Hi,

James Gray wrote:
>> == Looking for a successor for cgi.rb
>> Conditions are as follows:
>> * different name than cgi.rb
>> * using MVC separation
> 
> Was there any discussing of what's desired from the new version, beyond 
> the new architecture?

Only two requirement were discussed.  We are still considering what 
architecture is good for Ruby.  Advices about that is welcome, for 
exmaple we need this feature or that feature is good for web programers 
and so on.

> Are there encoding issues with cgi.rb?  (I ask because of the meeting it 
> was raised at.)

I didn't check but, if locale is set other than C, encoding differ 
LOCALE and ASCII-8BIT may happen.  If locale is C and process strings 
with ASCII-8BTI encoding, I don't know any critical issues yet.  Anyway 
the fact that cgi.rb doesn't think about encoding must be a critical 
issue XD

>> == String#gsub(regexp, hash)
>> Proposal to allow the following syntax:
>>   String#gsub(regexp, {"ä"=>"\u00C4", ..})
>> -> accepted
> 
> Can you explain how this works?
> 
> Does the Hash provide sequences of translations that are applied before 
> the match?  Are they reversed after the match?

Longer exmaple is following.

s = "<foo>aä & ∞</foo>"
TABLE_HTMLREF = {
  '&lt;' => '<',
  '&gt;' => '>',
  '&amp;' => '&',
  '&auml;' => "\u00C4",
  '&infin;' => "\u221E",
}
s.gsub(/&\w+;/, TABLE_HTMLREF)

This code equals following.
s.gsub(/&\w+;/){|k| TABLE_HTMLREF[k]}

But new gsub with hash is faster than this.

-- 
NARUSE, Yui  <naruse / airemix.com>
DBDB A476 FDBD 9450 02CD 0EFC BCE3 C388 472E C1EA