The PostgreSQL db.props file has 3 different ways to use booleans: BIT = boolean BOOLEANCHAR = char BOOLEANINT = int2
Back in May I was having problems testing for true against columns defined with the BIT attribute. Criteria was comparing the column to the integer 1, which fails. Postgres expects the comparison to be against the word true. John McNally was kind enough to add a new database adapter method to allow Criteria to use the correct method for each database. The postgres adapter was configured to use true/false instead of 1/0. On Aug 31 Jon Scott Stevens patched the postgres adapter so that it again used 1/0 instead of true/false. This was to fix a scarab bug: http://scarab.tigris.org/issues/show_bug.cgi?id=368 I don't have access to that bug tracker, I tried to create an account but so far it has not sent me a password. So I don't know exactly what that bug says, but my guess is that they are using BOOLEANINT instead of BIT for their boolean fields. Anyway, the patch breaks my BIT-based schema. I'm not sure what strings BOOLEANCHAR uses, I remember reading somewhere that it was 't' and 'f'. In which case, neither of the above comparisons will work. This is a real mess! It seems to me the only way to solve the problem is to have each database only support one method of storing BIT/BOOLEAN values. Of course this will cause problems for people who already have a schema in production that does not use that particular method. Thoughts? -james -- [] James A. Hillyerd <[EMAIL PROTECTED]> - Java Software Engineer [] PGP 1024D/D31BC40D F87B 7906 C0DA 32E8 B8F6 DE23 FBF6 4712 D31B C40D -- To unsubscribe, e-mail: <mailto:turbine-torque-dev-unsubscribe@;jakarta.apache.org> For additional commands, e-mail: <mailto:turbine-torque-dev-help@;jakarta.apache.org>
