I've found that under certain circumstances with Hibernate 3.6.10, a
PersistentCollection's queuedOperations will be executed *prior* to the
CollectionEntry::snapshot being taken. As a result, the snapshot includes
the results of any queued operations. (Including transient items added by a
Is this the intended behavior?
Can anyone point me to where I can do some more reading about the purpose of
queued operations, and get a better sense of Hibernate's collection-handling
architecture in general? I'm particularly interested in;
CollectionEntry::postIniitialize() method vs.
Relationship between CollectionEntry and the PersistentCollection itself.
PersistCollection Queued Operations: (purpose, etc)
Can someone tell to which events belong all the collection operation ?
Because for the moment I'm using this:
public class OgmEventListener extends DefaultFlushEventListener implements
And obviously any collection operations is called during the flush time
frame. Should I implement other event listener and to which event type
should it (they) be associated to ?
I am the author of the stack over flow question and can help out with testing. To be honest I am very new to Mongo DB and NoSQL db's in general but I am willing to learn and help code if necessary.
Here is another recent question about this implementation with a comment, so I know there is a need for it out there.
Let me know how I can help.
It seems it might be desirable for Heroku users to have a MongoDB
Directory for Lucene:
We're not far off from having a solution by using Guillaume's
contribution to Infinispan to a MongoDB enabled CacheStore: we use
Infinispan's Lucene Directory implementation, so that it benefit from
in-VM caching, and have it persist to a MongoDB backend for the slow
and long term.
What we should do is to make it trivial to setup: I don't expect the
average Heroku user to want to dive into the Infinispan manuals.
Any takers? I can help with the Infinispan configuration, and hope
Guillaume can as well as author of the CacheStore, but we likely need
a Heroku user to try it all out.
I'm having problems adding hibernate timeout. Sessions are not closed after
timeout. One of the requirements is not to use connection pool.
Below you can find a brief summary of our specification requirements.
- User logins in the web application using his own oracle database account;
- The web application only needs to access one database schema;
- The access rights are managed using oracle roles and grant privilegies;
- Only one database schema and multiple user oracle accounts;
- We only set the username and password when users tries to login in the
- The EntityManagerFactory is created using the username and password of
I tried to add the following lines but the sessions are not closed after
<property name="hibernate.c3p0.timeout">1800</property> ==> 30 minutes only
for test propose
Any ideas on how can I solve this problem?
I remember there was some issue on some environment that if we have hibernate.show_sql enabled during testing, but can we enable this "hibernate.format_sql" by default?
usually i'd like to compare the generated sql between old and new metamodel, and it is hared to read w/o formated sql
Strong Liu <stliu at hibernate.org>
Via some volunteers the works are starting to support Apache Lucene 4
in the Infinispan project,
basically to get the same functionality as we have today: to store the
indexes in Infinispan.
(So a requirement to upgrade Hibernate Search)
The Directory API changed a bit, and we might also decide that
Directory isn't the right abstraction layer anymore: there are more
options, like the new Coded API.
To tell the details we need some POC first, but it's clear already
that whatever we do we can't also keep compatibility with Lucene 3.x.
Still I want to keep the Lucene 3 Infinispan Directory, at least for a
while. Do we agree on that?
Maven-wise, the current module is identified as:
I plan to have both modules to live in the Infinispan project,
independent. Suggestions to label / identify these modules?
I noticed that "hibernate_dev" doesn't seem to be tweeting about our
pushed commits anymore.
Checking the service hooks it looks like enabled, but I have no means
to check if the token & secret are correct; who has access to this
it seems this is workable on annotation : org.hibernate.test.annotations.filter.subclass.tableperclass.Human
but i don't see how to do in in hbm, at least from the dtd/xsd, there is no such element for the "subclass", "joined-subclass", "union-subclass"
shall we change the xsd?
Strong Liu <stliu at hibernate.org>