Re: Funny timezone shift causes failure in test suite

[about a failure with a historical date in the JDBC driver's
test suite, when ran in the CET timezone]

On Tue, 13 Nov 2001 16:28:57 -0500, you wrote:
>At least) Sun's JRE seems to be braindead when it comes to
>timezones.. I made a small test program that runs through a sequence of
>years and for each year, it takes a date in the winter and a date in the
>summer and checks whether or not the calendar is in daylight savings
>time. From 1945 to 1977, it reports that in the summer, daylight savings
>time is observed, which it shouldn't be.
>
>After poking around some, I think that Java isn't concerned with
>historical behaviour (ICU [Internation Components for Unicode] isn't
>AFAICS, and the ICU Java classes developped by Taligent were integrated
>into Sun's JDK 1.1...)

Yes, you're right.

This bug description confirms that the JVM up to 1.3 does not
have a historically correct timezone implementation:
http://developer.java.sun.com/developer/bugParade/bugs/4257314.html
(registration required), and it says this is fixed in 1.4 aka
'merlin'.

I tested it with Sun's JDK 1.4 beta 3 and the Sun's J2EE 1.3
reference implementation on Red Hat Linux 7.1 and indeed the
test suite ran fine with 0 failures (I had to tweak some
unimplemented methods to be able to run the driver with this
JVM/J2EE).

Case closed. We don't need to do anything in the driver. Liam,
thanks a lot for your help!

Regards,
René Pijlman <rene(at)lab(dot)applinet(dot)nl>