/[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 1.74 by wakaba, Sun Mar 2 23:51:00 2008 UTC revision 1.75 by wakaba, Mon Mar 3 00:13:22 2008 UTC
# Line 346  sub _initialize_tokenizer ($) { Line 346  sub _initialize_tokenizer ($) {
346  ##   ->{tag_name} (START_TAG_TOKEN, END_TAG_TOKEN)  ##   ->{tag_name} (START_TAG_TOKEN, END_TAG_TOKEN)
347  ##   ->{public_identifier} (DOCTYPE_TOKEN)  ##   ->{public_identifier} (DOCTYPE_TOKEN)
348  ##   ->{system_identifier} (DOCTYPE_TOKEN)  ##   ->{system_identifier} (DOCTYPE_TOKEN)
349  ##   ->{correct} == 1 or 0 (DOCTYPE_TOKEN)  ##   ->{quirks} == 1 or 0 (DOCTYPE_TOKEN): "force-quirks" flag
350  ##   ->{attributes} isa HASH (START_TAG_TOKEN, END_TAG_TOKEN)  ##   ->{attributes} isa HASH (START_TAG_TOKEN, END_TAG_TOKEN)
351  ##        ->{name}  ##        ->{name}
352  ##        ->{value}  ##        ->{value}
# Line 1447  sub _get_next_token ($) { Line 1447  sub _get_next_token ($) {
1447          $self->{state} = DATA_STATE;          $self->{state} = DATA_STATE;
1448          !!!next-input-character;          !!!next-input-character;
1449    
1450          !!!emit ({type => DOCTYPE_TOKEN}); # incorrect          !!!emit ({type => DOCTYPE_TOKEN, quirks => 1});
1451    
1452          redo A;          redo A;
1453        } elsif ($self->{next_input_character} == -1) {        } elsif ($self->{next_input_character} == -1) {
# Line 1455  sub _get_next_token ($) { Line 1455  sub _get_next_token ($) {
1455          $self->{state} = DATA_STATE;          $self->{state} = DATA_STATE;
1456          ## reconsume          ## reconsume
1457    
1458          !!!emit ({type => DOCTYPE_TOKEN}); # incorrect          !!!emit ({type => DOCTYPE_TOKEN, quirks => 1});
1459    
1460          redo A;          redo A;
1461        } else {        } else {
1462          $self->{current_token}          $self->{current_token}
1463              = {type => DOCTYPE_TOKEN,              = {type => DOCTYPE_TOKEN,
1464                 name => chr ($self->{next_input_character}),                 name => chr ($self->{next_input_character}),
1465                 correct => 1};                 #quirks => 0,
1466                  };
1467  ## ISSUE: "Set the token's name name to the" in the spec  ## ISSUE: "Set the token's name name to the" in the spec
1468          $self->{state} = DOCTYPE_NAME_STATE;          $self->{state} = DOCTYPE_NAME_STATE;
1469          !!!next-input-character;          !!!next-input-character;
# Line 1490  sub _get_next_token ($) { Line 1491  sub _get_next_token ($) {
1491          $self->{state} = DATA_STATE;          $self->{state} = DATA_STATE;
1492          ## reconsume          ## reconsume
1493    
1494          delete $self->{current_token}->{correct};          $self->{current_token}->{quirks} = 1;
1495          !!!emit ($self->{current_token}); # DOCTYPE          !!!emit ($self->{current_token}); # DOCTYPE
1496    
1497          redo A;          redo A;
# Line 1522  sub _get_next_token ($) { Line 1523  sub _get_next_token ($) {
1523          $self->{state} = DATA_STATE;          $self->{state} = DATA_STATE;
1524          ## reconsume          ## reconsume
1525    
1526          delete $self->{current_token}->{correct};          $self->{current_token}->{quirks} = 1;
1527          !!!emit ($self->{current_token}); # DOCTYPE          !!!emit ($self->{current_token}); # DOCTYPE
1528    
1529          redo A;          redo A;
# Line 1586  sub _get_next_token ($) { Line 1587  sub _get_next_token ($) {
1587        }        }
1588    
1589        !!!parse-error (type => 'string after DOCTYPE name');        !!!parse-error (type => 'string after DOCTYPE name');
1590        delete $self->{current_token}->{correct};        $self->{current_token}->{quirks} = 1;
1591    
1592        $self->{state} = BOGUS_DOCTYPE_STATE;        $self->{state} = BOGUS_DOCTYPE_STATE;
1593        # next-input-character is already done        # next-input-character is already done
