Re: about REXML::Encoding

Hi.

> I think it was because the patch's solution breaks in cases where
> two  different encoded files are loaded one after the other, 
> but that's just a superficial guess and could be wrong. 
I write a script in which  'euc-jp' and 'utf-8' are tested.
I attach the test script to this mail.

# This script is written in UTF8. Also, it includes Japanese  
# characters. If you don't have Unicode font, You see odd characters.
 
Before testing, [ruby-core:1960]'s patch is needed. After patched,
rexml/encodings/euc-jp.rb is modified manually 
according to 'euc-jp.rb.rej'. And, If you don't have iconv or uconv,
you must install one of them.

In the script, REXML::Encoding is included in Foo class. 
REXML::Encoding#encoding=() is called two times and 'euc-jp' and 'UTF-8'
are assigned each time. In my environment, it seems this 
test script works well. 

> I looked at this when the patch was first proposed, and had
>  a reason for only accepting the part of the patch 
> that fixed the encoding's name... 
> although I don't the reason at the moment.
Yes. SHIFT-JIS and SHIFT_JIS problem is fixed in CVS. 

> I'm going to have to revisit it in more depth before I can give you
> a definitive answer.
Thank you for re-checking the patch. 
I am glad to hear that you review the patch.


__________________________________
Do You Yahoo!?
Upgrade Your Life
http://bb.yahoo.co.jp/