R CMD check and error in an \Sexpr in an Rd file

Martin,

I am pretty sure (but I will probably be proven wrong) that when
r-announce was created it was stated that every email got sent to both
r-help and r-devel,

and I see I have received emails from r-announce
in the past despite only being subscribed to r-devel and not r-help.
For example, I can find the announcement of R-2.13.0 (posted by P
Dalgaard) (and many earlier versions), but not R-2.13.1. ?While I am
pretty sure I have only subscribed to r-help for a brief period many
years ago, I did switch my r-devel subscription from one email address
to another.

I infer from your email that r-announce emails are no longer sent to
r-devel subscribers (which is consistent with the text on
?https://stat.ethz.ch/mailman/listinfo/r-announce
).

Kasper


On Mon, Oct 31, 2011 at 11:42 AM, Martin Maechler
<maechler at stat.math.ethz.ch> wrote:

? ?> Thank you very much Duncan, Uwe and Peter. ?Sorry if I
? ?> missed the announcement, I follow more r-devel than
? ?> r-help, which I find a bit hard to quickly read.

But please --- this concerns every one on R-devel ---
if you do not subscribe to R-help,
then do subscribe to R-announce -- that one has about
one posting *per month* and is really only for important
announcements (and basically reserved for R-core to post).

See a list of *all* ?R-announce postings of ?2011 :

?https://stat.ethz.ch/pipermail/r-announce/2011/date.html

with the 2.14.0 announcement at the end.

Martin Maechler, ETH Zurich


? ?> Will try now on the 2.14.0.

? ?> Renaud

? ?> On 31/10/2011 14:28, Prof Brian Ripley wrote:
? ?>> On Mon, 31 Oct 2011, Renaud Gaujoux wrote:
? ?>>
? ?>>> I do not see it on main CRAN home page.
? ?>>
? ?>> No, but see the announcement on R-announce this morning:
? ?>> https://stat.ethz.ch/pipermail/r-help/2011-October/294203.html
? ?>>
? ?>> Things which need human intervention can take some hours:
? ?>> the CRAN front page and binary distributions are two of
? ?>> those.
? ?>>
? ?>>> Do you mean
? ?>>> http://cran.r-project.org/src/base-prerelease/R-latest.tar.gz
? ?>>> ?
? ?>>>
? ?>>> On 31/10/2011 14:19, Uwe Ligges wrote:
? ?>>>>
? ?>>>>
? ?>>>> On 31.10.2011 13:13, Renaud Gaujoux wrote:
? ?>>>>> Thank you Duncan.
? ?>>>>>
? ?>>>>> I tried with: * using R version 2.14.0 RC (2011-10-24
? ?>>>>> r57417) * using platform: x86_64-unknown-linux-gnu
? ?>>>>> (64-bit)
? ?>>>>>
? ?>>>>> But I still get the errors for verbatim multiline and
? ?>>>>> the strange error if an error occurs in \Sexpr. ?Is
? ?>>>>> your patch included in this version? I will try now
? ?>>>>> with R-rc_2011-10-27_r57452.tar.gz.
? ?>>>>
? ?>>>> Try R-2.14.0, it is already released. Or R-devel.
? ?>>>>
? ?>>>> Best, Uwe Ligges
? ?>>>>
? ?>>>>
? ?>>>>>

Renaud

On 28/10/2011 17:25, Duncan Murdoch wrote:

On 28/10/2011 10:53 AM, Duncan Murdoch wrote:

On 28/10/2011 10:49 AM, Renaud Gaujoux wrote:

Hi,

another Rd related issue I encountered is that if an error occurs

in an

\Sexpr in an Rd file, then on get the following error:

* checking for portable compilation flags in Makevars ... OK
* checking for portable use of $(BLAS_LIBS) and $(LAPACK_LIBS)

... OK

* checking examples ... ERROR
Error in paste(before, x, after, sep = "") : object 'exfile'

not found

Execution halted

To reproduce, put a call like this in an Rd section:

\Sexpr[results=verbatim, stage=render]{x<- 1; stop("sexpr error")}

The strange thing is that it occurs at the example checking step.
Not sure why it does not break before.

Thank you.

Renaud

PS: I am on R version 2.13.2 (2011-09-30) - x86_64-pc-linux-gnu

(64-bit)

I would update to 2.13.2 patched, or the release candidate of
2.14.0.

Oops, sorry, 2.13.2 is "final", so I didn't backport the patch.
2.14.0
is what you should get.

Duncan Murdoch