[Rd] package vignettes, suggests and R CMD check

Paul Gilbert pgilbert at bank-banque-canada.ca
Wed Nov 22 16:05:45 CET 2006



Peter Ruckdeschel wrote:
> Dear R-devels,
> 
> We have encountered a problem concerning a dead lock in the dependency graph:
> 
> We have two packages A1 and A2 where A2 extends A1.
> 
> We have one package vignette for both packages, but which should be included
> into package A1.
> 
> So in the "suggests" field of the DESCRIPTION file of A1 we suggest A2
> to be able to produce the vignette at build/install time.
> 
> On the other hand this means that A2 has to be installed before being
> able to check A1, and A1, of course, has to be installed before being
> able to check A2.
> 
> How can this dead lock be resolved?
> 
> Is this a case for a bundle? And can a bundle have its own vignette?

Vignettes just go in one of the packages. There is not really a "bundle 
vignette", and also no really good place to put vignettes if they are 
more general than one of the packages. I'm not sure if a bundle will 
resolve the deadlock. (Kurt might know if it should.) The way I have 
handle this problem in dse is to put pieces of the vignettes in 
different packages, so the part that requires the additional 
functionality goes in A2, and A1 does not require A2. That allows me to 
check the packages individually. Then when I build the bundle I splice 
together the pdf files and put the combined pdf in one of the packages 
(which I think users may find a bit confusing, but there is no other place).

Paul

> 
> Thanks for listening
> Peter, Matthias, Florian, Thomas
> 
> ______________________________________________
> R-devel at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel
====================================================================================

La version française suit le texte anglais.

------------------------------------------------------------------------------------

This email may contain privileged and/or confidential inform...{{dropped}}



More information about the R-devel mailing list