Re: GPGET('terminal')

[Ethan M. <merritt@u.washington.edu>]

On Thursday 18 May 2006 09:35 am, Hans-Bernhard Br=F6ker wrote:
> Ethan Merritt wrote:
> > gnuplot> show xrange=20
> >
> >   set xdata time
> >   set xrange [ "01/06/93\t0000" : "01/11/93\t0000" ] noreverse
> > nowriteback
> >
> > gnuplot> print GPVAL_X_MIN
> > -207792000.0
> > gnuplot> print GPVAL_X_MAX
> > -194572800.0
>
> That's to be expected.  Time internally is kept in seconds since the
> millenium.  I.e. there's nothing wrong with the above.

I understand that the values are correct in some abstract sense
(apart from being printed as negative numbers).  But if the point
of these GPVAL_* variables is to make the information available to
a script, this format fails.  We have no mechanism in place to
use the value that is returned.

=46or example, if you want to lock the x2 axis to agree with x1,
the following seems the obvious way to do it:

  set x2range[ GPVAL_X_MIN : GPVAL_X_MAX ]

But that won't work in the above case. =20
I admit to being very unfamiliar with the time/date code.
I recently tried to use it for what I thought was a simple
case, but got thoroughly tangled up.  So I don't even have
a suggestion as to how this *should* work, or what would need
to be changed.

Perhaps it would be sufficient to provide conversion routines?
   internal-time-in-seconds  <=3D=3D>  timefmt string

Then (I think) it would work to say something like

  set x2range[ timestring(GPVAL_X_MIN) : timestring(GPVAL_X_MAX) ]


=2D-=20
Ethan A Merritt
Biomolecular Structure Center
University of Washington, Seattle WA