If one has the current stable release (ie from CRAN) of ggplot2, then attempting to pass a gtable (returned by arrangeGrob) to ggsave gives an error ("plot should be a ggplot2 plot"). Not your problem per se but I guess it could be considered a regression and I thought I'd just let you know.
I imagine this can be fixed by "faking" the class of the gtable but I am not an R developer by any means. Thanks for the useful software!
it works with the dev version of ggplot2, Hadley removed the class test in ggsave. Unfortunately it may be a little while before it gets to CRAN, in the meantime you could install ggplot2 from github.
If one has the current stable release (ie from CRAN) of ggplot2, then attempting to pass a gtable (returned by arrangeGrob) to ggsave gives an error ("plot should be a ggplot2 plot"). Not your problem per se but I guess it could be considered a regression and I thought I'd just let you know.
I imagine this can be fixed by "faking" the class of the gtable but I am not an R developer by any means. Thanks for the useful software!