $c,
+       ## tokenizer is appended to the parent element or the attribute value
-              };
+       ## as string, either literal string when no character reference or
-     } elsif ($match < 0) {
+       ## entity-replaced string otherwise, in this stage, since any characters
-       !!!parse-error (type => 'no refc', line => $l, column => $c);
+       ## that would not be consumed are appended in the data state or in an
-       if ($in_attr and $match < -1) {
+       ## appropriate attribute value state anyway.
-         !!!cp (1024);
-         return {type => CHARACTER_TOKEN, data => '&'.$entity_name,
+       if ($self->{prev_state} == DATA_STATE) {
-                 line => $l, column => $c,
+         !!!cp (986);
-                };
+         $self->{state} = $self->{prev_state};
-       } else {
+         ## Reconsume.
-         !!!cp (1025);
+         !!!emit ({type => CHARACTER_TOKEN,
-         return {type => CHARACTER_TOKEN, data => $value, has_reference => 1,
+                   data => $data,
-                 line => $l, column => $c,
+                   line => $self->{line_prev},
-                };
+                   column => $self->{column_prev} + 1 - length $self->{state_keyword},
+                  });
+         redo A;
+       } else {
+         !!!cp (985);
+         $self->{current_attribute}->{value} .= $data;
+         $self->{current_attribute}->{has_reference} = 1 if $has_ref;
+         $self->{state} = $self->{prev_state};
+         ## Reconsume.
+         redo A;
        }
      } else {
-       !!!cp (1026);
+       die "$0: $self->{state}: Unknown state";
-       !!!parse-error (type => 'bare ero', line => $l, column => $c);
-       ## NOTE: "No characters are consumed" in the spec.
-       return {type => CHARACTER_TOKEN, data => '&'.$value,
-               line => $l, column => $c,
-              };
      }
-   } else {
+   } # A
-     !!!cp (1027);
-     ## no characters are consumed
+   die "$0: _get_next_token: unexpected case";
-     !!!parse-error (type => 'bare ero', line => $l, column => $c);
+ } # _get_next_token
-     return undef;
-   }
- } # _tokenize_attempt_to_consume_an_entity
  sub _initialize_tree_constructor ($) {
    my $self = shift;
-Line 2729 
 sub _initialize_tree_constructor ($) {
+Line 3368 
 sub _initialize_tree_constructor ($) {
    ## TODO: Turn mutation events off # MUST
    ## TODO: Turn loose Document option (manakai extension) on
    $self->{document}->manakai_is_html (1); # MUST
+   $self->{document}->set_user_data (manakai_source_line => 1);
+   $self->{document}->set_user_data (manakai_source_column => 1);
  } # _initialize_tree_constructor
  sub _terminate_tree_constructor ($) {
-Line 2783 
 sub _tree_construction_initial ($) {
+Line 3424 
 sub _tree_construction_initial ($) {
        ## language.
        my $doctype_name = $token->{name};
        $doctype_name = '' unless defined $doctype_name;
-       $doctype_name =~ tr/a-z/A-Z/;
+       $doctype_name =~ tr/a-z/A-Z/; # ASCII case-insensitive
        if (not defined $token->{name} or # <!DOCTYPE>
-           defined $token->{public_identifier} or
            defined $token->{system_identifier}) {
          !!!cp ('t1');
          !!!parse-error (type => 'not HTML5', token => $token);
        } elsif ($doctype_name ne 'HTML') {
          !!!cp ('t2');
-         ## ISSUE: ASCII case-insensitive? (in fact it does not matter)
          !!!parse-error (type => 'not HTML5', token => $token);
+       } elsif (defined $token->{public_identifier}) {
+         if ($token->{public_identifier} eq 'XSLT-compat') {
+           !!!cp ('t1.2');
+           !!!parse-error (type => 'XSLT-compat', token => $token,
+                           level => $self->{level}->{should});
+         } else {
+           !!!parse-error (type => 'not HTML5', token => $token);
+         }
        } else {
          !!!cp ('t3');
+         #
        }
        my $doctype = $self->{document}->create_document_type_definition
-Line 2815 
 sub _tree_construction_initial ($) {
+Line 3463 
 sub _tree_construction_initial ($) {
        } elsif (defined $token->{public_identifier}) {
          my $pubid = $token->{public_identifier};
          $pubid =~ tr/a-z/A-z/;
-         if ({
+         my $prefix = [
-           "+//SILMARIL//DTD HTML PRO V0R11 19970101//EN" => 1,
+           "+//SILMARIL//DTD HTML PRO V0R11 19970101//",
-           "-//ADVASOFT LTD//DTD HTML 3.0 ASWEDIT + EXTENSIONS//EN" => 1,
+           "-//ADVASOFT LTD//DTD HTML 3.0 ASWEDIT + EXTENSIONS//",
-           "-//AS//DTD HTML 3.0 ASWEDIT + EXTENSIONS//EN" => 1,
+           "-//AS//DTD HTML 3.0 ASWEDIT + EXTENSIONS//",
-           "-//IETF//DTD HTML 2.0 LEVEL 1//EN" => 1,
+           "-//IETF//DTD HTML 2.0 LEVEL 1//",
-           "-//IETF//DTD HTML 2.0 LEVEL 2//EN" => 1,
+           "-//IETF//DTD HTML 2.0 LEVEL 2//",
-           "-//IETF//DTD HTML 2.0 STRICT LEVEL 1//EN" => 1,
+           "-//IETF//DTD HTML 2.0 STRICT LEVEL 1//",
-           "-//IETF//DTD HTML 2.0 STRICT LEVEL 2//EN" => 1,
+           "-//IETF//DTD HTML 2.0 STRICT LEVEL 2//",
-           "-//IETF//DTD HTML 2.0 STRICT//EN" => 1,
+           "-//IETF//DTD HTML 2.0 STRICT//",
-           "-//IETF//DTD HTML 2.0//EN" => 1,
+           "-//IETF//DTD HTML 2.0//",
-           "-//IETF//DTD HTML 2.1E//EN" => 1,
+           "-//IETF//DTD HTML 2.1E//",
-           "-//IETF//DTD HTML 3.0//EN" => 1,
+           "-//IETF//DTD HTML 3.0//",
-           "-//IETF//DTD HTML 3.0//EN//" => 1,
+           "-//IETF//DTD HTML 3.2 FINAL//",
-           "-//IETF//DTD HTML 3.2 FINAL//EN" => 1,
+           "-//IETF//DTD HTML 3.2//",
-           "-//IETF//DTD HTML 3.2//EN" => 1,
+           "-//IETF//DTD HTML 3//",
-           "-//IETF//DTD HTML 3//EN" => 1,
+           "-//IETF//DTD HTML LEVEL 0//",
-           "-//IETF//DTD HTML LEVEL 0//EN" => 1,
+           "-//IETF//DTD HTML LEVEL 1//",
-           "-//IETF//DTD HTML LEVEL 0//EN//2.0" => 1,
+           "-//IETF//DTD HTML LEVEL 2//",
-           "-//IETF//DTD HTML LEVEL 1//EN" => 1,
+           "-//IETF//DTD HTML LEVEL 3//",
-           "-//IETF//DTD HTML LEVEL 1//EN//2.0" => 1,
+           "-//IETF//DTD HTML STRICT LEVEL 0//",
-           "-//IETF//DTD HTML LEVEL 2//EN" => 1,
+           "-//IETF//DTD HTML STRICT LEVEL 1//",
-           "-//IETF//DTD HTML LEVEL 2//EN//2.0" => 1,
+           "-//IETF//DTD HTML STRICT LEVEL 2//",
-           "-//IETF//DTD HTML LEVEL 3//EN" => 1,
+           "-//IETF//DTD HTML STRICT LEVEL 3//",
-           "-//IETF//DTD HTML LEVEL 3//EN//3.0" => 1,
+           "-//IETF//DTD HTML STRICT//",
-           "-//IETF//DTD HTML STRICT LEVEL 0//EN" => 1,
+           "-//IETF//DTD HTML//",
-           "-//IETF//DTD HTML STRICT LEVEL 0//EN//2.0" => 1,
+           "-//METRIUS//DTD METRIUS PRESENTATIONAL//",
-           "-//IETF//DTD HTML STRICT LEVEL 1//EN" => 1,
+           "-//MICROSOFT//DTD INTERNET EXPLORER 2.0 HTML STRICT//",
-           "-//IETF//DTD HTML STRICT LEVEL 1//EN//2.0" => 1,
+           "-//MICROSOFT//DTD INTERNET EXPLORER 2.0 HTML//",
-           "-//IETF//DTD HTML STRICT LEVEL 2//EN" => 1,
+           "-//MICROSOFT//DTD INTERNET EXPLORER 2.0 TABLES//",
-           "-//IETF//DTD HTML STRICT LEVEL 2//EN//2.0" => 1,
+           "-//MICROSOFT//DTD INTERNET EXPLORER 3.0 HTML STRICT//",
-           "-//IETF//DTD HTML STRICT LEVEL 3//EN" => 1,
+           "-//MICROSOFT//DTD INTERNET EXPLORER 3.0 HTML//",
-           "-//IETF//DTD HTML STRICT LEVEL 3//EN//3.0" => 1,
+           "-//MICROSOFT//DTD INTERNET EXPLORER 3.0 TABLES//",
-           "-//IETF//DTD HTML STRICT//EN" => 1,
+           "-//NETSCAPE COMM. CORP.//DTD HTML//",
-           "-//IETF//DTD HTML STRICT//EN//2.0" => 1,
+           "-//NETSCAPE COMM. CORP.//DTD STRICT HTML//",
-           "-//IETF//DTD HTML STRICT//EN//3.0" => 1,
+           "-//O'REILLY AND ASSOCIATES//DTD HTML 2.0//",
-           "-//IETF//DTD HTML//EN" => 1,
+           "-//O'REILLY AND ASSOCIATES//DTD HTML EXTENDED 1.0//",
-           "-//IETF//DTD HTML//EN//2.0" => 1,
+           "-//O'REILLY AND ASSOCIATES//DTD HTML EXTENDED RELAXED 1.0//",
-           "-//IETF//DTD HTML//EN//3.0" => 1,
+           "-//SOFTQUAD SOFTWARE//DTD HOTMETAL PRO 6.0::19990601::EXTENSIONS TO HTML 4.0//",
-           "-//METRIUS//DTD METRIUS PRESENTATIONAL//EN" => 1,
+           "-//SOFTQUAD//DTD HOTMETAL PRO 4.0::19971010::EXTENSIONS TO HTML 4.0//",
-           "-//MICROSOFT//DTD INTERNET EXPLORER 2.0 HTML STRICT//EN" => 1,
+           "-//SPYGLASS//DTD HTML 2.0 EXTENDED//",
-           "-//MICROSOFT//DTD INTERNET EXPLORER 2.0 HTML//EN" => 1,
+           "-//SQ//DTD HTML 2.0 HOTMETAL + EXTENSIONS//",
-           "-//MICROSOFT//DTD INTERNET EXPLORER 2.0 TABLES//EN" => 1,
+           "-//SUN MICROSYSTEMS CORP.//DTD HOTJAVA HTML//",
-           "-//MICROSOFT//DTD INTERNET EXPLORER 3.0 HTML STRICT//EN" => 1,
+           "-//SUN MICROSYSTEMS CORP.//DTD HOTJAVA STRICT HTML//",
-           "-//MICROSOFT//DTD INTERNET EXPLORER 3.0 HTML//EN" => 1,
+           "-//W3C//DTD HTML 3 1995-03-24//",
-           "-//MICROSOFT//DTD INTERNET EXPLORER 3.0 TABLES//EN" => 1,
+           "-//W3C//DTD HTML 3.2 DRAFT//",
-           "-//NETSCAPE COMM. CORP.//DTD HTML//EN" => 1,
+           "-//W3C//DTD HTML 3.2 FINAL//",
-           "-//NETSCAPE COMM. CORP.//DTD STRICT HTML//EN" => 1,
+           "-//W3C//DTD HTML 3.2//",
-           "-//O'REILLY AND ASSOCIATES//DTD HTML 2.0//EN" => 1,
+           "-//W3C//DTD HTML 3.2S DRAFT//",
-           "-//O'REILLY AND ASSOCIATES//DTD HTML EXTENDED 1.0//EN" => 1,
+           "-//W3C//DTD HTML 4.0 FRAMESET//",
-           "-//O'REILLY AND ASSOCIATES//DTD HTML EXTENDED RELAXED 1.0//EN" => 1,
+           "-//W3C//DTD HTML 4.0 TRANSITIONAL//",
-           "-//SOFTQUAD SOFTWARE//DTD HOTMETAL PRO 6.0::19990601::EXTENSIONS TO HTML 4.0//EN" => 1,
+           "-//W3C//DTD HTML EXPERIMETNAL 19960712//",
-           "-//SOFTQUAD//DTD HOTMETAL PRO 4.0::19971010::EXTENSIONS TO HTML 4.0//EN" => 1,
+           "-//W3C//DTD HTML EXPERIMENTAL 970421//",
-           "-//SPYGLASS//DTD HTML 2.0 EXTENDED//EN" => 1,
+           "-//W3C//DTD W3 HTML//",
-           "-//SQ//DTD HTML 2.0 HOTMETAL + EXTENSIONS//EN" => 1,
+           "-//W3O//DTD W3 HTML 3.0//",
-           "-//SUN MICROSYSTEMS CORP.//DTD HOTJAVA HTML//EN" => 1,
+           "-//WEBTECHS//DTD MOZILLA HTML 2.0//",
-           "-//SUN MICROSYSTEMS CORP.//DTD HOTJAVA STRICT HTML//EN" => 1,
+           "-//WEBTECHS//DTD MOZILLA HTML//",
-           "-//W3C//DTD HTML 3 1995-03-24//EN" => 1,
+         ]; # $prefix
-           "-//W3C//DTD HTML 3.2 DRAFT//EN" => 1,
+         my $match;
-           "-//W3C//DTD HTML 3.2 FINAL//EN" => 1,
+         for (@$prefix) {
-           "-//W3C//DTD HTML 3.2//EN" => 1,
+           if (substr ($prefix, 0, length $_) eq $_) {
-           "-//W3C//DTD HTML 3.2S DRAFT//EN" => 1,
