Survex and dates

Wookey wookey at wookware.org
Wed Jul 28 12:45:55 BST 2010


+++ Olly Betts [2010-07-28 12:39 +0100]:
> On Wed, Jul 28, 2010 at 11:57:24AM +0100, Mark Shinwell wrote:
> > On Wed, Jul 28, 2010 at 12:36:02PM +0200, Graham Mullan wrote:
> > > Is it going to be possible to deal with the Unix date problem and get  
> > > Survex to recognise dates earlier than 1970, or are we tied down by the  
> > > vagaries of one operating system?
> > 
> > I'm not sure what you mean by the Unix date problem.  The usual practice is
> > for the standard Unix time type (time_t) to be signed, so you can go back
> > before the 1970 epoch.  A 32-bit time_t permits dates from 1901..2038 and the
> > 64-bit version goes back older than any cave.
> 
> Yes, there's no "Unix date problem" here - the only "Unix date problem" I
> am aware of is that signed 32-bit time_t wraps in 2038:
> 
> http://en.wikipedia.org/wiki/Year_2038_problem
> 
> Also, note that Microsoft Windows uses the 1970 epoch too:
> 
> http://msdn.microsoft.com/en-us/library/1f4c8f33%28VS.80%29.aspx
> 
> Currently the .3d format stores the date as seconds since 1970, and this
> is documented as unsigned, though cavern currently rejects any dates
> outside 1970-2038.  This is really just because I haven't had much feedback
> on what ranges of dates are useful, and what granularity of time stamping
> is useful, so it's just in the easiest format to store - that which we
> get from the system.
> 
> So anyway, why not tell us what dates you actually want to be able to use?

It's been mentioned on this list before. (Graham = file a bug on the
new Trac system - that might help avoid the details getting
forgotten). They have lots of 1960s data, and some 1950s IIRC and data
back to 1800s is plausible but in practice if we bottomed out at 1901
like 32-bit time_t that would be fine for all sensible purposoes.

Wookey
-- 
Principal hats:  Linaro, Emdebian, Wookware, Balloonboard, ARM
http://wookware.org/



More information about the Survex mailing list