square() is poor, very specific, clumsy and inconvenient. It could be removed with no regrets
Reported by Samuel GOUGEON (@sgougeon)
BUG DESCRIPTION:
----------------
What are square() strengths? Hard to say...
* square(rect) requires knowing a convenient rect. It is the most difficult task, and square() expects it from the user.
* square() completely fails setting isometric scales when a figure has multiple subplots.
* square() does not prevent the user resizing the figure and breaking isoviewing (when it is by chance... actually set)
* square() does not detect when the figure is docked and then resizes the whole desktop.
* If the largest dimension of the graphical window is bigger than the screen's height, Scilab limits its height to the screen height and square() silently fails.
Reframing of an axes with new bounds is the task of replot(). square() should no be used for that.
Is is worthwhile to keep square() just to make the window square(), while square() has almost nothing to do with isometrizing scales and isoview is already dedicated to proper isometrising?
We don't think so. This is why we recommend removing square().
To make square the current graphic window (if any), a "sqwin" option could be added to isoview() through the single syntax "isoview sqwin".