+             $match = 1;
-           "-//W3C//DTD HTML 4.0 FRAMESET//EN" => 1,
+             last;
-           "-//W3C//DTD HTML 4.0 TRANSITIONAL//EN" => 1,
+           }
-           "-//W3C//DTD HTML EXPERIMETNAL 19960712//EN" => 1,
+         }
-           "-//W3C//DTD HTML EXPERIMENTAL 970421//EN" => 1,
+         if ($match or
-           "-//W3C//DTD W3 HTML//EN" => 1,
+             $pubid eq "-//W3O//DTD W3 HTML STRICT 3.0//EN//" or
-           "-//W3O//DTD W3 HTML 3.0//EN" => 1,
+             $pubid eq "-/W3C/DTD HTML 4.0 TRANSITIONAL/EN" or
-           "-//W3O//DTD W3 HTML 3.0//EN//" => 1,
+             $pubid eq "HTML") {
-           "-//W3O//DTD W3 HTML STRICT 3.0//EN//" => 1,
-           "-//WEBTECHS//DTD MOZILLA HTML 2.0//EN" => 1,
-           "-//WEBTECHS//DTD MOZILLA HTML//EN" => 1,
-           "-/W3C/DTD HTML 4.0 TRANSITIONAL/EN" => 1,
-           "HTML" => 1,
-         }->{$pubid}) {
            !!!cp ('t5');
            $self->{document}->manakai_compat_mode ('quirks');
-         } elsif ($pubid eq "-//W3C//DTD HTML 4.01 FRAMESET//EN" or
+         } elsif ($pubid =~ m[^-//W3C//DTD HTML 4.01 FRAMESET//] or
-                  $pubid eq "-//W3C//DTD HTML 4.01 TRANSITIONAL//EN") {
+                  $pubid =~ m[^-//W3C//DTD HTML 4.01 TRANSITIONAL//]) {
            if (defined $token->{system_identifier}) {
              !!!cp ('t6');
              $self->{document}->manakai_compat_mode ('quirks');
-Line 2900 
 sub _tree_construction_initial ($) {
+Line 3542 
 sub _tree_construction_initial ($) {
              !!!cp ('t7');
              $self->{document}->manakai_compat_mode ('limited quirks');
            }
-         } elsif ($pubid eq "-//W3C//DTD XHTML 1.0 FRAMESET//EN" or
+         } elsif ($pubid =~ m[^-//W3C//DTD XHTML 1.0 FRAMESET//] or
-                  $pubid eq "-//W3C//DTD XHTML 1.0 TRANSITIONAL//EN") {
+                  $pubid =~ m[^-//W3C//DTD XHTML 1.0 TRANSITIONAL//]) {
            !!!cp ('t8');
            $self->{document}->manakai_compat_mode ('limited quirks');
          } else {
-Line 2914 
 sub _tree_construction_initial ($) {
+Line 3556 
 sub _tree_construction_initial ($) {
          my $sysid = $token->{system_identifier};
          $sysid =~ tr/A-Z/a-z/;
          if ($sysid eq "http://www.ibm.com/data/dtd/v11/ibmxhtml1-transitional.dtd") {
-           ## TODO: Check the spec: PUBLIC "(limited quirks)" "(quirks)"
+           ## NOTE: Ensure that |PUBLIC "(limited quirks)" "(quirks)"| is
+           ## marked as quirks.
            $self->{document}->manakai_compat_mode ('quirks');
            !!!cp ('t11');
          } else {
-Line 3086 
 sub _reset_insertion_mode ($) {
+Line 3729 
 sub _reset_insertion_mode ($) {
        if ($self->{open_elements}->[0]->[0] eq $node->[0]) {
          $last = 1;
          if (defined $self->{inner_html_node}) {
-           if ($self->{inner_html_node}->[1] & TABLE_CELL_EL) {
+           !!!cp ('t28');
-             !!!cp ('t27');
+           $node = $self->{inner_html_node};
-             #
+         } else {
-           } else {
+           die "_reset_insertion_mode: t27";
-             !!!cp ('t28');
-             $node = $self->{inner_html_node};
-           }
          }
        }
-     ## Step 4..14
+       ## Step 4..14
-     my $new_mode;
+       my $new_mode;
-     if ($node->[1] & FOREIGN_EL) {
+       if ($node->[1] & FOREIGN_EL) {
-       ## NOTE: Strictly spaking, the line below only applies to MathML and
+         !!!cp ('t28.1');
-       ## SVG elements.  Currently the HTML syntax supports only MathML and
+         ## NOTE: Strictly spaking, the line below only applies to MathML and
-       ## SVG elements as foreigners.
+         ## SVG elements.  Currently the HTML syntax supports only MathML and
-       $new_mode = $self->{insertion_mode} | IN_FOREIGN_CONTENT_IM;
+         ## SVG elements as foreigners.
-       ## ISSUE: What is set as the secondary insertion mode?
+         $new_mode = IN_BODY_IM | IN_FOREIGN_CONTENT_IM;
-     } else {
+       } elsif ($node->[1] & TABLE_CELL_EL) {
-       $new_mode = {
+         if ($last) {
+           !!!cp ('t28.2');
+           #
+         } else {
+           !!!cp ('t28.3');
+           $new_mode = IN_CELL_IM;
+         }
+       } else {
+         !!!cp ('t28.4');
+         $new_mode = {
                        select => IN_SELECT_IM,
                        ## NOTE: |option| and |optgroup| do not set
                        ## insertion mode to "in select" by themselves.
-                       td => IN_CELL_IM,
-                       th => IN_CELL_IM,
                        tr => IN_ROW_IM,
                        tbody => IN_TABLE_BODY_IM,
                        thead => IN_TABLE_BODY_IM,
-Line 3122 
 sub _reset_insertion_mode ($) {
+Line 3769 
 sub _reset_insertion_mode ($) {
                        body => IN_BODY_IM,
                        frameset => IN_FRAMESET_IM,
                       }->{$node->[0]->manakai_local_name};
-     }
+       }
-     $self->{insertion_mode} = $new_mode and return if defined $new_mode;
+       $self->{insertion_mode} = $new_mode and return if defined $new_mode;
        ## Step 15
        if ($node->[1] & HTML_EL) {
-Line 3297 
 sub _tree_construction_main ($) {
+Line 3944 
 sub _tree_construction_main ($) {
        ## NOTE: An end-of-file token.
        if ($content_model_flag == CDATA_CONTENT_MODEL) {
          !!!cp ('t43');
-         !!!parse-error (type => 'in CDATA:#'.$token->{type}, token => $token);
+         !!!parse-error (type => 'in CDATA:#eof', token => $token);
        } elsif ($content_model_flag == RCDATA_CONTENT_MODEL) {
          !!!cp ('t44');
-         !!!parse-error (type => 'in RCDATA:#'.$token->{type}, token => $token);
+         !!!parse-error (type => 'in RCDATA:#eof', token => $token);
        } else {
          die "$0: $content_model_flag in parse_rcdata";
        }
-Line 3337 
 sub _tree_construction_main ($) {
+Line 3984 
 sub _tree_construction_main ($) {
        ## Ignore the token
      } else {
        !!!cp ('t48');
-       !!!parse-error (type => 'in CDATA:#'.$token->{type}, token => $token);
+       !!!parse-error (type => 'in CDATA:#eof', token => $token);
        ## ISSUE: And ignore?
        ## TODO: mark as "already executed"
      }
-Line 3388 
 sub _tree_construction_main ($) {
+Line 4035 
 sub _tree_construction_main ($) {
        } # AFE
        unless (defined $formatting_element) {
          !!!cp ('t53');
-         !!!parse-error (type => 'unmatched end tag:'.$tag_name, token => $end_tag_token);
+         !!!parse-error (type => 'unmatched end tag', text => $tag_name, token => $end_tag_token);
          ## Ignore the token
          !!!next-token;
          return;
-Line 3405 
 sub _tree_construction_main ($) {
+Line 4052 
 sub _tree_construction_main ($) {
              last INSCOPE;
            } else { # in open elements but not in scope
              !!!cp ('t55');
-             !!!parse-error (type => 'unmatched end tag:'.$token->{tag_name},
+             !!!parse-error (type => 'unmatched end tag',
+                             text => $token->{tag_name},
                              token => $end_tag_token);
              ## Ignore the token
              !!!next-token;
-Line 3418 
 sub _tree_construction_main ($) {
+Line 4066 
 sub _tree_construction_main ($) {
        } # INSCOPE
        unless (defined $formatting_element_i_in_open) {
          !!!cp ('t57');
-         !!!parse-error (type => 'unmatched end tag:'.$token->{tag_name},
+         !!!parse-error (type => 'unmatched end tag',
+                         text => $token->{tag_name},
                          token => $end_tag_token);
          pop @$active_formatting_elements; # $formatting_element
          !!!next-token; ## TODO: ok?
-Line 3427 
 sub _tree_construction_main ($) {
+Line 4076 
 sub _tree_construction_main ($) {
        if (not $self->{open_elements}->[-1]->[0] eq $formatting_element->[0]) {
          !!!cp ('t58');
          !!!parse-error (type => 'not closed',
-                         value => $self->{open_elements}->[-1]->[0]
+                         text => $self->{open_elements}->[-1]->[0]
                              ->manakai_local_name,
                          token => $end_tag_token);
        }
-Line 3636 
 sub _tree_construction_main ($) {
+Line 4285 
 sub _tree_construction_main ($) {
    B: while (1) {
      if ($token->{type} == DOCTYPE_TOKEN) {
        !!!cp ('t73');
-       !!!parse-error (type => 'DOCTYPE in the middle', token => $token);
+       !!!parse-error (type => 'in html:#DOCTYPE', token => $token);
        ## Ignore the token
        ## Stay in the phase
        !!!next-token;
-Line 3645 
 sub _tree_construction_main ($) {
+Line 4294 
 sub _tree_construction_main ($) {
               $token->{tag_name} eq 'html') {
        if ($self->{insertion_mode} == AFTER_HTML_BODY_IM) {
          !!!cp ('t79');
-         !!!parse-error (type => 'after html:html', token => $token);
+         !!!parse-error (type => 'after html', text => 'html', token => $token);
          $self->{insertion_mode} = AFTER_BODY_IM;
        } elsif ($self->{insertion_mode} == AFTER_HTML_FRAMESET_IM) {
          !!!cp ('t80');
-         !!!parse-error (type => 'after html:html', token => $token);
+         !!!parse-error (type => 'after html', text => 'html', token => $token);
          $self->{insertion_mode} = AFTER_FRAMESET_IM;
        } else {
          !!!cp ('t81');
-Line 3699 
 sub _tree_construction_main ($) {
+Line 4348 
 sub _tree_construction_main ($) {
            !!!cp ('t87.2');
            #
          } elsif ({
-   ## TODO:
+                   b => 1, big => 1, blockquote => 1, body => 1, br => 1,
+                   center => 1, code => 1, dd => 1, div => 1, dl => 1, dt => 1,
+                   em => 1, embed => 1, font => 1, h1 => 1, h2 => 1, h3 => 1,
+                   h4 => 1, h5 => 1, h6 => 1, head => 1, hr => 1, i => 1,
+                   img => 1, li => 1, listing => 1, menu => 1, meta => 1,
+                   nobr => 1, ol => 1, p => 1, pre => 1, ruby => 1, s => 1,
+                   small => 1, span => 1, strong => 1, strike => 1, sub => 1,
+                   sup => 1, table => 1, tt => 1, u => 1, ul => 1, var => 1,
                   }->{$token->{tag_name}}) {
            !!!cp ('t87.2');
            !!!parse-error (type => 'not closed',
-                           value => $self->{open_elements}->[-1]->[0]
+                           text => $self->{open_elements}->[-1]->[0]
                                ->manakai_local_name,
                            token => $token);
            pop @{$self->{open_elements}}
                while $self->{open_elements}->[-1]->[1] & FOREIGN_EL;
-           $self->{insertion_mode} &= ~ $self->{insertion_mode};
+           $self->{insertion_mode} &= ~ IN_FOREIGN_CONTENT_IM;
