On Fri, 22 Jun 2007, hadley wickham wrote:
> On 6/22/07, Prof Brian Ripley <[EMAIL PROTECTED]> wrote:
>> On Fri, 22 Jun 2007, hadley wickham wrote:
>>
>> > What's the recommended way to check if an internet connection is
>> > available across platforms? I was using is.null(nsl()) but this does
>> > not work on windows.
>>
>> I don't know a way to do this on Windows without a possible very long
>> hang. It's complicated by the use of two alternative internet connection
>> mechanisms on Windows and having little control over the IE-based one.
>>
>> Another issue is software firewalls: they can take many minutes to
>> decide to let a new program talk to the Internet if they 'call home' to
>> ask advice.
>
> Thanks for those details, I hadn't considered the problems that might
> be associated with firewalls.
>
> To give a little more context, I have an example in one of my packages
> which downloads some data off the web. I don't really want to include
> the data in my package, since it's not my data (although I'm fairly
> certain I could get permission to include it).
There are several packages on CRAN which do that. Some do take a long
time, and as a result are only partially checked. If I recall right,
'celsius' is an example on the stoplist (it takes many tens of minutes to
get its data) and 'tseries' and a couple more fail occasionally when the
data are not available in time.
> So really it doesn't matter if there will be a long delay, provided
> that there won't be a long delay for the windows CRAN server. I could
> wrap the entire example in dontrun, but then that loses the useful
> checking I get locally when running r cmd check.
You could also not run the example on Windows by wrapping it in
if(.Platform$OS.type != "windows) {}
but I don't see that as necessary.
--
Brian D. Ripley, [EMAIL PROTECTED]
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
______________________________________________
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
