/[suikacvs]/markup/html/whatpm/Whatpm/ContentChecker.pm
Suika

Diff of /markup/html/whatpm/Whatpm/ContentChecker.pm

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

revision 1.97 by wakaba, Sun Sep 21 09:45:02 2008 UTC revision 1.109 by wakaba, Sat Aug 22 09:35:35 2009 UTC
# Line 121  our $AttrChecker = { Line 121  our $AttrChecker = {
121        ## no author requirement on conformance in the XML Base specification.        ## no author requirement on conformance in the XML Base specification.
122      },      },
123      id => sub {      id => sub {
124        my ($self, $attr) = @_;        my ($self, $attr, $item, $element_state) = @_;
125        my $value = $attr->value;        my $value = $attr->value;
126        $value =~ s/[\x09\x0A\x0D\x20]+/ /g;        $value =~ s/[\x09\x0A\x0D\x20]+/ /g;
127        $value =~ s/^\x20//;        $value =~ s/^\x20//;
# Line 135  our $AttrChecker = { Line 135  our $AttrChecker = {
135          push @{$self->{id}->{$value}}, $attr;          push @{$self->{id}->{$value}}, $attr;
136        } else {        } else {
137          $self->{id}->{$value} = [$attr];          $self->{id}->{$value} = [$attr];
138            $self->{id_type}->{$value} = $element_state->{id_type} || '';
139        }        }
140      },      },
141    },    },
# Line 220  $AttrStatus->{$XMLNS_NS}->{''} = FEATURE Line 221  $AttrStatus->{$XMLNS_NS}->{''} = FEATURE
221  ## TODO: xsi:schemaLocation for XHTML2 support (very, very low priority)  ## TODO: xsi:schemaLocation for XHTML2 support (very, very low priority)
222    
223  our %AnyChecker = (  our %AnyChecker = (
224      ## NOTE: |check_start| is invoked before anything on the element's
225      ## attributes and contents is checked.
226    check_start => sub { },    check_start => sub { },
227      ## NOTE: |check_attrs| and |check_attrs2| are invoked after
228      ## |check_start| and before anything on the element's contents is
229      ## checked.  |check_attrs| is invoked immediately before
230      ## |check_attrs2|.
231    check_attrs => sub {    check_attrs => sub {
232      my ($self, $item, $element_state) = @_;      my ($self, $item, $element_state) = @_;
233      for my $attr (@{$item->{node}->attributes}) {      for my $attr (@{$item->{node}->attributes}) {
# Line 255  our %AnyChecker = ( Line 262  our %AnyChecker = (
262        $self->_attr_status_info ($attr, $status);        $self->_attr_status_info ($attr, $status);
263      }      }
264    },    },
265      check_attrs2 => sub { },
266      ## NOTE: |check_child_element| is invoked for each occurence of
267      ## child elements.  It is invoked after |check_attrs| and before
268      ## |check_end|.  |check_child_element| and |check_child_text| are
269      ## invoked for each child elements and text nodes in tree order.
270    check_child_element => sub {    check_child_element => sub {
271      my ($self, $item, $child_el, $child_nsuri, $child_ln,      my ($self, $item, $child_el, $child_nsuri, $child_ln,
272          $child_is_transparent, $element_state) = @_;          $child_is_transparent, $element_state) = @_;
# Line 268  our %AnyChecker = ( Line 280  our %AnyChecker = (
280        #        #
281      }      }
282    },    },
283      ## NOTE: |check_child_text| is invoked for each occurence of child
284      ## text nodes.  It is invoked after |check_attrs| and before
285      ## |check_end|.  |check_child_element| and |check_child_text| are
286      ## invoked for each child elements and text nodes in tree order.
287    check_child_text => sub { },    check_child_text => sub { },
288      ## NOTE: |check_end| is invoked after everything on the element's
289      ## attributes and contents are checked.
290    check_end => sub {    check_end => sub {
291      my ($self, $item, $element_state) = @_;      my ($self, $item, $element_state) = @_;
292      ## NOTE: There is a modified copy of the code below for |html:ruby|.      ## NOTE: There is a modified copy of the code below for |html:ruby|.
