/[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 1.196 by wakaba, Sat Oct 4 07:58:58 2008 UTC revision 1.201 by wakaba, Sat Oct 4 12:20:35 2008 UTC
# Line 86  sub ALL_END_TAG_OPTIONAL_EL () { Line 86  sub ALL_END_TAG_OPTIONAL_EL () {
86    LI_EL |    LI_EL |
87    P_EL |    P_EL |
88    
89      ## ISSUE: option, optgroup, rt, rp?
90    
91    BODY_EL |    BODY_EL |
92    HTML_EL |    HTML_EL |
93    TABLE_CELL_EL |    TABLE_CELL_EL |
# Line 130  sub SPECIAL_EL () { Line 132  sub SPECIAL_EL () {
132    FORM_EL |    FORM_EL |
133    FRAMESET_EL |    FRAMESET_EL |
134    HEADING_EL |    HEADING_EL |
   OPTION_EL |  
   OPTGROUP_EL |  
135    SELECT_EL |    SELECT_EL |
136    TABLE_ROW_EL |    TABLE_ROW_EL |
137    TABLE_ROW_GROUP_EL |    TABLE_ROW_GROUP_EL |
# Line 249  my $el_category_f = { Line 249  my $el_category_f = {
249      mtext => FOREIGN_FLOW_CONTENT_EL,      mtext => FOREIGN_FLOW_CONTENT_EL,
250    },    },
251    $SVG_NS => {    $SVG_NS => {
252      foreignObject => FOREIGN_FLOW_CONTENT_EL,      foreignObject => FOREIGN_FLOW_CONTENT_EL | MISC_SCOPING_EL,
253      desc => FOREIGN_FLOW_CONTENT_EL,      desc => FOREIGN_FLOW_CONTENT_EL,
254      title => FOREIGN_FLOW_CONTENT_EL,      title => FOREIGN_FLOW_CONTENT_EL,
255    },    },
# Line 488  sub parse_byte_stream ($$$$;$$) { Line 488  sub parse_byte_stream ($$$$;$$) {
488      if (defined $charset_name) {      if (defined $charset_name) {
489        $charset = Message::Charset::Info->get_by_html_name ($charset_name);        $charset = Message::Charset::Info->get_by_html_name ($charset_name);
490    
       ## ISSUE: Unsupported encoding is not ignored according to the spec.  
491        require Whatpm::Charset::DecodeHandle;        require Whatpm::Charset::DecodeHandle;
492        $buffer = Whatpm::Charset::DecodeHandle::ByteBuffer->new        $buffer = Whatpm::Charset::DecodeHandle::ByteBuffer->new
493            ($byte_stream);            ($byte_stream);
# Line 2485  sub _get_next_token ($) { Line 2484  sub _get_next_token ($) {
2484          !!!cp (160);          !!!cp (160);
2485          $self->{ct}->{name} = chr $self->{nc};          $self->{ct}->{name} = chr $self->{nc};
2486          delete $self->{ct}->{quirks};          delete $self->{ct}->{quirks};
 ## ISSUE: "Set the token's name name to the" in the spec  
2487          $self->{state} = DOCTYPE_NAME_STATE;          $self->{state} = DOCTYPE_NAME_STATE;
2488          !!!next-input-character;          !!!next-input-character;
2489          redo A;          redo A;
# Line 3785  sub _tree_construction_root_element ($) Line 3783  sub _tree_construction_root_element ($)
3783      ## NOTE: Reprocess the token.      ## NOTE: Reprocess the token.
3784      !!!ack-later;      !!!ack-later;
3785      return; ## Go to the "before head" insertion mode.      return; ## Go to the "before head" insertion mode.
   
     ## ISSUE: There is an issue in the spec  
3786    } # B    } # B
3787    
3788    die "$0: _tree_construction_root_element: This should never be reached";    die "$0: _tree_construction_root_element: This should never be reached";
# Line 4522  sub _tree_construction_main ($) { Line 4518  sub _tree_construction_main ($) {
4518          pop @{$self->{open_elements}}          pop @{$self->{open_elements}}
4519              while $self->{open_elements}->[-1]->[1] & FOREIGN_EL;              while $self->{open_elements}->[-1]->[1] & FOREIGN_EL;
4520    
4521            ## NOTE: |<span><svg>| ... two parse errors, |<svg>| ... a parse error.
