Yes, I have test this but that doesn't functionned.
There is a bug of Nhibernate Linq ?
But if i do this :
var ldossier = (from dos in
session.LinqDossier().Expand(ListeLogicielDossier)
where dos.CodeDossier == 0114
select
I can't see a warning in the log.
It verifies that every (non-private) member is at least virtual
protected. internal protected is accepted by the session factory,
but doesn't seem to work.
On 22 Mrz., 17:12, Fabio Maulo fabioma...@gmail.com wrote:
I think you should have a WARNING in the log.
You can either specifiy eager loader with the criteria api by
using .SetFetchMode(YourOneToOneRelation, FetchMode.Eager). You can
also do stuff like this:
.SetFetchMode(RelationToA, FetchMode.Select)
.SetFetchMode(RelationToA.RelationAToB, FetchMode.Eager)
Which would not immediately result in
I have the following database model:
Data -- DataLanguage -- Language
In code I can access these like this:
Data.Languages
I want to do a eager load of these items on data status. The way I
would do that in SQL is this:
1st roundtrip: Fetch matching data items and the data language keys
Hi guys,
I keep getting IndexOutOfRangeException's on the current trunk (with code
that worked fine in NH 2.1) because it's looking for the primary key in the
resultset with square brackets round the column name ([QueuedTaskId]
instead of QueuedTaskId) - I've checked and the exception doesn't
which is the query ?
2010/3/24 James Crowley james.crow...@gmail.com
Hi guys,
I keep getting IndexOutOfRangeException's on the current trunk (with code
that worked fine in NH 2.1) because it's looking for the primary key in the
resultset with square brackets round the column name
The ` quotes on the column name are probably doing it.
On Wed, Mar 24, 2010 at 6:38 AM, James Crowley james.crow...@gmail.comwrote:
Hi guys,
I keep getting IndexOutOfRangeException's on the current trunk (with code
that worked fine in NH 2.1) because it's looking for the primary key in the
Hi,
Without questioning the implementation ;), it has nothing to do with
proxying as the proxy should have loaded all values as a result from
Assert.AreEqual(A, obj.PrivateSetter)
Or is it caused by the private set {} itself ?? (What is the mapping
here nosetter.camelcase ???), could be that
I checked and the `quotes are a valid syntax for escaping table names in the
NH mapping (it was discussed here:
http://code.google.com/p/fluent-nhibernate/issues/detail?id=55)
And just to follow up, it does look like a NamedQuery issue after all.
Calling
Session.GetQueuedJob((long)0); //
OK, I just tried with the castle dynamic proxies (before i had LinFu)
AND IT WORKS!! This means, that it is a problem with LinFu proxies!
The proxy did load the values, but they are not accessible by the non-
public members. NH seems to create two instances of the entity, one
that is derived from
If you put square brackets (quoting in the mapping) you should put square
brackets in your SP.
If you use a mapping-system that puts square brackets everywhere, instead
uses it only where really needed (as NH2.1.0 do), you should adapt your work
to your choose.
I don't need to do it because: I
I have a weird issue in a WebForms app that is using session per
request. On a certain screen we update some fields, update the entity
and then save it. The problem occurs if we set an entity to not
active. The select queries exclude entities that are not active. So,
looking at NH profiler, the
Fabio,
I'll get working on ConfORM and let you know.
The xml snippet is supposed be telling NH that the ID for the primary
table is also the key into the Plan object (which is my one-to-one
association).
On Mar 24, 12:01 am, Fabio Maulo fabioma...@gmail.com wrote:
Try to map your domain with
You speak of updating and saving an entity. Why is there a select
there at all? I think you need to explain more about how your code
works.
/Oskar
2010/3/24 Sal salbass...@hotmail.com:
I have a weird issue in a WebForms app that is using session per
request. On a certain screen we update some
The app seems to run the same without that snippet, so consider it
removed.
On Mar 24, 10:07 am, jprid64 john.pride...@gmail.com wrote:
Fabio,
I'll get working on ConfORM and let you know.
The xml snippet is supposed be telling NH that the ID for the primary
table is also the key into the
It needs to select to bind the entities to the view.
On Mar 24, 11:07 am, Oskar Berggren oskar.bergg...@gmail.com wrote:
You speak of updating and saving an entity. Why is there a select
there at all? I think you need to explain more about how your code
works.
/Oskar
2010/3/24 Sal
More specifically, to you mean rebind after the entity have been updated?
/Oskar
2010/3/24 Sal salbass...@hotmail.com:
It needs to select to bind the entities to the view.
On Mar 24, 11:07 am, Oskar Berggren oskar.bergg...@gmail.com wrote:
You speak of updating and saving an entity. Why is
Yes, indeed.
