/[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 1.224 by wakaba, Sun Aug 16 05:15:09 2009 UTC revision 1.229 by wakaba, Sat Sep 5 13:30:43 2009 UTC
# Line 911  sub _tree_construction_initial ($) { Line 911  sub _tree_construction_initial ($) {
911    
912    INITIAL: {    INITIAL: {
913      if ($token->{type} == DOCTYPE_TOKEN) {      if ($token->{type} == DOCTYPE_TOKEN) {
914        ## NOTE: Conformance checkers MAY, instead of reporting "not HTML5"        ## NOTE: Conformance checkers MAY, instead of reporting "not
915        ## error, switch to a conformance checking mode for another        ## HTML5" error, switch to a conformance checking mode for
916        ## language.        ## another language.  (We don't support such mode switchings; it
917          ## is nonsense to do anything different from what browsers do.)
918        my $doctype_name = $token->{name};        my $doctype_name = $token->{name};
919        $doctype_name = '' unless defined $doctype_name;        $doctype_name = '' unless defined $doctype_name;
920        $doctype_name =~ tr/a-z/A-Z/; # ASCII case-insensitive        my $doctype = $self->{document}->create_document_type_definition
921        if (not defined $token->{name} or # <!DOCTYPE>            ($doctype_name);
922            defined $token->{sysid}) {  
923          $doctype_name =~ tr/A-Z/a-z/; # ASCII case-insensitive
924          if ($doctype_name ne 'html') {
925          !!!cp ('t1');          !!!cp ('t1');
926          !!!parse-error (type => 'not HTML5', token => $token);          !!!parse-error (type => 'not HTML5', token => $token);
927        } elsif ($doctype_name ne 'HTML') {        } elsif (defined $token->{pubid}) {
928          !!!cp ('t2');          !!!cp ('t2');
929            ## XXX Obsolete permitted DOCTYPEs
930          !!!parse-error (type => 'not HTML5', token => $token);          !!!parse-error (type => 'not HTML5', token => $token);
931        } elsif (defined $token->{pubid}) {        } elsif (defined $token->{sysid}) {
932          if ($token->{pubid} eq 'XSLT-compat') {          if ($token->{sysid} eq 'about:legacy-compat') {
933            !!!cp ('t1.2');            !!!cp ('t1.2'); ## <!DOCTYPE HTML SYSTEM "about:legacy-compat">
934            !!!parse-error (type => 'XSLT-compat', token => $token,            !!!parse-error (type => 'XSLT-compat', token => $token,
935                            level => $self->{level}->{should});                            level => $self->{level}->{should});
936          } else {          } else {
937            !!!parse-error (type => 'not HTML5', token => $token);            !!!parse-error (type => 'not HTML5', token => $token);
938          }          }
939        } else {        } else { ## <!DOCTYPE HTML>
940          !!!cp ('t3');          !!!cp ('t3');
941          #          #
942        }        }
943                
       my $doctype = $self->{document}->create_document_type_definition  
         ($token->{name}); ## ISSUE: If name is missing (e.g. <!DOCTYPE>)?  
944        ## NOTE: Default value for both |public_id| and |system_id| attributes        ## NOTE: Default value for both |public_id| and |system_id| attributes
945        ## 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.
946        $doctype->public_id ($token->{pubid}) if defined $token->{pubid};        $doctype->public_id ($token->{pubid}) if defined $token->{pubid};
947        $doctype->system_id ($token->{sysid}) if defined $token->{sysid};        $doctype->system_id ($token->{sysid}) if defined $token->{sysid};
948    
949        ## NOTE: Other DocumentType attributes are null or empty lists.        ## NOTE: Other DocumentType attributes are null or empty lists.
950        ## In Firefox3, |internalSubset| attribute is set to the empty        ## In Firefox3, |internalSubset| attribute is set to the empty
951        ## string, while |null| is an allowed value for the attribute        ## string, while |null| is an allowed value for the attribute
952        ## according to DOM3 Core.        ## according to DOM3 Core.
