> OK, I removed did that, and then I got a bunch of errors like so:
> 
> ------------------%<------------------
> gnuplot> plot '-' notitle with linespoints
>                                ^ (originally under 'l' of linespoints)
>          line 0: warning: Skipping data file with no valid points
> 
> gnuplot> plot '-' notitle with linespoints
>                                            ^ (just after linespoints)
>          line 0: x range is invalid
> 
> 
> gnuplot> plot '-' notitle with linespoints
>                                ^ (under the 'l' again)
>          line 0: warning: Skipping data file with no valid points
> 
> gnuplot> plot '-' notitle with linespoints
>                                            ^ (again, just after
> linespoints) line 0: x range is invalid
> ------------------>%------------------
> 
> You get the idea. Given that gnuplot all by itself launches a shell,
> I'm not sure why it's trying to open a datafile, unless you're working
> with temporary files in the background (which is what I recommend for
> transmitting data to gnuplot based on my experience with Asymptote).
> 
> >__As an aside, you really should upgrade since 4.4 added a lot of
> >useful features.
> 
> Yes, but I don't actually use gnuplot. I'm just helping you test. :-)
> 
> > I'll add some checks to make sure gnuplot is runnable and to make
> > sure it uses --persist if it exists; probably will do this later
> > today.
> 
> Yeah, I figured a -v would do the trick, but my gnuplot doesn't look
> like it knows what -v means. Perhaps when you create the gnuplot
> object, you can do it in two stages. In the first, you open a script
> called gnuplot determine-version.gpl, or some such, which just prints
> out the current gnuplot version. Opening that with backticks would get
> the version. Then in the second stage---the real open---you could
> decide if you can use the --persist option and use the |- input
> redirect that you currently employ.
> 
> > Thanks for testing.
> 
> No problem. :-)

Hi David.

I wrote some feature-detection logic to handle the --persist issue.
Also, I just grabbed several older versions of gnuplot (4.0 and 4.2) to
try to reproduce your issue. So far I cannot. The older gnuplots fail
when tst.pl uses features they don't have ('with image', 'pallette',
pdf output), but they work otherwise. Is this what you're seeing? Do
the first few plots in tst.pl work at least?

Currently tst.pl has some dependencies on gnuplot 4.4, but
PDL::Graphics::Gnuplot is generally version agnostic (with one minor
exception I'll deal with). Thus gnuplot 4.2 should very much work. Tell
me if any of the plots work. If not, please remove all but the first
plot() commands in tst.pl, add 'dump => 1' as the first argument, and
run it. Instead of feeding gnuplot, the gnuplot commands should then go
to stdout. At that point you can save those to a file and run gnuplot
directly with that file as an argument. The success or failure of THAT
should be a good indication of what's wrong.

Thanks.
dima

_______________________________________________
Perldl mailing list
[email protected]
http://mailman.jach.hawaii.edu/mailman/listinfo/perldl

Reply via email to