Rainbows! Rack HTTP server user/dev discussion
 help / color / Atom feed
63d57cbca8928caef5d560e76c2405c38c145972 blob 2149 bytes (raw)

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
 
# -*- encoding: binary -*-

# Implements a worker thread pool model.  This is suited for platforms
# like Ruby 1.9, where the cost of dynamically spawning a new thread for
# every new client connection is higher than with the ThreadSpawn model,
# but the cost of an idle thread is low (e.g. NPTL under Linux).
#
# This model should provide a high level of compatibility with all Ruby
# implementations, and most libraries and applications.  Applications
# running under this model should be thread-safe but not necessarily
# reentrant.
#
# Applications using this model are required to be thread-safe.  Threads
# are never spawned dynamically under this model.
#
# If you're using green threads (MRI 1.8) and need to perform DNS lookups,
# consider using the "resolv-replace" library which replaces parts of the
# core Socket package with concurrent DNS lookup capabilities.
module Rainbows::ThreadPool
  include Rainbows::Base

  def worker_loop(worker) # :nodoc:
    init_worker_process(worker)
    pool = (1..worker_connections).map do
      Thread.new { LISTENERS.size == 1 ? sync_worker : async_worker }
    end

    while Rainbows.alive
      # if any worker dies, something is serious wrong, bail
      pool.each do |thr|
        Rainbows.tick or break
        thr.join(1) and Rainbows.quit!
      end
    end
    Rainbows::JoinThreads.acceptors(pool)
  end

  def sync_worker # :nodoc:
    s = LISTENERS[0]
    begin
      c = s.kgio_accept and c.process_loop
    rescue => e
      Rainbows::Error.listen_loop(e)
    end while Rainbows.alive
  end

  def async_worker # :nodoc:
    begin
      # TODO: check if select() or accept() is a problem on large
      # SMP systems under Ruby 1.9.  Hundreds of native threads
      # all working off the same socket could be a thundering herd
      # problem.  On the other hand, a thundering herd may not
      # even incur as much overhead as an extra Mutex#synchronize
      ret = select(LISTENERS) and ret[0].each do |s|
        s = s.kgio_tryaccept and s.process_loop
      end
    rescue Errno::EINTR
    rescue => e
      Rainbows::Error.listen_loop(e)
    end while Rainbows.alive
  end
end
debug log:

solving 63d57cb ...
found 63d57cb in http://bogomips.org/rainbows.git

Rainbows! Rack HTTP server user/dev discussion

Archives are clonable:
	git clone --mirror http://bogomips.org/rainbows-public
	git clone --mirror http://ou63pmih66umazou.onion/rainbows-public

Example config snippet for mirrors

Newsgroups are available over NNTP:
	nntp://news.public-inbox.org/inbox.comp.lang.ruby.rainbows
	nntp://ou63pmih66umazou.onion/inbox.comp.lang.ruby.rainbows

 note: .onion URLs require Tor: https://www.torproject.org/

AGPL code for this site: git clone https://public-inbox.org/public-inbox.git