[R-pkg-devel] Handling Not-Always-Needed Dependencies?

[Uwe Ligges]

On 03.08.2016 14:24, Dirk Eddelbuettel wrote:
>
> On 2 August 2016 at 19:45, Duncan Murdoch wrote:
> | Why not put together code to implement your idea, and see how big the
> | problem would be to phase it in, by seeing how many packages fail under it?
>
> Ahh, the old 'put-up-or-shut-up' gambit, very nice. Big fan of that myself.
>
> I have been sitting on the need to build better infrastructure for my reverse
> depends checks anyway, and may get there over the fall.  Or not as I have
> been saying that for a long time...  It will likely be Docker-based which may
> or may not be suitable for CRAN, but should be for r-hub which may get this
> all by itself too.  Not sure of this can be done in R alone.
>
> Then again, users of TravisCI know that just toggling
>
> _R_CHECK_FORCE_SUGGESTS_=FALSE

I was travelling, hence a delayed response:

Why users of TravisCI? It is documented in the manual. Setting it to 
true gives an error in the check if the suggested package is not 
available for the check. And typically you want to have it available to 
be able to actually check all parts of your package.
So this is not what you are looking for.

What you suggest means that we would have to run the checks twice, as 
Duncan explained already, once with and once without the suggetsed 
package installed.

Best,
Uwe






>
> does it too so in that sense it is already there, but not used?
>
> Dirk
>