            ## Reprocess.
            next B;
          } else {
-           ## TODO: case fixup
+           my $nsuri = $self->{open_elements}->[-1]->[0]->namespace_uri;
+           my $tag_name = $token->{tag_name};
+           if ($nsuri eq $SVG_NS) {
+             $tag_name = {
+                altglyph => 'altGlyph',
+                altglyphdef => 'altGlyphDef',
+                altglyphitem => 'altGlyphItem',
+                animatecolor => 'animateColor',
+                animatemotion => 'animateMotion',
+                animatetransform => 'animateTransform',
+                clippath => 'clipPath',
+                feblend => 'feBlend',
+                fecolormatrix => 'feColorMatrix',
+                fecomponenttransfer => 'feComponentTransfer',
+                fecomposite => 'feComposite',
+                feconvolvematrix => 'feConvolveMatrix',
+                fediffuselighting => 'feDiffuseLighting',
+                fedisplacementmap => 'feDisplacementMap',
+                fedistantlight => 'feDistantLight',
+                feflood => 'feFlood',
+                fefunca => 'feFuncA',
+                fefuncb => 'feFuncB',
+                fefuncg => 'feFuncG',
+                fefuncr => 'feFuncR',
+                fegaussianblur => 'feGaussianBlur',
+                feimage => 'feImage',
+                femerge => 'feMerge',
+                femergenode => 'feMergeNode',
+                femorphology => 'feMorphology',
+                feoffset => 'feOffset',
+                fepointlight => 'fePointLight',
+                fespecularlighting => 'feSpecularLighting',
+                fespotlight => 'feSpotLight',
+                fetile => 'feTile',
+                feturbulence => 'feTurbulence',
+                foreignobject => 'foreignObject',
+                glyphref => 'glyphRef',
+                lineargradient => 'linearGradient',
+                radialgradient => 'radialGradient',
+                #solidcolor => 'solidColor', ## NOTE: Commented in spec (SVG1.2)
+                textpath => 'textPath',
+             }->{$tag_name} || $tag_name;
+           }
+           ## "adjust SVG attributes" (SVG only) - done in insert-element-f
-           !!!insert-element-f ($self->{open_elements}->[-1]->[0]->namespace_uri, $token);
+           ## "adjust foreign attributes" - done in insert-element-f
+           !!!insert-element-f ($nsuri, $tag_name, $token->{attributes}, $token);
            if ($self->{self_closing}) {
              pop @{$self->{open_elements}};
-Line 3733 
 sub _tree_construction_main ($) {
+Line 4435 
 sub _tree_construction_main ($) {
          !!!cp ('t87.5');
          #
        } elsif ($token->{type} == END_OF_FILE_TOKEN) {
-         ## NOTE: "using the rules for secondary insertion mode" then "continue"
          !!!cp ('t87.6');
-         #
+         !!!parse-error (type => 'not closed',
-         ## TODO: ...
+                         text => $self->{open_elements}->[-1]->[0]
+                             ->manakai_local_name,
+                         token => $token);
+         pop @{$self->{open_elements}}
+             while $self->{open_elements}->[-1]->[1] & FOREIGN_EL;
+         $self->{insertion_mode} &= ~ IN_FOREIGN_CONTENT_IM;
+         ## Reprocess.
+         next B;
        } else {
          die "$0: $token->{type}: Unknown token type";
        }
-Line 3777 
 sub _tree_construction_main ($) {
+Line 4487 
 sub _tree_construction_main ($) {
            !!!cp ('t90');
            ## As if </noscript>
            pop @{$self->{open_elements}};
-           !!!parse-error (type => 'in noscript:#character', token => $token);
+           !!!parse-error (type => 'in noscript:#text', token => $token);
            ## Reprocess in the "in head" insertion mode...
            ## As if </head>
-Line 3813 
 sub _tree_construction_main ($) {
+Line 4523 
 sub _tree_construction_main ($) {
              !!!next-token;
              next B;
            } elsif ($self->{insertion_mode} == AFTER_HEAD_IM) {
-             !!!cp ('t94');
+             !!!cp ('t93.2');
-             #
+             !!!parse-error (type => 'after head', text => 'head',
+                             token => $token);
+             ## Ignore the token
+             !!!nack ('t93.3');
+             !!!next-token;
+             next B;
            } else {
              !!!cp ('t95');
-             !!!parse-error (type => 'in head:head', token => $token); # or in head noscript
+             !!!parse-error (type => 'in head:head',
+                             token => $token); # or in head noscript
              ## Ignore the token
              !!!nack ('t95.1');
              !!!next-token;
-Line 3842 
 sub _tree_construction_main ($) {
+Line 4558 
 sub _tree_construction_main ($) {
                  !!!cp ('t98');
                  ## As if </noscript>
                  pop @{$self->{open_elements}};
-                 !!!parse-error (type => 'in noscript:base', token => $token);
+                 !!!parse-error (type => 'in noscript', text => 'base',
+                                 token => $token);
                  $self->{insertion_mode} = IN_HEAD_IM;
                  ## Reprocess in the "in head" insertion mode...
-Line 3853 
 sub _tree_construction_main ($) {
+Line 4570 
 sub _tree_construction_main ($) {
                ## NOTE: There is a "as if in head" code clone.
                if ($self->{insertion_mode} == AFTER_HEAD_IM) {
                  !!!cp ('t100');
-                 !!!parse-error (type => 'after head:'.$token->{tag_name}, token => $token);
+                 !!!parse-error (type => 'after head',
+                                 text => $token->{tag_name}, token => $token);
                  push @{$self->{open_elements}},
                      [$self->{head_element}, $el_category->{head}];
                } else {
-Line 3870 
 sub _tree_construction_main ($) {
+Line 4588 
 sub _tree_construction_main ($) {
                ## NOTE: There is a "as if in head" code clone.
                if ($self->{insertion_mode} == AFTER_HEAD_IM) {
                  !!!cp ('t102');
-                 !!!parse-error (type => 'after head:'.$token->{tag_name}, token => $token);
+                 !!!parse-error (type => 'after head',
+                                 text => $token->{tag_name}, token => $token);
                  push @{$self->{open_elements}},
                      [$self->{head_element}, $el_category->{head}];
                } else {
-Line 3887 
 sub _tree_construction_main ($) {
+Line 4606 
 sub _tree_construction_main ($) {
                ## NOTE: There is a "as if in head" code clone.
                if ($self->{insertion_mode} == AFTER_HEAD_IM) {
                  !!!cp ('t104');
-                 !!!parse-error (type => 'after head:'.$token->{tag_name}, token => $token);
+                 !!!parse-error (type => 'after head',
+                                 text => $token->{tag_name}, token => $token);
                  push @{$self->{open_elements}},
                      [$self->{head_element}, $el_category->{head}];
                } else {
-Line 3897 
 sub _tree_construction_main ($) {
+Line 4617 
 sub _tree_construction_main ($) {
                my $meta_el = pop @{$self->{open_elements}}; ## ISSUE: This step is missing in the spec.
                unless ($self->{confident}) {
-                 if ($token->{attributes}->{charset}) { ## TODO: And if supported
+                 if ($token->{attributes}->{charset}) {
                    !!!cp ('t106');
+                   ## NOTE: Whether the encoding is supported or not is handled
+                   ## in the {change_encoding} callback.
                    $self->{change_encoding}
                        ->($self, $token->{attributes}->{charset}->{value},
                           $token);
-Line 3908 
 sub _tree_construction_main ($) {
+Line 4630 
 sub _tree_construction_main ($) {
                                             $token->{attributes}->{charset}
                                                 ->{has_reference});
                  } elsif ($token->{attributes}->{content}) {
-                   ## ISSUE: Algorithm name in the spec was incorrect so that not linked to the definition.
                    if ($token->{attributes}->{content}->{value}
-                       =~ /\A[^;]*;[\x09-\x0D\x20]*[Cc][Hh][Aa][Rr][Ss][Ee][Tt]
+                       =~ /[Cc][Hh][Aa][Rr][Ss][Ee][Tt]
                            [\x09-\x0D\x20]*=
                            [\x09-\x0D\x20]*(?>"([^"]*)"|'([^']*)'|
-                           ([^"'\x09-\x0D\x20][^\x09-\x0D\x20]*))/x) {
+                           ([^"'\x09-\x0D\x20][^\x09-\x0D\x20\x3B]*))/x) {
                      !!!cp ('t107');
+                     ## NOTE: Whether the encoding is supported or not is handled
+                     ## in the {change_encoding} callback.
                      $self->{change_encoding}
                          ->($self, defined $1 ? $1 : defined $2 ? $2 : $3,
                             $token);
-Line 3953 
 sub _tree_construction_main ($) {
+Line 4676 
 sub _tree_construction_main ($) {
                  !!!cp ('t111');
                  ## As if </noscript>
                  pop @{$self->{open_elements}};
-                 !!!parse-error (type => 'in noscript:title', token => $token);
+                 !!!parse-error (type => 'in noscript', text => 'title',
+                                 token => $token);
                  $self->{insertion_mode} = IN_HEAD_IM;
                  ## Reprocess in the "in head" insertion mode...
                } elsif ($self->{insertion_mode} == AFTER_HEAD_IM) {
                  !!!cp ('t112');
-                 !!!parse-error (type => 'after head:'.$token->{tag_name}, token => $token);
+                 !!!parse-error (type => 'after head',
+                                 text => $token->{tag_name}, token => $token);
                  push @{$self->{open_elements}},
                      [$self->{head_element}, $el_category->{head}];
                } else {
-Line 3973 
 sub _tree_construction_main ($) {
+Line 4698 
 sub _tree_construction_main ($) {
                pop @{$self->{open_elements}} # <head>
                    if $self->{insertion_mode} == AFTER_HEAD_IM;
                next B;
-             } elsif ($token->{tag_name} eq 'style') {
+             } elsif ($token->{tag_name} eq 'style' or
+                      $token->{tag_name} eq 'noframes') {
                ## NOTE: Or (scripting is enabled and tag_name eq 'noscript' and
                ## insertion mode IN_HEAD_IM)
                ## NOTE: There is a "as if in head" code clone.
                if ($self->{insertion_mode} == AFTER_HEAD_IM) {
                  !!!cp ('t114');
-                 !!!parse-error (type => 'after head:'.$token->{tag_name}, token => $token);
+                 !!!parse-error (type => 'after head',
+                                 text => $token->{tag_name}, token => $token);
                  push @{$self->{open_elements}},
                      [$self->{head_element}, $el_category->{head}];
                } else {
-Line 4000 
 sub _tree_construction_main ($) {
+Line 4727 
 sub _tree_construction_main ($) {
                  next B;
                } elsif ($self->{insertion_mode} == IN_HEAD_NOSCRIPT_IM) {
                  !!!cp ('t117');
-                 !!!parse-error (type => 'in noscript:noscript', token => $token);
+                 !!!parse-error (type => 'in noscript', text => 'noscript',
+                                 token => $token);
                  ## Ignore the token
                  !!!nack ('t117.1');
                  !!!next-token;
-Line 4014 
 sub _tree_construction_main ($) {
+Line 4742 
 sub _tree_construction_main ($) {
                  !!!cp ('t119');
                  ## As if </noscript>
                  pop @{$self->{open_elements}};
-                 !!!parse-error (type => 'in noscript:script', token => $token);
+                 !!!parse-error (type => 'in noscript', text => 'script',
+                                 token => $token);
                  $self->{insertion_mode} = IN_HEAD_IM;
