Hi Igor,

On 2010-Oct-13, at 12:17, Igor Stasenko wrote:

[snip]

> 
> yes. once i tried to implement double-linked list using CAS and failed..  :)

Me too.  For both trying and failing :).

[snip]
>> 
> 
> Okay, i must confess, that atomic swap seems inferior to CAS.
> So, lets stick with CAS.
> 
> Thanks, Tom for providing arguments and proving your point.

No problem.  The whole lock-free data structure problem is interesting and I don't have any nearby colleagues I can discuss it with, so I enjoyed the discussion with you.  The only thing we were missing was a couple of beers and a whiteboard to draw pictures and write code on :).

If you and or Eliot implement CAS, I'll try to find the time to update my Split-Order-List Hash Table code on Squeaksource to be the full multi-thread version.  I don't think anyone will have an immediate use for it, but the algorithm is a thing of beauty - I wish I could claim it was mine :).

> 
> -- 
> Best regards,
> Igor Stasenko AKA sig.
> 
Regards,

--
Tom Rushworth