[Rd] [R] x11() graphic device, displaying raster

[Duncan Murdoch]

On 12-01-25 10:51 AM, Uwe Ligges wrote:
>
>
> On 25.01.2012 16:44, Johannes Radinger wrote:
>>>>>>>> Hello,
>>>>>>>>
>>>>>>>> I am wondering about the X11() graphic device on Windows.
>>>>>>>> I try to plot a raster image but nothing gets displayed. I
>>>>>>>> found some pages where it is mentioned that x11() not
>>>>>>>> always supports raster rendering.
>>>>>>>> Is there any add on for x11, any update or any R-package
>>>>>>>> which solves that displaying problem in Windows?
>>>>>>>>
>>>>>>>> What I try to test it is an example from the
>>>>>>>> package {raster}:
>>>>>>>>
>>>>>>>> library(raster)
>>>>>>>> DEU_alt<- getData("alt", country="DEU", mask=TRUE)
>>>>>>>> x11()
>>>>>>>> plot(DEU_alt,axes=TRUE)
>>>>>>>>
>>>>>>>> best regards,
>>>>>>>>
>>>>>>>> /johannes
>>>>>>>
>>>>>>> Try an R version that is recent - it works for me.
>>>>>>>
>>>>>>> Uwe Ligges
>>>>>>
>>>>>> I first tried it with R 2.13.2 and raster version 1.9-64
>>>>> (16-January-2012).
>>>>>> Now I also installed the most recent version R 2.14.1 (Platform:
>>>>> i386-pc-mingw32/i386 (32-bit)) and raster version 1.9-64
>>> (16-January-2012).
>>>>>>
>>>>>> But in both cases no success. X11() opens and draws the axis and the
>>>>>> border for the scale but no raster...
>>>>>
>>>>> I do not really understand why you are using x11(). The Windows device
>>>>> is called windows() and you actually do not need to open it, since plot
>>>>> opens it anyway. Or are you under cygwin (which is not really Windows)?
>>>>> If not: Which version of Windows is this?
>>>>
>>>> Of course I do not need to open a graphic device (as you said it is open
>>> when calling plot). I just wanted to make it reproducable as e.g. Mac OS X
>>> opens Quartz as a standard device when calling plot on my Mac machine.
>>>>
>>>> So far as I understand the help are "all the devices (X11(), x11() and
>>> windows() implemented as variants of the same device". Thus it makes no
>>> difference if I call x11() or windows() or if 'plot' opens the device
>>> automatically. In all cases there is no raster displayed.
>>>>
>>>> Just some additional information from the 'grDevices' package:
>>>>> library(grDevices)
>>>>> dev.capabilities()
>>>> $semiTransparency
>>>> [1] TRUE
>>>>
>>>> $transparentBackground
>>>> [1] "fully"
>>>>
>>>> $rasterImage
>>>> [1] "yes"
>>>>
>>>> $capture
>>>> [1] TRUE
>>>>
>>>> $locator
>>>> [1] TRUE
>>>>
>>>> $events
>>>> [1] "MouseDown" "MouseMove" "MouseUp"   "Keybd"
>>>>
>>>> I am running R via the standard RGUI (so no Eclipse, Rtinn etc.). The
>>> machine I am working on is Windows 2008 Server Enterprise (Version 6.0) which
>>> I am remotely accessing from a ThinClient with WindowsXP-embedded.
>>>> This is the standard configuration of our institution. Maybe that
>>> configuration with ThinClients is a reason? But what should I ask our admin? What
>>> should he check etc?
>>>
>>> Yes! the remotesoftware is probably unable to send the rasterimage!
>>>
>>> See ?image:
>>>
>>> "Problems with the rendering of raster images have been reported using
>>> ‘windows()’ devices under Remote Desktop."
>>>
>>
>>
>> I shifted this corresponce over to the more suitable r-devel-list.
>>
>> @Uwe Ligges: Thank you for that usefull but very disappointing message. So it seems that these are know problems caused by the RemoteDesktop Setup.
>>
>> I just wanted to know if the source of the problem is caused by the R-code or the RemoteDesktop Software. If it is caused by the RemoteDesktop (as you stated)..are there any settings which need to be set by our IT-admin? Or is there anything I/we can do to solve that problem?
>
> Since it is correctly displayed locally, we can assume this is a problem
> of the RDP connection. We do not know of a setting that enables the
> feature via RemoteDesktop, but if you know any, please let us know and
> we will include this in the documentation as a hint for others.
>

I would add to this:  please report this as a bug to your vendor 
(presumably Microsoft).  They may have a workaround, or may point out 
something R could do to avoid it.

Duncan Murdoch