This is a fascinating issue. The MD5 string is produced by the SQL
statement and all the prepared statements. So the chances of getting a
collision is extremely low and i mean REALLY low.
So i am wondering if the issue isn't somewhere else.
Jari Ketola wrote:
Hi Nitai,
Thank you for your input. We are currently not using CACHENAME or
CACHEDOMAIN attributes, so queries are cached "the old fashioned way"
- based on the query string etc.
internalCacheName = CacheFactory.createCacheKey(serverName +
queryString + ((preparedDataList != null) ?
preparedDataList.hashCode() : 0) + thisDataSource.getUsername() +
thisDataSource.getPassword() + thisDataSource.getDataSourceName());
Using CACHENAME and CACHEDOMAIN would likely solve this problem, but
it would be nice keeping the code as close to ColdFusion specs as
possible.
We don't flush queries - we expire them using dynamic variables that
are included in the queries and updated as sections are updated. Sort
of a poor man's CACHEDOMAIN
--
official tag/function reference: http://openbd.org/manual/
mailing list - http://groups.google.com/group/openbd?hl=en
