/[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 1.78 by wakaba,
Mon Mar  3 11:56:18 2008 UTC
+revision 1.153 by wakaba,
Fri Aug 15 08:32:41 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>;
- my $permitted_slash_tag_name = {
+ my $SVG_NS = q<http://www.w3.org/2000/svg>;
-   base => 1,
+ my $XLINK_NS = q<http://www.w3.org/1999/xlink>;
-   link => 1,
+ my $XML_NS = q<http://www.w3.org/XML/1998/namespace>;
-   meta => 1,
+ my $XMLNS_NS = q<http://www.w3.org/2000/xmlns/>;
-   hr => 1,
-   br => 1,
+ sub A_EL () { 0b1 }
-   img => 1,
+ sub ADDRESS_EL () { 0b10 }
-   embed => 1,
+ sub BODY_EL () { 0b100 }
-   param => 1,
+ sub BUTTON_EL () { 0b1000 }
-   area => 1,
+ sub CAPTION_EL () { 0b10000 }
-   col => 1,
+ sub DD_EL () { 0b100000 }
-   input => 1,
+ sub DIV_EL () { 0b1000000 }
+ sub DT_EL () { 0b10000000 }
+ sub FORM_EL () { 0b100000000 }
+ sub FORMATTING_EL () { 0b1000000000 }
+ sub FRAMESET_EL () { 0b10000000000 }
+ sub HEADING_EL () { 0b100000000000 }
+ sub HTML_EL () { 0b1000000000000 }
+ sub LI_EL () { 0b10000000000000 }
+ sub NOBR_EL () { 0b100000000000000 }
+ sub OPTION_EL () { 0b1000000000000000 }
+ sub OPTGROUP_EL () { 0b10000000000000000 }
+ sub P_EL () { 0b100000000000000000 }
+ sub SELECT_EL () { 0b1000000000000000000 }
+ sub TABLE_EL () { 0b10000000000000000000 }
+ sub TABLE_CELL_EL () { 0b100000000000000000000 }
+ sub TABLE_ROW_EL () { 0b1000000000000000000000 }
+ sub TABLE_ROW_GROUP_EL () { 0b10000000000000000000000 }
+ sub MISC_SCOPING_EL () { 0b100000000000000000000000 }
+ sub MISC_SPECIAL_EL () { 0b1000000000000000000000000 }
+ 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 |
+   TABLE_ROW_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 |
+   RUBY_COMPONENT_EL
+ }
+ ## NOTE: Used in </body> and EOF algorithms.
+ sub ALL_END_TAG_OPTIONAL_EL () {
+   DD_EL |
+   DT_EL |
+   LI_EL |
+   P_EL |
+   BODY_EL |
+   HTML_EL |
+   TABLE_CELL_EL |
+   TABLE_ROW_EL |
+   TABLE_ROW_GROUP_EL
+ }
+ sub SCOPING_EL () {
+   BUTTON_EL |
+   CAPTION_EL |
+   HTML_EL |
+   TABLE_EL |
+   TABLE_CELL_EL |
+   MISC_SCOPING_EL
+ }
+ sub TABLE_SCOPING_EL () {
+   HTML_EL |
+   TABLE_EL
+ }
+ sub TABLE_ROWS_SCOPING_EL () {
+   HTML_EL |
+   TABLE_ROW_GROUP_EL
+ }
+ sub TABLE_ROW_SCOPING_EL () {
+   HTML_EL |
+   TABLE_ROW_EL
+ }
+ sub SPECIAL_EL () {
+   ADDRESS_EL |
+   BODY_EL |
+   DIV_EL |
+   DD_EL |
+   DT_EL |
+   LI_EL |
+   P_EL |
+   FORM_EL |
+   FRAMESET_EL |
+   HEADING_EL |
+   OPTION_EL |
+   OPTGROUP_EL |
+   SELECT_EL |
+   TABLE_ROW_EL |
+   TABLE_ROW_GROUP_EL |
+   MISC_SPECIAL_EL
+ }
+ my $el_category = {
+   a => A_EL | FORMATTING_EL,
+   address => ADDRESS_EL,
+   applet => MISC_SCOPING_EL,
+   area => MISC_SPECIAL_EL,
+   b => FORMATTING_EL,
+   base => MISC_SPECIAL_EL,
+   basefont => MISC_SPECIAL_EL,
+   bgsound => MISC_SPECIAL_EL,
+   big => FORMATTING_EL,
+   blockquote => MISC_SPECIAL_EL,
+   body => BODY_EL,
+   br => MISC_SPECIAL_EL,
+   button => BUTTON_EL,
+   caption => CAPTION_EL,
+   center => MISC_SPECIAL_EL,
+   col => MISC_SPECIAL_EL,
+   colgroup => MISC_SPECIAL_EL,
+   dd => DD_EL,
+   dir => MISC_SPECIAL_EL,
+   div => DIV_EL,
+   dl => MISC_SPECIAL_EL,
+   dt => DT_EL,
+   em => FORMATTING_EL,
+   embed => MISC_SPECIAL_EL,
+   fieldset => MISC_SPECIAL_EL,
+   font => FORMATTING_EL,
+   form => FORM_EL,
+   frame => MISC_SPECIAL_EL,
+   frameset => FRAMESET_EL,
+   h1 => HEADING_EL,
+   h2 => HEADING_EL,
+   h3 => HEADING_EL,
+   h4 => HEADING_EL,
+   h5 => HEADING_EL,
+   h6 => HEADING_EL,
+   head => MISC_SPECIAL_EL,
+   hr => MISC_SPECIAL_EL,
+   html => HTML_EL,
+   i => FORMATTING_EL,
+   iframe => MISC_SPECIAL_EL,
+   img => MISC_SPECIAL_EL,
+   input => MISC_SPECIAL_EL,
+   isindex => MISC_SPECIAL_EL,
+   li => LI_EL,
+   link => MISC_SPECIAL_EL,
+   listing => MISC_SPECIAL_EL,
+   marquee => MISC_SCOPING_EL,
+   menu => MISC_SPECIAL_EL,
+   meta => MISC_SPECIAL_EL,
+   nobr => NOBR_EL | FORMATTING_EL,
+   noembed => MISC_SPECIAL_EL,
+   noframes => MISC_SPECIAL_EL,
+   noscript => MISC_SPECIAL_EL,
+   object => MISC_SCOPING_EL,
+   ol => MISC_SPECIAL_EL,
+   optgroup => OPTGROUP_EL,
+   option => OPTION_EL,
+   p => P_EL,
+   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,
+   small => FORMATTING_EL,
+   spacer => MISC_SPECIAL_EL,
+   strike => FORMATTING_EL,
+   strong => FORMATTING_EL,
+   style => MISC_SPECIAL_EL,
+   table => TABLE_EL,
+   tbody => TABLE_ROW_GROUP_EL,
+   td => TABLE_CELL_EL,
+   textarea => MISC_SPECIAL_EL,
+   tfoot => TABLE_ROW_GROUP_EL,
+   th => TABLE_CELL_EL,
+   thead => TABLE_ROW_GROUP_EL,
+   title => MISC_SPECIAL_EL,
+   tr => TABLE_ROW_EL,
+   tt => FORMATTING_EL,
+   u => FORMATTING_EL,
+   ul => MISC_SPECIAL_EL,
+   wbr => MISC_SPECIAL_EL,
+ };
+ my $el_category_f = {
+   $MML_NS => {
+     'annotation-xml' => MML_AXML_EL,
+     mi => FOREIGN_FLOW_CONTENT_EL,
+     mo => FOREIGN_FLOW_CONTENT_EL,
+     mn => FOREIGN_FLOW_CONTENT_EL,
+     ms => FOREIGN_FLOW_CONTENT_EL,
+     mtext => FOREIGN_FLOW_CONTENT_EL,
+   },
+   $SVG_NS => {
+     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 61 
 my $c1_entity_char = {
+Line 347 
 my $c1_entity_char = {
 x9F => 0x0178,
  }; # $c1_entity_char
- my $special_category = {
-   address => 1, area => 1, base => 1, basefont => 1, bgsound => 1,
-   blockquote => 1, body => 1, br => 1, center => 1, col => 1, colgroup => 1,
-   dd => 1, dir => 1, div => 1, dl => 1, dt => 1, embed => 1, fieldset => 1,
-   form => 1, frame => 1, frameset => 1, h1 => 1, h2 => 1, h3 => 1,
-   h4 => 1, h5 => 1, h6 => 1, head => 1, hr => 1, iframe => 1, image => 1,
-   img => 1, input => 1, isindex => 1, li => 1, link => 1, listing => 1,
-   menu => 1, meta => 1, noembed => 1, noframes => 1, noscript => 1,
-   ol => 1, optgroup => 1, option => 1, p => 1, param => 1, plaintext => 1,
-   pre => 1, script => 1, select => 1, spacer => 1, style => 1, tbody => 1,
-   textarea => 1, tfoot => 1, thead => 1, title => 1, tr => 1, ul => 1, wbr => 1,
- };
- my $scoping_category = {
-   button => 1, caption => 1, html => 1, marquee => 1, object => 1,
-   table => 1, td => 1, th => 1,
- };
- my $formatting_category = {
-   a => 1, b => 1, big => 1, em => 1, font => 1, i => 1, nobr => 1,
-   s => 1, small => 1, strile => 1, strong => 1, tt => 1, u => 1,
- };
- # $phrasing_category: all other elements
  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 = 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;
-   if (defined $charset) {
-     require Encode; ## TODO: decode(utf8) don't delete BOM
-     $s = \ (Encode::decode ($charset, $$bytes_s));
-     $self->{input_encoding} = lc $charset; ## TODO: normalize name
-     $self->{confident} = 1;
-   } else {
-     ## TODO: Implement HTML5 detection algorithm
-     require Whatpm::Charset::UniversalCharDet;
-     $charset = Whatpm::Charset::UniversalCharDet->detect_byte_string
-         (substr ($$bytes_s, 0, 1024));
-     $charset ||= 'windows-1252';
-     $s = \ (Encode::decode ($charset, $$bytes_s));
-     $self->{input_encoding} = $charset;
-     $self->{confident} = 0;
-   }
-   $self->{change_encoding} = sub {
+   my $onerror = $_[2] || sub {
-     my $self = shift;
+     my (%opt) = @_;
-     my $charset = lc shift;
+     warn "Parse error ($opt{type})\n";
-     ## TODO: if $charset is supported
+   };
-     ## TODO: normalize charset name
+   $self->{parse_error} = $onerror; # updated later by parse_char_string
-     ## "Change the encoding" algorithm:
+   ## HTML5 encoding sniffing algorithm
+   require Message::Charset::Info;
-     ## Step 1
+   my $charset;
-     if ($charset eq 'utf-16') { ## ISSUE: UTF-16BE -> UTF-8? UTF-16LE -> UTF-8?
+   my $buffer;
-       $charset = 'utf-8';
+   my ($char_stream, $e_status);
+   SNIFFING: {
+     ## Step 1
+     if (defined $charset_name) {
+       $charset = Message::Charset::Info->get_by_iana_name ($charset_name);
+       ## 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
-     if (defined $self->{input_encoding} and
+     my $byte_buffer = '';
-         $self->{input_encoding} eq $charset) {
+     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_iana_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;
-       return;
+       last SNIFFING;
+     } elsif ($byte_buffer =~ /^\xFF\xFE/) {
+       $charset = Message::Charset::Info->get_by_iana_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_iana_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;
      }
-     !!!parse-error (type => 'charset label detected:'.$self->{input_encoding}.
+     ## Step 4
-         ':'.$charset, level => 'w');
+     ## TODO: <meta charset>
-     ## Step 3
+     ## Step 5
-     # if (can) {
+     ## TODO: from history
-       ## change the encoding on the fly.
-       #$self->{confident} = 1;
-       #return;
-     # }
-     ## Step 4
+     ## Step 6
-     throw Whatpm::HTML::RestartParser (charset => $charset);
+     require Whatpm::Charset::UniversalCharDet;
+     $charset_name = Whatpm::Charset::UniversalCharDet->detect_byte_string
+         ($byte_buffer);
+     if (defined $charset_name) {
+       $charset = Message::Charset::Info->get_by_iana_name ($charset_name);
+       ## 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_iana_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
+   $self->{input_encoding} = $charset->get_iana_name;
+   if ($e_status & Message::Charset::Info::FALLBACK_ENCODING_IMPL ()) {
+     !!!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())) {
+     !!!parse-error (type => 'chardecode:no error',
+                     text => $self->{input_encoding},
+                     level => $self->{level}->{uncertain},
+                     line => 1, column => 1,
+                     layer => 'encode');
+   }
+   $self->{change_encoding} = sub {
+     my $self = shift;
+     $charset_name = shift;
+     my $token = shift;
+     $charset = Message::Charset::Info->get_by_iana_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
+       if ($charset->{category} &
+           Message::Charset::Info::CHARSET_CATEGORY_UTF16 ()) {
+         $charset = Message::Charset::Info->get_by_iana_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;
+       }
+       !!!parse-error (type => 'charset label detected',
+                       text => $self->{input_encoding},
+                       value => $charset_name,
+                       level => $self->{level}->{warn},
+                       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}
+   my $char_onerror = sub {
+     my (undef, $type, %opt) = @_;
+     !!!parse-error (layer => 'encode',
+                     %opt, type => $type,
+                     line => $self->{line}, column => $self->{column} + 1);
+     if ($opt{octets}) {
+       ${$opt{octets}} = "\x{FFFD}"; # relacement character
+     }
+   };
+   $char_stream->onerror ($char_onerror);
    my @args = @_; shift @args; # $s
    my $return;
    try {
-     $return = $self->parse_char_string ($s, @args);
+     $return = $self->parse_char_stream ($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
+     $self->{input_encoding} = $charset->get_iana_name;
+     if ($e_status & Message::Charset::Info::FALLBACK_ENCODING_IMPL ()) {
+       !!!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())) {
+       !!!parse-error (type => 'chardecode:no error',
+                       text => $self->{input_encoding},
+                       level => $self->{level}->{uncertain},
+                       line => 1, column => 1,
+                       layer => 'encode');
+     }
      $self->{confident} = 1;
-     $return = $self->parse_char_string ($s, @args);
+     $char_stream->onerror ($char_onerror);
+     $return = $self->parse_char_stream ($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 162 
