Re: [boost] [thread-pool/futures] Explicit yielding

From: Peter Dimov (pdimov_at_[hidden])
Date: 2008-06-06 16:20:46

Johan Torp:

> In a previous post you said "Limiting the reuse to pool threads eliminates
> most other (theoretical IMO)
> problems.". What theoretical problems has been identified with thread
> re-use?

"State" leaking from one thread to another via thread-local variables, 
thread priority, or thread stack size.

> Does existing threadpools (such as java.util.concurrency.ExecutorService)
> offer thread re-use and if so is it considered an success?

I don't see this in the documentation of ThreadPoolExecutor, so I'd say that 
it doesn't reuse threads in this way. It only reuses the current thread in 
'submit' if the task would be otherwise rejected and the rejection policy is 
set to CallerRunsPolicy.