Request for guidance on compiling C++ using RInside within Xcode 5, OSX 10.9.4, R 3.1.1 (mavericks version)
To be precise this seems like a bug in Rcpp/RInside - whichever is adding the -I../Rcpp include. Since Rcpp contains String.h it will break everything due to the conflict with standard string.h (HFS is case-insensitive by default). Cheers, Simon
On Jul 28, 2014, at 2:13 AM, Kevin Ushey <kevinushey at gmail.com> wrote:
Hi Roger, Did you see this StackOverflow post? http://stackoverflow.com/questions/15489065/rcppeclipse-on-mac-os-x/17054405 It sounds like the poster describes the same problem as what you're seeing, and various prescriptions are offered. Cheers, Kevin On Fri, Jul 25, 2014 at 6:44 AM, Roger Dean <Roger.Dean at uws.edu.au> wrote:
Request for guidance on compiling C++ using RInside within Xcode 5, OSX
10.9.4, R 3.1.1 (mavericks version).
Hi R-SIG-mac experts,
I know this question is close to some asked previously, and also to
comment 2.10 of the FAQ document on RInside provided by Dirk Eddelbuettel
and Romain Francois, but I still cannot solve it, nor find examples of
others doing so that are clear (or at least, clear to me, apologies if
appropriate).
I would like to compile C++ code that uses RInside, but within Xcode 5,
and using OSX 10.9.4, R 3.1.1 (mavericks version). The reason I wish to do
this is that I am writing time series analysis/generative externals for
MAXMSP, a widely used MIDI and audio processing platform, and it is built
on C and focused on Xcode as the developer tool (it is difficult to find
people compiling MAX externals using clang++ and not Xcode, though I
continue to search also towards this possible alternative route.)
So the helpful examples in the RInside source include rinside_sample0.cpp,
and this compiles/builds fine for me with the Make process, or with clang++
directly. And it runs, as do all but one of the other examples. It is very
simple as intended, and I'm sure many of you will have looked at it:
// -*- mode: C++; c-indent-level: 4; c-basic-offset: 4; tab-width: 8; -*-
//
// Simple example showing how to do the standard 'hello, world' using
embedded R
//
// Copyright (C) 2009 Dirk Eddelbuettel
// Copyright (C) 2010 Dirk Eddelbuettel and Romain Francois
//
// GPL'ed
#include <RInside.h> // for the embedded R via RInside
int main(int argc, char *argv[]) {
RInside R(argc, argv); // create an embedded R instance
R["txt"] = "Hello, world!\n"; // assign a char* (string) to 'txt'
R.parseEvalQ("cat(txt)"); // eval the init string, ignoring
any returns
exit(0);
}
I have taken note of the comments about the .R/Makevars, including the
statements about -stdlib=libstdc++ -mmacosx-version-min=10.6 from Sturla
Molden (Feb 8 2014) and others. I am providing to Xcode libR.dylib, and
paths to headers and other libraries. I saw a reference to an Rcpp library,
but in my RCpp source and R 3.1.1 there is only Rcpp.so, not .dylib. I have
added this to the Xcode project, though I am not clear that it will be used.
After innumerable permutations in my attempts, I can only get as far as
reproducible problems with the string.h code from the Rcpp source (see
listing below taken directly from the Xcode 'issues' listing). I realise
there is a historic issue about the LLVM default compiler (now labelled
Apple LLVM 5.0), but I hoped that was what Makevars could solve.
Can anyone advise whether it is currently feasible to do what I want, and
ideally give more specific cues, please?
many thanks in advance
Roger Dean
roger.dean at uws.edu.au<mailto:roger.dean at uws.edu.au>
Reported Issues:
/Library/Frameworks/R.framework/Versions/3.1/Resources/library/Rcpp/include/Rcpp/String.h
/Library/Frameworks/R.framework/Versions/3.1/Resources/library/Rcpp/include/Rcpp/String.h:52:17:
Use of undeclared identifier 'internal'
/Library/Frameworks/R.framework/Versions/3.1/Resources/library/Rcpp/include/Rcpp/String.h:52:39:
Expected member name or ';' after declaration specifiers
/Library/Frameworks/R.framework/Versions/3.1/Resources/library/Rcpp/include/Rcpp/String.h:53:17:
Use of undeclared identifier 'internal'
/Library/Frameworks/R.framework/Versions/3.1/Resources/library/Rcpp/include/Rcpp/String.h:53:45:
Expected member name or ';' after declaration specifiers
/Library/Frameworks/R.framework/Versions/3.1/Resources/library/Rcpp/include/Rcpp/String.h:66:16:
Unknown type name 'SEXP'
/Library/Frameworks/R.framework/Versions/3.1/Resources/library/Rcpp/include/Rcpp/String.h:71:23:
Unknown type name 'StringProxy'
/Library/Frameworks/R.framework/Versions/3.1/Resources/library/Rcpp/include/Rcpp/String.h:75:23:
Unknown type name 'const_StringProxy'
/Library/Frameworks/R.framework/Versions/3.1/Resources/library/Rcpp/include/Rcpp/String.h:102:17:
Unknown type name 'Rcomplex'; did you mean '_Complex'?