                  ## Reprocess in the "in head" insertion mode...
                } elsif ($self->{insertion_mode} == AFTER_HEAD_IM) {
                  !!!cp ('t120');
-                 !!!parse-error (type => 'after head:'.$token->{tag_name}, token => $token);
+                 !!!parse-error (type => 'after head',
+                                 text => $token->{tag_name}, token => $token);
                  push @{$self->{open_elements}},
                      [$self->{head_element}, $el_category->{head}];
                } else {
-Line 4038 
 sub _tree_construction_main ($) {
+Line 4768 
 sub _tree_construction_main ($) {
                  !!!cp ('t122');
                  ## As if </noscript>
                  pop @{$self->{open_elements}};
-                 !!!parse-error (type => 'in noscript:'.$token->{tag_name}, token => $token);
+                 !!!parse-error (type => 'in noscript',
+                                 text => $token->{tag_name}, token => $token);
                  ## Reprocess in the "in head" insertion mode...
                  ## As if </head>
-Line 4077 
 sub _tree_construction_main ($) {
+Line 4808 
 sub _tree_construction_main ($) {
                !!!cp ('t129');
                ## As if </noscript>
                pop @{$self->{open_elements}};
-               !!!parse-error (type => 'in noscript:/'.$token->{tag_name}, token => $token);
+               !!!parse-error (type => 'in noscript:/',
+                               text => $token->{tag_name}, token => $token);
                ## Reprocess in the "in head" insertion mode...
                ## As if </head>
-Line 4120 
 sub _tree_construction_main ($) {
+Line 4852 
 sub _tree_construction_main ($) {
                  !!!cp ('t133');
                  ## As if </noscript>
                  pop @{$self->{open_elements}};
-                 !!!parse-error (type => 'in noscript:/head', token => $token);
+                 !!!parse-error (type => 'in noscript:/',
+                                 text => 'head', token => $token);
                  ## Reprocess in the "in head" insertion mode...
                  pop @{$self->{open_elements}};
-Line 4133 
 sub _tree_construction_main ($) {
+Line 4866 
 sub _tree_construction_main ($) {
                  $self->{insertion_mode} = AFTER_HEAD_IM;
                  !!!next-token;
                  next B;
+               } elsif ($self->{insertion_mode} == AFTER_HEAD_IM) {
+                 !!!cp ('t134.1');
+                 !!!parse-error (type => 'unmatched end tag', text => 'head',
+                                 token => $token);
+                 ## Ignore the token
+                 !!!next-token;
+                 next B;
                } else {
-                 !!!cp ('t135');
+                 die "$0: $self->{insertion_mode}: Unknown insertion mode";
-                 #
                }
              } elsif ($token->{tag_name} eq 'noscript') {
                if ($self->{insertion_mode} == IN_HEAD_NOSCRIPT_IM) {
-Line 4144 
 sub _tree_construction_main ($) {
+Line 4883 
 sub _tree_construction_main ($) {
                  $self->{insertion_mode} = IN_HEAD_IM;
                  !!!next-token;
                  next B;
-               } elsif ($self->{insertion_mode} == BEFORE_HEAD_IM) {
+               } elsif ($self->{insertion_mode} == BEFORE_HEAD_IM or
+                        $self->{insertion_mode} == AFTER_HEAD_IM) {
                  !!!cp ('t137');
-                 !!!parse-error (type => 'unmatched end tag:noscript', token => $token);
+                 !!!parse-error (type => 'unmatched end tag',
+                                 text => 'noscript', token => $token);
                  ## Ignore the token ## ISSUE: An issue in the spec.
                  !!!next-token;
                  next B;
-Line 4157 
 sub _tree_construction_main ($) {
+Line 4898 
 sub _tree_construction_main ($) {
              } elsif ({
                        body => 1, html => 1,
                       }->{$token->{tag_name}}) {
-               if ($self->{insertion_mode} == BEFORE_HEAD_IM) {
+               if ($self->{insertion_mode} == BEFORE_HEAD_IM or
-                 !!!cp ('t139');
+                   $self->{insertion_mode} == IN_HEAD_IM or
-                 ## As if <head>
+                   $self->{insertion_mode} == IN_HEAD_NOSCRIPT_IM) {
-                 !!!create-element ($self->{head_element}, $HTML_NS, 'head',, $token);
-                 $self->{open_elements}->[-1]->[0]->append_child ($self->{head_element});
-                 push @{$self->{open_elements}},
-                     [$self->{head_element}, $el_category->{head}];
-                 $self->{insertion_mode} = IN_HEAD_IM;
-                 ## Reprocess in the "in head" insertion mode...
-               } elsif ($self->{insertion_mode} == IN_HEAD_NOSCRIPT_IM) {
                  !!!cp ('t140');
-                 !!!parse-error (type => 'unmatched end tag:'.$token->{tag_name}, token => $token);
+                 !!!parse-error (type => 'unmatched end tag',
+                                 text => $token->{tag_name}, token => $token);
+                 ## Ignore the token
+                 !!!next-token;
+                 next B;
+               } elsif ($self->{insertion_mode} == AFTER_HEAD_IM) {
+                 !!!cp ('t140.1');
+                 !!!parse-error (type => 'unmatched end tag',
+                                 text => $token->{tag_name}, token => $token);
                  ## Ignore the token
                  !!!next-token;
                  next B;
                } else {
-                 !!!cp ('t141');
+                 die "$0: $self->{insertion_mode}: Unknown insertion mode";
                }
+             } elsif ($token->{tag_name} eq 'p') {
-               #
+               !!!cp ('t142');
-             } elsif ({
+               !!!parse-error (type => 'unmatched end tag',
-                       p => 1, br => 1,
+                               text => $token->{tag_name}, token => $token);
-                      }->{$token->{tag_name}}) {
+               ## Ignore the token
+               !!!next-token;
+               next B;
+             } elsif ($token->{tag_name} eq 'br') {
                if ($self->{insertion_mode} == BEFORE_HEAD_IM) {
-                 !!!cp ('t142');
+                 !!!cp ('t142.2');
-                 ## As if <head>
+                 ## (before head) as if <head>, (in head) as if </head>
                  !!!create-element ($self->{head_element}, $HTML_NS, 'head',, $token);
                  $self->{open_elements}->[-1]->[0]->append_child ($self->{head_element});
-                 push @{$self->{open_elements}},
+                 $self->{insertion_mode} = AFTER_HEAD_IM;
-                     [$self->{head_element}, $el_category->{head}];
+                 ## Reprocess in the "after head" insertion mode...
+               } elsif ($self->{insertion_mode} == IN_HEAD_IM) {
+                 !!!cp ('t143.2');
+                 ## As if </head>
+                 pop @{$self->{open_elements}};
+                 $self->{insertion_mode} = AFTER_HEAD_IM;
+                 ## Reprocess in the "after head" insertion mode...
+               } elsif ($self->{insertion_mode} == IN_HEAD_NOSCRIPT_IM) {
+                 !!!cp ('t143.3');
+                 ## ISSUE: Two parse errors for <head><noscript></br>
+                 !!!parse-error (type => 'unmatched end tag',
+                                 text => 'br', token => $token);
+                 ## As if </noscript>
+                 pop @{$self->{open_elements}};
                  $self->{insertion_mode} = IN_HEAD_IM;
                  ## Reprocess in the "in head" insertion mode...
-               } else {
+                 ## As if </head>
-                 !!!cp ('t143');
+                 pop @{$self->{open_elements}};
-               }
+                 $self->{insertion_mode} = AFTER_HEAD_IM;
-               #
+                 ## Reprocess in the "after head" insertion mode...
-             } else {
+               } elsif ($self->{insertion_mode} == AFTER_HEAD_IM) {
-               if ($self->{insertion_mode} == AFTER_HEAD_IM) {
+                 !!!cp ('t143.4');
-                 !!!cp ('t144');
                  #
                } else {
-                 !!!cp ('t145');
+                 die "$0: $self->{insertion_mode}: Unknown insertion mode";
-                 !!!parse-error (type => 'unmatched end tag:'.$token->{tag_name}, token => $token);
-                 ## Ignore the token
-                 !!!next-token;
-                 next B;
                }
+               ## ISSUE: does not agree with IE7 - it doesn't ignore </br>.
+               !!!parse-error (type => 'unmatched end tag',
+                               text => 'br', token => $token);
+               ## Ignore the token
+               !!!next-token;
+               next B;
+             } else {
+               !!!cp ('t145');
+               !!!parse-error (type => 'unmatched end tag',
+                               text => $token->{tag_name}, token => $token);
+               ## Ignore the token
+               !!!next-token;
+               next B;
              }
              if ($self->{insertion_mode} == IN_HEAD_NOSCRIPT_IM) {
                !!!cp ('t146');
                ## As if </noscript>
                pop @{$self->{open_elements}};
-               !!!parse-error (type => 'in noscript:/'.$token->{tag_name}, token => $token);
+               !!!parse-error (type => 'in noscript:/',
+                               text => $token->{tag_name}, token => $token);
                ## Reprocess in the "in head" insertion mode...
                ## As if </head>
-Line 4229 
 sub _tree_construction_main ($) {
+Line 4998 
 sub _tree_construction_main ($) {
              } elsif ($self->{insertion_mode} == BEFORE_HEAD_IM) {
  ## ISSUE: This case cannot be reached?
                !!!cp ('t148');
-               !!!parse-error (type => 'unmatched end tag:'.$token->{tag_name}, token => $token);
+               !!!parse-error (type => 'unmatched end tag',
+                               text => $token->{tag_name}, token => $token);
                ## Ignore the token ## ISSUE: An issue in the spec.
                !!!next-token;
                next B;
-Line 4340 
 sub _tree_construction_main ($) {
+Line 5110 
 sub _tree_construction_main ($) {
                  !!!cp ('t153');
                  !!!parse-error (type => 'start tag not allowed',
-                     value => $token->{tag_name}, token => $token);
+                     text => $token->{tag_name}, token => $token);
                  ## Ignore the token
                  !!!nack ('t153.1');
                  !!!next-token;
                  next B;
                } elsif ($self->{insertion_mode} == IN_CAPTION_IM) {
-                 !!!parse-error (type => 'not closed:caption', token => $token);
+                 !!!parse-error (type => 'not closed', text => 'caption',
+                                 token => $token);