# Line 311  our $IsInHTMLInteractiveContent = sub { Line 329  our $IsInHTMLInteractiveContent = sub {
329    ## Whatpm::ContentChecler::HTML for the list of all (conditionally    ## Whatpm::ContentChecler::HTML for the list of all (conditionally
330    ## or permanently) interactive content.    ## or permanently) interactive content.
331    
332      ## The variable name is not good, since this method also returns
333      ## true for non-interactive content as long as the element cannot be
334      ## interactive content.
335    
336    if ($nsuri eq $HTML_NS and ($ln eq 'video' or $ln eq 'audio')) {    if ($nsuri eq $HTML_NS and ($ln eq 'video' or $ln eq 'audio')) {
337      return $el->has_attribute ('controls');      return $el->has_attribute_ns (undef, 'controls');
338    } elsif ($nsuri eq $HTML_NS and $ln eq 'menu') {    } elsif ($nsuri eq $HTML_NS and $ln eq 'menu') {
339      my $value = $el->get_attribute ('type');      my $value = $el->get_attribute_ns (undef, 'type');
340      $value =~ tr/A-Z/a-z/; # ASCII case-insensitive      $value =~ tr/A-Z/a-z/; # ASCII case-insensitive
341      return ($value eq 'toolbar');      return ($value eq 'toolbar');
342    } else {    } else {
343      return 1;      return 1;
344    }    }
345    
346      ## TODO: input:not([type=hidden])
347  }; # $IsInHTMLInteractiveContent  }; # $IsInHTMLInteractiveContent
348    
349  my $HTMLTransparentElements = {  my $HTMLTransparentElements = {
350    $HTML_NS => {qw/ins 1 del 1 font 1 noscript 1 canvas 1 a 1/},    $HTML_NS => {
351    ## NOTE: |html:noscript| is transparent if scripting is disabled      ins => 1, del => 1,
352    ## and not in |head|.      font => 1, ## dropped from the spec
353  };      noscript => 1,
354          ## NOTE: |html:noscript| is transparent if scripting is disabled
355          ## and not in |head|.
356        canvas => 1,
357        a => 1,
358        map => 1,
359      },
360    }; # $HTMLTransparentElements
361    
362    ## NOTE: Now that the term "semi-transparent content model" is dropped
363    ## from the spec, but the concept is not.
364  my $HTMLSemiTransparentElements = {  my $HTMLSemiTransparentElements = {
365    $HTML_NS => {object => 1, video => 1, audio => 1},    $HTML_NS => {object => 1, video => 1, audio => 1},
366  };  }; # $HTMLSemiTransparentElements
367    
368  our $Element = {};  our $Element = {};
369    
# Line 390  my $default_error_level = { Line 423  my $default_error_level = {
423    
424    html4_fact => 'm',    html4_fact => 'm',
425    html5_no_may => 'm',    html5_no_may => 'm',
426      html5_fact => 'm',
427    
428    xml_error => 'm', ## TODO: correct?    xml_error => 'm', ## TODO: correct?
429    xml_id_error => 'm', ## TODO: ?    xml_id_error => 'm', ## TODO: ?
# Line 487  sub check_document ($$$;$) { Line 521  sub check_document ($$$;$) {
521      if ($doc->manakai_is_html) {      if ($doc->manakai_is_html) {
522        if (not $doc->manakai_has_bom and        if (not $doc->manakai_has_bom and
523            not defined $doc->manakai_charset) {            not defined $doc->manakai_charset) {
524          unless ($charset->{is_html_ascii_superset}) {          unless ($charset->{category}
525                      & Message::Charset::Info::CHARSET_CATEGORY_ASCII_COMPAT ()) {
526            $onerror->(node => $doc,            $onerror->(node => $doc,
527                       level => $self->{level}->{must},                       level => $self->{level}->{must},
528                       type => 'non ascii superset',                       type => 'non ascii superset',
# Line 515  sub check_document ($$$;$) { Line 550  sub check_document ($$$;$) {
550                     level => $self->{level}->{should},                     level => $self->{level}->{should},
551                     layer => 'encode');                     layer => 'encode');
552        } elsif ($charset->{iana_names}->{'cesu-8'} or        } elsif ($charset->{iana_names}->{'cesu-8'} or
553                 $charset->{iana_names}->{'utf-8'} or ## ISSUE: UNICODE-1-1-UTF-7?                 $charset->{iana_names}->{'utf-7'} or ## ISSUE: UNICODE-1-1-UTF-7?
