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

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

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 1.215 by wakaba, Thu Jul 2 22:11:51 2009 UTC revision 1.242 by wakaba, Sun Sep 6 13:02:21 2009 UTC
# Line 179  my $el_category = { Line 179  my $el_category = {
179    dt => DTDD_EL,    dt => DTDD_EL,
180    em => FORMATTING_EL,    em => FORMATTING_EL,
181    embed => MISC_SPECIAL_EL,    embed => MISC_SPECIAL_EL,
   eventsource => MISC_SPECIAL_EL,  
182    fieldset => MISC_SPECIAL_EL,    fieldset => MISC_SPECIAL_EL,
183    figure => MISC_SPECIAL_EL,    figure => MISC_SPECIAL_EL,
184    font => FORMATTING_EL,    font => FORMATTING_EL,
# Line 195  my $el_category = { Line 194  my $el_category = {
194    h6 => HEADING_EL,    h6 => HEADING_EL,
195    head => MISC_SPECIAL_EL,    head => MISC_SPECIAL_EL,
196    header => MISC_SPECIAL_EL,    header => MISC_SPECIAL_EL,
197      hgroup => MISC_SPECIAL_EL,
198    hr => MISC_SPECIAL_EL,    hr => MISC_SPECIAL_EL,
199    html => HTML_EL,    html => HTML_EL,
200    i => FORMATTING_EL,    i => FORMATTING_EL,
# Line 203  my $el_category = { Line 203  my $el_category = {
203    #image => MISC_SPECIAL_EL, ## NOTE: Commented out in the spec.    #image => MISC_SPECIAL_EL, ## NOTE: Commented out in the spec.
204    input => MISC_SPECIAL_EL,    input => MISC_SPECIAL_EL,
205    isindex => MISC_SPECIAL_EL,    isindex => MISC_SPECIAL_EL,
206      ## XXX keygen? (Whether a void element is in Special or not does not
207      ## affect to the processing, however.)
208    li => LI_EL,    li => LI_EL,
209    link => MISC_SPECIAL_EL,    link => MISC_SPECIAL_EL,
210    listing => MISC_SPECIAL_EL,    listing => MISC_SPECIAL_EL,
# Line 247  my $el_category = { Line 249  my $el_category = {
249    u => FORMATTING_EL,    u => FORMATTING_EL,
250    ul => MISC_SPECIAL_EL,    ul => MISC_SPECIAL_EL,
251    wbr => MISC_SPECIAL_EL,    wbr => MISC_SPECIAL_EL,
252      xmp => MISC_SPECIAL_EL,
253  };  };
254    
255  my $el_category_f = {  my $el_category_f = {
# Line 647  sub parse_char_stream ($$$;$$) { Line 650  sub parse_char_stream ($$$;$$) {
650    
651    ## NOTE: |set_inner_html| copies most of this method's code    ## NOTE: |set_inner_html| copies most of this method's code
652    
653      ## Confidence: irrelevant.
654    $self->{confident} = 1 unless exists $self->{confident};    $self->{confident} = 1 unless exists $self->{confident};
655    
656    $self->{document}->input_encoding ($self->{input_encoding})    $self->{document}->input_encoding ($self->{input_encoding})
657        if defined $self->{input_encoding};        if defined $self->{input_encoding};
658  ## TODO: |{input_encoding}| is needless?  ## TODO: |{input_encoding}| is needless?
# Line 864  sub _initialize_tree_constructor ($) { Line 869  sub _initialize_tree_constructor ($) {
869    $self->{document}->manakai_is_html (1); # MUST    $self->{document}->manakai_is_html (1); # MUST
870    $self->{document}->set_user_data (manakai_source_line => 1);    $self->{document}->set_user_data (manakai_source_line => 1);
871    $self->{document}->set_user_data (manakai_source_column => 1);    $self->{document}->set_user_data (manakai_source_column => 1);
872    
873      $self->{frameset_ok} = 1;
874  } # _initialize_tree_constructor  } # _initialize_tree_constructor
875    
876  sub _terminate_tree_constructor ($) {  sub _terminate_tree_constructor ($) {
# Line 911  sub _tree_construction_initial ($) { Line 918  sub _tree_construction_initial ($) {
918    
919    INITIAL: {    INITIAL: {
920      if ($token->{type} == DOCTYPE_TOKEN) {      if ($token->{type} == DOCTYPE_TOKEN) {
921        ## NOTE: Conformance checkers MAY, instead of reporting "not HTML5"        ## NOTE: Conformance checkers MAY, instead of reporting "not
922        ## error, switch to a conformance checking mode for another        ## HTML5" error, switch to a conformance checking mode for
923        ## language.        ## another language.  (We don't support such mode switchings; it
924          ## is nonsense to do anything different from what browsers do.)
925        my $doctype_name = $token->{name};        my $doctype_name = $token->{name};
926        $doctype_name = '' unless defined $doctype_name;        $doctype_name = '' unless defined $doctype_name;
927        $doctype_name =~ tr/a-z/A-Z/; # ASCII case-insensitive        my $doctype = $self->{document}->create_document_type_definition
928        if (not defined $token->{name} or # <!DOCTYPE>            ($doctype_name);
929            defined $token->{sysid}) {  
930          $doctype_name =~ tr/A-Z/a-z/; # ASCII case-insensitive
931          if ($doctype_name ne 'html') {
932          !!!cp ('t1');          !!!cp ('t1');
933          !!!parse-error (type => 'not HTML5', token => $token);          !!!parse-error (type => 'not HTML5', token => $token);
934        } elsif ($doctype_name ne 'HTML') {        } elsif (defined $token->{pubid}) {
935          !!!cp ('t2');          !!!cp ('t2');
936            ## XXX Obsolete permitted DOCTYPEs
937          !!!parse-error (type => 'not HTML5', token => $token);          !!!parse-error (type => 'not HTML5', token => $token);
938        } elsif (defined $token->{pubid}) {        } elsif (defined $token->{sysid}) {
939          if ($token->{pubid} eq 'XSLT-compat') {          if ($token->{sysid} eq 'about:legacy-compat') {
940            !!!cp ('t1.2');            !!!cp ('t1.2'); ## <!DOCTYPE HTML SYSTEM "about:legacy-compat">
941            !!!parse-error (type => 'XSLT-compat', token => $token,            !!!parse-error (type => 'XSLT-compat', token => $token,
942                            level => $self->{level}->{should});                            level => $self->{level}->{should});
943          } else {          } else {
944            !!!parse-error (type => 'not HTML5', token => $token);            !!!parse-error (type => 'not HTML5', token => $token);
945          }          }
946        } else {        } else { ## <!DOCTYPE HTML>
947          !!!cp ('t3');          !!!cp ('t3');
948          #          #
949        }        }
950                
       my $doctype = $self->{document}->create_document_type_definition  
         ($token->{name}); ## ISSUE: If name is missing (e.g. <!DOCTYPE>)?  
951        ## NOTE: Default value for both |public_id| and |system_id| attributes        ## NOTE: Default value for both |public_id| and |system_id| attributes
952        ## are empty strings, so that we don't set any value in missing cases.        ## are empty strings, so that we don't set any value in missing cases.
953        $doctype->public_id ($token->{pubid}) if defined $token->{pubid};        $doctype->public_id ($token->{pubid}) if defined $token->{pubid};
954        $doctype->system_id ($token->{sysid}) if defined $token->{sysid};        $doctype->system_id ($token->{sysid}) if defined $token->{sysid};
955    
956        ## NOTE: Other DocumentType attributes are null or empty lists.        ## NOTE: Other DocumentType attributes are null or empty lists.
957        ## In Firefox3, |internalSubset| attribute is set to the empty        ## In Firefox3, |internalSubset| attribute is set to the empty
958        ## string, while |null| is an allowed value for the attribute        ## string, while |null| is an allowed value for the attribute
959        ## according to DOM3 Core.        ## according to DOM3 Core.
960        $self->{document}->append_child ($doctype);        $self->{document}->append_child ($doctype);
961                
962        if ($token->{quirks} or $doctype_name ne 'HTML') {        if ($token->{quirks} or $doctype_name ne 'html') {
963          !!!cp ('t4');          !!!cp ('t4');
964          $self->{document}->manakai_compat_mode ('quirks');          $self->{document}->manakai_compat_mode ('quirks');
965        } elsif (defined $token->{pubid}) {        } elsif (defined $token->{pubid}) {
# Line 1419  sub _tree_construction_main ($) { Line 1429  sub _tree_construction_main ($) {
1429      ## Step 3      ## Step 3
1430      ## TODO: Mark as "already executed", if ...      ## TODO: Mark as "already executed", if ...
