From: Nathaniel Talbott <nathaniel@...>
Date: 2003-12-11T01:10:37+09:00
Subject: Re: Segfault in 1.8.1p3

On Dec 10, 2003, at 05:33, ts wrote:

>>>>>> "N" == Nathaniel Talbott <nathaniel@talbott.ws> writes:
>
> N> to track this down. I have the server stopped in gdb at the SIGSEGV
> N> point, so I can give information about the state of data structures 
> in
> N> memory if that would help.
>
>  It will be useless : the best is to give an example to reproduce the 
> bug

OK... didn't realize it would be this easy:

server.rb:

   require 'webrick'
   include WEBrick

   s = HTTPServer.new(
     :Port => 2000,
     :Logger => BasicLog::new('/dev/null'),
     :AccessLog => []
   )

   class HelloServlet < HTTPServlet::AbstractServlet
     def do_GET(req, res)
       res.body = "<HTML>hello, world.</HTML>"
       res['Content-Type'] = "text/html"
     end
   end
   s.mount("/hello", HelloServlet)

   trap("INT"){ s.shutdown }
   s.start


killer.rb:

   require 'thread'
   require 'net/http'

   threads = []

   200.times do
     threads << Thread.new do
       h = Net::HTTP::new('localhost', 2000)
       loop do
         h.get('/hello')
       end
     end
   end

   threads.each{|t| t.join}

Running:

   ntalbott@proxytest:~/sandbox$ ruby server.rb &
   [1] 20130
   ntalbott@proxytest:~/sandbox$ ruby killer.rb
   /usr/lib/ruby/1.8/net/http.rb:1699: [BUG] Segmentation fault
   ruby 1.8.1 (2003-12-05) [i386-linux]

   Aborted

Interestingly enough, it's not the server, but the killer that dies, so 
I can't guarantee this is the same segfault that my larger program has. 
It's a good place to start, however.

Thanks,


Nathaniel

<:((><