query corruption for complexer queries on -current?

Hi,
using the structure from a.sql, populating with some entries, using the query in
sel_reise.sh (with hitting some tuples) I get only a subset (1 entry!) of the
matching tuples. Delting all lines in sel_reise.sh where 'e.*' is used
(3.lines), I get all tuples. The same script works without errors on pgsql 6.3
using the same data.

Some stats with my original data:

abr=> select count(*) from reise where begt > 'Feb 01 00:00:00 1999';
count
-----
13
(1 row)

abr=> select distinct mid from reise where begt > 'Feb 01 00:00:00 1999';
mid
---
605
(1 row)

abr=> select distinct mid from emp;
mid
---
0
605
(2 rows)

Any clues?

Bye!
----
Michael Reifenberger
Plaut Software GmbH, R/3 Basis