eddelbuettel / bh Goto Github PK
View Code? Open in Web Editor NEWR package providing Boost Header files
R package providing Boost Header files
When I try to install the BH package in R version 3.4.2, I get the following errors:
``
install.packages("BH")
Installing package into ‘C:/NotBackedUp/software/R/3.4’
(as ‘lib’ is unspecified)
trying URL 'https://mirror.aarnet.edu.au/pub/CRAN/bin/windows/contrib/3.4/BH_1.66.0-1.zip'
Content type 'application/zip' length 17880018 bytes (17.1 MB)
downloaded 16.9 MB
Warning in install.packages :
downloaded length 17682432 != reported length 17880018
Warning in install.packages :
error 1 in extracting from zip file
Warning in install.packages :
cannot open compressed file 'BH/DESCRIPTION', probable reason 'No such file or directory'
Error in install.packages : cannot open the connection
Or if I try to download and then install it as a tar.gz file:
``
package.name = "BH"
download.path = "C:/NotBackedUp/Temp"
pfname = download.packages(package.name, download.path)
trying URL 'https://mirror.aarnet.edu.au/pub/CRAN/src/contrib/BH_1.66.0-1.tar.gz'
Content type 'application/x-gzip' length 11583445 bytes (11.0 MB)
downloaded 10.9 MB
Warning message:
In download.file(url, destfile, method, mode = "wb", ...) :
downloaded length 11427840 != reported length 11583445
install.packages(repos=NULL, pkgs=pfname[2])
Installing package into ‘C:/NotBackedUp/software/R/3.4’
(as ‘lib’ is unspecified)
'\svrau100qsm01.oceania.corp.anz.com\PFSMB PBWMg\Deposits\Cash&TDs\Data Science\Reporting\TDs\Tasks'
CMD.EXE was started with the above path as the current directory.
UNC paths are not supported. Defaulting to Windows directory.
Error in untar2(tarfile, files, list, exdir, restore_times) :
incomplete block on file
Warning in install.packages :
running command '"C:/PROGRA1/R/R-341.2/bin/x64/R" CMD INSTALL -l "C:\NotBackedUp\software\R\3.4" "C:/NotBackedUp/Temp/BH_1.66.0-1.tar.gz"' had status 1
Warning in install.packages :
installation of package ‘C:/NotBackedUp/Temp/BH_1.66.0-1.tar.gz’ had non-zero exit status
Or if I try to download and install it as a windows binary file:
``
package.name = "BH"
download.path = "C:/NotBackedUp/Temp"
pfname = download.packages(package.name, download.path, type="win.binary")
trying URL 'https://mirror.aarnet.edu.au/pub/CRAN/bin/windows/contrib/3.4/BH_1.66.0-1.zip'
Content type 'application/zip' length 17880018 bytes (17.1 MB)
downloaded 16.9 MB
Warning message:
In download.file(url, destfile, method, mode = "wb", ...) :
downloaded length 17743872 != reported length 17880018
install.packages(repos=NULL, pkgs=pfname[2])
Installing package into ‘C:/NotBackedUp/software/R/3.4’
(as ‘lib’ is unspecified)
Warning in install.packages :
error 1 in extracting from zip file
Warning in install.packages :
cannot open compressed file 'BH/DESCRIPTION', probable reason 'No such file or directory'
Error in install.packages : cannot open the connection
I've tried the following 2 repositories and get the same error:
Global (CDN) - RStudio
https://cran.rstudio.com/src/contrib/BH_1.66.0-1.tar.gz
Australia (Melbourne 1) [https] - AARNET
https://mirror.aarnet.edu.au/pub/CRAN/bin/windows/contrib/3.4/BH_1.66.0-1.zip
We have a package (rstan package) that currently uses BH for Boost, but is soon going to need MPI for parallelization.
Boost's MPI libraries are not header only---they require the Boost build tools and then linking.
Is there any chance these could be added to BH or should we just go ahead and add the relevant libs ourselves in the stan headers package package?
RcppStream also needs boost/graph/graphviz.hpp
Looking at http://www.boost.org/doc/libs/1_57_0/libs/graph/doc/index.html:
The Boost Graph Library is a header-only library and does not need to be built to be used. The only exceptions are the GraphViz input parser and the GraphML parser.
so we can probably do as we've done in Boost Date.Time and include the headers only.
