Thanks for that Patrick.
Yes I altered the code to remove the Hungarian notation from the index
names, which didn't call correctName().
I'll have ago at getting a debug session running tonight and see if I
cant track down this issue, before I post the patch.
Regards
Ben
On 11/1/07, Patrick Linskey <[EMAIL PROTECTED]> wrote:
> Hi,
>
> (Moving to dev list since this is becoming more dev-related than
> user-related.)
>
> I'm making some guesses about what your code is doing, but my
> suspicion is that you've done what I suggested and overridden
> correctName(). The ordering of the words in the SQL version of that
> name looks a little suspect; I'd guess that the code is going through
> a different pathway for constructing names for foreign keys in
> relations.
>
> I'd guess that somewhere, the code is not calling correctName() in one
> of those methods. Alternately, maybe the problem is is the
> getTableName() method, which seems to grab the field mapping's name
> and use it in the computation of table names for join tables without
> using correctName() first. Maybe this name then gets incorporated into
> the foreign key col name.
>
> The thing is, the field mapping's name is used directly in a lot of
> places (via FieldMapping.getName()), and that needs to remain as the
> name of the field itself. This is a bit unfortunate, since it makes it
> harder to have just one place to put the code that would do the
> translation for all fields. I believe, however, that the
> MappingDefaults stuff is appropriately designed to make it possible to
> do this, even if we need to catch a few more cases in
> MappingDefaultsImpl.
>
> If you can post a patch somewhere, I can help take a look at some of
> the cases that are failing to be caught.
>
> -Patrick
>
> On Oct 31, 2007 3:50 PM, ben short <[EMAIL PROTECTED]> wrote:
> > Hi Patrick,
> >
> > Im struggling to find out why field m_clzFooBar6 of the
> > HungarianNotationFields class does not get it name changed. The
> > classes and the sql executed follows.
> >
> > @Entity
> > public class HungarianNotationFields
> > {
> > private String mFooBar1;
> >
> > private String strFooBar2;
> >
> > private Integer intFooBar3;
> >
> > private Long lgFooBar4;
> >
> > private int m_intFooBar5;
> >
> > @ManyToOne(targetEntity = OtherClass.class)
> > private OtherClass m_clzFooBar6;
> > }
> >
> > @Entity
> > public class OtherClass
> > {
> > @Id
> > private Integer m_intBarFoo1;
> > }
> >
> > CREATE TABLE HungarianNotationFields (ID BIGINT NOT NULL, FOOBAR3
> > INTEGER, FOOBAR4 BIGINT, FOOBAR1 VARCHAR(255), FOOBAR5 INTEGER,
> > FOOBAR2 VARCHAR(255), m_clzFooBar6_BARFOO1 INTEGER, PRIMARY KEY (ID));
> >
> > CREATE TABLE OtherClass (BARFOO1 INTEGER NOT NULL, PRIMARY KEY (BARFOO1));
> >
> > Any Idea why this might be?
> >
> > Regards
> >
> > Ben
> >
> >
> > On 10/31/07, Patrick Linskey <[EMAIL PROTECTED]> wrote:
> > > Backing state for MappingDefaults.defaultMissing():
> > >
> > > Whether to fill in missing mapping information at runtime with the
> > > default values supplied by this plugin. A value of false means that
> > > all mapping information must be present at runtime.
> > >
> > > -Patrick
> > >
> > > On Oct 30, 2007 4:27 PM, ben short <[EMAIL PROTECTED]> wrote:
> > > > In the MappingDefaultsImpl class there is a property _defMissing, any
> > > > idea what thats for?
> > > >
> > > >
> > > > On 10/30/07, Patrick Linskey <[EMAIL PROTECTED]> wrote:
> > > > > Cool. FYI, we end up putting most of our test cases in
> > > > > openjpa-persistence-jdbc, for bootstrapping reasons.
> > > > >
> > > > > -Patrick
> > > > >
> > > > > On Oct 30, 2007 3:43 PM, ben short <[EMAIL PROTECTED]> wrote:
> > > > > > Hi Patrick,
> > > > > >
> > > > > > I managed to get it going... I was trying to get idea to create the
> > > > > > project from the pom. As you say using mvn idea:idea works much
> > > > > > better.
> > > > > >
> > > > > > Thanks
> > > > > >
> > > > > > Ben
> > > > > >
> > > > > >
> > > > > > On 10/30/07, Patrick Linskey <[EMAIL PROTECTED]> wrote:
> > > > > > > I use IDEA mostly.
> > > > > > >
> > > > > > > I set things up initially with mvn at the command line, and then
> > > > > > > ran
> > > > > > > mvn idea:idea to build project files. But the craziness around
> > > > > > > the jjt
> > > > > > > and javacc mvn plugins meant that I needed to manually modify
> > > > > > > something to get the project files just so.
> > > > > > >
> > > > > > > I do compilations from IDEA directly, not via the mvn
> > > > > > > integration. And
> > > > > > > I use TeamCity to fire off remote tests and checkins.
> > > > > > >
> > > > > > > -Patrick
> > > > > > >
> > > > > > > On Oct 30, 2007 2:54 PM, ben short <[EMAIL PROTECTED]> wrote:
> > > > > > > > What IDE do you guys use, as im having a hell of a time getting
> > > > > > > > it to
> > > > > > > > work with idea 7
> > > > > > > >
> > > > > > > >
> > > > > > > > On 10/30/07, Patrick Linskey <[EMAIL PROTECTED]> wrote:
> > > > > > > > > Best thing to do is create a JIRA and attach the diff (svn
> > > > > > > > > diff >
> > > > > > > > > openjpa-<jiranumber>.patch). If you don't have a JIRA
> > > > > > > > > account, let me
> > > > > > > > > know; I'll create an issue for you.
