[infinispan-dev] Set of keys modified in tx

[Vladimir Blagojevic]

On 5/20/09 1:57 PM, Manik Surtani wrote:
>
> Hmm, I wonder if this is all a lot of hassle for an early unlock.  
> Perhaps it does make sense - as discussed on a separate email thread - 
> to just stipulate that any explicit locks acquired using lock() from 
> within a TX get released when the TX completes.  Much easier for all 
> involved.

Yeah I think it is. I already looked into implementing this, not hard, 
just wiring a small change through a half-dozen classes. Will postpone 
this since we are removing unlock.

Cheers