Hello. This is Jisu KIM, a maintainer of R package TDA
(https://cran.r-project.org/web/packages/TDA/index.html)
I used to hold local copy of boost for my R package, but I decided to use R package BH. Then CRAN check started to complain that my package produces '_ZSt4cout', which is from std::cout.
(e.g. https://www.r-project.org/nosvn/R.check/r-devel-linux-x86_64-debian-gcc/TDA-00check.html)
I checked my source code and verified that I commented out all std::cout from my code. I used to comment out all std::cout from boost code as well. I looked at source code of R package BH, and I found out that many std::cout lines are not commented out.
Is there any ways that my R package ignore those std::cout code in your R package BH? Or do you have any plans to comment out std::cout in next version?
Thanks,
Jisu KIM
When releasing BH 1.75 we had to move to C++14 (#76). This did not create any trouble except for windows oldrel (R 3.6). For what I've seen googlePolylines
(@dcooley), rlas
, lidR
and wellknown
(@sckott) are affected with the same errors which are
BH/include/boost/geometry/index/detail/varray_detail.hpp:131:13: error: 'is_trivially_copyable' is not a member of 'std'
BH/include/boost/geometry/index/detail/varray_detail.hpp:483:46: error: 'is_trivially_constructible' is not a member of 'std'
BH/include/boost/geometry/strategies/compare.hpp:175:63: error: call to non-constexpr function 'const _Tp& std::min(const _Tp&, const _Tp&) [with _Tp = unsigned int]
After spending too much time on it I post my patch here for other maitainers and future maitainers.
The fist two are easy. R 3.6 on windows uses gcc 4.9 and gcc 4.9 has a partial support of C++14. is_trivially_copyable
and is_trivially_constructible
are missing (SO question). Consequently I added substitutes of these properties before to include boost headers
namespace std
{
template <typename T>
class is_trivially_copyable
{
public:
bool operator()() { return value; };
static constexpr bool value = true;
};
template <typename T>
class is_trivially_constructible
{
public:
bool operator()() { return value; };
static constexpr bool value = true;
};
}
Here we can see that I added substitutes that return true
. Well, ok it is bad but anyway if we write bad code with non trivially copyable/constructible objects this will fail on all other compilers that have a full support of C++14
The third issue however was super tricky to me. L170-175 of boost/geometry/strategies/compare.hpp
we can see that we need constexpr const T& min( const T& a, const T& b );
return compare::detail::compare_loop
<
ComparePolicy,
0,
((std::min)(geometry::dimension<Point1>::value,
geometry::dimension<Point2>::value))
>::apply(left, right);
However constexpr min
is missing too in gcc 4.9. There is only const min
. constexpr
is resolved at compile time and there is no way to redefine min
as constexpr
without compilation error (AFAIK). So I copied compare.hpp
in my package and patched it like that
return compare::detail::compare_loop <ComparePolicy, 0, 2>::apply(left, right);
And I included my patched header before all other boost header to avoid inclusion of the actuall boost header.
It is ok to me because I know that my points have 2 dimensions in the code involved so hardcoding it is unsafe but ok.
To finished I wrapped that in precompiler directives to do not mess up everything with recent compilers.
#if defined(__GNUC__) && (__GNUC__ < 5) && !defined(__clang__) && !defined(__llvm__) && !defined(__INTEL_COMPILER)
#include "gcc_4_9_patch/std_is_trivially.hpp"
#include "gcc_4_9_patch/boost_compare_patched.hpp"
#endif
#include <boost/geometry.hpp>
#include <boost/geometry/geometries/geometries.hpp>
I tested this patch with rlas
on CRAN winbuilder oldrel and it passed ok. I now need to do the same with lidR
.