/Library/Frameworks/R.framework/Versions/3.1/Resources/library/Rcpp/include/Rcpp/String.h:103:17:
Unknown type name 'Rbyte'
/Library/Frameworks/R.framework/Versions/3.1/Resources/library/Rcpp/include/Rcpp/String.h:108:35:
Unknown type name 'Rbyte'
/Library/Frameworks/R.framework/Versions/3.1/Resources/library/Rcpp/include/Rcpp/String.h:110:35:
Unknown type name 'Rcomplex'; did you mean '_Complex'?
/Library/Frameworks/R.framework/Versions/3.1/Resources/library/Rcpp/include/Rcpp/String.h:111:35:
Unknown type name 'SEXP'
/Library/Frameworks/R.framework/Versions/3.1/Resources/library/Rcpp/include/Rcpp/String.h:112:41:
Unknown type name 'StringProxy'
/Library/Frameworks/R.framework/Versions/3.1/Resources/library/Rcpp/include/Rcpp/String.h:170:42:
Unknown type name 'StringProxy'
/Library/Frameworks/R.framework/Versions/3.1/Resources/library/Rcpp/include/Rcpp/String.h:178:36:
Unknown type name 'SEXP'
/Library/Frameworks/R.framework/Versions/3.1/Resources/library/Rcpp/include/Rcpp/String.h:313:16:
Unknown type name 'SEXP'
/Library/Frameworks/R.framework/Versions/3.1/Resources/library/Rcpp/include/Rcpp/String.h:318:16:
Unknown type name 'SEXP'
/Library/Frameworks/R.framework/Versions/3.1/Resources/library/Rcpp/include/Rcpp/String.h:355:9:
Unknown type name 'SEXP'
/Library/Frameworks/R.framework/Versions/3.1/Resources/library/Rcpp/include/Rcpp/String.h:358:21:
Implicit instantiation of undefined template 'std::basic_string<char,
std::char_traits<char>, std::allocator<char> >'
Too many errors emitted, stopping now
Roger Dean, Founder and Artistic Director, austraLYSIS. Mobile: +61 (0)481
309 612.
Mail: austraLYSIS, PO Box 225, Milperra, NSW 2214, Australia.
Email<mailto:dr.metagroove at mindless.com> ... Web<
http://www.australysis.com/> ? Recent CD release: MultiPiano (1978-2012),
solo and computer interactive piano, Tall Poppies TP 225 (double CD). John
Shand's review: 'a trail-blazer', 'exquisite', 'surprising and
disquieting', 'willing grooves' ...
Prof Roger Dean,PhD, DSc, DLitt, FAHA, Research Professor of Sonic
Communication, Music, Cognition and Action Group, MARCS Institute.
Mail: MARCS Institute, University of Western Sydney, Locked Bag 1797,
PENRITH, NSW 2751, Australia.
Emai<mailto:roger.dean at uws.edu.au>l ? Phone: +61 2 9772 6902 ... Fax: +61
2 9772 6040
We are in Milperra ('Bankstown Campus'), just by the M5 motorway. Check
the campus maps to visit.
Web<http://marcs.uws.edu.au> ... Articles list<
http://tinyurl.com/RTDeanCitations> ... Wikipedia<
http://en.wikipedia.org/wiki/Roger_Dean_(musician)> ? I am founding
co-editor of the sound intermedia journal soundsRite<
http://soundsite.uws.edu.au> and currently co-editing the forthcoming
Oxford Handbook of Algorithmic Music.
I try to minimise 'Email Distraction', in order to get on with research
and creative work. So I reply to all appropriate emails, but in batches. If
yours needs a response, please expect to get one normally within 2-4
working days.
[[alternative HTML version deleted]]
_______________________________________________ R-SIG-Mac mailing list R-SIG-Mac at r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-mac
[[alternative HTML version deleted]]
_______________________________________________ R-SIG-Mac mailing list R-SIG-Mac at r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-mac