953        $self->{document}->append_child ($doctype);        $self->{document}->append_child ($doctype);
954                
955        if ($token->{quirks} or $doctype_name ne 'HTML') {        if ($token->{quirks} or $doctype_name ne 'html') {
956          !!!cp ('t4');          !!!cp ('t4');
957          $self->{document}->manakai_compat_mode ('quirks');          $self->{document}->manakai_compat_mode ('quirks');
958        } elsif (defined $token->{pubid}) {        } elsif (defined $token->{pubid}) {
# Line 1435  sub _tree_construction_main ($) { Line 1438  sub _tree_construction_main ($) {
1438    }; # $script_start_tag    }; # $script_start_tag
1439    
1440    ## NOTE: $open_tables->[-1]->[0] is the "current table" element node.    ## NOTE: $open_tables->[-1]->[0] is the "current table" element node.
1441    ## NOTE: $open_tables->[-1]->[1] is the "tainted" flag.    ## NOTE: $open_tables->[-1]->[1] is the "tainted" flag (OBSOLETE; unused).
1442    ## 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.
1443    my $open_tables = [[$self->{open_elements}->[0]->[0]]];    my $open_tables = [[$self->{open_elements}->[0]->[0]]];
1444    
# Line 2265  sub _tree_construction_main ($) { Line 2268  sub _tree_construction_main ($) {
2268    
2269            ## NOTE: There is a "as if in head" code clone.            ## NOTE: There is a "as if in head" code clone.
2270            $parse_rcdata->(RCDATA_CONTENT_MODEL);            $parse_rcdata->(RCDATA_CONTENT_MODEL);
2271            ## ISSUE: A spec bug [Bug 6038]  
2272              ## NOTE: At this point the stack of open elements contain
2273              ## the |head| element (index == -2) and the |script| element
2274              ## (index == -1).  In the "after head" insertion mode the
2275              ## |head| element is inserted only for the purpose of
2276              ## providing the context for the |script| element, and
2277              ## therefore we can now and have to remove the element from
2278              ## the stack.
2279            splice @{$self->{open_elements}}, -2, 1, () # <head>            splice @{$self->{open_elements}}, -2, 1, () # <head>
2280                if ($self->{insertion_mode} & IM_MASK) == AFTER_HEAD_IM;                if ($self->{insertion_mode} & IM_MASK) == AFTER_HEAD_IM;
2281            next B;            next B;
# Line 3011  sub _tree_construction_main ($) { Line 3021  sub _tree_construction_main ($) {
3021        $insert = $insert_to_current;        $insert = $insert_to_current;
3022        #        #
3023      } elsif ($self->{insertion_mode} & TABLE_IMS) {      } elsif ($self->{insertion_mode} & TABLE_IMS) {
3024        if ($token->{type} == CHARACTER_TOKEN) {        C: {
3025          if (not $open_tables->[-1]->[1] and # tainted          my $s;
3026              $token->{data} =~ s/^([\x09\x0A\x0C\x20]+)//) {          if ($token->{type} == CHARACTER_TOKEN) {
3027            $self->{open_elements}->[-1]->[0]->manakai_append_text ($1);            !!!cp ('t194');
3028                            $self->{pending_chars} ||= [];
3029            unless (length $token->{data}) {            push @{$self->{pending_chars}}, $token;
3030              !!!cp ('t194');            !!!next-token;
3031              !!!next-token;            next B;
3032              next B;          } else {
3033              if ($self->{pending_chars}) {
3034                $s = join '', map { $_->{data} } @{$self->{pending_chars}};
3035                delete $self->{pending_chars};
3036                if ($s =~ /[^\x09\x0A\x0C\x0D\x20]/) {
3037                  !!!cp ('t195');
3038                  #
3039                } else {
3040                  !!!cp ('t195.1');
3041                  #$self->{open_elements}->[-1]->[0]->manakai_append_text ($s);
3042                  $self->{open_elements}->[-1]->[0]->append_child
3043                      ($self->{document}->create_text_node ($s));
3044                  last C;
3045                }
3046            } else {            } else {
3047              !!!cp ('t195');              !!!cp ('t195.2');
3048                last C;
3049            }            }
3050          }          }
3051    
3052            ## Foster parenting
3053          !!!parse-error (type => 'in table:#text', token => $token);          !!!parse-error (type => 'in table:#text', token => $token);
3054    
3055          ## NOTE: As if in body, but insert into the foster parent element.          ## NOTE: As if in body, but insert into the foster parent element.
