Datamapper DataMapper → 0.10.2

Comments and changes to this ticket

• 

  Jonathan Stott (namelessjon) October 14th, 2009 @ 06:45 PM

  • State changed from “unconfirmed” to “confirmed”

  According to git bisect, the problem commit is in dm-core (there are no dm-timestamp changes in the time period):

  6b9ab6d A dirty sibling should not make a resource dirty

  Attached is the script used to run the bisection.

  • test_timestamps.rb 712 Bytes

• You flagged this item as spam.
  

  mpd October 14th, 2009 @ 11:22 PM

  I believe this is what's causing the issue in #1077 as well. I never make the dm-timestamps connection, though, which is why I was never able to come up with a toy script to repeat it.

  Awesome work finding this one!

• 

  Sindre Aarsaether October 15th, 2009 @ 12:14 AM

  Yeah I actually spent atleast 2.5 hours trying to locate the bug. Just found it after I moved all my big models out into a separate script and started removing everything I could. Guess I will do that a lot sooner the next time!

• 

  Dan Kubb (dkubb) October 22nd, 2009 @ 09:26 PM

  • State changed from “confirmed” to “resolved”

  (from [d4ec6c90871f47bddd0cef3e6ba43b32c1117d62]) Fixed stack-too deep error thrown by saving a circular reference

  • This solution involves tracking which resources have been saved in a specific session, and not resaving them. This will require further refactoring since I would like to remove the "resources" shared state and perhaps track if an object is saved in an ivar.

  [#1094 state:resolved] http://github.com/datamapper/dm-core/commit/d4ec6c90871f47bddd0cef3...

• You flagged this item as spam.
  

  mpd October 23rd, 2009 @ 07:37 PM

  bummer, guess my problem wasn't this. still happening.

• 

  Sindre Aarsaether October 25th, 2009 @ 07:03 AM

  Just a heads up, I'm still experiencing very similar stack-level too deep problems, even though it might be a different bug.

  /usr/local/ruby-1.9.1-p243/lib/ruby/gems/1.9.1/gems/extlib-0.9.14/lib/extlib/assertions.rb:4:in each': stack level too deep (SystemStackError)

  from /usr/local/ruby-1.9.1-p243/lib/ruby/gems/1.9.1/gems/extlib-0.9.14/lib/extlib/assertions.rb:4:in `assert_kind_of'
  from /usr/local/ruby-1.9.1-p243/lib/ruby/gems/1.9.1/gems/dm-core-0.10.2/lib/dm-core/associations/relationship.rb:306:in `loaded?'
  from /usr/local/ruby-1.9.1-p243/lib/ruby/gems/1.9.1/gems/dm-core-0.10.2/lib/dm-core/resource.rb:805:in `block in parent_relationships'
  from /usr/local/ruby-1.9.1-p243/lib/ruby/gems/1.9.1/gems/dm-core-0.10.2/lib/dm-core/resource.rb:804:in `each_value'
  from /usr/local/ruby-1.9.1-p243/lib/ruby/gems/1.9.1/gems/dm-core-0.10.2/lib/dm-core/resource.rb:804:in `parent_relationships'
  from /usr/local/ruby-1.9.1-p243/lib/ruby/gems/1.9.1/gems/dm-core-0.10.2/lib/dm-core/resource.rb:841:in `parent_resources'
  from /usr/local/ruby-1.9.1-p243/lib/ruby/gems/1.9.1/gems/dm-core-0.10.2/lib/dm-core/resource.rb:618:in `dirty_parents?'
  from /usr/local/ruby-1.9.1-p243/lib/ruby/gems/1.9.1/gems/dm-core-0.10.2/lib/dm-core/resource.rb:619:in `block in dirty_parents?'
   ... 5945 levels...
  from /usr/local/ruby-1.9.1-p243/lib/ruby/gems/1.9.1/gems/thor-0.9.9/lib/thor.rb:90:in `start'
  from /usr/local/ruby-1.9.1-p243/lib/ruby/gems/1.9.1/gems/thor-0.9.9/bin/thor:7:in `<top (required)>'
  from /usr/local/ruby/bin/thor:19:in `load'
  from /usr/local/ruby/bin/thor:19:in `<main>'
  

