R devel - Sep 2013 - Importing packages in Depend

Hi,

I am currently preparing a new version of my package papeR. When I run R CMD
check using the development version of R I get the following note:

Package in Depends field not imported from: ?nlme?, ?lme4?, ?survival?
  These packages needs to imported from for the case when
  this namespace is loaded but not attached.

I now have problems to fix this issue. It is easy to get rid of two of the
three package warnings by using 

import(lme4)
import(survival) 

in my NAMESPACE. Yet, I cannot import nlme as I then get the following
warnings:

Warning: replacing previous import ?VarCorr? when loading ?lme4?
Warning: replacing previous import ?lmList? when loading ?lme4?

I know that nlme is imported in lme4 but I cannot import both at the same
time. Can anyone help me here? I do *not* want or can use importFrom as I
rely on multiple functions and also rely on non-exported methods such as
summary.lme().

Thanks and all the best,
Benjamin



--
View this message in context:
http://r.789695.n4.nabble.com/Importing-packages-in-Depend-tp4675947.html
Sent from the R devel mailing list archive at Nabble.com.

On Thu, Sep 12, 2013 at 6:49 AM, Benjamin Hofner <benjamin.hofner at
fau.de> wrote:
> Hi,
>
> I am currently preparing a new version of my package papeR. When I run R
CMD
> check using the development version of R I get the following note:
>
> Package in Depends field not imported from: ?nlme?, ?lme4?, ?survival?
>   These packages needs to imported from for the case when
>   this namespace is loaded but not attached.
>
> I now have problems to fix this issue. It is easy to get rid of two of the
> three package warnings by using
>
> import(lme4)
> import(survival)
>
> in my NAMESPACE. Yet, I cannot import nlme as I then get the following
> warnings:
>
> Warning: replacing previous import ?VarCorr? when loading ?lme4?
> Warning: replacing previous import ?lmList? when loading ?lme4?
This happens because 'lme4' re-exports the generic function VarCorr()
that 'nlme' already exports.  R considers this to be a conflict and
generates a warnings, which is detected by 'R CMD check'.   I've
been
trying to argue that this warning is unnecessary/not correct because
in these cases, the two objects in conflict are actually the same
identical object.  You can see this by:
> nlme::VarCorr
function (x, sigma = 1, rdig = 3)
UseMethod("VarCorr")
<bytecode: 0x00000000111eb728>
<environment: namespace:nlme>
> lme4::VarCorr
function (x, sigma = 1, rdig = 3)
UseMethod("VarCorr")
<bytecode: 0x00000000111eb728>
<environment: namespace:nlme>

Note how they are both in the same namespace, i.e.
> identical(lme4::VarCorr, nlme::VarCorr)
[1] TRUE


I've submitted a patch for this 'Bug 15451 - PATCH:
namespaceImportFrom() not to warn when the identical object is
imported twice' on 2013-09-10
[https://bugs.r-project.org/bugzilla3/show_bug.cgi?id=15451].  The
tricky part is to convince R-core.

Otherwise, the only solution I know of is either to use importFrom():s
or to convince the author of 'lme4' to remove that re-export, which
may not be possible.

/Henrik
>
> I know that nlme is imported in lme4 but I cannot import both at the same
> time. Can anyone help me here? I do *not* want or can use importFrom as I
> rely on multiple functions and also rely on non-exported methods such as
> summary.lme().
>
> Thanks and all the best,
> Benjamin
>
>
>
> --
> View this message in context:
http://r.789695.n4.nabble.com/Importing-packages-in-Depend-tp4675947.html
> Sent from the R devel mailing list archive at Nabble.com.
>
> ______________________________________________
> R-devel at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel