Re: Help needed with printing bug

[Ben Abbott]

> On May 4, 2015, at 2:27 PM, John W. Eaton <address@hidden> wrote:
> 
> On 05/04/2015 12:10 PM, Dmitri A. Sergatskov wrote:
>> ​fltk backend never really worked for me. In rc-4 (Fedora 21 x86_64) it
>> segfaults.
>> I.e.:
>> ./run-octave (or --no-gui)
>> graphics_toolkit fltk
>> ​plot(1:10)
>> a gray fltk windows pops up​ and then it crashes with sig.11.
>> 
>> Perhaps it is due to a faulty (free) nvidia driver (the computer
>> has quadro 2000 GF106GL card), so it is technically not
>> octave fault. I suspect this could be the issue with Windows
> 
> Yes, but I'm getting reports that 3.8.2 works while the current release 
> candidates don't on the same system.
> 
> So maybe we were doing something differently in the past that avoided the 
> bug(s).
> 
> jwe

Perhaps the problem is related to the z-order of various objects? In some 
instances, maybe the change below results in the polygon used to color the axes 
box being rendered on top of the lines of the axes box?

        http://hg.savannah.gnu.org/hgweb/octave/rev/abd8d8c686c9

Has anyone who can reproduce the problem tried removing the axes polygon to see 
if the plot box is/was hidden behind it?

        plot (rand (3))
        print -depsc withcolor.eps
        set (gca (), “color", “none")
        print -depsc withoutcolor.eps

If not, this 

Ben