Dag H. Wanvik wrote:
"Mamta A. Satoor (JIRA)" <[EMAIL PROTECTED]> writes:
I would like reviewers to take note of following comment at line 125 in the new
class
/* Believe it or not, a String might have more
* collation elements than characters.
* So, we handle that case by increasing the int array
* by 5 and copying array elements.
*/
Basically, when the characters in the string are converted to collation
elements, one character can get translated into say 2 collation elements (more
info can be found on it in this url
http://www.developertutorials.com/tutorials/java/text-searching-java-050524/page3.html
Would it not be easier/better to just use an ArrayList<int> to handle
any needed expansion for you (in stead of coding it yourself) and
finally do a toArray()?
Possibly, but I'm not sure an int[] is the long term solution. Looking
at RuleBasedCollator it seems to me the better object to cache would be
the CollationElementIterator itself, rather than the result of iterating
over it. I would say stay with the int[] and the simple array expansion
until the functionality is working and then look at improving it. An
ArrayList<int> could be expensive, is every int going to be expanded to
an Integer object?
Dan.
