Error in datestr.m in Octave-3.2.0 MingW

[John W. Eaton]

On 24-Jun-2009, Michael Goffioul wrote:

| On Tue, Jun 23, 2009 at 11:14 PM, John W. Eaton<jwe at octave.org> wrote:
| > On 23-Jun-2009, Philip Nienhuis wrote:
| >
| > | Digging a bit further, and echoeing strftime's arguments, i.e. variable:
| > | - df
| > | and expression:
| > | - localtime(mktime(tm))
| > | in the erroneous datestr.m, I see:
| >
| > Then maybe the bug is in mktime or localtime?  What are the values of
| > tm and mktime (tm) here?  I expect
| 
| >  tm =
| >  {
| >    year =  69
| >    mon = 0
| >    mday =  1
| >    hour = 0
| >    min = 0
| >    sec = 0
| >    usec = 0
| >  }
| >
| > and
| >
| >  ans = -31518000
| >
| > If these are not the values returned, then what does
| >
| >  localtime (-31518000)
| >
| > return on your Windows system?
| 
| ans =
| {
|   usec = 0
|   sec = 0
|   min = 0
|   hour = 0
|   mday = 0
|   mon = 0
|   year = 0
|   wday = 0
|   yday = 0
|   isdst = 0
|   zone = unknown
| }

Can you also send the values of tm and mktime (tm) that are computed
in datestr (datenum (1969, 1, 1))?

I'd be happy to have replacements for broken versions of mktime and/or
localtime and a configure test to determine when we should use them...

jwe