It is a trick problem.  This was more in the way of a feature request than a proposal.
 
With respect to the transaction buffering issue that you mentioned elsewhere, I think that per page probably makes sense.  This would also be a good use of a fixed size array-based native hash map -- if you plan to flush when N page are filled.
 
-bryan
-----Original Message-----
From: jdbm-developer-admin@lists.sourceforge.net [mailto:jdbm-developer-admin@lists.sourceforge.net] On Behalf Of Kevin Day
Sent: Tuesday, September 27, 2005 12:35 AM
To: JDBM Developer listserv
Subject: re[2]: [Jdbm-developer] nio support

Bryan-
How would you see the following working?  Are you talking about separate log files per transaction?
 
How do you reconcile the fact that multiple threads could be attempting to modify the same block?
 
In my mind, this has always been the problem with simultaneous transactions.  I could see how you could have different offsets in the log file for each block in each transaction - but how do you determine who wins?
 
Seems like a tricky problem to me...
 
- K


 > I would also like to see support for full transactional logging with
multiple readers and writers, e.g., multiple logs and not flushing the
log until the transaction was complete. <
------------------------------------------------------- SF.Net email is sponsored by: Tame your development challenges with Apache's Geronimo App Server. Download it for free - -and be entered to win a 42" plasma tv or your very own Sony(tm)PSP. Click here to play: http://sourceforge.net/geronimo.php _______________________________________________ Jdbm-developer mailing list Jdbm-developer@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/jdbm-developer