Georgi Boshnakov
2019-Jun-28 10:23 UTC
[Rd] tools::package_native_routine_registration_skeleton?
> ... reports that the fortran names of all the registered fortran functions are ?undocumented objects?...Those symbols should not be exported. In quantreg v5.33, NAMESPACE has 'exportPattern(".")'. Maybe it is a good time to remove that and export explicitly the symbols that are meant for export. Georgi Boshnakov ------------------------------ Message: 6 Date: Fri, 28 Jun 2019 09:06:52 +0000 From: "Koenker, Roger W" <rkoenker at illinois.edu> To: "r-devel at r-project.org" <r-devel at r-project.org> Subject: [Rd] tools::package_native_routine_registration_skeleton? Message-ID: <00702A4F-67B0-4B0C-A791-E72AF51C4F9C at illinois.edu> Content-Type: text/plain; charset="utf-8" I?m trying to sort out the consequences of adding registration for my quantreg package. I?ve generated a quantreg_init.c file with the directive in the message title, and placed it in my src directory. The first issue is that 3 of the 20, or so registered functions have the same name in fortran as the calling function in R. This seems easy to fix. More puzzling is that R CMD check now reports that the fortran names of all the registered fortran functions are ?undocumented objects?. Can someone point me to the officially sanctioned way to deal with this: presumably I don?t need to document them as if they were R functions. Roger Koenker r.koenker at ucl.ac.uk<mailto:r.koenker at ucl.ac.uk> Department of Economics, UCL London WC1H 0AX. [[alternative HTML version deleted]] ------------------------------ Subject: Digest Footer _______________________________________________ R-devel at r-project.org mailing list DIGESTED https://stat.ethz.ch/mailman/listinfo/r-devel ------------------------------ End of R-devel Digest, Vol 196, Issue 25 ****************************************
Koenker, Roger W
2019-Jun-28 10:27 UTC
[Rd] tools::package_native_routine_registration_skeleton?
Thanks, I was just coming to that conclusion and beginning to look for a way to
make a list of
exportable objects.
Roger Koenker
r.koenker at ucl.ac.uk<mailto:r.koenker at ucl.ac.uk>
Department of Economics, UCL
London WC1H 0AX.
On Jun 28, 2019, at 11:23 AM, Georgi Boshnakov <georgi.boshnakov at
manchester.ac.uk<mailto:georgi.boshnakov at manchester.ac.uk>> wrote:
... reports that the fortran names of all the registered fortran functions are
?undocumented objects?...
Those symbols should not be exported. In quantreg v5.33, NAMESPACE has
'exportPattern(".")'. Maybe it is a good time to remove that
and export explicitly the symbols that are meant for export.
Georgi Boshnakov
------------------------------
Message: 6
Date: Fri, 28 Jun 2019 09:06:52 +0000
From: "Koenker, Roger W" <rkoenker at
illinois.edu<mailto:rkoenker at illinois.edu>>
To: "r-devel at r-project.org<mailto:r-devel at r-project.org>"
<r-devel at r-project.org<mailto:r-devel at r-project.org>>
Subject: [Rd] tools::package_native_routine_registration_skeleton?
Message-ID: <00702A4F-67B0-4B0C-A791-E72AF51C4F9C at
illinois.edu<mailto:00702A4F-67B0-4B0C-A791-E72AF51C4F9C at
illinois.edu>>
Content-Type: text/plain; charset="utf-8"
I?m trying to sort out the consequences of adding registration for my quantreg
package.
I?ve generated a quantreg_init.c file with the directive in the message title,
and placed it
in my src directory. The first issue is that 3 of the 20, or so registered
functions have
the same name in fortran as the calling function in R. This seems easy to fix.
More
puzzling is that R CMD check now reports that the fortran names of all the
registered fortran
functions are ?undocumented objects?. Can someone point me to the officially
sanctioned
way to deal with this: presumably I don?t need to document them as if they were
R functions.