 sub parse_byte_string ($$$$;$) {
+Line 591 
 sub parse_byte_string ($$$$;$) {
  ## such as |parse_byte_string| in this module, must ensure that it does
  ## strip the BOM and never strip any ZWNBSP.
- *parse_char_string = \&parse_string;
+ sub parse_char_string ($$$;$) {
+   my $self = shift;
+   require utf8;
+   my $s = ref $_[0] ? $_[0] : \($_[0]);
+   open my $input, '<' . (utf8::is_utf8 ($$s) ? ':utf8' : ''), $s;
+   return $self->parse_char_stream ($input, @_[1..$#_]);
+ } # parse_char_string
+ *parse_string = \&parse_char_string;
- sub parse_string ($$$;$) {
+ 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 177 
 sub parse_string ($$$;$) {
+Line 613 
 sub parse_string ($$$;$) {
        if defined $self->{input_encoding};
    my $i = 0;
-   my $line = 1;
+   $self->{line_prev} = $self->{line} = 1;
-   my $column = 0;
+   $self->{column_prev} = $self->{column} = 0;
    $self->{set_next_char} = sub {
      my $self = shift;
      pop @{$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}) {
-     $column++;
+       $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
-       $line++;
+       !!!cp ('j1');
-       $column = 0;
+       $self->{line}++;
+       $self->{column} = 0;
      } elsif ($self->{next_char} == 0x000D) { # CR
-       $i++ if substr ($$s, $i, 1) eq "\x0A";
+       !!!cp ('j2');
+       my $next = $input->getc;
+       if (defined $next and $next ne "\x0A") {
+         $self->{next_next_char} = $next;
+       }
        $self->{next_char} = 0x000A; # LF # MUST
-       $line++;
+       $self->{line}++;
-       $column = 0;
+       $self->{column} = 0;
      } elsif ($self->{next_char} > 0x10FFFF) {
+       !!!cp ('j3');
        $self->{next_char} = 0xFFFD; # REPLACEMENT CHARACTER # MUST
      } elsif ($self->{next_char} == 0x0000) { # NULL
+       !!!cp ('j4');
        !!!parse-error (type => 'NULL');
        $self->{next_char} = 0xFFFD; # REPLACEMENT CHARACTER # MUST
+     } elsif ($self->{next_char} <= 0x0008 or
+              (0x000E <= $self->{next_char} and $self->{next_char} <= 0x001F) or
+              (0x007F <= $self->{next_char} and $self->{next_char} <= 0x009F) or
+              (0xD800 <= $self->{next_char} and $self->{next_char} <= 0xDFFF) or
+              (0xFDD0 <= $self->{next_char} and $self->{next_char} <= 0xFDDF) or
+              {
+xFFFE => 1, 0xFFFF => 1, 0x1FFFE => 1, 0x1FFFF => 1,
+x2FFFE => 1, 0x2FFFF => 1, 0x3FFFE => 1, 0x3FFFF => 1,
+x4FFFE => 1, 0x4FFFF => 1, 0x5FFFE => 1, 0x5FFFF => 1,
+x6FFFE => 1, 0x6FFFF => 1, 0x7FFFE => 1, 0x7FFFF => 1,
+x8FFFE => 1, 0x8FFFF => 1, 0x9FFFE => 1, 0x9FFFF => 1,
+xAFFFE => 1, 0xAFFFF => 1, 0xBFFFE => 1, 0xBFFFF => 1,
+xCFFFE => 1, 0xCFFFF => 1, 0xDFFFE => 1, 0xDFFFF => 1,
+xEFFFE => 1, 0xEFFFF => 1, 0xFFFFE => 1, 0xFFFFF => 1,
+x10FFFE => 1, 0x10FFFF => 1,
+              }->{$self->{next_char}}) {
+       !!!cp ('j5');
+       if ($self->{next_char} < 0x10000) {
+         !!!parse-error (type => 'control char',
+                         text => (sprintf 'U+%04X', $self->{next_char}));
+       } else {
+         !!!parse-error (type => 'control char',
+                         text => (sprintf 'U-%08X', $self->{next_char}));
+       }
      }
    };
    $self->{prev_char} = [-1, -1, -1];
-Line 209 
 sub parse_string ($$$;$) {
+Line 686 
 sub parse_string ($$$;$) {
    my $onerror = $_[2] || sub {
      my (%opt) = @_;
-     warn "Parse error ($opt{type}) at line $opt{line} column $opt{column}\n";
+     my $line = $opt{token} ? $opt{token}->{line} : $opt{line};
+     my $column = $opt{token} ? $opt{token}->{column} : $opt{column};
+     warn "Parse error ($opt{type}) at line $line column $column\n";
    };
    $self->{parse_error} = sub {
-     $onerror->(@_, line => $line, column => $column);
+     $onerror->(line => $self->{line}, column => $self->{column}, @_);
    };
    $self->_initialize_tokenizer;
-Line 220 
 sub parse_string ($$$;$) {
+Line 699 
 sub parse_string ($$$;$) {
    $self->_construct_tree;
    $self->_terminate_tree_constructor;
+   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',
+               warn => 'w',
+               info => 'i',
+               uncertain => 'u'},
+   }, $class;
    $self->{set_next_char} = sub {
      $self->{next_char} = -1;
    };
-Line 287 
 sub DOCTYPE_SYSTEM_IDENTIFIER_SINGLE_QUO
+Line 773 
 sub DOCTYPE_SYSTEM_IDENTIFIER_SINGLE_QUO
  sub AFTER_DOCTYPE_SYSTEM_IDENTIFIER_STATE () { 31 }
  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 DOCTYPE_TOKEN () { 1 }
  sub COMMENT_TOKEN () { 2 }
-Line 303 
 sub TABLE_IMS ()      { 0b1000000 }
+Line 791 
 sub TABLE_IMS ()      { 0b1000000 }
  sub ROW_IMS ()        { 0b10000000 }
  sub BODY_AFTER_IMS () { 0b100000000 }
  sub FRAME_IMS ()      { 0b1000000000 }
+ sub SELECT_IMS ()     { 0b10000000000 }
+ sub IN_FOREIGN_CONTENT_IM () { 0b100000000000 }
+     ## NOTE: "in foreign content" insertion mode is special; it is combined
+     ## with the secondary insertion mode.  In this parser, they are stored
+     ## together in the bit-or'ed form.
+ ## NOTE: "initial" and "before html" insertion modes have no constants.
+ ## NOTE: "after after body" insertion mode.
  sub AFTER_HTML_BODY_IM () { AFTER_HTML_IMS | BODY_AFTER_IMS }
+ ## NOTE: "after after frameset" insertion mode.
  sub AFTER_HTML_FRAMESET_IM () { AFTER_HTML_IMS | FRAME_IMS }
  sub IN_HEAD_IM () { HEAD_IMS | 0b00 }
  sub IN_HEAD_NOSCRIPT_IM () { HEAD_IMS | 0b01 }
  sub AFTER_HEAD_IM () { HEAD_IMS | 0b10 }
-Line 319 
 sub IN_TABLE_IM () { TABLE_IMS }
+Line 818 
 sub IN_TABLE_IM () { TABLE_IMS }
  sub AFTER_BODY_IM () { BODY_AFTER_IMS }
  sub IN_FRAMESET_IM () { FRAME_IMS | 0b01 }
  sub AFTER_FRAMESET_IM () { FRAME_IMS | 0b10 }
- sub IN_SELECT_IM () { 0b01 }
+ sub IN_SELECT_IM () { SELECT_IMS | 0b01 }
+ sub IN_SELECT_IN_TABLE_IM () { SELECT_IMS | 0b10 }
  sub IN_COLUMN_GROUP_IM () { 0b10 }
  ## Implementations MUST act as if state machine in the spec
-Line 332 
 sub _initialize_tokenizer ($) {
+Line 832 
 sub _initialize_tokenizer ($) {
    undef $self->{current_attribute};
    undef $self->{last_emitted_start_tag_name};
    undef $self->{last_attribute_value_state};
+   delete $self->{self_closing};
    $self->{char} = [];
    # $self->{next_char}
    !!!next-input-character;
-Line 352 
 sub _initialize_tokenizer ($) {
+Line 853 
 sub _initialize_tokenizer ($) {
  ##        ->{value}
  ##        ->{has_reference} == 1 or 0
  ##   ->{data} (COMMENT_TOKEN, CHARACTER_TOKEN)
+ ## NOTE: The "self-closing flag" is hold as |$self->{self_closing}|.
+ ##     |->{self_closing}| is used to save the value of |$self->{self_closing}|
+ ##     while the token is pushed back to the stack.
  ## Emitted token MUST immediately be handled by the tree construction state.
-Line 378 
 sub _initialize_tokenizer ($) {
+Line 882 
 sub _initialize_tokenizer ($) {
  sub _get_next_token ($) {
    my $self = shift;
+   if ($self->{self_closing}) {
+     !!!parse-error (type => 'nestc', token => $self->{current_token});
+     ## NOTE: The |self_closing| flag is only set by start tag token.
+     ## In addition, when a start tag token is emitted, it is always set to
+     ## |current_token|.
+     delete $self->{self_closing};
+   }
    if (@{$self->{token}}) {
+     $self->{self_closing} = $self->{token}->[0]->{self_closing};
      return shift @{$self->{token}};
    }
-Line 441 
 sub _get_next_token ($) {
+Line 955 
 sub _get_next_token ($) {
          #
        } elsif ($self->{next_char} == -1) {
          !!!cp (11);
-         !!!emit ({type => END_OF_FILE_TOKEN});
+         !!!emit ({type => END_OF_FILE_TOKEN,
+                   line => $self->{line}, column => $self->{column}});
          last A; ## TODO: ok?
        } else {
          !!!cp (12);
        }
        # Anything else
        my $token = {type => CHARACTER_TOKEN,
-                    data => chr $self->{next_char}};
+                    data => chr $self->{next_char},
+                    line => $self->{line}, column => $self->{column},
+                   };
        ## Stay in the data state
        !!!next-input-character;
-Line 457 
 sub _get_next_token ($) {
+Line 974 
 sub _get_next_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);
-Line 465 
 sub _get_next_token ($) {
+Line 984 
 sub _get_next_token ($) {
        unless (defined $token) {
          !!!cp (13);
-         !!!emit ({type => CHARACTER_TOKEN, data => '&'});
+         !!!emit ({type => CHARACTER_TOKEN, data => '&',
+                   line => $l, column => $c,
+                  });
        } else {
          !!!cp (14);
          !!!emit ($token);
-Line 484 
 sub _get_next_token ($) {
+Line 1005 
 sub _get_next_token ($) {
            ## reconsume
            $self->{state} = DATA_STATE;
-           !!!emit ({type => CHARACTER_TOKEN, data => '<'});
+           !!!emit ({type => CHARACTER_TOKEN, data => '<',
+                     line => $self->{line_prev},
+                     column => $self->{column_prev},
+                    });
            redo A;
          }
-Line 504 
 sub _get_next_token ($) {
+Line 1028 
 sub _get_next_token ($) {
            !!!cp (19);
            $self->{current_token}
              = {type => START_TAG_TOKEN,
-                tag_name => chr ($self->{next_char} + 0x0020)};
+                tag_name => chr ($self->{next_char} + 0x0020),
+                line => $self->{line_prev},
+                column => $self->{column_prev}};
            $self->{state} = TAG_NAME_STATE;
            !!!next-input-character;
            redo A;
-Line 512 
 sub _get_next_token ($) {
+Line 1038 
 sub _get_next_token ($) {
                   $self->{next_char} <= 0x007A) { # a..z
            !!!cp (20);
            $self->{current_token} = {type => START_TAG_TOKEN,
-                             tag_name => chr ($self->{next_char})};
+                                     tag_name => chr ($self->{next_char}),
+                                     line => $self->{line_prev},
+                                     column => $self->{column_prev}};
            $self->{state} = TAG_NAME_STATE;
            !!!next-input-character;
            redo A;
          } elsif ($self->{next_char} == 0x003E) { # >
            !!!cp (21);
-           !!!parse-error (type => 'empty start tag');
+           !!!parse-error (type => 'empty start tag',
+                           line => $self->{line_prev},
+                           column => $self->{column_prev});
            $self->{state} = DATA_STATE;
            !!!next-input-character;
-           !!!emit ({type => CHARACTER_TOKEN, data => '<>'});
+           !!!emit ({type => CHARACTER_TOKEN, data => '<>',
+                     line => $self->{line_prev},
+                     column => $self->{column_prev},
+                    });
            redo A;
          } elsif ($self->{next_char} == 0x003F) { # ?