                  ## NOTE: As if </caption>.
                  ## have a table element in table scope
-Line 4366 
 sub _tree_construction_main ($) {
+Line 5137 
 sub _tree_construction_main ($) {
                    !!!cp ('t157');
                    !!!parse-error (type => 'start tag not allowed',
-                                   value => $token->{tag_name}, token => $token);
+                                   text => $token->{tag_name}, token => $token);
                    ## Ignore the token
                    !!!nack ('t157.1');
                    !!!next-token;
-Line 4383 
 sub _tree_construction_main ($) {
+Line 5154 
 sub _tree_construction_main ($) {
                  unless ($self->{open_elements}->[-1]->[1] & CAPTION_EL) {
                    !!!cp ('t159');
                    !!!parse-error (type => 'not closed',
-                                   value => $self->{open_elements}->[-1]->[0]
+                                   text => $self->{open_elements}->[-1]->[0]
                                        ->manakai_local_name,
                                    token => $token);
                  } else {
-Line 4425 
 sub _tree_construction_main ($) {
+Line 5196 
 sub _tree_construction_main ($) {
                  } # INSCOPE
                    unless (defined $i) {
                      !!!cp ('t165');
-                     !!!parse-error (type => 'unmatched end tag:'.$token->{tag_name}, token => $token);
+                     !!!parse-error (type => 'unmatched end tag',
+                                     text => $token->{tag_name},
+                                     token => $token);
                      ## Ignore the token
                      !!!next-token;
                      next B;
-Line 4442 
 sub _tree_construction_main ($) {
+Line 5215 
 sub _tree_construction_main ($) {
                          ne $token->{tag_name}) {
                    !!!cp ('t167');
                    !!!parse-error (type => 'not closed',
-                                   value => $self->{open_elements}->[-1]->[0]
+                                   text => $self->{open_elements}->[-1]->[0]
                                        ->manakai_local_name,
                                    token => $token);
                  } else {
-Line 4459 
 sub _tree_construction_main ($) {
+Line 5232 
 sub _tree_construction_main ($) {
                  next B;
                } elsif ($self->{insertion_mode} == IN_CAPTION_IM) {
                  !!!cp ('t169');
-                 !!!parse-error (type => 'unmatched end tag:'.$token->{tag_name}, token => $token);
+                 !!!parse-error (type => 'unmatched end tag',
+                                 text => $token->{tag_name}, token => $token);
                  ## Ignore the token
                  !!!next-token;
                  next B;
-Line 4486 
 sub _tree_construction_main ($) {
+Line 5260 
 sub _tree_construction_main ($) {
                    !!!cp ('t173');
                    !!!parse-error (type => 'unmatched end tag',
-                                   value => $token->{tag_name}, token => $token);
+                                   text => $token->{tag_name}, token => $token);
                    ## Ignore the token
                    !!!next-token;
                    next B;
-Line 4502 
 sub _tree_construction_main ($) {
+Line 5276 
 sub _tree_construction_main ($) {
                  unless ($self->{open_elements}->[-1]->[1] & CAPTION_EL) {
                    !!!cp ('t175');
                    !!!parse-error (type => 'not closed',
-                                   value => $self->{open_elements}->[-1]->[0]
+                                   text => $self->{open_elements}->[-1]->[0]
                                        ->manakai_local_name,
                                    token => $token);
                  } else {
-Line 4519 
 sub _tree_construction_main ($) {
+Line 5293 
 sub _tree_construction_main ($) {
                  next B;
                } elsif ($self->{insertion_mode} == IN_CELL_IM) {
                  !!!cp ('t177');
-                 !!!parse-error (type => 'unmatched end tag:'.$token->{tag_name}, token => $token);
+                 !!!parse-error (type => 'unmatched end tag',
+                                 text => $token->{tag_name}, token => $token);
                  ## Ignore the token
                  !!!next-token;
                  next B;
-Line 4562 
 sub _tree_construction_main ($) {
+Line 5337 
 sub _tree_construction_main ($) {
                  !!!cp ('t182');
                  !!!parse-error (type => 'unmatched end tag',
-                     value => $token->{tag_name}, token => $token);
+                     text => $token->{tag_name}, token => $token);
                  ## Ignore the token
                  !!!next-token;
                  next B;
                } # INSCOPE
              } elsif ($token->{tag_name} eq 'table' and
                       $self->{insertion_mode} == IN_CAPTION_IM) {
-               !!!parse-error (type => 'not closed:caption', token => $token);
+               !!!parse-error (type => 'not closed', text => 'caption',
+                               token => $token);
                ## As if </caption>
                ## have a table element in table scope
-Line 4587 
 sub _tree_construction_main ($) {
+Line 5363 
 sub _tree_construction_main ($) {
                } # INSCOPE
                unless (defined $i) {
                  !!!cp ('t186');
-                 !!!parse-error (type => 'unmatched end tag:caption', token => $token);
+                 !!!parse-error (type => 'unmatched end tag',
+                                 text => 'caption', token => $token);
                  ## Ignore the token
                  !!!next-token;
                  next B;
-Line 4602 
 sub _tree_construction_main ($) {
+Line 5379 
 sub _tree_construction_main ($) {
                unless ($self->{open_elements}->[-1]->[1] & CAPTION_EL) {
                  !!!cp ('t188');
                  !!!parse-error (type => 'not closed',
-                                 value => $self->{open_elements}->[-1]->[0]
+                                 text => $self->{open_elements}->[-1]->[0]
                                      ->manakai_local_name,
                                  token => $token);
                } else {
-Line 4622 
 sub _tree_construction_main ($) {
+Line 5399 
 sub _tree_construction_main ($) {
                       }->{$token->{tag_name}}) {
                if ($self->{insertion_mode} & BODY_TABLE_IMS) {
                  !!!cp ('t190');
-                 !!!parse-error (type => 'unmatched end tag:'.$token->{tag_name}, token => $token);
+                 !!!parse-error (type => 'unmatched end tag',
+                                 text => $token->{tag_name}, token => $token);
                  ## Ignore the token
                  !!!next-token;
                  next B;
-Line 4636 
 sub _tree_construction_main ($) {
+Line 5414 
 sub _tree_construction_main ($) {
                       }->{$token->{tag_name}} and
                       $self->{insertion_mode} == IN_CAPTION_IM) {
                !!!cp ('t192');
-               !!!parse-error (type => 'unmatched end tag:'.$token->{tag_name}, token => $token);
+               !!!parse-error (type => 'unmatched end tag',
+                               text => $token->{tag_name}, token => $token);
                ## Ignore the token
                !!!next-token;
                next B;
-Line 4676 
 sub _tree_construction_main ($) {
+Line 5455 
 sub _tree_construction_main ($) {
            }
          }
-             !!!parse-error (type => 'in table:#character', token => $token);
+         !!!parse-error (type => 'in table:#text', token => $token);
              ## As if in body, but insert into foster parent element
              ## ISSUE: Spec says that "whenever a node would be inserted
-Line 4727 
 sub _tree_construction_main ($) {
+Line 5506 
 sub _tree_construction_main ($) {
          !!!next-token;
          next B;
        } elsif ($token->{type} == START_TAG_TOKEN) {
-             if ({
+         if ({
-                  tr => ($self->{insertion_mode} != IN_ROW_IM),
+              tr => ($self->{insertion_mode} != IN_ROW_IM),
-                  th => 1, td => 1,
+              th => 1, td => 1,
-                 }->{$token->{tag_name}}) {
+             }->{$token->{tag_name}}) {
-               if ($self->{insertion_mode} == IN_TABLE_IM) {
+           if ($self->{insertion_mode} == IN_TABLE_IM) {
-                 ## Clear back to table context
+             ## Clear back to table context
-                 while (not ($self->{open_elements}->[-1]->[1]
+             while (not ($self->{open_elements}->[-1]->[1]
-                                 & TABLE_SCOPING_EL)) {
+                             & TABLE_SCOPING_EL)) {
-                   !!!cp ('t201');
+               !!!cp ('t201');
-                   pop @{$self->{open_elements}};
+               pop @{$self->{open_elements}};
-                 }
+             }
-                 !!!insert-element ('tbody',, $token);
+             !!!insert-element ('tbody',, $token);
-                 $self->{insertion_mode} = IN_TABLE_BODY_IM;
+             $self->{insertion_mode} = IN_TABLE_BODY_IM;
-                 ## reprocess in the "in table body" insertion mode...
+             ## reprocess in the "in table body" insertion mode...
-               }
+           }
-               if ($self->{insertion_mode} == IN_TABLE_BODY_IM) {
+           if ($self->{insertion_mode} == IN_TABLE_BODY_IM) {
-                 unless ($token->{tag_name} eq 'tr') {
+             unless ($token->{tag_name} eq 'tr') {
-                   !!!cp ('t202');
+               !!!cp ('t202');
-                   !!!parse-error (type => 'missing start tag:tr', token => $token);
+               !!!parse-error (type => 'missing start tag:tr', token => $token);
-                 }
+             }
-                 ## Clear back to table body context
+             ## Clear back to table body context
-                 while (not ($self->{open_elements}->[-1]->[1]
+             while (not ($self->{open_elements}->[-1]->[1]
-                                 & TABLE_ROWS_SCOPING_EL)) {
+                             & TABLE_ROWS_SCOPING_EL)) {
-                   !!!cp ('t203');
+               !!!cp ('t203');
-                   ## ISSUE: Can this case be reached?
+               ## ISSUE: Can this case be reached?
-                   pop @{$self->{open_elements}};
+               pop @{$self->{open_elements}};
-                 }
+             }
                  $self->{insertion_mode} = IN_ROW_IM;
                  if ($token->{tag_name} eq 'tr') {
-Line 4812 
 sub _tree_construction_main ($) {
+Line 5591 
 sub _tree_construction_main ($) {
                  unless (defined $i) {
                    !!!cp ('t210');
  ## TODO: This type is wrong.
-                   !!!parse-error (type => 'unmacthed end tag:'.$token->{tag_name}, token => $token);
+                   !!!parse-error (type => 'unmacthed end tag',
+                                   text => $token->{tag_name}, token => $token);
                    ## Ignore the token
                    !!!nack ('t210.1');
                    !!!next-token;
-Line 4856 
 sub _tree_construction_main ($) {
+Line 5636 
 sub _tree_construction_main ($) {
                  } # INSCOPE
                  unless (defined $i) {
                    !!!cp ('t216');
- ## TODO: This erorr type ios wrong.
+ ## TODO: This erorr type is wrong.
-                   !!!parse-error (type => 'unmatched end tag:'.$token->{tag_name}, token => $token);
+                   !!!parse-error (type => 'unmatched end tag',
+                                   text => $token->{tag_name}, token => $token);
                    ## Ignore the token
                    !!!nack ('t216.1');
                    !!!next-token;
-Line 4932 
 sub _tree_construction_main ($) {
+Line 5713 
 sub _tree_construction_main ($) {
                }
              } elsif ($token->{tag_name} eq 'table') {
                !!!parse-error (type => 'not closed',
-                               value => $self->{open_elements}->[-1]->[0]
+                               text => $self->{open_elements}->[-1]->[0]
                                    ->manakai_local_name,
                                token => $token);
-Line 4953 
 sub _tree_construction_main ($) {
+Line 5734 
 sub _tree_construction_main ($) {
                unless (defined $i) {
                  !!!cp ('t223');
  ## TODO: The following is wrong, maybe.
-                 !!!parse-error (type => 'unmatched end tag:table', token => $token);
+                 !!!parse-error (type => 'unmatched end tag', text => 'table',
+                                 token => $token);
                  ## Ignore tokens </table><table>
                  !!!nack ('t223.1');
                  !!!next-token;
                  next B;
                }
  ## TODO: Followings are removed from the latest spec.
                ## generate implied end tags
                while ($self->{open_elements}->[-1]->[1] & END_TAG_OPTIONAL_EL) {
                  !!!cp ('t224');
-Line 4971 
 sub _tree_construction_main ($) {
+Line 5753 
 sub _tree_construction_main ($) {
                  !!!cp ('t225');
                  ## NOTE: |<table><tr><table>|
                  !!!parse-error (type => 'not closed',
-                                 value => $self->{open_elements}->[-1]->[0]
+                                 text => $self->{open_elements}->[-1]->[0]
                                      ->manakai_local_name,
                                  token => $token);
                } else {
-Line 5012 
 sub _tree_construction_main ($) {
+Line 5794 
 sub _tree_construction_main ($) {
                my $type = lc $token->{attributes}->{type}->{value};
                if ($type eq 'hidden') {
                  !!!cp ('t227.3');
-                 !!!parse-error (type => 'in table:'.$token->{tag_name}, token => $token);
+                 !!!parse-error (type => 'in table',
+                                 text => $token->{tag_name}, token => $token);
                  !!!insert-element ($token->{tag_name}, $token->{attributes}, $token);
-Line 5040 
 sub _tree_construction_main ($) {
+Line 5823 
 sub _tree_construction_main ($) {
            #
          }
-         !!!parse-error (type => 'in table:'.$token->{tag_name}, token => $token);
+         !!!parse-error (type => 'in table', text => $token->{tag_name},
+                         token => $token);
          $insert = $insert_to_foster;
          #
-Line 5062 
 sub _tree_construction_main ($) {
+Line 5846 
 sub _tree_construction_main ($) {
                } # INSCOPE
                unless (defined $i) {
                  !!!cp ('t230');
-                 !!!parse-error (type => 'unmatched end tag:'.$token->{tag_name}, token => $token);
+                 !!!parse-error (type => 'unmatched end tag',
+                                 text => $token->{tag_name}, token => $token);
                  ## Ignore the token
                  !!!nack ('t230.1');
                  !!!next-token;
-Line 5103 
 sub _tree_construction_main ($) {
+Line 5888 
 sub _tree_construction_main ($) {
                  unless (defined $i) {
                    !!!cp ('t235');
