hibernate-devel — development

[Hibernate] Changes and additions to the XDoclet module.

[Joel Rosi-S. <Joe...@Et...>]

First off, today I committed into cvs changes to the Hibernate XDoclet 
module. These changes should bring the module into compliance with the 
hibernate-mapping-1.1.dtd. I would appreciate it if those of you who are 
using XDoclet would give this version a go and let me know if there are 
any problems that need addressing. See the Readme file for details on 
what has been done.

In the next round of work I plan on adding some new functionality. I 
have parsed through the 1.1 schema at a high level and there actually 
does not seem to be that much that still requires support. Here is my 
hit list:

   1. Top level collections - is it reasonable to leave these in the
      merge file or are class level tags desirable?
   2. <joined-subclass>
   3. <composite-id>
   4. <key-many-to-one>
   5. <generated-key>
   6. <index-many-to-many>

Has anyone noticed anything else that would be useful to support?

It is also my intention to get the documentation current and have a 
reproducable build process in place by the time I am done with this work.

If all goes well, I should be able to knock it all off by the end of 
next week

- joel

[Hibernate] OUTER JOIN and AS in query language

[Gavin K. <ga...@ap...>]

Does anyone have any objections / contributions to make 
to this proposed new FROM clause syntax:

https://sourceforge.net/forum/forum.php?thread_id=774491&forum_id=128638

[Hibernate] Re: Becoming an official XDoclet module

[Joel Rosi-S. <Joe...@Et...>]

Good, I am pleased that you support the proposition. It would be fine by 
me whoever the committer was; my goal is to have a viable module and I 
am willing to support that which ever way is deemed to be most appropriate.

- joel

Ara Abrahamian wrote:

>Well, Gavin once requested commit access to xdoclet's cvs for the
>hibernate module. In that period of time we've been discussing an
>approach for handling new modules without the need to put them in
>xdoclet's cvs. Now I'm +1 on hibernate in xdoclet's core and opening cvs
>access for its maintainer. So if Gavin agrees too I can candidate you
>(or maybe Gavin himself, whomever you find more appropriate for this
>role).
>
>Ara. 
>
>  
>
>>-----Original Message-----
>>From: Joel Rosi-Schwartz [mailto:Joe...@Et...]
>>Sent: Friday, December 06, 2002 5:15 PM
>>To: hib...@li...
>>Cc: ar...@ya...; Gav...@ex...
>>Subject: Becoming an official XDoclet module
>>
>>I know that this has come up before in the forum and dev list, but I
>>think it is worth re-evaluation. I just did an experiment of adding
>>    
>>
>the
>  
>
>>hibernate module to the latest XDoclet cvs source. It took all of 5
>>minutes to integrate the module into the Xdoclet build. The results
>>    
>>
>were
>  
>
>>as I expected; the build went smoothly and all of the documentation
>>    
>>
>was
>  
>
>>produced without any pain. So I am wondering if maintaining the
>>Hibernate module seperately is really worth the extra pain and effort
>>that is required.
>>
>>As I understand it the main consideration is not having a maintainer
>>    
>>
>for
>  
>
>>the module that has commit priviledges on the XDoclet project. I agree
>>that handling changes via patches is awkward. So what does it take to
>>get accepted as a module and as a commiter to the Xdoclet project? If
>>    
>>
>I
>  
>
>>had assurances that Gavin and Ara were both willing to give me a bit
>>    
>>
>of
>  
>
>>support as needed, I would be willing to take on being the maintainer.
>>
>>- joel
>>    
>>
>
>
>
>  
>

Re: [Hibernate] Java-based query evaluator

[Juozas B. <ba...@ce...>]

<snip>
> I'm curious how feasible it would be to write a Java-based evaluator of
> Hibernate queries, or at least leverage the existing parser code to gain
> access to the abstract syntax tree.
>
> One possibility is that certain subsets of queries (most filters, I
> think) could be run within the JVM, saving a db query.
possible it can be a good optimization, I am not sure:
 if   "SELECT * FROM TABLE" was executed in current session and all objects
are in cache , execute query on session cache.
It is possible to implement "If  current query result will be subset of some
query executed before ... ".
But it is not trivial.
Not all databases support standard sql, "NULL==NULL" can evalute to true on
databases, possible it can be more problems.

>
> Also, a translator could allow Hibernate query language to be used with
> other persistence engines (ok, this is not so compelling since Hibernate
> is the best :)
>
<snip>
> Thanks,
> Chris
>
>
>
>
>
> -------------------------------------------------------
> This sf.net email is sponsored by:ThinkGeek
> Welcome to geek heaven.
> http://thinkgeek.com/sf
> _______________________________________________
> hibernate-devel mailing list
> hib...@li...
> https://lists.sourceforge.net/lists/listinfo/hibernate-devel

[Hibernate] RE: Becoming an official XDoclet module

[Ara A. <ar...@ya...>]

Well, Gavin once requested commit access to xdoclet's cvs for the
hibernate module. In that period of time we've been discussing an
approach for handling new modules without the need to put them in
xdoclet's cvs. Now I'm +1 on hibernate in xdoclet's core and opening cvs
access for its maintainer. So if Gavin agrees too I can candidate you
(or maybe Gavin himself, whomever you find more appropriate for this
role).

Ara. 

