[Rd] R 3.1.1 and 3.1.2 both fail their test suites

Martin Maechler maechler at stat.math.ethz.ch
Mon Nov 10 17:52:39 CET 2014


>>>>> Duncan Murdoch <murdoch.duncan at gmail.com>
>>>>>     on Wed, 5 Nov 2014 10:22:09 -0500 writes:

    > On 05/11/2014 9:36 AM, Peter Simons wrote:
    >> Hi Duncan,
    >> 
    >> > I don't think we should be removing tests for everybody to allow a few
    >> > people to test a build of R that none of us actually use.
    >> 
    >> no tests need to be removed.

    > My response was to Martin, who proposed exactly that.

    >> All that needs to be done is to distinguish
    >> tests that require the recommended packages from those that don't. Then
    >> users can choose which test set they want to run.

    > Go ahead and submit a patch that does this, and I expect it would be 
    > accepted.
    > Duncan Murdoch

I have committed changes (svn 66943 and 66951) which more or
less achieves this.
Basic idea:  example() newly has an argument  'run.dontcheck = interactive()'
      whereas till now it implicitely had  'run.dontcheck = TRUE'
so using example(..) in our examples or strict tests should now
be safe(r).

So, in the current development version of R,
indeed  'make check' passes -- on one platform at least :-) --
even when R was configured to *not* install the recommended
packages alongside.

BTW: In one place, I've used

   base.and.rec <- .packages(all.available=TRUE, lib=.Library)
   example(glm, run.dontcheck = any("MASS" == base.and.rec))

for checking the presence of one (and by implication "all",
almost surely) recommended package.

Martin Maechler

    >> It would be particularly nice if "make check" would do the right thing
    >> automatically based on the choice of --with{,out}-recommended-packages
    >> at ./configure time. Offering two separate "check" targets would be
    >> equally good, though.
    >> 
    >> Best regards,
    >> Peter



More information about the R-devel mailing list