4522    
4523          $self->{insertion_mode} &= ~ IN_FOREIGN_CONTENT_IM;          $self->{insertion_mode} &= ~ IN_FOREIGN_CONTENT_IM;
4524          ## Reprocess.          ## Reprocess.
4525          next B;          next B;
# Line 4715  sub _tree_construction_main ($) { Line 4713  sub _tree_construction_main ($) {
4713                  !!!cp ('t105');                  !!!cp ('t105');
4714                }                }
4715                !!!insert-element ($token->{tag_name}, $token->{attributes}, $token);                !!!insert-element ($token->{tag_name}, $token->{attributes}, $token);
4716                my $meta_el = pop @{$self->{open_elements}}; ## ISSUE: This step is missing in the spec.                my $meta_el = pop @{$self->{open_elements}};
4717    
4718                unless ($self->{confident}) {                unless ($self->{confident}) {
4719                  if ($token->{attributes}->{charset}) {                  if ($token->{attributes}->{charset}) {
# Line 5172  sub _tree_construction_main ($) { Line 5170  sub _tree_construction_main ($) {
5170        } else {        } else {
5171          die "$0: $token->{type}: Unknown token type";          die "$0: $token->{type}: Unknown token type";
5172        }        }
   
           ## ISSUE: An issue in the spec.  
5173      } elsif ($self->{insertion_mode} & BODY_IMS) {      } elsif ($self->{insertion_mode} & BODY_IMS) {
5174            if ($token->{type} == CHARACTER_TOKEN) {            if ($token->{type} == CHARACTER_TOKEN) {
5175              !!!cp ('t150');              !!!cp ('t150');
# Line 6801  sub _tree_construction_main ($) { Line 6797  sub _tree_construction_main ($) {
6797        } else {        } else {
6798          die "$0: $token->{type}: Unknown token type";          die "$0: $token->{type}: Unknown token type";
6799        }        }
   
       ## ISSUE: An issue in spec here  
6800      } else {      } else {
6801        die "$0: $self->{insertion_mode}: Unknown insertion mode";        die "$0: $self->{insertion_mode}: Unknown insertion mode";
6802      }      }
# Line 7064  sub _tree_construction_main ($) { Line 7058  sub _tree_construction_main ($) {
7058          }          }
7059    
7060          ## 5. (a) has a |p| element in scope          ## 5. (a) has a |p| element in scope
         ## ISSUE: Is this step really necessary?  
7061          INSCOPE: for (reverse @{$self->{open_elements}}) {          INSCOPE: for (reverse @{$self->{open_elements}}) {
7062            if ($_->[1] & P_EL) {            if ($_->[1] & P_EL) {
7063              !!!cp ('t353');              !!!cp ('t353');
7064    
7065                ## NOTE: |<p><li>|, for example.
7066    
7067              !!!back-token; # <x>              !!!back-token; # <x>
7068              $token = {type => END_TAG_TOKEN, tag_name => 'p',              $token = {type => END_TAG_TOKEN, tag_name => 'p',
7069                        line => $token->{line}, column => $token->{column}};                        line => $token->{line}, column => $token->{column}};
# Line 7144  sub _tree_construction_main ($) { Line 7140  sub _tree_construction_main ($) {
7140          }          }
7141    
7142          ## 5. (a) has a |p| element in scope          ## 5. (a) has a |p| element in scope