# Line 3034  sub _tree_construction_main ($) { Line 3059  sub _tree_construction_main ($) {
3059            # MUST            # MUST
3060            my $foster_parent_element;            my $foster_parent_element;
3061            my $next_sibling;            my $next_sibling;
3062            my $prev_sibling;            #my $prev_sibling;
3063            OE: for (reverse 0..$#{$self->{open_elements}}) {            OE: for (reverse 0..$#{$self->{open_elements}}) {
3064              if ($self->{open_elements}->[$_]->[1] == TABLE_EL) {              if ($self->{open_elements}->[$_]->[1] == TABLE_EL) {
3065                my $parent = $self->{open_elements}->[$_]->[0]->parent_node;                my $parent = $self->{open_elements}->[$_]->[0]->parent_node;
# Line 3042  sub _tree_construction_main ($) { Line 3067  sub _tree_construction_main ($) {
3067                  $foster_parent_element = $parent;                  $foster_parent_element = $parent;
3068                  !!!cp ('t196');                  !!!cp ('t196');
3069                  $next_sibling = $self->{open_elements}->[$_]->[0];                  $next_sibling = $self->{open_elements}->[$_]->[0];
3070                  $prev_sibling = $next_sibling->previous_sibling;            #      $prev_sibling = $next_sibling->previous_sibling;
3071                  #                  #
3072                } else {                } else {
3073                  !!!cp ('t197');                  !!!cp ('t197');
3074                  $foster_parent_element = $self->{open_elements}->[$_ - 1]->[0];                  $foster_parent_element = $self->{open_elements}->[$_ - 1]->[0];
3075                  $prev_sibling = $foster_parent_element->last_child;            #      $prev_sibling = $foster_parent_element->last_child;
3076                  #                  #
3077                }                }
3078                last OE;                last OE;
3079              }              }
3080            } # OE            } # OE
3081            $foster_parent_element = $self->{open_elements}->[0]->[0] and            $foster_parent_element = $self->{open_elements}->[0]->[0] #and
3082            $prev_sibling = $foster_parent_element->last_child            #$prev_sibling = $foster_parent_element->last_child
3083                unless defined $foster_parent_element;                unless defined $foster_parent_element;
3084            undef $prev_sibling unless $open_tables->[-1]->[2]; # ~node inserted            #undef $prev_sibling unless $open_tables->[-1]->[2]; # ~node inserted
3085            if (defined $prev_sibling and            #if (defined $prev_sibling and
3086                $prev_sibling->node_type == 3) {            #    $prev_sibling->node_type == 3) {
3087              !!!cp ('t198');            #  !!! cp ('t198');
3088              $prev_sibling->manakai_append_text ($token->{data});            #  $prev_sibling->manakai_append_text ($s);
3089            } else {            #} else {
3090              !!!cp ('t199');              !!!cp ('t199');
3091              $foster_parent_element->insert_before              $foster_parent_element->insert_before
3092                  ($self->{document}->create_text_node ($token->{data}),                  ($self->{document}->create_text_node ($s), $next_sibling);
3093                   $next_sibling);            #}
           }  
3094            $open_tables->[-1]->[1] = 1; # tainted            $open_tables->[-1]->[1] = 1; # tainted
3095            $open_tables->[-1]->[2] = 1; # ~node inserted            $open_tables->[-1]->[2] = 1; # ~node inserted
3096          } else {          } else {
# Line 3077  sub _tree_construction_main ($) { Line 3101  sub _tree_construction_main ($) {
3101            ## are discarded and fragment parsing does not invoke any            ## are discarded and fragment parsing does not invoke any
3102            ## script.            ## script.
