Re: 2.9.9?

[John W. Eaton]

On 26-Sep-2006, David Bateman wrote:

| David Bateman wrote:
| > John W. Eaton wrote:
| >   
| >> On 26-Sep-2006, David Bateman wrote:
| >>
| >> | John W. Eaton wrote:
| >> | > Are there any pending bugs that should be fixed or any other reasons
| >> | > to delay a new snapshot?
| >> | >
| >> | > Thanks,
| >> | >
| >> | > jwe
| >> | >
| >> | >   
| >> | I'd hoped to have eigs ready, but I still have a rare seg-fault to
| >> | address. When were you thinking of doing the release? By the weekend I
| >> | hope to have eigs in a better shape.
| >>
| >> I can wait a short while, but would like to make a new snapshot by
| >> next week at least.
| >>
| >> jwe
| >>
| >>   
| >>     
| > There is one bug that was identify in 2.9.8 that I don't believe is
| > fixed yet. The fact that plot3 and mesh are ignoring the "hold on" I
| > looked at this and the problem appears to be in __gnuplot_splot__ itself
| > as running
| >
| >     __gnuplot_set__ parametric;
| >     __gnuplot_raw__ ("set nohidden3d;\n");
| >
| > and then running
| >
| > x = rand(10,1);
| > y = rand(10,1);
| > z = rand(10,1);
| > tmp = [([x; NaN*ones(1,size(x,2))])(:),([y;
| > NaN*ones(1,size(y,2))])(:),([z; NaN*ones(1,size(z,2))])(:)];
| >
| > __gnuplot_splot__ tmp
| >
| > multiple times doesn't result in multiple plots on the same figure.
| > However, its really not obvious to me why this is happening.
| >
| > D.
| >
| >   
| Ok I believe the following fixes this issue.

Thanks.  I was headed in this direction.  So now you also understand
that it is up to the plot commands to save data when hold is "on".  So
I think we will also need some fixes for mesh, or we will need to
somehow merge all the 3D plotting functions so they go through one
function, similar to the way 2D plotting is supposed to work, and I
think mostly does as functions like semilogx eventually call __plt__.

Your patch is the right idea, but I think I want to hide some of these
details in some functions since the code is mostly duplicated in both
__plt__ and __plt3__ now.  I'll take a look at this and commit a
patch.

Thanks,

jwe