All right. We did talk about different things. Now I see the problem. Whoever 
initially programmed PostMethod did not take content encoding into consideration when 
performing URL encoding, and I blundered by failing to properly audit that bit of 
code. 

I'll fix it right away

Thanks for pinning it down for me

Cheers

Oleg 

>Let me make some definitions to ensure we are talking about the same 
>thing. There are two encodings envolved:
>
>                  Enc 1                          Enc 2
>[unicode string] ----------> URL encoded form --------> byte[]
>    "�"       URLEnc w/ UTF-8       %C3%B6      ASCII     0x25, 0x43 ...
>    "�"       URLEnc w/ ISO-8859-1    %F6       ASCII     0x25, 0x46 ...
>
>The focus is on Enc 1 and NOT on Enc 2.
>
>[EMAIL PROTECTED] wrote:
>> Odi, GET is an entirely different ball game. RFC is quite explicit
>> about it: it's US-ASCII all over the place, except for
>
>Sure, it's ASCII for what actually goes over the wire (enc 2). But 
>wether (how's that damn word spelt??) I encode � as %C3%B6 or %F6 in the 
>request parameter (enc 1) is quite different and *both* versions do 
>completely comply with the RFC, since the RFC only specifies enc 2 and 
>not enc 1. Correct me if I am wrong.
>
>> request/response body. That's why it takes URL-encoding in the very
>> first place in order to comply with the spec. Only request/response
>> body may be encoded with a different encoding
>
>You are talking about encoding 2 here.
>
>> It still does not make URIUtil#toUsingCharset make any sense, as far
>> as I can see.
>
>Of course not. I was only trying to figure out what was going on in 
>Sung-Gu's mind :-)
>
>
>---------------------------------------------------------------------
>To unsubscribe, e-mail: [EMAIL PROTECTED]
>For additional commands, e-mail: [EMAIL PROTECTED]
>
>

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to