Re: something like perl's 'require 5.4'

[Paul van Delst <paul.vandelst(at)noaa.gov>]

JD Smith wrote:
> 
> Paul van Delst wrote:
> >
> > JD Smith wrote:
> > >
> > > Paul van Delst wrote:
> > > >
> > > > wot about
> > > >
> > > > IDL> print, double('5.4.1') ge 5.4d0
> > > >    1
> > > >
> > >
> > > Because it's exactly the same!  Yes it's ge, but is it gt?
> > >
> > > IDL>  print, double('5.4.1') gt 5.4d0
> > >    0
> > >
> > > IDL>  print, double('5.4.1') eq 5.4d0
> > >    1
> > >
> > > No it's not, it's eq.  Same problem.  So use this if you don't care
> > > about the last digit and don't want to be open about it (it's not
> > > exactly obvious this is the case).  Use the string compare method
> > > otherwise.
> > >
> > > > Doesn't assuage your other concerns regarding the significance of the last digit however.
> > > > I use comparisons like the above for code that contains BREAK, CONTINUE, SWITCH, etc
> > > > statements. Or similar for the version in which pointers and objects were introduced (5.2?
> > > > can't remember).
> > >
> > > The problem here is you'll not err cleanly... unknown control statements
> > > will cause compile errors.  Not a lot that we can do about this.
> >
> > Nuh-uh. They're interpreted as user functions/procedures.
> 
> Right.  I stand corrected.  Just hope they don't have any such named
> routines lying about on their path.

Me too. :o)

> Dereferencing pointers will cause
> compile error in old versions though, right?

I would hope so. But I do have utility and wrapper codes that simply pass pointers along,
i.e. doesn't dereference them.

paulv

-- 
Paul van Delst           A little learning is a dangerous thing;
CIMSS @ NOAA/NCEP        Drink deep, or taste not the Pierian spring;
Ph: (301)763-8000 x7274  There shallow draughts intoxicate the brain,
Fax:(301)763-8545        And drinking largely sobers us again.
paul.vandelst@noaa.gov                   Alexander Pope.