1431    
1432      ## Step 4      ## Step 4 (HTML5 revision 2702)
1433      $insert->($script_el);      $insert->($script_el);
   
     ## ISSUE: $script_el is not put into the stack  
1434      push @{$self->{open_elements}}, [$script_el, $el_category->{script}];      push @{$self->{open_elements}}, [$script_el, $el_category->{script}];
1435    
1436      ## Step 5      ## Step 5
# Line 1437  sub _tree_construction_main ($) { Line 1445  sub _tree_construction_main ($) {
1445    }; # $script_start_tag    }; # $script_start_tag
1446    
1447    ## NOTE: $open_tables->[-1]->[0] is the "current table" element node.    ## NOTE: $open_tables->[-1]->[0] is the "current table" element node.
1448    ## NOTE: $open_tables->[-1]->[1] is the "tainted" flag.    ## NOTE: $open_tables->[-1]->[1] is the "tainted" flag (OBSOLETE; unused).
1449    ## NOTE: $open_tables->[-1]->[2] is set false when non-Text node inserted.    ## NOTE: $open_tables->[-1]->[2] is set false when non-Text node inserted.
1450    my $open_tables = [[$self->{open_elements}->[0]->[0]]];    my $open_tables = [[$self->{open_elements}->[0]->[0]]];
1451    
# Line 1607  sub _tree_construction_main ($) { Line 1615  sub _tree_construction_main ($) {
1615                
1616        ## Step 8        ## Step 8
1617        if ($common_ancestor_node->[1] & TABLE_ROWS_EL) {        if ($common_ancestor_node->[1] & TABLE_ROWS_EL) {
1618            ## Foster parenting.
1619          my $foster_parent_element;          my $foster_parent_element;
1620          my $next_sibling;          my $next_sibling;
1621          OE: for (reverse 0..$#{$self->{open_elements}}) {          OE: for (reverse 0..$#{$self->{open_elements}}) {
1622            if ($self->{open_elements}->[$_]->[1] == TABLE_EL) {            if ($self->{open_elements}->[$_]->[1] == TABLE_EL) {
1623                               my $parent = $self->{open_elements}->[$_]->[0]->parent_node;              !!!cp ('t65.2');
1624                               if (defined $parent and $parent->node_type == 1) {              $foster_parent_element = $self->{open_elements}->[$_ - 1]->[0];
1625                                 !!!cp ('t65.1');              $next_sibling = $self->{open_elements}->[$_]->[0];
1626                                 $foster_parent_element = $parent;              undef $next_sibling
1627                                 $next_sibling = $self->{open_elements}->[$_]->[0];                  unless $next_sibling->parent_node eq $foster_parent_element;
1628                               } else {              last OE;
1629                                 !!!cp ('t65.2');            }
1630                                 $foster_parent_element          } # OE
1631                                   = $self->{open_elements}->[$_ - 1]->[0];          $foster_parent_element ||= $self->{open_elements}->[0]->[0];
1632                               }  
                              last OE;  
                            }  
                          } # OE  
                          $foster_parent_element = $self->{open_elements}->[0]->[0]  
                            unless defined $foster_parent_element;  
1633          $foster_parent_element->insert_before ($last_node->[0], $next_sibling);          $foster_parent_element->insert_before ($last_node->[0], $next_sibling);
1634          $open_tables->[-1]->[1] = 1; # tainted          $open_tables->[-1]->[1] = 1; # tainted
1635        } else {        } else {
# Line 1681  sub _tree_construction_main ($) { Line 1685  sub _tree_construction_main ($) {
1685      $self->{open_elements}->[-1]->[0]->append_child ($_[0]);      $self->{open_elements}->[-1]->[0]->append_child ($_[0]);
1686    }; # $insert_to_current    }; # $insert_to_current
1687    
1688      ## Foster parenting.  Note that there are three "foster parenting"
1689      ## code in the parser: for elements (this one), for texts, and for
1690      ## elements in the AAA code.
1691    my $insert_to_foster = sub {    my $insert_to_foster = sub {
1692      my $child = shift;      my $child = shift;
1693      if ($self->{open_elements}->[-1]->[1] & TABLE_ROWS_EL) {      if ($self->{open_elements}->[-1]->[1] & TABLE_ROWS_EL) {
# Line 1689  sub _tree_construction_main ($) { Line 1696  sub _tree_construction_main ($) {
1696        my $next_sibling;        my $next_sibling;
1697        OE: for (reverse 0..$#{$self->{open_elements}}) {        OE: for (reverse 0..$#{$self->{open_elements}}) {
1698          if ($self->{open_elements}->[$_]->[1] == TABLE_EL) {          if ($self->{open_elements}->[$_]->[1] == TABLE_EL) {
1699                               my $parent = $self->{open_elements}->[$_]->[0]->parent_node;            !!!cp ('t71');
1700                               if (defined $parent and $parent->node_type == 1) {            $foster_parent_element = $self->{open_elements}->[$_ - 1]->[0];
1701                                 !!!cp ('t70');            $next_sibling = $self->{open_elements}->[$_]->[0];
1702                                 $foster_parent_element = $parent;            undef $next_sibling
1703                                 $next_sibling = $self->{open_elements}->[$_]->[0];                unless $next_sibling->parent_node eq $foster_parent_element;
1704                               } else {            last OE;
1705                                 !!!cp ('t71');          }
1706                                 $foster_parent_element        } # OE
1707                                   = $self->{open_elements}->[$_ - 1]->[0];        $foster_parent_element ||= $self->{open_elements}->[0]->[0];
1708                               }  
1709                               last OE;        $foster_parent_element->insert_before ($child, $next_sibling);
                            }  
                          } # OE  
                          $foster_parent_element = $self->{open_elements}->[0]->[0]  
                            unless defined $foster_parent_element;  
                          $foster_parent_element->insert_before  
                            ($child, $next_sibling);  
1710        $open_tables->[-1]->[1] = 1; # tainted        $open_tables->[-1]->[1] = 1; # tainted
1711      } else {      } else {
1712        !!!cp ('t72');        !!!cp ('t72');
# Line 1733  sub _tree_construction_main ($) { Line 1734  sub _tree_construction_main ($) {
1734    ## document.write ("b")</script>|    ## document.write ("b")</script>|
1735    
1736    B: while (1) {    B: while (1) {
1737    
1738        ## The "in table text" insertion mode.
1739        if ($self->{insertion_mode} & TABLE_IMS and
1740            not $self->{insertion_mode} & IN_FOREIGN_CONTENT_IM and
1741            not $self->{insertion_mode} & IN_CDATA_RCDATA_IM) {
1742          C: {
1743            my $s;
1744            if ($token->{type} == CHARACTER_TOKEN) {
1745              !!!cp ('t194');
1746              $self->{pending_chars} ||= [];
1747              push @{$self->{pending_chars}}, $token;
1748              !!!next-token;
1749              next B;
1750            } else {
1751              if ($self->{pending_chars}) {
1752                $s = join '', map { $_->{data} } @{$self->{pending_chars}};
1753                delete $self->{pending_chars};
1754                if ($s =~ /[^\x09\x0A\x0C\x0D\x20]/) {
1755                  !!!cp ('t195');
1756                  #
1757                } else {
1758                  !!!cp ('t195.1');
1759                  #$self->{open_elements}->[-1]->[0]->manakai_append_text ($s);
1760                  $self->{open_elements}->[-1]->[0]->append_child
1761                      ($self->{document}->create_text_node ($s));
1762                  last C;
1763                }
1764              } else {
1765                !!!cp ('t195.2');
1766                last C;
1767              }
1768            }
1769    
1770            ## Foster parenting.
1771            !!!parse-error (type => 'in table:#text', token => $token);
1772    
1773            ## NOTE: As if in body, but insert into the foster parent element.
1774            $reconstruct_active_formatting_elements->($insert_to_foster);
1775                
1776            if ($self->{open_elements}->[-1]->[1] & TABLE_ROWS_EL) {
1777              # MUST
1778              my $foster_parent_element;
1779              my $next_sibling;
1780              OE: for (reverse 0..$#{$self->{open_elements}}) {
1781                if ($self->{open_elements}->[$_]->[1] == TABLE_EL) {
1782                  !!!cp ('t197');
1783                  $foster_parent_element = $self->{open_elements}->[$_ - 1]->[0];
1784                  $next_sibling = $self->{open_elements}->[$_]->[0];
1785                  undef $next_sibling
1786                    unless $next_sibling->parent_node eq $foster_parent_element;
1787                  last OE;
1788                }
1789              } # OE
1790              $foster_parent_element ||= $self->{open_elements}->[0]->[0];
1791    
1792              !!!cp ('t199');
1793              $foster_parent_element->insert_before
1794                  ($self->{document}->create_text_node ($s), $next_sibling);
1795    
1796              $open_tables->[-1]->[1] = 1; # tainted
1797              $open_tables->[-1]->[2] = 1; # ~node inserted
1798            } else {
1799              ## NOTE: Fragment case or in a foster parent'ed element
1800              ## (e.g. |<table><span>a|).  In fragment case, whether the
1801              ## character is appended to existing node or a new node is
1802              ## created is irrelevant, since the foster parent'ed nodes
1803              ## are discarded and fragment parsing does not invoke any
1804              ## script.
1805              !!!cp ('t200');
1806              $self->{open_elements}->[-1]->[0]->manakai_append_text ($s);
1807            }
1808          } # C
1809        } # TABLE_IMS
1810    
1811      if ($token->{type} == DOCTYPE_TOKEN) {      if ($token->{type} == DOCTYPE_TOKEN) {
1812        !!!cp ('t73');        !!!cp ('t73');
1813        !!!parse-error (type => 'in html:#DOCTYPE', token => $token);        !!!parse-error (type => 'in html:#DOCTYPE', token => $token);
# Line 1874  sub _tree_construction_main ($) { Line 1949  sub _tree_construction_main ($) {
1949          } elsif ({          } elsif ({
1950                    b => 1, big => 1, blockquote => 1, body => 1, br => 1,                    b => 1, big => 1, blockquote => 1, body => 1, br => 1,
1951                    center => 1, code => 1, dd => 1, div => 1, dl => 1, dt => 1,                    center => 1, code => 1, dd => 1, div => 1, dl => 1, dt => 1,
1952                    em => 1, embed => 1, font => 1, h1 => 1, h2 => 1, h3 => 1,                    em => 1, embed => 1, h1 => 1, h2 => 1, h3 => 1,
1953                    h4 => 1, h5 => 1, h6 => 1, head => 1, hr => 1, i => 1,                    h4 => 1, h5 => 1, h6 => 1, head => 1, hr => 1, i => 1,
1954                    img => 1, li => 1, listing => 1, menu => 1, meta => 1,                    img => 1, li => 1, listing => 1, menu => 1, meta => 1,
1955                    nobr => 1, ol => 1, p => 1, pre => 1, ruby => 1, s => 1,                    nobr => 1, ol => 1, p => 1, pre => 1, ruby => 1, s => 1,
1956                    small => 1, span => 1, strong => 1, strike => 1, sub => 1,                    small => 1, span => 1, strong => 1, strike => 1, sub => 1,
1957                    sup => 1, table => 1, tt => 1, u => 1, ul => 1, var => 1,                    sup => 1, table => 1, tt => 1, u => 1, ul => 1, var => 1,
1958                   }->{$token->{tag_name}}) {                   }->{$token->{tag_name}} or
1959                     ($token->{tag_name} eq 'font' and
1960                      ($token->{attributes}->{color} or
1961                       $token->{attributes}->{face} or
1962                       $token->{attributes}->{size}))) {
1963            !!!cp ('t87.2');            !!!cp ('t87.2');
1964            !!!parse-error (type => 'not closed',            !!!parse-error (type => 'not closed',
1965                            text => $self->{open_elements}->[-1]->[0]                            text => $self->{open_elements}->[-1]->[0]
# Line 1956  sub _tree_construction_main ($) { Line 2035  sub _tree_construction_main ($) {
2035          }          }
2036        } elsif ($token->{type} == END_TAG_TOKEN) {        } elsif ($token->{type} == END_TAG_TOKEN) {
2037          ## NOTE: "using the rules for secondary insertion mode" then "continue"          ## NOTE: "using the rules for secondary insertion mode" then "continue"
2038          !!!cp ('t87.5');          if ($token->{tag_name} eq 'script') {
2039          #            !!!cp ('t87.41');
2040              #
2041              ## XXXscript: Execute script here.
2042            } else {
2043              !!!cp ('t87.5');
2044              #
2045            }
2046        } elsif ($token->{type} == END_OF_FILE_TOKEN) {        } elsif ($token->{type} == END_OF_FILE_TOKEN) {
2047          !!!cp ('t87.6');          !!!cp ('t87.6');
2048          !!!parse-error (type => 'not closed',          !!!parse-error (type => 'not closed',
# Line 2141  sub _tree_construction_main ($) { Line 2226  sub _tree_construction_main ($) {
2226            !!!ack ('t103.1');            !!!ack ('t103.1');
2227            !!!next-token;            !!!next-token;
2228            next B;            next B;
2229          } elsif ($token->{tag_name} eq 'command' or          } elsif ($token->{tag_name} eq 'command') {
                  $token->{tag_name} eq 'eventsource') {  
2230            if ($self->{insertion_mode} == IN_HEAD_IM) {            if ($self->{insertion_mode} == IN_HEAD_IM) {
2231              ## NOTE: If the insertion mode at the time of the emission              ## NOTE: If the insertion mode at the time of the emission
2232              ## of the token was "before head", $self->{insertion_mode}              ## of the token was "before head", $self->{insertion_mode}
# Line 2257  sub _tree_construction_main ($) { Line 2341  sub _tree_construction_main ($) {
2341    
2342            ## NOTE: There is a "as if in head" code clone.            ## NOTE: There is a "as if in head" code clone.
2343            $parse_rcdata->(RCDATA_CONTENT_MODEL);            $parse_rcdata->(RCDATA_CONTENT_MODEL);
2344            ## ISSUE: A spec bug [Bug 6038]  
2345              ## NOTE: At this point the stack of open elements contain
2346              ## the |head| element (index == -2) and the |script| element
2347              ## (index == -1).  In the "after head" insertion mode the
2348              ## |head| element is inserted only for the purpose of
2349              ## providing the context for the |script| element, and
2350              ## therefore we can now and have to remove the element from
2351              ## the stack.
2352            splice @{$self->{open_elements}}, -2, 1, () # <head>            splice @{$self->{open_elements}}, -2, 1, () # <head>
2353                if ($self->{insertion_mode} & IM_MASK) == AFTER_HEAD_IM;                if ($self->{insertion_mode} & IM_MASK) == AFTER_HEAD_IM;
2354            next B;            next B;
# Line 2400  sub _tree_construction_main ($) { Line 2491  sub _tree_construction_main ($) {
2491              ## reprocess              ## reprocess
2492              !!!ack-later;              !!!ack-later;
2493              next B;              next B;
2494            } elsif ($token->{type} == END_TAG_TOKEN) {        } elsif ($token->{type} == END_TAG_TOKEN) {
2495              if ($token->{tag_name} eq 'head') {          ## "Before head", "in head", and "after head" insertion modes
2496                if ($self->{insertion_mode} == BEFORE_HEAD_IM) {          ## ignore most of end tags.  Exceptions are "body", "html",
2497                  !!!cp ('t132');          ## and "br" end tags.  "Before head" and "in head" insertion
2498                  ## As if <head>          ## modes also recognize "head" end tag.  "In head noscript"
2499                  !!!create-element ($self->{head_element}, $HTML_NS, 'head',, $token);          ## insertion modes ignore end tags except for "noscript" and
2500                  $self->{open_elements}->[-1]->[0]->append_child ($self->{head_element});          ## "br".
                 push @{$self->{open_elements}},  
                     [$self->{head_element}, $el_category->{head}];  
2501    
2502                  ## Reprocess in the "in head" insertion mode...          if ($token->{tag_name} eq 'head') {
2503                  pop @{$self->{open_elements}};            if ($self->{insertion_mode} == BEFORE_HEAD_IM) {
2504                  $self->{insertion_mode} = AFTER_HEAD_IM;              !!!cp ('t132');
2505                  !!!next-token;              ## As if <head>
2506                  next B;              !!!create-element ($self->{head_element}, $HTML_NS, 'head',, $token);
2507                } elsif ($self->{insertion_mode} == IN_HEAD_NOSCRIPT_IM) {              $self->{open_elements}->[-1]->[0]->append_child ($self->{head_element});
2508                  !!!cp ('t133');              push @{$self->{open_elements}},
2509                  ## As if </noscript>                  [$self->{head_element}, $el_category->{head}];
2510                  pop @{$self->{open_elements}};  
2511                  !!!parse-error (type => 'in noscript:/',              ## Reprocess in the "in head" insertion mode...
