help-octave
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: printing to large size weirdness (qt backend)


From: Ben Abbott
Subject: Re: printing to large size weirdness (qt backend)
Date: Wed, 24 Aug 2016 08:17:07 -0400

On Aug 24, 2016, at 12:07 AM, Dmitri A. Sergatskov <address@hidden> wrote:

I see it on both 4.0.3 and recent dev version. Only with qt backend
(never with fltk). It is not 100% reproducible.

plot(1:10) ; # More complex plots like sombrero() might be better
# No try with different Xsize and Ysize. The "fun" starts when
# either of them exceeds the size of the physical screen.
# png or eps

print ("-SXsize, Ysize", "t.eps","-depsc")

Essentially, sometimes the figure gets clipped to the screen size.
(attached is the outputs of
octave:1> plot(randn(10))
octave:2> print("-S4000,4000","qt4000.eps", "-depsc")
octave:3> print("-S2000,2000","qt2000.eps", "-depsc")
on a computer with 1920x1200 screen.
)

Also notice that the size and the location of the on-screen plot
changes.

On the computer with non-intel graphics the clipped plots sometimes
obviously corrupted.

Dmitri.

I have a suspicion.

Unless osmesa is used, OpenGL printing requires on screen rendering. During printing, the figure is resized such that the size in pixels equals the output size in points (72pts/in). The resizing and repositioning may place a portion of the figure off screen, which would not render during the print process. 

You can prevent the resizing by setting the -SXsize,Ysize to the figure size.

Can you confirm that the clipping is always present when the figure is moved off screen. 

If my guess is correct, a fix would be to use osmesa when printing (even when the figure is visible).

In any event, please file a bug report.

Ben

reply via email to

[Prev in Thread] Current Thread [Next in Thread]