            !!!cp (22);
-           !!!parse-error (type => 'pio');
+           !!!parse-error (type => 'pio',
+                           line => $self->{line_prev},
+                           column => $self->{column_prev});
            $self->{state} = BOGUS_COMMENT_STATE;
+           $self->{current_token} = {type => COMMENT_TOKEN, data => '',
+                                     line => $self->{line_prev},
+                                     column => $self->{column_prev},
+                                    };
            ## $self->{next_char} is intentionally left as is
            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
-           !!!emit ({type => CHARACTER_TOKEN, data => '<'});
+           !!!emit ({type => CHARACTER_TOKEN, data => '<',
+                     line => $self->{line_prev},
+                     column => $self->{column_prev},
+                    });
            redo A;
          }
-Line 545 
 sub _get_next_token ($) {
+Line 1089 
 sub _get_next_token ($) {
          die "$0: $self->{content_model} in tag open";
        }
      } elsif ($self->{state} == CLOSE_TAG_OPEN_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}) {
            ## NOTE: <http://krijnhoetmer.nl/irc-logs/whatwg/20070626#l-564>
            my @next_char;
            TAGNAME: for (my $i = 0; $i < length $self->{last_emitted_start_tag_name}; $i++) {
-Line 563 
 sub _get_next_token ($) {
+Line 1109 
 sub _get_next_token ($) {
                !!!back-next-input-character (@next_char);
                $self->{state} = DATA_STATE;
-               !!!emit ({type => CHARACTER_TOKEN, data => '</'});
+               !!!emit ({type => CHARACTER_TOKEN, data => '</',
+                         line => $l, column => $c,
+                        });
                redo A;
              }
-Line 582 
 sub _get_next_token ($) {
+Line 1130 
 sub _get_next_token ($) {
              $self->{next_char} = shift @next_char; # reconsume
              !!!back-next-input-character (@next_char);
              $self->{state} = DATA_STATE;
-             !!!emit ({type => CHARACTER_TOKEN, data => '</'});
+             !!!emit ({type => CHARACTER_TOKEN, data => '</',
+                       line => $l, column => $c,
+                      });
              redo A;
            } else {
              !!!cp (27);
-Line 595 
 sub _get_next_token ($) {
+Line 1145 
 sub _get_next_token ($) {
            !!!cp (28);
            # next-input-character is already done
            $self->{state} = DATA_STATE;
-           !!!emit ({type => CHARACTER_TOKEN, data => '</'});
+           !!!emit ({type => CHARACTER_TOKEN, data => '</',
+                     line => $l, column => $c,
+                    });
            redo A;
          }
        }
-Line 603 
 sub _get_next_token ($) {
+Line 1155 
 sub _get_next_token ($) {
        if (0x0041 <= $self->{next_char} and
            $self->{next_char} <= 0x005A) { # A..Z
          !!!cp (29);
-         $self->{current_token} = {type => END_TAG_TOKEN,
+         $self->{current_token}
-                           tag_name => chr ($self->{next_char} + 0x0020)};
+             = {type => END_TAG_TOKEN,
+                tag_name => chr ($self->{next_char} + 0x0020),
+                line => $l, column => $c};
          $self->{state} = TAG_NAME_STATE;
          !!!next-input-character;
          redo A;
-Line 612 
 sub _get_next_token ($) {
+Line 1166 
 sub _get_next_token ($) {
                 $self->{next_char} <= 0x007A) { # a..z
          !!!cp (30);
          $self->{current_token} = {type => END_TAG_TOKEN,
-                           tag_name => chr ($self->{next_char})};
+                                   tag_name => chr ($self->{next_char}),
+                                   line => $l, column => $c};
          $self->{state} = TAG_NAME_STATE;
          !!!next-input-character;
          redo A;
        } elsif ($self->{next_char} == 0x003E) { # >
          !!!cp (31);
-         !!!parse-error (type => 'empty end tag');
+         !!!parse-error (type => 'empty end tag',
+                         line => $self->{line_prev}, ## "<" in "</>"
+                         column => $self->{column_prev} - 1);
          $self->{state} = DATA_STATE;
          !!!next-input-character;
          redo A;
-Line 628 
 sub _get_next_token ($) {
+Line 1185 
 sub _get_next_token ($) {
          $self->{state} = DATA_STATE;
          # reconsume
-         !!!emit ({type => CHARACTER_TOKEN, data => '</'});
+         !!!emit ({type => CHARACTER_TOKEN, data => '</',
+                   line => $l, column => $c,
+                  });
          redo A;
        } else {
          !!!cp (33);
          !!!parse-error (type => 'bogus end tag');
          $self->{state} = BOGUS_COMMENT_STATE;
+         $self->{current_token} = {type => COMMENT_TOKEN, data => '',
+                                   line => $self->{line_prev}, # "<" of "</"
+                                   column => $self->{column_prev} - 1,
+                                  };
          ## $self->{next_char} is intentionally left as is
          redo A;
        }
-Line 651 
 sub _get_next_token ($) {
+Line 1214 
 sub _get_next_token ($) {
        } elsif ($self->{next_char} == 0x003E) { # >
          if ($self->{current_token}->{type} == START_TAG_TOKEN) {
            !!!cp (35);
-           $self->{current_token}->{first_start_tag}
-               = not defined $self->{last_emitted_start_tag_name};
            $self->{last_emitted_start_tag_name} = $self->{current_token}->{tag_name};
          } elsif ($self->{current_token}->{type} == END_TAG_TOKEN) {
            $self->{content_model} = PCDATA_CONTENT_MODEL; # MUST
-Line 684 
 sub _get_next_token ($) {
+Line 1245 
 sub _get_next_token ($) {
          !!!parse-error (type => 'unclosed tag');
          if ($self->{current_token}->{type} == START_TAG_TOKEN) {
            !!!cp (39);
-           $self->{current_token}->{first_start_tag}
-               = not defined $self->{last_emitted_start_tag_name};
            $self->{last_emitted_start_tag_name} = $self->{current_token}->{tag_name};
          } elsif ($self->{current_token}->{type} == END_TAG_TOKEN) {
            $self->{content_model} = PCDATA_CONTENT_MODEL; # MUST
-Line 706 
 sub _get_next_token ($) {
+Line 1265 
 sub _get_next_token ($) {
          redo A;
        } elsif ($self->{next_char} == 0x002F) { # /
+         !!!cp (42);
+         $self->{state} = SELF_CLOSING_START_TAG_STATE;
          !!!next-input-character;
-         if ($self->{next_char} == 0x003E and # >
-             $self->{current_token}->{type} == START_TAG_TOKEN and
-             $permitted_slash_tag_name->{$self->{current_token}->{tag_name}}) {
-           # permitted slash
-           !!!cp (42);
-           #
-         } else {
-           !!!cp (43);
-           !!!parse-error (type => 'nestc');
-         }
-         $self->{state} = BEFORE_ATTRIBUTE_NAME_STATE;
-         # next-input-character is already done
          redo A;
        } else {
          !!!cp (44);
-Line 741 
 sub _get_next_token ($) {
+Line 1290 
 sub _get_next_token ($) {
        } elsif ($self->{next_char} == 0x003E) { # >
          if ($self->{current_token}->{type} == START_TAG_TOKEN) {
            !!!cp (46);
-           $self->{current_token}->{first_start_tag}
-               = not defined $self->{last_emitted_start_tag_name};
            $self->{last_emitted_start_tag_name} = $self->{current_token}->{tag_name};
          } elsif ($self->{current_token}->{type} == END_TAG_TOKEN) {
            $self->{content_model} = PCDATA_CONTENT_MODEL; # MUST
-Line 764 
 sub _get_next_token ($) {
+Line 1311 
 sub _get_next_token ($) {
        } elsif (0x0041 <= $self->{next_char} and
                 $self->{next_char} <= 0x005A) { # A..Z
          !!!cp (49);
-         $self->{current_attribute} = {name => chr ($self->{next_char} + 0x0020),
+         $self->{current_attribute}
-                               value => ''};
+             = {name => chr ($self->{next_char} + 0x0020),
+                value => '',
+                line => $self->{line}, column => $self->{column}};
          $self->{state} = ATTRIBUTE_NAME_STATE;
          !!!next-input-character;
          redo A;
        } elsif ($self->{next_char} == 0x002F) { # /
+         !!!cp (50);
+         $self->{state} = SELF_CLOSING_START_TAG_STATE;
          !!!next-input-character;
-         if ($self->{next_char} == 0x003E and # >
-             $self->{current_token}->{type} == START_TAG_TOKEN and
-             $permitted_slash_tag_name->{$self->{current_token}->{tag_name}}) {
-           # permitted slash
-           !!!cp (50);
-           #
-         } else {
-           !!!cp (51);
-           !!!parse-error (type => 'nestc');
-         }
-         ## Stay in the state
-         # next-input-character is already done
          redo A;
        } elsif ($self->{next_char} == -1) {
          !!!parse-error (type => 'unclosed tag');
          if ($self->{current_token}->{type} == START_TAG_TOKEN) {
            !!!cp (52);
-           $self->{current_token}->{first_start_tag}
-               = not defined $self->{last_emitted_start_tag_name};
            $self->{last_emitted_start_tag_name} = $self->{current_token}->{tag_name};
          } elsif ($self->{current_token}->{type} == END_TAG_TOKEN) {
            $self->{content_model} = PCDATA_CONTENT_MODEL; # MUST
-Line 819 
 sub _get_next_token ($) {
+Line 1356 
 sub _get_next_token ($) {
          } else {
            !!!cp (56);
          }
-         $self->{current_attribute} = {name => chr ($self->{next_char}),
+         $self->{current_attribute}
-                               value => ''};
+             = {name => chr ($self->{next_char}),
+                value => '',
+                line => $self->{line}, column => $self->{column}};
          $self->{state} = ATTRIBUTE_NAME_STATE;
          !!!next-input-character;
          redo A;
-Line 830 
 sub _get_next_token ($) {
+Line 1369 
 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});
+           !!!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 859 
 sub _get_next_token ($) {
+Line 1398 
 sub _get_next_token ($) {
          $before_leave->();
          if ($self->{current_token}->{type} == START_TAG_TOKEN) {
            !!!cp (61);
-           $self->{current_token}->{first_start_tag}
-               = not defined $self->{last_emitted_start_tag_name};
            $self->{last_emitted_start_tag_name} = $self->{current_token}->{tag_name};
          } elsif ($self->{current_token}->{type} == END_TAG_TOKEN) {
            !!!cp (62);
-Line 885 
 sub _get_next_token ($) {
+Line 1422 
 sub _get_next_token ($) {
          !!!next-input-character;
          redo A;
        } elsif ($self->{next_char} == 0x002F) { # /
+         !!!cp (64);
          $before_leave->();
+         $self->{state} = SELF_CLOSING_START_TAG_STATE;
          !!!next-input-character;
-         if ($self->{next_char} == 0x003E and # >
-             $self->{current_token}->{type} == START_TAG_TOKEN and
-             $permitted_slash_tag_name->{$self->{current_token}->{tag_name}}) {
-           # permitted slash
-           !!!cp (64);
-           #
-         } else {
-           !!!cp (65);
-           !!!parse-error (type => 'nestc');
-         }
-         $self->{state} = BEFORE_ATTRIBUTE_NAME_STATE;
-         # next-input-character is already done
          redo A;
        } elsif ($self->{next_char} == -1) {
          !!!parse-error (type => 'unclosed tag');
          $before_leave->();
          if ($self->{current_token}->{type} == START_TAG_TOKEN) {
            !!!cp (66);
-           $self->{current_token}->{first_start_tag}
-               = not defined $self->{last_emitted_start_tag_name};
            $self->{last_emitted_start_tag_name} = $self->{current_token}->{tag_name};
          } elsif ($self->{current_token}->{type} == END_TAG_TOKEN) {
            $self->{content_model} = PCDATA_CONTENT_MODEL; # MUST
-Line 957 
 sub _get_next_token ($) {
+Line 1482 
 sub _get_next_token ($) {
        } elsif ($self->{next_char} == 0x003E) { # >
          if ($self->{current_token}->{type} == START_TAG_TOKEN) {
            !!!cp (73);
-           $self->{current_token}->{first_start_tag}
-               = not defined $self->{last_emitted_start_tag_name};
            $self->{last_emitted_start_tag_name} = $self->{current_token}->{tag_name};
          } elsif ($self->{current_token}->{type} == END_TAG_TOKEN) {
            $self->{content_model} = PCDATA_CONTENT_MODEL; # MUST
-Line 981 
 sub _get_next_token ($) {
+Line 1504 
 sub _get_next_token ($) {
        } elsif (0x0041 <= $self->{next_char} and
                 $self->{next_char} <= 0x005A) { # A..Z
          !!!cp (76);
-         $self->{current_attribute} = {name => chr ($self->{next_char} + 0x0020),
+         $self->{current_attribute}
-                               value => ''};
+             = {name => chr ($self->{next_char} + 0x0020),
+                value => '',
+                line => $self->{line}, column => $self->{column}};
          $self->{state} = ATTRIBUTE_NAME_STATE;
          !!!next-input-character;
          redo A;
        } elsif ($self->{next_char} == 0x002F) { # /
+         !!!cp (77);
+         $self->{state} = SELF_CLOSING_START_TAG_STATE;
          !!!next-input-character;
-         if ($self->{next_char} == 0x003E and # >
-             $self->{current_token}->{type} == START_TAG_TOKEN and
-             $permitted_slash_tag_name->{$self->{current_token}->{tag_name}}) {
-           # permitted slash
-           !!!cp (77);
-           #
-         } else {
-           !!!cp (78);
-           !!!parse-error (type => 'nestc');
-           ## TODO: Different error type for <aa / bb> than <aa/>
-         }
-         $self->{state} = BEFORE_ATTRIBUTE_NAME_STATE;
-         # next-input-character is already done
          redo A;
        } elsif ($self->{next_char} == -1) {
          !!!parse-error (type => 'unclosed tag');
          if ($self->{current_token}->{type} == START_TAG_TOKEN) {
            !!!cp (79);
-           $self->{current_token}->{first_start_tag}
-               = not defined $self->{last_emitted_start_tag_name};
            $self->{last_emitted_start_tag_name} = $self->{current_token}->{tag_name};
          } elsif ($self->{current_token}->{type} == END_TAG_TOKEN) {
            $self->{content_model} = PCDATA_CONTENT_MODEL; # MUST
-Line 1029 
 sub _get_next_token ($) {
+Line 1541 
 sub _get_next_token ($) {
          redo A;
        } else {
          !!!cp (82);
-         $self->{current_attribute} = {name => chr ($self->{next_char}),
+         $self->{current_attribute}
-                               value => ''};
+             = {name => chr ($self->{next_char}),
+                value => '',
+                line => $self->{line}, column => $self->{column}};
          $self->{state} = ATTRIBUTE_NAME_STATE;
          !!!next-input-character;
          redo A;
-Line 1063 
 sub _get_next_token ($) {
+Line 1577 
 sub _get_next_token ($) {
        } elsif ($self->{next_char} == 0x003E) { # >
          if ($self->{current_token}->{type} == START_TAG_TOKEN) {
            !!!cp (87);
-           $self->{current_token}->{first_start_tag}
-               = not defined $self->{last_emitted_start_tag_name};
            $self->{last_emitted_start_tag_name} = $self->{current_token}->{tag_name};
          } elsif ($self->{current_token}->{type} == END_TAG_TOKEN) {
            $self->{content_model} = PCDATA_CONTENT_MODEL; # MUST
-Line 1088 
 sub _get_next_token ($) {
+Line 1600 
 sub _get_next_token ($) {
          !!!parse-error (type => 'unclosed tag');
          if ($self->{current_token}->{type} == START_TAG_TOKEN) {
            !!!cp (90);
-           $self->{current_token}->{first_start_tag}
-               = not defined $self->{last_emitted_start_tag_name};
            $self->{last_emitted_start_tag_name} = $self->{current_token}->{tag_name};
          } elsif ($self->{current_token}->{type} == END_TAG_TOKEN) {
            $self->{content_model} = PCDATA_CONTENT_MODEL; # MUST
-Line 1137 
 sub _get_next_token ($) {
+Line 1647 
 sub _get_next_token ($) {
          !!!parse-error (type => 'unclosed attribute value');
          if ($self->{current_token}->{type} == START_TAG_TOKEN) {
            !!!cp (97);
-           $self->{current_token}->{first_start_tag}
-               = not defined $self->{last_emitted_start_tag_name};
            $self->{last_emitted_start_tag_name} = $self->{current_token}->{tag_name};
          } elsif ($self->{current_token}->{type} == END_TAG_TOKEN) {
            $self->{content_model} = PCDATA_CONTENT_MODEL; # MUST
-Line 1181 
 sub _get_next_token ($) {
+Line 1689 
 sub _get_next_token ($) {
          !!!parse-error (type => 'unclosed attribute value');
          if ($self->{current_token}->{type} == START_TAG_TOKEN) {
            !!!cp (103);
-           $self->{current_token}->{first_start_tag}
-               = not defined $self->{last_emitted_start_tag_name};
            $self->{last_emitted_start_tag_name} = $self->{current_token}->{tag_name};
          } elsif ($self->{current_token}->{type} == END_TAG_TOKEN) {
            $self->{content_model} = PCDATA_CONTENT_MODEL; # MUST
-Line 1228 
 sub _get_next_token ($) {
+Line 1734 
 sub _get_next_token ($) {
        } elsif ($self->{next_char} == 0x003E) { # >
          if ($self->{current_token}->{type} == START_TAG_TOKEN) {
            !!!cp (109);
-           $self->{current_token}->{first_start_tag}
-               = not defined $self->{last_emitted_start_tag_name};
            $self->{last_emitted_start_tag_name} = $self->{current_token}->{tag_name};
          } elsif ($self->{current_token}->{type} == END_TAG_TOKEN) {
            $self->{content_model} = PCDATA_CONTENT_MODEL; # MUST
-Line 1253 
 sub _get_next_token ($) {
+Line 1757 
 sub _get_next_token ($) {
          !!!parse-error (type => 'unclosed tag');
          if ($self->{current_token}->{type} == START_TAG_TOKEN) {
            !!!cp (112);
-           $self->{current_token}->{first_start_tag}
-               = not defined $self->{last_emitted_start_tag_name};
            $self->{last_emitted_start_tag_name} = $self->{current_token}->{tag_name};
          } elsif ($self->{current_token}->{type} == END_TAG_TOKEN) {
            $self->{content_model} = PCDATA_CONTENT_MODEL; # MUST
-Line 1325 
 sub _get_next_token ($) {
+Line 1827 
 sub _get_next_token ($) {
        } elsif ($self->{next_char} == 0x003E) { # >
          if ($self->{current_token}->{type} == START_TAG_TOKEN) {
            !!!cp (119);
-           $self->{current_token}->{first_start_tag}
-               = not defined $self->{last_emitted_start_tag_name};
            $self->{last_emitted_start_tag_name} = $self->{current_token}->{tag_name};
          } elsif ($self->{current_token}->{type} == END_TAG_TOKEN) {
            $self->{content_model} = PCDATA_CONTENT_MODEL; # MUST
-Line 1347 
 sub _get_next_token ($) {
+Line 1847 
 sub _get_next_token ($) {
          redo A;
        } elsif ($self->{next_char} == 0x002F) { # /
+         !!!cp (122);
+         $self->{state} = SELF_CLOSING_START_TAG_STATE;
          !!!next-input-character;
-         if ($self->{next_char} == 0x003E and # >
+         redo A;
-             $self->{current_token}->{type} == START_TAG_TOKEN and
+       } elsif ($self->{next_char} == -1) {
-             $permitted_slash_tag_name->{$self->{current_token}->{tag_name}}) {
+         !!!parse-error (type => 'unclosed tag');
-           # permitted slash
+         if ($self->{current_token}->{type} == START_TAG_TOKEN) {
-           !!!cp (122);
+           !!!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 {
-           !!!cp (123);
+           die "$0: $self->{current_token}->{type}: Unknown token type";
-           !!!parse-error (type => 'nestc');
          }
-         $self->{state} = BEFORE_ATTRIBUTE_NAME_STATE;
+         $self->{state} = DATA_STATE;
-         # next-input-character is already done
+         ## Reconsume.
