[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: JULDAY 5.4 not same as 5.3?
"Don Woodraska" <firstname.lastname@example.org> wrote in message
> I've pinpointed the problem.
> The bug appears only with unsigned-long and unsigned-64-bit-long hour
> to JULDAY.
> The bug appears on lines 178-179 of JULDAY.PRO. Here it is:
Well done, Dan. To correct my earlier post, this DOES also occur on windows.
In the list of data values in your original post, i.e...
>>> UMM LONG = 2
>>> UDD LONG = 16
>>> UYEAR LONG = 2001
>>> UHR ULONG = 0
>>> UMIN ULONG = 0
>>> USEC ULONG = 0
...I missed the fact that the hour minute & second values are ULONG and not
LONG. (Though I did wonder why you put a "U" prefix on the variable names!)
I wonder how many other routines can be broken if you pass unsigned-integer
values to them. And how paranoid does a person writing a routine have to be?
Final comment: Isn't JULDAY's argument order ghastly? Month, day, year,
hour, minute, second? I mean, really!
National Institute for Water and Atmospheric Research