Hello Oliver, yes of course is anyone interested! We have the same problem implementing a store for our xml database. Our current solution is to disable cacheing and that is not adequate.
Thanks in advance, Stefan
Oliver Zeigermann wrote:
Hi Duncan and Martin,
I was thinking of a very similar alternative cache also using LRUMap for the transactional file store. I also noticed StandardStore seems to completely lack the concept of isolation of caching. The only thing it does is to reset the store upon error in
void delist(Service service, boolean success)
This is barely sufficient as in case of error all kinds of inconsistent data can be read from concurrently running transactions until the cache is reset.
Maybe StandardStore's caching needs a complete revision?!
If anyone is interested I could implement our transactional caching in StandardStore instead of inside the transactional file store incorporating Martins patch. So everyone could benefit from it.
Oliver
Martin Holz wrote:
Hello Duncan,
you are right. This is a known bug: http://nagoya.apache.org/bugzilla/show_bug.cgi?id=13142
Martin
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
.
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
