[Changeset] Add data sources and line series

[John W. Eaton]

On 28-Aug-2008, David Bateman wrote:

| The attached changeset adds data sources and the line series objects. 
| Basically a line series wraps the underlying line object with exactly 
| the same data available from the hggroup as the line object itself. So 
| its use is not immediately obvious. However when combined with 
| datasources it makes sense. Consider the code
| 
|            x = 0:0.1:10;
|            y = sin (x);
|            plot (x, y, "ydatasource", "y");
|            for i = 1 : 100
|              pause(0.1)
|              y = sin (x + 0.1 * i);
|              refreshdata();
|            endfor

This seems to work correctly for me with the gnuplot backend, so I've
applied the patch.

But now when I try plotting with the fltk backend, only the first plot
is displayed and even an explicit drawnow does not refresh the plot
window.  For example,

  octave:1> backend ("fltk")
  octave:2> somebrero

works, then after

  octave:3> peaks

the sombrero is still displayed.  Is it just me?  I doubt that this is
related to your patch, but I just happened to notice it now.

| The y variable is linked to the plot and the plot is updated every time 
| the refreshdata function is called. This will be useful to implement the 
| "linkdata" function as well, though to do that needs some thought on 
| where/when the drawnow/refresh should be called to update the figures, 
| so I haven't done that yet. Perhaps the linking should be the 
| responsibility of the backend in any case.

I have no experience with the linkdata function.  It appears to be a
relatively recent addition.

| I've also included a partially working version of the linkprop function. 
| It is partially working in that the linking is done correctly, but 
| linkprop is supposed to return an object that when it is destroyed 
| removes the link. This is related to the OnCleanUp function and so can't 
| be implemented without similar functionality in Octave.

OK.

| Finally this patch exposes the zlim issue mentioned in by previous 
| e-mail.

First, I will try to find out what is happening with the strange
hggroup -> text conversion that I'm seeing.  Do you not see that
behavior?

| The only group object still to implement are contour, errorbar, 
| surface and scatter groups. I'll do this soon.

OK.  Thanks for looking at all of these.

Thanks,

jwe