> -----Original Message-----
> From: Joel Rosi-Schwartz [mailto:Joe...@Et...]
> Sent: Friday, December 06, 2002 5:15 PM
> To: hib...@li...
> Cc: ar...@ya...; Gav...@ex...
> Subject: Becoming an official XDoclet module
> 
> I know that this has come up before in the forum and dev list, but I
> think it is worth re-evaluation. I just did an experiment of adding
the
> hibernate module to the latest XDoclet cvs source. It took all of 5
> minutes to integrate the module into the Xdoclet build. The results
were
> as I expected; the build went smoothly and all of the documentation
was
> produced without any pain. So I am wondering if maintaining the
> Hibernate module seperately is really worth the extra pain and effort
> that is required.
> 
> As I understand it the main consideration is not having a maintainer
for
> the module that has commit priviledges on the XDoclet project. I agree
> that handling changes via patches is awkward. So what does it take to
> get accepted as a module and as a commiter to the Xdoclet project? If
I
> had assurances that Gavin and Ara were both willing to give me a bit
of
> support as needed, I would be willing to take on being the maintainer.
> 
> - joel

[Hibernate] Becoming an official XDoclet module

[Joel Rosi-S. <Joe...@Et...>]

I know that this has come up before in the forum and dev list, but I 
think it is worth re-evaluation. I just did an experiment of adding the 
hibernate module to the latest XDoclet cvs source. It took all of 5 
minutes to integrate the module into the Xdoclet build. The results were 
as I expected; the build went smoothly and all of the documentation was 
produced without any pain. So I am wondering if maintaining the 
Hibernate module seperately is really worth the extra pain and effort 
that is required.

As I understand it the main consideration is not having a maintainer for 
the module that has commit priviledges on the XDoclet project. I agree 
that handling changes via patches is awkward. So what does it take to 
get accepted as a module and as a commiter to the Xdoclet project? If I 
had assurances that Gavin and Ara were both willing to give me a bit of 
support as needed, I would be willing to take on being the maintainer.

- joel

Re: [Hibernate] Re: [Opensymphony-developers] A Hibranate provider for OsUser?

[Joel Rosi-S. <Joe...@Et...>]

Hi Christoph,

Yes, this is exactly the point I was bringing up. You see I *will* be 
using a Session bean, not doubts about it, that is what is right for 
*my* architecture. But other designs will call for different solutions. 
So what is the best solution as far as designing a provider or hierarchy 
of providers so that the various common usage scenarios are accounted 
for. Ideas?

- joel

Christoph Sturm wrote:

>Hi Joel!
>
>it would be great if you work on a security provider for osuser, but IMHO it
>would be good not to use a session bean for it. Then it would be also usable
>in pure servlet containers, for example resin.
>
>regards
> chris
>----- Original Message -----
>From: "Joel Rosi-Schwartz" <Joe...@Et...>
>To: "Patrick Lightbody" <pli...@ci...>
>Cc: <hib...@li...>;
><ope...@li...>
>Sent: Thursday, December 05, 2002 10:08 PM
>Subject: Re: [Hibernate] Re: [Opensymphony-developers] A Hibranate provider
>for OsUser?
>
>
>  
>
>>Patrick,
>>
>>Fine, I'll be happy to make this contribution. My plans are to use a
>>Session bean and have it retieve the the Hibernate SessionFactory via
>>JNDI. Does this suit your needs? One of the problems with integrating
>>Hibernate with OsUser is that Hibernate is so flexible and that there
>>are several applicable use cases. I have not really begun to explore how
>>to best design a provider such that it is easy to use in all them and I
>>would really appreciate any suggestion?
>>
>>- joel
>>
>>Patrick Lightbody wrote:
>>    
>>

Re: [Hibernate] Re: [Opensymphony-developers] A Hibranate provider for OsUser?

[Christoph S. <ch...@mc...>]

Hi Joel!

it would be great if you work on a security provider for osuser, but IMHO it
would be good not to use a session bean for it. Then it would be also usable
in pure servlet containers, for example resin.

regards
 chris
----- Original Message -----
From: "Joel Rosi-Schwartz" <Joe...@Et...>
To: "Patrick Lightbody" <pli...@ci...>
Cc: <hib...@li...>;
<ope...@li...>
Sent: Thursday, December 05, 2002 10:08 PM
Subject: Re: [Hibernate] Re: [Opensymphony-developers] A Hibranate provider
for OsUser?


> Patrick,
>
> Fine, I'll be happy to make this contribution. My plans are to use a
> Session bean and have it retieve the the Hibernate SessionFactory via
> JNDI. Does this suit your needs? One of the problems with integrating
> Hibernate with OsUser is that Hibernate is so flexible and that there
> are several applicable use cases. I have not really begun to explore how
> to best design a provider such that it is easy to use in all them and I
> would really appreciate any suggestion?
>
> - joel
>
> Patrick Lightbody wrote:
>
> >Joel,
> >I don't think one has been created yet. If you could write one, I'll get
it
> >in CVS right away (I will need it very soon as well, so that'd be great
if
> >you could get it done).
> >
> >-Pat
> >
> >----- Original Message -----
> >From: "Joel Rosi-Schwartz" <Joe...@Et...>
> >To: <hib...@li...>;
> ><ope...@li...>;
> ><ope...@li...>
> >Sent: Thursday, December 05, 2002 4:31 AM
> >Subject: [Opensymphony-developers] A Hibranate provider for OsUser?
> >
> >
> >
> >
> >>Has anyone created a Hibernate provider for Open Symphony's OsUser
> >>module that they are willing to share? If so I would very much like to
> >>have a copy. If not I will put one together in the next few days and
> >>share it back to the OsUser project.
> >>
> >>Thanks,
> >>Joel
> >>
> >>
>
>
>
>
> -------------------------------------------------------
> This sf.net email is sponsored by:ThinkGeek
> Welcome to geek heaven.
> http://thinkgeek.com/sf
> _______________________________________________
> hibernate-devel mailing list
> hib...@li...
> https://lists.sourceforge.net/lists/listinfo/hibernate-devel
>

[Hibernate] Java-based query evaluator

[Chris N. <ch...@si...>]

I'm curious how feasible it would be to write a Java-based evaluator of
Hibernate queries, or at least leverage the existing parser code to gain
access to the abstract syntax tree. 

