The purpose of this email is not to argue that there is a need to
have the option of killing a single response if it
takes longer than say 30 seconds. This is a worthwhile feature, years
of work in php has shown me that it is.
Thanks to a bit of help from various source in the community, I've
been able to find something that works with fastcgi
and surely should work with webbrick.
I tried -idle-timeout with the FastCgiConfig/FastCgiServer directives
and they failed, eg
FastCgiServer /www/vhosts/whatever/public/dispatch.fcgi -idle-timeout 30
left me with dispatch.fcgi still using up 98% of the cpu on my
intentional infinite loop test.
I tried require "timeout" and Timeout::timeout(2){ logger.info('im
here') } in dispatch.fcgi and that failed (pretty sure).
What did work came by suggestion.
In an ApplicationController do:
def process(*args)
timeout(30) { logger.info('killed by me'); super }
end
and it seems that it only takes care of one response (did some testing).
But, I would like if someone could point out the error in my
FastCgiServer attempt.
Thanks.
------------------------------------------------------------------------
Hunter Peress
[EMAIL PROTECTED]
Web Programmer
The Santa Fe New Mexican, Inc.
www.freenewmexican.com
_______________________________________________
Rails-core mailing list
Rails-core@lists.rubyonrails.org
http://lists.rubyonrails.org/mailman/listinfo/rails-core