         ## ISSUE: Is this step really necessary?  
7143          INSCOPE: for (reverse @{$self->{open_elements}}) {          INSCOPE: for (reverse @{$self->{open_elements}}) {
7144            if ($_->[1] & P_EL) {            if ($_->[1] & P_EL) {
7145              !!!cp ('t353.1');              !!!cp ('t353.1');
# Line 7403  sub _tree_construction_main ($) { Line 7398  sub _tree_construction_main ($) {
7398          }          }
7399          !!!next-token;          !!!next-token;
7400          next B;          next B;
7401          } elsif ($token->{tag_name} eq 'optgroup' or
7402                   $token->{tag_name} eq 'option') {
7403            ## has an |option| element in scope
7404            INSCOPE: for (reverse 0..$#{$self->{open_elements}}) {
7405              my $node = $self->{open_elements}->[$_];
7406              if ($node->[1] & OPTION_EL) {
7407                !!!cp ('t397.1');
7408                ## NOTE: As if </option>
7409                !!!back-token; # <option> or <optgroup>
7410                $token = {type => END_TAG_TOKEN, tag_name => 'option',
7411                          line => $token->{line}, column => $token->{column}};
7412                next B;
7413              } elsif ($node->[1] & SCOPING_EL) {
7414                !!!cp ('t397.2');
7415                last INSCOPE;
7416              }
7417            } # INSCOPE
7418    
7419            $reconstruct_active_formatting_elements->($insert_to_current);
7420    
7421            !!!insert-element-t ($token->{tag_name}, $token->{attributes}, $token);
7422    
7423            !!!nack ('t397.3');
7424            !!!next-token;
7425            redo B;
7426        } elsif ($token->{tag_name} eq 'rt' or        } elsif ($token->{tag_name} eq 'rt' or
7427                 $token->{tag_name} eq 'rp') {                 $token->{tag_name} eq 'rp') {
7428          ## has a |ruby| element in scope          ## has a |ruby| element in scope
# Line 7450  sub _tree_construction_main ($) { Line 7470  sub _tree_construction_main ($) {
7470                    
7471          if ($self->{self_closing}) {          if ($self->{self_closing}) {
7472            pop @{$self->{open_elements}};            pop @{$self->{open_elements}};
7473            !!!ack ('t398.1');            !!!ack ('t398.6');
7474          } else {          } else {
7475            !!!cp ('t398.2');            !!!cp ('t398.7');
7476            $self->{insertion_mode} |= IN_FOREIGN_CONTENT_IM;            $self->{insertion_mode} |= IN_FOREIGN_CONTENT_IM;
7477            ## NOTE: |<body><math><mi><svg>| -> "in foreign content" insertion            ## NOTE: |<body><math><mi><svg>| -> "in foreign content" insertion
7478            ## mode, "in body" (not "in foreign content") secondary insertion            ## mode, "in body" (not "in foreign content") secondary insertion
# Line 7463  sub _tree_construction_main ($) { Line 7483  sub _tree_construction_main ($) {
7483          next B;          next B;
7484        } elsif ({        } elsif ({
7485                  caption => 1, col => 1, colgroup => 1, frame => 1,                  caption => 1, col => 1, colgroup => 1, frame => 1,
7486                  frameset => 1, head => 1, option => 1, optgroup => 1,                  frameset => 1, head => 1,
7487                  tbody => 1, td => 1, tfoot => 1, th => 1,                  tbody => 1, td => 1, tfoot => 1, th => 1,
7488                  thead => 1, tr => 1,                  thead => 1, tr => 1,
7489                 }->{$token->{tag_name}}) {                 }->{$token->{tag_name}}) {
# Line 7474  sub _tree_construction_main ($) { Line 7494  sub _tree_construction_main ($) {
7494          !!!nack ('t401.1'); ## NOTE: |<col/>| or |<frame/>| here is an error.          !!!nack ('t401.1'); ## NOTE: |<col/>| or |<frame/>| here is an error.
