coatless-rpkg / rcppensmallen Goto Github PK

Maybe I don't understand the plugin system, but shouldn't inlineCxxPlugin be exported in order for other packages to use it?

In PR #25's 29ed79b , we lost the ability to fully link to the Ensmallen commits on GitHub due to the NEWS.md file having too many links to pages on GitHub. e.g.

From: NEWS.md
Status: 429
Message: Too Many Requests

From Slack, @eddelbuettel mentioned that Kurt Hornik is working to add a GITHUB_PAT/GITHUB_TOKEN to allow for increased traffic rates to GitHub. Once this goes through, we can restore by running:

# Read in news items
items = readLines("NEWS.md")

# Convert
items = sub(
          "mlpack/ensmallen#([0-9]+)",
          "[#\1](https://github.com/mlpack/ensmallen/pulls/\1)", 
          items
         )

# Write out fixed items
writeLines(items, "NEWS.md")

In all that exitement about GH Actions it seems to have been overlooked that the most current R version is not 3.6 anymore.

• Move package instructions out into separate vignette.
• Rename existing using to be a case study.

Hi, while installing the package I got the following error messages

installing source package ‘RcppEnsmallen’ ...
** package ‘RcppEnsmallen’ successfully unpacked and MD5 sums checked
** using staged installation
** libs
/opt/local/bin/clang++-mp-13 -std=gnu++11 -I"/opt/local/Library/Frameworks/R.framework/Resources/include" -DNDEBUG -I'/Users/gpasa/Library/R/x86_64/4.1/library/Rcpp/include' -I'/Users/gpasa/Library/R/x86_64/4.1/library/RcppArmadillo/include' -I/opt/local/include -isysroot/Library/Developer/CommandLineTools/SDKs/MacOSX12.sdk -I../inst/include -fopenmp -fPIC -pipe -Os -stdlib=libc++ -isysroot/Library/Developer/CommandLineTools/SDKs/MacOSX12.sdk -arch x86_64 -c RcppExports.cpp -o RcppExports.o
In file included from RcppExports.cpp:4:
In file included from ./../inst/include/RcppEnsmallen.h:25:
In file included from ../inst/include/ensmallen.hpp:75:
../inst/include/ensmallen_bits/callbacks/callbacks.hpp:765:26: error: no member named 'HasGenerationalStepTakenSignature' in namespace 'ens::callbacks::traits'
callbacks::traits::HasGenerationalStepTakenSignature<
~~~~~~~~~~~~~~~~~~~^
../inst/include/ensmallen_bits/callbacks/callbacks.hpp:766:7: error: 'CallbackType' does not refer to a value
CallbackType, OptimizerType, FunctionType, MatType, ObjectivesVecType,
^
../inst/include/ensmallen_bits/callbacks/callbacks.hpp:758:21: note: declared here
template<typename CallbackType,
^
../inst/include/ensmallen_bits/callbacks/callbacks.hpp:767:28: error: expected member name or ';' after declaration specifiers
IndicesType>::hasBool, bool>::type
~~~~~~~~~~~~~~~~~~~~~^

Probably unintentional and would-be-nice-if-turned-off, but each time I comment on an issue I get an email with

Run failed for master (e7a2f28)

Repository: coatless/rcppensmallen
Workflow: Commands
Duration: 3.0 seconds
Finished: 2020-01-11 12:33:43 UTC

View results: https://github.com/coatless/rcppensmallen/commit/e7a2f2803f70eb22f08d0f7b2cd578a97c63021a/checks?check_suite_id=395491589

Jobs:

• Surface the optimizers into R.
• Consider implementing our own configure.ac to check for:
  • C++11-compliant compiler
  • Armadillo version > 6.500
  • macOS OpenMP issues

See https://sourceforge.net/p/arma/news/2023/10/ensmallen-220-released---c-library-for-numerical-optimisation/

It looks like you just gave the automated procedure some love last night but it also looks like it didn't catch this.

Link to paper: https://arxiv.org/abs/1810.09361
Citation:

@ARTICLE{2018arXiv181009361B,
   author = {{Bhardwaj}, S. and {Curtin}, R.~R. and {Edel}, M. and {Mentekidis}, Y. and 
	{Sanderson}, C.},
    title = "{ensmallen: a flexible C++ library for efficient function optimization}",
  journal = {ArXiv e-prints},
archivePrefix = "arXiv",
   eprint = {1810.09361},
 primaryClass = "cs.MS",
 keywords = {Computer Science - Mathematical Software, Computer Science - Machine Learning, Mathematics - Optimization and Control, 65K10, 68N99, 68W99, 90C53, G.1.6, G.4},
     year = 2018,
    month = oct,
   adsurl = {http://adsabs.harvard.edu/abs/2018arXiv181009361B},
  adsnote = {Provided by the SAO/NASA Astrophysics Data System}
}

/cc @conradsnicta

Thank you for creating a great package that is easy to use. However, when creating a R package with RcppEnsmallen I am encountering an error that I'm not sure how to address. In the RcppExports.cpp file there are include statements for both Rcpp and RcppArmadillo, which is leading to the error "Do not include either 'Rcpp.h' or 'RcppArmadillo.h'. Please only use 'RcppEnsmallen.h'." Because the RcppExports.cpp file is generated by the Rcpp::compileAttributes() function I am not sure how to correct this. This is only an issue I've encountered when creating a package, not when running individual script files via Rcpp::sourceCpp().

I am a beginner to using Rcpp so there might be something simple that I am missing. I use devtools and RStudio for creating packages, and from my understanding devtools::document(), as well as the RStudio package build/reload tools, automatically calls the Rcpp::compileAttributes() function. My .cpp file containing functions only begins with the following:

#include <RcppEnsmallen.h>
#include <RcppArmadilloExtensions/sample.h>

// [[Rcpp::plugins("cpp11")]]

// [[Rcpp::depends(RcppEnsmallen)]]

using namespace Rcpp;

In my use of RcppEnsmallen by just calling Rcpp::sourceCpp() on scripts rather than developing a package, I do not encounter any errors. I'm not sure what is the proper way to address the RcppExports file however.

Any help would be greatly appreciated, I look forward to continue to use RcppEnsmallen in my research.