+         !!!emit ($self->{current_token}); # start tag or end tag
          redo A;
        } else {
-         !!!cp (124);
+         !!!cp ('124.1');
          !!!parse-error (type => 'no space between attributes');
          $self->{state} = BEFORE_ATTRIBUTE_NAME_STATE;
          ## reconsume
          redo A;
        }
+     } elsif ($self->{state} == SELF_CLOSING_START_TAG_STATE) {
+       if ($self->{next_char} == 0x003E) { # >
+         if ($self->{current_token}->{type} == END_TAG_TOKEN) {
+           !!!cp ('124.2');
+           !!!parse-error (type => 'nestc', token => $self->{current_token});
+           ## TODO: Different type than slash in start tag
+           $self->{content_model} = PCDATA_CONTENT_MODEL; # MUST
+           if ($self->{current_token}->{attributes}) {
+             !!!cp ('124.4');
+             !!!parse-error (type => 'end tag attribute');
+           } else {
+             !!!cp ('124.5');
+           }
+           ## TODO: Test |<title></title/>|
+         } else {
+           !!!cp ('124.3');
+           $self->{self_closing} = 1;
+         }
+         $self->{state} = DATA_STATE;
+         !!!next-input-character;
+         !!!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');
+         ## TODO: This error type is wrong.
+         $self->{state} = BEFORE_ATTRIBUTE_NAME_STATE;
+         ## Reconsume.
+         redo A;
+       }
      } elsif ($self->{state} == BOGUS_COMMENT_STATE) {
        ## (only happen if PCDATA state)
-       my $token = {type => COMMENT_TOKEN, data => ''};
+       ## NOTE: Set by the previous state
+       #my $token = {type => COMMENT_TOKEN, data => ''};
        BC: {
          if ($self->{next_char} == 0x003E) { # >
-Line 1379 
 sub _get_next_token ($) {
+Line 1943 
 sub _get_next_token ($) {
            $self->{state} = DATA_STATE;
            !!!next-input-character;
-           !!!emit ($token);
+           !!!emit ($self->{current_token}); # comment
            redo A;
          } elsif ($self->{next_char} == -1) {
-Line 1387 
 sub _get_next_token ($) {
+Line 1951 
 sub _get_next_token ($) {
            $self->{state} = DATA_STATE;
            ## reconsume
-           !!!emit ($token);
+           !!!emit ($self->{current_token}); # comment
            redo A;
          } else {
            !!!cp (126);
-           $token->{data} .= chr ($self->{next_char});
+           $self->{current_token}->{data} .= chr ($self->{next_char}); # comment
            !!!next-input-character;
            redo BC;
          }
-Line 1402 
 sub _get_next_token ($) {
+Line 1966 
 sub _get_next_token ($) {
      } 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};
-Line 1410 
 sub _get_next_token ($) {
+Line 1976 
 sub _get_next_token ($) {
          push @next_char, $self->{next_char};
          if ($self->{next_char} == 0x002D) { # -
            !!!cp (127);
-           $self->{current_token} = {type => COMMENT_TOKEN, data => ''};
+           $self->{current_token} = {type => COMMENT_TOKEN, data => '',
+                                     line => $l, column => $c,
+                                    };
            $self->{state} = COMMENT_START_STATE;
            !!!next-input-character;
            redo A;
-Line 1446 
 sub _get_next_token ($) {
+Line 2014 
 sub _get_next_token ($) {
                      !!!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 {
-Line 1466 
 sub _get_next_token ($) {
+Line 2038 
 sub _get_next_token ($) {
          } else {
            !!!cp (135);
          }
+       } elsif ($self->{insertion_mode} & IN_FOREIGN_CONTENT_IM and
+                $self->{open_elements}->[-1]->[1] & FOREIGN_EL and
+                $self->{next_char} == 0x005B) { # [
+         !!!next-input-character;
+         push @next_char, $self->{next_char};
+         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);
        }
-Line 1474 
 sub _get_next_token ($) {
+Line 2089 
 sub _get_next_token ($) {
        $self->{next_char} = shift @next_char;
        !!!back-next-input-character (@next_char);
        $self->{state} = BOGUS_COMMENT_STATE;
+       $self->{current_token} = {type => COMMENT_TOKEN, data => '',
+                                 line => $l, column => $c,
+                                };
        redo A;
        ## ISSUE: typos in spec: chacacters, is is a parse error
-Line 1597 
 sub _get_next_token ($) {
+Line 2215 
 sub _get_next_token ($) {
          redo A;
        } elsif ($self->{next_char} == 0x002D) { # -
          !!!cp (152);
-         !!!parse-error (type => 'dash in comment');
+         !!!parse-error (type => 'dash in comment',
+                         line => $self->{line_prev},
+                         column => $self->{column_prev});
          $self->{current_token}->{data} .= '-'; # comment
          ## Stay in the state
          !!!next-input-character;
-Line 1613 
 sub _get_next_token ($) {
+Line 2233 
 sub _get_next_token ($) {
          redo A;
        } else {
          !!!cp (154);
-         !!!parse-error (type => 'dash in comment');
+         !!!parse-error (type => 'dash in comment',
+                         line => $self->{line_prev},
+                         column => $self->{column_prev});
          $self->{current_token}->{data} .= '--' . chr ($self->{next_char}); # comment
          $self->{state} = COMMENT_STATE;
          !!!next-input-character;
-Line 1652 
 sub _get_next_token ($) {
+Line 2274 
 sub _get_next_token ($) {
          $self->{state} = DATA_STATE;
          !!!next-input-character;
-         !!!emit ({type => DOCTYPE_TOKEN, quirks => 1});
+         !!!emit ($self->{current_token}); # DOCTYPE (quirks)
          redo A;
        } elsif ($self->{next_char} == -1) {
-Line 1661 
 sub _get_next_token ($) {
+Line 2283 
 sub _get_next_token ($) {
          $self->{state} = DATA_STATE;
          ## reconsume
-         !!!emit ({type => DOCTYPE_TOKEN, quirks => 1});
+         !!!emit ($self->{current_token}); # DOCTYPE (quirks)
          redo A;
        } else {
          !!!cp (160);
-         $self->{current_token}
+         $self->{current_token}->{name} = chr $self->{next_char};
-             = {type => DOCTYPE_TOKEN,
+         delete $self->{current_token}->{quirks};
-                name => chr ($self->{next_char}),
-                #quirks => 0,
-               };
  ## ISSUE: "Set the token's name name to the" in the spec
          $self->{state} = DOCTYPE_NAME_STATE;
          !!!next-input-character;
-Line 2061 
 sub _get_next_token ($) {
+Line 2680 
 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 2097 
 sub _get_next_token ($) {
+Line 2716 
 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 2145 
 sub _get_next_token ($) {
+Line 2764 
 sub _get_next_token ($) {
        } elsif ($self->{next_char} == -1) {
          !!!cp (217);
          !!!parse-error (type => 'unclosed DOCTYPE');
          $self->{state} = DATA_STATE;
          ## reconsume
-Line 2186 
 sub _get_next_token ($) {
+Line 2804 
 sub _get_next_token ($) {
          !!!next-input-character;
          redo A;
        }
+     } elsif ($self->{state} == CDATA_BLOCK_STATE) {
+       my $s = '';
+       my ($l, $c) = ($self->{line}, $self->{column});
+       CS: while ($self->{next_char} != -1) {
+         if ($self->{next_char} == 0x005D) { # ]
+           !!!next-input-character;
+           if ($self->{next_char} == 0x005D) { # ]
+             !!!next-input-character;
+             MDC: {
+               if ($self->{next_char} == 0x003E) { # >
+                 !!!cp (221.1);
+                 !!!next-input-character;
+                 last CS;
+               } elsif ($self->{next_char} == 0x005D) { # ]
+                 !!!cp (221.2);
+                 $s .= ']';
+                 !!!next-input-character;
+                 redo MDC;
+               } else {
+                 !!!cp (221.3);
+                 $s .= ']]';
+                 #
+               }
+             } # MDC
+           } else {
+             !!!cp (221.4);
+             $s .= ']';
+             #
+           }
+         } else {
+           !!!cp (221.5);
+           #
+         }
+         $s .= chr $self->{next_char};
+         !!!next-input-character;
+       } # CS
+       $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);
+       }
+       redo A;
+       ## ISSUE: "text tokens" in spec.
+       ## TODO: Streaming support
      } else {
        die "$0: $self->{state}: Unknown state";
      }
-Line 2197 
 sub _get_next_token ($) {
+Line 2869 
 sub _get_next_token ($) {
  sub _tokenize_attempt_to_consume_an_entity ($$$) {
    my ($self, $in_attr, $additional) = @_;
+   my ($l, $c) = ($self->{line_prev}, $self->{column_prev});
    if ({
 x0009 => 1, 0x000A => 1, 0x000B => 1, 0x000C => 1, # HT, LF, VT, FF,
 x0020 => 1, 0x003C => 1, 0x0026 => 1, -1 => 1, # SP, <, & # 0x000D # CR
-Line 2237 
 sub _tokenize_attempt_to_consume_an_enti
+Line 2911 
 sub _tokenize_attempt_to_consume_an_enti
            redo X;
          } elsif (not defined $code) { # no hexadecimal digit
            !!!cp (1005);
-           !!!parse-error (type => 'bare hcro');
+           !!!parse-error (type => 'bare hcro', line => $l, column => $c);
            !!!back-next-input-character ($x_char, $self->{next_char});
            $self->{next_char} = 0x0023; # #
            return undef;
-Line 2246 
 sub _tokenize_attempt_to_consume_an_enti
+Line 2920 
 sub _tokenize_attempt_to_consume_an_enti
            !!!next-input-character;
          } else {
            !!!cp (1007);
-           !!!parse-error (type => 'no refc');
+           !!!parse-error (type => 'no refc', line => $l, column => $c);
          }
          if ($code == 0 or (0xD800 <= $code and $code <= 0xDFFF)) {
            !!!cp (1008);
-           !!!parse-error (type => sprintf 'invalid character reference:U+%04X', $code);
+           !!!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 => sprintf 'invalid character reference:U-%08X', $code);
+           !!!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');
+           !!!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);
+           !!!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};
+                 has_reference => 1,
+                 line => $l, column => $c,
+                };
        } # X
      } elsif (0x0030 <= $self->{next_char} and
               $self->{next_char} <= 0x0039) { # 0..9
-Line 2289 
 sub _tokenize_attempt_to_consume_an_enti
+Line 2969 
 sub _tokenize_attempt_to_consume_an_enti
          !!!next-input-character;
        } else {
          !!!cp (1014);
-         !!!parse-error (type => 'no refc');
+         !!!parse-error (type => 'no refc', line => $l, column => $c);
        }
        if ($code == 0 or (0xD800 <= $code and $code <= 0xDFFF)) {
          !!!cp (1015);
-         !!!parse-error (type => sprintf 'invalid character reference:U+%04X', $code);
+         !!!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);
+         !!!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');
+         !!!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);
+         !!!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};
+       return {type => CHARACTER_TOKEN, data => chr $code, has_reference => 1,
+               line => $l, column => $c,
+              };
      } else {
        !!!cp (1019);
-       !!!parse-error (type => 'bare nero');
+       !!!parse-error (type => 'bare nero', line => $l, column => $c);
        !!!back-next-input-character ($self->{next_char});
        $self->{next_char} = 0x0023; # #
        return undef;
-Line 2330 
 sub _tokenize_attempt_to_consume_an_enti
+Line 3019 
 sub _tokenize_attempt_to_consume_an_enti
      require Whatpm::_NamedEntityList;
      our $EntityChar;
-     while (length $entity_name < 10 and
+     while (length $entity_name < 30 and
             ## NOTE: Some number greater than the maximum length of entity name
             ((0x0041 <= $self->{next_char} and # a
               $self->{next_char} <= 0x005A) or # x
-Line 2363 
 sub _tokenize_attempt_to_consume_an_enti
+Line 3052 
 sub _tokenize_attempt_to_consume_an_enti
      if ($match > 0) {
        !!!cp (1023);
-       return {type => CHARACTER_TOKEN, data => $value, has_reference => 1};
+       return {type => CHARACTER_TOKEN, data => $value, has_reference => 1,
+               line => $l, column => $c,
+              };
      } elsif ($match < 0) {
-       !!!parse-error (type => 'no refc');
+       !!!parse-error (type => 'no refc', line => $l, column => $c);
        if ($in_attr and $match < -1) {
          !!!cp (1024);
-         return {type => CHARACTER_TOKEN, data => '&'.$entity_name};
+         return {type => CHARACTER_TOKEN, data => '&'.$entity_name,
+                 line => $l, column => $c,
+                };
        } else {
          !!!cp (1025);
-         return {type => CHARACTER_TOKEN, data => $value, has_reference => 1};
+         return {type => CHARACTER_TOKEN, data => $value, has_reference => 1,
+                 line => $l, column => $c,
+                };
        }
      } else {
        !!!cp (1026);
-       !!!parse-error (type => 'bare ero');
+       !!!parse-error (type => 'bare ero', line => $l, column => $c);
        ## NOTE: "No characters are consumed" in the spec.
-       return {type => CHARACTER_TOKEN, data => '&'.$value};
+       return {type => CHARACTER_TOKEN, data => '&'.$value,
+               line => $l, column => $c,
+              };
      }
    } else {
      !!!cp (1027);
      ## no characters are consumed
-     !!!parse-error (type => 'bare ero');
+     !!!parse-error (type => 'bare ero', line => $l, column => $c);
      return undef;
    }
  } # _tokenize_attempt_to_consume_an_entity
-Line 2420 
 sub _construct_tree ($) {
+Line 3117 
 sub _construct_tree ($) {
    !!!next-token;
-   $self->{insertion_mode} = BEFORE_HEAD_IM;
    undef $self->{form_element};
    undef $self->{head_element};
    $self->{open_elements} = [];
    undef $self->{inner_html_node};
+   ## NOTE: The "initial" insertion mode.
    $self->_tree_construction_initial; # MUST
+   ## NOTE: The "before html" insertion mode.
    $self->_tree_construction_root_element;
+   $self->{insertion_mode} = BEFORE_HEAD_IM;
+   ## NOTE: The "before head" insertion mode and so on.
