[Rd] Problem to execute a function loading by my package that I created
Hello,
I'm a beginner on R.
I should create a package on R. The name of my package is SGeMS.
In my package, I have a big program writed in C++. This program compiles in
C++. I put this program in SGeMS/src.
In my program, I added a header file and a source file named respectively
CreateFile.h and CreateFile.cpp.
CreateFile contains 5 functions : File, FileGeneral, FileData, FileVariogram
and FileDistribution.
I use the package Rcpp to compile this. So in CreateFile.h, I added
'RcppExport' in front of all the functions.
I compiled my package with R CMD INSTALL SGeMS. The compilation works.
In R, I put :
> library("SGeMS")
Loading required package: Rcpp
> is.loaded("File")
[1] TRUE
> is.loaded("FileGeneral")
[1] TRUE
> is.loaded("FileData")
[1] TRUE
> is.loaded("FileVariogram")
[1] TRUE
> is.loaded("FileDistribution")
[1] TRUE
> FileGeneral("parameters.par",general,3.0)
Error in .Call("FileGeneral", NameFile, general, SK_mean, PACKAGE = "SGeMS")
:
C symbol name "FileGeneral" not in DLL for package "SGeMS"
So I have a mistake. But I didn't know why.
Is it possible to put folders and subfolders in /src ?
Awaiting reply,
Pauline
--
View this message in context:
http://r.789695.n4.nabble.com/Problem-to-execute-a-function-loading-by-my-package-that-I-created-tp3568081p3568081.html
Sent from the R devel mailing list archive at Nabble.com.
__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
Re: [Rd] [R-SIG-Mac] check leads to .o files and to packagename-Ex.R
On 06/02/2011 06:47 AM, Benilton Carvalho wrote: > 'R CMD check' should be applied on the .tar.gz, not on the source directory. Why? The help says: "Check R packages from package sources, which can be directories or package 'tar' archives with extension '.tar.gz', '.tar.bz2' or '.tgz'." I just skimmed through the relevant section (1.3.1) in the R Extensions manual, and it doesn't say anything about running on tarballs being preferred. Add my vote to the wishlist that the src directory should get cleaned after R CMD check. Ben Bolker > > So, it'd be something like: > > R CMD build pkg > R CMD check pkg_version.tar.gz > > b > > On 2 June 2011 06:54, Marius Hofert wrote: >> Dear expeRts, >> >> I work on the R package "nacopula" >> (https://r-forge.r-project.org/projects/nacopula/) under Mac OS X 10.6.7 >> (MacBook Pro). The session info is: >> R version 2.14.0 Under development (unstable) (2011-05-02 r55730) >> Platform: x86_64-apple-darwin9.8.0/x86_64 (64-bit) >> >> When I apply "R CMD check nacopula" my source directory is filled with .o >> files (which leads to >> a warning for the next check). Further, if I use R 2.14.0, the check >> produces a file "nacopula-Ex.R" >> on the top level, i.e., where folders like R, src, tests, man, inst, and >> demo reside. >> Shouldn't the source directory stay nice-and-clean, without being filled >> with .o >> files and example files? >> >> Cheers, >> >> Marius >> >> Attachment: screen shot >> >> >> ___ >> R-SIG-Mac mailing list >> r-sig-...@r-project.org >> https://stat.ethz.ch/mailman/listinfo/r-sig-mac >> >> > > > __ R-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-devel
Re: [Rd] [R-SIG-Mac] check leads to .o files and to packagename-Ex.R
On 11-06-02 9:06 AM, Ben Bolker wrote: On 06/02/2011 06:47 AM, Benilton Carvalho wrote: 'R CMD check' should be applied on the .tar.gz, not on the source directory. Why? Because that is what you will send to others. The reason to check a directory instead of a tar file is to save time: you don't need to recompile all the files, as a normal check does. The help says: "Check R packages from package sources, which can be directories or package 'tar' archives with extension '.tar.gz', '.tar.bz2' or '.tgz'." I just skimmed through the relevant section (1.3.1) in the R Extensions manual, and it doesn't say anything about running on tarballs being preferred. Add my vote to the wishlist that the src directory should get cleaned after R CMD check. Then the advantage of checking a directory would be lost. Duncan Murdoch Ben Bolker So, it'd be something like: R CMD build pkg R CMD check pkg_version.tar.gz b On 2 June 2011 06:54, Marius Hofert wrote: Dear expeRts, I work on the R package "nacopula" (https://r-forge.r-project.org/projects/nacopula/) under Mac OS X 10.6.7 (MacBook Pro). The session info is: R version 2.14.0 Under development (unstable) (2011-05-02 r55730) Platform: x86_64-apple-darwin9.8.0/x86_64 (64-bit) When I apply "R CMD check nacopula" my source directory is filled with .o files (which leads to a warning for the next check). Further, if I use R 2.14.0, the check produces a file "nacopula-Ex.R" on the top level, i.e., where folders like R, src, tests, man, inst, and demo reside. Shouldn't the source directory stay nice-and-clean, without being filled with .o files and example files? Cheers, Marius Attachment: screen shot ___ R-SIG-Mac mailing list r-sig-...@r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-mac __ R-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-devel __ R-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-devel
Re: [Rd] [R-SIG-Mac] check leads to .o files and to packagename-Ex.R
