Re: [Gnuplot-bugs] filledcurve bug?

[Hans-Bernhard B. <br...@ph...>]

On Mon, 23 Feb 2004, Eric Gullichsen wrote:

> 
> There appears to be a bug related to filledcurve.
> 
> I believe:
> 
>    plot [0:1000] [-10:10] \
>      (x>=400 && x<600 ? 2  : 1/0) with filledcurve y1=0 lt 1, \
>      (x>600 && x<650  ? 2  : 1/0) with filledcurve y1=0 lt 2
> 
> should result in the colored rectangles being adjacent at x=600.

Not as you wrote it.  Note that for x==600, *both* plotted functions
are undefined, and that gnuplot doesn't have infinite resolution. 
Functions are sampled at a finite number of points (--> "set samples").
If there doesn't happen to be a sample at exactly x==600, this can't
work as you wanted it to.

So: make those comparision x<=600 and x>=600, and set samples to
a value that guarantees a sample at x==600, and you might be getting
somewhere.  (set samples 101 works).

-- 
Hans-Bernhard Broeker (br...@ph...)
Even if all the snow were burnt, ashes would remain.