On Mar 24, 11:27 am, Oskar Berggren oskar.bergg...@gmail.com wrote:
More specifically, to you mean rebind after the entity have been updated?
/Oskar
2010/3/24 Sal salbass...@hotmail.com:
It needs to select to bind the entities to the view.
On Mar 24, 11:07 am, Oskar
Depending on how abstracted your NH usage is, you could call flush after the
update, which would force the update to fire. Then the selects afterward
would return the right results.
- Robert
On Wed, Mar 24, 2010 at 10:29 AM, Sal salbass...@hotmail.com wrote:
Yes, indeed.
On Mar 24, 11:27 am,
Depending on how abstracted your NH usage is, you could call flush after the
update, which would force the update to fire. Then the selects afterward
would return the right results.
Yeah, I did that in one other place. I guess it just seems
weird...meaning, I must be doing something wrong with
For many-to-one you can use batch-size in the referenced class declaration.
Diego
On Wed, Mar 24, 2010 at 00:14, jprid64 john.pride...@gmail.com wrote:
Thanks, Diego. That worked (sort of) for my one-to-many. I set the
batch size to 20 (knowing I would only get 15 based on the data I'm
Your current FlushMode affects this behavior as well. FlushMode.Auto
should cause the update to happen before the select.
/Oskar
2010/3/24 Robert Rudduck rob...@rpowered.net:
Depending on how abstracted your NH usage is, you could call flush after the
update, which would force the update to
I am using the default (assuming Fluent Nhibernate uses the same
default) which I believe is Commit? I wonder what other side effects
would appear if I changed that to Auto.
On Mar 24, 11:34 am, Oskar Berggren oskar.bergg...@gmail.com wrote:
Your current FlushMode affects this behavior as well.
Thanks for the update :), i personally don't trust anything with
foo(fu) or bar in it :P
--
You received this message because you are subscribed to the Google Groups
nhusers group.
To post to this group, send email to nhus...@googlegroups.com.
To unsubscribe from this group, send email to
That mapping does not mean nothing well only mean that *you* will set
the ID.
2010/3/24 jprid64 john.pride...@gmail.com
Fabio,
I'll get working on ConfORM and let you know.
The xml snippet is supposed be telling NH that the ID for the primary
table is also the key into the Plan object
Thanks, that worked. I thought I had read that .NET assumes a class is
serializable by default, so I didn't think I needed the attribute.
--
You received this message because you are subscribed to the Google Groups
nhusers group.
To post to this group, send email to nhus...@googlegroups.com.
Well, my unit of work is around the request and the commit happens at
the end, so I can't really commit it early. Someone on our team did
set the flushmode to commit - we weren't using auto. I am tempted to
set it to auto, but I have also ready many negative things about using
auto, performance
Sorry -- I wasn't clear. I removed the param tag with no side
effects. Yes, I am responsible for the ID with generator
class=assigned.
On Mar 24, 11:02 am, Fabio Maulo fabioma...@gmail.com wrote:
That mapping does not mean nothing well only mean that *you* will set
the ID.
2010/3/24
it shows an 88k file and gives me errors. Does anyone know someplace
other than sourceforge I can get it?
tqii
--
You received this message because you are subscribed to the Google Groups
nhusers group.
To post to this group, send email to nhus...@googlegroups.com.
To unsubscribe from this
Are you using windows zip utility?
Tuna Toksöz
Eternal sunshine of the open source mind.
http://devlicio.us/blogs/tuna_toksoz
http://tunatoksoz.com
http://twitter.com/tehlike
On Wed, Mar 24, 2010 at 6:40 PM, tqwhite tqwh...@gmail.com wrote:
it shows an 88k file and gives me errors. Does
If you can get access to the ISession, or some wrapper, you could call
Flush() on it to force the flush before select.
/Oskar
2010/3/24 Sal salbass...@hotmail.com:
Well, my unit of work is around the request and the commit happens at
the end, so I can't really commit it early. Someone on our
If you can get access to the ISession, or some wrapper, you could call
Flush() on it to force the flush before select.
Yep, I will do that. Is this considered bad practice?
On Mar 24, 12:42 pm, Oskar Berggren oskar.bergg...@gmail.com wrote:
If you can get access to the ISession, or some
It is somewhat not great practice, as the view manager must now be fully
NHibernate aware, rather than having NHibernate abstracted.
John Davidson
On Wed, Mar 24, 2010 at 12:54 PM, Sal salbass...@hotmail.com wrote:
If you can get access to the ISession, or some wrapper, you could call
Instead try ConfORM you can have a look to my post and learn how ConfORM map
a one-to-one...
welll... there you can see even how you can map a one-to-one in XML ;)
http://fabiomaulo.blogspot.com/2010/03/conform-mapping-one-to-one.html
2010/3/24 jprid64 john.pride...@gmail.com
Sorry -- I wasn't
It is somewhat not great practice, as the view manager must now be fully
NHibernate aware, rather than having NHibernate abstracted.