2512                                  text => 'head', token => $token);              pop @{$self->{open_elements}};
2513                                $self->{insertion_mode} = AFTER_HEAD_IM;
2514                  ## Reprocess in the "in head" insertion mode...              !!!next-token;
2515                  pop @{$self->{open_elements}};              next B;
2516                  $self->{insertion_mode} = AFTER_HEAD_IM;            } elsif ($self->{insertion_mode} == IN_HEAD_NOSCRIPT_IM) {
2517                  !!!next-token;              !!!cp ('t133');
2518                  next B;              #
2519                } elsif ($self->{insertion_mode} == IN_HEAD_IM) {            } elsif ($self->{insertion_mode} == IN_HEAD_IM) {
2520                  !!!cp ('t134');              !!!cp ('t134');
2521                  pop @{$self->{open_elements}};              pop @{$self->{open_elements}};
2522                  $self->{insertion_mode} = AFTER_HEAD_IM;              $self->{insertion_mode} = AFTER_HEAD_IM;
2523                  !!!next-token;              !!!next-token;
2524                  next B;              next B;
2525                } elsif ($self->{insertion_mode} == AFTER_HEAD_IM) {            } elsif ($self->{insertion_mode} == AFTER_HEAD_IM) {
2526                  !!!cp ('t134.1');              !!!cp ('t134.1');
2527                  !!!parse-error (type => 'unmatched end tag', text => 'head',              #
2528                                  token => $token);            } else {
2529                  ## Ignore the token              die "$0: $self->{insertion_mode}: Unknown insertion mode";
2530                  !!!next-token;            }
2531                  next B;          } elsif ($token->{tag_name} eq 'noscript') {
2532                } else {            if ($self->{insertion_mode} == IN_HEAD_NOSCRIPT_IM) {
2533                  die "$0: $self->{insertion_mode}: Unknown insertion mode";              !!!cp ('t136');
2534                }              pop @{$self->{open_elements}};
2535              } elsif ($token->{tag_name} eq 'noscript') {              $self->{insertion_mode} = IN_HEAD_IM;
2536                if ($self->{insertion_mode} == IN_HEAD_NOSCRIPT_IM) {              !!!next-token;
2537                  !!!cp ('t136');              next B;
2538                  pop @{$self->{open_elements}};            } else {
2539                  $self->{insertion_mode} = IN_HEAD_IM;              !!!cp ('t138');
2540                  !!!next-token;              #
2541                  next B;            }
2542                } elsif ($self->{insertion_mode} == BEFORE_HEAD_IM or          } elsif ({
2543                         $self->{insertion_mode} == AFTER_HEAD_IM) {              body => ($self->{insertion_mode} != IN_HEAD_NOSCRIPT_IM),
2544                  !!!cp ('t137');              html => ($self->{insertion_mode} != IN_HEAD_NOSCRIPT_IM),
2545                  !!!parse-error (type => 'unmatched end tag',              br => 1,
2546                                  text => 'noscript', token => $token);          }->{$token->{tag_name}}) {
2547                  ## Ignore the token ## ISSUE: An issue in the spec.            if ($self->{insertion_mode} == BEFORE_HEAD_IM) {
2548                  !!!next-token;              !!!cp ('t142.2');
2549                  next B;              ## (before head) as if <head>, (in head) as if </head>
2550                } else {              !!!create-element ($self->{head_element}, $HTML_NS, 'head',, $token);
2551                  !!!cp ('t138');              $self->{open_elements}->[-1]->[0]->append_child ($self->{head_element});
2552                  #              $self->{insertion_mode} = AFTER_HEAD_IM;
               }  
             } elsif ({  
                       body => 1, html => 1,  
                      }->{$token->{tag_name}}) {  
               ## TODO: This branch is entirely redundant.  
               if ($self->{insertion_mode} == BEFORE_HEAD_IM or  
                   $self->{insertion_mode} == IN_HEAD_IM or  
                   $self->{insertion_mode} == IN_HEAD_NOSCRIPT_IM) {  
                 !!!cp ('t140');  
                 !!!parse-error (type => 'unmatched end tag',  
                                 text => $token->{tag_name}, token => $token);  
                 ## Ignore the token  
                 !!!next-token;  
                 next B;  
               } elsif ($self->{insertion_mode} == AFTER_HEAD_IM) {  
                 !!!cp ('t140.1');  
                 !!!parse-error (type => 'unmatched end tag',  
                                 text => $token->{tag_name}, token => $token);  
                 ## Ignore the token  
                 !!!next-token;  
                 next B;  
               } else {  
                 die "$0: $self->{insertion_mode}: Unknown insertion mode";  
               }  
             } elsif ($token->{tag_name} eq 'p') {  
               !!!cp ('t142');  
               !!!parse-error (type => 'unmatched end tag',  
                               text => $token->{tag_name}, token => $token);  
               ## Ignore the token  
               !!!next-token;  
               next B;  
             } elsif ($token->{tag_name} eq 'br') {  
               if ($self->{insertion_mode} == BEFORE_HEAD_IM) {  
                 !!!cp ('t142.2');  
                 ## (before head) as if <head>, (in head) as if </head>  
                 !!!create-element ($self->{head_element}, $HTML_NS, 'head',, $token);  
                 $self->{open_elements}->[-1]->[0]->append_child ($self->{head_element});  
                 $self->{insertion_mode} = AFTER_HEAD_IM;  
2553        
2554                  ## Reprocess in the "after head" insertion mode...              ## Reprocess in the "after head" insertion mode...
2555                } elsif ($self->{insertion_mode} == IN_HEAD_IM) {            } elsif ($self->{insertion_mode} == IN_HEAD_IM) {
2556                  !!!cp ('t143.2');              !!!cp ('t143.2');
2557                  ## As if </head>              ## As if </head>
2558                  pop @{$self->{open_elements}};              pop @{$self->{open_elements}};
2559                  $self->{insertion_mode} = AFTER_HEAD_IM;              $self->{insertion_mode} = AFTER_HEAD_IM;
2560        
2561                  ## Reprocess in the "after head" insertion mode...              ## Reprocess in the "after head" insertion mode...
2562                } elsif ($self->{insertion_mode} == IN_HEAD_NOSCRIPT_IM) {            } elsif ($self->{insertion_mode} == IN_HEAD_NOSCRIPT_IM) {
2563                  !!!cp ('t143.3');              !!!cp ('t143.3');
2564                  ## NOTE: Two parse errors for <head><noscript></br>              ## NOTE: Two parse errors for <head><noscript></br>
2565                  !!!parse-error (type => 'unmatched end tag',              !!!parse-error (type => 'unmatched end tag',
2566                                  text => 'br', token => $token);                              text => $token->{tag_name}, token => $token);
2567                  ## As if </noscript>              ## As if </noscript>
2568                  pop @{$self->{open_elements}};              pop @{$self->{open_elements}};
2569                  $self->{insertion_mode} = IN_HEAD_IM;              $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...  
               } elsif ($self->{insertion_mode} == AFTER_HEAD_IM) {  
                 !!!cp ('t143.4');  
                 #  
               } else {  
                 die "$0: $self->{insertion_mode}: Unknown insertion mode";  
               }  
   
               ## 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 => $token);  
               ## Ignore the token  
               !!!next-token;  
               next B;  
             }  
2570    
2571              if ($self->{insertion_mode} == IN_HEAD_NOSCRIPT_IM) {              ## Reprocess in the "in head" insertion mode...
2572                !!!cp ('t146');              ## As if </head>
2573                ## As if </noscript>              pop @{$self->{open_elements}};
2574                pop @{$self->{open_elements}};              $self->{insertion_mode} = AFTER_HEAD_IM;
               !!!parse-error (type => 'in noscript:/',  
                               text => $token->{tag_name}, token => $token);  
                 
               ## Reprocess in the "in head" insertion mode...  
               ## As if </head>  
               pop @{$self->{open_elements}};  
2575    
2576                ## Reprocess in the "after head" insertion mode...              ## Reprocess in the "after head" insertion mode...
2577              } elsif ($self->{insertion_mode} == IN_HEAD_IM) {            } elsif ($self->{insertion_mode} == AFTER_HEAD_IM) {
2578                !!!cp ('t147');              !!!cp ('t143.4');
2579                ## As if </head>              #
2580                pop @{$self->{open_elements}};            } else {
2581                die "$0: $self->{insertion_mode}: Unknown insertion mode";
2582              }
2583    
2584                ## Reprocess in the "after head" insertion mode...            ## "after head" insertion mode
2585              } elsif ($self->{insertion_mode} == BEFORE_HEAD_IM) {            ## As if <body>
2586  ## ISSUE: This case cannot be reached?            !!!insert-element ('body',, $token);
2587                !!!cp ('t148');            $self->{insertion_mode} = IN_BODY_IM;
2588                !!!parse-error (type => 'unmatched end tag',            ## Reprocess.
2589                                text => $token->{tag_name}, token => $token);            next B;
2590                ## Ignore the token ## ISSUE: An issue in the spec.          }
               !!!next-token;  
               next B;  
             } else {  
               !!!cp ('t149');  
             }  
2591    
2592              ## "after head" insertion mode          ## End tags are ignored by default.
