# Thread Deadlock on Ruby 1.8.7

**URL:** <https://www.ruby-forum.com/t/thread-deadlock-on-ruby-1-8-7/161149>\
**Category:** Ruby\
**Created:** [February 23, 2009, 9:40pm UTC](https://www.ruby-forum.com/t/thread-deadlock-on-ruby-1-8-7/161149 "2009-02-23T21:40:40Z")\
**Posts on this page:** 6\
**Page:** 1

<div class="post-metadata">

**Author:** ![Jayce\_M](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/j/e274bd/32.png) [@Jayce\_M](https://www.ruby-forum.com/u/Jayce_M)\
**Post date:** [February 23, 2009, 9:40pm UTC](https://www.ruby-forum.com/t/thread-deadlock-on-ruby-1-8-7/161149/1 "2009-02-23T21:40:40Z")

</div>

Hey everyone.

I’m running into a bit of a problem, what I’m trying to do is  
basically use one ruby script to initiate the main program, so that when  
I want to restart the program, it literally exits, going all the way  
back to the starter script so that all the files loaded by the main  
script are loaded anew when the starter script runs the program again.

I have achieved this by using system(), but it started two ruby  
processes so I tried using Thread.new{} instead since all thread  
variables are local to that thread, thus all my files and variables are  
unassigned when the thread terminates. It runs wonderfully, but when I  
call the quit command, I get a deadlock error from the main program’s  
quit method. Here’s the terminal output from basic start to error:

# [System] Welcome to Ecko! [System] Loading configuration… [12:24:10 PM][System] Configuration loaded. Running integrity checks… [12:24:10 PM][System] System data is stable. Displaying info… [12:24:10 PM][System] Please read the release notes for important information. [12:24:10 PM][System] Starting Ecko 0.3.83 Beta RC2. [12:24:10 PM][Connection] Checking for stored authtoken. [12:24:10 PM][Connection] Found stored authtoken. [12:24:10 PM][Connection] Logged in as v-Ecko-v [12:24:10 PM][Connection] Joining default chatrooms: Botdom [12:24:10 PM][Connection] Joined #Botdom [ok] [12:24:39 PM][Botdom] \*\* AzimuthDragon has joined. [12:24:41 PM][Botdom] |]quit deadlock 0x294c1fc: sleep:J(0x2e6e07c) (main) - Ecko.rb:32 deadlock 0x2e6e07c: sleep:- - C:/Documents and Settings/User/My Documents/NetBe ansProjects/Ecko Beta 2 RC2/lib/System/Core/Communication.class.rb:555 C:/Documents and Settings/User/My Documents/NetBeansProjects/Ecko Beta 2 RC2/lib /System/Core/Communication.class.rb:555: Thread(0x2e6e07c): deadlock (fatal)

Ecko has shut down. Would you like to run it again? (y/n)

This program logs into a chat room, by the way.

Here’s the starting sequence: Run Ecko.bat -\> Ecko.rb -\> main.rb

The batch file displays the ‘Run again?’ messages, for when the ruby  
script itself terminates, Ecko.rb is the initation script, and main.rb  
is the program’s main file. The program can be run just by using  
main.rb, but the terminal closes when the program terminates.

I’m also using this initiation script to recognize errors, before it was  
just checking the exit code. Here’s the code for the initation script  
and the method that quits the main program when used:

ecko = Thread.new { require ‘main.rb’ } # \< Starts the main script, the  
main script runs itself when it’s loaded, so requiring it is all that  
needs to be done.  
ecko.join # \< I think this is being used right, it’s so that the  
initiation script doesn’t continue running while the thread is alive.  
This is also as far as the initation script gets.  
case ecko[“exitstatus”] # \< Exit code is stored here manually by the  
main program, see the below code for more info.

That’s all the relevant code for the initiation script, here’s the main  
program’s quit code:

when ‘quit’  
say(c, ‘Exiting…’)  
@joined.each { |chan| log(chan, “\n----- Left #{chan} at  
#{Time.now.strftime(”%I:%M:%S %p")} (Quit command was used) -----\n")} #  
\< Logs the quit to the chatroom logs.  
self.disconnect() # \< Disconnects from chatroom.  
Thread.current[“exitstatus”] = 0 # \< Sets the exitstatus for 0 since  
this is a normal exit.  
Thread.stop # \< Kills the thread, if I did it right, the program dies  
when this is executed.  
true

That’s the code used to close the main program, and it dies on  
Thread.stop.

What am I doing wrong?