Well, it's actually happening in a service class. So, this solution
does not leak NH.
On Mar 24, 1:01 pm, John Davidson jwdavid...@gmail.com wrote:
It is
I'm using whatever Explorer chooses. I assume it's the Windows
utility.
It's only downloading about 88k. When I try to download it on a
Macintosh, it gives me a download error. This happened to me before
when I tried to download NH itself. That error just went away. (Nice
of time, too, I'd have
Well if you can call Flush() in the service component, why couldn't you call
Commit(), which would then do a Flush()?
John Davidson
--
You received this message because you are subscribed to the Google Groups
nhusers group.
To post to this group, send email to nhus...@googlegroups.com.
To
I have only two questions.
Why a query to DB, in the middle of a UoW, should include changes suspended
in the UoW ?
The UoW's definition is here
http://martinfowler.com/eaaCatalog/unitOfWork.html
Which is your interpretation of UoW pattern ?
2010/3/24 Sal salbass...@hotmail.com
It is
Well if you can call Flush() in the service component, why couldn't you call
Commit(), which would then do a Flush()?
Yeah, I guess I could. This a bit of a departure from our current
approach, so I need to think it through. My current unit of work
implementation is handled at the http request
Well if you can call Flush() in the service component, why couldn't you call
Commit(), which would then do a Flush()?
Yeah, I guess I could. This a bit of a departure from our current
approach, so I need to think it through. My current unit of work
implementation is handled at the http request
Well if you can call Flush() in the service component, why couldn't you call
Commit(), which would then do a Flush()?
Yeah, I guess I could. This a bit of a departure from our current
approach, so I need to think it through. My current unit of work
implementation is handled at the http request
Well if you can call Flush() in the service component, why couldn't you call
Commit(), which would then do a Flush()?
Yeah, I guess I could. This a bit of a departure from our current
approach, so I need to think it through. My current unit of work
implementation is handled at the http request
As Fabio suggests, a UoW is not equivalent to an http request. A UofW should
be a complete action within a database context. Normally a query, or very
infrequently a series of queries, may make a unit of work. You should call
Commit at that point.
Then when you start your update you begin a new
I have only two questions.
Why a query to DB, in the middle of a UoW, should include changes suspended
in the UoW ?
The UoW's definition is herehttp://martinfowler.com/eaaCatalog/unitOfWork.html
Which is your interpretation of UoW pattern ?
Fabio,
That is a great point. To be honest, I
Another detail, in case there are any Masters watching, is that the
current version, 2.1.2, is only 88.1kB. There previous version, 1.0.0,
is 2.3MB. Nothing slims down that much.
tqii
On Mar 24, 12:33 pm, tqwhite tqwh...@gmail.com wrote:
I'm using whatever Explorer chooses. I assume it's the
As Fabio suggests, a UoW is not equivalent to an http request. A UofW should
be a complete action within a database context. Normally a query, or very
infrequently a series of queries, may make a unit of work. You should call
Commit at that point.
So, are you saying that I should view this as
I had issues with the download as well. The download I did was just the
2.1.2-GA-Bin zip file from sourceforge. The file size for me was 87KB, however
using windows compression tool to unzip caused an error saying it could not
open the folder. I actually had to download winrar and open it
Hey,
I'm trying to get my site to run on a webhost with medium trust, I
followed the blog post at (http://nhforge.org/wikis/howtonh/pre-
generate-lazy-loading-proxies.aspx) to pre-generate the proxies. The
proxy generates ok but when I run it it throws the following
exception.
Method
Yes, just displaying data requires 1 transaction. to update date there is a
minimum of 2 transactions. If you create data and want the user to see it
then there is also 2 transactions minimum.
there is only one ISession object per http request
John Davidson
On Wed, Mar 24, 2010 at 2:17 PM, Sal
I am spiking an example of CQRS as described by Udi. I am using rhino
service bus to process commands and then fire events in the domain.
I'm using a 'To Do List' for the spike
public void Consume(NewQuickTask message)
{
var task=
session.GetUser(message.UserId).AddTask(message.Title);
If you are spiking a greenfield app with CQRS using nhibernate or any
orm will lead to a complete and utter failure.
On Wed, Mar 24, 2010 at 8:48 PM, Jason Meckley jasonmeck...@gmail.com wrote:
I am spiking an example of CQRS as described by Udi. I am using rhino
service bus to process commands
Winrar did the job. Thanks Michael!
Does anyone know why?
tqii
On Mar 24, 1:18 pm, Michael A. Bell mb...@interactivesoftworks.com
wrote:
I had issues with the download as well. The download I did was just the
2.1.2-GA-Bin zip file from sourceforge. The file size for me was 87KB,
however
Greg, thank you the *concise* response :) care to elaborate? I can see
how my design may fail, but how can a tool fail?
