/[suikacvs]/markup/html/whatpm/t/HTML-tokenizer.t
Suika

Diff of /markup/html/whatpm/t/HTML-tokenizer.t

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

revision 1.34 by wakaba, Sun Apr 13 06:44:27 2008 UTC revision 1.43 by wakaba, Sat Sep 20 09:28:32 2008 UTC
# Line 26  BEGIN { Line 26  BEGIN {
26  }  }
27    
28  use Test;  use Test;
29  BEGIN { plan tests => 477 }  BEGIN { plan tests => 1097 }
30    
31  use Data::Dumper;  use Data::Dumper;
32  $Data::Dumper::Useqq = 1;  $Data::Dumper::Useqq = 1;
33    $Data::Dumper::Sortkeys = 1;
34  sub Data::Dumper::qquote {  sub Data::Dumper::qquote {
35    my $s = shift;    my $s = shift;
36    $s =~ s/([^\x20\x21-\x26\x28-\x5B\x5D-\x7E])/sprintf '\x{%02X}', ord $1/ge;    $s =~ s/([^\x20\x21-\x26\x28-\x5B\x5D-\x7E])/sprintf '\x{%02X}', ord $1/ge;
# Line 106  for my $file_name (grep {$_} split /\s+/ Line 107  for my $file_name (grep {$_} split /\s+/
107        my $p = Whatpm::HTML->new;        my $p = Whatpm::HTML->new;
108        my $i = 0;        my $i = 0;
109        my @token;        my @token;
110        $p->{set_next_char} = sub {        $p->{set_nc} = sub {
111          my $self = shift;          my $self = shift;
112    
113          pop @{$self->{prev_char}};  #        pop @{$self->{prev_char}};
114          unshift @{$self->{prev_char}}, $self->{next_char};  #        unshift @{$self->{prev_char}}, $self->{nc};
115    
116          $self->{next_char} = -1 and return if $i >= length $s;          $self->{nc} = -1 and return if $i >= length $s;
117          $self->{next_char} = ord substr $s, $i++, 1;          $self->{nc} = ord substr $s, $i++, 1;
118    
119          if ($self->{next_char} == 0x000D) { # CR          if ($self->{nc} == 0x000D) { # CR
120            $i++ if substr ($s, $i, 1) eq "\x0A";            $i++ if substr ($s, $i, 1) eq "\x0A";
121            $self->{next_char} = 0x000A; # LF # MUST            $self->{nc} = 0x000A; # LF # MUST
122          } elsif ($self->{next_char} > 0x10FFFF) {          } elsif ($self->{nc} > 0x10FFFF) {
123            $self->{next_char} = 0xFFFD; # REPLACEMENT CHARACTER # MUST            $self->{nc} = 0xFFFD; # REPLACEMENT CHARACTER # MUST
124            push @token, 'ParseError';            push @token, 'ParseError';
125          } elsif ($self->{next_char} == 0x0000) { # NULL          } elsif ($self->{nc} == 0x0000) { # NULL
126            $self->{next_char} = 0xFFFD; # REPLACEMENT CHARACTER # MUST            $self->{nc} = 0xFFFD; # REPLACEMENT CHARACTER # MUST
127              push @token, 'ParseError';
128            } elsif ($self->{nc} <= 0x0008 or
129                     (0x000E <= $self->{nc} and
130                      $self->{nc} <= 0x001F) or
131                     (0x007F <= $self->{nc} and
132                      $self->{nc} <= 0x009F) or
133                     (0xD800 <= $self->{nc} and
134                      $self->{nc} <= 0xDFFF) or
135                     (0xFDD0 <= $self->{nc} and
136                      $self->{nc} <= 0xFDDF) or
137                     {
138                       0xFFFE => 1, 0xFFFF => 1, 0x1FFFE => 1, 0x1FFFF => 1,
139                       0x2FFFE => 1, 0x2FFFF => 1, 0x3FFFE => 1, 0x3FFFF => 1,
140                       0x4FFFE => 1, 0x4FFFF => 1, 0x5FFFE => 1, 0x5FFFF => 1,
