Boost logo

Boost :

From: Jeff Garland (jeff_at_[hidden])
Date: 2003-01-22 08:26:35


> > Seems like there will be cases where you will want to print
> > the 'not_a_date' value instead of throwing an exception.
> > Without a bit more detail, however, it is unclear to me
> > how this bears on a default constructor. You can already
> > test a date to find out if it is not_a_date_time?
>
> Well, I want to construct a datetime structure with "not a date" for the
> date, but a valid time. And similarly, for a date value, but not a valid
> time. I'm not sure what "not_a_date_time" means. Does it mean the
> entire datetime is invalid? Just the date? The time?

Ah, I see -- the intent is that the whole thing is 'not_a_date_time',
but you are right the interface isn't totally clear on that. Looks like
we really need a query for this on the time directly -- not to mention
a constructor.

Jeff


Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk