Re: splot broken in rc4?

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

On Wednesday 14 February 2007 01:38, Jacques Le Bourlot wrote:
> Bonjour,
> 
> I'm experiencing a possible bug with 4.2rc4. This used to work fine with
> rc2 and previous releases. I try to do a elementary 3D (s)plot:
> 
> splot "ex_splot.dat" u 1:2:3 w l
> 
> I get an error message:
> "Wrong number of columns in input data - line 1"

I cannot reproduce this with your test file, and really I can't imagine
how you could get this error message with any input file.  

If the program does not like your splot command, it should say:
  "Need 1 or 3 columns for cartesian data"

If the program has trouble reading the file for some reason, 
it should rather give you an error message about 
  "Skipping data file with no valid points".

In no case that I can think of would it give the error message you
report.  So I am mystified.  It makes no sense to me.

What happens if you bypass the actual file contents by saying?

   splot 'ex_splot.dat' u (rand(0)):(rand(0)):(rand(0)) w l
  

Is it possible that you have some previous commands active, 
read in from ~/.gnuplot when you started the program?

> PS: I compiled on a SuSE 9.3 (with all latest patches and upgrades). wxt
> works fine, x11 too, etc...

Could you please try again after compiling with   CFLAGS  -g -O0 ?

 

-- 
Ethan A Merritt
Biomolecular Structure Center
University of Washington, Seattle 98195-7742