554                 $charset->{iana_names}->{'bocu-1'} or                 $charset->{iana_names}->{'bocu-1'} or
555                 $charset->{iana_names}->{'scsu'}) {                 $charset->{iana_names}->{'scsu'}) {
556          $onerror->(node => $doc,          $onerror->(node => $doc,
# Line 558  sub check_element ($$$;$) { Line 593  sub check_element ($$$;$) {
593    $self->{plus_elements} = {};    $self->{plus_elements} = {};
594    $self->{minus_elements} = {};    $self->{minus_elements} = {};
595    $self->{id} = {};    $self->{id} = {};
596    $self->{form} = {};    $self->{id_type} = {}; # 'form' / 'labelable' / 'menu'
597      $self->{form} = {}; # form/@name
598      #$self->{has_autofocus};
599      $self->{idref} = []; # @form, @for, @contextmenu
600    $self->{term} = {};    $self->{term} = {};
601    $self->{usemap} = [];    $self->{usemap} = [];
602    $self->{ref} = []; # datetemplate data references    $self->{ref} = []; # datetemplate data references
603    $self->{template} = []; # datatemplate template references    $self->{template} = []; # datatemplate template references
   $self->{contextmenu} = [];  
604    $self->{map} = {};    $self->{map} = {};
   $self->{menu} = {};  
605    $self->{has_link_type} = {};    $self->{has_link_type} = {};
606    $self->{flag} = {};    $self->{flag} = {};
607    #$self->{has_uri_attr};    #$self->{has_uri_attr};
# Line 635  next unless $code;## TODO: temp. Line 671  next unless $code;## TODO: temp.
671        my @new_item;        my @new_item;
672        push @new_item, [$eldef->{check_start}, $self, $item, $element_state];        push @new_item, [$eldef->{check_start}, $self, $item, $element_state];
673        push @new_item, [$eldef->{check_attrs}, $self, $item, $element_state];        push @new_item, [$eldef->{check_attrs}, $self, $item, $element_state];
674          push @new_item, [$eldef->{check_attrs2}, $self, $item, $element_state];
675                
676        my @child = @{$item->{node}->child_nodes};        my @child = @{$item->{node}->child_nodes};
677        while (@child) {        while (@child) {
# Line 777  next unless $code;## TODO: temp. Line 814  next unless $code;## TODO: temp.
814      }      }
815    }    }
816    
817    for (@{$self->{contextmenu}}) {    for (@{$self->{idref}}) {
818      unless ($self->{menu}->{$_->[0]}) {      if ($self->{id}->{$_->[1]} and $self->{id_type}->{$_->[1]} eq $_->[0]) {
819        $self->{onerror}->(node => $_->[1], type => 'no referenced menu',        #
820        } elsif ($_->[0] eq 'any' and $self->{id}->{$_->[1]}) {
821          #
822        } else {
823          $self->{onerror}->(node => $_->[2],
824                             type => {
825                                      any => 'no referenced element', ## TODOC: type
826                                      form => 'no referenced form',
827                                      labelable => 'no referenced control',
828                                      menu => 'no referenced menu',
829                                      datalist => 'no referenced datalist', ## TODOC: type
830                                     }->{$_->[0]},
831                             value => $_->[1],
832                           level => $self->{level}->{must});                           level => $self->{level}->{must});
833      }      }
834    }    }
# Line 788  next unless $code;## TODO: temp. Line 837  next unless $code;## TODO: temp.
837    delete $self->{minus_elements};    delete $self->{minus_elements};
838    delete $self->{onerror};    delete $self->{onerror};
839    delete $self->{id};    delete $self->{id};
840      delete $self->{id_type};
841    delete $self->{form};    delete $self->{form};
842      delete $self->{has_autofocus};
843      delete $self->{idref};
844    delete $self->{usemap};    delete $self->{usemap};
845    delete $self->{ref};    delete $self->{ref};
846    delete $self->{template};    delete $self->{template};
Legend:
Removed from v.1.97  
changed lines
  Added in v.1.109
[email protected]
 ViewVC Help
Powered by ViewVC 1.1.24