From tom.haber at uhasselt.be  Thu Mar 14 16:23:02 2019
From: tom.haber at uhasselt.be (Tom HABER)
Date: Thu, 14 Mar 2019 16:23:02 +0100
Subject: [Rcpp-devel] How to provide a custom generator/constructor for an
	Rcpp class?
In-Reply-To: <CAH2GPm2_PiMjbC6E1puqfF9kPB96XWKWBwDHyHRZG+zf+SGQqg@mail.gmail.com>
References: <CAH2GPm2_PiMjbC6E1puqfF9kPB96XWKWBwDHyHRZG+zf+SGQqg@mail.gmail.com>
Message-ID: <CAH2GPm1PmUbnYzAp2J704G8f2WmKBXoL-WrrDkY0cSC+GtAWXA@mail.gmail.com>

Hey,

I asked this question on stackoverflow (
https://stackoverflow.com/questions/54234934/how-to-provide-a-custom-generator-constructor-for-an-rcpp-class)
a while back but never got an answer, so I'm trying my luck here.

-----

I've written R bindings for my C++ library, but for usability I would like
to change the behavior of the methods and constructors somewhat in R:
adding defaults and packing variables into a list.

For example: instead of calling the generator ClassName(name, list(var1=1,
var2=2)), I would like to have ClassName(name, var1=1, var2=2)

Initially I was just using loadModules and using the C++ methods directly,
but then switched to setRcppClass which allows me to override the methods
including initialize. At first I tried using callNextMethod to no avail and
then basically resorted to copy-pasting what the Rcpp provided
initialize method
does with some small adjustments.

setRcppClass("ClassName", module="mymodule",
        methods=list(
            initialize=function(.Object, name, ...) {
                args <- list(...)
                argNames <- allNames(args)
                cppArgs <- nzchar(argNames)
                .CppObject <<- .CppGenerator$new(name, args[cppArgs])
            },
            show=function() {
                cat('ClassName of type:', .CppObject$name(), '\n')
            }))

This does almost what I want, except that I need to explicitly name the name
 parameter: ClassName(name="Foo", var1=1, var2=2)

Obviously I also want to avoid copying the Rcpp code, but instead call it
somehow.

Additionally R CMD check is complaining about all references to .CppGenerator
.CppObject allNames
-------

Thanks a lot in advance

Tom Haber
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.r-forge.r-project.org/pipermail/rcpp-devel/attachments/20190314/561b5fa5/attachment-0001.html>

From edd at debian.org  Tue Mar 26 01:10:37 2019
From: edd at debian.org (Dirk Eddelbuettel)
Date: Mon, 25 Mar 2019 19:10:37 -0500
Subject: [Rcpp-devel] macOS and RcppGSL
Message-ID: <23705.28157.414533.380728@rob.eddelbuettel.com>


I got a bug report for RcppGSL on macOS:

    https://github.com/eddelbuettel/rcppgsl/issues/19

and I can't make progress on this for
    a) lack of a mac machine,
    b) inability to reproduce on rhub (no GSL installed; ticked filed
       earlier), and of course
    c) the fact that is passes at CRAN on platforms.

If someone with a mac could chime and re-affirm that RcppGSL 0.3.6 builds for
them I'd appreciate it. If it fails I'd appreciate pointers too :)

Dirk

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

From hankin.robin at gmail.com  Tue Mar 26 01:33:12 2019
From: hankin.robin at gmail.com (robin hankin)
Date: Tue, 26 Mar 2019 13:33:12 +1300
Subject: [Rcpp-devel] macOS and RcppGSL
In-Reply-To: <23705.28157.414533.380728@rob.eddelbuettel.com>
References: <23705.28157.414533.380728@rob.eddelbuettel.com>
Message-ID: <CAHHjBM7zLv5BLsHR8y8rjbZB8HUD6rO-Hjsvz_BdNAFtFqJxqA@mail.gmail.com>

Hi Dirk,

R CMD check RcppGSL_0.3.6.tar.gz clean for me under R-3.5.2, mac os 10.13.6

HTH

Robin


<hankin.robin at gmail.com>
hankin.robin at gmail.com

<hankin.robin at gmail.com>

<hankin.robin at gmail.com>


On Tue, Mar 26, 2019 at 1:10 PM Dirk Eddelbuettel <edd at debian.org> wrote:

>
> I got a bug report for RcppGSL on macOS:
>
>     https://github.com/eddelbuettel/rcppgsl/issues/19
>
> and I can't make progress on this for
>     a) lack of a mac machine,
>     b) inability to reproduce on rhub (no GSL installed; ticked filed
>        earlier), and of course
>     c) the fact that is passes at CRAN on platforms.
>
> If someone with a mac could chime and re-affirm that RcppGSL 0.3.6 builds
> for
> them I'd appreciate it. If it fails I'd appreciate pointers too :)
>
> Dirk
>
> --
> http://dirk.eddelbuettel.com | @eddelbuettel | edd at debian.org
> _______________________________________________
> Rcpp-devel mailing list
> Rcpp-devel at lists.r-forge.r-project.org
> https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/rcpp-devel
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.r-forge.r-project.org/pipermail/rcpp-devel/attachments/20190326/9fb06f49/attachment.html>

From edd at debian.org  Tue Mar 26 02:05:22 2019
From: edd at debian.org (Dirk Eddelbuettel)
Date: Mon, 25 Mar 2019 20:05:22 -0500
Subject: [Rcpp-devel] macOS and RcppGSL
In-Reply-To: <CAHHjBM7zLv5BLsHR8y8rjbZB8HUD6rO-Hjsvz_BdNAFtFqJxqA@mail.gmail.com>
References: <23705.28157.414533.380728@rob.eddelbuettel.com>
 <CAHHjBM7zLv5BLsHR8y8rjbZB8HUD6rO-Hjsvz_BdNAFtFqJxqA@mail.gmail.com>
Message-ID: <23705.31442.978819.997267@rob.eddelbuettel.com>


Hi Robin.

On 26 March 2019 at 13:33, robin hankin wrote:
| R CMD check RcppGSL_0.3.6.tar.gz clean for me under R-3.5.2, mac os 10.13.6

Thanks a lot -- and also to Kevin who chimed in over at GitHub (and for whom
it also works).

Dirk

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