One possibility is that certain subsets of queries (most filters, I
think) could be run within the JVM, saving a db query.

Also, a translator could allow Hibernate query language to be used with
other persistence engines (ok, this is not so compelling since Hibernate
is the best :)

But really, the query language seems very clean and simple, and I
believe has the potential to be useful in scenarios even outside of
persistence. Kind of like JXPath, but more SQL-like. This way all the
queries in my application could use the same language, regardless of
whether there is an underlying database.

I've looked at the cirrus.hibernate.query package and most of it is
pretty generic, but there are some obvious places where the persistence
stuff creeps in. Is it worth persuing reusing this code? If not, do you
anticipate many changes to the query language going forward?

Thanks,
Chris

RE: [Hibernate] Suggestion for better error reports...

[Gavin K. <Gav...@ex...>]

Do It! :)

> -----Original Message-----
> From: Max Rydahl Andersen [mailto:ma...@eo...]=20
> Sent: Friday, 6 December 2002 2:27 AM
> To: hib...@li...
> Subject: [Hibernate] Suggestion for better error reports...
>=20
>=20
> Hi!
>=20
> By "accident" I updated an object instance which id was=20
> associated with another loaded instance in the current=20
> instance and got the following error:
>=20
> "Another object was associated with this id (the object with=20
> the given id was already loaded)"
>=20
> That is all fine, but would it not be nice if the exception=20
> message contained some more information about this object ?=20
> Currently we have to enable trace to see which class and id=20
> that is related to the exception.
>=20
> With trace enabled you would get:
>  "updating eg.FooBar#42"
>=20
> This is actually a typical scenario for a couple of dozens=20
> exceptions thrown when saving, updating, etc.
>=20
> How about adding that vital information to the exception=20
> message without enabling the trace ?
>=20
> So we instead would get:
>=20
> "Another object was associated with this id (the object with=20
> the given id was already loaded) [eg.FooBar#42]"
>=20
> Any comments ?
>=20
> Otherwise I will send a patch for this change ?
>=20
> /max
>=20
>=20
>=20
> -------------------------------------------------------
> This sf.net email is sponsored by:ThinkGeek
> Welcome to geek heaven.
> http://thinkgeek.com/sf=20
> _______________________________________________
> hibernate-devel mailing list hib...@li...
> https://lists.sourceforge.net/lists/listinfo/hibernate-devel
>=20


********** CAUTION - Disclaimer **********
This message may contain privileged and confidential
information. If you are not the intended recipient of this
message (or responsible for delivery of the message to
such person) you are hereby notified that any use,
dissemination, distribution or reproduction of this message
is prohibited. If you have received this message in error,
you should destroy it and kindly notify the sender by reply
e-mail. Please advise immediately if you or your employer
do not consent to Internet e-mail for messages of this kind.
Opinions, conclusions and other information in this
message that do not relate to the official business of
Expert Information Services Pty Ltd ("The Company")
shall be understood as neither given nor endorsed by it.

The Company advises that this e-mail and any attached
files should be scanned to detect viruses. The Company
accepts no liability for loss or damage (whether caused
by negligence or not) resulting from the use of any
attached files.
**EIS******** End of Disclaimer **********

Re: [Hibernate] Re: [Opensymphony-developers] A Hibranate provider for OsUser?

[Joel Rosi-S. <Joe...@Et...>]

Patrick,

Fine, I'll be happy to make this contribution. My plans are to use a 
Session bean and have it retieve the the Hibernate SessionFactory via 
JNDI. Does this suit your needs? One of the problems with integrating 
Hibernate with OsUser is that Hibernate is so flexible and that there 
are several applicable use cases. I have not really begun to explore how 
to best design a provider such that it is easy to use in all them and I 
would really appreciate any suggestion?

- joel

Patrick Lightbody wrote:

>Joel,
>I don't think one has been created yet. If you could write one, I'll get it
>in CVS right away (I will need it very soon as well, so that'd be great if
>you could get it done).
>
>-Pat
>
>----- Original Message -----
>From: "Joel Rosi-Schwartz" <Joe...@Et...>
>To: <hib...@li...>;
><ope...@li...>;
><ope...@li...>
>Sent: Thursday, December 05, 2002 4:31 AM
>Subject: [Opensymphony-developers] A Hibranate provider for OsUser?
>
>
>  
>
>>Has anyone created a Hibernate provider for Open Symphony's OsUser
>>module that they are willing to share? If so I would very much like to
>>have a copy. If not I will put one together in the next few days and
>>share it back to the OsUser project.
>>
>>Thanks,
>>Joel
>>    
>>

[Hibernate] Re: [Opensymphony-developers] A Hibranate provider for OsUser?

[Patrick L. <pli...@ci...>]

Joel,
I don't think one has been created yet. If you could write one, I'll get it
in CVS right away (I will need it very soon as well, so that'd be great if
you could get it done).

-Pat

----- Original Message -----
From: "Joel Rosi-Schwartz" <Joe...@Et...>
To: <hib...@li...>;
<ope...@li...>;
<ope...@li...>
Sent: Thursday, December 05, 2002 4:31 AM
Subject: [Opensymphony-developers] A Hibranate provider for OsUser?


