I'm picking up this 5-year old thread.
1. About the four memory allocations without a stacktrace
I think the four memory allocations without a stacktrace reported by Rprofmem():
> Rprofmem(); x <- raw(2000); Rprofmem("")
> cat(readLines("Rprofmem.out", n=5, warn=FALSE), sep="\n")
192 :360 :360 :
>From this bug report (it's a proposal for speedup only, not a bug),
https://bugs.r-project.org/bugzilla/show_bug.cgi?id=16895#c6
the fact that you can construct factors with non-unique aka
"duplicated" levels in R has been re-raised. As mentioned there,
we had a small discussion here (on 'R-d
Thank you, Prof Brian Ripley. That is a helpful resource!
Since my only non-ASCII characters are in an example data frame, I think I
will use the iconv() function to convert that example data frame into
UTF-8. Then, I will indicate that R version >=2.10 must be in the Depends
field of the DESCRIPT