I'm missing on.exit()
in C++ ;-)
Header-only, available in 1.60.0: http://www.boost.org/doc/libs/1_60_0/libs/scope_exit/doc/html/index.html
package failed to extract while trying to download as a dependency to dplyr package on R gui for windows ( 32 bit), R 3.2.0.
following is the error message
utils:::menuInstallPkgs()
trying URL 'http://ftp.iitm.ac.in/cran/bin/windows/contrib/3.2/BH_1.55.0-3.zip'
Content type 'application/zip' length 12472065 bytes (11.9 MB)
downloaded 173 KB
Error in read.dcf(file.path(pkgname, "DESCRIPTION"), c("Package", "Type")) :
cannot open the connection
In addition: Warning messages:
1: In download.file(url, destfile, method, mode = "wb", ...) :
downloaded length 177890 != reported length 12472065
2: In unzip(zipname, exdir = dest) : error 1 in extracting from zip file
3: In read.dcf(file.path(pkgname, "DESCRIPTION"), c("Package", "Type")) :
cannot open compressed file 'BH/DESCRIPTION', probable reason 'No such file or directory'
Hi,
I have noticed that @eddelbuettel is the maintainer for both BH and AsioHeaders.
I am currently working on an R package that uses HTTP and (will use) Websockets. I am planning on switching from my current http library to Boost Beast.
I have managed to get working examples to compile and run using BH and AsioHeaders: https://github.com/nx10/r-boost-beast-example
However, there are some headers missing from Boost and some from Asio which i had to include manually:
Would you consider to add these headers to the BH and AsioHeaders packages?
Or alternatively would you consider creating and maintaining a similar package for Boost Beast (That includes the headers)?
This would ensure that the versions of all headers are compatible.
Boost Beast would offer an alternative to libraries used by for example: https://github.com/rstudio/websocket or https://github.com/rstudio/httpuv
Thank you for your time!
I noticed that in the align
folder, some files are missing compared with the official Boost repository, see
eddelbuettle/bh and boostorg/align.
Is it intentional to reduce package size?
The old tarballs have been removed from the working tree but are still present in the history. They make the repo size much larger than it needs to be, as the tarballs can be downloaded from boost directly.
Stripping them from the history take the bare repo size (e.g. git clone --mirror
) from ~350 MB to 16MB on my machine. This will make cloning this repo much faster!
You can do so by downloading the BFG repo cleaner and running the following commands.
git clone --mirror [email protected]:eddelbuettel/bh.git
java -jar bfg.jar --strip-blobs-bigger-than 10MB bh.git
cd bh.git
git reflog expire --expire=now --all && git gc --prune=now --aggressive
git push --force
Hi,
First of all thank you for this very useful package.
I would like to use it inside a package to read a gzip file.
At the end of the compilation, I get the following error message:
** R
** preparing package for lazy loading
** help
*** installing help indices
** building package indices
** testing if installed package can be loaded
Error in dyn.load(file, DLLpath = DLLpath, ...) :
unable to load shared object '/xxxxxxxxxxxxx/NeutronMultiplicityCounting.so':
/xxxxxxxxxxxxxxxxxxxxx/NeutronMultiplicityCounting/libs/NeutronMultiplicityCounting.so: undefined symbol: _ZN5boost9iostreams4zlib8deflatedE
Error: loading failed
Execution halted
ERROR: loading failed
I believe the error come from the header file <boost/iostreams/filter/gzip.hpp> that I need. It uses the zlib.hpp file that declare the following variable:
BOOST_IOSTREAMS_DECL extern const int deflated;
My DESCRIPTION file declare BH under the LinkingTo section.
I use other functions from BH without any problem.
I can solve the problem by adding the following line to my Makevars file but I guess this is not the right way to do:
PKG_LIBS = -L/usr/lib/x86_64-linux-gnu/ -lboost_iostreams -lboost_filesystem
Wilfried
Is it possible to add Boost Flyweight at some point? The memoisation example looks really attractive for a recursive function I'm trying to write using Rcpp.
sorry, probably i'm just confused, but why is property_tree/json_parser.hpp not part of BH?
is it a non-header-only? or is there any other reason?
(i'm definitely confused as my local update of RcppShark compiles well, but build_win cannot find the above file. )
Would it be possible to add the boost.compute headers to the BH package? These headers would be very helpful for GPU development and seems only right that they should be part of the BH package? Although I am sure you could find them, the github page is https://github.com/boostorg/compute
I'm not sure what you policy is exactly on adding additional boost components but it couldn't hurt to ask.
Hello there!
I'm working on a R package that uses boost-process to pipe/ spawn/ stop processes. It's nice since this works across different platforms. It's header-only, so would you consider including it in BH?
I am trying to install BH, but it is taking too much time to install even after successful donwload.
