ruby-dev summary 23841-23877

[Takaaki Tateishi]

Hello,

I summarized interesting discussions which had been posted to ruby-dev
for the past two weeks.


[ruby-dev:23864] avalue and break in C block

Nakada suggested the following APIs, since there is no way to pass
a block which always receives an array and to return a value using
'break' in the C layer.
(1) rb_iterate() for handling avalue
(2) rb_iterate() for handling a Proc object
(3) rb_iter_break() to return a value
Matz replied that he had a plan to re-design specifications about
(1) and (2) with considering multiple assignment, and he turn out
not to be eager to improve rb_iter_break() since he doesn't want
us to use it in the C layer.


[ruby-dev:23876] cgi, cgi/session

Horikawa posted a set of patches to fix/improve cgi.rb and
cgi/session.rb. He
(1) fixed an error about Session#delete(),
(2) added a code to validate data before doing untaint,
(3) fixed an error when receiving an empty data from a form and
(4) simplified CGI#[].

 

[gabriele renzi]

[ruby-dev:23876] cgi, cgi/session

Horikawa posted a set of patches to fix/improve cgi.rb and
cgi/session.rb. He
(1) fixed an error about Session#delete(),
(2) added a code to validate data before doing untaint,
(3) fixed an error when receiving an empty data from a form and
(4) simplified CGI#[].

talking about cgi/session... is it impossible to think of providing
some magic tpo be able to use cgi/session with webrick?
I mean, if we could provide something like:
Session.new(CgiWrapper.new(webrick_request),params)

This would be really useful imho, even if probably would require
further unification of WEBrick::Cookie and CGI::Cookie