2593              ## As if <body>          !!!cp ('t145');
2594              !!!insert-element ('body',, $token);          !!!parse-error (type => 'unmatched end tag',
2595              $self->{insertion_mode} = IN_BODY_IM;                          text => $token->{tag_name}, token => $token);
2596              ## reprocess          ## Ignore the token.
2597              next B;          !!!next-token;
2598            next B;
2599        } elsif ($token->{type} == END_OF_FILE_TOKEN) {        } elsif ($token->{type} == END_OF_FILE_TOKEN) {
2600          if ($self->{insertion_mode} == BEFORE_HEAD_IM) {          if ($self->{insertion_mode} == BEFORE_HEAD_IM) {
2601            !!!cp ('t149.1');            !!!cp ('t149.1');
# Line 3008  sub _tree_construction_main ($) { Line 3023  sub _tree_construction_main ($) {
3023        $insert = $insert_to_current;        $insert = $insert_to_current;
3024        #        #
3025      } elsif ($self->{insertion_mode} & TABLE_IMS) {      } elsif ($self->{insertion_mode} & TABLE_IMS) {
3026        if ($token->{type} == CHARACTER_TOKEN) {        if ($token->{type} == START_TAG_TOKEN) {
         if (not $open_tables->[-1]->[1] and # tainted  
             $token->{data} =~ s/^([\x09\x0A\x0C\x20]+)//) {  
           $self->{open_elements}->[-1]->[0]->manakai_append_text ($1);  
                 
           unless (length $token->{data}) {  
             !!!cp ('t194');  
             !!!next-token;  
             next B;  
           } else {  
             !!!cp ('t195');  
           }  
         }  
   
         !!!parse-error (type => 'in table:#text', token => $token);  
   
         ## NOTE: As if in body, but insert into the foster parent element.  
         $reconstruct_active_formatting_elements->($insert_to_foster);  
               
         if ($self->{open_elements}->[-1]->[1] & TABLE_ROWS_EL) {  
           # MUST  
           my $foster_parent_element;  
           my $next_sibling;  
           my $prev_sibling;  
           OE: for (reverse 0..$#{$self->{open_elements}}) {  
             if ($self->{open_elements}->[$_]->[1] == TABLE_EL) {  
               my $parent = $self->{open_elements}->[$_]->[0]->parent_node;  
               if (defined $parent and $parent->node_type == 1) {  
                 $foster_parent_element = $parent;  
                 !!!cp ('t196');  
                 $next_sibling = $self->{open_elements}->[$_]->[0];  
                 $prev_sibling = $next_sibling->previous_sibling;  
                 #  
               } else {  
                 !!!cp ('t197');  
                 $foster_parent_element = $self->{open_elements}->[$_ - 1]->[0];  
                 $prev_sibling = $foster_parent_element->last_child;  
                 #  
               }  
               last OE;  
             }  
           } # OE  
           $foster_parent_element = $self->{open_elements}->[0]->[0] and  
           $prev_sibling = $foster_parent_element->last_child  
               unless defined $foster_parent_element;  
           undef $prev_sibling unless $open_tables->[-1]->[2]; # ~node inserted  
           if (defined $prev_sibling and  
               $prev_sibling->node_type == 3) {  
             !!!cp ('t198');  
             $prev_sibling->manakai_append_text ($token->{data});  
           } else {  
             !!!cp ('t199');  
             $foster_parent_element->insert_before  
                 ($self->{document}->create_text_node ($token->{data}),  
                  $next_sibling);  
           }  
           $open_tables->[-1]->[1] = 1; # tainted  
           $open_tables->[-1]->[2] = 1; # ~node inserted  
         } else {  
           ## NOTE: Fragment case or in a foster parent'ed element  
           ## (e.g. |<table><span>a|).  In fragment case, whether the  
           ## character is appended to existing node or a new node is  
           ## created is irrelevant, since the foster parent'ed nodes  
           ## are discarded and fragment parsing does not invoke any  
           ## script.  
           !!!cp ('t200');  
           $self->{open_elements}->[-1]->[0]->manakai_append_text  
               ($token->{data});  
         }  
               
         !!!next-token;  
         next B;  
       } elsif ($token->{type} == START_TAG_TOKEN) {  
3027          if ({          if ({
3028               tr => (($self->{insertion_mode} & IM_MASK) != IN_ROW_IM),               tr => (($self->{insertion_mode} & IM_MASK) != IN_ROW_IM),
3029               th => 1, td => 1,               th => 1, td => 1,
# Line 3348  sub _tree_construction_main ($) { Line 3291  sub _tree_construction_main ($) {
3291            !!!ack-later;            !!!ack-later;
3292            next B;            next B;
3293          } elsif ($token->{tag_name} eq 'style') {          } elsif ($token->{tag_name} eq 'style') {
3294            if (not $open_tables->[-1]->[1]) { # tainted            !!!cp ('t227.8');
3295              !!!cp ('t227.8');            ## NOTE: This is a "as if in head" code clone.
3296              ## NOTE: This is a "as if in head" code clone.            $parse_rcdata->(CDATA_CONTENT_MODEL);
3297              $parse_rcdata->(CDATA_CONTENT_MODEL);            $open_tables->[-1]->[2] = 0 if @$open_tables; # ~node inserted
3298              $open_tables->[-1]->[2] = 0 if @$open_tables; # ~node inserted            next B;
             next B;  
           } else {  
             !!!cp ('t227.7');  
             #  
           }  
3299          } elsif ($token->{tag_name} eq 'script') {          } elsif ($token->{tag_name} eq 'script') {
3300            if (not $open_tables->[-1]->[1]) { # tainted            !!!cp ('t227.6');
3301              !!!cp ('t227.6');            ## NOTE: This is a "as if in head" code clone.
3302              ## NOTE: This is a "as if in head" code clone.            $script_start_tag->();
3303              $script_start_tag->();            $open_tables->[-1]->[2] = 0 if @$open_tables; # ~node inserted
3304              $open_tables->[-1]->[2] = 0 if @$open_tables; # ~node inserted            next B;
             next B;  
           } else {  
             !!!cp ('t227.5');  
             #  
           }  
3305          } elsif ($token->{tag_name} eq 'input') {          } elsif ($token->{tag_name} eq 'input') {
3306            if (not $open_tables->[-1]->[1]) { # tainted            if ($token->{attributes}->{type}) {
3307              if ($token->{attributes}->{type}) { ## TODO: case              my $type = $token->{attributes}->{type}->{value};
3308                my $type = lc $token->{attributes}->{type}->{value};              $type =~ tr/A-Z/a-z/; ## ASCII case-insensitive.
3309                if ($type eq 'hidden') {              if ($type eq 'hidden') {
3310                  !!!cp ('t227.3');                !!!cp ('t227.3');
3311                  !!!parse-error (type => 'in table',                !!!parse-error (type => 'in table',
3312                                  text => $token->{tag_name}, token => $token);                                text => $token->{tag_name}, token => $token);
3313    
3314                  !!!insert-element ($token->{tag_name}, $token->{attributes}, $token);                !!!insert-element ($token->{tag_name}, $token->{attributes}, $token);
3315                  $open_tables->[-1]->[2] = 0 if @$open_tables; # ~node inserted                $open_tables->[-1]->[2] = 0 if @$open_tables; # ~node inserted
3316    
3317                  ## TODO: form element pointer                ## TODO: form element pointer
3318    
3319                  pop @{$self->{open_elements}};                pop @{$self->{open_elements}};
3320    
3321                  !!!next-token;                !!!next-token;
3322                  !!!ack ('t227.2.1');                !!!ack ('t227.2.1');
3323                  next B;                next B;
               } else {  
                 !!!cp ('t227.2');  
                 #  
               }  
3324              } else {              } else {
3325                !!!cp ('t227.1');                !!!cp ('t227.1');
3326                #                #
# Line 3837  sub _tree_construction_main ($) { Line 3766  sub _tree_construction_main ($) {
3766            !!!next-token;            !!!next-token;
3767            next B;            next B;
3768          } elsif ({          } elsif ({
3769                     select => 1, input => 1, textarea => 1,                     select => 1, input => 1, textarea => 1, keygen => 1,
3770                   }->{$token->{tag_name}} or                   }->{$token->{tag_name}} or
3771                   (($self->{insertion_mode} & IM_MASK)                   (($self->{insertion_mode} & IM_MASK)
3772                        == IN_SELECT_IN_TABLE_IM and                        == IN_SELECT_IN_TABLE_IM and
# Line 3846  sub _tree_construction_main ($) { Line 3775  sub _tree_construction_main ($) {
3775                     tbody => 1, tfoot => 1, thead => 1,                     tbody => 1, tfoot => 1, thead => 1,
3776                     tr => 1, td => 1, th => 1,                     tr => 1, td => 1, th => 1,
3777                    }->{$token->{tag_name}})) {                    }->{$token->{tag_name}})) {
3778            ## TODO: The type below is not good - <select> is replaced by </select>  
3779            !!!parse-error (type => 'not closed', text => 'select',            ## 1. Parse error.
