[R-pkg-devel] doMC dependency

Maxime Turgeon maxime.turgeon at mail.mcgill.ca
Mon May 15 21:08:54 CEST 2017


If you look at the CRAN page for the package doMC, you'll notice that in the reverse dependencies, there are only "reverse suggests" and "reverse enhances" (no "reverse depends"). Looking at the source code for some of the packages in "reverse suggests" should give you an answer to your question.


Max

________________________________
From: R-package-devel <r-package-devel-bounces at r-project.org> on behalf of Brian G. Peterson <brian at braverock.com>
Sent: May 15, 2017 3:04:10 PM
To: Christopher Lalansingh; r-package-devel at r-project.org
Subject: Re: [R-pkg-devel] doMC dependency

On Mon, 2017-05-15 at 16:09 +0000, Christopher Lalansingh wrote:
> Hi everyone,
>
>
> I'm preparing to release a package which uses doMC for
> parallelization in certain functions. I've used require(doMC) to
> specify this for these functions, and importFrom("doMC",
> "registerDoMC") in my NAMESPACE.
>
>
> I ran all the R CMD --as-cran checks under 3.4.0 and R-devel on
> Debian 8 sucessfully, but when I submitted to CRAN the package failed
> when built on Windows. This is because doMC is not available for
> Windows. It was recommended by Uwe Ligges that I move doMC to
> "Suggests", but since doMC is required for functions in my package
> this does not seem to work:

I'd suggest using doParallel instead of doMC if you want your package
to work on Windows too.  the 'parallel' package uses forking clusters
on *nix and TCP/snow style clusters on Windows.

You'll need to do a little work to detect things, or give users
instructions for initializing a cluster, but it should work.


> * checking package dependencies ... ERROR
> Namespace dependency not required: 'doMC'
>
>
> Is there a way to keep doMC in 'Suggests' by changing how I
> attach/load doMC in my package?

______________________________________________
R-package-devel at r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-package-devel

	[[alternative HTML version deleted]]



More information about the R-package-devel mailing list