I saw the below message,
But after that it has been trying to install it for a very very long time. :-(. Is there a way to speed up ?
Requested by @thk686 over private email which is the least preferred way.
Appears to be needed for g++ 4.6.3 on Windows and Boost Fusion code
So we can import it, and suppress the R CRAN check
NOTE:
* checking dependencies in R code ... NOTE
Namespace in Imports field not imported from: ‘BH’
All declared Imports should be used.
See the information on DESCRIPTION files in the chapter ‘Creating R
packages’ of the ‘Writing R Extensions’ manual.
The README.md
file for this project says "Not that this can be used solely by headers-only Boost libraries. This covers most of Boost, but excludes some libraries which require linking for parts or all of their functionality."
Supposed to be "Note that", I assume?
FWIW, I've decided that whenever I use "note that " in my own writing, I can usually replace it with the empty string. =)
Linking against the BH package installed from CRAN for building GPU/CUDA R packages fails with the following error:
/usr/lib64/R/library/BH/include/boost/config/suffix.hpp(510): error: identifier "__float128" is undefined
This is due to a bug in current version of Boost (1.60) used in BH package in CRAN.
https://svn.boost.org/trac/boost/ticket/11852
Can the current version of BH in CRAN based on Boost 1.60 be patched or bumped to the next version?
Hi!
I realize you are reluctant to modify boost source / backporting things for stability reasons, but I still wonder if it's possible to backport this single line fix to silence compile warnings like this:
In file included from /usr/lib/R/site-library/BH/include/boost/random/detail/integer_log2.hpp:19:0,
from /usr/lib/R/site-library/BH/include/boost/random/detail/large_arithmetic.hpp:19,
from /usr/lib/R/site-library/BH/include/boost/random/detail/const_mod.hpp:23,
from /usr/lib/R/site-library/BH/include/boost/random/linear_congruential.hpp:30,
from /usr/lib/R/site-library/BH/include/boost/random/additive_combine.hpp:27,
from /usr/lib/R/site-library/rstan/include/rstan/stan_fit.hpp:15,
from /usr/lib/R/site-library/rstan/include/rstan/rstaninc.hpp:3,
from stan_files/ctsmgen.hpp:18,
from stan_files/ctsmgen.cc:3:
/usr/lib/R/site-library/BH/include/boost/pending/integer_log2.hpp:7:89: note: #pragma message: This header is deprecated. Use <boost/integer/integer_log2.hpp> instead.
This is needed by RcppStreams
First of all thank you for all your contributions especially for the Rcpp, RcppArmadillo and BH packages.
A few months back, I created a text processing package textTinyR using the boost library. Although I faced some problems due to the boost-locale, I partly overcame those by not using locale on Windows (I also came across a similar issue) and by adding a configure file for linux and macintosh (on macintosh it worked only for the latest release, Sierra). Last week I submitted an updated version to CRAN and I realized that it doesn't work for macintosh any more.
I wonder if it's possible to add full support for boost-locale in the BH package? In such a case I would restructure my package accordingly, so that it works on all OS's.
Which appears to be not a lot:
Running this on a Windows 7. Tried to download via CRAN and got the following error:
Installing package into ‘C:/Users/jreid/Documents/R/win-library/3.2’
(as ‘lib’ is unspecified)
trying URL 'https://cran.rstudio.com/bin/windows/contrib/3.2/BH_1.58.0-1.zip'
Content type 'application/zip' length 13846694 bytes (13.2 MB)
downloaded 13.1 MB
Warning in install.packages :
downloaded length 13713408 != reported length 13846694
Warning in install.packages :
error 1 in extracting from zip file
Warning in install.packages :
cannot open compressed file 'BH/DESCRIPTION', probable reason 'No such file or directory'
Error in install.packages : cannot open the connection
Hi.
I can't found some boost headers in this package. Please tell me how can I use a missing headers in my package?
Now I install the boost headers in my OS. But I look for a way to do it without external system wide dependencies.
