/[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 1.58 by wakaba, Tue Sep 4 11:19:07 2007 UTC revision 1.62 by wakaba, Sun Nov 11 04:59:35 2007 UTC
# Line 147  sub new ($) { Line 147  sub new ($) {
147    $self->{parse_error} = sub {    $self->{parse_error} = sub {
148      #      #
149    };    };
150      $self->{application_cache_selection} = sub {
151        #
152      };
153    return $self;    return $self;
154  } # new  } # new
155    
# Line 263  sub _initialize_tokenizer ($) { Line 266  sub _initialize_tokenizer ($) {
266  ## has completed loading.  If one has, then it MUST be executed  ## has completed loading.  If one has, then it MUST be executed
267  ## and removed from the list.  ## and removed from the list.
268    
269    ## NOTE: HTML5 "Writing HTML documents" section, applied to
270    ## documents and not to user agents and conformance checkers,
271    ## contains some requirements that are not detected by the
272    ## parsing algorithm:
273    ## - Some requirements on character encoding declarations. ## TODO
274    ## - "Elements MUST NOT contain content that their content model disallows."
275    ##   ... Some are parse error, some are not (will be reported by c.c.).
276    ## - Polytheistic slash SHOULD NOT be used. (Applied only to atheists.) ## TODO
277    ## - Text (in elements, attributes, and comments) SHOULD NOT contain
278    ##   control characters other than space characters. ## TODO: (what is control character? C0, C1 and DEL?  Unicode control character?)
279    
280    ## TODO: HTML5 poses authors two SHOULD-level requirements that cannot
281    ## be detected by the HTML5 parsing algorithm:
282    ## - Text,
283    
284  sub _get_next_token ($) {  sub _get_next_token ($) {
285    my $self = shift;    my $self = shift;
286    if (@{$self->{token}}) {    if (@{$self->{token}}) {
# Line 2080  sub _tree_construction_root_element ($) Line 2098  sub _tree_construction_root_element ($)
2098              redo B;              redo B;
2099            }            }
2100          }          }
2101    
2102            $self->{application_cache_selection}->(undef);
2103    
2104            #
2105          } elsif ($token->{type} == START_TAG_TOKEN) {
2106            if ($token->{tag_name} eq 'html' and
2107                $token->{attributes}->{manifest}) { ## ISSUE: Spec spells as "application"
2108              $self->{application_cache_selection}
2109                   ->($token->{attributes}->{manifest}->{value});
2110              ## ISSUE: No relative reference resolution?
2111            } else {
2112              $self->{application_cache_selection}->(undef);
2113            }
2114    
2115            ## ISSUE: There is an issue in the spec
2116          #          #
2117        } elsif ({        } elsif ({
                 START_TAG_TOKEN, 1,  
2118                  END_TAG_TOKEN, 1,                  END_TAG_TOKEN, 1,
2119                  END_OF_FILE_TOKEN, 1,                  END_OF_FILE_TOKEN, 1,
2120                 }->{$token->{type}}) {                 }->{$token->{type}}) {
2121            $self->{application_cache_selection}->(undef);
2122    
2123          ## ISSUE: There is an issue in the spec          ## ISSUE: There is an issue in the spec
2124          #          #
2125        } else {        } else {
2126          die "$0: $token->{type}: Unknown token type";          die "$0: $token->{type}: Unknown token type";
2127        }        }
2128    
2129        my $root_element; !!!create-element ($root_element, 'html');        my $root_element; !!!create-element ($root_element, 'html');
2130        $self->{document}->append_child ($root_element);        $self->{document}->append_child ($root_element);
2131        push @{$self->{open_elements}}, [$root_element, 'html'];        push @{$self->{open_elements}}, [$root_element, 'html'];
# Line 5331  sub set_inner_html ($$$) { Line 5366  sub set_inner_html ($$$) {
5366    
5367  } # tree construction stage  } # tree construction stage
5368    
 sub get_inner_html ($$$) {  
   my (undef, $node, $on_error) = @_;  
   