7495          !!!next-token;          !!!next-token;
7496          next B;          next B;
7497                  } elsif ($token->{tag_name} eq 'param' or
7498          ## ISSUE: An issue on HTML5 new elements in the spec.                 $token->{tag_name} eq 'source') {
7499            !!!insert-element-t ($token->{tag_name}, $token->{attributes}, $token);
7500            pop @{$self->{open_elements}};
7501    
7502            !!!ack ('t398.5');
7503            !!!next-token;
7504            redo B;
7505        } else {        } else {
7506          if ($token->{tag_name} eq 'image') {          if ($token->{tag_name} eq 'image') {
7507            !!!cp ('t384');            !!!cp ('t384');
# Line 7511  sub _tree_construction_main ($) { Line 7537  sub _tree_construction_main ($) {
7537            !!!ack ('t388.2');            !!!ack ('t388.2');
7538          } elsif ({          } elsif ({
7539                    area => 1, basefont => 1, bgsound => 1, br => 1,                    area => 1, basefont => 1, bgsound => 1, br => 1,
7540                    embed => 1, img => 1, param => 1, spacer => 1, wbr => 1,                    embed => 1, img => 1, spacer => 1, wbr => 1,
                   #image => 1,  
7541                   }->{$token->{tag_name}}) {                   }->{$token->{tag_name}}) {
7542            !!!cp ('t388.1');            !!!cp ('t388.1');
7543            pop @{$self->{open_elements}};            pop @{$self->{open_elements}};
# Line 7553  sub _tree_construction_main ($) { Line 7578  sub _tree_construction_main ($) {
7578              }              }
7579            }            }
7580    
7581            !!!parse-error (type => 'start tag not allowed',            ## NOTE: |<marquee></body>|, |<svg><foreignobject></body>|
7582    
7583              !!!parse-error (type => 'unmatched end tag',
7584                            text => $token->{tag_name}, token => $token);                            text => $token->{tag_name}, token => $token);
7585            ## NOTE: Ignore the token.            ## NOTE: Ignore the token.
7586            !!!next-token;            !!!next-token;
# Line 7580  sub _tree_construction_main ($) { Line 7607  sub _tree_construction_main ($) {
7607          ## up-to-date, though it has same effect as speced.          ## up-to-date, though it has same effect as speced.
7608          if (@{$self->{open_elements}} > 1 and          if (@{$self->{open_elements}} > 1 and
7609              $self->{open_elements}->[1]->[1] & BODY_EL) {              $self->{open_elements}->[1]->[1] & BODY_EL) {
           ## ISSUE: There is an issue in the spec.  
7610            unless ($self->{open_elements}->[-1]->[1] & BODY_EL) {            unless ($self->{open_elements}->[-1]->[1] & BODY_EL) {
7611              !!!cp ('t406');              !!!cp ('t406');
7612              !!!parse-error (type => 'not closed',              !!!parse-error (type => 'not closed',
# Line 7616  sub _tree_construction_main ($) { Line 7642  sub _tree_construction_main ($) {
7642    
7643                  applet => 1, button => 1, marquee => 1, object => 1,                  applet => 1, button => 1, marquee => 1, object => 1,
7644                 }->{$token->{tag_name}}) {                 }->{$token->{tag_name}}) {
7645            ## NOTE: Code for <li> start tags includes "as if </li>" code.
7646            ## Code for <dt> or <dd> start tags includes "as if </dt> or
7647            ## </dd>" code.
7648    
7649          ## has an element in scope          ## has an element in scope
7650          my $i;          my $i;
7651          INSCOPE: for (reverse 0..$#{$self->{open_elements}}) {          INSCOPE: for (reverse 0..$#{$self->{open_elements}}) {
# Line 7774  sub _tree_construction_main ($) { Line 7804  sub _tree_construction_main ($) {
7804          ## NOTE: As normal, except </p> implies <p> and ...          ## NOTE: As normal, except </p> implies <p> and ...