  ## TODO: The following is wrong.
-                   !!!parse-error (type => 'unmatched end tag:'.$token->{type}, token => $token);
+                   !!!parse-error (type => 'unmatched end tag',
+                                   text => $token->{type}, token => $token);
                    ## Ignore the token
                    !!!nack ('t236.1');
                    !!!next-token;
-Line 5139 
 sub _tree_construction_main ($) {
+Line 5925 
 sub _tree_construction_main ($) {
                  } # INSCOPE
                  unless (defined $i) {
                    !!!cp ('t239');
-                   !!!parse-error (type => 'unmatched end tag:'.$token->{tag_name}, token => $token);
+                   !!!parse-error (type => 'unmatched end tag',
+                                   text => $token->{tag_name}, token => $token);
                    ## Ignore the token
                    !!!nack ('t239.1');
                    !!!next-token;
-Line 5185 
 sub _tree_construction_main ($) {
+Line 5972 
 sub _tree_construction_main ($) {
                } # INSCOPE
                unless (defined $i) {
                  !!!cp ('t243');
-                 !!!parse-error (type => 'unmatched end tag:'.$token->{tag_name}, token => $token);
+                 !!!parse-error (type => 'unmatched end tag',
+                                 text => $token->{tag_name}, token => $token);
                  ## Ignore the token
                  !!!nack ('t243.1');
                  !!!next-token;
-Line 5219 
 sub _tree_construction_main ($) {
+Line 6007 
 sub _tree_construction_main ($) {
                  } # INSCOPE
                    unless (defined $i) {
                      !!!cp ('t249');
-                     !!!parse-error (type => 'unmatched end tag:'.$token->{tag_name}, token => $token);
+                     !!!parse-error (type => 'unmatched end tag',
+                                     text => $token->{tag_name}, token => $token);
                      ## Ignore the token
                      !!!nack ('t249.1');
                      !!!next-token;
-Line 5242 
 sub _tree_construction_main ($) {
+Line 6031 
 sub _tree_construction_main ($) {
                  } # INSCOPE
                    unless (defined $i) {
                      !!!cp ('t252');
-                     !!!parse-error (type => 'unmatched end tag:tr', token => $token);
+                     !!!parse-error (type => 'unmatched end tag',
+                                     text => 'tr', token => $token);
                      ## Ignore the token
                      !!!nack ('t252.1');
                      !!!next-token;
-Line 5277 
 sub _tree_construction_main ($) {
+Line 6067 
 sub _tree_construction_main ($) {
                } # INSCOPE
                unless (defined $i) {
                  !!!cp ('t256');
-                 !!!parse-error (type => 'unmatched end tag:'.$token->{tag_name}, token => $token);
+                 !!!parse-error (type => 'unmatched end tag',
+                                 text => $token->{tag_name}, token => $token);
                  ## Ignore the token
                  !!!nack ('t256.1');
                  !!!next-token;
-Line 5304 
 sub _tree_construction_main ($) {
+Line 6095 
 sub _tree_construction_main ($) {
                        tbody => 1, tfoot => 1, thead => 1, # $self->{insertion_mode} == IN_TABLE_IM
                       }->{$token->{tag_name}}) {
            !!!cp ('t258');
-           !!!parse-error (type => 'unmatched end tag:'.$token->{tag_name}, token => $token);
+           !!!parse-error (type => 'unmatched end tag',
+                           text => $token->{tag_name}, token => $token);
            ## Ignore the token
            !!!nack ('t258.1');
             !!!next-token;
            next B;
          } else {
            !!!cp ('t259');
-           !!!parse-error (type => 'in table:/'.$token->{tag_name}, token => $token);
+           !!!parse-error (type => 'in table:/',
+                           text => $token->{tag_name}, token => $token);
            $insert = $insert_to_foster;
            #
-Line 5361 
 sub _tree_construction_main ($) {
+Line 6154 
 sub _tree_construction_main ($) {
              if ($token->{tag_name} eq 'colgroup') {
                if ($self->{open_elements}->[-1]->[1] & HTML_EL) {
                  !!!cp ('t264');
-                 !!!parse-error (type => 'unmatched end tag:colgroup', token => $token);
+                 !!!parse-error (type => 'unmatched end tag',
+                                 text => 'colgroup', token => $token);
                  ## Ignore the token
                  !!!next-token;
                  next B;
-Line 5374 
 sub _tree_construction_main ($) {
+Line 6168 
 sub _tree_construction_main ($) {
                }
              } elsif ($token->{tag_name} eq 'col') {
                !!!cp ('t266');
-               !!!parse-error (type => 'unmatched end tag:col', token => $token);
+               !!!parse-error (type => 'unmatched end tag',
+                               text => 'col', token => $token);
                ## Ignore the token
                !!!next-token;
                next B;
-Line 5404 
 sub _tree_construction_main ($) {
+Line 6199 
 sub _tree_construction_main ($) {
            if ($self->{open_elements}->[-1]->[1] & HTML_EL) {
              !!!cp ('t269');
  ## TODO: Wrong error type?
-             !!!parse-error (type => 'unmatched end tag:colgroup', token => $token);
+             !!!parse-error (type => 'unmatched end tag',
+                             text => 'colgroup', token => $token);
              ## Ignore the token
              !!!nack ('t269.1');
              !!!next-token;
-Line 5458 
 sub _tree_construction_main ($) {
+Line 6254 
 sub _tree_construction_main ($) {
            !!!nack ('t277.1');
            !!!next-token;
            next B;
-         } elsif ($token->{tag_name} eq 'select' or
+         } elsif ({
-                  $token->{tag_name} eq 'input' or
+                    select => 1, input => 1, textarea => 1,
+                  }->{$token->{tag_name}} or
                   ($self->{insertion_mode} == IN_SELECT_IN_TABLE_IM and
                    {
                     caption => 1, table => 1,
-Line 5467 
 sub _tree_construction_main ($) {
+Line 6264 
 sub _tree_construction_main ($) {
                     tr => 1, td => 1, th => 1,
                    }->{$token->{tag_name}})) {
            ## TODO: The type below is not good - <select> is replaced by </select>
-           !!!parse-error (type => 'not closed:select', token => $token);
+           !!!parse-error (type => 'not closed', text => 'select',
+                           token => $token);
            ## NOTE: As if the token were </select> (<select> case) or
            ## as if there were </select> (otherwise).
            ## have an element in table scope
-Line 5485 
 sub _tree_construction_main ($) {
+Line 6283 
 sub _tree_construction_main ($) {
            } # INSCOPE
            unless (defined $i) {
              !!!cp ('t280');
-             !!!parse-error (type => 'unmatched end tag:select', token => $token);
+             !!!parse-error (type => 'unmatched end tag',
+                             text => 'select', token => $token);
              ## Ignore the token
              !!!nack ('t280.1');
              !!!next-token;
-Line 5509 
 sub _tree_construction_main ($) {
+Line 6308 
 sub _tree_construction_main ($) {
            }
          } else {
            !!!cp ('t282');
-           !!!parse-error (type => 'in select:'.$token->{tag_name}, token => $token);
+           !!!parse-error (type => 'in select',
+                           text => $token->{tag_name}, token => $token);
            ## Ignore the token
            !!!nack ('t282.1');
            !!!next-token;
-Line 5527 
 sub _tree_construction_main ($) {
+Line 6327 
 sub _tree_construction_main ($) {
              pop @{$self->{open_elements}};
            } else {
              !!!cp ('t285');
-             !!!parse-error (type => 'unmatched end tag:'.$token->{tag_name}, token => $token);
+             !!!parse-error (type => 'unmatched end tag',
+                             text => $token->{tag_name}, token => $token);
              ## Ignore the token
            }
            !!!nack ('t285.1');
-Line 5539 
 sub _tree_construction_main ($) {
+Line 6340 
 sub _tree_construction_main ($) {
              pop @{$self->{open_elements}};
            } else {
              !!!cp ('t287');
-             !!!parse-error (type => 'unmatched end tag:'.$token->{tag_name}, token => $token);
+             !!!parse-error (type => 'unmatched end tag',
+                             text => $token->{tag_name}, token => $token);
              ## Ignore the token
            }
            !!!nack ('t287.1');
-Line 5561 
 sub _tree_construction_main ($) {
+Line 6363 
 sub _tree_construction_main ($) {
            } # INSCOPE
            unless (defined $i) {
              !!!cp ('t290');
-             !!!parse-error (type => 'unmatched end tag:'.$token->{tag_name}, token => $token);
+             !!!parse-error (type => 'unmatched end tag',
+                             text => $token->{tag_name}, token => $token);
              ## Ignore the token
              !!!nack ('t290.1');
              !!!next-token;
-Line 5582 
 sub _tree_construction_main ($) {
+Line 6385 
 sub _tree_construction_main ($) {
                    tfoot => 1, thead => 1, tr => 1, td => 1, th => 1,
                   }->{$token->{tag_name}}) {
  ## TODO: The following is wrong?
-           !!!parse-error (type => 'unmatched end tag:'.$token->{tag_name}, token => $token);
+           !!!parse-error (type => 'unmatched end tag',
+                           text => $token->{tag_name}, token => $token);
            ## have an element in table scope
            my $i;
-Line 5623 
 sub _tree_construction_main ($) {
+Line 6427 
 sub _tree_construction_main ($) {
            unless (defined $i) {
              !!!cp ('t297');
  ## TODO: The following error type is correct?
-             !!!parse-error (type => 'unmatched end tag:select', token => $token);
+             !!!parse-error (type => 'unmatched end tag',
+                             text => 'select', token => $token);
              ## Ignore the </select> token
              !!!nack ('t297.1');
              !!!next-token; ## TODO: ok?
-Line 5640 
 sub _tree_construction_main ($) {
+Line 6445 
 sub _tree_construction_main ($) {
            next B;
          } else {
            !!!cp ('t299');
-           !!!parse-error (type => 'in select:/'.$token->{tag_name}, token => $token);
+           !!!parse-error (type => 'in select:/',
+                           text => $token->{tag_name}, token => $token);
            ## Ignore the token
            !!!nack ('t299.3');
            !!!next-token;
-Line 5678 
 sub _tree_construction_main ($) {
+Line 6484 
 sub _tree_construction_main ($) {
          if ($self->{insertion_mode} == AFTER_HTML_BODY_IM) {
            !!!cp ('t301');
-           !!!parse-error (type => 'after html:#character', token => $token);
+           !!!parse-error (type => 'after html:#text', token => $token);
            ## Reprocess in the "after body" insertion mode.
          } else {
-Line 5686 
 sub _tree_construction_main ($) {
+Line 6492 
 sub _tree_construction_main ($) {
          }
          ## "after body" insertion mode
-         !!!parse-error (type => 'after body:#character', token => $token);
+         !!!parse-error (type => 'after body:#text', token => $token);
          $self->{insertion_mode} = IN_BODY_IM;
          ## reprocess
-Line 5694 
 sub _tree_construction_main ($) {
+Line 6500 
 sub _tree_construction_main ($) {
        } elsif ($token->{type} == START_TAG_TOKEN) {
          if ($self->{insertion_mode} == AFTER_HTML_BODY_IM) {
            !!!cp ('t303');
-           !!!parse-error (type => 'after html:'.$token->{tag_name}, token => $token);
+           !!!parse-error (type => 'after html',
+                           text => $token->{tag_name}, token => $token);
            ## Reprocess in the "after body" insertion mode.
          } else {
-Line 5702 
 sub _tree_construction_main ($) {
+Line 6509 
 sub _tree_construction_main ($) {
          }
          ## "after body" insertion mode
-         !!!parse-error (type => 'after body:'.$token->{tag_name}, token => $token);
+         !!!parse-error (type => 'after body',
+                         text => $token->{tag_name}, token => $token);
          $self->{insertion_mode} = IN_BODY_IM;
          !!!ack-later;
-Line 5711 
 sub _tree_construction_main ($) {
+Line 6519 
 sub _tree_construction_main ($) {
        } elsif ($token->{type} == END_TAG_TOKEN) {
          if ($self->{insertion_mode} == AFTER_HTML_BODY_IM) {
            !!!cp ('t305');
-           !!!parse-error (type => 'after html:/'.$token->{tag_name}, token => $token);
+           !!!parse-error (type => 'after html:/',
+                           text => $token->{tag_name}, token => $token);
            $self->{insertion_mode} = AFTER_BODY_IM;
            ## Reprocess in the "after body" insertion mode.
-Line 5723 
 sub _tree_construction_main ($) {
+Line 6532 
 sub _tree_construction_main ($) {
          if ($token->{tag_name} eq 'html') {
            if (defined $self->{inner_html_node}) {
              !!!cp ('t307');
-             !!!parse-error (type => 'unmatched end tag:html', token => $token);
+             !!!parse-error (type => 'unmatched end tag',
+                             text => 'html', token => $token);
              ## Ignore the token
              !!!next-token;
              next B;
-Line 5735 
 sub _tree_construction_main ($) {
+Line 6545 
 sub _tree_construction_main ($) {
            }
          } else {
            !!!cp ('t309');
-           !!!parse-error (type => 'after body:/'.$token->{tag_name}, token => $token);
+           !!!parse-error (type => 'after body:/',
+                           text => $token->{tag_name}, token => $token);
            $self->{insertion_mode} = IN_BODY_IM;
            ## reprocess
-Line 5763 
 sub _tree_construction_main ($) {
+Line 6574 
 sub _tree_construction_main ($) {
          if ($token->{data} =~ s/^[^\x09\x0A\x0B\x0C\x20]+//) {
            if ($self->{insertion_mode} == IN_FRAMESET_IM) {
              !!!cp ('t311');
-             !!!parse-error (type => 'in frameset:#character', token => $token);
+             !!!parse-error (type => 'in frameset:#text', token => $token);
            } elsif ($self->{insertion_mode} == AFTER_FRAMESET_IM) {
              !!!cp ('t312');
-             !!!parse-error (type => 'after frameset:#character', token => $token);
+             !!!parse-error (type => 'after frameset:#text', token => $token);
-           } else { # "after html frameset"
+           } else { # "after after frameset"
              !!!cp ('t313');
-             !!!parse-error (type => 'after html:#character', token => $token);
+             !!!parse-error (type => 'after html:#text', token => $token);
-             $self->{insertion_mode} = AFTER_FRAMESET_IM;
-             ## Reprocess in the "after frameset" insertion mode.
