Class used to manage timeout handlers across multiple threads.
Timeout handlers should be managed by using the class methods which are synchronized.
id = TimeoutHandler.register(10, Timeout::Error)
begin
sleep 20
puts 'foo'
ensure
TimeoutHandler.cancel(id)
end
will raise Timeout::Error
id = TimeoutHandler.register(10, Timeout::Error)
begin
sleep 5
puts 'foo'
ensure
TimeoutHandler.cancel(id)
end
will print 'foo'
| TimeoutMutex | = | Mutex.new # :nodoc: |
Mutex used to synchronize access across threads |
||
Cancels the timeout handler id
Source: show
# File lib/webrick/utils.rb, line 154 def TimeoutHandler.cancel(id) TimeoutMutex.synchronize{ instance.cancel(Thread.current, id) } end
Creates a new TimeoutHandler. You should use ::register and ::cancel instead of creating the timeout handler directly.
Source: show
# File lib/webrick/utils.rb, line 163 def initialize @timeout_info = Hash.new Thread.start{ while true now = Time.now @timeout_info.keys.each{|thread| ary = @timeout_info[thread] next unless ary ary.dup.each{|info| time, exception = *info interrupt(thread, info.object_id, exception) if time < now } } sleep 0.5 end } end
Registers a new timeout handler
Source: show
# File lib/webrick/utils.rb, line 146 def TimeoutHandler.register(seconds, exception) TimeoutMutex.synchronize{ instance.register(Thread.current, Time.now + seconds, exception) } end
Cancels the timeout handler id
Source: show
# File lib/webrick/utils.rb, line 204 def cancel(thread, id) if ary = @timeout_info[thread] ary.delete_if{|info| info.object_id == id } if ary.empty? @timeout_info.delete(thread) end return true end return false end
Interrupts the timeout handler id and raises
exception
Source: show
# File lib/webrick/utils.rb, line 183 def interrupt(thread, id, exception) TimeoutMutex.synchronize{ if cancel(thread, id) && thread.alive? thread.raise(exception, "execution timeout") end } end