FreeTDS Development Group

["Schörk, Andreas" <andreas.schoerk AT infor.de>]

You are right, it is the part which is saving.
this part subtracts the hour instead of adding it.
The retrieving part also subtracts because it thinks the 
sender did an adding. 
Therefore the 2 hour difference.
When I correct the binding part by letting it add how suggested,
the differences disappear.
What is funny is, that java does not impose a 
day light saving in 1998. This I do not understand, 
but I suspect this is another problem.
I see the variables are used twice. So I should have corrected
the constants in the statements as well. 
Then the lines 2 and 4 would show the same results as 1 and 3.

Andreas

-----Ursprüngliche Nachricht-----
Von: Stefan Bodewig [mailto:bodewig AT bost.de]
Gesendet: Donnerstag, 9. November 2000 18:15
An: TDS Development Group
Betreff: [freetds] Re: JDBC: Problem with DSTOffset in binding
date-values


Andreas Schörk <andreas.schoerk AT infor.de> wrote:

> To show the differences you have to be in an area where daylight
> saving is done during August 2007. For instance GMT + 1.

Thats mine!!! ;-)

OK, I've run your tests and got 

got 2007-09-09 13:35:06.4, expected 2007-09-09 15:35:06.4
got 1998-03-09 15:35:05.4, expected 2007-09-09 15:35:06.4
got 2007-09-09 13:35:00.0, expected 2007-09-09 15:35:00.0
got 1998-03-09 15:35:00.0, expected 2007-09-09 15:35:00.0

but they already hold these values in the DB, i.e. when I query them
using QueryAnalyzer. So it cannot be the part that is retrieving the
data, or am I missing something?

Stefan

---
You are currently subscribed to freetds as: [andreas.schoerk AT infor.de]
To unsubscribe, forward this message to
$subst('Email.Unsub')

Attachment: t0056.java
Description: Binary data