Re: [Plplot-devel] Re: Error building PLplot on Windows, relatedto GD

[Arjen M. <arj...@wl...>]

"Alan W. Irwin" wrote:
> 
> On 2005-07-05 20:48-0700 Walt Brainerd wrote:
> 
> > call plptex(3.0_p, 0.5_p, 1.0_p, 0.0_p, 1.0, &
> >      "Graph of y = f(x)")
> 
> Hi Walt,
> 
> I don't yet know much of anything about modern fortran.  In fact, your
> example was about the first I had ever looked at closely. I assume the _p
> prefix means you are typing all those constants in the above argument list
> as real*8.  But the fifth argument above is 1.0, not 1.0_p.  I suspect
> that typo is the source of your problem with plptex.
> 
> If that is not the source of your problem with plptex, then you will have to
> rely on Arjen to answer your further questions about getting plptex to work
> since he knows modern fortran.
> 

Alan,

the postfix "_p" is Fortran 90/95's way of describing the _kind_ of a 
literal constant. Where FORTRAN77 allows you to do:

double precision x
x = 1.0d00

Fortran 90 and beyond has generalised this notion:

integer, parameter :: p = kind(1.0d00)    ! An alternative to
selected_real_kind()
real (kind=p) :: x

x = 1.0_p

What you noticed, seems to be correct indeed: the one argument is
missing
this postfix. As PLplot uses double precision reals by default, this is
bound
to cause trouble (which is why a module with the PLplot routines'
interfaces
would be so nice!)

Regards,

Arjen