Roger Koenker
r.koenker at ucl.ac.uk<mailto:r.koenker at ucl.ac.uk><mailto:r.koenker
at ucl.ac.uk>
Department of Economics, UCL
London WC1H 0AX.
[[alternative HTML version deleted]]
------------------------------
Subject: Digest Footer
_______________________________________________
R-devel at r-project.org<mailto:R-devel at r-project.org> mailing list
DIGESTED
https://stat.ethz.ch/mailman/listinfo/r-devel
------------------------------
End of R-devel Digest, Vol 196, Issue 25
****************************************
[[alternative HTML version deleted]]
Duncan Murdoch
2019-Jun-28 12:11 UTC
[Rd] tools::package_native_routine_registration_skeleton?
On 28/06/2019 6:27 a.m., Koenker, Roger W wrote:> Thanks, I was just coming to that conclusion and beginning to look for a way to make a list of > exportable objects.After library(quantreg), ls("package:quantreg") will list all the names you currently export. So cat(ls("package:quantreg"), sep = ", ") will print the list in a form suitable for including in the export() directive in your NAMESPACE file. Just delete the names that are meant for internal use only. Duncan Murdoch> > Roger Koenker > r.koenker at ucl.ac.uk<mailto:r.koenker at ucl.ac.uk> > Department of Economics, UCL > London WC1H 0AX. > > > On Jun 28, 2019, at 11:23 AM, Georgi Boshnakov <georgi.boshnakov at manchester.ac.uk<mailto:georgi.boshnakov at manchester.ac.uk>> wrote: > > ... reports that the fortran names of all the registered fortran functions are ?undocumented objects?... > > Those symbols should not be exported. In quantreg v5.33, NAMESPACE has 'exportPattern(".")'. Maybe it is a good time to remove that and export explicitly the symbols that are meant for export. > > Georgi Boshnakov > > > ------------------------------ > > Message: 6 > Date: Fri, 28 Jun 2019 09:06:52 +0000 > From: "Koenker, Roger W" <rkoenker at illinois.edu<mailto:rkoenker at illinois.edu>> > To: "r-devel at r-project.org<mailto:r-devel at r-project.org>" <r-devel at r-project.org<mailto:r-devel at r-project.org>> > Subject: [Rd] tools::package_native_routine_registration_skeleton? > Message-ID: <00702A4F-67B0-4B0C-A791-E72AF51C4F9C at illinois.edu<mailto:00702A4F-67B0-4B0C-A791-E72AF51C4F9C at illinois.edu>> > Content-Type: text/plain; charset="utf-8" > > I?m trying to sort out the consequences of adding registration for my quantreg package. > I?ve generated a quantreg_init.c file with the directive in the message title, and placed it > in my src directory. The first issue is that 3 of the 20, or so registered functions have > the same name in fortran as the calling function in R. This seems easy to fix. More > puzzling is that R CMD check now reports that the fortran names of all the registered fortran > functions are ?undocumented objects?. Can someone point me to the officially sanctioned > way to deal with this: presumably I don?t need to document them as if they were R functions. > > Roger Koenker > r.koenker at ucl.ac.uk<mailto:r.koenker at ucl.ac.uk><mailto:r.koenker at ucl.ac.uk> > Department of Economics, UCL > London WC1H 0AX. > > > > [[alternative HTML version deleted]] > > > > ------------------------------ > > Subject: Digest Footer > > _______________________________________________ > R-devel at r-project.org<mailto:R-devel at r-project.org> mailing list DIGESTED > https://stat.ethz.ch/mailman/listinfo/r-devel > > > ------------------------------ > > End of R-devel Digest, Vol 196, Issue 25 > **************************************** > > > [[alternative HTML version deleted]] > > ______________________________________________ > R-devel at r-project.org mailing list > https://stat.ethz.ch/mailman/listinfo/r-devel >