For instance, here:
> library(simmer)
/.../R/test-3.4/BH/include/boost/container/detail/pair.hpp:85:30: runtime error: reference binding to null pointer of type 'const struct piecewise_construct_t'
/.../R/test-3.4/BH/include/boost/container/detail/pair.hpp:85:30: runtime error: reference binding to null pointer of type 'const struct piecewise_construct_t'
/.../R/test-3.4/BH/include/boost/container/detail/pair.hpp:85:30: runtime error: reference binding to null pointer of type 'const struct piecewise_construct_t'
/.../R/test-3.4/BH/include/boost/container/detail/pair.hpp:85:30: runtime error: reference binding to null pointer of type 'const struct piecewise_construct_t'
Solved since v1.63.0. It is a very minor issue, because the offending pointer is never used (see ticket 12577), so no hurries at all. I just want to put this on the record, just in case someone struggles with the same issue.
When compiling with gcc-8 (currently being deployed on CRAN's Fedora machines) many packages that include BH see this warning:
warning: 'template<class> class std::auto_ptr' is deprecated [-Wdeprecated-declarations]
It looks like CRAN will start flagging these warnings in CMD check, so they will have to be fixed.
Some packages that have this include anytime, dplyr, later, svglite, purrrlyr, highlight, ldat, readr, and a bunch more.
In DESCRIPTION, README.md and maybe also the manual pages
I also posted an issue on https://stackoverflow.com/questions/59572516/org-renjin-cran-libraries-missing-from-repositories inquiring if anyone has run into this from the org.renjin.cran standpoint.
It seems that these dependencies were removed from repositories. Note following the above error that https://repository.mulesoft.org/nexus/content/repositories/public/org/renjin/cran/ no longer contains a BH folder, and https://nexus.bedatadriven.com/content/groups/public/org/renjin/cran/BH/ contains only signatures of files, but no actual library .pom files.
This is weird because http://packages.renjin.org/package/org.renjin.cran/BH/1.69.0-1 specifically claims that those packages exist.
I have been unable to find any information about planned changes to this.
I tried to use boost::filesystem
, which is listed as available in BH, but it seems that it requires linking to boost::system
:
Rcpp::sourceCpp(code='
#include <Rcpp.h>
// Rcpp::depends("BH")
#include <boost/filesystem.hpp>
// [[Rcpp::export]]
int hello() { return 0; }
')
gives undefined symbol: boost::system::generic_category()
.
Would it be possible to add the Boost Mp11 library to BH? I believe it‘s header-only.
Dear Mr. Eddelbuettel:
Thank you for updating great libraries(Rcpp, RInside, RQuantLib etc).
I could not find Boost::Geometry in BH when I tried to use it.
I can easily imagine that you did not add it for its simplicity.
Can I ask you that you have any plan to add "Geometry" library in BH in the near future?
I'm currently using boost/polygon/voronoi.hpp
to make a fast Delaunay triangulation of large point cloud. Everything was good and I was very happy until I tested the package on CRAN. It does not compile because fatal error: boost/polygon/voronoi.hpp: No such file or directory
I was perplex until I found that on my machine (linux) all the boost header are installed but some are missing in BH
. This is why it compiles on my machine but not on CRAN.
I understand that some boost
stuff are not in BH
but it seems to me that the voronoi
stuff is header only. Actually only 3 voronoi_*.hpp
headers are missing but most of them are actually present in BH
.
So I'm wondering why they are missing? if they can be added in BH
? and if it is a viable option to include these files in my own package as a workaround.
Add Boost Multiprecision as per rcpp-devel today.
Dear Dirk -
Thank you for providing this great package bh to allow us efficiently using boost in Rcpp. It seems that BH only includes some of the libraries.
Are we going to include odeint which is under boost/numeric/odeint to solve differential equations.
We are working on drug developments, and use lots of differential equations to model the dynamic process. It will be GREAT that odeint can be included in the future version.
PS: Odeint is a header only C++ library. If this can be included, I would be more than happy to contribute an example for your review.
Best,
Haoda Fu
Would it be possible to include a single R function like "bh_version" or something so that we can put the package into the "Include:" directive in the DESCRIPTION file? This is to work around CRAN checks. There are cases where BH is needed for compiling test code and it seems to work if BH is Included, but not always if it is in Suggests.
Package BH has been updated to 1.75.0 (jumping from 1.72.0) and a release candidate is available for installation (see the brief tweet from yesterday, the command is a simple install.packages("BH", repos="https://ghrr.github.io/drat")
.