-             !!!parse-error (type => 'after frameset:#character', token => $token);
            }
            ## Ignore the token.
-Line 5789 
 sub _tree_construction_main ($) {
+Line 6596 
 sub _tree_construction_main ($) {
          die qq[$0: Character "$token->{data}"];
        } elsif ($token->{type} == START_TAG_TOKEN) {
-         if ($self->{insertion_mode} == AFTER_HTML_FRAMESET_IM) {
-           !!!cp ('t316');
-           !!!parse-error (type => 'after html:'.$token->{tag_name}, token => $token);
-           $self->{insertion_mode} = AFTER_FRAMESET_IM;
-           ## Process in the "after frameset" insertion mode.
-         } else {
-           !!!cp ('t317');
-         }
          if ($token->{tag_name} eq 'frameset' and
              $self->{insertion_mode} == IN_FRAMESET_IM) {
            !!!cp ('t318');
-Line 5816 
 sub _tree_construction_main ($) {
+Line 6613 
 sub _tree_construction_main ($) {
            next B;
          } elsif ($token->{tag_name} eq 'noframes') {
            !!!cp ('t320');
-           ## NOTE: As if in body.
+           ## NOTE: As if in head.
            $parse_rcdata->(CDATA_CONTENT_MODEL);
            next B;
+           ## NOTE: |<!DOCTYPE HTML><frameset></frameset></html><noframes></noframes>|
+           ## has no parse error.
          } else {
            if ($self->{insertion_mode} == IN_FRAMESET_IM) {
              !!!cp ('t321');
-             !!!parse-error (type => 'in frameset:'.$token->{tag_name}, token => $token);
+             !!!parse-error (type => 'in frameset',
-           } else {
+                             text => $token->{tag_name}, token => $token);
+           } elsif ($self->{insertion_mode} == AFTER_FRAMESET_IM) {
              !!!cp ('t322');
-             !!!parse-error (type => 'after frameset:'.$token->{tag_name}, token => $token);
+             !!!parse-error (type => 'after frameset',
+                             text => $token->{tag_name}, token => $token);
+           } else { # "after after frameset"
+             !!!cp ('t322.2');
+             !!!parse-error (type => 'after after frameset',
+                             text => $token->{tag_name}, token => $token);
            }
            ## Ignore the token
            !!!nack ('t322.1');
-Line 5833 
 sub _tree_construction_main ($) {
+Line 6639 
 sub _tree_construction_main ($) {
            next B;
          }
        } elsif ($token->{type} == END_TAG_TOKEN) {
-         if ($self->{insertion_mode} == AFTER_HTML_FRAMESET_IM) {
-           !!!cp ('t323');
-           !!!parse-error (type => 'after html:/'.$token->{tag_name}, token => $token);
-           $self->{insertion_mode} = AFTER_FRAMESET_IM;
-           ## Process in the "after frameset" insertion mode.
-         } else {
-           !!!cp ('t324');
-         }
          if ($token->{tag_name} eq 'frameset' and
              $self->{insertion_mode} == IN_FRAMESET_IM) {
            if ($self->{open_elements}->[-1]->[1] & HTML_EL and
                @{$self->{open_elements}} == 1) {
              !!!cp ('t325');
-             !!!parse-error (type => 'unmatched end tag:'.$token->{tag_name}, token => $token);
+             !!!parse-error (type => 'unmatched end tag',
+                             text => $token->{tag_name}, token => $token);
              ## Ignore the token
              !!!next-token;
            } else {
-Line 5874 
 sub _tree_construction_main ($) {
+Line 6671 
 sub _tree_construction_main ($) {
          } else {
            if ($self->{insertion_mode} == IN_FRAMESET_IM) {
              !!!cp ('t330');
-             !!!parse-error (type => 'in frameset:/'.$token->{tag_name}, token => $token);
+             !!!parse-error (type => 'in frameset:/',
-           } else {
+                             text => $token->{tag_name}, token => $token);
+           } elsif ($self->{insertion_mode} == AFTER_FRAMESET_IM) {
+             !!!cp ('t330.1');
+             !!!parse-error (type => 'after frameset:/',
+                             text => $token->{tag_name}, token => $token);
+           } else { # "after after html"
              !!!cp ('t331');
-             !!!parse-error (type => 'after frameset:/'.$token->{tag_name}, token => $token);
+             !!!parse-error (type => 'after after frameset:/',
+                             text => $token->{tag_name}, token => $token);
            }
            ## Ignore the token
            !!!next-token;
-Line 5931 
 sub _tree_construction_main ($) {
+Line 6734 
 sub _tree_construction_main ($) {
          my $meta_el = pop @{$self->{open_elements}}; ## ISSUE: This step is missing in the spec.
          unless ($self->{confident}) {
-           if ($token->{attributes}->{charset}) { ## TODO: And if supported
+           if ($token->{attributes}->{charset}) {
              !!!cp ('t335');
+             ## NOTE: Whether the encoding is supported or not is handled
+             ## in the {change_encoding} callback.
              $self->{change_encoding}
                  ->($self, $token->{attributes}->{charset}->{value}, $token);
-Line 5941 
 sub _tree_construction_main ($) {
+Line 6746 
 sub _tree_construction_main ($) {
                                       $token->{attributes}->{charset}
                                           ->{has_reference});
            } elsif ($token->{attributes}->{content}) {
-             ## ISSUE: Algorithm name in the spec was incorrect so that not linked to the definition.
              if ($token->{attributes}->{content}->{value}
-                 =~ /\A[^;]*;[\x09-\x0D\x20]*[Cc][Hh][Aa][Rr][Ss][Ee][Tt]
+                 =~ /[Cc][Hh][Aa][Rr][Ss][Ee][Tt]
                      [\x09-\x0D\x20]*=
                      [\x09-\x0D\x20]*(?>"([^"]*)"|'([^']*)'|
-                     ([^"'\x09-\x0D\x20][^\x09-\x0D\x20]*))/x) {
+                     ([^"'\x09-\x0D\x20][^\x09-\x0D\x20\x3B]*))/x) {
                !!!cp ('t336');
+               ## NOTE: Whether the encoding is supported or not is handled
+               ## in the {change_encoding} callback.
                $self->{change_encoding}
                    ->($self, defined $1 ? $1 : defined $2 ? $2 : $3, $token);
                $meta_el->[0]->get_attribute_node_ns (undef, 'content')
-Line 5982 
 sub _tree_construction_main ($) {
+Line 6788 
 sub _tree_construction_main ($) {
          $parse_rcdata->(RCDATA_CONTENT_MODEL);
          next B;
        } elsif ($token->{tag_name} eq 'body') {
-         !!!parse-error (type => 'in body:body', token => $token);
+         !!!parse-error (type => 'in body', text => 'body', token => $token);
          if (@{$self->{open_elements}} == 1 or
              not ($self->{open_elements}->[1]->[1] & BODY_EL)) {
-Line 6102 
 sub _tree_construction_main ($) {
+Line 6908 
 sub _tree_construction_main ($) {
              if ($i != -1) {
                !!!cp ('t355');
                !!!parse-error (type => 'not closed',
-                               value => $self->{open_elements}->[-1]->[0]
+                               text => $self->{open_elements}->[-1]->[0]
                                    ->manakai_local_name,
                                token => $token);
              } else {
-Line 6256 
 sub _tree_construction_main ($) {
+Line 7062 
 sub _tree_construction_main ($) {
                  xmp => 1,
                  iframe => 1,
                  noembed => 1,
-                 noframes => 1,
+                 noframes => 1, ## NOTE: This is an "as if in head" code clone.
                  noscript => 0, ## TODO: 1 if scripting is enabled
                 }->{$token->{tag_name}}) {
          if ($token->{tag_name} eq 'xmp') {
-Line 6278 
 sub _tree_construction_main ($) {
+Line 7084 
 sub _tree_construction_main ($) {
            !!!next-token;
            next B;
          } else {
+           !!!ack ('t391.1');
            my $at = $token->{attributes};
            my $form_attrs;
            $form_attrs->{action} = $at->{action} if $at->{action};
-Line 6321 
 sub _tree_construction_main ($) {
+Line 7129 
 sub _tree_construction_main ($) {
                           line => $token->{line}, column => $token->{column}},
                          {type => END_TAG_TOKEN, tag_name => 'form',
                           line => $token->{line}, column => $token->{column}};
-           !!!nack ('t391.1'); ## NOTE: Not acknowledged.
            !!!back-token (@tokens);
            !!!next-token;
            next B;
-Line 6369 
 sub _tree_construction_main ($) {
+Line 7176 
 sub _tree_construction_main ($) {
            ## Ignore the token
          } else {
            !!!cp ('t398');
-           !!!parse-error (type => 'in RCDATA:#'.$token->{type}, token => $token);
+           !!!parse-error (type => 'in RCDATA:#eof', token => $token);
          }
          !!!next-token;
          next B;
+       } elsif ($token->{tag_name} eq 'rt' or
+                $token->{tag_name} eq 'rp') {
+         ## has a |ruby| element in scope
+         INSCOPE: for (reverse 0..$#{$self->{open_elements}}) {
+           my $node = $self->{open_elements}->[$_];
+           if ($node->[1] & RUBY_EL) {
+             !!!cp ('t398.1');
+             ## generate implied end tags
+             while ($self->{open_elements}->[-1]->[1] & END_TAG_OPTIONAL_EL) {
+               !!!cp ('t398.2');
+               pop @{$self->{open_elements}};
+             }
+             unless ($self->{open_elements}->[-1]->[1] & RUBY_EL) {
+               !!!cp ('t398.3');
+               !!!parse-error (type => 'not closed',
+                               text => $self->{open_elements}->[-1]->[0]
+                                   ->manakai_local_name,
+                               token => $token);
+               pop @{$self->{open_elements}}
+                   while not $self->{open_elements}->[-1]->[1] & RUBY_EL;
+             }
+             last INSCOPE;
+           } elsif ($node->[1] & SCOPING_EL) {
+             !!!cp ('t398.4');
+             last INSCOPE;
+           }
+         } # INSCOPE
+         !!!insert-element-t ($token->{tag_name}, $token->{attributes}, $token);
+         !!!nack ('t398.5');
+         !!!next-token;
+         redo B;
        } elsif ($token->{tag_name} eq 'math' or
                 $token->{tag_name} eq 'svg') {
          $reconstruct_active_formatting_elements->($insert_to_current);
+         ## "Adjust MathML attributes" ('math' only) - done in insert-element-f
+         ## "adjust SVG attributes" ('svg' only) - done in insert-element-f
+         ## "adjust foreign attributes" - done in insert-element-f
-         !!!insert-element-f ($token->{tag_name} eq 'math' ? $MML_NS : $SVG_NS, $token);
+         !!!insert-element-f ($token->{tag_name} eq 'math' ? $MML_NS : $SVG_NS, $token->{tag_name}, $token->{attributes}, $token);
          if ($self->{self_closing}) {
            pop @{$self->{open_elements}};
-Line 6399 
 sub _tree_construction_main ($) {
+Line 7245 
 sub _tree_construction_main ($) {
                  thead => 1, tr => 1,
                 }->{$token->{tag_name}}) {
          !!!cp ('t401');
-         !!!parse-error (type => 'in body:'.$token->{tag_name}, token => $token);
+         !!!parse-error (type => 'in body',
+                         text => $token->{tag_name}, token => $token);
          ## Ignore the token
          !!!nack ('t401.1'); ## NOTE: |<col/>| or |<frame/>| here is an error.
