Question

I have a weird problem with aRuby program I’m writing.

Basically the idea is for the program to run constantly in the background. The program checks my browser history every 30 seconds and uploads any new history items to a server.

# client.rb
history = HistoryUploader.new(Chrome)
# Run everything
loop do
  history.run
  sleep 30
end

The important part of the HistoryUploader class looks like this

class HistoryUploader
  def run
    upload_history until local.last_seen_history_item == server.last_seen_history_item
  end

  def upload_history
    # POST batches of history items to the server
  end
end

The main problem I see with this code is that if HistoryUploader.run takes more than 30 seconds to complete (which it very well may since it is sending multiple http requests), the outside loop in client.rb will attempt to call run again and I could get parallel requests going to the server which would really confuse things.

Is there a way that I can block the run method from being called twice until it has finished?

Answer 1

I would use a request queue which executes requests one after another. You can also put a simple boolean flag inside HistoryUploader e.g. @is_uploading:

class HistoryUploader

  attr_accessor :is_uploading

  def initialize
    @is_uploading = false
  end

  def run
    if @is_uploading
        return
    end     
    upload_history until local.last_seen_history_item == server.last_seen_history_item
  end

  def upload_history
    @is_uploading = true
    # POST batches of history items to the server
    # On uploading finished: 
    @is_uploading = false
  end
end

If you really want to block the main loop until the uploading is finished you can dispatch a thread and wait for it to finish using join:

require 'thread'

t = Thread.new do
    #post to server
end
t.join