I think this is a quartz problem interacting with emacs/ESS. When I run the following lines in the *R* buffer,
## ---------- library(grid) library(lattice) library(png) Rlogo.png <- readPNG("/Library/Frameworks/R.framework/Versions/2.15/Resources/library/png/img/Rlogo.png", TRUE) Rlogo.grob <- rasterGrob(Rlogo.png, width=unit(.64,"in"), y=unit(0,"npc")) Rlogo.legend <- list(left=list(fun=Rlogo.grob)) ## example xyplot(1:10 ~ 1:10, legend=Rlogo.legend) ## ---------- the plot shows up on the quartz device ok. When I print from the quartz device, the sample image in the print window has two superimposed graphs. One is the actual image. The other is an upside down mirror image about twice as big. The printed page shows the doubled image. When I print directly to the pdf() device on the Mac, it behaves normally. When I run the same lines above on the windows() device in Windows, it also behaves ok. When I run the same lines from the R gui on the Mac, it behaves normally. I saw this first with R version 2.15.2 RC (2012-10-18 r60960) and then when I updated a few minutes ago to R version 2.15.2 (2012-10-26) -- "Trick or Treat", I still see it. Sometimes, when I attempt to print from the quartz device while running under ESS, I get the pinwheel and this message in the *R* buffer. > Error: REAL() can only be applied to a 'numeric', not a 'NULL' Killing R is the only recourse at this point. Rich [[alternative HTML version deleted]] _______________________________________________ R-SIG-Mac mailing list R-SIG-Mac@r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-mac