scripts/gb-convert-entities.pl

   1 #!/usr/bin/perl
   2
   3 while( <> ) {
   4     if( /"[^">]*&([^;]+);/ && $1 !~ /^link/ ) {
   5         warn "Entity $1 in attribute";
   6     }
   7     s{\&(apos|nbsp|iexcl|cent|pound|curren|yen|brvbar|sect|uml|copy|ordf|laquo|not|shy|reg|macr|deg|plusmn|sup2|sup3|acute|micro|para|middot|cedil|sup1|ordm|raquo|frac14|frac12|frac34|iquest|Agrave|Aacute|Acirc|Atilde|Auml|Aring|AElig|Ccedil|Egrave|Eacute|Ecirc|Euml|Igrave|Iacute|Icirc|Iuml|ETH|Ntilde|Ograve|Oacute|Ocirc|Otilde|Ouml|times|Oslash|Ugrave|Uacute|Ucirc|Uuml|Yacute|THORN|szlig|agrave|aacute|acirc|atilde|auml|aring|aelig|ccedil|egrave|eacute|ecirc|euml|igrave|iacute|icirc|iuml|eth|ntilde|ograve|oacute|ocirc|otilde|ouml|divide|oslash|ugrave|uacute|ucirc|uuml|yacute|thorn|yuml|ampersand|lsquot|rsquot|ldquot|rdquot|minus|endash|emdash|ellips|lellips|blankline|percent|thinspace|frac116|plus);}{<ch.$1/>}g;
   8
   9     s{\240}{<ch.nbsp/>}g;
  10     s{\241}{<ch.iexcl/>}g;
  11     s{\242}{<ch.cent/>}g;
  12     s{\243}{<ch.pound/>}g;
  13     s{\244}{<ch.curren/>}g;
  14     s{\245}{<ch.yen/>}g;
  15     s{\246}{<ch.brvbar/>}g;
  16     s{\247}{<ch.sect/>}g;
  17     s{\250}{<ch.uml/>}g;
  18     s{\251}{<ch.copy/>}g;
  19     s{\252}{<ch.ordf/>}g;
  20     s{\253}{<ch.laquo/>}g;
  21     s{\254}{<ch.not/>}g;
  22     s{\255}{<ch.shy/>}g;
  23     s{\256}{<ch.reg/>}g;
  24     s{\257}{<ch.macr/>}g;
  25     s{\260}{<ch.deg/>}g;
  26     s{\261}{<ch.plusmn/>}g;
  27     s{\262}{<ch.sup2/>}g;
  28     s{\263}{<ch.sup3/>}g;
  29     s{\264}{<ch.acute/>}g;
  30     s{\265}{<ch.micro/>}g;
  31     s{\266}{<ch.para/>}g;
  32     s{\267}{<ch.middot/>}g;
  33     s{\270}{<ch.cedil/>}g;
  34     s{\271}{<ch.sup1/>}g;
  35     s{\272}{<ch.ordm/>}g;
  36     s{\273}{<ch.raquo/>}g;
  37     s{\274}{<ch.frac14/>}g;
  38     s{\275}{<ch.frac12/>}g;
  39     s{\276}{<ch.frac34/>}g;
  40     s{\277}{<ch.iquest/>}g;
  41     s{\300}{<ch.Agrave/>}g;
  42     s{\301}{<ch.Aacute/>}g;
  43     s{\302}{<ch.Acirc/>}g;
  44     s{\303}{<ch.Atilde/>}g;
  45     s{\304}{<ch.Auml/>}g;
  46     s{\305}{<ch.Aring/>}g;
  47     s{\306}{<ch.AElig/>}g;
  48     s{\307}{<ch.Ccedil/>}g;
  49     s{\310}{<ch.Egrave/>}g;
  50     s{\311}{<ch.Eacute/>}g;
  51     s{\312}{<ch.Ecirc/>}g;
  52     s{\313}{<ch.Euml/>}g;
  53     s{\314}{<ch.Igrave/>}g;
  54     s{\315}{<ch.Iacute/>}g;
  55     s{\316}{<ch.Icirc/>}g;
  56     s{\317}{<ch.Iuml/>}g;
  57     s{\320}{<ch.ETH/>}g;
  58     s{\321}{<ch.Ntilde/>}g;
  59     s{\322}{<ch.Ograve/>}g;
  60     s{\323}{<ch.Oacute/>}g;
  61     s{\324}{<ch.Ocirc/>}g;
  62     s{\325}{<ch.Otilde/>}g;
  63     s{\326}{<ch.Ouml/>}g;
  64     s{\327}{<ch.times/>}g;
  65     s{\330}{<ch.Oslash/>}g;
  66     s{\331}{<ch.Ugrave/>}g;
  67     s{\332}{<ch.Uacute/>}g;
  68     s{\333}{<ch.Ucirc/>}g;
  69     s{\334}{<ch.Uuml/>}g;
  70     s{\335}{<ch.Yacute/>}g;
  71     s{\336}{<ch.THORN/>}g;
  72     s{\337}{<ch.szlig/>}g;
  73     s{\340}{<ch.agrave/>}g;
  74     s{\341}{<ch.aacute/>}g;
  75     s{\342}{<ch.acirc/>}g;
  76     s{\343}{<ch.atilde/>}g;
  77     s{\344}{<ch.auml/>}g;
  78     s{\345}{<ch.aring/>}g;
  79     s{\346}{<ch.aelig/>}g;
  80     s{\347}{<ch.ccedil/>}g;
  81     s{\350}{<ch.egrave/>}g;
  82     s{\351}{<ch.eacute/>}g;
  83     s{\352}{<ch.ecirc/>}g;
  84     s{\353}{<ch.euml/>}g;
  85     s{\354}{<ch.igrave/>}g;
  86     s{\355}{<ch.iacute/>}g;
  87     s{\356}{<ch.icirc/>}g;
  88     s{\357}{<ch.iuml/>}g;
  89     s{\360}{<ch.eth/>}g;
  90     s{\361}{<ch.ntilde/>}g;
  91     s{\362}{<ch.ograve/>}g;
  92     s{\363}{<ch.oacute/>}g;
  93     s{\364}{<ch.ocirc/>}g;
  94     s{\365}{<ch.otilde/>}g;
  95     s{\366}{<ch.ouml/>}g;
  96     s{\367}{<ch.divide/>}g;
  97     s{\370}{<ch.oslash/>}g;
  98     s{\371}{<ch.ugrave/>}g;
  99     s{\372}{<ch.uacute/>}g;
 100     s{\373}{<ch.ucirc/>}g;
 101     s{\374}{<ch.uuml/>}g;
 102     s{\375}{<ch.yacute/>}g;
 103     s{\376}{<ch.thorn/>}g;
 104     s{\377}{<ch.yuml/>}g;
 105
 106     print;
 107 }