    $self->_tree_construction_main;
  } # _construct_tree
  sub _tree_construction_initial ($) {
    my $self = shift;
+   ## NOTE: "initial" insertion mode
    INITIAL: {
      if ($token->{type} == DOCTYPE_TOKEN) {
        ## NOTE: Conformance checkers MAY, instead of reporting "not HTML5"
-Line 2444 
 sub _tree_construction_initial ($) {
+Line 3149 
 sub _tree_construction_initial ($) {
        if (not defined $token->{name} or # <!DOCTYPE>
            defined $token->{public_identifier} or
            defined $token->{system_identifier}) {
-         !!!parse-error (type => 'not HTML5');
+         !!!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');
+         !!!parse-error (type => 'not HTML5', token => $token);
+       } else {
+         !!!cp ('t3');
        }
        my $doctype = $self->{document}->create_document_type_definition
          ($token->{name}); ## ISSUE: If name is missing (e.g. <!DOCTYPE>)?
+       ## NOTE: Default value for both |public_id| and |system_id| attributes
+       ## are empty strings, so that we don't set any value in missing cases.
        $doctype->public_id ($token->{public_identifier})
            if defined $token->{public_identifier};
        $doctype->system_id ($token->{system_identifier})
-Line 2461 
 sub _tree_construction_initial ($) {
+Line 3172 
 sub _tree_construction_initial ($) {
        $self->{document}->append_child ($doctype);
        if ($token->{quirks} or $doctype_name ne 'HTML') {
+         !!!cp ('t4');
          $self->{document}->manakai_compat_mode ('quirks');
        } 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,
+             $pubid eq "HTML") {
-           "-//W3O//DTD W3 HTML STRICT 3.0//EN//" => 1,
+           !!!cp ('t5');
-           "-//WEBTECHS//DTD MOZILLA HTML 2.0//EN" => 1,
-           "-//WEBTECHS//DTD MOZILLA HTML//EN" => 1,
-           "-/W3C/DTD HTML 4.0 TRANSITIONAL/EN" => 1,
-           "HTML" => 1,
-         }->{$pubid}) {
            $self->{document}->manakai_compat_mode ('quirks');
-         } elsif ($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');
            } else {
+             !!!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 {
+           !!!cp ('t9');
          }
+       } else {
+         !!!cp ('t10');
        }
        if (defined $token->{system_identifier}) {
          my $sysid = $token->{system_identifier};
          $sysid =~ tr/A-Z/a-z/;
          if ($sysid eq "http://www.ibm.com/data/dtd/v11/ibmxhtml1-transitional.dtd") {
+           ## NOTE: Ensure that |PUBLIC "(limited quirks)" "(quirks)"| is
+           ## marked as quirks.
            $self->{document}->manakai_compat_mode ('quirks');
+           !!!cp ('t11');
+         } else {
+           !!!cp ('t12');
          }
+       } else {
+         !!!cp ('t13');
        }
-       ## Go to the root element phase.
+       ## Go to the "before html" insertion mode.
        !!!next-token;
        return;
      } elsif ({
-Line 2568 
 sub _tree_construction_initial ($) {
+Line 3289 
 sub _tree_construction_initial ($) {
                END_TAG_TOKEN, 1,
                END_OF_FILE_TOKEN, 1,
               }->{$token->{type}}) {
-       !!!parse-error (type => 'no DOCTYPE');
+       !!!cp ('t14');
+       !!!parse-error (type => 'no DOCTYPE', token => $token);
        $self->{document}->manakai_compat_mode ('quirks');
-       ## Go to the root element phase
+       ## Go to the "before html" insertion mode.
        ## reprocess
+       !!!ack-later;
        return;
      } elsif ($token->{type} == CHARACTER_TOKEN) {
        if ($token->{data} =~ s/^([\x09\x0A\x0B\x0C\x20]+)//) { # \x0D
          ## Ignore the token
          unless (length $token->{data}) {
-           ## Stay in the phase
+           !!!cp ('t15');
+           ## Stay in the insertion mode.
            !!!next-token;
            redo INITIAL;
+         } else {
+           !!!cp ('t16');
          }
+       } else {
+         !!!cp ('t17');
        }
-       !!!parse-error (type => 'no DOCTYPE');
+       !!!parse-error (type => 'no DOCTYPE', token => $token);
        $self->{document}->manakai_compat_mode ('quirks');
-       ## Go to the root element phase
+       ## Go to the "before html" insertion mode.
        ## reprocess
        return;
      } elsif ($token->{type} == COMMENT_TOKEN) {
+       !!!cp ('t18');
        my $comment = $self->{document}->create_comment ($token->{data});
        $self->{document}->append_child ($comment);
-       ## Stay in the phase.
+       ## Stay in the insertion mode.
        !!!next-token;
        redo INITIAL;
      } else {
        die "$0: $token->{type}: Unknown token type";
      }
    } # INITIAL
+   die "$0: _tree_construction_initial: This should be never reached";
  } # _tree_construction_initial
  sub _tree_construction_root_element ($) {
    my $self = shift;
+   ## NOTE: "before html" insertion mode.
    B: {
        if ($token->{type} == DOCTYPE_TOKEN) {
-         !!!parse-error (type => 'in html:#DOCTYPE');
+         !!!cp ('t19');
+         !!!parse-error (type => 'in html:#DOCTYPE', token => $token);
          ## Ignore the token
-         ## Stay in the phase
+         ## Stay in the insertion mode.
          !!!next-token;
          redo B;
        } elsif ($token->{type} == COMMENT_TOKEN) {
+         !!!cp ('t20');
          my $comment = $self->{document}->create_comment ($token->{data});
          $self->{document}->append_child ($comment);
-         ## Stay in the phase
+         ## Stay in the insertion mode.
          !!!next-token;
          redo B;
        } elsif ($token->{type} == CHARACTER_TOKEN) {
-Line 2623 
 sub _tree_construction_root_element ($)
+Line 3358 
 sub _tree_construction_root_element ($)
            ## Ignore the token.
            unless (length $token->{data}) {
-             ## Stay in the phase
+             !!!cp ('t21');
+             ## Stay in the insertion mode.
              !!!next-token;
              redo B;
+           } else {
+             !!!cp ('t22');
            }
+         } else {
+           !!!cp ('t23');
          }
          $self->{application_cache_selection}->(undef);
          #
        } elsif ($token->{type} == START_TAG_TOKEN) {
-         if ($token->{tag_name} eq 'html' and
+         if ($token->{tag_name} eq 'html') {
-             $token->{attributes}->{manifest}) {
+           my $root_element;
-           $self->{application_cache_selection}
+           !!!create-element ($root_element, $HTML_NS, $token->{tag_name}, $token->{attributes}, $token);
-                ->($token->{attributes}->{manifest}->{value});
+           $self->{document}->append_child ($root_element);
-           ## ISSUE: No relative reference resolution?
+           push @{$self->{open_elements}},
+               [$root_element, $el_category->{html}];
+           if ($token->{attributes}->{manifest}) {
+             !!!cp ('t24');
+             $self->{application_cache_selection}
+                 ->($token->{attributes}->{manifest}->{value});
+             ## ISSUE: Spec is unclear on relative references.
+             ## According to Hixie (#whatwg 2008-03-19), it should be
+             ## resolved against the base URI of the document in HTML
+             ## or xml:base of the element in XHTML.
+           } else {
+             !!!cp ('t25');
+             $self->{application_cache_selection}->(undef);
+           }
+           !!!nack ('t25c');
+           !!!next-token;
+           return; ## Go to the "before head" insertion mode.
          } else {
-           $self->{application_cache_selection}->(undef);
+           !!!cp ('t25.1');
+           #
          }
-         ## ISSUE: There is an issue in the spec
-         #
        } elsif ({
                  END_TAG_TOKEN, 1,
                  END_OF_FILE_TOKEN, 1,
                 }->{$token->{type}}) {
-         $self->{application_cache_selection}->(undef);
+         !!!cp ('t26');
-         ## ISSUE: There is an issue in the spec
          #
        } else {
          die "$0: $token->{type}: Unknown token type";
        }
-       my $root_element; !!!create-element ($root_element, 'html');
+     my $root_element;
-       $self->{document}->append_child ($root_element);
+     !!!create-element ($root_element, $HTML_NS, 'html',, $token);
-       push @{$self->{open_elements}}, [$root_element, 'html'];
+     $self->{document}->append_child ($root_element);
-       ## reprocess
+     push @{$self->{open_elements}}, [$root_element, $el_category->{html}];
-       #redo B;
-       return; ## Go to the main phase.
+     $self->{application_cache_selection}->(undef);
+     ## NOTE: Reprocess the token.
+     !!!ack-later;
+     return; ## Go to the "before head" insertion mode.
+     ## ISSUE: There is an issue in the spec
    } # B
+   die "$0: _tree_construction_root_element: This should never be reached";
  } # _tree_construction_root_element
  sub _reset_insertion_mode ($) {
-Line 2677 
 sub _reset_insertion_mode ($) {
+Line 3440 
 sub _reset_insertion_mode ($) {
      ## Step 3
      S3: {
-       ## ISSUE: Oops! "If node is the first node in the stack of open
-       ## elements, then set last to true. If the context element of the
-       ## HTML fragment parsing algorithm is neither a td element nor a
-       ## th element, then set node to the context element. (fragment case)":
-       ## The second "if" is in the scope of the first "if"!?
        if ($self->{open_elements}->[0]->[0] eq $node->[0]) {
          $last = 1;
          if (defined $self->{inner_html_node}) {
-           if ($self->{inner_html_node}->[1] eq 'td' or
+           !!!cp ('t28');
-               $self->{inner_html_node}->[1] eq 'th') {
+           $node = $self->{inner_html_node};
-             #
+         } else {
-           } else {
+           die "_reset_insertion_mode: t27";
-             $node = $self->{inner_html_node};
-           }
          }
        }
-       ## Step 4..13
+       ## Step 4..14
-       my $new_mode = {
+       my $new_mode;
+       if ($node->[1] & FOREIGN_EL) {
+         !!!cp ('t28.1');
+         ## NOTE: Strictly spaking, the line below only applies to MathML and
+         ## SVG elements.  Currently the HTML syntax supports only MathML and
+         ## SVG elements as foreigners.
+         $new_mode = IN_BODY_IM | IN_FOREIGN_CONTENT_IM;
+       } elsif ($node->[1] & TABLE_CELL_EL) {
+         if ($last) {
+           !!!cp ('t28.2');
+           #
+         } else {
+           !!!cp ('t28.3');
+           $new_mode = IN_CELL_IM;
+         }
+       } else {
+         !!!cp ('t28.4');
+         $new_mode = {
                        select => IN_SELECT_IM,
-                       td => IN_CELL_IM,
+                       ## NOTE: |option| and |optgroup| do not set
-                       th => IN_CELL_IM,
+                       ## insertion mode to "in select" by themselves.
                        tr => IN_ROW_IM,
                        tbody => IN_TABLE_BODY_IM,
                        thead => IN_TABLE_BODY_IM,
-Line 2709 
 sub _reset_insertion_mode ($) {
+Line 3482 
 sub _reset_insertion_mode ($) {
                        head => IN_BODY_IM, # not in head!
                        body => IN_BODY_IM,
                        frameset => IN_FRAMESET_IM,
-                      }->{$node->[1]};
+                      }->{$node->[0]->manakai_local_name};
+       }
        $self->{insertion_mode} = $new_mode and return if defined $new_mode;
-       ## Step 14
+       ## Step 15
-       if ($node->[1] eq 'html') {
+       if ($node->[1] & HTML_EL) {
          unless (defined $self->{head_element}) {
+           !!!cp ('t29');
            $self->{insertion_mode} = BEFORE_HEAD_IM;
          } else {
+           ## ISSUE: Can this state be reached?
+           !!!cp ('t30');
            $self->{insertion_mode} = AFTER_HEAD_IM;
          }
          return;
+       } else {
+         !!!cp ('t31');
        }
-       ## Step 15
+       ## Step 16
        $self->{insertion_mode} = IN_BODY_IM and return if $last;
-       ## Step 16
+       ## Step 17
        $i--;
        $node = $self->{open_elements}->[$i];
-       ## Step 17
+       ## Step 18
        redo S3;
      } # S3
+   die "$0: _reset_insertion_mode: This line should never be reached";
  } # _reset_insertion_mode
  sub _tree_construction_main ($) {
-Line 2753 
 sub _tree_construction_main ($) {
+Line 3534 
 sub _tree_construction_main ($) {
      return if $entry->[0] eq '#marker';
      for (@{$self->{open_elements}}) {
        if ($entry->[0] eq $_->[0]) {
+         !!!cp ('t32');
          return;
        }
      }
-Line 2767 
 sub _tree_construction_main ($) {
+Line 3549 
 sub _tree_construction_main ($) {
        ## Step 6
        if ($entry->[0] eq '#marker') {
+         !!!cp ('t33_1');
          #
        } else {
          my $in_open_elements;
          OE: for (@{$self->{open_elements}}) {
            if ($entry->[0] eq $_->[0]) {
+             !!!cp ('t33');
              $in_open_elements = 1;
              last OE;
            }
          }
          if ($in_open_elements) {
+           !!!cp ('t34');
            #
          } else {
+           ## NOTE: <!DOCTYPE HTML><p><b><i><u></p> <p>X
+           !!!cp ('t35');
            redo S4;
          }
        }
-Line 2801 
 sub _tree_construction_main ($) {
+Line 3588 
 sub _tree_construction_main ($) {
        ## Step 11
        unless ($clone->[0] eq $active_formatting_elements->[-1]->[0]) {
+         !!!cp ('t36');
          ## Step 7'
          $i++;
          $entry = $active_formatting_elements->[$i];
          redo S7;
        }
+       !!!cp ('t37');
      } # S7
    }; # $reconstruct_active_formatting_elements
    my $clear_up_to_marker = sub {
      for (reverse 0..$#$active_formatting_elements) {
        if ($active_formatting_elements->[$_]->[0] eq '#marker') {
+         !!!cp ('t38');
          splice @$active_formatting_elements, $_;
          return;
        }
      }
+     !!!cp ('t39');
    }; # $clear_up_to_marker
-   my $parse_rcdata = sub ($$) {
+   my $insert;
-     my ($content_model_flag, $insert) = @_;
+   my $parse_rcdata = sub ($) {
+     my ($content_model_flag) = @_;
      ## Step 1
      my $start_tag_name = $token->{tag_name};
      my $el;
-     !!!create-element ($el, $start_tag_name, $token->{attributes});
+     !!!create-element ($el, $HTML_NS, $start_tag_name, $token->{attributes}, $token);
      ## Step 2
-     $insert->($el); # /context node/->append_child ($el)
+     $insert->($el);
      ## Step 3
      $self->{content_model} = $content_model_flag; # CDATA or RCDATA
-Line 2836 
 sub _tree_construction_main ($) {
+Line 3631 
 sub _tree_construction_main ($) {
      ## Step 4
      my $text = '';
+     !!!nack ('t40.1');
      !!!next-token;
      while ($token->{type} == CHARACTER_TOKEN) { # or until stop tokenizing
+       !!!cp ('t40');
        $text .= $token->{data};
        !!!next-token;
      }
      ## Step 5
      if (length $text) {
+       !!!cp ('t41');
        my $text = $self->{document}->create_text_node ($text);
        $el->append_child ($text);
      }
-Line 2852 
 sub _tree_construction_main ($) {
+Line 3650 
 sub _tree_construction_main ($) {
      $self->{content_model} = PCDATA_CONTENT_MODEL;
      ## Step 7
-     if ($token->{type} == END_TAG_TOKEN and $token->{tag_name} eq $start_tag_name) {
+     if ($token->{type} == END_TAG_TOKEN and
+         $token->{tag_name} eq $start_tag_name) {
+       !!!cp ('t42');
        ## Ignore the token
-     } elsif ($content_model_flag == CDATA_CONTENT_MODEL) {
-       !!!parse-error (type => 'in CDATA:#'.$token->{type});
-     } elsif ($content_model_flag == RCDATA_CONTENT_MODEL) {
-       !!!parse-error (type => 'in RCDATA:#'.$token->{type});
      } else {
-       die "$0: $content_model_flag in parse_rcdata";
+       ## NOTE: An end-of-file token.