3103            !!!cp ('t200');            !!!cp ('t200');
3104            $self->{open_elements}->[-1]->[0]->manakai_append_text            $self->{open_elements}->[-1]->[0]->manakai_append_text ($s);
               ($token->{data});  
3105          }          }
3106                      } # C
3107          !!!next-token;  
3108          next B;        if ($token->{type} == START_TAG_TOKEN) {
       } elsif ($token->{type} == START_TAG_TOKEN) {  
3109          if ({          if ({
3110               tr => (($self->{insertion_mode} & IM_MASK) != IN_ROW_IM),               tr => (($self->{insertion_mode} & IM_MASK) != IN_ROW_IM),
3111               th => 1, td => 1,               th => 1, td => 1,
# Line 3907  sub _tree_construction_main ($) { Line 3929  sub _tree_construction_main ($) {
3929              ## Reprocess the token.              ## Reprocess the token.
3930              next B;              next B;
3931            }            }
3932            } elsif ($token->{tag_name} eq 'script') {
3933              !!!cp ('t281.3');
3934              ## NOTE: This is an "as if in head" code clone
3935              $script_start_tag->();
3936              next B;
3937          } else {          } else {
3938            !!!cp ('t282');            !!!cp ('t282');
3939            !!!parse-error (type => 'in select',            !!!parse-error (type => 'in select',
# Line 4426  sub _tree_construction_main ($) { Line 4453  sub _tree_construction_main ($) {
4453                  table => 1,                  table => 1,
4454                  hr => 1,                  hr => 1,
4455                 }->{$token->{tag_name}}) {                 }->{$token->{tag_name}}) {
4456    
4457            ## 1. When there is an opening |form| element:
4458          if ($token->{tag_name} eq 'form' and defined $self->{form_element}) {          if ($token->{tag_name} eq 'form' and defined $self->{form_element}) {
4459            !!!cp ('t350');            !!!cp ('t350');
4460            !!!parse-error (type => 'in form:form', token => $token);            !!!parse-error (type => 'in form:form', token => $token);
# Line 4435  sub _tree_construction_main ($) { Line 4464  sub _tree_construction_main ($) {
4464            next B;            next B;
4465          }          }
4466    
4467            ## 2. Close the |p| element, if any.
4468          if ($token->{tag_name} ne 'table' or # The Hixie Quirk          if ($token->{tag_name} ne 'table' or # The Hixie Quirk
4469              $self->{document}->manakai_compat_mode ne 'quirks') {              $self->{document}->manakai_compat_mode ne 'quirks') {
4470            ## has a p element in scope            ## has a p element in scope
# Line 4451  sub _tree_construction_main ($) { Line 4481  sub _tree_construction_main ($) {
4481              }              }
4482            } # INSCOPE            } # INSCOPE
4483          }          }
4484              
4485            ## 3. Close the opening <hn> element, if any.
4486            if ({h1 => 1, h2 => 1, h3 => 1,
4487                 h4 => 1, h5 => 1, h6 => 1}->{$token->{tag_name}}) {
4488              if ($self->{open_elements}->[-1]->[1] == HEADING_EL) {
4489                !!!parse-error (type => 'not closed',
4490                                text => $self->{open_elements}->[-1]->[0]->manakai_local_name,
4491                                token => $token);
4492                pop @{$self->{open_elements}};
4493              }
4494            }
4495    
4496            ## 4. Insertion.