3780                            token => $token);            if ($token->{tag_name} eq 'select') {
3781            ## NOTE: As if the token were </select> (<select> case) or                !!!parse-error (type => 'select in select', ## XXX: documentation
3782            ## as if there were </select> (otherwise).                                token => $token);
3783            ## have an element in table scope            } else {
3784                !!!parse-error (type => 'not closed', text => 'select',
3785                                token => $token);
3786              }
3787    
3788              ## 2./<select>-1. Unless "have an element in table scope" (select):
3789            my $i;            my $i;
3790            INSCOPE: for (reverse 0..$#{$self->{open_elements}}) {            INSCOPE: for (reverse 0..$#{$self->{open_elements}}) {
3791              my $node = $self->{open_elements}->[$_];              my $node = $self->{open_elements}->[$_];
# Line 3866  sub _tree_construction_main ($) { Line 3800  sub _tree_construction_main ($) {
3800            } # INSCOPE            } # INSCOPE
3801            unless (defined $i) {            unless (defined $i) {
3802              !!!cp ('t280');              !!!cp ('t280');
3803              !!!parse-error (type => 'unmatched end tag',              if ($token->{tag_name} eq 'select') {
3804                              text => 'select', token => $token);                ## NOTE: This error would be raised when
3805              ## Ignore the token                ## |select.innerHTML = '<select>'| is executed; in this
3806                  ## case two errors, "select in select" and "unmatched
3807                  ## end tags" are reported to the user, the latter might
3808                  ## be confusing but this is what the spec requires.
3809                  !!!parse-error (type => 'unmatched end tag',
3810                                  text => 'select',
3811                                  token => $token);
3812                }
3813                ## Ignore the token.
3814              !!!nack ('t280.1');              !!!nack ('t280.1');
3815              !!!next-token;              !!!next-token;
3816              next B;              next B;
3817            }            }
3818    
3819              ## 3. Otherwise, as if there were <select>:
3820                                
3821            !!!cp ('t281');            !!!cp ('t281');
3822            splice @{$self->{open_elements}}, $i;            splice @{$self->{open_elements}}, $i;
# Line 3889  sub _tree_construction_main ($) { Line 3833  sub _tree_construction_main ($) {
3833              ## Reprocess the token.              ## Reprocess the token.
3834              next B;              next B;
3835            }            }
3836            } elsif ($token->{tag_name} eq 'script') {
3837              !!!cp ('t281.3');
3838              ## NOTE: This is an "as if in head" code clone
3839              $script_start_tag->();
3840              next B;
3841          } else {          } else {
3842            !!!cp ('t282');            !!!cp ('t282');
3843            !!!parse-error (type => 'in select',            !!!parse-error (type => 'in select',
# Line 4300  sub _tree_construction_main ($) { Line 4249  sub _tree_construction_main ($) {
4249          $parse_rcdata->(CDATA_CONTENT_MODEL);          $parse_rcdata->(CDATA_CONTENT_MODEL);
4250          next B;          next B;
4251        } elsif ({        } elsif ({
4252                  base => 1, command => 1, eventsource => 1, link => 1,                  base => 1, command => 1, link => 1,
4253                 }->{$token->{tag_name}}) {                 }->{$token->{tag_name}}) {
4254          !!!cp ('t334');          !!!cp ('t334');
4255          ## NOTE: This is an "as if in head" code clone, only "-t" differs          ## NOTE: This is an "as if in head" code clone, only "-t" differs
# Line 4390  sub _tree_construction_main ($) { Line 4339  sub _tree_construction_main ($) {
4339          !!!nack ('t343.1');          !!!nack ('t343.1');
4340          !!!next-token;          !!!next-token;
4341          next B;          next B;
4342          } elsif ($token->{tag_name} eq 'frameset') {
4343            !!!parse-error (type => 'in body', text => $token->{tag_name},
4344                            token => $token);
4345    
4346            if (@{$self->{open_elements}} == 1 or
4347                not ($self->{open_elements}->[1]->[1] == BODY_EL)) {
4348              !!!cp ('t343.2');
4349              ## Ignore the token.
4350            } elsif (not $self->{frameset_ok}) {
4351              !!!cp ('t343.3');
4352              ## Ignore the token.
4353            } else {
4354              !!!cp ('t343.4');
4355              
4356              ## 1. Remove the second element.
4357              my $body = $self->{open_elements}->[1]->[0];
4358              my $body_parent = $body->parent_node;
4359              $body_parent->remove_child ($body) if $body_parent;
4360    
4361              ## 2. Pop nodes.
4362              splice @{$self->{open_elements}}, 1;
4363    
4364              ## 3. Insert.
4365              !!!insert-element-t ($token->{tag_name}, $token->{attributes}, $token);
4366    
4367              ## 4. Switch.
4368              $self->{insertion_mode} = IN_FRAMESET_IM;
4369            }
4370    
4371            !!!nack ('t343.5');
4372            !!!next-token;
4373            next B;
4374        } elsif ({        } elsif ({
4375                  ## NOTE: Start tags for non-phrasing flow content elements                  ## NOTE: Start tags for non-phrasing flow content elements
4376    
# Line 4398  sub _tree_construction_main ($) { Line 4379  sub _tree_construction_main ($) {
4379                  center => 1, datagrid => 1, details => 1, dialog => 1,                  center => 1, datagrid => 1, details => 1, dialog => 1,
4380                  dir => 1, div => 1, dl => 1, fieldset => 1, figure => 1,                  dir => 1, div => 1, dl => 1, fieldset => 1, figure => 1,
4381                  footer => 1, h1 => 1, h2 => 1, h3 => 1, h4 => 1, h5 => 1,                  footer => 1, h1 => 1, h2 => 1, h3 => 1, h4 => 1, h5 => 1,
4382                  h6 => 1, header => 1, menu => 1, nav => 1, ol => 1, p => 1,                  h6 => 1, header => 1, hgroup => 1,
4383                    menu => 1, nav => 1, ol => 1, p => 1,
4384                  section => 1, ul => 1,                  section => 1, ul => 1,
4385                  ## NOTE: As normal, but drops leading newline                  ## NOTE: As normal, but drops leading newline
4386                  pre => 1, listing => 1,                  pre => 1, listing => 1,
# Line 4408  sub _tree_construction_main ($) { Line 4390  sub _tree_construction_main ($) {
4390                  table => 1,                  table => 1,
4391                  hr => 1,                  hr => 1,
4392                 }->{$token->{tag_name}}) {                 }->{$token->{tag_name}}) {
4393    
4394            ## 1. When there is an opening |form| element:
4395          if ($token->{tag_name} eq 'form' and defined $self->{form_element}) {          if ($token->{tag_name} eq 'form' and defined $self->{form_element}) {
4396            !!!cp ('t350');            !!!cp ('t350');
4397            !!!parse-error (type => 'in form:form', token => $token);            !!!parse-error (type => 'in form:form', token => $token);
# Line 4417  sub _tree_construction_main ($) { Line 4401  sub _tree_construction_main ($) {
4401            next B;            next B;
4402          }          }
4403    
4404          ## has a p element in scope          ## 2. Close the |p| element, if any.
4405          INSCOPE: for (reverse @{$self->{open_elements}}) {          if ($token->{tag_name} ne 'table' or # The Hixie Quirk
4406            if ($_->[1] == P_EL) {              $self->{document}->manakai_compat_mode ne 'quirks') {
4407              !!!cp ('t344');            ## has a p element in scope
4408              !!!back-token; # <form>            INSCOPE: for (reverse @{$self->{open_elements}}) {
4409              $token = {type => END_TAG_TOKEN, tag_name => 'p',              if ($_->[1] == P_EL) {
4410                        line => $token->{line}, column => $token->{column}};                !!!cp ('t344');
4411              next B;                !!!back-token; # <form>
4412            } elsif ($_->[1] & SCOPING_EL) {                $token = {type => END_TAG_TOKEN, tag_name => 'p',
4413              !!!cp ('t345');                          line => $token->{line}, column => $token->{column}};
4414              last INSCOPE;                next B;
4415                } elsif ($_->[1] & SCOPING_EL) {
4416                  !!!cp ('t345');
4417                  last INSCOPE;
4418                }
4419              } # INSCOPE
4420            }
4421    
4422            ## 3. Close the opening <hn> element, if any.