141                       0x6FFFE => 1, 0x6FFFF => 1, 0x7FFFE => 1, 0x7FFFF => 1,
142                       0x8FFFE => 1, 0x8FFFF => 1, 0x9FFFE => 1, 0x9FFFF => 1,
143                       0xAFFFE => 1, 0xAFFFF => 1, 0xBFFFE => 1, 0xBFFFF => 1,
144                       0xCFFFE => 1, 0xCFFFF => 1, 0xDFFFE => 1, 0xDFFFF => 1,
145                       0xEFFFE => 1, 0xEFFFF => 1, 0xFFFFE => 1, 0xFFFFF => 1,
146                       0x10FFFE => 1, 0x10FFFF => 1,
147                      }->{$self->{nc}}) {
148            push @token, 'ParseError';            push @token, 'ParseError';
149          }          }
150        };        };
151        $p->{prev_char} = [-1, -1, -1];  
152        $p->{next_char} = -1;  
153    
154          $p->{read_until} = sub { return 0 };
155                
156        $p->{parse_error} = sub {        $p->{parse_error} = sub {
157          push @token, 'ParseError';          push @token, 'ParseError';
# Line 140  for my $file_name (grep {$_} split /\s+/ Line 164  for my $file_name (grep {$_} split /\s+/
164          PCDATA => Whatpm::HTML::PCDATA_CONTENT_MODEL (),          PCDATA => Whatpm::HTML::PCDATA_CONTENT_MODEL (),
165          PLAINTEXT => Whatpm::HTML::PLAINTEXT_CONTENT_MODEL (),          PLAINTEXT => Whatpm::HTML::PLAINTEXT_CONTENT_MODEL (),
166        }->{$cm};        }->{$cm};
167        $p->{last_emitted_start_tag_name} = $last_start_tag;        $p->{last_stag_name} = $last_start_tag;
168    
169        while (1) {        while (1) {
170          my $token = $p->_get_next_token;          my $token = $p->_get_next_token;
# Line 159  for my $file_name (grep {$_} split /\s+/ Line 183  for my $file_name (grep {$_} split /\s+/
183          $test_token->[1] = $token->{data} if defined $token->{data};          $test_token->[1] = $token->{data} if defined $token->{data};
184          if ($token->{type} == Whatpm::HTML::START_TAG_TOKEN ()) {          if ($token->{type} == Whatpm::HTML::START_TAG_TOKEN ()) {
185            $test_token->[2] = {map {$_->{name} => $_->{value}} values %{$token->{attributes}}};            $test_token->[2] = {map {$_->{name} => $_->{value}} values %{$token->{attributes}}};
186            if ({            $test_token->[3] = 1 if $p->{self_closing};
187                 ## NOTE: Permitted slash            delete $p->{self_closing};
                br => 1, link => 1,  
               }->{$token->{tag_name}}) {  
             delete $p->{self_closing};  
           }  
188          } elsif ($token->{type} == Whatpm::HTML::DOCTYPE_TOKEN ()) {          } elsif ($token->{type} == Whatpm::HTML::DOCTYPE_TOKEN ()) {
189            $test_token->[1] = $token->{name};            $test_token->[1] = $token->{name};
190            $test_token->[2] = $token->{public_identifier};            $test_token->[2] = $token->{pubid};
191            $test_token->[3] = $token->{system_identifier};            $test_token->[3] = $token->{sysid};
192            $test_token->[4] = $token->{quirks} ? 0 : 1;            $test_token->[4] = $token->{quirks} ? 0 : 1;
193          }          }
194    
Legend:
Removed from v.1.34  
changed lines
  Added in v.1.43
[email protected]
 ViewVC Help
Powered by ViewVC 1.1.24