colorbar() rotation/translation bugs [message #84245] |
Sat, 11 May 2013 16:48  |
dg86
Messages: 118 Registered: September 2012
|
Senior Member |
|
|
Dear Folks,
The built-in colorbar() function graphics routine appears to exhibit
bugs when the colorbar is moved around programmatically. The problem
is particularly evident if the graphic subsequently is saved. Here's an example
a = surface(/test) ; create a graphic
b = colorbar(target = a) ; the colorbar is off at the bottom of the screen. Oh well.
b.rotate, 20 ; rotate the colorbar -- this works on the screen
b.save, 'badgraphic.png'
When I open badgraphic.png, the colorbar is not rotated, it's just fatter.
Similarly, if I try
b.translate, /reset
instead of the rotation, badgraphic.png has the colortable above the
surface, whereas the screen shows it occluded by the surface. In this case,
the bug is more subtle, but should be counted as a bug nonetheless.
I'm running IDL 8.2.2 (64 bit) under Mac OS X 10.7.5.
Is this a known bug? Does anyone have workarounds?
Thanks,
David
|
|
|
Re: colorbar() rotation/translation bugs [message #84330 is a reply to message #84245] |
Mon, 13 May 2013 09:50  |
dg86
Messages: 118 Registered: September 2012
|
Senior Member |
|
|
> As an aside: I hadn't really thought of this before, but the NG colorbar is always "on glass"; it doesn't get transformed into the 3D coordinate system of a surface or contour plot.
Dear Mark,
That's exactly how this came up -- I was trying to rotate the colorbar onto one of the back
surfaces of a 3D plot. It looked great on the screen, but couldn't be saved. I ended up
shearing it into place with the Gimp.
TTFN,
David
|
|
|
Re: colorbar() rotation/translation bugs [message #84334 is a reply to message #84245] |
Mon, 13 May 2013 08:28  |
Mark Piper
Messages: 198 Registered: December 2009
|
Senior Member |
|
|
On Saturday, May 11, 2013 5:48:18 PM UTC-6, David Grier wrote:
> Is this a known bug? Does anyone have workarounds?
This is definitely incorrect behavior, which is still present in 8.2.3. I'll log a report.
As an aside: I hadn't really thought of this before, but the NG colorbar is always "on glass"; it doesn't get transformed into the 3D coordinate system of a surface or contour plot.
mp
|
|
|