+       if ($content_model_flag == CDATA_CONTENT_MODEL) {
+         !!!cp ('t43');
+         !!!parse-error (type => 'in CDATA:#eof', token => $token);
+       } elsif ($content_model_flag == RCDATA_CONTENT_MODEL) {
+         !!!cp ('t44');
+         !!!parse-error (type => 'in RCDATA:#eof', token => $token);
+       } else {
+         die "$0: $content_model_flag in parse_rcdata";
+       }
      }
      !!!next-token;
    }; # $parse_rcdata
-   my $script_start_tag = sub ($) {
+   my $script_start_tag = sub () {
-     my $insert = $_[0];
      my $script_el;
-     !!!create-element ($script_el, 'script', $token->{attributes});
+     !!!create-element ($script_el, $HTML_NS, 'script', $token->{attributes}, $token);
      ## TODO: mark as "parser-inserted"
      $self->{content_model} = CDATA_CONTENT_MODEL;
      delete $self->{escape}; # MUST
      my $text = '';
+     !!!nack ('t45.1');
      !!!next-token;
      while ($token->{type} == CHARACTER_TOKEN) {
+       !!!cp ('t45');
        $text .= $token->{data};
        !!!next-token;
      } # stop if non-character token or tokenizer stops tokenising
      if (length $text) {
+       !!!cp ('t46');
        $script_el->manakai_append_text ($text);
      }
-Line 2887 
 sub _tree_construction_main ($) {
+Line 3694 
 sub _tree_construction_main ($) {
      if ($token->{type} == END_TAG_TOKEN and
          $token->{tag_name} eq 'script') {
+       !!!cp ('t47');
        ## Ignore the token
      } else {
-       !!!parse-error (type => 'in CDATA:#'.$token->{type});
+       !!!cp ('t48');
+       !!!parse-error (type => 'in CDATA:#eof', token => $token);
        ## ISSUE: And ignore?
        ## TODO: mark as "already executed"
      }
      if (defined $self->{inner_html_node}) {
+       !!!cp ('t49');
        ## TODO: mark as "already executed"
      } else {
+       !!!cp ('t50');
        ## TODO: $old_insertion_point = current insertion point
        ## TODO: insertion point = just before the next input character
-Line 2910 
 sub _tree_construction_main ($) {
+Line 3721 
 sub _tree_construction_main ($) {
      !!!next-token;
    }; # $script_start_tag
+   ## NOTE: $open_tables->[-1]->[0] is the "current table" element node.
+   ## NOTE: $open_tables->[-1]->[1] is the "tainted" flag.
+   my $open_tables = [[$self->{open_elements}->[0]->[0]]];
    my $formatting_end_tag = sub {
-     my $tag_name = shift;
+     my $end_tag_token = shift;
+     my $tag_name = $end_tag_token->{tag_name};
+     ## NOTE: The adoption agency algorithm (AAA).
      FET: {
        ## Step 1
        my $formatting_element;
        my $formatting_element_i_in_active;
        AFE: for (reverse 0..$#$active_formatting_elements) {
-         if ($active_formatting_elements->[$_]->[1] eq $tag_name) {
+         if ($active_formatting_elements->[$_]->[0] eq '#marker') {
+           !!!cp ('t52');
+           last AFE;
+         } elsif ($active_formatting_elements->[$_]->[0]->manakai_local_name
+                      eq $tag_name) {
+           !!!cp ('t51');
            $formatting_element = $active_formatting_elements->[$_];
            $formatting_element_i_in_active = $_;
            last AFE;
-         } elsif ($active_formatting_elements->[$_]->[0] eq '#marker') {
-           last AFE;
          }
        } # AFE
        unless (defined $formatting_element) {
-         !!!parse-error (type => 'unmatched end tag:'.$tag_name);
+         !!!cp ('t53');
+         !!!parse-error (type => 'unmatched end tag', text => $tag_name, token => $end_tag_token);
          ## Ignore the token
          !!!next-token;
          return;
-Line 2939 
 sub _tree_construction_main ($) {
+Line 3761 
 sub _tree_construction_main ($) {
          my $node = $self->{open_elements}->[$_];
          if ($node->[0] eq $formatting_element->[0]) {
            if ($in_scope) {
+             !!!cp ('t54');
              $formatting_element_i_in_open = $_;
              last INSCOPE;
            } else { # in open elements but not in scope
-             !!!parse-error (type => 'unmatched end tag:'.$token->{tag_name});
+             !!!cp ('t55');
+             !!!parse-error (type => 'unmatched end tag',
+                             text => $token->{tag_name},
+                             token => $end_tag_token);
              ## Ignore the token
              !!!next-token;
              return;
            }
-         } elsif ({
+         } elsif ($node->[1] & SCOPING_EL) {
-                   table => 1, caption => 1, td => 1, th => 1,
+           !!!cp ('t56');
-                   button => 1, marquee => 1, object => 1, html => 1,
-                  }->{$node->[1]}) {
            $in_scope = 0;
          }
        } # INSCOPE
        unless (defined $formatting_element_i_in_open) {
-         !!!parse-error (type => 'unmatched end tag:'.$token->{tag_name});
+         !!!cp ('t57');
+         !!!parse-error (type => 'unmatched end tag',
+                         text => $token->{tag_name},
+                         token => $end_tag_token);
          pop @$active_formatting_elements; # $formatting_element
          !!!next-token; ## TODO: ok?
          return;
        }
        if (not $self->{open_elements}->[-1]->[0] eq $formatting_element->[0]) {
-         !!!parse-error (type => 'not closed:'.$self->{open_elements}->[-1]->[1]);
+         !!!cp ('t58');
+         !!!parse-error (type => 'not closed',
+                         text => $self->{open_elements}->[-1]->[0]
+                             ->manakai_local_name,
+                         token => $end_tag_token);
        }
        ## Step 2
-Line 2969 
 sub _tree_construction_main ($) {
+Line 3800 
 sub _tree_construction_main ($) {
        my $furthest_block_i_in_open;
        OE: for (reverse 0..$#{$self->{open_elements}}) {
          my $node = $self->{open_elements}->[$_];
-         if (not $formatting_category->{$node->[1]} and
+         if (not ($node->[1] & FORMATTING_EL) and
              #not $phrasing_category->{$node->[1]} and
-             ($special_category->{$node->[1]} or
+             ($node->[1] & SPECIAL_EL or
-              $scoping_category->{$node->[1]})) {
+              $node->[1] & SCOPING_EL)) { ## Scoping is redundant, maybe
+           !!!cp ('t59');
            $furthest_block = $node;
            $furthest_block_i_in_open = $_;
          } elsif ($node->[0] eq $formatting_element->[0]) {
+           !!!cp ('t60');
            last OE;
          }
        } # OE
        ## Step 3
        unless (defined $furthest_block) { # MUST
+         !!!cp ('t61');
          splice @{$self->{open_elements}}, $formatting_element_i_in_open;
          splice @$active_formatting_elements, $formatting_element_i_in_active, 1;
          !!!next-token;
-Line 2994 
 sub _tree_construction_main ($) {
+Line 3828 
 sub _tree_construction_main ($) {
        ## Step 5
        my $furthest_block_parent = $furthest_block->[0]->parent_node;
        if (defined $furthest_block_parent) {
+         !!!cp ('t62');
          $furthest_block_parent->remove_child ($furthest_block->[0]);
        }
-Line 3016 
 sub _tree_construction_main ($) {
+Line 3851 
 sub _tree_construction_main ($) {
          S7S2: {
            for (reverse 0..$#$active_formatting_elements) {
              if ($active_formatting_elements->[$_]->[0] eq $node->[0]) {
+               !!!cp ('t63');
                $node_i_in_active = $_;
                last S7S2;
              }
-Line 3029 
 sub _tree_construction_main ($) {
+Line 3865 
 sub _tree_construction_main ($) {
          ## Step 4
          if ($last_node->[0] eq $furthest_block->[0]) {
+           !!!cp ('t64');
            $bookmark_prev_el = $node->[0];
          }
          ## Step 5
          if ($node->[0]->has_child_nodes ()) {
+           !!!cp ('t65');
            my $clone = [$node->[0]->clone_node (0), $node->[1]];
            $active_formatting_elements->[$node_i_in_active] = $clone;
            $self->{open_elements}->[$node_i_in_open] = $clone;
-Line 3051 
 sub _tree_construction_main ($) {
+Line 3889 
 sub _tree_construction_main ($) {
        } # S7
        ## Step 8
-       $common_ancestor_node->[0]->append_child ($last_node->[0]);
+       if ($common_ancestor_node->[1] & TABLE_ROWS_EL) {
+         my $foster_parent_element;
+         my $next_sibling;
+         OE: for (reverse 0..$#{$self->{open_elements}}) {
+           if ($self->{open_elements}->[$_]->[1] & TABLE_EL) {
+                              my $parent = $self->{open_elements}->[$_]->[0]->parent_node;
+                              if (defined $parent and $parent->node_type == 1) {
+                                !!!cp ('t65.1');
+                                $foster_parent_element = $parent;
+                                $next_sibling = $self->{open_elements}->[$_]->[0];
+                              } else {
+                                !!!cp ('t65.2');
+                                $foster_parent_element
+                                  = $self->{open_elements}->[$_ - 1]->[0];
+                              }
+                              last OE;
+                            }
+                          } # OE
+                          $foster_parent_element = $self->{open_elements}->[0]->[0]
+                            unless defined $foster_parent_element;
+         $foster_parent_element->insert_before ($last_node->[0], $next_sibling);
+         $open_tables->[-1]->[1] = 1; # tainted
+       } else {
+         !!!cp ('t65.3');
+         $common_ancestor_node->[0]->append_child ($last_node->[0]);
+       }
        ## Step 9
        my $clone = [$formatting_element->[0]->clone_node (0),
-Line 3068 
 sub _tree_construction_main ($) {
+Line 3931 
 sub _tree_construction_main ($) {
        my $i;
        AFE: for (reverse 0..$#$active_formatting_elements) {
          if ($active_formatting_elements->[$_]->[0] eq $formatting_element->[0]) {
+           !!!cp ('t66');
            splice @$active_formatting_elements, $_, 1;
            $i-- and last AFE if defined $i;
          } elsif ($active_formatting_elements->[$_]->[0] eq $bookmark_prev_el) {
+           !!!cp ('t67');
            $i = $_;
          }
        } # AFE
-Line 3080 
 sub _tree_construction_main ($) {
+Line 3945 
 sub _tree_construction_main ($) {
        undef $i;
        OE: for (reverse 0..$#{$self->{open_elements}}) {
          if ($self->{open_elements}->[$_]->[0] eq $formatting_element->[0]) {
+           !!!cp ('t68');
            splice @{$self->{open_elements}}, $_, 1;
            $i-- and last OE if defined $i;
          } elsif ($self->{open_elements}->[$_]->[0] eq $furthest_block->[0]) {
+           !!!cp ('t69');
            $i = $_;
          }
        } # OE
-Line 3093 
 sub _tree_construction_main ($) {
+Line 3960 
 sub _tree_construction_main ($) {
      } # FET
    }; # $formatting_end_tag
-   my $insert_to_current = sub {
+   $insert = my $insert_to_current = sub {
      $self->{open_elements}->[-1]->[0]->append_child ($_[0]);
    }; # $insert_to_current
    my $insert_to_foster = sub {
-                        my $child = shift;
+     my $child = shift;
-                        if ({
+     if ($self->{open_elements}->[-1]->[1] & TABLE_ROWS_EL) {
-                             table => 1, tbody => 1, tfoot => 1,
+       # MUST
-                             thead => 1, tr => 1,
+       my $foster_parent_element;
-                            }->{$self->{open_elements}->[-1]->[1]}) {
+       my $next_sibling;
-                          # MUST
+       OE: for (reverse 0..$#{$self->{open_elements}}) {
-                          my $foster_parent_element;
+         if ($self->{open_elements}->[$_]->[1] & TABLE_EL) {
-                          my $next_sibling;
-                          OE: for (reverse 0..$#{$self->{open_elements}}) {
-                            if ($self->{open_elements}->[$_]->[1] eq 'table') {
                               my $parent = $self->{open_elements}->[$_]->[0]->parent_node;
                               if (defined $parent and $parent->node_type == 1) {
+                                !!!cp ('t70');
                                 $foster_parent_element = $parent;
                                 $next_sibling = $self->{open_elements}->[$_]->[0];
                               } else {
+                                !!!cp ('t71');
                                 $foster_parent_element
                                   = $self->{open_elements}->[$_ - 1]->[0];
                               }
-Line 3123 
 sub _tree_construction_main ($) {
+Line 3989 
 sub _tree_construction_main ($) {
                             unless defined $foster_parent_element;
                           $foster_parent_element->insert_before
                             ($child, $next_sibling);
-                        } else {
+       $open_tables->[-1]->[1] = 1; # tainted
-                          $self->{open_elements}->[-1]->[0]->append_child ($child);
+     } else {
-                        }
+       !!!cp ('t72');
+       $self->{open_elements}->[-1]->[0]->append_child ($child);
+     }
    }; # $insert_to_foster
-   my $insert;
+   B: while (1) {
-   B: {
      if ($token->{type} == DOCTYPE_TOKEN) {
-       !!!parse-error (type => 'DOCTYPE in the middle');
+       !!!cp ('t73');
+       !!!parse-error (type => 'in html:#DOCTYPE', token => $token);
        ## Ignore the token
        ## Stay in the phase
        !!!next-token;
-       redo B;
+       next B;
-     } elsif ($token->{type} == END_OF_FILE_TOKEN) {
-       if ($self->{insertion_mode} & AFTER_HTML_IMS) {
-         #
-       } else {
-         ## Generate implied end tags
-         if ({
-              dd => 1, dt => 1, li => 1, p => 1, td => 1, th => 1, tr => 1,
-              tbody => 1, tfoot=> 1, thead => 1,
-             }->{$self->{open_elements}->[-1]->[1]}) {
-           !!!back-token;
-           $token = {type => END_TAG_TOKEN, tag_name => $self->{open_elements}->[-1]->[1]};
-           redo B;
-         }
-         if (@{$self->{open_elements}} > 2 or
-             (@{$self->{open_elements}} == 2 and $self->{open_elements}->[1]->[1] ne 'body')) {
-           !!!parse-error (type => 'not closed:'.$self->{open_elements}->[-1]->[1]);
-         } elsif (defined $self->{inner_html_node} and
-                  @{$self->{open_elements}} > 1 and
-                  $self->{open_elements}->[1]->[1] ne 'body') {
-           !!!parse-error (type => 'not closed:'.$self->{open_elements}->[-1]->[1]);
-         }
-         ## ISSUE: There is an issue in the spec.