4423            if ({h1 => 1, h2 => 1, h3 => 1,
4424                 h4 => 1, h5 => 1, h6 => 1}->{$token->{tag_name}}) {
4425              if ($self->{open_elements}->[-1]->[1] == HEADING_EL) {
4426                !!!parse-error (type => 'not closed',
4427                                text => $self->{open_elements}->[-1]->[0]->manakai_local_name,
4428                                token => $token);
4429                pop @{$self->{open_elements}};
4430            }            }
4431          } # INSCOPE          }
4432              
4433            ## 4. Insertion.
4434          !!!insert-element-t ($token->{tag_name}, $token->{attributes}, $token);          !!!insert-element-t ($token->{tag_name}, $token->{attributes}, $token);
4435          if ($token->{tag_name} eq 'pre' or $token->{tag_name} eq 'listing') {          if ($token->{tag_name} eq 'pre' or $token->{tag_name} eq 'listing') {
4436            !!!nack ('t346.1');            !!!nack ('t346.1');
# Line 4464  sub _tree_construction_main ($) { Line 4464  sub _tree_construction_main ($) {
4464            !!!cp ('t386');            !!!cp ('t386');
4465            pop @{$self->{open_elements}};            pop @{$self->{open_elements}};
4466                    
4467            !!!nack ('t386.1');            !!!ack ('t386.1');
4468            !!!next-token;            !!!next-token;
4469          } else {          } else {
4470            !!!nack ('t347.1');            !!!nack ('t347.1');
# Line 4474  sub _tree_construction_main ($) { Line 4474  sub _tree_construction_main ($) {
4474        } elsif ($token->{tag_name} eq 'li') {        } elsif ($token->{tag_name} eq 'li') {
4475          ## NOTE: As normal, but imply </li> when there's another <li> ...          ## NOTE: As normal, but imply </li> when there's another <li> ...
4476    
4477          ## NOTE: Special, Scope (<li><foo><li> == <li><foo><li/></foo></li>)          ## NOTE: Special, Scope (<li><foo><li> == <li><foo><li/></foo></li>)::
4478            ## Interpreted as <li><foo/></li><li/> (non-conforming)            ## Interpreted as <li><foo/></li><li/> (non-conforming):
4479            ## blockquote (O9.27), center (O), dd (Fx3, O, S3.1.2, IE7),            ## blockquote (O9.27), center (O), dd (Fx3, O, S3.1.2, IE7),
4480            ## dt (Fx, O, S, IE), dl (O), fieldset (O, S, IE), form (Fx, O, S),            ## dt (Fx, O, S, IE), dl (O), fieldset (O, S, IE), form (Fx, O, S),
4481            ## hn (O), pre (O), applet (O, S), button (O, S), marquee (Fx, O, S),            ## hn (O), pre (O), applet (O, S), button (O, S), marquee (Fx, O, S),
4482            ## object (Fx)            ## object (Fx)
4483            ## Generate non-tree (non-conforming)            ## Generate non-tree (non-conforming):
4484            ## basefont (IE7 (where basefont is non-void)), center (IE),            ## basefont (IE7 (where basefont is non-void)), center (IE),
4485            ## form (IE), hn (IE)            ## form (IE), hn (IE)
4486          ## address, div, p (<li><foo><li> == <li><foo/></li><li/>)          ## address, div, p (<li><foo><li> == <li><foo/></li><li/>)::
4487            ## Interpreted as <li><foo><li/></foo></li> (non-conforming)            ## Interpreted as <li><foo><li/></foo></li> (non-conforming):
4488            ## div (Fx, S)            ## div (Fx, S)
4489    
4490          my $non_optional;          my $non_optional;
# Line 4831  sub _tree_construction_main ($) { Line 4831  sub _tree_construction_main ($) {
4831            next B;            next B;
4832          }          }
4833        } elsif ($token->{tag_name} eq 'textarea') {        } elsif ($token->{tag_name} eq 'textarea') {
4834          ## Step 1          ## 1. Insert
4835          !!!insert-element-t ($token->{tag_name}, $token->{attributes}, $token);          !!!insert-element-t ($token->{tag_name}, $token->{attributes}, $token);
4836                    
4837          ## Step 2          ## Step 2 # XXX
4838          ## TODO: $self->{form_element} if defined          ## TODO: $self->{form_element} if defined
4839    
4840          ## Step 3          ## 2. Drop U+000A LINE FEED
4841          $self->{ignore_newline} = 1;          $self->{ignore_newline} = 1;
4842    
4843          ## Step 4          ## 3. RCDATA
         ## ISSUE: This step is wrong. (r2302 enbugged)  
   
         ## Step 5  
4844          $self->{content_model} = RCDATA_CONTENT_MODEL;          $self->{content_model} = RCDATA_CONTENT_MODEL;
4845          delete $self->{escape}; # MUST          delete $self->{escape}; # MUST
4846    
4847          ## Step 6-7          ## 4., 6. Insertion mode
4848          $self->{insertion_mode} |= IN_CDATA_RCDATA_IM;          $self->{insertion_mode} |= IN_CDATA_RCDATA_IM;
4849    
4850            ## XXX: 5. frameset-ok flag
4851    
4852          !!!nack ('t392.1');          !!!nack ('t392.1');
4853          !!!next-token;          !!!next-token;
4854          next B;          next B;
# Line 4940  sub _tree_construction_main ($) { Line 4939  sub _tree_construction_main ($) {
4939          next B;          next B;
4940        } elsif ({        } elsif ({
4941                  caption => 1, col => 1, colgroup => 1, frame => 1,                  caption => 1, col => 1, colgroup => 1, frame => 1,
4942                  frameset => 1, head => 1,                  head => 1,
4943                  tbody => 1, td => 1, tfoot => 1, th => 1,                  tbody => 1, td => 1, tfoot => 1, th => 1,
4944                  thead => 1, tr => 1,                  thead => 1, tr => 1,
4945                 }->{$token->{tag_name}}) {                 }->{$token->{tag_name}}) {
# Line 4995  sub _tree_construction_main ($) { Line 4994  sub _tree_construction_main ($) {
4994          } elsif ({          } elsif ({
4995                    area => 1, basefont => 1, bgsound => 1, br => 1,                    area => 1, basefont => 1, bgsound => 1, br => 1,
4996                    embed => 1, img => 1, spacer => 1, wbr => 1,                    embed => 1, img => 1, spacer => 1, wbr => 1,
4997                      keygen => 1,
4998                   }->{$token->{tag_name}}) {                   }->{$token->{tag_name}}) {
4999            !!!cp ('t388.1');            !!!cp ('t388.1');
5000            pop @{$self->{open_elements}};            pop @{$self->{open_elements}};
# Line 5020  sub _tree_construction_main ($) { Line 5020  sub _tree_construction_main ($) {
5020          next B;          next B;
5021        }        }
5022      } elsif ($token->{type} == END_TAG_TOKEN) {      } elsif ($token->{type} == END_TAG_TOKEN) {
5023        if ($token->{tag_name} eq 'body') {        if ($token->{tag_name} eq 'body' or $token->{tag_name} eq 'html') {
5024          ## has a |body| element in scope  
5025            ## 1. If not "have an element in scope":
5026            ## "has a |body| element in scope"
5027          my $i;          my $i;
5028          INSCOPE: {          INSCOPE: {
5029            for (reverse @{$self->{open_elements}}) {            for (reverse @{$self->{open_elements}}) {
# Line 5035  sub _tree_construction_main ($) { Line 5037  sub _tree_construction_main ($) {
5037              }              }
5038            }            }
5039    
5040            ## NOTE: |<marquee></body>|, |<svg><foreignobject></body>|            ## NOTE: |<marquee></body>|, |<svg><foreignobject></body>|,
5041              ## and fragment cases.
5042    
5043            !!!parse-error (type => 'unmatched end tag',            !!!parse-error (type => 'unmatched end tag',
5044                            text => $token->{tag_name}, token => $token);                            text => $token->{tag_name}, token => $token);
5045            ## NOTE: Ignore the token.            ## Ignore the token.  (</body> or </html>)
5046            !!!next-token;            !!!next-token;
5047            next B;            next B;
5048          } # INSCOPE          } # INSCOPE
5049    
5050            ## 2. If unclosed elements:
5051          for (@{$self->{open_elements}}) {          for (@{$self->{open_elements}}) {
5052            unless ($_->[1] & ALL_END_TAG_OPTIONAL_EL) {            unless ($_->[1] & ALL_END_TAG_OPTIONAL_EL ||
5053                      $_->[1] == OPTGROUP_EL ||
5054                      $_->[1] == OPTION_EL ||
5055                      $_->[1] == RUBY_COMPONENT_EL) {
5056              !!!cp ('t403');              !!!cp ('t403');
5057              !!!parse-error (type => 'not closed',              !!!parse-error (type => 'not closed',
5058                              text => $_->[0]->manakai_local_name,                              text => $_->[0]->manakai_local_name,
# Line 5056  sub _tree_construction_main ($) { Line 5063  sub _tree_construction_main ($) {
5063            }            }
5064          }          }
5065    
5066            ## 3. Switch the insertion mode.