> Has anyone created a Hibernate provider for Open Symphony's OsUser
> module that they are willing to share? If so I would very much like to
> have a copy. If not I will put one together in the next few days and
> share it back to the OsUser project.
>
> Thanks,
> Joel
>
>
>
> -------------------------------------------------------
> This sf.net email is sponsored by:ThinkGeek
> Welcome to geek heaven.
> http://thinkgeek.com/sf
> _______________________________________________
> Opensymphony-developers mailing list
> Ope...@li...
> https://lists.sourceforge.net/lists/listinfo/opensymphony-developers

[Hibernate] Suggestion for better error reports...

[Max R. A. <ma...@eo...>]

Hi!

By "accident" I updated an object instance which id was associated with
another loaded instance in the current instance and got the following error:

"Another object was associated with this id (the object with the given id
was already loaded)"

That is all fine, but would it not be nice if the exception message
contained some more information about this object ?
Currently we have to enable trace to see which class and id that is related
to the exception.

With trace enabled you would get:
 "updating eg.FooBar#42"

This is actually a typical scenario for a couple of dozens exceptions thrown
when saving, updating, etc.

How about adding that vital information to the exception message without
enabling the trace ?

So we instead would get:

"Another object was associated with this id (the object with the given id
was already loaded) [eg.FooBar#42]"

Any comments ?

Otherwise I will send a patch for this change ?

/max

[Hibernate] A Hibranate provider for OsUser?

[Joel Rosi-S. <Joe...@Et...>]

Has anyone created a Hibernate provider for Open Symphony's OsUser 
module that they are willing to share? If so I would very much like to 
have a copy. If not I will put one together in the next few days and 
share it back to the OsUser project.

Thanks,
Joel

RE: [Hibernate] Re: Hibernate vs. Castor example

[Gavin K. <Gav...@ex...>]

I have performed some extra testing ... this functionality
seems to work correctly....

You'll have to give some more details.

> -----Original Message-----
> From: Dave Johnson [mailto:sno...@nc...]=20
> Sent: Wednesday, 4 December 2002 11:41 PM
> To: Gavin King
> Subject: Re: [Hibernate] Re: Hibernate vs. Castor example
>=20
>=20
> I don't have any top level collections.  Top level=20
> collections are those=20
> that exist outside of <class> mapping, right?  When I take out the=20
> ses.load(), my tests fail. When I put it back in they work again.=20
>  Perhaps a bug?
>=20
> - Dave
>=20
>=20
> Gavin King wrote:
>=20
> >It isn't usually; not unless you have toplevel collections...
> >
> > =20
> >
> >>-----Original Message-----
> >>From: Dave Johnson [mailto:sno...@nc...]
> >>Sent: Wednesday, 4 December 2002 4:09 PM
> >>To: hib...@li...
> >>Subject: [Hibernate] Re: Hibernate vs. Castor example
> >>
> >>
> >>
> >>In the Hibernate implementation of the DAO, I implement the
> >>removeNewsfeed method shown below. Why is the call to=20
> >>ses.load() necessary?
> >>
> >>- Dave
> >>
> >>
> >>   public void removeNewsfeed( Newsfeed feed ) throws DAOException {
> >>     removeObject( ag.Newsfeed.class, feed.getId(), feed );
> >>   }
> >>
> >>   private void removeObject( Class clazz, String id, Object obj )
> >>     throws DAOException {
> >>     Session ses =3D null;
> >>     try {
> >>       ses =3D sessionFactory.openSession();
> >>       obj =3D ses.load(clazz,id);
> >>       ses.delete(obj);
> >>       ses.flush();
> >>       ses.connection().commit();
> >>     }
> >>     catch (Exception e) {
> >>       try { ses.connection().rollback(); }
> >>       catch (Exception ex) { e.printStackTrace(); };
> >>       throw new DAOException(e);
> >>     }
> >>     finally {
> >>       try { ses.close(); }
> >>       catch (Exception ex) { ex.printStackTrace(); };
> >>     }
> >>   }
> >>
> >>
> >>
> >>
> >>
> >>-------------------------------------------------------
> >>This SF.net email is sponsored by: Microsoft Visual Studio.NET
> >>comprehensive development tool, built to increase your=20
> >>productivity. Try a free online hosted session at:=20
> >>   =20
> >>
> >http://ads.sourceforge.net/cgi-bin/redirect.pl?micr0003en
> >_______________________________________________
> >hibernate-devel mailing list hib...@li...
> >https://lists.sourceforge.net/lists/listinfo/hibernate-devel
> >
> >
> >********** CAUTION - Disclaimer **********
> >This message may contain privileged and confidential information. If=20
> >you are not the intended recipient of this message (or=20
> responsible for=20
> >delivery of the message to such person) you are hereby notified that=20
> >any use, dissemination, distribution or reproduction of this message
> >is prohibited. If you have received this message in error,
> >you should destroy it and kindly notify the sender by reply
> >e-mail. Please advise immediately if you or your employer
> >do not consent to Internet e-mail for messages of this kind.
> >Opinions, conclusions and other information in this
> >message that do not relate to the official business of
> >Expert Information Services Pty Ltd ("The Company")
> >shall be understood as neither given nor endorsed by it.
> >
> >The Company advises that this e-mail and any attached
> >files should be scanned to detect viruses. The Company
> >accepts no liability for loss or damage (whether caused
> >by negligence or not) resulting from the use of any
> >attached files.
> >**EIS******** End of Disclaimer **********
> >
> >
> > =20
> >
>=20
>=20
>=20


********** CAUTION - Disclaimer **********
This message may contain privileged and confidential
information. If you are not the intended recipient of this
message (or responsible for delivery of the message to
such person) you are hereby notified that any use,
dissemination, distribution or reproduction of this message
is prohibited. If you have received this message in error,
you should destroy it and kindly notify the sender by reply
e-mail. Please advise immediately if you or your employer
do not consent to Internet e-mail for messages of this kind.
Opinions, conclusions and other information in this
message that do not relate to the official business of
Expert Information Services Pty Ltd ("The Company")
shall be understood as neither given nor endorsed by it.

