As I dig into this issue, it has become apparent that the BOOLEAN
datatype was removed so that Derby would be compatible with DB2. The
regression test lang/db2Compatibility.sql monitors this behavior.
The IBM folks clearly invested a fair amount of effort in building a
DB2-compatible Derby. I don't want to simply undo that work. Would it be
reasonable to introduce a startup property which causes Derby to operate
in a DB2-compatible mode? The default for this property would be false,
but it might be useful for developers who want to use Derby as a baby DB2.
-Rick
Rick Hillegas wrote:
I have assigned this issue (bug 499) to myself. I plan to do the
following:
1) Re-enable the BOOLEAN datatype by removing the parser short-circuit.
2) Re-enable the TRUE and FALSE literals.
3) Add appropriate unit tests.
Cheers,
-Rick
Jeffrey Lichtman wrote:
However, before someone undisables the lines identified by Jeff, I
will mention that more work than meets the eye went into it - at
least into adjusting the tests. . .
Of course any new feature should have tests written for it. I didn't
mean that someone should hack out a line of code to enable a feature
without writing tests for it.
Can't someone at IBM resurrect the tests for the boolean type? I
would expect them to be accessible in whatever source code control
system IBM uses. Or perhaps IBM considers the reinstatement of the
disabled features to be against their interests.
- Jeff Lichtman
[EMAIL PROTECTED]
Check out Swazoo Koolak's Web Jukebox at
http://swazoo.com/