5067          $self->{insertion_mode} = AFTER_BODY_IM;          $self->{insertion_mode} = AFTER_BODY_IM;
5068          !!!next-token;          if ($token->{tag_name} eq 'body') {
         next B;  
       } elsif ($token->{tag_name} eq 'html') {  
         ## TODO: Update this code.  It seems that the code below is not  
         ## up-to-date, though it has same effect as speced.  
         if (@{$self->{open_elements}} > 1 and  
             $self->{open_elements}->[1]->[1] == BODY_EL) {  
           unless ($self->{open_elements}->[-1]->[1] == BODY_EL) {  
             !!!cp ('t406');  
             !!!parse-error (type => 'not closed',  
                             text => $self->{open_elements}->[1]->[0]  
                                 ->manakai_local_name,  
                             token => $token);  
           } else {  
             !!!cp ('t407');  
           }  
           $self->{insertion_mode} = AFTER_BODY_IM;  
           ## reprocess  
           next B;  
         } else {  
           !!!cp ('t408');  
           !!!parse-error (type => 'unmatched end tag',  
                           text => $token->{tag_name}, token => $token);  
           ## Ignore the token  
5069            !!!next-token;            !!!next-token;
5070            next B;          } else { # html
5071              ## Reprocess.
5072          }          }
5073            next B;
5074        } elsif ({        } elsif ({
5075                  ## NOTE: End tags for non-phrasing flow content elements                  ## NOTE: End tags for non-phrasing flow content elements
5076    
# Line 5091  sub _tree_construction_main ($) { Line 5078  sub _tree_construction_main ($) {
5078                  address => 1, article => 1, aside => 1, blockquote => 1,                  address => 1, article => 1, aside => 1, blockquote => 1,
5079                  center => 1, datagrid => 1, details => 1, dialog => 1,                  center => 1, datagrid => 1, details => 1, dialog => 1,
5080                  dir => 1, div => 1, dl => 1, fieldset => 1, figure => 1,                  dir => 1, div => 1, dl => 1, fieldset => 1, figure => 1,
5081                  footer => 1, header => 1, listing => 1, menu => 1, nav => 1,                  footer => 1, header => 1, hgroup => 1,
5082                    listing => 1, menu => 1, nav => 1,
5083                  ol => 1, pre => 1, section => 1, ul => 1,                  ol => 1, pre => 1, section => 1, ul => 1,
5084    
5085                  ## NOTE: As normal, but ... optional tags                  ## NOTE: As normal, but ... optional tags
# Line 5442  sub _tree_construction_main ($) { Line 5430  sub _tree_construction_main ($) {
5430    ## TODO: script stuffs    ## TODO: script stuffs
5431  } # _tree_construct_main  } # _tree_construct_main
5432    
5433    ## XXX: How this method is organized is somewhat out of date, although
5434    ## it still does what the current spec documents.
5435  sub set_inner_html ($$$$;$) {  sub set_inner_html ($$$$;$) {
5436    my $class = shift;    my $class = shift;
5437    my $node = shift;    my $node = shift; # /context/
5438    #my $s = \$_[0];    #my $s = \$_[0];
5439    my $onerror = $_[1];    my $onerror = $_[1];
5440    my $get_wrapper = $_[2] || sub ($) { return $_[0] };    my $get_wrapper = $_[2] || sub ($) { return $_[0] };
5441    
   ## ISSUE: Should {confident} be true?  
   
5442    my $nt = $node->node_type;    my $nt = $node->node_type;
5443    if ($nt == 9) {    if ($nt == 9) { # Document (invoke the algorithm with no /context/ element)
5444      # MUST      # MUST
5445            
5446      ## Step 1 # MUST      ## Step 1 # MUST
# Line 5467  sub set_inner_html ($$$$;$) { Line 5455  sub set_inner_html ($$$$;$) {
5455    
5456      ## Step 3, 4, 5 # MUST      ## Step 3, 4, 5 # MUST
5457      $class->parse_char_string ($_[0] => $node, $onerror, $get_wrapper);      $class->parse_char_string ($_[0] => $node, $onerror, $get_wrapper);
5458    } elsif ($nt == 1) {    } elsif ($nt == 1) { # Element (invoke the algorithm with /context/ element)
5459      ## TODO: If non-html element      ## TODO: If non-html element
5460    
5461      ## NOTE: Most of this code is copied from |parse_string|      ## NOTE: Most of this code is copied from |parse_string|
5462    
5463  ## TODO: Support for $get_wrapper  ## TODO: Support for $get_wrapper
5464    
5465      ## Step 1 # MUST      ## F1. Create an HTML document.
5466      my $this_doc = $node->owner_document;      my $this_doc = $node->owner_document;
5467      my $doc = $this_doc->implementation->create_document;      my $doc = $this_doc->implementation->create_document;
5468      $doc->manakai_is_html (1);      $doc->manakai_is_html (1);
5469    
5470        ## F2. Propagate quirkness flag
5471        my $node_doc = $node->owner_document;
5472        $doc->manakai_compat_mode ($node_doc->manakai_compat_mode);
5473    
5474        ## F3. Create an HTML parser
5475      my $p = $class->new;      my $p = $class->new;
5476      $p->{document} = $doc;      $p->{document} = $doc;
5477    
# Line 5605  sub set_inner_html ($$$$;$) { Line 5599  sub set_inner_html ($$$$;$) {
5599      $p->_initialize_tokenizer;      $p->_initialize_tokenizer;
5600      $p->_initialize_tree_constructor;      $p->_initialize_tree_constructor;
5601    
5602      ## Step 2      ## F4. If /context/ is not undef...
5603    
5604        ## F4.1. content model flag
5605      my $node_ln = $node->manakai_local_name;      my $node_ln = $node->manakai_local_name;
5606      $p->{content_model} = {      $p->{content_model} = {
5607        title => RCDATA_CONTENT_MODEL,        title => RCDATA_CONTENT_MODEL,
# Line 5625  sub set_inner_html ($$$$;$) { Line 5621  sub set_inner_html ($$$$;$) {
5621      $p->{inner_html_node} = [$node, $el_category->{$node_ln}];      $p->{inner_html_node} = [$node, $el_category->{$node_ln}];
5622        ## TODO: Foreign element OK?        ## TODO: Foreign element OK?
5623    
5624      ## Step 3      ## F4.2. Root |html| element
5625      my $root = $doc->create_element_ns      my $root = $doc->create_element_ns
5626        ('http://www.w3.org/1999/xhtml', [undef, 'html']);        ('http://www.w3.org/1999/xhtml', [undef, 'html']);
5627    
5628      ## Step 4 # MUST      ## F4.3.
5629      $doc->append_child ($root);      $doc->append_child ($root);
5630    
5631      ## Step 5 # MUST      ## F4.4.
5632      push @{$p->{open_elements}}, [$root, $el_category->{html}];      push @{$p->{open_elements}}, [$root, $el_category->{html}];
5633    
5634      undef $p->{head_element};      undef $p->{head_element};
5635      undef $p->{head_element_inserted};      undef $p->{head_element_inserted};
5636    
5637      ## Step 6 # MUST      ## F4.5.
5638      $p->_reset_insertion_mode;      $p->_reset_insertion_mode;
5639    
5640      ## Step 7 # MUST      ## F4.6.
5641      my $anode = $node;      my $anode = $node;
5642      AN: while (defined $anode) {      AN: while (defined $anode) {
5643        if ($anode->node_type == 1) {        if ($anode->node_type == 1) {
# Line 5656  sub set_inner_html ($$$$;$) { Line 5652  sub set_inner_html ($$$$;$) {
5652        }        }
5653        $anode = $anode->parent_node;        $anode = $anode->parent_node;
5654      } # AN      } # AN
5655        
5656      ## Step 9 # MUST      ## F.5. Set the input stream.
5657        $p->{confident} = 1; ## Confident: irrelevant.
5658    
5659        ## F.6. Start the parser.
5660      {      {
5661        my $self = $p;        my $self = $p;
5662        !!!next-token;        !!!next-token;
5663      }      }
5664      $p->_tree_construction_main;      $p->_tree_construction_main;
5665    
5666      ## Step 10 # MUST      ## F.7.
5667      my @cn = @{$node->child_nodes};      my @cn = @{$node->child_nodes};
5668      for (@cn) {      for (@cn) {
5669        $node->remove_child ($_);        $node->remove_child ($_);
Legend:
Removed from v.1.215  
changed lines
  Added in v.1.242
[email protected]
 ViewVC Help
Powered by ViewVC 1.1.24