[Octave-bug-tracker] [bug #57418] [wish] Simplify legend column number computation for concistency

[Pantxo Diribarne]

URL:
  <https://savannah.gnu.org/bugs/?57418>

                 Summary: [wish] Simplify legend column number computation for
concistency
                 Project: GNU Octave
            Submitted by: pantxo
            Submitted on: dim. 15 déc. 2019 18:50:07 UTC
                Category: Plotting with OpenGL
                Severity: 3 - Normal
                Priority: 5 - Normal
              Item Group: Inaccurate Result
                  Status: None
             Assigned to: None
         Originator Name: 
        Originator Email: 
             Open/Closed: Open
         Discussion Lock: Any
                 Release: dev
        Operating System: Any

    _______________________________________________________

Details:

The new legend code currently tries to implement the old Octave specific
feature of computing the "best" number of columns. This  leads to inconsistent
results when changing the "location" after creation  since the legend is not
redrawn, only its position is adjusted.


figure (89); clf;
plot (rand (3, 20));
hl = legend ("fontsize", 30, "location", "northoutside");

figure (90); clf;
plot (rand (3, 20));
hl = legend ("fontsize", 30);
set (hl, "location", "northoutside");


Note that changing the peer axes size will lead to the same kind of
inconsistency. 
In order to solve this, we would need to force a legend redraw, which is much
slower, even for small legends (most of them). 

Now that we implement the numcolmuns* properties, there is a simple way to
adjust the number of columns manually for large legends. So I'd like remove
this complicated automatic column number feature and set "ncol = 1" for
vertical orientation and "ncol = nitem" for horizontal (this is also what ML
does).

I'll attach a patch that implements the above when I have a bug number.
 




    _______________________________________________________

Reply to this item at:

  <https://savannah.gnu.org/bugs/?57418>

_______________________________________________
  Message posté via Savannah
  https://savannah.gnu.org/