The Company advises that this e-mail and any attached
files should be scanned to detect viruses. The Company
accepts no liability for loss or damage (whether caused
by negligence or not) resulting from the use of any
attached files.
**EIS******** End of Disclaimer **********

RE: [Hibernate] Enhancements to the Hibernate Xdoclet subtasks

[Gavin K. <Gav...@ex...>]

>Okay, I have gone through the Hibernate schema and the=20
>Xdoclet xdt and have I made the changes that I *think*=20
>brings the module in line with changes made for 1.1. I=20
>have attached the xdt and would appreciate if those of=20
>you who are more familiar with these could double check=20
>what I have done.

Looks all fine to me!


>After I get some feed back on this I will update the unit=20
>test to reflect these changes and make certain everything=20
>actually runs. Then I will commit this stage back to cvs.

Go ahead...


Thanks heaps

Gavin



********** CAUTION - Disclaimer **********
This message may contain privileged and confidential
information. If you are not the intended recipient of this
message (or responsible for delivery of the message to
such person) you are hereby notified that any use,
dissemination, distribution or reproduction of this message
is prohibited. If you have received this message in error,
you should destroy it and kindly notify the sender by reply
e-mail. Please advise immediately if you or your employer
do not consent to Internet e-mail for messages of this kind.
Opinions, conclusions and other information in this
message that do not relate to the official business of
Expert Information Services Pty Ltd ("The Company")
shall be understood as neither given nor endorsed by it.

The Company advises that this e-mail and any attached
files should be scanned to detect viruses. The Company
accepts no liability for loss or damage (whether caused
by negligence or not) resulting from the use of any
attached files.
**EIS******** End of Disclaimer **********

RE: [Hibernate] Re: Hibernate and JCS Questions

[Gavin K. <Gav...@ex...>]

Oh .... I hadn't realized that the original message was so=20
widely cross-posted.

I answered the original post more completely on the Hibernate
list, but I should also just confirm James' speculation that=20
the reason we don't support use of JCS distributed caching with=20
Hibernate is that we are extra fussy about ensuring transaction=20
isolation and that implies some kind of locking of cached data.

Note that Hibernate has a pluggable cache architecture so it
is actually possible to plug in a caching strategy that has
weaker consistency guarantees than those assumed by Hibernate's
built-in strategies.

Gavin King
http://hibernate.sourceforge.net


> -----Original Message-----
> From: James Taylor [mailto:ja...@ja...]=20
> Sent: Thursday, 5 December 2002 1:56 PM
> To: Turbine JCS Users List
> Cc: 'hib...@li...'; Blair Harrison;=20
> Tatum Lade; Lee Parayno
> Subject: [Hibernate] Re: Hibernate and JCS Questions
>=20
>=20
> > 1) How mature is JCS? Who is using it? On how many machines? With=20
> > which of the three major configurations (fully distributed,=20
> > client/server, client/cluster)?
>=20
> I am using it. Two machines, memory caches on each. Put only=20
> lateral distribution between them with javagroups. Quite basic really.
>=20
> Others? (I'm curious too)
>=20
> > 2) What was the process by which Gavin gained sufficient=20
> confidence in=20
> > JCS to include it in Hibernate?
> >=20
> > 3) Why does Gavin indicate that JCS can't be used in a clustered=20
> > read-write environment, in contradiction to the JCS documentation?
>=20
> I suspect the biggest problem is that there is no distributed=20
> locking, for JCS, which makes it difficult to use for O/R=20
> mapping. In my case, we cache immutable read mostly data=20
> objects, and when the underlying data store is updated for an=20
> entity, it is purged from the cache.=20
>=20
> For more mutable data, I do not cache.
>=20
> > 4) If we were able to commit to using Hibernate in a=20
> read-only fashion=20
> > and doing our data writing some other way, changing the=20
> data out from=20
> > under Hibernate, which various legacy admin tools will do=20
> anyway, does=20
> > this change the preference profile between Hibernate's=20
> session-level=20
> > cache vs. JCS? If so, how?
> >=20
> > The bottom line, quite frankly, is that we're nervous about=20
> relying on=20
> > JCS in the absence of some reassurances about its maturity=20
> and use in=20
> > the real world, as JCS is clearly attempting to solve difficult=20
> > engineering issues,
>=20
> I wouldn't call JCS mature, but it works reasonably well for=20
> some peoples needs, and is a good place to start for a=20
> distributed cache. But because it has no consistency=20
> guarantee between nodes in the cache, I'm not sure it is=20
> workable for O/R mapping and multiple nodes.
>=20
> (However, I believe one could implement an auxilliary which=20
> provided more consistency -- all the hooks should be there)
>=20
>=20
>=20
> -------------------------------------------------------
> This SF.net email is sponsored by: Microsoft Visual Studio.NET=20
> comprehensive development tool, built to increase your=20
> productivity. Try a free online hosted session at:=20
http://ads.sourceforge.net/cgi-bin/redirect.pl?micr0003en
_______________________________________________
hibernate-devel mailing list hib...@li...
https://lists.sourceforge.net/lists/listinfo/hibernate-devel


********** CAUTION - Disclaimer **********
This message may contain privileged and confidential
information. If you are not the intended recipient of this
message (or responsible for delivery of the message to
such person) you are hereby notified that any use,
dissemination, distribution or reproduction of this message
is prohibited. If you have received this message in error,
you should destroy it and kindly notify the sender by reply
e-mail. Please advise immediately if you or your employer
do not consent to Internet e-mail for messages of this kind.
Opinions, conclusions and other information in this
message that do not relate to the official business of
Expert Information Services Pty Ltd ("The Company")
shall be understood as neither given nor endorsed by it.