          !!!next-token;
-Line 6484 
 sub _tree_construction_main ($) {
+Line 7331 
 sub _tree_construction_main ($) {
            }
            !!!parse-error (type => 'start tag not allowed',
-                           value => $token->{tag_name}, token => $token);
+                           text => $token->{tag_name}, token => $token);
            ## NOTE: Ignore the token.
            !!!next-token;
            next B;
-Line 6494 
 sub _tree_construction_main ($) {
+Line 7341 
 sub _tree_construction_main ($) {
            unless ($_->[1] & ALL_END_TAG_OPTIONAL_EL) {
              !!!cp ('t403');
              !!!parse-error (type => 'not closed',
-                             value => $_->[0]->manakai_local_name,
+                             text => $_->[0]->manakai_local_name,
                              token => $token);
              last;
            } else {
-Line 6514 
 sub _tree_construction_main ($) {
+Line 7361 
 sub _tree_construction_main ($) {
            unless ($self->{open_elements}->[-1]->[1] & BODY_EL) {
              !!!cp ('t406');
              !!!parse-error (type => 'not closed',
-                             value => $self->{open_elements}->[1]->[0]
+                             text => $self->{open_elements}->[1]->[0]
                                  ->manakai_local_name,
                              token => $token);
            } else {
-Line 6525 
 sub _tree_construction_main ($) {
+Line 7372 
 sub _tree_construction_main ($) {
            next B;
          } else {
            !!!cp ('t408');
-           !!!parse-error (type => 'unmatched end tag:'.$token->{tag_name}, token => $token);
+           !!!parse-error (type => 'unmatched end tag',
+                           text => $token->{tag_name}, token => $token);
            ## Ignore the token
            !!!next-token;
            next B;
-Line 6553 
 sub _tree_construction_main ($) {
+Line 7401 
 sub _tree_construction_main ($) {
          unless (defined $i) { # has an element in scope
            !!!cp ('t413');
-           !!!parse-error (type => 'unmatched end tag:'.$token->{tag_name}, token => $token);
+           !!!parse-error (type => 'unmatched end tag',
+                           text => $token->{tag_name}, token => $token);
+           ## NOTE: Ignore the token.
          } else {
            ## Step 1. generate implied end tags
            while ({
+                   ## END_TAG_OPTIONAL_EL
                    dd => ($token->{tag_name} ne 'dd'),
                    dt => ($token->{tag_name} ne 'dt'),
                    li => ($token->{tag_name} ne 'li'),
                    p => 1,
+                   rt => 1,
+                   rp => 1,
                   }->{$self->{open_elements}->[-1]->[0]->manakai_local_name}) {
              !!!cp ('t409');
              pop @{$self->{open_elements}};
-Line 6571 
 sub _tree_construction_main ($) {
+Line 7424 
 sub _tree_construction_main ($) {
                    ne $token->{tag_name}) {
              !!!cp ('t412');
              !!!parse-error (type => 'not closed',
-                             value => $self->{open_elements}->[-1]->[0]
+                             text => $self->{open_elements}->[-1]->[0]
                                  ->manakai_local_name,
                              token => $token);
            } else {
-Line 6608 
 sub _tree_construction_main ($) {
+Line 7461 
 sub _tree_construction_main ($) {
          unless (defined $i) { # has an element in scope
            !!!cp ('t421');
-           !!!parse-error (type => 'unmatched end tag:'.$token->{tag_name}, token => $token);
+           !!!parse-error (type => 'unmatched end tag',
+                           text => $token->{tag_name}, token => $token);
+           ## NOTE: Ignore the token.
          } else {
            ## Step 1. generate implied end tags
            while ($self->{open_elements}->[-1]->[1] & END_TAG_OPTIONAL_EL) {
-Line 6621 
 sub _tree_construction_main ($) {
+Line 7476 
 sub _tree_construction_main ($) {
                    ne $token->{tag_name}) {
              !!!cp ('t417.1');
              !!!parse-error (type => 'not closed',
-                             value => $self->{open_elements}->[-1]->[0]
+                             text => $self->{open_elements}->[-1]->[0]
                                  ->manakai_local_name,
                              token => $token);
            } else {
-Line 6653 
 sub _tree_construction_main ($) {
+Line 7508 
 sub _tree_construction_main ($) {
          unless (defined $i) { # has an element in scope
            !!!cp ('t425.1');
-           !!!parse-error (type => 'unmatched end tag:'.$token->{tag_name}, token => $token);
+           !!!parse-error (type => 'unmatched end tag',
+                           text => $token->{tag_name}, token => $token);
+           ## NOTE: Ignore the token.
          } else {
            ## Step 1. generate implied end tags
            while ($self->{open_elements}->[-1]->[1] & END_TAG_OPTIONAL_EL) {
-Line 6665 
 sub _tree_construction_main ($) {
+Line 7522 
 sub _tree_construction_main ($) {
            if ($self->{open_elements}->[-1]->[0]->manakai_local_name
                    ne $token->{tag_name}) {
              !!!cp ('t425');
-             !!!parse-error (type => 'unmatched end tag:'.$token->{tag_name}, token => $token);
+             !!!parse-error (type => 'unmatched end tag',
+                             text => $token->{tag_name}, token => $token);
            } else {
              !!!cp ('t426');
            }
-Line 6696 
 sub _tree_construction_main ($) {
+Line 7554 
 sub _tree_construction_main ($) {
                    ne $token->{tag_name}) {
              !!!cp ('t412.1');
              !!!parse-error (type => 'not closed',
-                             value => $self->{open_elements}->[-1]->[0]
+                             text => $self->{open_elements}->[-1]->[0]
                                  ->manakai_local_name,
                              token => $token);
            } else {
-Line 6706 
 sub _tree_construction_main ($) {
+Line 7564 
 sub _tree_construction_main ($) {
            splice @{$self->{open_elements}}, $i;
          } else {
            !!!cp ('t413.1');
-           !!!parse-error (type => 'unmatched end tag:'.$token->{tag_name}, token => $token);
+           !!!parse-error (type => 'unmatched end tag',
+                           text => $token->{tag_name}, token => $token);
            !!!cp ('t415.1');
            ## As if <p>, then reprocess the current token
-Line 6729 
 sub _tree_construction_main ($) {
+Line 7588 
 sub _tree_construction_main ($) {
          next B;
        } elsif ($token->{tag_name} eq 'br') {
          !!!cp ('t428');
-         !!!parse-error (type => 'unmatched end tag:br', token => $token);
+         !!!parse-error (type => 'unmatched end tag',
+                         text => 'br', token => $token);
          ## As if <br>
          $reconstruct_active_formatting_elements->($insert_to_current);
-Line 6754 
 sub _tree_construction_main ($) {
+Line 7614 
 sub _tree_construction_main ($) {
                  noscript => 0, ## TODO: if scripting is enabled
                 }->{$token->{tag_name}}) {
          !!!cp ('t429');
-         !!!parse-error (type => 'unmatched end tag:'.$token->{tag_name}, token => $token);
+         !!!parse-error (type => 'unmatched end tag',
+                         text => $token->{tag_name}, token => $token);
          ## Ignore the token
          !!!next-token;
          next B;
-Line 6773 
 sub _tree_construction_main ($) {
+Line 7634 
 sub _tree_construction_main ($) {
              ## generate implied end tags
              while ($self->{open_elements}->[-1]->[1] & END_TAG_OPTIONAL_EL) {
                !!!cp ('t430');
-               ## ISSUE: Can this case be reached?
+               ## NOTE: |<ruby><rt></ruby>|.
+               ## ISSUE: <ruby><rt></rt> will also take this code path,
+               ## which seems wrong.
                pop @{$self->{open_elements}};
+               $node_i++;
              }
              ## Step 2
-Line 6783 
 sub _tree_construction_main ($) {
+Line 7647 
 sub _tree_construction_main ($) {
                !!!cp ('t431');
                ## NOTE: <x><y></x>
                !!!parse-error (type => 'not closed',
-                               value => $self->{open_elements}->[-1]->[0]
+                               text => $self->{open_elements}->[-1]->[0]
                                    ->manakai_local_name,
                                token => $token);
              } else {
-Line 6791 
 sub _tree_construction_main ($) {
+Line 7655 
 sub _tree_construction_main ($) {
              }
              ## Step 3
-             splice @{$self->{open_elements}}, $node_i;
+             splice @{$self->{open_elements}}, $node_i if $node_i < 0;
              !!!next-token;
              last S2;
-Line 6802 
 sub _tree_construction_main ($) {
+Line 7666 
 sub _tree_construction_main ($) {
                  ($node->[1] & SPECIAL_EL or
                   $node->[1] & SCOPING_EL)) {
                !!!cp ('t433');
-               !!!parse-error (type => 'unmatched end tag:'.$token->{tag_name}, token => $token);
+               !!!parse-error (type => 'unmatched end tag',
+                               text => $token->{tag_name}, token => $token);
                ## Ignore the token
                !!!next-token;
                last S2;
-Line 6848 
 sub _tree_construction_main ($) {
+Line 7713 
 sub _tree_construction_main ($) {
    ## TODO: script stuffs
  } # _tree_construct_main
- sub set_inner_html ($$$) {
+ sub set_inner_html ($$$;$) {
    my $class = shift;
    my $node = shift;
    my $s = \$_[0];
    my $onerror = $_[1];
+   my $get_wrapper = $_[2] || sub ($) { return $_[0] };
    ## ISSUE: Should {confident} be true?
-Line 6871 
 sub set_inner_html ($$$) {
+Line 7737 
 sub set_inner_html ($$$) {
      }
      ## Step 3, 4, 5 # MUST
-     $class->parse_string ($$s => $node, $onerror);
+     $class->parse_char_string ($$s => $node, $onerror, $get_wrapper);
    } elsif ($nt == 1) {
      ## TODO: If non-html element
      ## NOTE: Most of this code is copied from |parse_string|
+ ## TODO: Support for $get_wrapper
      ## Step 1 # MUST
      my $this_doc = $node->owner_document;
      my $doc = $this_doc->implementation->create_document;
-Line 6917 
 sub set_inner_html ($$$) {
+Line 7785 
 sub set_inner_html ($$$) {
          !!!cp ('i4');
          !!!parse-error (type => 'NULL');
          $self->{next_char} = 0xFFFD; # REPLACEMENT CHARACTER # MUST
+       } elsif ($self->{next_char} <= 0x0008 or
+                (0x000E <= $self->{next_char} and
+                 $self->{next_char} <= 0x001F) or
+                (0x007F <= $self->{next_char} and
+                 $self->{next_char} <= 0x009F) or
+                (0xD800 <= $self->{next_char} and
+                 $self->{next_char} <= 0xDFFF) or
+                (0xFDD0 <= $self->{next_char} and
+                 $self->{next_char} <= 0xFDDF) or
+                {
+xFFFE => 1, 0xFFFF => 1, 0x1FFFE => 1, 0x1FFFF => 1,
+x2FFFE => 1, 0x2FFFF => 1, 0x3FFFE => 1, 0x3FFFF => 1,
+x4FFFE => 1, 0x4FFFF => 1, 0x5FFFE => 1, 0x5FFFF => 1,
+x6FFFE => 1, 0x6FFFF => 1, 0x7FFFE => 1, 0x7FFFF => 1,
+x8FFFE => 1, 0x8FFFF => 1, 0x9FFFE => 1, 0x9FFFF => 1,
+xAFFFE => 1, 0xAFFFF => 1, 0xBFFFE => 1, 0xBFFFF => 1,
+xCFFFE => 1, 0xCFFFF => 1, 0xDFFFE => 1, 0xDFFFF => 1,
+xEFFFE => 1, 0xEFFFF => 1, 0xFFFFE => 1, 0xFFFFF => 1,
+x10FFFE => 1, 0x10FFFF => 1,
+                }->{$self->{next_char}}) {
+         !!!cp ('i4.1');
+         if ($self->{next_char} < 0x10000) {
+           !!!parse-error (type => 'control char',
+                           text => (sprintf 'U+%04X', $self->{next_char}));
+         } else {
+           !!!parse-error (type => 'control char',
+                           text => (sprintf 'U-%08X', $self->{next_char}));
+         }
        }
      };
      $p->{prev_char} = [-1, -1, -1];
      $p->{next_char} = -1;
+     $p->{getc_until} = sub {
+       ## TODO: ...
+       return undef;
+     }; # $p->{getc_until};
      my $ponerror = $onerror || sub {
        my (%opt) = @_;
        my $line = $opt{line};

 Legend:



Removed from v.1.129
 


changed lines


 
Added in v.1.171
 Legend:



Removed from v.1.129
 


changed lines


 
Added in v.1.171
-Removed from v.1.129
+Added in v.1.171

[email protected]	ViewVC Help
Powered by ViewVC 1.1.24