4497          !!!insert-element-t ($token->{tag_name}, $token->{attributes}, $token);          !!!insert-element-t ($token->{tag_name}, $token->{attributes}, $token);
4498          if ($token->{tag_name} eq 'pre' or $token->{tag_name} eq 'listing') {          if ($token->{tag_name} eq 'pre' or $token->{tag_name} eq 'listing') {
4499            !!!nack ('t346.1');            !!!nack ('t346.1');
# Line 4495  sub _tree_construction_main ($) { Line 4537  sub _tree_construction_main ($) {
4537        } elsif ($token->{tag_name} eq 'li') {        } elsif ($token->{tag_name} eq 'li') {
4538          ## NOTE: As normal, but imply </li> when there's another <li> ...          ## NOTE: As normal, but imply </li> when there's another <li> ...
4539    
4540          ## NOTE: Special, Scope (<li><foo><li> == <li><foo><li/></foo></li>)          ## NOTE: Special, Scope (<li><foo><li> == <li><foo><li/></foo></li>)::
4541            ## Interpreted as <li><foo/></li><li/> (non-conforming)            ## Interpreted as <li><foo/></li><li/> (non-conforming):
4542            ## blockquote (O9.27), center (O), dd (Fx3, O, S3.1.2, IE7),            ## blockquote (O9.27), center (O), dd (Fx3, O, S3.1.2, IE7),
4543            ## 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),
4544            ## 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),
4545            ## object (Fx)            ## object (Fx)
4546            ## Generate non-tree (non-conforming)            ## Generate non-tree (non-conforming):
4547            ## basefont (IE7 (where basefont is non-void)), center (IE),            ## basefont (IE7 (where basefont is non-void)), center (IE),
4548            ## form (IE), hn (IE)            ## form (IE), hn (IE)
4549          ## address, div, p (<li><foo><li> == <li><foo/></li><li/>)          ## address, div, p (<li><foo><li> == <li><foo/></li><li/>)::
4550            ## Interpreted as <li><foo><li/></foo></li> (non-conforming)            ## Interpreted as <li><foo><li/></foo></li> (non-conforming):
4551            ## div (Fx, S)            ## div (Fx, S)
4552    
4553          my $non_optional;          my $non_optional;
# Line 5041  sub _tree_construction_main ($) { Line 5083  sub _tree_construction_main ($) {
5083        }        }
5084      } elsif ($token->{type} == END_TAG_TOKEN) {      } elsif ($token->{type} == END_TAG_TOKEN) {
5085        if ($token->{tag_name} eq 'body') {        if ($token->{tag_name} eq 'body') {
5086          ## has a |body| element in scope  
5087            ## 1. If not "have an element in scope":
5088            ## "has a |body| element in scope"
5089          my $i;          my $i;
5090          INSCOPE: {          INSCOPE: {
5091            for (reverse @{$self->{open_elements}}) {            for (reverse @{$self->{open_elements}}) {
# Line 5064  sub _tree_construction_main ($) { Line 5108  sub _tree_construction_main ($) {
5108            next B;            next B;
5109          } # INSCOPE          } # INSCOPE
5110    
5111            ## 2. If unclosed elements:
5112          for (@{$self->{open_elements}}) {          for (@{$self->{open_elements}}) {
5113            unless ($_->[1] & ALL_END_TAG_OPTIONAL_EL ||            unless ($_->[1] & ALL_END_TAG_OPTIONAL_EL ||
5114                    $_->[1] == OPTGROUP_EL ||                    $_->[1] == OPTGROUP_EL ||
# Line 5079  sub _tree_construction_main ($) { Line 5124  sub _tree_construction_main ($) {
5124            }            }
5125          }          }
5126    
5127            ## 3. Switch the insertion mode.
5128          $self->{insertion_mode} = AFTER_BODY_IM;          $self->{insertion_mode} = AFTER_BODY_IM;
5129          !!!next-token;          !!!next-token;
5130          next B;          next B;
Legend:
Removed from v.1.224  
changed lines
  Added in v.1.229
[email protected]
 ViewVC Help
Powered by ViewVC 1.1.24