A reverse-dependency check (result summary is here) shows that a handful of packages need (often simple) changes:
#define
(and @rcurtin is already aware)I will file courtesy issue on the simple C++14 compilation change which I have verified in all three cases. I can also make PRs and possibly test that a change to C++14 does affect builds under current BH which I have not yet done. Both TDA and wellknown are still 'open' and I would very much appreciate it if the maintainers looked into it as well.
CRAN will close for the winter break on December 18 so it is unlikely we will get all this sorted out by then. CRAN reopens on January 4 and it would be fantastic if all packages could be updated by then allowing BH itself to migrate without breakage.
Please reach out (here or in email) if anything is unclear. I have access to one dedicated (if old/slow) machine where I can test this.
Hi Dirk,
This is David Silkworth from years back.
I am porting a significant enhancement to my FaultTree package and have run into the following FATAL ERROR.
bdd.cc:23:49: fatal error: boost/multiprecision/miller_rabin.hpp: No such file or directory
I have other issues, like awaiting C++14 enabled compiler in Rtools for Windows, but I would appreciate if this could be put on the list. I have confirmed my error goes away if I simply place this header into my local library for BH, but that wouldn't do for package release.
Hi, I was trying to install the library this way:
library('devtools') install_github("eddelbuettel/bh")
but got the following error:
Error : /tmp/Rtmp1M5W5j/devtools10b83412e428/eddelbuettel-bh-ea401b9/man/BH-package.Rd: Sections \title, and \name must exist and be unique in Rd files ERROR: installing Rd objects failed for package ‘BH’ removing ‘/home/federico/R/x86_64-pc-linux-gnu-library/3.2/BH’ Installation failed: Command failed (1)
R version 3.2.2, Ubuntu 16.04
I first got the error trying to install sqldf, so decided to try from github
For example, in numeric/odeint.hpp
:
bh/inst/include/boost/numeric/odeint.hpp
Line 31 in f84296a
...but after path length adjustment, the actual file is numeric/odeint/stepper/runge_kutta_cash_karp54_cl.hpp
: https://github.com/eddelbuettel/bh/blob/master/inst/include/boost/numeric/odeint/stepper/runge_kutta_cash_karp54_cl.hpp
There may be other examples, but this is the first one that I ran into when compiling.
Hi,
Im running R with these settings using Windows7 64 bit through Rstudio Version 1.0.143 :
R version 3.4.0 (2017-04-21)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows 7 x64 (build 7601) Service Pack 1
I tried installing BH - and it failed with the error: expected download length != download length
[as a side note: For whatever reason after that failure that seems to have messed up my default connection with CRAN so further install.packages()
returned an error of : missing where expected TRUE or FALSE ... and only after restarting my Windows session did that go away.]
I tried manually downloading the default current version download from CRAN and wasn't able to download it (error...). { r-release: BH_1.62.0-1.zip}
In the end I downloaded the older release and that solved my problems. {r-oldrel: BH_1.62.0-1.zip}
I see they are the same version number so I dont know what the difference is, I can only guess my mirror version of that file is somehow not good...
Some packages on CRAN that use {BH} (e.g. {bigmemory}) get some warnings for the devel version on windows (see these CRAN checks). Some packages get rejected by CRAN because of this.
Found the following significant warnings:
d:/RCompile/CRANpkg/lib/3.5/BH/include/boost/interprocess/detail/win32_api.hpp:145:9: warning: ISO C++ prohibits anonymous structs [-Wpedantic]
d:/RCompile/CRANpkg/lib/3.5/BH/include/boost/interprocess/detail/win32_api.hpp:153:9: warning: ISO C++ prohibits anonymous structs [-Wpedantic]
d:/RCompile/CRANpkg/lib/3.5/BH/include/boost/interprocess/detail/win32_api.hpp:178:13: warning: ISO C++ prohibits anonymous structs [-Wpedantic]
d:/RCompile/CRANpkg/lib/3.5/BH/include/boost/interprocess/detail/win32_api.hpp:180:7: warning: ISO C++ prohibits anonymous structs [-Wpedantic]
In a recent post (http://dirk.eddelbuettel.com/blog/2018/02/13/#bh_1.66.0-1), you seem to talk about this problem, but I'm not sure how to fix this. Can you please help us figure out how to fix these warnings?
