nfedyashev / retryable Goto Github PK

The default configuration object includes both sleep and sleep_method but the "Sleeping" section shows a lambda being given to sleep: and not sleep_method:

Is that an error in the Sleeping section? (I came here today because it seems that sleep with a lambda used to work but doesn't seem to be working now.

It looks like Rubygems has version 3.0.1 of retryable available (https://rubygems.org/gems/retryable/versions/3.0.1), but that version doesn't exist in this repository. That's concerning for security reasons.
Could you check if that was an accident or a malicious upload?

Some kind of "reset time" option for tries would be nice.
When this option is set it should reset the tries count back to 0.
Why could this be useful?

I wrote a script that connects to a mysql server, sometimes i need to restart the server, so i wrapped my code with this gem.
When the server is restarting, a exception is raised, then it waits 10 seconds, the server is online again end everything is fine,
except i restart the server multiple times a week then the tries count is reached an the script stops.
but if i could define a timeout after the tries count is resetted this would work perfect.

sorry for my bad english, i hope you can understand what i mean.

Related commit
68c71b1

@nfedyashev Is there a recently published version of this gem from this fork? If not, would you be willing to publish one?

Thanks!

Hi,

The version 3.0.5 is from 2019.
There are changes being made since then. I am interested specifically the ruby compatibility for the v3.1 and v3.2 that was done in Jan 2023.
Are those changes inside of 3.0.5 version? Do you plan on releasing a new image version?

Thanks!

Is there a way to force an exception that Retryable would pick up?
I'm trying to retry if HTTParty response.code is not a 200 or 204. I can get it to raise, but that goes to my begin/rescue, and skips out of Retryable's logic.

def httpPost(results, url = @url)
    #ensure_cb = Proc.new { |retries| puts "Total Retries: #{retries}" }
    exception_cb = Proc.new { |exception| puts "For URL: #{url}\nRetry-rescued error: #{exception.inspect}"}
    begin
      Retryable.retryable(:exception_cb => exception_cb, tries: 10, sleep: lambda { |n| 4**n }, on: [Errno::ECONNREFUSED, SocketError, Net::ReadTimeout, HTTParty::Error, Timeout::Error, Errno::EPIPE, OpenSSL::SSL::SSLError, EOFError]) do |retries, exception|
        response = HTTParty.post(url, body: results.to_json, headers: {'Content-Type' => 'application/json', 'User-Agent' => 'SendGrid 1.0'})
        puts "Attempt #{retries} failed with exception: #{exception}" if retries > 0
        exception if not [200, 204].include?(response.code) #throw http-code-level error if not a 200 or 204
        puts "Total Attempts: #{retries + 1}" if retries > 0
        return response
      end
    rescue StandardError => e
      puts "#{Time.now} - httpPost error: #{e.inspect}"
    end
  end

All references to TimeoutError should be changed to Timeout::Error, at least for Ruby versions newer than 2.3. It seems that TimeoutError was deprecated.