# Line 1615  sub _get_next_token ($) { Line 1616  sub _get_next_token ($) {
1616          $self->{state} = DATA_STATE;          $self->{state} = DATA_STATE;
1617          !!!next-input-character;          !!!next-input-character;
1618    
1619          delete $self->{current_token}->{correct};          $self->{current_token}->{quirks} = 1;
1620          !!!emit ($self->{current_token}); # DOCTYPE          !!!emit ($self->{current_token}); # DOCTYPE
1621    
1622          redo A;          redo A;
# Line 1625  sub _get_next_token ($) { Line 1626  sub _get_next_token ($) {
1626          $self->{state} = DATA_STATE;          $self->{state} = DATA_STATE;
1627          ## reconsume          ## reconsume
1628    
1629          delete $self->{current_token}->{correct};          $self->{current_token}->{quirks} = 1;
1630          !!!emit ($self->{current_token}); # DOCTYPE          !!!emit ($self->{current_token}); # DOCTYPE
1631    
1632          redo A;          redo A;
1633        } else {        } else {
1634          !!!parse-error (type => 'string after PUBLIC');          !!!parse-error (type => 'string after PUBLIC');
1635          delete $self->{current_token}->{correct};          $self->{current_token}->{quirks} = 1;
1636    
1637          $self->{state} = BOGUS_DOCTYPE_STATE;          $self->{state} = BOGUS_DOCTYPE_STATE;
1638          !!!next-input-character;          !!!next-input-character;
# Line 1648  sub _get_next_token ($) { Line 1649  sub _get_next_token ($) {
1649          $self->{state} = DATA_STATE;          $self->{state} = DATA_STATE;
1650          !!!next-input-character;          !!!next-input-character;
1651    
1652          delete $self->{current_token}->{correct};          $self->{current_token}->{quirks} = 1;
1653          !!!emit ($self->{current_token}); # DOCTYPE          !!!emit ($self->{current_token}); # DOCTYPE
1654    
1655          redo A;          redo A;
# Line 1658  sub _get_next_token ($) { Line 1659  sub _get_next_token ($) {
1659          $self->{state} = DATA_STATE;          $self->{state} = DATA_STATE;
1660          ## reconsume          ## reconsume
1661    
1662          delete $self->{current_token}->{correct};          $self->{current_token}->{quirks} = 1;
1663          !!!emit ($self->{current_token}); # DOCTYPE          !!!emit ($self->{current_token}); # DOCTYPE
1664    
1665          redo A;          redo A;
# Line 1680  sub _get_next_token ($) { Line 1681  sub _get_next_token ($) {
1681          $self->{state} = DATA_STATE;          $self->{state} = DATA_STATE;
1682          !!!next-input-character;          !!!next-input-character;
1683    
1684          delete $self->{current_token}->{correct};          $self->{current_token}->{quirks} = 1;
1685          !!!emit ($self->{current_token}); # DOCTYPE          !!!emit ($self->{current_token}); # DOCTYPE
1686    
1687          redo A;          redo A;
# Line 1690  sub _get_next_token ($) { Line 1691  sub _get_next_token ($) {
1691          $self->{state} = DATA_STATE;          $self->{state} = DATA_STATE;
1692          ## reconsume          ## reconsume
1693    
1694          delete $self->{current_token}->{correct};          $self->{current_token}->{quirks} = 1;
1695          !!!emit ($self->{current_token}); # DOCTYPE          !!!emit ($self->{current_token}); # DOCTYPE
1696    
1697          redo A;          redo A;
# Line 1732  sub _get_next_token ($) { Line 1733  sub _get_next_token ($) {
1733          $self->{state} = DATA_STATE;          $self->{state} = DATA_STATE;
1734          ## reconsume          ## reconsume
1735    
1736          delete $self->{current_token}->{correct};          $self->{current_token}->{quirks} = 1;
1737          !!!emit ($self->{current_token}); # DOCTYPE          !!!emit ($self->{current_token}); # DOCTYPE
1738    
1739          redo A;          redo A;
1740        } else {        } else {