7805    
7806          ## has an element in scope          ## has an element in scope
7807            my $non_optional;
7808          my $i;          my $i;
7809          INSCOPE: for (reverse 0..$#{$self->{open_elements}}) {          INSCOPE: for (reverse 0..$#{$self->{open_elements}}) {
7810            my $node = $self->{open_elements}->[$_];            my $node = $self->{open_elements}->[$_];
# Line 7784  sub _tree_construction_main ($) { Line 7815  sub _tree_construction_main ($) {
7815            } elsif ($node->[1] & SCOPING_EL) {            } elsif ($node->[1] & SCOPING_EL) {
7816              !!!cp ('t411.1');              !!!cp ('t411.1');
7817              last INSCOPE;              last INSCOPE;
7818              } elsif ($node->[1] & END_TAG_OPTIONAL_EL) {
7819                ## NOTE: |END_TAG_OPTIONAL_EL| includes "p"
7820                !!!cp ('t411.2');
7821                #
7822              } else {
7823                !!!cp ('t411.3');
7824                $non_optional ||= $node;
7825                #
7826            }            }
7827          } # INSCOPE          } # INSCOPE
7828    
7829          if (defined $i) {          if (defined $i) {
7830            if ($self->{open_elements}->[-1]->[0]->manakai_local_name            ## 1. Generate implied end tags
7831                    ne $token->{tag_name}) {            #
7832    
7833              ## 2. If current node != "p", parse error
7834              if ($non_optional) {
7835              !!!cp ('t412.1');              !!!cp ('t412.1');
7836              !!!parse-error (type => 'not closed',              !!!parse-error (type => 'not closed',
7837                              text => $self->{open_elements}->[-1]->[0]                              text => $non_optional->[0]->manakai_local_name,
                                 ->manakai_local_name,  
7838                              token => $token);                              token => $token);
7839            } else {            } else {
7840              !!!cp ('t414.1');              !!!cp ('t414.1');
7841            }            }
7842    
7843              ## 3. Pop
7844            splice @{$self->{open_elements}}, $i;            splice @{$self->{open_elements}}, $i;
7845          } else {          } else {
7846            !!!cp ('t413.1');            !!!cp ('t413.1');
# Line 7839  sub _tree_construction_main ($) { Line 7881  sub _tree_construction_main ($) {
7881          ## Ignore the token.          ## Ignore the token.
7882          !!!next-token;          !!!next-token;
7883          next B;          next B;
       } elsif ({  
                 caption => 1, col => 1, colgroup => 1, frame => 1,  
                 frameset => 1, head => 1, option => 1, optgroup => 1,  
                 tbody => 1, td => 1, tfoot => 1, th => 1,  
                 thead => 1, tr => 1,  
                 area => 1, basefont => 1, bgsound => 1,  
                 embed => 1, hr => 1, iframe => 1, image => 1,  
                 img => 1, input => 1, isindex => 1, noembed => 1,  
                 noframes => 1, param => 1, select => 1, spacer => 1,  
                 table => 1, textarea => 1, wbr => 1,  
                 noscript => 0, ## TODO: if scripting is enabled  
                }->{$token->{tag_name}}) {  
         !!!cp ('t429');  
         !!!parse-error (type => 'unmatched end tag',  
                         text => $token->{tag_name}, token => $token);  
         ## Ignore the token  
         !!!next-token;  
         next B;  
7884        } else {        } else {
7885          if ($token->{tag_name} eq 'sarcasm') {          if ($token->{tag_name} eq 'sarcasm') {
7886            sleep 0.001; # take a deep breath            sleep 0.001; # take a deep breath
# Line 7868  sub _tree_construction_main ($) { Line 7892  sub _tree_construction_main ($) {
7892    
7893          ## Step 2          ## Step 2
7894          S2: {          S2: {
7895            if ($node->[0]->manakai_local_name eq $token->{tag_name}) {            my $node_tag_name = $node->[0]->manakai_local_name;
7896              $node_tag_name =~ tr/A-Z/a-z/; # for SVG camelCase tag names
7897              if ($node_tag_name eq $token->{tag_name}) {
7898              ## Step 1              ## Step 1
7899              ## generate implied end tags              ## generate implied end tags
7900              while ($self->{open_elements}->[-1]->[1] & END_TAG_OPTIONAL_EL) {              while ($self->{open_elements}->[-1]->[1] & END_TAG_OPTIONAL_EL) {
# Line 7881  sub _tree_construction_main ($) { Line 7907  sub _tree_construction_main ($) {
7907              }              }
7908                    
7909              ## Step 2              ## Step 2
7910              if ($self->{open_elements}->[-1]->[0]->manakai_local_name              my $current_tag_name
7911                      ne $token->{tag_name}) {                  = $self->{open_elements}->[-1]->[0]->manakai_local_name;
7912                $current_tag_name =~ tr/A-Z/a-z/;
7913                if ($current_tag_name ne $token->{tag_name}) {
7914                !!!cp ('t431');                !!!cp ('t431');
7915                ## NOTE: <x><y></x>                ## NOTE: <x><y></x>
7916                !!!parse-error (type => 'not closed',                !!!parse-error (type => 'not closed',
Legend:
Removed from v.1.196  
changed lines
  Added in v.1.201
[email protected]
 ViewVC Help
Powered by ViewVC 1.1.24