According to the BSL-1.0, a license file needs to be shipped with the software but no such file appears in this repo. Can you please add one?
in DESCRIPTION
and/or README.md
Current installation is partial, and e.g. RcppQuantuccia fails on Windows for lack of min.h:
*** arch - i386
make[1]: Entering directory `/cygdrive/d/temp/RtmpYZm7Z4/R.INSTALL2cd94790a7dcf/RcppQuantuccia/src-i386'
d:/Compiler/gcc-4.9.3/mingw_32/bin/g++ -std=gnu++11 -I"D:/RCompile/recent/R/include" -DNDEBUG -I../inst/include/ -I. -DBOOST_NO_AUTO_PTR -I'D:/RCompile/CRANpkg/lib/4.0/Rcpp/include' -I'D:/RCompile/CRANpkg/lib/4.0/BH/include' -I"d:/Compiler/gcc-4.9.3/local330/include" -O2 -Wall -mtune=core2 -c RcppExports.cpp -o RcppExports.o
In file included from ../inst/include/ql/math/statistics/all.hpp:1:0,
from ../inst/include/ql/math/all.hpp:21,
from ../inst/include/ql/quantlib.hpp:14,
from ../inst/include/RcppQuantuccia_types.h:6,
from ../inst/include/RcppQuantuccia_RcppExports.h:7,
from ../inst/include/RcppQuantuccia.h:7,
from RcppExports.cpp:4:
../inst/include/ql/math/statistics/incrementalstatistics.hpp:42:49: fatal error: boost/accumulators/statistics/min.hpp: No such file or directory
#include <boost/accumulators/statistics/min.hpp>
^
compilation terminated.
make[1]: *** [RcppExports.o] Error 1
make[1]: Leaving directory `/cygdrive/d/temp/RtmpYZm7Z4/R.INSTALL2cd94790a7dcf/RcppQuantuccia/src-i386'
ERROR: compilation failed for package 'RcppQuantuccia'
* removing 'd:/Rcompile/CRANpkg/lib/4.0/RcppQuantuccia'
I've prepared this script to remove and purge unwanted files from git history. You can use it to remove those big tar.gz as follows:
./git-remove.sh [email protected]:eddelbuettel/bh.git tar.gz
You'll be asked for confirmation before removing anything, and then, if everything went ok, changes will be automatically pushed. I've already tried it with a fork (check it) and it worked nicely (310.61 MiB -> 16.49 MiB).
Hello there,
I am experiencing strange error messages when I try and install the BH package.
Ubuntu 16.04.2 LTS
R version 3.4.0 (2017-04-21) -- "You Stupid Darkness"
Steps to reproduce:
library(devtools)
install_version("BH", version = "1.62.0-1")
This gives me the following:
trying URL 'https://mirrors.ebi.ac.uk/CRAN/src/contrib/BH_1.62.0-1.tar.gz'
Content type 'application/x-gzip' length 10181096 bytes (9.7 MB)
==================================================
downloaded 9.7 MB
* installing *source* package ‘BH’ ...
** package ‘BH’ successfully unpacked and MD5 sums checked
** inst
Warning in file.append(to[okay], from[okay]) :
write error during file append
Warning in file.append(to[okay], from[okay]) :
write error during file append
Warning in file.append(to[okay], from[okay]) :
write error during file append
...
<redacted-for-brevity>
...
Warning in file.append(to[okay], from[okay]) :
write error during file append
Warning in file.append(to[okay], from[okay]) :
write error during file append
** help
Error : /tmp/RtmpN1qAz1/R.INSTALL47aa53b0cfe7/BH/man/BH-package.Rd: Sections \title, and \name must exist and be unique in Rd files
ERROR: installing Rd objects failed for package ‘BH’
Any ideas?
Thanks!
Background:
Not really an issue but I have been developing standalone "software" using shiny and a not-so-insignificant number of packages (including shiny) have BH somewhere in their dependency tree. This causes significant "bloat" to the created software and installation time. In an impatient world, users can feel like installers are broken when they take a long time.
Question:
First off- I am totally against functions/packages that modify another package. However, I am hoping there might be some way to programmatically-prune BH to only what a calling-package actually requires. There was some talk of this here: stan-dev/math#545
Since I'm creating GUI-interface software that users are not developing with then I think this would be OK.
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.