[Hibernate-devel] Probable bug

[Gavin_King/Cirrus%<CI...@ci...>]

I need some help here:

Daniel Wunsch reports that the following query:

  session.find(
    "FROM person IN CLASS com.consol.loan.bo.Person WHERE person.login
= ?",
    login,
    Hibernate.STRING
  );

Generated the following SQL in 1.0.1:

SELECT DISTINCT person.id as id0, person.id as id1, person.password as
password0, person.phone as phone0, person.login as login0, person.lastName
as lastName0, person.email as email0, person.firstName as firstName0,
person.phoneExt as phoneExt0, person.password as password1, person.phone as
phone1, person.firstName as firstName1, person.phoneExt as phoneExt1 FROM
Person person WHERE 1=1 AND ( person.login=?  ) AND ( person.login=?  )

and then failed with an exception, whereas in 1.0 it generated

SELECT DISTINCT person.id, person.password, person.phone, person.login,
person.lastName, person.email, person.firstName, person.phoneExt FROM
Person person WHERE 1=1 AND ( person.login=?  )

(the second is correct, clearly)

His mapping looks like:

<class name="com.consol.loan.bo.Person" table="Person" select="distinct">
  <id name="id" type="long" column="id">
    <generator class="hilo.long" />
  </id>
  <property name="sunIdent" column="sunIdent" type="string" />
  <property name="password" column="password" type="string" />
  <property name="phone" column="phone" type="string" />
  <property name="login" column="login" type="string" />
  <property name="lastName" column="lastName" type="string" />
  <property name="email" column="email" type="string" />
  <property name="firstName" column="firstName" type="string" />
  <property name="phoneExt" column="phoneExt" type="string" />
  <list role="roles">
    <key column="person" />
    <index column="i" />
    <many-to-many class="com.consol.loan.bo.Role" />
  </list>
</class>

As a result I pulled 1.0.1 from the downloads page. Unfortunately I have
not been able to reproduce this bug and his code is sufficiently close
to so many things in the test suite that I should be seeing test failures
all over the place but, actually, I'm not seeing any.

On the other hand, this is the exact area of the code that changed between
1.0 and 1.0.1 so his bug report is utterly credible.

If people would take a CVS snapshot (a current snapshot or the version
tagged v1001) and test their systems against that, I would *very* much
appreciate it. I'm completely stumped.

Gavin