[Zope-ZEO] ZODB database corruption under multiple connections

[Jim Fulton]

"John D. Heintz" wrote:
> 
> Jim,
> 
> Why wouldn't the transaction do any necessary sync()ing for us when it
> detects a conflict?  It would seem that the transaction object would be
> in a good place to catch the ConflictError, call sync(), and re-raise
> the exception.

Because sync implies aborting the current transaction.  It's at least
concievable that an application could recover from a conflict error without
aborting the current transaction.  I can't give a great example, but
I also don't feel confortable assuming that this will always be the right
thing to do.

I suppose that this automatic behavior could be a configuration setting.

Jim

--
Jim Fulton           mailto:jim@digicool.com   Python Powered!        
Technical Director   (888) 344-4332            http://www.python.org  
Digital Creations    http://www.digicool.com   http://www.zope.org