antville/helma
antville/helma
1
1
Fork 0
Code Issues 5 Pull requests 5 Projects Releases 5 Activity Actions

Helma encode() functions now do entity encoding

Browse source 
again, but the right way. It transforms special Windows characters (such as smart quotes)
top valid Unicode, uses symbolic entities for ISO-8859-1 characters and numeric entities
fpr everything above.
This commit is contained in:
hns 2002-05-27 16:52:34 +00:00
parent 5a589d4bd6
commit c424e54261
1 changed files with 180 additions and 122 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

298
src/helma/util/HtmlEncoder.java
View file

@ -17,109 +17,146 @@ import java.text.*;
public final class HtmlEncoder {
// transformation table for characters 128 to 255. These actually fall into two
// groups, put together for efficiency: "Windows" chacacters 128-159 such as
// "smart quotes", which are encoded to valid Unicode entities, and
// valid ISO-8859 caracters 160-255, which are encoded to the symbolic HTML
// entity. Everything >= 256 is encoded to a numeric entity.
//
// for mor on HTML entities see http://www.pemberley.com/janeinfo/latin1.html and
// ftp://ftp.unicode.org/Public/MAPPINGS/VENDORS/MICSFT/WINDOWS/CP1252.TXT
//
static final String[] transform = {
"€", // 128
"", // empty string means character is undefined in unicode
"‚",
"ƒ",
"„",
"…",
"†",
"‡",
"ˆ",
"‰",
"Š",
"‹",
"Œ",
"",
"Ž",
"",
"",
"‘",
"’",
"“",
"”",
"•",
"–",
"—",
"˜",
"™",
"š",
"›",
"œ",
"",
"ž",
"Ÿ", // 159
" ", // 160
"¡",
"¢",
"£",
"¤",
"¥",
"¦",
"§",
"¨",
"©",
"ª",
"«",
"¬",
"­",
"®",
"¯",
"°",
"±",
"²",
"³",
"´",
"µ",
"¶",
"·",
"¸",
"¹",
"º",
"»",
"¼",
"½",
"¾",
"¿",
"À",
"Á",
"Â",
"Ã",
"Ä",
"Å",
"Æ",
"Ç",
"È",
"É",
"Ê",
"Ë",
"Ì",
"Í",
"Î",
"Ï",
"Ð",
"Ñ",
"Ò",
"Ó",
"Ô",
"Õ",
"Ö",
"×",
"Ø",
"Ù",
"Ú",
"Û",
"Ü",
"Ý",
"Þ",
"ß",
"à",
"á",
"â",
"ã",
"ä",
"å",
"æ",
"ç",
"è",
"é",
"ê",
"ë",
"ì",
"í",
"î",
"ï",
"ð",
"ñ",
"ò",
"ó",
"ô",
"õ",
"ö",
"÷",
"ø",
"ù",
"ú",
"û",
"ü",
"ý",
"þ",
"ÿ" // 255
};
/*
static final Hashtable convertor = new Hashtable (128);
// conversion table
static {
convertor.put(new Integer(160), " ");
convertor.put(new Integer(161), "¡");
convertor.put(new Integer(162), "¢");
convertor.put(new Integer(163), "£");
convertor.put(new Integer(164), "¤");
convertor.put(new Integer(165), "¥");
convertor.put(new Integer(166), "¦");
convertor.put(new Integer(167), "§");
convertor.put(new Integer(168), "¨");
convertor.put(new Integer(169), "©");
convertor.put(new Integer(170), "ª");
convertor.put(new Integer(171), "«");
convertor.put(new Integer(172), "¬");
convertor.put(new Integer(173), "­");
convertor.put(new Integer(174), "®");
convertor.put(new Integer(175), "¯");
convertor.put(new Integer(176), "°");
convertor.put(new Integer(177), "±");
convertor.put(new Integer(178), "²");
convertor.put(new Integer(179), "³");
convertor.put(new Integer(180), "´");
convertor.put(new Integer(181), "µ");
convertor.put(new Integer(182), "¶");
convertor.put(new Integer(183), "·");
convertor.put(new Integer(184), "¸");
convertor.put(new Integer(185), "¹");
convertor.put(new Integer(186), "º");
convertor.put(new Integer(187), "»");
convertor.put(new Integer(188), "¼");
convertor.put(new Integer(189), "½");
convertor.put(new Integer(190), "¾");
convertor.put(new Integer(191), "¿");
convertor.put(new Integer(192), "À");
convertor.put(new Integer(193), "Á");
convertor.put(new Integer(194), "Â");
convertor.put(new Integer(195), "Ã");
convertor.put(new Integer(196), "Ä");
convertor.put(new Integer(197), "Å");
convertor.put(new Integer(198), "Æ");
convertor.put(new Integer(199), "Ç");
convertor.put(new Integer(200), "È");