- Jayce

---

<div class="post-metadata">

**Author:** ![Jayce\_M](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/j/e274bd/32.png) [@Jayce\_M](https://www.ruby-forum.com/u/Jayce_M)\
**Post date:** [February 24, 2009, 1:35am UTC](https://www.ruby-forum.com/t/thread-deadlock-on-ruby-1-8-7/161149/2 "2009-02-24T01:35:14Z")

</div>

Eh, probably. I was looking through it all, was more curious about the  
deadlock, I went to Google for threading answers, and the page I found  
didn’t mention Thread#exit, so, didn’t think to look up other methods.  
I’ll  
give it a try.

Thank you!

* * *

From: “Joel VanderWerf” [removed\_email\_address@domain.invalid](mailto:removed_email_address@domain.invalid)  
Sent: Monday, February 23, 2009 4:21 PM  
To: “ruby-talk ML” [removed\_email\_address@domain.invalid](mailto:removed_email_address@domain.invalid)  
Subject: Re: Thread Deadlock on Ruby 1.8.7

---

<div class="post-metadata">

**Author:** ![Jayce\_M](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/j/e274bd/32.png) [@Jayce\_M](https://www.ruby-forum.com/u/Jayce_M)\
**Post date:** [February 24, 2009, 1:22am UTC](https://www.ruby-forum.com/t/thread-deadlock-on-ruby-1-8-7/161149/3 "2009-02-24T01:22:23Z")

</div>

Jayce M. wrote:

> when ‘quit’  
> say(c, ‘Exiting…’)  
> @joined.each { |chan| log(chan, “\n----- Left #{chan} at #{Time.now.strftime(”%I:%M:%S %p")} (Quit command was used) -----\n")} # \< Logs the quit to the chatroom logs.  
> self.disconnect() # \< Disconnects from chatroom.  
> Thread.current[“exitstatus”] = 0 # \< Sets the exitstatus for 0 since this is a normal exit.  
> Thread.stop # \< Kills the thread, if I did it right, the program dies when this is executed.  
> true
> 
> That’s the code used to close the main program, and it dies on Thread.stop.

Given your description, wouldn’t Thread.exit be the right method, not  
Thread.stop ?

---

<div class="post-metadata">

**Author:** ![Jayce\_M](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/j/e274bd/32.png) [@Jayce\_M](https://www.ruby-forum.com/u/Jayce_M)\
**Post date:** [February 24, 2009, 1:46am UTC](https://www.ruby-forum.com/t/thread-deadlock-on-ruby-1-8-7/161149/4 "2009-02-24T01:46:20Z")

</div>

It worked, thank you. I am curious though, what exactly causes the  
deadlock,  
in most cases?

* * *

From: “Jayce M.” [removed\_email\_address@domain.invalid](mailto:removed_email_address@domain.invalid)  
Sent: Monday, February 23, 2009 4:33 PM  
To: “ruby-talk ML” [removed\_email\_address@domain.invalid](mailto:removed_email_address@domain.invalid)  
Subject: Re: Thread Deadlock on Ruby 1.8.7

---

<div class="post-metadata">

**Author:** ![Jayce\_M](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/j/e274bd/32.png) [@Jayce\_M](https://www.ruby-forum.com/u/Jayce_M)\
**Post date:** [February 24, 2009, 1:56am UTC](https://www.ruby-forum.com/t/thread-deadlock-on-ruby-1-8-7/161149/5 "2009-02-24T01:56:16Z")

</div>

Jayce M. wrote:

> It worked, thank you. I am curious though, what exactly causes the  
> deadlock, in most cases?

Perhaps all threads were in the sleep/stop state?

---

<div class="post-metadata">

**Author:** ![Jayce\_M](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/j/e274bd/32.png) [@Jayce\_M](https://www.ruby-forum.com/u/Jayce_M)\
**Post date:** [February 24, 2009, 2:00am UTC](https://www.ruby-forum.com/t/thread-deadlock-on-ruby-1-8-7/161149/6 "2009-02-24T02:00:23Z")

</div>

True, if I had the starter script paused so that the main program could  
execute, an improper exit would kill the running thread and never resume  
the  
main one, would it?

🙂 Thanks for the help, I really appreciate it.

* * *

From: “Joel VanderWerf” [removed\_email\_address@domain.invalid](mailto:removed_email_address@domain.invalid)  
Sent: Monday, February 23, 2009 4:53 PM  
To: “ruby-talk ML” [removed\_email\_address@domain.invalid](mailto:removed_email_address@domain.invalid)  
Subject: Re: Thread Deadlock on Ruby 1.8.7