[Tikzdevice-bugs] Encoding problems in package vignettes

[Prof Brian Ripley]

This concerns packages

Animal CNVassoc DOSim MAMA NMF POT SpatialExtremes actuar adehabitatHR 
alphahull amer aplpack caret catnet chemometrics compareGroups doBy dr 
expm fitdistrplus gRbase geepack genoPlotR ghyp hwde interval marelac 
mhsmm mhurdle nlstools nonrandom nsRFA psych rangeMapper scatterplot3d 
soiltexture spBayes svcR tikzDevice varcompci

As you can see from the CRAN package check page at
http://cran.r-project.org/web/checks/check_summary.html, R-pre and 
R-devel are reporting that one or more of the vignettes from your 
package have an undeclared encoding.  This may be causing problems on 
your own machine or on one of the check machines.  In quite a few 
cases this seems to be unintentional, with e.g. directional quotes or 
times sign in CP1252 being used instead of the correct LaTeX markup 
and hence the PDF contains incorrect glyphs.

To find non-ASCII input, you can use 
tools::showNonASCII(readLines('foo.Rnw'))

It is most portable if you can have ASCII vignettes, but if you really 
need another encoding the way to declare it is

\usepackage[foo]{inputenc}

in the preamble of your vignette, where 'foo' is preferably latin1 or 
utf8 (but latin9 and ansinew seem to work for existing packages).

The 'Writing R Extensions' manual in 2.13.0 beta covers all this.

Can we please have updated versions with these problems corrected.  It 
is also a good time to check your package under 2.13.0 beta (or 
R-devel), and Kurt will expect other problems shown on the check page 
to be resolved.

Brian Ripley

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595