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

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

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

revision 1.27 by wakaba, Thu Jul 2 22:24:28 2009 UTC revision 1.30 by wakaba, Sun Aug 16 05:24:47 2009 UTC
# Line 948  sub _get_next_token ($) { Line 948  sub _get_next_token ($) {
948          if ({          if ({
949               0x0022 => 1, # "               0x0022 => 1, # "
950               0x0027 => 1, # '               0x0027 => 1, # '
951                 0x003C => 1, # <
952               0x003D => 1, # =               0x003D => 1, # =
953              }->{$self->{nc}}) {              }->{$self->{nc}}) {
954            !!!cp (55);            !!!cp (55);
# Line 1070  sub _get_next_token ($) { Line 1071  sub _get_next_token ($) {
1071    
1072          redo A;          redo A;
1073        } else {        } else {
1074          if ($self->{nc} == 0x0022 or # "          if ({
1075              $self->{nc} == 0x0027) { # '               0x0022 => 1, # "
1076                 0x0027 => 1, # '
1077                 0x003C => 1, # <
1078                }->{$self->{nc}}) {
1079            !!!cp (69);            !!!cp (69);
1080            ## XML5: Not a parse error.            ## XML5: Not a parse error.
1081            !!!parse-error (type => 'bad attribute name');            !!!parse-error (type => 'bad attribute name');
# Line 1182  sub _get_next_token ($) { Line 1186  sub _get_next_token ($) {
1186            !!!cp (78.2);            !!!cp (78.2);
1187          }          }
1188    
1189          if ($self->{nc} == 0x0022 or # "          if ({
1190              $self->{nc} == 0x0027) { # '               0x0022 => 1, # "
1191                 0x0027 => 1, # '
1192                 0x003C => 1, # <
1193                }->{$self->{nc}}) {
1194            !!!cp (78);            !!!cp (78);
1195            ## XML5: Not a parse error.            ## XML5: Not a parse error.
1196            !!!parse-error (type => 'bad attribute name');            !!!parse-error (type => 'bad attribute name');
# Line 2130  sub _get_next_token ($) { Line 2137  sub _get_next_token ($) {
2137          $self->{state} = BEFORE_DOCTYPE_NAME_STATE;          $self->{state} = BEFORE_DOCTYPE_NAME_STATE;
2138          !!!next-input-character;          !!!next-input-character;
2139          redo A;          redo A;
2140          } elsif ($self->{nc} == -1) {
2141            !!!cp (155.1);
2142            !!!parse-error (type => 'unclosed DOCTYPE');
2143            $self->{ct}->{quirks} = 1;
2144    
2145            $self->{state} = DATA_STATE;
2146            ## Reconsume.
2147            !!!emit ($self->{ct}); # DOCTYPE (quirks)
2148    
2149            redo A;
2150        } else {        } else {
2151          !!!cp (156);          !!!cp (156);
2152          ## XML5: Unless EOF, swith to the bogus comment state.          ## XML5: Swith to the bogus comment state.
2153          !!!parse-error (type => 'no space before DOCTYPE name');          !!!parse-error (type => 'no space before DOCTYPE name');
2154          $self->{state} = BEFORE_DOCTYPE_NAME_STATE;          $self->{state} = BEFORE_DOCTYPE_NAME_STATE;
2155          ## reconsume          ## reconsume
# Line 2157  sub _get_next_token ($) { Line 2174  sub _get_next_token ($) {
2174          !!!emit ($self->{ct}); # DOCTYPE (quirks)          !!!emit ($self->{ct}); # DOCTYPE (quirks)
2175    
2176          redo A;          redo A;
2177          } elsif (0x0041 <= $self->{nc} and $self->{nc} <= 0x005A) { # A..Z
2178            !!!cp (158.1);
2179            $self->{ct}->{name} # DOCTYPE
2180                = chr ($self->{nc} + ($self->{is_xml} ? 0 : 0x0020));
2181            delete $self->{ct}->{quirks};
2182            $self->{state} = DOCTYPE_NAME_STATE;
2183            !!!next-input-character;
2184            redo A;
2185        } elsif ($self->{nc} == -1) {        } elsif ($self->{nc} == -1) {
2186          !!!cp (159);          !!!cp (159);
2187          !!!parse-error (type => 'no DOCTYPE name');          !!!parse-error (type => 'no DOCTYPE name');
# Line 2203  sub _get_next_token ($) { Line 2228  sub _get_next_token ($) {
2228          !!!emit ($self->{ct}); # DOCTYPE          !!!emit ($self->{ct}); # DOCTYPE
2229    
2230          redo A;          redo A;
2231          } elsif (0x0041 <= $self->{nc} and $self->{nc} <= 0x005A) { # A..Z
2232            !!!cp (162.1);
2233            $self->{ct}->{name} # DOCTYPE
2234                .= chr ($self->{nc} + ($self->{is_xml} ? 0 : 0x0020));
2235            delete $self->{ct}->{quirks};
2236            ## Stay in the state.
2237            !!!next-input-character;
2238            redo A;
2239        } elsif ($self->{nc} == -1) {        } elsif ($self->{nc} == -1) {
2240          !!!cp (163);          !!!cp (163);
2241          !!!parse-error (type => 'unclosed DOCTYPE');          !!!parse-error (type => 'unclosed DOCTYPE');
# Line 2224  sub _get_next_token ($) { Line 2257  sub _get_next_token ($) {
2257          redo A;          redo A;
2258        } else {        } else {
2259          !!!cp (164);          !!!cp (164);
2260          $self->{ct}->{name}          $self->{ct}->{name} .= chr ($self->{nc}); # DOCTYPE
2261            .= chr ($self->{nc}); # DOCTYPE          ## Stay in the state.
         ## Stay in the state  
2262          !!!next-input-character;          !!!next-input-character;
2263          redo A;          redo A;
2264        }        }
Legend:
Removed from v.1.27  
changed lines
  Added in v.1.30
admin@suikawiki.org
 ViewVC Help
Powered by ViewVC 1.1.24