/**
* Given a string, return an array of tokens. The separator can be escaped with the '\' character.
* The '\' character may also be escaped by the '\' character.
*
* @param s the string to tokenize.
* @param separator the separator char.
* @param maxTokens the maxmimum number of tokens returned. If the max is reached, the remaining
* part of s is appended to the end of the last token.
* @return an array of tokens.
*/
public static String[] tokenize(String s, char separator, int maxTokens) {
List tokens = new ArrayList();
StringBuilder token = new StringBuilder();
boolean prevIsEscapeChar = false;
for (int i = 0; i < s.length(); i += Character.charCount(i)) {
int currentChar = s.codePointAt(i);
if (prevIsEscapeChar) {
// Case 1: escaped character
token.appendCodePoint(currentChar);
prevIsEscapeChar = false;
} else if (currentChar == separator && tokens.size() < maxTokens - 1) {
// Case 2: separator
tokens.add(token.toString());
token = new StringBuilder();
} else if (currentChar == '\\') {
// Case 3: escape character
prevIsEscapeChar = true;
} else {
// Case 4: regular character
token.appendCodePoint(currentChar);
}
}
if (token.length() > 0) {
tokens.add(token.toString());
}
return (String[]) tokens.toArray(new String[] {});
}
protected String makeSpace(int len) {
if (len <= 0) {
return "";
}
StringBuilder sb = new StringBuilder(len);
for (int i = 0; i < len; i++) {
sb.append(' ');
}
return sb.toString();
}
/**
* Convert the name to a valid HTML name.
*
* @param name the name that needs to be converted to valid HTML name.
* @return a valid HTML name string.
*/
public String getName(String name) {
StringBuilder sb = new StringBuilder();
char ch;
/* The HTML 4 spec at http://www.w3.org/TR/html4/types.html#h-6.2 mentions
* that the name/id should begin with a letter followed by other valid characters.
* The HTML 5 spec (draft) is more permissive on names/ids where the only restriction
* is that it should be at least one character long and should not contain spaces.
* The spec draft is @ http://www.w3.org/html/wg/drafts/html/master/dom.html#the-id-attribute.
*
* For HTML 4, we need to check for non-characters at the beginning of the name and
* substitute it accordingly, "_" and "$" can appear at the beginning of a member name.
* The method substitutes "$" with "Z:Z:D" and will prefix "_" with "Z:Z".
*/
for (int i = 0; i < name.length(); i++) {
ch = name.charAt(i);
switch (ch) {
case '(':
case ')':
case '<':
case '>':
case ',':
sb.append('-');
break;
case ' ':
case '[':
break;
case ']':
sb.append(":A");
break;
// Any appearance of $ needs to be substituted with ":D" and not with hyphen
// since a field name "P$$ and a method P(), both valid member names, can end
// up as "P--". A member name beginning with $ needs to be substituted with
// "Z:Z:D".
case '$':
if (i == 0) sb.append("Z:Z");
sb.append(":D");
break;
// A member name beginning with _ needs to be prefixed with "Z:Z" since valid anchor
// names can only begin with a letter.
case '_':
if (i == 0) sb.append("Z:Z");
sb.append(ch);
break;
default:
sb.append(ch);
}
}
return sb.toString();
}