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 06/02/2011 09:12 AM, Duncan Murdoch wrote: > On 11-06-02 9:06 AM, Ben Bolker wrote: >> On 06/02/2011 06:47 AM, Benilton Carvalho wrote: >>> 'R CMD check' should be applied on the .tar.gz, not on the source >>> directory. >> >>Why? > > Because that is what you will send to others. The reason to check a > directory instead of a tar file is to save time: you don't need to > recompile all the files, as a normal check does. > > The help says: "Check R packages from package sources, which can >> be directories or package 'tar' archives with extension '.tar.gz', >> '.tar.bz2' or '.tgz'." I just skimmed through the relevant section >> (1.3.1) in the R Extensions manual, and it doesn't say anything about >> running on tarballs being preferred. >> >>Add my vote to the wishlist that the src directory should get cleaned >> after R CMD check. > > Then the advantage of checking a directory would be lost. > > Duncan Murdoch OK, that makes sense. The logic was not clear to me. Do you know offhand if R CMD check uses "make"-like logic to determine whether the sources should be recompiled even if there are existing .o/.so files, or is that up to the user ... ? Ben > >> >>Ben Bolker >> >>> >>> So, it'd be something like: >>> >>> R CMD build pkg >>> R CMD check pkg_version.tar.gz >>> >>> b >> >> >> >>> >>> On 2 June 2011 06:54, Marius Hofert wrote: Dear expeRts, I work on the R package "nacopula" (https://r-forge.r-project.org/projects/nacopula/) under Mac OS X 10.6.7 (MacBook Pro). The session info is: R version 2.14.0 Under development (unstable) (2011-05-02 r55730) Platform: x86_64-apple-darwin9.8.0/x86_64 (64-bit) When I apply "R CMD check nacopula" my source directory is filled with .o files (which leads to a warning for the next check). Further, if I use R 2.14.0, the check produces a file "nacopula-Ex.R" on the top level, i.e., where folders like R, src, tests, man, inst, and demo reside. Shouldn't the source directory stay nice-and-clean, without being filled with .o files and example files? Cheers, Marius Attachment: screen shot ___ R-SIG-Mac mailing list r-sig-...@r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-mac >>> >>> >>> >> >> __ >> R-devel@r-project.org mailing list >> https://stat.ethz.ch/mailman/listinfo/r-devel > -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.10 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAk3nmqcACgkQc5UpGjwzenOP7wCeJ/qoUed60giuBeBrJm3jiHhm DRUAoIXl8cZE9f3TMpJ909puSgSgE8ZJ =9zaZ -END PGP SIGNATURE- __ R-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-devel
Re: [Rd] [R-SIG-Mac] check leads to .o files and to packagename-Ex.R
On 02/06/2011 10:13 AM, Ben Bolker wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 06/02/2011 09:12 AM, Duncan Murdoch wrote: > On 11-06-02 9:06 AM, Ben Bolker wrote: >> On 06/02/2011 06:47 AM, Benilton Carvalho wrote: >>> 'R CMD check' should be applied on the .tar.gz, not on the source >>> directory. >> >> Why? > > Because that is what you will send to others. The reason to check a > directory instead of a tar file is to save time: you don't need to > recompile all the files, as a normal check does. > >The help says: "Check R packages from package sources, which can >> be directories or package 'tar' archives with extension '.tar.gz', >> '.tar.bz2' or '.tgz'." I just skimmed through the relevant section >> (1.3.1) in the R Extensions manual, and it doesn't say anything about >> running on tarballs being preferred. >> >> Add my vote to the wishlist that the src directory should get cleaned >> after R CMD check. > > Then the advantage of checking a directory would be lost. > > Duncan Murdoch OK, that makes sense. The logic was not clear to me. Do you know offhand if R CMD check uses "make"-like logic to determine whether the sources should be recompiled even if there are existing .o/.so files, or is that up to the user ... ? It uses make, with some cleverly written default rules. For Windows, that's GNU make, but other systems might have a different make involved. Duncan Murdoch __ R-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-devel
Re: [Rd] possibly invalid assertion in setRefClass?
Right, but Michael's point is valid. It's the standard bug from having
to explicitly check for a non-empty list before taking names(x).
Will fix. Thanks for the catch.
John
On 6/1/11 9:21 AM, William Dunlap wrote:
-Original Message-
From: r-devel-boun...@r-project.org
[mailto:r-devel-boun...@r-project.org] On Behalf Of Michael Lawrence
Sent: Wednesday, June 01, 2011 9:15 AM
To: r-devel@r-project.org
Subject: [Rd] possibly invalid assertion in setRefClass?
setRefClass("Foo", fields = list())
Error in setRefClass("Foo", fields = list()) :
A list argument for fields must have nonempty names for all
the fields
In my opinion, the above should not fail. There are no fields.
You can work around it by attaching an empty names attribute
to fields:
> setRefClass("Foo", fields = structure(list(), names=character()))
Generator object for class "Foo":
No fields defined
Class Methods:
"callSuper", "copy", "export", "field", "getClass", "getRefClass",
"import", "initFields"
Reference Superclasses:
"envRefClass"
Bill Dunlap
Spotfire, TIBCO Software
wdunlap tibco.com
Thanks,
Michael
[[alternative HTML version deleted]]
__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
