Re: BUG #4849: intermittent future timestamps

David Leppik  wrote: 
> Never mind.  Turns out the bug was in our own code (read:  me,  
> personally, being stupid) to convert a java.sql.Timestamp to  
> java.sql.Date.  Why it works at all in MySQL... I don't even want
> to know.
 
java.sql.Date or java.util.Date?  (You don't show your imports, so
it's impossible to tell from the code snippet.)  If it's
java.util.Date, I can't immediately see why your errors would be
greater than one second.  If it's java.sql.Date, you're probably in
territory where the behavior is undefined, but it's hard to see where
you would get the results you showed.  Perhaps there's an opportunity
for us to make the PostgreSQL JDBC driver behave more sanely in this
circumstance?
 
> Why is it we can spend weeks looking at a bug, and we can't find
> it until we decide to blame it on someone else?
 
It's probably a corollary to the tendency to see our own gaffs when
reading the post coming back from the list much more clearly than they
appeared before clicking "send".  :-/
 
-Kevin