1741          !!!parse-error (type => 'string after PUBLIC literal');          !!!parse-error (type => 'string after PUBLIC literal');
1742          delete $self->{current_token}->{correct};          $self->{current_token}->{quirks} = 1;
1743    
1744          $self->{state} = BOGUS_DOCTYPE_STATE;          $self->{state} = BOGUS_DOCTYPE_STATE;
1745          !!!next-input-character;          !!!next-input-character;
# Line 1767  sub _get_next_token ($) { Line 1768  sub _get_next_token ($) {
1768          $self->{state} = DATA_STATE;          $self->{state} = DATA_STATE;
1769          !!!next-input-character;          !!!next-input-character;
1770    
1771          delete $self->{current_token}->{correct};          $self->{current_token}->{quirks} = 1;
1772          !!!emit ($self->{current_token}); # DOCTYPE          !!!emit ($self->{current_token}); # DOCTYPE
1773    
1774          redo A;          redo A;
# Line 1777  sub _get_next_token ($) { Line 1778  sub _get_next_token ($) {
1778          $self->{state} = DATA_STATE;          $self->{state} = DATA_STATE;
1779          ## reconsume          ## reconsume
1780    
1781          delete $self->{current_token}->{correct};          $self->{current_token}->{quirks} = 1;
1782          !!!emit ($self->{current_token}); # DOCTYPE          !!!emit ($self->{current_token}); # DOCTYPE
1783    
1784          redo A;          redo A;
1785        } else {        } else {
1786          !!!parse-error (type => 'string after SYSTEM');          !!!parse-error (type => 'string after SYSTEM');
1787          delete $self->{current_token}->{correct};          $self->{current_token}->{quirks} = 1;
1788    
1789          $self->{state} = BOGUS_DOCTYPE_STATE;          $self->{state} = BOGUS_DOCTYPE_STATE;
1790          !!!next-input-character;          !!!next-input-character;
# Line 1800  sub _get_next_token ($) { Line 1801  sub _get_next_token ($) {
1801          $self->{state} = DATA_STATE;          $self->{state} = DATA_STATE;
1802          !!!next-input-character;          !!!next-input-character;
1803    
1804          delete $self->{current_token}->{correct};          $self->{current_token}->{quirks} = 1;
1805          !!!emit ($self->{current_token}); # DOCTYPE          !!!emit ($self->{current_token}); # DOCTYPE
1806    
1807          redo A;          redo A;
# Line 1810  sub _get_next_token ($) { Line 1811  sub _get_next_token ($) {
1811          $self->{state} = DATA_STATE;          $self->{state} = DATA_STATE;
1812          ## reconsume          ## reconsume
1813    
1814          delete $self->{current_token}->{correct};          $self->{current_token}->{quirks} = 1;
1815          !!!emit ($self->{current_token}); # DOCTYPE          !!!emit ($self->{current_token}); # DOCTYPE
1816    
1817          redo A;          redo A;
# Line 1832  sub _get_next_token ($) { Line 1833  sub _get_next_token ($) {
1833          $self->{state} = DATA_STATE;          $self->{state} = DATA_STATE;
1834          !!!next-input-character;          !!!next-input-character;
1835    
1836          delete $self->{current_token}->{correct};          $self->{current_token}->{quirks} = 1;
1837          !!!emit ($self->{current_token}); # DOCTYPE          !!!emit ($self->{current_token}); # DOCTYPE
1838    
1839          redo A;          redo A;
# Line 1842  sub _get_next_token ($) { Line 1843  sub _get_next_token ($) {
1843          $self->{state} = DATA_STATE;          $self->{state} = DATA_STATE;
1844          ## reconsume          ## reconsume
1845    
1846          delete $self->{current_token}->{correct};          $self->{current_token}->{quirks} = 1;
1847          !!!emit ($self->{current_token}); # DOCTYPE          !!!emit ($self->{current_token}); # DOCTYPE
1848    
1849          redo A;          redo A;
# Line 1874  sub _get_next_token ($) { Line 1875  sub _get_next_token ($) {
1875          $self->{state} = DATA_STATE;          $self->{state} = DATA_STATE;
1876          ## reconsume          ## reconsume
1877