On Tue, Oct 13, 2015 at 1:50 AM, John Pruitt <jpruitt@doozer.com> wrote:
> Okay, I'll acknowledge that the issue has been discussed before, however I
> fail to see how the issue can be considered resolved. The function takes
> time zone aware arguments, ignores that input, and returns plainly incorrect
> answers - the very definition of a bug.
I also feel like that it is a bug. Here I attached a patch that
corrects the problem.
The patch just adds the difference in DST to the hours of difference
between the two
timestamp. The drawback in this approach is that, it produces the result as
24:00:00 instead of 1 day.
Any better approach to handle this problem?
Regards,
Hari Babu
Fujitsu Australia
