nhibernate-issues

[nhibernate-issues] [JIRA] Created: (NH-853) Inconsistent need of the Flush method invokation

[NHibernate J. <mik...@us...>]

Inconsistent need of the Flush method invokation
------------------------------------------------

         Key: NH-853
         URL: http://jira.nhibernate.org/browse/NH-853
     Project: NHibernate
        Type: Bug
  Components: Core =20
    Versions: 1.2.0.Beta2   =20
    Reporter: Andr=C3=A9s G. Aragoneses
    Priority: Trivial
 Attachments: NHibBugAboutFlush.zip

I have managed to create a simple NHibernate testcase which creates some ob=
jects and all the operations are done successfully without the FLUSH method=
 invokation *except* the many-to-many relation creations.

I am attaching the testcase as a VS2005 solution. Note the following line i=
n Program.cs:

            //if this line is commented, the table TERMINALS_TERMINALGROUPS=
 is not persisted! but the rest yes
            oSession.Flush();

P.S.: Be sure to add the NHibernate DLL's to the 'Lib' directory (and to th=
e bin/Debug ones) when uncompressing.

P.S.2: Don't pay attention to the name of the solution. It's for another bu=
g I am going to open.

--=20
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
   http://jira.nhibernate.org/secure/Administrators.jspa
-
For more information on JIRA, see:
   http://www.atlassian.com/software/jira

[nhibernate-issues] [JIRA] Closed: (NH-853) Inconsistent need of the Flush method invokation

[NHibernate J. <mik...@us...>]

     [ http://jira.nhibernate.org/browse/NH-853?page=3Dall ]
    =20
Sergey Koshcheyev closed NH-853:
--------------------------------

    Resolution: Not an Issue

This is not an issue with NHibernate, NH has to perform identity inserts ri=
ght away to be able to return an identifier to you. Even in that case, you =
are advised to call Flush explicitly.

> Inconsistent need of the Flush method invokation
> ------------------------------------------------
>
>          Key: NH-853
>          URL: http://jira.nhibernate.org/browse/NH-853
>      Project: NHibernate
>         Type: Bug
>   Components: Core
>     Versions: 1.2.0.Beta2
>     Reporter: Andr=C3=A9s G. Aragoneses
>     Priority: Trivial
>  Attachments: NHibBugAboutFlush.zip
>
> I have managed to create a simple NHibernate testcase which creates some =
objects and all the operations are done successfully without the FLUSH meth=
od invokation *except* the many-to-many relation creations.
> I am attaching the testcase as a VS2005 solution. Note the following line=
 in Program.cs:
>             //if this line is commented, the table TERMINALS_TERMINALGROU=
PS is not persisted! but the rest yes
>             oSession.Flush();
> P.S.: Be sure to add the NHibernate DLL's to the 'Lib' directory (and to =
the bin/Debug ones) when uncompressing.
> P.S.2: Don't pay attention to the name of the solution. It's for another =
bug I am going to open.

--=20
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
   http://jira.nhibernate.org/secure/Administrators.jspa
-
For more information on JIRA, see:
   http://www.atlassian.com/software/jira

[nhibernate-issues] [JIRA] Commented: (NH-853) Inconsistent need of the Flush method invokation

[NHibernate J. <mik...@us...>]

     [ http://jira.nhibernate.org/browse/NH-853?page=3Dcomments#action_1473=
7 ]
    =20
Ayende Rahien commented on NH-853:
----------------------------------

Hehe,
I had the _exact_ same response as you did about a year ago, for the same r=
eason.
To expand a bit on Sergey's comment, the issue is that NH has to go to the =
DB as soon as you Save() an object that has Identity PK.
This may lead into situations where stuff seems to work, even without a flu=
sh. Consider such things as bugs in your code.
In the future, NHibernate is going to move to a mode where it will only do =
Identity Inserts when in a transaction, so this is something to consider.

> Inconsistent need of the Flush method invokation
> ------------------------------------------------
>
>          Key: NH-853
>          URL: http://jira.nhibernate.org/browse/NH-853
>      Project: NHibernate
>         Type: Bug
>   Components: Core
>     Versions: 1.2.0.Beta2
>     Reporter: Andr=C3=A9s G. Aragoneses
>     Priority: Trivial
>  Attachments: NHibBugAboutFlush.zip
>
> I have managed to create a simple NHibernate testcase which creates some =
objects and all the operations are done successfully without the FLUSH meth=
od invokation *except* the many-to-many relation creations.
> I am attaching the testcase as a VS2005 solution. Note the following line=
 in Program.cs:
>             //if this line is commented, the table TERMINALS_TERMINALGROU=
PS is not persisted! but the rest yes
>             oSession.Flush();
> P.S.: Be sure to add the NHibernate DLL's to the 'Lib' directory (and to =
the bin/Debug ones) when uncompressing.
> P.S.2: Don't pay attention to the name of the solution. It's for another =
bug I am going to open.

--=20
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
   http://jira.nhibernate.org/secure/Administrators.jspa
-
For more information on JIRA, see:
   http://www.atlassian.com/software/jira

[nhibernate-issues] [JIRA] Commented: (NH-853) Inconsistent need of the Flush method invokation

[NHibernate J. <mik...@us...>]

     [ http://jira.nhibernate.org/browse/NH-853?page=3Dcomments#action_1475=
8 ]
    =20
Andr=C3=A9s G. Aragoneses commented on NH-853:
-----------------------------------------

Thanks for the comments.

BTW, regarding the Ayende's one: "In the future, NHibernate is going to mov=
e to a mode where it will only do Identity Inserts when in a transaction", =
why don't we close then this issue in that future? I think this behaviour m=
ay be confusing for newcomers so, at least, we could reference them to this=
 "UNRESOLVED" feature.

> Inconsistent need of the Flush method invokation
> ------------------------------------------------
>
>          Key: NH-853
>          URL: http://jira.nhibernate.org/browse/NH-853
>      Project: NHibernate
>         Type: Bug
>   Components: Core
>     Versions: 1.2.0.Beta2
>     Reporter: Andr=C3=A9s G. Aragoneses
>     Priority: Trivial
>  Attachments: NHibBugAboutFlush.zip
>
> I have managed to create a simple NHibernate testcase which creates some =
objects and all the operations are done successfully without the FLUSH meth=
od invokation *except* the many-to-many relation creations.
> I am attaching the testcase as a VS2005 solution. Note the following line=
 in Program.cs:
>             //if this line is commented, the table TERMINALS_TERMINALGROU=
PS is not persisted! but the rest yes
>             oSession.Flush();
> P.S.: Be sure to add the NHibernate DLL's to the 'Lib' directory (and to =
the bin/Debug ones) when uncompressing.
> P.S.2: Don't pay attention to the name of the solution. It's for another =
bug I am going to open.

--=20
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
   http://jira.nhibernate.org/secure/Administrators.jspa
-
For more information on JIRA, see:
   http://www.atlassian.com/software/jira