   ## Step 1  
   my $s = '';  
   
   my $in_cdata;  
   my $parent = $node;  
   while (defined $parent) {  
     if ($parent->node_type == 1 and  
         $parent->namespace_uri eq 'http://www.w3.org/1999/xhtml' and  
         {  
           style => 1, script => 1, xmp => 1, iframe => 1,  
           noembed => 1, noframes => 1, noscript => 1,  
         }->{$parent->local_name}) { ## TODO: case thingy  
       $in_cdata = 1;  
     }  
     $parent = $parent->parent_node;  
   }  
   
   ## Step 2  
   my @node = @{$node->child_nodes};  
   C: while (@node) {  
     my $child = shift @node;  
     unless (ref $child) {  
       if ($child eq 'cdata-out') {  
         $in_cdata = 0;  
       } else {  
         $s .= $child; # end tag  
       }  
       next C;  
     }  
       
     my $nt = $child->node_type;  
     if ($nt == 1) { # Element  
       my $tag_name = $child->tag_name; ## TODO: manakai_tag_name  
       $s .= '<' . $tag_name;  
       ## NOTE: Non-HTML case:  
       ## <http://permalink.gmane.org/gmane.org.w3c.whatwg.discuss/11191>  
   
       my @attrs = @{$child->attributes}; # sort order MUST be stable  
       for my $attr (@attrs) { # order is implementation dependent  
         my $attr_name = $attr->name; ## TODO: manakai_name  
         $s .= ' ' . $attr_name . '="';  
         my $attr_value = $attr->value;  
         ## escape  
         $attr_value =~ s/&/&amp;/g;  
         $attr_value =~ s/</&lt;/g;  
         $attr_value =~ s/>/&gt;/g;  
         $attr_value =~ s/"/&quot;/g;  
         $s .= $attr_value . '"';  
       }  
       $s .= '>';  
         
       next C if {  
         area => 1, base => 1, basefont => 1, bgsound => 1,  
         br => 1, col => 1, embed => 1, frame => 1, hr => 1,  
         img => 1, input => 1, link => 1, meta => 1, param => 1,  
         spacer => 1, wbr => 1,  
       }->{$tag_name};  
   
       $s .= "\x0A" if $tag_name eq 'pre' or $tag_name eq 'textarea';  
   
       if (not $in_cdata and {  
         style => 1, script => 1, xmp => 1, iframe => 1,  
         noembed => 1, noframes => 1, noscript => 1,  
         plaintext => 1,  
       }->{$tag_name}) {  
         unshift @node, 'cdata-out';  
         $in_cdata = 1;  
       }  
   
       unshift @node, @{$child->child_nodes}, '</' . $tag_name . '>';  
     } elsif ($nt == 3 or $nt == 4) {  
       if ($in_cdata) {  
         $s .= $child->data;  
       } else {  
         my $value = $child->data;  
         $value =~ s/&/&amp;/g;  
         $value =~ s/</&lt;/g;  
         $value =~ s/>/&gt;/g;  
         $value =~ s/"/&quot;/g;  
         $s .= $value;  
       }  
     } elsif ($nt == 8) {  
       $s .= '<!--' . $child->data . '-->';  
     } elsif ($nt == 10) {  
       $s .= '<!DOCTYPE ' . $child->name . '>';  
     } elsif ($nt == 5) { # entrefs  
       push @node, @{$child->child_nodes};  
     } else {  
       $on_error->($child) if defined $on_error;  
     }  
     ## ISSUE: This code does not support PIs.  
   } # C  
     
   ## Step 3  
   return \$s;  
 } # get_inner_html  
   
5369  1;  1;
5370  # $Date$  # $Date$
Legend:
Removed from v.1.58  
changed lines
  Added in v.1.62
[email protected]
 ViewVC Help
Powered by ViewVC 1.1.24