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

[Dirk Eddelbuettel]

On 3 August 2016 at 17:39, Uwe Ligges wrote:
| On 03.08.2016 17:17, Dirk Eddelbuettel wrote:
| > On 3 August 2016 at 16:21, Uwe Ligges wrote:
| >
| > But then you are treating Suggests as Depends and installing irregardless.
| 
| Yes, because I want to check all the runtime code on CRAN and not only 
| subsets.

To me, the intent of Suggests is to create such subsets.

If you prohibit its use, may as well remove it (or rename it to, say,
"VignetteDepends" or whatnot).

| > When we let R do its business Suggests is (currently) interpreted as
| > Depends.
| 
| No, as we check installtion without suggetsed packages.

I would do the same for Suggests.  Test without them.

| > As Thomas pointed out, Suggests: seems to be overloaded already to two
| > distinct use cases.
| 
| Yes, and I know that a package should work without the suggestes 
| installed, we do not disagree there.

Hooray. Agreement.

| So we agree that it would be nice to test with and without installed 
| suggested packages? And the checks should give fine results in both 
| cases, right?

I would suggest to modify the behaviour to test without suggested packages as
a default.  Should the suggested package be present on the box a test could
be conditional.

A side effect would be to make things faster as fewer packages are involved.

Dirk

-- 
http://dirk.eddelbuettel.com | @eddelbuettel | edd at debian.org