3/30/2017 5:25:24 AM
calls already populate caches. We now do something similar
once the transaction for the create call returns.
Signed-off-by: Pierre-Alexandre Meyer <pierre@mouraf.org>
|
3/30/2017 4:42:21 AM
a bad ehcache.xml is supplied, simply revert to default configuration
(throwing errors in a Guice provider doesn't work well).
Signed-off-by: Pierre-Alexandre Meyer <pierre@mouraf.org>
|
3/30/2017 4:34:43 AM
is local only.
See also https://github.com/killbill/killbill/issues/247.
Signed-off-by: Pierre-Alexandre Meyer <pierre@mouraf.org>
|
3/30/2017 3:42:42 AM
most cases, we've already had to fetch the latest state
to populate the audit and history tables.
Return it back to the transaction wrapper: this lets us avoid yet
another query when the DAO layer has to access it.
Signed-off-by: Pierre-Alexandre Meyer <pierre@mouraf.org>
|
3/29/2017 11:47:47 AM
full object to check if it exists already, just
look-up the record id.
Signed-off-by: Pierre-Alexandre Meyer <pierre@mouraf.org>
|
3/29/2017 10:58:03 AM
the entity and the entity record id
both before and after executing the JDBC transaction.
In reality, because we want the latest state for the history tables,
we need to query for the current state before the JDBC transaction
only in case of DELETE operations.
Additionally, no need to retrieve the entity record id separately
(it comes for free from the entity).
This saves 3 JDBC calls for any CREATE or DELETE operation, 2 for any UPDATE operation.
Signed-off-by: Pierre-Alexandre Meyer <pierre@mouraf.org>
|
3/29/2017 8:10:46 AM
throws NullPointerException if the provided key is null.
Signed-off-by: Pierre-Alexandre Meyer <pierre@mouraf.org>
|
3/29/2017 8:07:24 AM
is in preparation of upcoming performance-related history and audit changes.
See also https://github.com/killbill/killbill/issues/335.
Signed-off-by: Pierre-Alexandre Meyer <pierre@mouraf.org>
|
|
|
3/28/2017 2:56:31 PM
Meyer <pierre@mouraf.org>
|