> > > > > > > > >
> > > > > > > > > -Patrick
> > > > > > > > >
> > > > > > > > > On Oct 30, 2007 2:07 AM, ben short <[EMAIL PROTECTED]> wrote:
> > > > > > > > > > OK I'll give it ago. I assume that you'll need a patch of
> > > > > > > > > > the changes
> > > > > > > > > > I make. Where is the best place to send it once I'm done?
> > > > > > > > > >
> > > > > > > > > >
> > > > > > > > > > On 10/29/07, Patrick Linskey <[EMAIL PROTECTED]> wrote:
> > > > > > > > > > > Hi,
> > > > > > > > > > >
> > > > > > > > > > > You can do this by providing your own MappingDefaults
> > > > > > > > > > > implementation
> > > > > > > > > > > (via the openjpa.MappingDefaults property). You'll
> > > > > > > > > > > probably want to
> > > > > > > > > > > extend PersistenceMappingDefaults and override
> > > > > > > > > > > correctName(), to stay
> > > > > > > > > > > in line with the rest of the JPA defaulting rules and to
> > > > > > > > > > > simplify the
> > > > > > > > > > > implementation.
> > > > > > > > > > >
> > > > > > > > > > > Alternately, you could improve MappingDefaultsImpl and
> > > > > > > > > > > contribute the
> > > > > > > > > > > change to OpenJPA. Such an improvement might look like so:
> > > > > > > > > > >
> > > > > > > > > > > 1. add a new setter/getter pair called
> > > > > > > > > > > 'removeHungarianNotation' (a
> > > > > > > > > > > boolean) or 'removeLeadingPrefixes' (a list of strings).
> > > > > > > > > > >
> > > > > > > > > > > 2. implement logic in MappingDefaultsImpl.correctName()
> > > > > > > > > > > to either
> > > > > > > > > > > guess about hungarian-notation trimming or to remove the
> > > > > > > > > > > prefixes
> > > > > > > > > > > identified in 'removeLeadingPrefixes'.
> > > > > > > > > > >
> > > > > > > > > > > 3. write a test case that that extends SingleEMFTestCase
> > > > > > > > > > > and looks
> > > > > > > > > > > something like so:
> > > > > > > > > > >
> > > > > > > > > > > public TestColumnNameTruncation extends SingleEMFTestCase
> > > > > > > > > > > {
> > > > > > > > > > > public void setUp() {
> > > > > > > > > > > setUp(HungarianNotationFields.class, CLEAR_TABLES,
> > > > > > > > > > > "openjpa.MappingDefaults",
> > > > > > > > > > > "RemoveLeadingPrefixes=\"m,_\"");
> > > > > > > > > > > }
> > > > > > > > > > >
> > > > > > > > > > > public void testColumnNameTruncation() {
> > > > > > > > > > > ClassMapping cm = (ClassMapping)
> > > > > > > > > > > JPAFacadeHelper.getMetaData(emf,
> > > > > > > > > > > HungarianNotationFields.class);
> > > > > > > > > > > // check that the associated FieldMappings in cm
> > > > > > > > > > > have the
> > > > > > > > > > > right column names
> > > > > > > > > > > }
> > > > > > > > > > > }
> > > > > > > > > > >
> > > > > > > > > > > -Patrick
> > > > > > > > > > >
> > > > > > > > > > > On 10/25/07, ben short <[EMAIL PROTECTED]> wrote:
> > > > > > > > > > > > Hi,
> > > > > > > > > > > >
> > > > > > > > > > > > I name my class member variables (fields) with the
> > > > > > > > > > > > prefix m, such as
> > > > > > > > > > > > mId. Is it possible to tell OpenJPA to create the table
> > > > > > > > > > > > columns
> > > > > > > > > > > > without the m prefix?
> > > > > > > > > > > >
> > > > > > > > > > > > I know this can be done using the @Column annotation,
> > > > > > > > > > > > but it would be
> > > > > > > > > > > > nice to be able to set a global flag, like I think you
> > > > > > > > > > > > can with
> > > > > > > > > > > > hibernate.
> > > > > > > > > > > >
> > > > > > > > > > > > Regards
> > > > > > > > > > > >
> > > > > > > > > > > > Ben
> > > > > > > > > > > >
> > > > > > > > > > >
> > > > > > > > > > >
> > > > > > > > > > > --
> > > > > > > > > > > Patrick Linskey
> > > > > > > > > > > 202 669 5907
> > > > > > > > > > >
> > > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > > --
> > > > > > > > > Patrick Linskey
> > > > > > > > > 202 669 5907
> > > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > > >
> > > > > > >
> > > > > > > --
> > > > > > > Patrick Linskey
> > > > > > > 202 669 5907
> > > > > > >
> > > > > >
> > > > >
> > > > >
> > > > >
> > > > > --
> > > > > Patrick Linskey
> > > > > 202 669 5907
> > > > >
> > > >
> > >
> > >
> > >
> > > --
> > > Patrick Linskey
> > > 202 669 5907
> > >
> >
>
>
>
> --
> Patrick Linskey
> 202 669 5907
>
