[Zope3-dev] i18n, unicode, and the underline

[Shane Hathaway]

Stephan Richter wrote:
> On Monday 14 April 2003 09:56, Shane Hathaway wrote:
> 
>>While we're on the subject, I'd like to discuss some thoughts I've had
>>on Zope I/O and Unicode.  It seems to me that the HTTP server should
>>expect to work with pure byte arrays, since HTTP has only minimal
>>provisions for Unicode.  I think the publisher should be the component
>>that does the work of converting byte-array requests into Unicode, and
>>of converting Unicode into byte arrays for the response.  What do you
>>think?
> 
> 
> But it does that already. Request (which does the conversion) is part of the 
> Publisher.  I think the conversion is exactely at the right spot right now; 
> earlier would be bad, since we would not know which data is binary and which 
> is text and later would be too deep in the system.

That's a definite improvement over the way it used to be.  Thanks.

Shane