Re: [R-pkg-devel] apparent S3 methods exported but not registered

[Uwe Ligges]

On 01.03.2018 18:40, Dason Kurkiewicz wrote:

They're just note NOTES. Although I don't like the names your functions
take because of the possibility to confuse with s3 if they have been in the
package as long as you say I don't see why this would be an issue. Just
send an explanation when you submit and I don't think there should be any
issues.


It is an issue, as not only you but also R will treat these as S3 methods.
As these are not S3 methods, you shoudl ideally
- Give them new names
- Deprecate the old usage (but still allow it), so users get warnings 
and get migrate to the new names.


Best,
Uwe Ligges






-Dason Kurkiewicz

On Mar 1, 2018 11:36 AM, "Brian G. Peterson"  wrote:

... bumping this in the hopes that someone has an answer ...

On Fri, 2018-01-26 at 10:35 -0600, Brian G. Peterson wrote:

I'm preparing a new release of PerformanceAnalytics, and as usual
this brings the host of CRAN checks.

I get this note:

* checking S3 generic/method consistency ... NOTE
Found the following apparent S3 methods exported but not registered:
   mean.LCL mean.UCL mean.geometric mean.stderr
See section 'Registering S3 methods' in the 'Writing R Extensions'
manual.

these are not S3 methods for 'mean', but rather helper functions that
do other things related to means.

These functions have been in the package for over a decade, are used
in
coursework at a number of universities, and we assume are in user
code.
  So it seems that changing them would be disruptive.

Now, all that said, what to do about it?

We could register them as S3method{} , and I think calling them
explicitly will still work, though this seems inelegant.

I can't find any way to tell R that these *aren't* S3 methods ...

Any insight would be greatly appreciated.

Regards,

Brian




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

[[alternative HTML version deleted]]

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



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

[R-pkg-devel] Does win-builder restrict where binaries can be downloaded from?

[Jialin Ma]

Dear all,

I have the following error on win-builder, which says the url from
anaconda.org can not be opened. The url is valid, does win-builder
restrict the domain where binaries can be downloaded? I have previously
tried download from github, where it works.

* installing *source* package 'symengine' ...
** libs

*** arch - i386
rm -f interface.o symengine.dll 
"D:/RCompile/recent/R/bin/i386/Rscript.exe" "../tools/winlibs.R"
trying URL 'https://anaconda.org/symengine/symengine/0.3.0.353.g22cef70
/download/win-32/symengine-0.3.0.353.g22cef70-vc14_0.tar.bz2'
Error in download.file(link, "symenginelib.tar.bz2", quiet = FALSE) : 
  cannot open URL 'https://anaconda.org/symengine/symengine/0.3.0.353.g
22cef70/download/win-32/symengine-0.3.0.353.g22cef70-vc14_0.tar.bz2'
Calls: download_symengine -> download.file
In addition: Warning message:
In download.file(link, "symenginelib.tar.bz2", quiet = FALSE) :
  InternetOpenUrl failed: 'Die Serververbindung konnte nicht
hergestellt werden.'
Execution halted
make: *** [winlibs] Error 1
ERROR: compilation failed for package 'symengine'
* removing 'd:/RCompile/CRANguest/R-devel/lib/symengine'
In R CMD INSTALL

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