The Company advises that this e-mail and any attached
files should be scanned to detect viruses. The Company
accepts no liability for loss or damage (whether caused
by negligence or not) resulting from the use of any
attached files.
**EIS******** End of Disclaimer **********

[Hibernate] Re: Hibernate and JCS Questions

[James T. <ja...@ja...>]

> 1) How mature is JCS? Who is using it? On how many machines? With which of
> the three major configurations (fully distributed, client/server,
> client/cluster)?

I am using it. Two machines, memory caches on each. Put only lateral
distribution between them with javagroups. Quite basic really.

Others? (I'm curious too)

> 2) What was the process by which Gavin gained sufficient confidence in JCS
> to include it in Hibernate?
> 
> 3) Why does Gavin indicate that JCS can't be used in a clustered read-write
> environment, in contradiction to the JCS documentation?

I suspect the biggest problem is that there is no distributed locking,
for JCS, which makes it difficult to use for O/R mapping. In my case, we
cache immutable read mostly data objects, and when the underlying data
store is updated for an entity, it is purged from the cache. 

For more mutable data, I do not cache.

> 4) If we were able to commit to using Hibernate in a read-only fashion and
> doing our data writing some other way, changing the data out from under
> Hibernate, which various legacy admin tools will do anyway, does this change
> the preference profile between Hibernate's session-level cache vs. JCS? If
> so, how?
> 
> The bottom line, quite frankly, is that we're nervous about relying on JCS
> in the absence of some reassurances about its maturity and use in the real
> world, as JCS is clearly attempting to solve difficult engineering issues,

I wouldn't call JCS mature, but it works reasonably well for some
peoples needs, and is a good place to start for a distributed cache. But
because it has no consistency guarantee between nodes in the cache, I'm
not sure it is workable for O/R mapping and multiple nodes.

(However, I believe one could implement an auxilliary which provided
more consistency -- all the hooks should be there)

RE: [Hibernate] Hibernate and JCS Questions

[Gavin K. <Gav...@ex...>]

> 2) What was the process by which Gavin gained sufficient=20
> confidence in JCS to include it in Hibernate?

Christian Muenier has done quite a lot of research and testing
with JCS and it was him who recommended use of JCS in Hibernate.
He has implied to me (privately) that he does not fully trust
JCS distributed caching.

> 3) Why does Gavin indicate that JCS can't be used in a=20
> clustered read-write environment, in contradiction to the JCS=20
> documentation?

Firstly, please understand that Hibernate is geared towards
highly transactional systems and is very very fussy about
ensuring that transaction isolation is always preserved. JCS
distributed caching was not designed with this kind of usage
in mind.

Secondly, to support that kind of strict transaction isolation,
Hibernate sits some metadata in on top of the cached
object. The way this is done is *not* safe for use with a
distributed cache.

Now, I fully understand that not all applications require this=20
level of transaction isolation and it sounds like your application=20
is one of those. Its quite easy to add a new caching strategy that=20
implements read-write caching with weaker isolation semantics.
(This should not take much development effort on your part.)
=20
> 4) If we were able to commit to using Hibernate in a=20
> read-only fashion and doing our data writing some other way,=20
> changing the data out from under Hibernate, which various=20
> legacy admin tools will do anyway, does this change the=20
> preference profile between Hibernate's session-level cache=20
> vs. JCS? If so, how?

Oh...nothing so extreme as that! All you would need is two
session factories, one with no cache, one with a read-only
cache (and a sufficiently short timeout) ... you don't need
To scrap Hibernate all together! But the correct solution
is simply to add a new implementation of=20
CacheConcurrencyStrategy.

> The bottom line, quite frankly, is that we're nervous about=20
> relying on JCS in the absence of some reassurances about its=20
> maturity and use in the real world, as JCS is clearly=20
> attempting to solve difficult engineering issues, and this=20

Hibernate has a pluggable caching architecture, though that's
not currently exposed to the user. If this is a critical issue
for you, you should probably spend a little while investigating
other cache implementations. It would *not* be difficult to plug
them into Hibernate.

Gavin


********** CAUTION - Disclaimer **********
This message may contain privileged and confidential
information. If you are not the intended recipient of this
message (or responsible for delivery of the message to
such person) you are hereby notified that any use,
dissemination, distribution or reproduction of this message
is prohibited. If you have received this message in error,
you should destroy it and kindly notify the sender by reply
e-mail. Please advise immediately if you or your employer
do not consent to Internet e-mail for messages of this kind.
Opinions, conclusions and other information in this
message that do not relate to the official business of
Expert Information Services Pty Ltd ("The Company")
shall be understood as neither given nor endorsed by it.

The Company advises that this e-mail and any attached
files should be scanned to detect viruses. The Company
accepts no liability for loss or damage (whether caused
by negligence or not) resulting from the use of any
attached files.
**EIS******** End of Disclaimer **********

[Hibernate] Hibernate and JCS Questions

[Paul S. <psn...@IF...>]

Gentlemen,

My name is Paul Snively. I'm the senior Java architect for
http://www.ifilm.com, the seventh-largest video-streaming service in the
world. We are in the process of redeploying our portal using Java
technology. To that end, we have been investigating Hibernate extensively,
and are now dealing with questions surrounding data caching. Our expected
infrastructure for this redeployment involves ~10 application servers,
possibly with another layer of a hopefully smaller number of machines, but
still >1, acting as data servers.

