• Open Menu Close Menu
  • Apple
  • Shopping Bag
  • Apple
  • Mac
  • iPad
  • iPhone
  • Watch
  • TV
  • Music
  • Support
  • Search apple.com
  • Shopping Bag

Lists

Open Menu Close Menu
  • Terms and Conditions
  • Lists hosted on this site
  • Email the Postmaster
  • Tips for posting to public mailing lists
Re: Let's play "Spot the Deadlock" ;-)
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Let's play "Spot the Deadlock" ;-)


  • Subject: Re: Let's play "Spot the Deadlock" ;-)
  • From: Kieran Kelleher <email@hidden>
  • Date: Wed, 7 Oct 2009 17:35:51 -0400

Hmmm..... not sure if it is Wonder. This is an app deployed with Wonder source version svn#9978 (which is from Sep 27th).

There are two traces with EOObjectStoreCoordinator.lock() .... the 3rd last in my listing and that's happening in ERXEC.finalize() in the FinalizerThread. The other is in ERXEC.objectsWithFetchSpec(...) on a regular WorkerThread in appendToResponse in some page that I cannot identify, not that I think it matters in this case anyway since the finalizer is sth I have no control of ... that's the 7th last stack trace. I am baffled as to how to resolve this one though.

Anyone WOExperts got any insight?

Regards, Kieran

On Oct 7, 2009, at 5:08 PM, Henrique Gomes wrote:


On Oct 7, 2009, at 9:06 PM, Kieran Kelleher wrote:

Thread t@76291: (state = BLOCKED)
- sun.misc.Unsafe.park(boolean, long) @bci=0 (Interpreted frame)

Let me guess, you have just updated wonder, right?

The reason I'm saying this is that I've spent the day solving a similar deadlock, similar on the 'unsafe park' at least.
I am using docmosis (http://docmosis.com/) to generate PDFs and .DOCs. I've cooked up a simple 'data provider' that uses KVC to supply the data on the templates. I just pass an EO to it and presto. Since all my code is i the request/response loop, I never gave it a second thought. It turns out that a thread is spun to do the actual work, and the latest changes in locking behaviour in wonder exposed the faulty assumptions.
I now pass a GlobalID and lazily initialize an EditingContetx and fetch the EO on the first access to the data provider, that way every acces is on the same thread.


So my point really is that Anjo's latest changes on ERXEC might have exposed something, or there is a bug on those changes (naaaaahhh)

Regards,

Henrique Gomes

_______________________________________________
Do not post admin requests to the list. They will be ignored.
Webobjects-dev mailing list      (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden

_______________________________________________ Do not post admin requests to the list. They will be ignored. Webobjects-dev mailing list (email@hidden) Help/Unsubscribe/Update your Subscription: This email sent to email@hidden
  • Follow-Ups:
    • Re: Let's play "Spot the Deadlock" ;-)
      • From: Lachlan Deck <email@hidden>
References: 
 >Let's play "Spot the Deadlock" ;-) (From: Kieran Kelleher <email@hidden>)
 >Re: Let's play "Spot the Deadlock" ;-) (From: Henrique Gomes <email@hidden>)

  • Prev by Date: Re: NSTimestamp bug in timestampByAddingGregorianUnits
  • Next by Date: Re: NSTimestamp bug in timestampByAddingGregorianUnits
  • Previous by thread: Re: Let's play "Spot the Deadlock" ;-)
  • Next by thread: Re: Let's play "Spot the Deadlock" ;-)
  • Index(es):
    • Date
    • Thread