"Too many raster images" in devPS.c

Hi,

I am finding the recently added [1] functionality of embedding raster 
images into plots on R devices very useful! Thanks to Paul Murrell and 
others for providing that. I noted that in 
https://svn.r-project.org/R/trunk/src/library/grDevices/src/devPS.c
a macro is defined: #define MAX_RASTERS 64, and consequently, I get

Error in grid.Call.graphics("L_raster", x$raster, x$x, x$y, x$width, 
x$height,  :
   Too many raster images

even for relatively innocent graphics, such as extensions of [2] (which 
I made with Bioconductor's "splots" package). Besides that, I imagine 
that raster images could be useful as 'glyphs' in various types of plots.

Besides the not so helpful option of patching that macro in my private 
copy of R, is there an intention to extend this functionality to 
accommodate for larger plots more generally?

[1] http://developer.r-project.org/Raster/raster-RFC.html
[2] http://www.ebi.ac.uk/~huber/pub/Druggable_ratio_1_before.pdf

Thank you and best wishes,
      Wolfgang

--
Wolfgang Huber
EMBL
http://www.embl.de/research/units/genome_biology/huber/contact

Hi,
I am finding the recently added [1] functionality of embedding raster images into plots on R devices very useful! Thanks to Paul Murrell and others for providing that. I noted that in https://svn.r-project.org/R/trunk/src/library/grDevices/src/devPS.c
a macro is defined: #define MAX_RASTERS 64, and consequently, I get
Error in grid.Call.graphics("L_raster", x$raster, x$x, x$y, x$width, x$height,  :
  Too many raster images
even for relatively innocent graphics, such as extensions of [2] (which I made with Bioconductor's "splots" package). Besides that, I imagine that raster images could be useful as 'glyphs' in various types of plots.
Besides the not so helpful option of patching that macro in my private copy of R, is there an intention to extend this functionality to accommodate for larger plots more generally?

[1] http://developer.r-project.org/Raster/raster-RFC.html
[2] http://www.ebi.ac.uk/~huber/pub/Druggable_ratio_1_before.pdf
Thank you and best wishes,
     Wolfgang
--
Wolfgang Huber
EMBL
http://www.embl.de/research/units/genome_biology/huber/contact

Hi Paul

from a user perspective, this sounds reasonable.

         Thank you very much, 
               Wolfgang


Il giorno Jan 28, 2010, alle ore 2:44 AM, Paul Murrell ha scritto:

Wolfgang Huber wrote:

Hi,
I am finding the recently added [1] functionality of embedding raster images into plots on R devices very useful! Thanks to Paul Murrell and others for providing that. I noted that in https://svn.r-project.org/R/trunk/src/library/grDevices/src/devPS.c
a macro is defined: #define MAX_RASTERS 64, and consequently, I get
Error in grid.Call.graphics("L_raster", x$raster, x$x, x$y, x$width, x$height,  :
  Too many raster images
even for relatively innocent graphics, such as extensions of [2] (which I made with Bioconductor's "splots" package). Besides that, I imagine that raster images could be useful as 'glyphs' in various types of plots.
Besides the not so helpful option of patching that macro in my private copy of R, is there an intention to extend this functionality to accommodate for larger plots more generally?

A simple solution (given the current implementation) would be to allow the user to specify the max number of raster images when starting a PDF file, e.g., ...

pdf("plotwithlotsofimages.pdf", maxRaster=1024)

Would that suffice?

Paul

[1] http://developer.r-project.org/Raster/raster-RFC.html
[2] http://www.ebi.ac.uk/~huber/pub/Druggable_ratio_1_before.pdf
Thank you and best wishes,
     Wolfgang
--
Wolfgang Huber
EMBL
http://www.embl.de/research/units/genome_biology/huber/contact