The Hibernate documentation goes into some detail regarding caching options,
and Gavin King has been kind enough to expand on that information still
further. In particular, our current understanding is that the use of JCS in
Hibernate is recommended for clustered read-only environments, but cannot be
used in clustered read-write environments. On the other hand, the JCS
documents make quite clear that clustered "put" operations are supported and
expected to work. Following Gavin's lead, one alternative that suggested
itself to us was simply to rely on Hibernate's session-level cache
exclusively and not use JCS at all, but feedback from Gavin suggests in no
uncertain terms that such an approach is counterindicated.

So our questions basically boil down to the following:

1) How mature is JCS? Who is using it? On how many machines? With which of
the three major configurations (fully distributed, client/server,
client/cluster)?

2) What was the process by which Gavin gained sufficient confidence in JCS
to include it in Hibernate?

3) Why does Gavin indicate that JCS can't be used in a clustered read-write
environment, in contradiction to the JCS documentation?

4) If we were able to commit to using Hibernate in a read-only fashion and
doing our data writing some other way, changing the data out from under
Hibernate, which various legacy admin tools will do anyway, does this change
the preference profile between Hibernate's session-level cache vs. JCS? If
so, how?

The bottom line, quite frankly, is that we're nervous about relying on JCS
in the absence of some reassurances about its maturity and use in the real
world, as JCS is clearly attempting to solve difficult engineering issues,
and this component will prove critical to our success should we choose to
use it. The clearest alternative to us is to rely on Hibernate's own
session-level caching and to do whatever we need to do in order to make that
as efficient as humanly possible. Any feedback about our options would be
most greatly appreciated.

Many thanks,
Paul Snively

RE: [Hibernate] Re: Hibernate vs. Castor example

[Gavin K. <Gav...@ex...>]

> I don't have any top level collections.  Top level=20
> collections are those=20
> that exist outside of <class> mapping, right?  When I take out the=20
> ses.load(), my tests fail. When I put it back in they work again.=20
>  Perhaps a bug?

Hmmmm. Perhaps. How does the test fail? Exception stack trace?


********** CAUTION - Disclaimer **********
This message may contain privileged and confidential
information. If you are not the intended recipient of this
message (or responsible for delivery of the message to
such person) you are hereby notified that any use,
dissemination, distribution or reproduction of this message
is prohibited. If you have received this message in error,
you should destroy it and kindly notify the sender by reply
e-mail. Please advise immediately if you or your employer
do not consent to Internet e-mail for messages of this kind.
Opinions, conclusions and other information in this
message that do not relate to the official business of
Expert Information Services Pty Ltd ("The Company")
shall be understood as neither given nor endorsed by it.

The Company advises that this e-mail and any attached
files should be scanned to detect viruses. The Company
accepts no liability for loss or damage (whether caused
by negligence or not) resulting from the use of any
attached files.
**EIS******** End of Disclaimer **********

RE: [Hibernate] Global Components

[Gavin K. <Gav...@ex...>]

Just use an XML external entity to save yourself the copy/paste.

:)

> -----Original Message-----
> From: Aapo Laakkonen [mailto:aap...@pr...]=20
> Sent: Thursday, 5 December 2002 6:09 AM
> To: hib...@li...
> Subject: [Hibernate] Global Components
>=20
>=20
> I have need for global components. Here is example of one:
>=20
> <component name=3D"metadata" class=3D"eg.Metadata">
>   <property name=3D"createdOn" column=3D"created_on"=20
> type=3D"timestamp" unique=3D"false"/>
>   <many-to-one name=3D"createdBy" class=3D"eg.User"=20
> column=3D"created_by" unique=3D"false"/>
>   <property name=3D"accepted" column=3D"accepted" type=3D"boolean"=20
> unique=3D"false"/>
>   <many-to-one name=3D"acceptedBy" class=3D"eg.User"=20
> column=3D"accepted_by" unique=3D"false"/>
>   <property name=3D"archived" column=3D"archived" type=3D"boolean"=20
> unique=3D"false"/>
>   <property name=3D"archivedOn" column=3D"archived_on"=20
> type=3D"timestamp" unique=3D"false"/>
>   <many-to-one name=3D"archivedBy" class=3D"eg.User"=20
> column=3D"archived_by" unique=3D"false"/>
>   <property name=3D"modifiedOn" column=3D"modified_on"=20
> type=3D"timestamp" unique=3D"false"/>
>   <many-to-one name=3D"modifiedBy" class=3D"eg.User"=20
> column=3D"modified_by" unique=3D"false"/>
>   <property name=3D"deleted" column=3D"deleted" type=3D"boolean"=20
> unique=3D"false"/>
>   <property name=3D"deletedOn" column=3D"deleted_on"=20
> type=3D"timestamp" unique=3D"false"/>
>   <many-to-one name=3D"deletedBy" class=3D"eg.User"=20
> column=3D"deleted_by" unique=3D"false"/> </component>
>=20
>=20
> With current implementation I have to copy this definition to=20
> all of my mapping files that need to provide metadata. Is=20
> there workaround to this? Or is it just me who ever likes to=20
> see this kind of functionality?
>=20
>=20
>=20
> Kind Regards
> Aapo <b/>ungle Laakkonen
>=20
>=20
>=20
>=20
> -------------------------------------------------------
> This SF.net email is sponsored by: Microsoft Visual Studio.NET=20
> comprehensive development tool, built to increase your=20
> productivity. Try a free online hosted session at:=20
http://ads.sourceforge.net/cgi-bin/redirect.pl?micr0003en
_______________________________________________
hibernate-devel mailing list hib...@li...
https://lists.sourceforge.net/lists/listinfo/hibernate-devel


********** CAUTION - Disclaimer **********
This message may contain privileged and confidential
information. If you are not the intended recipient of this
message (or responsible for delivery of the message to
such person) you are hereby notified that any use,
dissemination, distribution or reproduction of this message
is prohibited. If you have received this message in error,
you should destroy it and kindly notify the sender by reply
e-mail. Please advise immediately if you or your employer
do not consent to Internet e-mail for messages of this kind.
Opinions, conclusions and other information in this
message that do not relate to the official business of
Expert Information Services Pty Ltd ("The Company")
shall be understood as neither given nor endorsed by it.

