[Octave-bug-tracker] [bug #40114] pcolor with NaN produces black grid cells and invalid EPS files

[Alexander Barth]

Follow-up Comment #3, bug #40114 (project octave):



I tested the current version of octave with change
(http://hg.savannah.gnu.org/hgweb/octave/rev/e48a22b94727).
There are still difference in the way NaNs are handeld compared to matlab
(R2013a). Matlab tries to plot as much as possible while this is not the case
for the current dev version of octave (compare pcolor_octave_current.png and
pcolor_matlab.png generated by test_pcolor.m).
In particular a NaN in the parameter c is handeld different than a NaN in the
coordinates. Therefore I think we should not use the clip variable in both
cases.

The attached patch (pcolor_NaN2.patch) makes the behaviour compatible (see
pcolor_octave_with_patch_NaN2.png).

Can somebody have a look to the new patch?
Thanks!


(file #29318, file #29319, file #29320, file #29321)
    _______________________________________________________

Additional Item Attachment:

File name: test_pcolor.m                  Size:0 KB
File name: pcolor_octave_current.png      Size:166 KB
File name: pcolor_octave_with_patch_NaN2.png Size:168 KB
File name: pcolor_matlab.png              Size:48 KB


    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?40114>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/