convertor.put(new Integer(201), "É");
convertor.put(new Integer(202), "Ê");
convertor.put(new Integer(203), "Ë");
convertor.put(new Integer(204), "Ì");
convertor.put(new Integer(205), "Í");
convertor.put(new Integer(206), "Î");
convertor.put(new Integer(207), "Ï");
convertor.put(new Integer(208), "Ð");
convertor.put(new Integer(209), "Ñ");
convertor.put(new Integer(210), "Ò");
convertor.put(new Integer(211), "Ó");
convertor.put(new Integer(212), "Ô");
convertor.put(new Integer(213), "Õ");
convertor.put(new Integer(214), "Ö");
convertor.put(new Integer(215), "×");
convertor.put(new Integer(216), "Ø");
convertor.put(new Integer(217), "Ù");
convertor.put(new Integer(218), "Ú");
convertor.put(new Integer(219), "Û");
convertor.put(new Integer(220), "Ü");
convertor.put(new Integer(221), "Ý");
convertor.put(new Integer(222), "Þ");
convertor.put(new Integer(223), "ß");
convertor.put(new Integer(224), "à");
convertor.put(new Integer(225), "á");
convertor.put(new Integer(226), "â");
convertor.put(new Integer(227), "ã");
convertor.put(new Integer(228), "ä");
convertor.put(new Integer(229), "å");
convertor.put(new Integer(230), "æ");
convertor.put(new Integer(231), "ç");
convertor.put(new Integer(232), "è");
convertor.put(new Integer(233), "é");
convertor.put(new Integer(234), "ê");
convertor.put(new Integer(235), "ë");
convertor.put(new Integer(236), "ì");
convertor.put(new Integer(237), "í");
convertor.put(new Integer(238), "î");
convertor.put(new Integer(239), "ï");
convertor.put(new Integer(240), "ð");
convertor.put(new Integer(241), "ñ");
convertor.put(new Integer(242), "ò");
convertor.put(new Integer(243), "ó");
convertor.put(new Integer(244), "ô");
convertor.put(new Integer(245), "õ");
convertor.put(new Integer(246), "ö");
convertor.put(new Integer(247), "÷");
convertor.put(new Integer(248), "ø");
convertor.put(new Integer(249), "ù");
convertor.put(new Integer(250), "ú");
convertor.put(new Integer(251), "û");
convertor.put(new Integer(252), "ü");
convertor.put(new Integer(253), "ý");
convertor.put(new Integer(254), "þ");
convertor.put(new Integer(255), "ÿ");
} */
/**
*
@ -197,16 +234,16 @@ public final class HtmlEncoder {
ret.append ('>');
break;
default:
ret.append (c);
// if (c < 160)
// ret.append ((char) c);
// else if (c >= 160 && c <= 255)
// ret.append (convertor.get(new Integer(c)));
// else {
// ret.append ("&#");
// ret.append (c);
// ret.append (";");
// }
if (c < 128)
ret.append (c);
else if (c >= 128 && c < 256)
ret.append (transform[c-128]);
else {
ret.append ("&#");
ret.append ((int) c);
ret.append (";");
}
if (!tagOpen && !Character.isWhitespace (c))
swallowOneNewline = false;
}
@ -271,16 +308,16 @@ public final class HtmlEncoder {
}
break;
default:
ret.append (c);
// if (c < 160)
// ret.append ((char) c);
// else if (c >= 160 && c <= 255)
// ret.append (convertor.get(new Integer(c)));
// else {
// ret.append ("&#");
// ret.append (c);
// ret.append (";");
// }
if (c < 128)
ret.append (c);
else if (c >= 128 && c < 256)
ret.append (transform[c-128]);
else {
ret.append ("&#");
ret.append ((int) c);
ret.append (";");
}
}
}
}
@ -315,5 +352,26 @@ public final class HtmlEncoder {
}
}
// test method
public static String printCharRange (int from, int to) {
StringBuffer response = new StringBuffer();
for (int i=from;i<to;i++) {
response.append (i);
response.append (" ");
response.append ((char) i);
response.append (" ");
if (i < 128)
response.append ((char) i);
else if (i >= 128 && i < 256)
response.append (transform[i-128]);
else {
response.append ("&#");
response.append (i);
response.append (";");
}
response.append ("\r\n");
}
return response.toString();
}
} // end of class