Description Christophe Telep 2005-12-08 16:25:17 EST

+++ This bug was initially created as a clone of Bug #119410 +++

The following is the text from comment #1 (Comment #1 From Steven R. Shaw 
2005-10-27 13:19) in bug #112789 :

Eclipse will incorporate and package ICU. However there are no packaging 
details as of yet. But it will be in a plugin and intended for eclipse 3.2 
platform.

The changes that one might incurr could be as simply changing an import 
statement to use the ICU class as oppose to the Java class.
As will the following java.text.* clases only an import statement would be 
changed:
 - BreakIterator
- CollationKey
 - Collator
 - DateFormat
 - DecimalFormat
 - NumberFormat
 - SimpleDateFormat
 - StringCharacterIterator
 - StringTokenizer

The biggest concern of the presentation comes as an effort to support unicode 
4.1:
 - Problem areas: using combining characters and using supplementary 
characters which are larger than a single char
 - Use BreakIterator to iterate through a string. Cannot use charAt(..) 
because actual characters could be 1 or 2 chars in length. Except for 
characters like "/" which are always one char ASCII. Use of charAt(..) must be 
fixed / changed.

java.text classes used in external interfaces must not be changed. Do not use 
ICU for external APIs.

Someone brought up a question about current test plans for the changes, and as 
of yet there are none.

So to as far as changes for us are required would be to use the ICU classes 
instead of the java classes and make sure string manipulations also use the 
ICU classes.
But since eclipse does not currently package ICU the changes cannot be done 
yet I suppose.

The ICU website is: http://icu.sourceforge.net/ (much more info that this 
presentation has provided)
There are more links in the presentation slides which I can send out if you do 
not receive a copy (and once I receive my copy).