  /usr/local/ruby-1.9.1-p243/lib/ruby/gems/1.9.1/gems/extlib-0.9.14/lib/extlib/assertions.rb:4:in each': stack level too deep (SystemStackError)

  from /usr/local/ruby-1.9.1-p243/lib/ruby/gems/1.9.1/gems/extlib-0.9.14/lib/extlib/assertions.rb:4:in `assert_kind_of'
  from /usr/local/ruby-1.9.1-p243/lib/ruby/gems/1.9.1/gems/dm-core-0.10.2/lib/dm-core/associations/relationship.rb:306:in `loaded?'
  from /usr/local/ruby-1.9.1-p243/lib/ruby/gems/1.9.1/gems/dm-core-0.10.2/lib/dm-core/resource.rb:805:in `block in parent_relationships'
  from /usr/local/ruby-1.9.1-p243/lib/ruby/gems/1.9.1/gems/dm-core-0.10.2/lib/dm-core/resource.rb:804:in `each_value'
  from /usr/local/ruby-1.9.1-p243/lib/ruby/gems/1.9.1/gems/dm-core-0.10.2/lib/dm-core/resource.rb:804:in `parent_relationships'
  from /usr/local/ruby-1.9.1-p243/lib/ruby/gems/1.9.1/gems/dm-core-0.10.2/lib/dm-core/resource.rb:841:in `parent_resources'
  from /usr/local/ruby-1.9.1-p243/lib/ruby/gems/1.9.1/gems/dm-core-0.10.2/lib/dm-core/resource.rb:618:in `dirty_parents?'
  from /usr/local/ruby-1.9.1-p243/lib/ruby/gems/1.9.1/gems/dm-core-0.10.2/lib/dm-core/resource.rb:619:in `block in dirty_parents?'
   ... 5945 levels...
  from /usr/local/ruby-1.9.1-p243/lib/ruby/gems/1.9.1/gems/thor-0.9.9/lib/thor.rb:90:in `start'
  from /usr/local/ruby-1.9.1-p243/lib/ruby/gems/1.9.1/gems/thor-0.9.9/bin/thor:7:in `<top (required)>'
  from /usr/local/ruby/bin/thor:19:in `load'
  from /usr/local/ruby/bin/thor:19:in `<main>'
  

  Will try to reduce it.

• 

  Sindre Aarsaether October 25th, 2009 @ 07:30 AM

  I'm having a hard time reducing it, as it happens in a complex environment. But, it is easy to show that it is definately a bug, and most likely linked to the fix of this one.

  invoice = company.invoices.first(*some criteria*)
  payment.reload # I have a saved payment, just reloading to be certain
  
  # THIS THROWS STACK LEVEL TOO DEEP
  payment.update :invoice => invoice
  
  # WHILE THIS DOES NOT
  payment.invoice = invoice
  payment.save
  

  So I am guessing that this is really the existing bug, where the fix did not take #update into account somehow?

• 

  Dan Kubb (dkubb) October 25th, 2009 @ 04:51 PM

  • State changed from “resolved” to “unconfirmed”

  Interesting. Sometimes it helps to attach the whole stacktrace to the ticket, not just the abbreviated version that ruby outputs when it throws a SystemStackError exception.

• 

  Dan Kubb (dkubb) October 26th, 2009 @ 05:51 AM

  • Assigned user set to “Dan Kubb (dkubb)”

  mpd: I just pushed some fixes for a recursion bug I found in another ticket. Can you try updating to edge (git) dm-core and see if it resolves your issue?

• You flagged this item as spam.
  

  mpd October 26th, 2009 @ 11:45 AM

  I can confirm that my issue (stack too deep/coredumps on freebsd) is resolved on all platforms I have access to.

  thanks!

• 

  Dan Kubb (dkubb) October 26th, 2009 @ 04:30 PM

  • Milestone set to “0.10.2”
  • State changed from “unconfirmed” to “resolved”