For what its worth... I tried adding this for y'all before I left, but
it actually does not at all affect the tests. Not sure what tests you
saw failing because the id sequence/table is not there, but I tried a
few tests that use sequences and they all fail trying to build the
MetaData object.
My guess is that most likely nothing here really changed. We had slews
of test failures that tbh hid lots of problems.
Nonetheless, returning null instead of empty string is the convention we
decided on in Austin. It just does not make sense to keep doing that
isEmpty() style checks higher an
Just a reminder that I am on vacation this next week.
If you want this in the meantime, you just need to alter the
org.hibernate.metamodel.spi.relational.Database#generateSchemaCreationScript
and
org.hibernate.metamodel.spi.relational.Database#generateDropSchemaScript
methods. There is a
or
Yep I have not hooked my new schema management tool stuff into metamodel
creation yet.
On Jul 29, 2012 1:24 PM, "Strong Liu" wrote:
> btw, some tests are failing due to the identifier generator required
> sequence are not created.
>
> org.hibernate.id.PersistentIdentifierGenerator#sqlCreateString
btw, some tests are failing due to the identifier generator required sequence
are not created.
org.hibernate.id.PersistentIdentifierGenerator#sqlCreateStrings is never get
called
(yes, i see there is a org.hibernate.id.SequenceGenerator#registerExportables)
and others are failing with stacktr
I was not aware of that, the root issue is
"org.hibernate.metamodel.internal.source.annotations.attribute.Column#name" has
a default empty string value.
On Jul 30, 2012, at 1:30 AM, Steve Ebersole wrote:
> Hadn't we all agreed though that the convention was for sources to return
> null and n
Hadn't we all agreed though that the convention was for sources to return
null and not empty strings?
On Jul 29, 2012 11:11 AM, "Strong Liu" wrote:
> https://hibernate.onjira.com/browse/HHH-7480
>
> don't know which change caused this, but it is fixed now
>
> On Jul 28, 2012, at 10:27 AM, Steve E
https://hibernate.onjira.com/browse/HHH-7480
don't know which change caused this, but it is fixed now
On Jul 28, 2012, at 10:27 AM, Steve Ebersole wrote:
> Not sure if the expectation at this point is that all tests on the
> metamodel branch pass when hibernate.test.new_metadata_mappings is se