If I choose Guid as my id, my approach might work. If I choose
Identity (or some other db POID) this will definitely work. I'm hoping
there's is 3rd option which I'm not aware
How do you prove that the events that you publish to the reporting
model (and every other model) from your domain match what nhibernate
(or any other ORM) believes your changes to be? It is a difficult
problem and the models,your database model and your eventing model,
will diverge over time
I created an entity via the default constructor and added it my unit
of work immediately via Session.Save. I then made subsequent changes
to this entity prior to commit. I would have expected a single insert
with the changes aggregated. However, what I see is an insert with the
default profile
No idea, but I have seen this on several zip files in the past that have
nothing to do with NHibernate or eachother.
-Original Message-
From: nhusers@googlegroups.com [mailto:nhus...@googlegroups.com] On Behalf Of
tqwhite
Sent: Wednesday, March 24, 2010 2:58 PM
To: nhusers
Subject:
Your expectation is not correct. If you Save a transient instance and then
modify it, it will be saved with the original values first and then updated.
You'll need to make sure you don't call save on an invalid object.
Diego
On Wed, Mar 24, 2010 at 15:29, ragamuf raga...@gmail.com wrote:
I
the possibility does exist for the pvm and domain to be out of sync. I
understood this as eventual consistency. Are you talking about
permanently loosing consistency. Like entire handlers not succeeding?
Lets assume that events will succeed. in my current architecture the
domain and pvm are
This has nothing to do with eventual consistency...
This has to do with the fact that you have 2 models and 2 ways of
detecting changes. How do you insure they find the same changes?
Greg
On Wed, Mar 24, 2010 at 10:09 PM, Jason Meckley jasonmeck...@gmail.com wrote:
the possibility does exist
yep. I came to the same conclusion painfully.
On Mar 24, 4:04 pm, Diego Mijelshon di...@mijelshon.com.ar wrote:
Your expectation is not correct. If you Save a transient instance and then
modify it, it will be saved with the original values first and then updated.
You'll need to make sure you
Use session.Persist instead session.Save but take care with session.Persist
if you aren't using XML to map NH well I mean if you are not using
neither XML nor ConfORM.
2010/3/24 Jason Meckley jasonmeck...@gmail.com
I am spiking an example of CQRS as described by Udi. I am using rhino
Thanks for all the info, but I am getting following exception:
NHibernate.HibernateException was unHandeled
Message=Could not create the driver from
NHibernate.Driver.OracleLiteDataClientDriver.
I have downloaded the NHibernate-2.1.2.GA-bin.zip , I hope this has Oracle
Lite support as it is a
I have one way commands coming from the client. the commands are
consumed. consuming includes 1. Nh getting entity(s) and updating 2.
sending events in the domain stating this happened. it seems natural
to me, what am I missing?
On Mar 24, 4:10 pm, Greg Young gregoryyou...@gmail.com wrote:
This
what does nhibernate save to the database?
what were your events to to the reporting side?
are they the same? can you prove it?
what if they aren't?
On Wed, Mar 24, 2010 at 11:35 PM, Jason Meckley jasonmeck...@gmail.com wrote:
I have one way commands coming from the client. the commands are
Fabio, are you saying session.persist for saving the task (HiLo POID)
or the pvm?
On Mar 24, 5:29 pm, Fabio Maulo fabioma...@gmail.com wrote:
Use session.Persist instead session.Save but take care with session.Persist
if you aren't using XML to map NH well I mean if you are not using
/// summary
/// Make a transient instance persistent. This operation cascades to
associated
/// instances if the association is mapped with
ttcascade=persist/tt.br/
/// The semantics of this method are defined by JSR-220.
/// /summary
/// param name=obja transient instance to be made persistent
Wondering if this is officially a bug and will it get fixed?
We have no listener on Flush events, only on the PostInsert/Update/
Delete and PostCollectionRecreate/Update/Remove.
We have implemented a Unit of Work pattern, in which we set FlushMode
to Never and explicitly call Flush();
Our
Probably is Not a bug.
Are you loading a collection inside your listeners ?
2010/3/24 Michael diSibio michael.disi...@gmail.com
Wondering if this is officially a bug and will it get fixed?
We have no listener on Flush events, only on the PostInsert/Update/
Delete and
I'd just like to bump this. I posted originally on a late Friday
afternoon, so I guess everyone went home.
On Mar 12, 5:02 pm, Michael diSibio michael.disi...@gmail.com wrote:
What we observed:
- An explicit timeout set by 'Connect Timeout' clause in a connection
string in the configuration
69 matches
Mail list logo
