[Rd] \link{} to help pages in Debian

Prof Brian Ripley ripley at stats.ox.ac.uk
Wed Dec 1 08:38:38 CET 2004


On Wed, 1 Dec 2004, Iago Mosqueira wrote:

> On Tue, 2004-11-30 at 17:56, Prof Brian Ripley wrote:
>
>> Yes, and it is true as they are all linked to the same tree when
>> help.start is run.  You do need the alternative trees to be in your
>> .libPath() at that time.
>
> My question is then what to do about it from the package maintainer
> point of view. I want to ensure that links in help pages work regardless
> of library location.

That's guaranteed by the R install scripts, just by following `Writing R 
Extensions'.  *If* it is not working for you, you are doing something 
which you are not telling us.

> For example, my Debian system shows
>
>> .libPaths()
> [1] "/usr/local/lib/R/site-library" "/usr/lib/R/site-library"
> [3] "/usr/lib/R/library"
>
> So when running R CMD INSTALL I am placing packages by default in
> .libPaths()[1], but debian packages get installed in .libPaths()[3]. I
> assume this is a Debian package maintainer decision.
>
> Should I use this information in any way for packae installation? Or
> should R install routine take care of it? The information in R
> administration manual was not clear enough for me, I am afraid.

No, yes, your messages are not clear enough for your helpers.

>> Only by default, just as by default they are under Linux, so
> `obviously'
>> you don't know what happens under either.  Is Windows relevant here
> (the
>> picture is a lot more complicated there)?
>
> The point of windows was the same package works fine in Windows, in a
> default installation with one library.

Ah, but that is not what you actually said (an example of the lack of 
clarity).

I still don't know if you are imagining that there might be a problem that 
you want to write a package to avoid, or that you have a current problem. 
If the latter, please start again with the full details: which packages, 
the commands you used to install them, what happens and what the links in 
the html file concerned are.

-- 
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



More information about the R-devel mailing list