Checking package help file examples

I am trying to figure out the sanctioned way for
'R CMD check pkg' to make sure that the examples
in help files give the expected results.

Writing R Extensions says that check runs the help
file examples (which INSTALL extracts from pkg/man/*.Rd
and puts into files in pkg/R-ex).  It looks like check
concatenates all the example files, pkg/R-ex/*.R, into one
file, pkg/pkg-Ex.R, which it runs to produce pkg/pkg-Ex.Rout.
If there is an error running these files then check aborts.

I don't seen anything in Writing R Extensions about
checking that correct results were produced.  However,
in R 2.7.0, check will compare pkg/pkg-Ex.Rout to
pkg/tests/Examples/pkg/pkg-Ex.Rout.save, if the latter
file is in your package source, and report any differences.

I don't see any mention of this and was wondering if this
method should be recommended or if there was a preferred
way to check help file examples for correctness.  (It would
be nicer if the example() function could also check for
correctness, and the above method doesn't allow for that.)

I use *.R and corresponding *.Rout.save files in the subdirectory "tests" for 
checking packages (see e.g., packages systemfit, micEcon, 
sampleSelection, ...). I think that it is better to separate examples from 
testing, because IMHO there should be a few simple examples but tests should 
contain many different function calls with (partly) unusual arguments to 
check (nearly) the entire functionality of each function.

See also the manual "Writing R Extensions"
http://cran.r-project.org/doc/manuals/R-exts.html#Package-subdirectories
======================================
Subdirectory tests is for additional package-specific test code, similar to 
the specific tests that come with the R distribution. Test code can either be 
provided directly in a .R file, or via a .Rin file containing code which in 
turn creates the corresponding .R file (e.g., by collecting all function 
objects in the package and then calling them with the strangest arguments). 
The results of running a .R file are written to a .Rout file. If there is a 
corresponding .Rout.save file, these two are compared, with differences being 
reported but not causing an error. The directory tests is copied to the check 
area, and the tests are run with the copy as the working directory and with 
R_LIBS set to ensure that the copy of the package installed during testing 
will be found by library(pkg_name). 
======================================

Best wishes,
Arne

Arne Henningsen
http://www.arne-henningsen.name