-       }
-       ## Stop parsing
-       last B;
      } elsif ($token->{type} == START_TAG_TOKEN and
               $token->{tag_name} eq 'html') {
        if ($self->{insertion_mode} == AFTER_HTML_BODY_IM) {
-         ## Turn into the main phase
+         !!!cp ('t79');
-         !!!parse-error (type => 'after html:html');
+         !!!parse-error (type => 'after html', text => 'html', token => $token);
          $self->{insertion_mode} = AFTER_BODY_IM;
        } elsif ($self->{insertion_mode} == AFTER_HTML_FRAMESET_IM) {
-         ## Turn into the main phase
+         !!!cp ('t80');
-         !!!parse-error (type => 'after html:html');
+         !!!parse-error (type => 'after html', text => 'html', token => $token);
          $self->{insertion_mode} = AFTER_FRAMESET_IM;
+       } else {
+         !!!cp ('t81');
        }
- ## ISSUE: "aa<html>" is not a parse error.
+       !!!cp ('t82');
- ## ISSUE: "<html>" in fragment is not a parse error.
+       !!!parse-error (type => 'not first start tag', token => $token);
-       unless ($token->{first_start_tag}) {
-         !!!parse-error (type => 'not first start tag');
-       }
        my $top_el = $self->{open_elements}->[0]->[0];
        for my $attr_name (keys %{$token->{attributes}}) {
          unless ($top_el->has_attribute_ns (undef, $attr_name)) {
+           !!!cp ('t84');
            $top_el->set_attribute_ns
              (undef, [undef, $attr_name],
               $token->{attributes}->{$attr_name}->{value});
          }
        }
+       !!!nack ('t84.1');
        !!!next-token;
-       redo B;
+       next B;
      } elsif ($token->{type} == COMMENT_TOKEN) {
        my $comment = $self->{document}->create_comment ($token->{data});
        if ($self->{insertion_mode} & AFTER_HTML_IMS) {
+         !!!cp ('t85');
          $self->{document}->append_child ($comment);
        } elsif ($self->{insertion_mode} == AFTER_BODY_IM) {
+         !!!cp ('t86');
          $self->{open_elements}->[0]->[0]->append_child ($comment);
        } else {
+         !!!cp ('t87');
          $self->{open_elements}->[-1]->[0]->append_child ($comment);
        }
        !!!next-token;
-       redo B;
+       next B;
-     } elsif ($self->{insertion_mode} & HEAD_IMS) {
+     } elsif ($self->{insertion_mode} & IN_FOREIGN_CONTENT_IM) {
+       if ($token->{type} == CHARACTER_TOKEN) {
+         !!!cp ('t87.1');
+         $self->{open_elements}->[-1]->[0]->manakai_append_text ($token->{data});
+         !!!next-token;
+         next B;
+       } elsif ($token->{type} == START_TAG_TOKEN) {
+         if ((not {mglyph => 1, malignmark => 1}->{$token->{tag_name}} and
+              $self->{open_elements}->[-1]->[1] & FOREIGN_FLOW_CONTENT_EL) or
+             not ($self->{open_elements}->[-1]->[1] & FOREIGN_EL) or
+             ($token->{tag_name} eq 'svg' and
+              $self->{open_elements}->[-1]->[1] & MML_AXML_EL)) {
+           ## NOTE: "using the rules for secondary insertion mode"then"continue"
+           !!!cp ('t87.2');
+           #
+         } elsif ({
+                   b => 1, big => 1, blockquote => 1, body => 1, br => 1,
+                   center => 1, code => 1, dd => 1, div => 1, dl => 1, dt => 1,
+                   em => 1, embed => 1, font => 1, h1 => 1, h2 => 1, h3 => 1,
+                   h4 => 1, h5 => 1, h6 => 1, head => 1, hr => 1, i => 1,
+                   img => 1, li => 1, listing => 1, menu => 1, meta => 1,
+                   nobr => 1, ol => 1, p => 1, pre => 1, ruby => 1, s => 1,
+                   small => 1, span => 1, strong => 1, strike => 1, sub => 1,
+                   sup => 1, table => 1, tt => 1, u => 1, ul => 1, var => 1,
+                  }->{$token->{tag_name}}) {
+           !!!cp ('t87.2');
+           !!!parse-error (type => 'not closed',
+                           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 {
+           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} || $tag_name;
+           }
+           ## "adjust SVG attributes" (SVG only) - done in insert-element-f
+           ## "adjust foreign attributes" - done in insert-element-f
+           !!!insert-element-f ($nsuri, $tag_name, $token->{attributes}, $token);
+           if ($self->{self_closing}) {
+             pop @{$self->{open_elements}};
+             !!!ack ('t87.3');
+           } else {
+             !!!cp ('t87.4');
+           }
+           !!!next-token;
+           next B;
+         }
+       } elsif ($token->{type} == END_TAG_TOKEN) {
+         ## NOTE: "using the rules for secondary insertion mode" then "continue"
+         !!!cp ('t87.5');
+         #
+       } elsif ($token->{type} == END_OF_FILE_TOKEN) {
+         !!!cp ('t87.6');
+         !!!parse-error (type => 'not closed',
+                         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: $token->{type}: Unknown token type";
+       }
+     }
+     if ($self->{insertion_mode} & HEAD_IMS) {
        if ($token->{type} == CHARACTER_TOKEN) {
          if ($token->{data} =~ s/^([\x09\x0A\x0B\x0C\x20]+)//) {
-           $self->{open_elements}->[-1]->[0]->manakai_append_text ($1);
+           unless ($self->{insertion_mode} == BEFORE_HEAD_IM) {
+             !!!cp ('t88.2');
+             $self->{open_elements}->[-1]->[0]->manakai_append_text ($1);
+           } else {
+             !!!cp ('t88.1');
+             ## Ignore the token.
+             !!!next-token;
+             next B;
+           }
            unless (length $token->{data}) {
+             !!!cp ('t88');
              !!!next-token;
-             redo B;
+             next B;
            }
          }
          if ($self->{insertion_mode} == BEFORE_HEAD_IM) {
+           !!!cp ('t89');
            ## As if <head>
-           !!!create-element ($self->{head_element}, 'head');
+           !!!create-element ($self->{head_element}, $HTML_NS, 'head',, $token);
            $self->{open_elements}->[-1]->[0]->append_child ($self->{head_element});
-           push @{$self->{open_elements}}, [$self->{head_element}, 'head'];
+           push @{$self->{open_elements}},
+               [$self->{head_element}, $el_category->{head}];
            ## Reprocess in the "in head" insertion mode...
            pop @{$self->{open_elements}};
            ## Reprocess in the "after head" insertion mode...
          } elsif ($self->{insertion_mode} == IN_HEAD_NOSCRIPT_IM) {
+           !!!cp ('t90');
            ## As if </noscript>
            pop @{$self->{open_elements}};
-           !!!parse-error (type => 'in noscript:#character');
+           !!!parse-error (type => 'in noscript:#text', token => $token);
            ## Reprocess in the "in head" insertion mode...
            ## As if </head>
-Line 3234 
 sub _tree_construction_main ($) {
+Line 4209 
 sub _tree_construction_main ($) {
            ## Reprocess in the "after head" insertion mode...
          } elsif ($self->{insertion_mode} == IN_HEAD_IM) {
+           !!!cp ('t91');
            pop @{$self->{open_elements}};
            ## Reprocess in the "after head" insertion mode...
+         } else {
+           !!!cp ('t92');
          }
-             ## "after head" insertion mode
+         ## "after head" insertion mode
-             ## As if <body>
+         ## As if <body>
-             !!!insert-element ('body');
+         !!!insert-element ('body',, $token);
-             $self->{insertion_mode} = IN_BODY_IM;
+         $self->{insertion_mode} = IN_BODY_IM;
-             ## reprocess
+         ## reprocess
-             redo B;
+         next B;
-           } elsif ($token->{type} == START_TAG_TOKEN) {
+       } elsif ($token->{type} == START_TAG_TOKEN) {
-             if ($token->{tag_name} eq 'head') {
+         if ($token->{tag_name} eq 'head') {
-               if ($self->{insertion_mode} == BEFORE_HEAD_IM) {
+           if ($self->{insertion_mode} == BEFORE_HEAD_IM) {
-                 !!!create-element ($self->{head_element}, $token->{tag_name}, $token->{attributes});
+             !!!cp ('t93');
-                 $self->{open_elements}->[-1]->[0]->append_child ($self->{head_element});
+             !!!create-element ($self->{head_element}, $HTML_NS, $token->{tag_name}, $token->{attributes}, $token);
-                 push @{$self->{open_elements}}, [$self->{head_element}, $token->{tag_name}];
+             $self->{open_elements}->[-1]->[0]->append_child
-                 $self->{insertion_mode} = IN_HEAD_IM;
+                 ($self->{head_element});
-                 !!!next-token;
+             push @{$self->{open_elements}},
-                 redo B;
+                 [$self->{head_element}, $el_category->{head}];
-               } elsif ($self->{insertion_mode} == AFTER_HEAD_IM) {
+             $self->{insertion_mode} = IN_HEAD_IM;
-                 #
+             !!!nack ('t93.1');
-               } else {
+             !!!next-token;
-                 !!!parse-error (type => 'in head:head'); # or in head noscript
+             next B;
-                 ## Ignore the token
+           } elsif ($self->{insertion_mode} == AFTER_HEAD_IM) {
-                 !!!next-token;
+             !!!cp ('t93.2');
-                 redo B;
+             !!!parse-error (type => 'after head', text => 'head',
-               }
+                             token => $token);
-             } elsif ($self->{insertion_mode} == BEFORE_HEAD_IM) {
+             ## Ignore the token
-               ## As if <head>
+             !!!nack ('t93.3');
-               !!!create-element ($self->{head_element}, 'head');
+             !!!next-token;
-               $self->{open_elements}->[-1]->[0]->append_child ($self->{head_element});
+             next B;
-               push @{$self->{open_elements}}, [$self->{head_element}, 'head'];
+           } else {
+             !!!cp ('t95');
+             !!!parse-error (type => 'in head:head',
+                             token => $token); # or in head noscript
+             ## Ignore the token
+             !!!nack ('t95.1');
+             !!!next-token;
+             next B;
+           }
+         } elsif ($self->{insertion_mode} == BEFORE_HEAD_IM) {
+           !!!cp ('t96');
+           ## 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->{head_element}, $el_category->{head}];
-               $self->{insertion_mode} = IN_HEAD_IM;
+           $self->{insertion_mode} = IN_HEAD_IM;
-               ## Reprocess in the "in head" insertion mode...
+           ## Reprocess in the "in head" insertion mode...
-             }
+         } else {
+           !!!cp ('t97');
+         }
              if ($token->{tag_name} eq 'base') {
                if ($self->{insertion_mode} == IN_HEAD_NOSCRIPT_IM) {
+                 !!!cp ('t98');
                  ## As if </noscript>
                  pop @{$self->{open_elements}};
-                 !!!parse-error (type => 'in noscript:base');
+                 !!!parse-error (type => 'in noscript', text => 'base',
+                                 token => $token);
                  $self->{insertion_mode} = IN_HEAD_IM;
                  ## Reprocess in the "in head" insertion mode...
+               } else {
+                 !!!cp ('t99');
                }
                ## NOTE: There is a "as if in head" code clone.
                if ($self->{insertion_mode} == AFTER_HEAD_IM) {
-                 !!!parse-error (type => 'after head:'.$token->{tag_name});
+                 !!!cp ('t100');
-                 push @{$self->{open_elements}}, [$self->{head_element}, 'head'];
+                 !!!parse-error (type => 'after head',
+                                 text => $token->{tag_name}, token => $token);
+                 push @{$self->{open_elements}},
+                     [$self->{head_element}, $el_category->{head}];
+               } else {
+                 !!!cp ('t101');
                }
-               !!!insert-element ($token->{tag_name}, $token->{attributes});
+               !!!insert-element ($token->{tag_name}, $token->{attributes}, $token);
                pop @{$self->{open_elements}}; ## ISSUE: This step is missing in the spec.
-               pop @{$self->{open_elements}}
+               pop @{$self->{open_elements}} # <head>
                    if $self->{insertion_mode} == AFTER_HEAD_IM;
+               !!!nack ('t101.1');
                !!!next-token;
-               redo B;
+               next B;
              } elsif ($token->{tag_name} eq 'link') {
                ## NOTE: There is a "as if in head" code clone.
                if ($self->{insertion_mode} == AFTER_HEAD_IM) {
-                 !!!parse-error (type => 'after head:'.$token->{tag_name});
+                 !!!cp ('t102');
-                 push @{$self->{open_elements}}, [$self->{head_element}, 'head'];
+                 !!!parse-error (type => 'after head',
+                                 text => $token->{tag_name}, token => $token);
+                 push @{$self->{open_elements}},
+                     [$self->{head_element}, $el_category->{head}];
+               } else {
+                 !!!cp ('t103');
                }
-               !!!insert-element ($token->{tag_name}, $token->{attributes});
+               !!!insert-element ($token->{tag_name}, $token->{attributes}, $token);
                pop @{$self->{open_elements}}; ## ISSUE: This step is missing in the spec.
-               pop @{$self->{open_elements}}
+               pop @{$self->{open_elements}} # <head>
                    if $self->{insertion_mode} == AFTER_HEAD_IM;
+               !!!ack ('t103.1');
                !!!next-token;
-               redo B;
+               next B;
              } elsif ($token->{tag_name} eq 'meta') {
                ## NOTE: There is a "as if in head" code clone.
                if ($self->{insertion_mode} == AFTER_HEAD_IM) {
-                 !!!parse-error (type => 'after head:'.$token->{tag_name});
+                 !!!cp ('t104');
-                 push @{$self->{open_elements}}, [$self->{head_element}, 'head'];
+                 !!!parse-error (type => 'after head',
+                                 text => $token->{tag_name}, token => $token);
+                 push @{$self->{open_elements}},
+                     [$self->{head_element}, $el_category->{head}];
+               } else {
+                 !!!cp ('t105');
                }
-               !!!insert-element ($token->{tag_name}, $token->{attributes});
+               !!!insert-element ($token->{tag_name}, $token->{attributes}, $token);
                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, $token->{attributes}->{charset}->{value});
+                       ->($self, $token->{attributes}->{charset}->{value},
+                          $token);
                    $meta_el->[0]->get_attribute_node_ns (undef, 'charset')
                        ->set_user_data (manakai_has_reference =>
                                             $token->{attributes}->{charset}
                                                 ->{has_reference});
                  } elsif ($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 : $3);
+                         ->($self, defined $1 ? $1 : defined $2 ? $2 : $3,
+                            $token);
                      $meta_el->[0]->get_attribute_node_ns (undef, 'content')
                          ->set_user_data (manakai_has_reference =>
                                               $token->{attributes}->{content}
                                                     ->{has_reference});
+                   } else {
+                     !!!cp ('t108');
                    }
                  }
                } else {
                  if ($token->{attributes}->{charset}) {
+                   !!!cp ('t109');
                    $meta_el->[0]->get_attribute_node_ns (undef, 'charset')
                        ->set_user_data (manakai_has_reference =>
                                             $token->{attributes}->{charset}
                                                 ->{has_reference});
                  }
                  if ($token->{attributes}->{content}) {
+                   !!!cp ('t110');
                    $meta_el->[0]->get_attribute_node_ns (undef, 'content')
                        ->set_user_data (manakai_has_reference =>
                                             $token->{attributes}->{content}
-Line 3353 
 sub _tree_construction_main ($) {
+Line 4380 
 sub _tree_construction_main ($) {
                  }
                }
-               pop @{$self->{open_elements}}
+               pop @{$self->{open_elements}} # <head>
                    if $self->{insertion_mode} == AFTER_HEAD_IM;
+               !!!ack ('t110.1');
                !!!next-token;
-               redo B;
+               next B;
              } elsif ($token->{tag_name} eq 'title') {
                if ($self->{insertion_mode} == IN_HEAD_NOSCRIPT_IM) {
+                 !!!cp ('t111');
                  ## As if </noscript>
                  pop @{$self->{open_elements}};
-                 !!!parse-error (type => 'in noscript:title');
+                 !!!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) {
-                 !!!parse-error (type => 'after head:'.$token->{tag_name});
+                 !!!cp ('t112');
-                 push @{$self->{open_elements}}, [$self->{head_element}, 'head'];
+                 !!!parse-error (type => 'after head',
+                                 text => $token->{tag_name}, token => $token);
+                 push @{$self->{open_elements}},
+                     [$self->{head_element}, $el_category->{head}];
+               } else {
+                 !!!cp ('t113');
