Name: paC48320 Date: 02/20/98


firstSentence() method of JavaDoc cannot parse Japanese sentences.
In JDK1.1.x,
sun.tools.javadoc.DocumentationGenerator#firstSentence(String)
uses no i18n coding.
Full stop characters are hard-coded.
So javadoc cannot get a fisrt sentence
from Japanese sentences correctly.

Javadoc should use java.text.BreakIterator#getSentence().
But BreakIterator of JDK1.1.6G can iterate sentences correctly.
I send a bug report about this BreakIterator today.
When BreakIterator is modified as the bug report,
you can use BreakIterator safely.

I recommend BreakIterator will be modified,
and Javadoc will use new BreakIterator.
And I also hope this bug is applied to JDK1.2 Javadoc.
(In JDK1.2beta3_G,
 sun.tools.javadoc.doclets.HtmlDocWriter#firstSentence(String))

But if BreakIterator is not modified,
you can use a workaround as following code.
But it is only considered of English and Japanese.
So you should not use this workaround.

for (int i = 0 ; i < len ; i++) {
switch (s.charAt(i)) {
case '\u3002': // PUNCTUATION IDEOGRAPHIC FULL STOP
return s.substring(0, i + 1);
case '.':
case '\uff0e': // FULLWIDTH FULL STOP
period = true;
break;
case '@':
return s.substring(0, i);
case ' ':
case '\t':
case '\n':
if (period) {
return s.substring(0, i);
}
break;
default:
period = false;
}
}
(Review ID: 25410)
======================================================================