[changeset] colorbar lefts multiplot

[Ben Abbott]

On Apr 1, 2009, at 5:44 PM, Petr Mikulik wrote:

> 	Hello Ben,
>
>>> I've just discovered yet another bug (seems to be bug in 3.1.55,  
>>> not in
>>> 3.0.3):
>>>
>>> hold on
>>> colorbar
>>> imagesc(1./hilb(2));
>>> hold off
>>>
>>> hold on
>>> imagesc(1./hilb(4));
>>> hold off
>>>
>>> Note that it does not hurt saying
>>> 	unset multiplot
>>> if 'multiplot' was used somewhere in past and there is such a  
>>> danger of
>>> having multiplot on after some premature crash.
>>
>> It may be a few days before I have time to dig into this. Do you  
>> have a
>> suggestion for what should be changed Octave's sources?
>
> As far as I know you are now avoiding "set colorbox; replot" of  
> gnuplot,
> then I don't know.

Petr,

I verified that "unset multiplot" was in place. The problem has to do  
with a missing semicolon.

The line below produced the gnuplot error.

plot "-" binary array=2x2 scan=yx origin=(1,1) dx=1 dy=1 using 1 title  
"" with image plot "-" binary array=4x4 scan=yx origin=(1,1) dx=1 dy=1  
using 1 title "" with image ;

If I understand correctly, it should be ...

plot "-" binary array=2x2 scan=yx origin=(1,1) dx=1 dy=1 using 1 title  
"" with image; plot "-" binary array=4x4 scan=yx origin=(1,1) dx=1  
dy=1 using 1 title "" with image ;

Please verify this works for you. I'm unfamiliar with this part of the  
code, so any comments are welcome.

Ben

-------------- next part --------------
A non-text attachment was scrubbed...
Name: changeset-image.patch
Type: application/octet-stream
Size: 1660 bytes
Desc: not available
Url : https://www-old.cae.wisc.edu/pipermail/octave-maintainers/attachments/20090402/256a5bcf/attachment.obj 
-------------- next part --------------