The Company advises that this e-mail and any attached
files should be scanned to detect viruses. The Company
accepts no liability for loss or damage (whether caused
by negligence or not) resulting from the use of any
attached files.
**EIS******** End of Disclaimer **********

[Hibernate] RE: Lazy Collections (to maverick users)

[Aapo L. <aap...@pr...>]

> What guarantees does Maverick make about when / wether discard() will
be
> called? Is it *guaranteed* to be called in the case of an exception,
etc?

I checked this today, and found out that it gives tou no guarantees,
until I posted a patch.

> -----Original Message-----
> From: Aapo Laakkonen [mailto:aap...@pr...]
> Sent: Wednesday, December 04, 2002 5:02 AM
> To: mav...@li...
> Subject: [Mav-user] FW: ModelLifeTime / discard() not guaranteed
> 
> I think that Maverick should guarantee that discard() method of 
> ModelLifetime implementing classes get's called whatever happens
> during rendering view or executing command.
> 
> Here is my modifications to CommandBase class go() method:
>
> ...


Good idea.  Now checked into CVS.

Jeff

[Hibernate] Transaction Commit & Session Close

[Aapo L. <aap...@pr...>]

Here is my settings:

<hibernate-configuration>

  <session-factory name="java:comp/env/jndi/HibernateSessionFactory">

    <property
name="connection.datasource">java:comp/env/jdbc/xa/postgresql/somedb</pr
operty>
    <property
name="dialect">cirrus.hibernate.sql.PostgreSQLDialect</property>

    ... mapping files ...

    </session-factory>

</hibernate-configuration>

hibernate.use_outer_join=true
hibernate.show_sql=false
hibernate.jdbc.batch_size=10
hibernate.statement_cache.size=20
hibernate.transaction.factory_class=cirrus.hibernate.transaction.JTATran
sactionFactory
hibernate.transaction.manager_lookup_class=cirrus.hibernate.transaction.
ResinTransactionManagerLookup
hibernate.query.substitutions yes 'Y', no 'N'
jta.UserTransaction=java:comp/UserTransaction


I start transaction with this code:

Session session = null;
cirrus.hibernate.Transaction tx = null;

try {

  if (log.isDebugEnabled())
    log.debug("Getting Hibernate session.");

  session = this.sessionFactory.openSession();

  if (log.isInfoEnabled())
    log.info("Starting database transaction.");

  tx = session.beginTransaction();

And I end the transaction like this:

  if (log.isInfoEnabled())
    log.info("Committing database transaction.");

    tx.commit();

    if (log.isInfoEnabled())
      log.info("Database transaction committed successfully.");

  } catch (SQLException e) {
    log.error(e);
    this.rollbackTransaction(tx); // just an exception handling helper
func
    throw new ServletException(e);
  } catch (HibernateException e) {
    log.error(e);
    this.rollbackTransaction(tx); // just an exception handling helper
func
    throw new ServletException(e);
  } catch (Exception e) {
    log.error(e);
    this.rollbackTransaction(tx); // just an exception handling helper
func
    throw new ServletException(e);
  } finally {

    try { session.close();
    } catch (SQLException e) {
    } catch (HibernateException e) {}

    if (log.isDebugEnabled())
      log.debug("Closed Hibernate session.");
  }
}

Everything goes fine until finally block. It gives me warnings:

21:35:24,381 WARN  [JDBCExceptionReporter] SQL Warning: 0, SQLState:
null
21:35:24,381 WARN  [JDBCExceptionReporter] WARNING:  COMMIT: no
transaction in progress

What causes these? Am I doing something wrong? If I tx.commit, do I
still need to close session? If I do, then how do I close session
without the warnings mentioned above?


Regards
Aapo <b/>ungle Laakkonen

[Hibernate] Global Components

[Aapo L. <aap...@pr...>]

I have need for global components. Here is example of one:

<component name="metadata" class="eg.Metadata">
  <property name="createdOn" column="created_on" type="timestamp"
unique="false"/>
  <many-to-one name="createdBy" class="eg.User" column="created_by"
unique="false"/>
  <property name="accepted" column="accepted" type="boolean"
unique="false"/>
  <many-to-one name="acceptedBy" class="eg.User" column="accepted_by"
unique="false"/>
  <property name="archived" column="archived" type="boolean"
unique="false"/>
  <property name="archivedOn" column="archived_on" type="timestamp"
unique="false"/>
  <many-to-one name="archivedBy" class="eg.User" column="archived_by"
unique="false"/>
  <property name="modifiedOn" column="modified_on" type="timestamp"
unique="false"/>
  <many-to-one name="modifiedBy" class="eg.User" column="modified_by"
unique="false"/>
  <property name="deleted" column="deleted" type="boolean"
unique="false"/>
  <property name="deletedOn" column="deleted_on" type="timestamp"
unique="false"/>
  <many-to-one name="deletedBy" class="eg.User" column="deleted_by"
unique="false"/>
</component>


With current implementation I have to copy this definition to all of my
mapping files that need to provide metadata. Is there workaround to
this? Or is it just me who ever likes to see this kind of functionality?



Kind Regards
Aapo <b/>ungle Laakkonen