                }
                ## NOTE: There is a "as if in head" code clone.
                my $parent = defined $self->{head_element} ? $self->{head_element}
                    : $self->{open_elements}->[-1]->[0];
-               $parse_rcdata->(RCDATA_CONTENT_MODEL,
+               $parse_rcdata->(RCDATA_CONTENT_MODEL);
-                               sub { $parent->append_child ($_[0]) });
+               pop @{$self->{open_elements}} # <head>
-               pop @{$self->{open_elements}}
                    if $self->{insertion_mode} == AFTER_HEAD_IM;
-               redo B;
+               next B;
-             } elsif ($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) {
-                 !!!parse-error (type => 'after head:'.$token->{tag_name});
+                 !!!cp ('t114');
-                 push @{$self->{open_elements}}, [$self->{head_element}, 'head'];
+                 !!!parse-error (type => 'after head',
+                                 text => $token->{tag_name}, token => $token);
+                 push @{$self->{open_elements}},
+                     [$self->{head_element}, $el_category->{head}];
+               } else {
+                 !!!cp ('t115');
                }
-               $parse_rcdata->(CDATA_CONTENT_MODEL, $insert_to_current);
+               $parse_rcdata->(CDATA_CONTENT_MODEL);
-               pop @{$self->{open_elements}}
+               pop @{$self->{open_elements}} # <head>
                    if $self->{insertion_mode} == AFTER_HEAD_IM;
-               redo B;
+               next B;
              } elsif ($token->{tag_name} eq 'noscript') {
                if ($self->{insertion_mode} == IN_HEAD_IM) {
+                 !!!cp ('t116');
                  ## NOTE: and scripting is disalbed
-                 !!!insert-element ($token->{tag_name}, $token->{attributes});
+                 !!!insert-element ($token->{tag_name}, $token->{attributes}, $token);
                  $self->{insertion_mode} = IN_HEAD_NOSCRIPT_IM;
+                 !!!nack ('t116.1');
                  !!!next-token;
-                 redo B;
+                 next B;
                } elsif ($self->{insertion_mode} == IN_HEAD_NOSCRIPT_IM) {
-                 !!!parse-error (type => 'in noscript:noscript');
+                 !!!cp ('t117');
+                 !!!parse-error (type => 'in noscript', text => 'noscript',
+                                 token => $token);
                  ## Ignore the token
+                 !!!nack ('t117.1');
                  !!!next-token;
-                 redo B;
+                 next B;
                } else {
+                 !!!cp ('t118');
                  #
                }
              } elsif ($token->{tag_name} eq 'script') {
                if ($self->{insertion_mode} == IN_HEAD_NOSCRIPT_IM) {
+                 !!!cp ('t119');
                  ## As if </noscript>
                  pop @{$self->{open_elements}};
-                 !!!parse-error (type => 'in noscript:script');
+                 !!!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) {
-                 !!!parse-error (type => 'after head:'.$token->{tag_name});
+                 !!!cp ('t120');
-                 push @{$self->{open_elements}}, [$self->{head_element}, 'head'];
+                 !!!parse-error (type => 'after head',
+                                 text => $token->{tag_name}, token => $token);
+                 push @{$self->{open_elements}},
+                     [$self->{head_element}, $el_category->{head}];
+               } else {
+                 !!!cp ('t121');
                }
                ## NOTE: There is a "as if in head" code clone.
-               $script_start_tag->($insert_to_current);
+               $script_start_tag->();
-               pop @{$self->{open_elements}}
+               pop @{$self->{open_elements}} # <head>
                    if $self->{insertion_mode} == AFTER_HEAD_IM;
-               redo B;
+               next B;
              } elsif ($token->{tag_name} eq 'body' or
                       $token->{tag_name} eq 'frameset') {
                if ($self->{insertion_mode} == IN_HEAD_NOSCRIPT_IM) {
+                 !!!cp ('t122');
                  ## As if </noscript>
                  pop @{$self->{open_elements}};
-                 !!!parse-error (type => 'in noscript:'.$token->{tag_name});
+                 !!!parse-error (type => 'in noscript',
+                                 text => $token->{tag_name}, token => $token);
                  ## Reprocess in the "in head" insertion mode...
                  ## As if </head>
-Line 3436 
 sub _tree_construction_main ($) {
+Line 4491 
 sub _tree_construction_main ($) {
                  ## Reprocess in the "after head" insertion mode...
                } elsif ($self->{insertion_mode} == IN_HEAD_IM) {
+                 !!!cp ('t124');
                  pop @{$self->{open_elements}};
                  ## Reprocess in the "after head" insertion mode...
+               } else {
+                 !!!cp ('t125');
                }
                ## "after head" insertion mode
-               !!!insert-element ($token->{tag_name}, $token->{attributes});
+               !!!insert-element ($token->{tag_name}, $token->{attributes}, $token);
                if ($token->{tag_name} eq 'body') {
+                 !!!cp ('t126');
                  $self->{insertion_mode} = IN_BODY_IM;
                } elsif ($token->{tag_name} eq 'frameset') {
+                 !!!cp ('t127');
                  $self->{insertion_mode} = IN_FRAMESET_IM;
                } else {
                  die "$0: tag name: $self->{tag_name}";
                }
+               !!!nack ('t127.1');
                !!!next-token;
-               redo B;
+               next B;
              } else {
+               !!!cp ('t128');
                #
              }
              if ($self->{insertion_mode} == IN_HEAD_NOSCRIPT_IM) {
+               !!!cp ('t129');
                ## As if </noscript>
                pop @{$self->{open_elements}};
-               !!!parse-error (type => 'in noscript:/'.$token->{tag_name});
+               !!!parse-error (type => 'in noscript:/',
+                               text => $token->{tag_name}, token => $token);
                ## Reprocess in the "in head" insertion mode...
                ## As if </head>
-Line 3467 
 sub _tree_construction_main ($) {
+Line 4531 
 sub _tree_construction_main ($) {
                ## Reprocess in the "after head" insertion mode...
              } elsif ($self->{insertion_mode} == IN_HEAD_IM) {
+               !!!cp ('t130');
                ## As if </head>
                pop @{$self->{open_elements}};
                ## Reprocess in the "after head" insertion mode...
+             } else {
+               !!!cp ('t131');
              }
              ## "after head" insertion mode
              ## As if <body>
-             !!!insert-element ('body');
+             !!!insert-element ('body',, $token);
              $self->{insertion_mode} = IN_BODY_IM;
              ## reprocess
-             redo B;
+             !!!ack-later;
+             next B;
            } elsif ($token->{type} == END_TAG_TOKEN) {
              if ($token->{tag_name} eq 'head') {
                if ($self->{insertion_mode} == BEFORE_HEAD_IM) {
+                 !!!cp ('t132');
                  ## As if <head>
-                 !!!create-element ($self->{head_element}, 'head');
+                 !!!create-element ($self->{head_element}, $HTML_NS, 'head',, $token);
                  $self->{open_elements}->[-1]->[0]->append_child ($self->{head_element});
-                 push @{$self->{open_elements}}, [$self->{head_element}, 'head'];
+                 push @{$self->{open_elements}},
+                     [$self->{head_element}, $el_category->{head}];
                  ## Reprocess in the "in head" insertion mode...
                  pop @{$self->{open_elements}};
                  $self->{insertion_mode} = AFTER_HEAD_IM;
                  !!!next-token;
-                 redo B;
+                 next B;
                } elsif ($self->{insertion_mode} == IN_HEAD_NOSCRIPT_IM) {
+                 !!!cp ('t133');
                  ## As if </noscript>
                  pop @{$self->{open_elements}};
-                 !!!parse-error (type => 'in noscript:script');
+                 !!!parse-error (type => 'in noscript:/',
+                                 text => 'head', token => $token);
                  ## Reprocess in the "in head" insertion mode...
                  pop @{$self->{open_elements}};
                  $self->{insertion_mode} = AFTER_HEAD_IM;
                  !!!next-token;
-                 redo B;
+                 next B;
                } elsif ($self->{insertion_mode} == IN_HEAD_IM) {
+                 !!!cp ('t134');
                  pop @{$self->{open_elements}};
                  $self->{insertion_mode} = AFTER_HEAD_IM;
                  !!!next-token;
-                 redo B;
+                 next B;
+               } elsif ($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 {
-                 #
+                 die "$0: $self->{insertion_mode}: Unknown insertion mode";
                }
              } elsif ($token->{tag_name} eq 'noscript') {
                if ($self->{insertion_mode} == IN_HEAD_NOSCRIPT_IM) {
+                 !!!cp ('t136');
                  pop @{$self->{open_elements}};
                  $self->{insertion_mode} = IN_HEAD_IM;
                  !!!next-token;
-                 redo B;
+                 next B;
-               } elsif ($self->{insertion_mode} == BEFORE_HEAD_IM) {
+               } elsif ($self->{insertion_mode} == BEFORE_HEAD_IM or
-                 !!!parse-error (type => 'unmatched end tag:noscript');
+                        $self->{insertion_mode} == AFTER_HEAD_IM) {
+                 !!!cp ('t137');
+                 !!!parse-error (type => 'unmatched end tag',
+                                 text => 'noscript', token => $token);
                  ## Ignore the token ## ISSUE: An issue in the spec.
                  !!!next-token;
-                 redo B;
+                 next B;
                } else {
+                 !!!cp ('t138');
                  #
                }
              } elsif ({
                        body => 1, html => 1,
                       }->{$token->{tag_name}}) {
-               if ($self->{insertion_mode} == BEFORE_HEAD_IM) {
+               if ($self->{insertion_mode} == BEFORE_HEAD_IM or
-                 ## As if <head>
+                   $self->{insertion_mode} == IN_HEAD_IM or
-                 !!!create-element ($self->{head_element}, 'head');
+                   $self->{insertion_mode} == IN_HEAD_NOSCRIPT_IM) {
-                 $self->{open_elements}->[-1]->[0]->append_child ($self->{head_element});
+                 !!!cp ('t140');
-                 push @{$self->{open_elements}}, [$self->{head_element}, 'head'];
+                 !!!parse-error (type => 'unmatched end tag',
+                                 text => $token->{tag_name}, token => $token);
-                 $self->{insertion_mode} = IN_HEAD_IM;
-                 ## Reprocess in the "in head" insertion mode...
-               } elsif ($self->{insertion_mode} == IN_HEAD_NOSCRIPT_IM) {
-                 !!!parse-error (type => 'unmatched end tag:'.$token->{tag_name});
                  ## Ignore the token
                  !!!next-token;
-                 redo B;
+                 next B;
+               } elsif ($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 {
+                 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) {
-                 ## As if <head>
+                 !!!cp ('t142.2');
-                 !!!create-element ($self->{head_element}, 'head');
+                 ## (before head) as if <head>, (in head) as if </head>
+                 !!!create-element ($self->{head_element}, $HTML_NS, 'head',, $token);
                  $self->{open_elements}->[-1]->[0]->append_child ($self->{head_element});
-                 push @{$self->{open_elements}}, [$self->{head_element}, 'head'];
+                 $self->{insertion_mode} = AFTER_HEAD_IM;
+                 ## Reprocess in the "after head" insertion mode...
+               } elsif ($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}};
                  $self->{insertion_mode} = IN_HEAD_IM;
                  ## Reprocess in the "in head" insertion mode...
-               }
+                 ## As if </head>
+                 pop @{$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 ($self->{insertion_mode} == AFTER_HEAD_IM) {
+                 !!!cp ('t143.4');
                  #
                } else {
-                 !!!parse-error (type => 'unmatched end tag:'.$token->{tag_name});
+                 die "$0: $self->{insertion_mode}: Unknown insertion mode";
-                 ## Ignore the token
-                 !!!next-token;
-                 redo 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});
+               !!!parse-error (type => 'in noscript:/',
+                               text => $token->{tag_name}, token => $token);
                ## Reprocess in the "in head" insertion mode...
                ## As if </head>
-Line 3579 
 sub _tree_construction_main ($) {
+Line 4704 
 sub _tree_construction_main ($) {
                ## Reprocess in the "after head" insertion mode...
              } elsif ($self->{insertion_mode} == IN_HEAD_IM) {
+               !!!cp ('t147');
                ## As if </head>
                pop @{$self->{open_elements}};
                ## Reprocess in the "after head" insertion mode...
              } elsif ($self->{insertion_mode} == BEFORE_HEAD_IM) {
-               !!!parse-error (type => 'unmatched end tag:'.$token->{tag_name});
+ ## ISSUE: This case cannot be reached?
+               !!!cp ('t148');
+               !!!parse-error (type => 'unmatched end tag',
+                               text => $token->{tag_name}, token => $token);
                ## Ignore the token ## ISSUE: An issue in the spec.
                !!!next-token;
-               redo B;
+               next B;
+             } else {
+               !!!cp ('t149');
              }
              ## "after head" insertion mode
              ## As if <body>
-             !!!insert-element ('body');
+             !!!insert-element ('body',, $token);
              $self->{insertion_mode} = IN_BODY_IM;
              ## reprocess
-             redo B;
+             next B;
-           } else {
+       } elsif ($token->{type} == END_OF_FILE_TOKEN) {
-             die "$0: $token->{type}: Unknown token type";
+         if ($self->{insertion_mode} == BEFORE_HEAD_IM) {
-           }
+           !!!cp ('t149.1');
+           ## NOTE: 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->{head_element}, $el_category->{head}];
+           #$self->{insertion_mode} = IN_HEAD_IM;
+           ## NOTE: Reprocess.
+           ## NOTE: As if </head>
+           #pop @{$self->{open_elements}};
+           #$self->{insertion_mode} = IN_AFTER_HEAD_IM;
+           ## NOTE: Reprocess.
+           #
+         } elsif ($self->{insertion_mode} == IN_HEAD_IM) {
+           !!!cp ('t149.2');
+           ## NOTE: As if </head>
+           pop @{$self->{open_elements}};
+           #$self->{insertion_mode} = IN_AFTER_HEAD_IM;
+           ## NOTE: Reprocess.
+           #
+         } elsif ($self->{insertion_mode} == IN_HEAD_NOSCRIPT_IM) {
+           !!!cp ('t149.3');
+           !!!parse-error (type => 'in noscript:#eof', token => $token);
+           ## As if </noscript>
+           pop @{$self->{open_elements}};
+           #$self->{insertion_mode} = IN_HEAD_IM;
+           ## NOTE: Reprocess.
+           ## NOTE: As if </head>
+           pop @{$self->{open_elements}};
+           #$self->{insertion_mode} = IN_AFTER_HEAD_IM;
+           ## NOTE: Reprocess.
+           #
+         } else {