Build warnings #5

mvuorre · 2016-09-29T18:44:40Z

Currently, bmlm does not pass build cleanly, as detailed in cran-comments.md. I would like to get it to CRAN but am afraid it wont go through with the warnings.

Test environments

local OS X install, R 3.3.1
win-builder (devel and release)

R CMD check results

0 errors | 1 warning | 0 notes

Status: 1 WARNING checking compiled code ... WARNING

File ‘bmlm/libs/bmlm.so’:
Found ‘___assert_rtn’, possibly from ‘assert’ (C)
Object: ‘Modules.o’

Compiled code should not call entry points which might terminate R nor write to stdout/stderr instead of to the console, nor the system RNG.

Win-builder results

Installation time in seconds: 197 Check time in seconds: 108 Status: 1 WARNING

checking whether package 'bmlm' can be installed ... WARNING

Found the following significant warnings:

d:/RCompile/CRANpkg/lib/3.4/BH/include/boost/math/special_functions/detail/bernoulli_details.hpp:127:36: warning: ISO C++ 1998 does not support 'long long' [-Wlong-long]

d:/RCompile/CRANpkg/lib/3.4/BH/include/boost/math/special_functions/detail/bernoulli_details.hpp:151:9: warning: ISO C++ 1998 does not support 'long long' [-Wlong-long]

See 'd:/RCompile/CRANguest/R-devel/bmlm.Rcheck/00install.out' for details.

Kindly asking if @jgabry has insight on these, what I think are C++ and rstanarm related issues.

The text was updated successfully, but these errors were encountered:

jgabry · 2016-09-29T19:06:59Z

Hey Matti, I should be able to take a look tonight or tomorrow and help
you sort it out. And yeah, definitely hold off on submitting to CRAN until
we deal with this.

Jonah

On Thursday, September 29, 2016, Matti Vuorre notifications@github.com
wrote:

Currently, bmlm does not pass build cleanly, as detailed in
cran-comments.md
https://github.com/mvuorre/bmlm/blob/dev/cran-comments.md. I would like
to get it to CRAN but am afraid it wont go through with the warnings.
Test environments

local OS X install, R 3.3.1
win-builder (devel and release)

R CMD check results

0 errors | 1 warning | 0 notes

Status: 1 WARNING checking compiled code ... WARNING

File ‘bmlm/libs/bmlm.so’:
Found ‘___assert_rtn’, possibly from ‘assert’ (C)
Object: ‘Modules.o’

Compiled code should not call entry points which might terminate R nor
write to stdout/stderr instead of to the console, nor the system RNG.
Win-builder results

Installation time in seconds: 197 Check time in seconds: 108 Status: 1
WARNING

checking whether package 'bmlm' can be installed ... WARNING

Found the following significant warnings:

d:/RCompile/CRANpkg/lib/3.4/BH/include/boost/math/special_
functions/detail/bernoulli_details.hpp:127:36: warning: ISO C++ 1998 does
not support 'long long' [-Wlong-long]

d:/RCompile/CRANpkg/lib/3.4/BH/include/boost/math/special_
functions/detail/bernoulli_details.hpp:151:9: warning: ISO C++ 1998 does
not support 'long long' [-Wlong-long]

See 'd:/RCompile/CRANguest/R-devel/bmlm.Rcheck/00install.out' for details.

Kindly asking if @jgabry https://github.com/jgabry has insight on
these, what I think are C++ and rstanarm related issues.

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
#5, or mute the thread
https://github.com/notifications/unsubscribe-auth/AHb4Qz_VPjG1WqSC7oLyvh4X446xkPOUks5qvAcZgaJpZM4KKVrQ
.

jgabry · 2016-09-30T02:45:22Z

I haven't fully sorted it out yet. @bgoodri have you seen this check error
from a package made with rstan.package.skeleton? Looks like maybe something
to do with the shared object but I haven't seen this before.

On another note, unless I missed them the package doesn't seem to have
any unit tests. I would be very hesitant about releasing a package on
CRAN until you have a pretty thorough and public testing framework. I
can't recommend that enough.

Jonah

On Thursday, September 29, 2016, Matti Vuorre notifications@github.com
wrote:

Currently, bmlm does not pass build cleanly, as detailed in
cran-comments.md
https://github.com/mvuorre/bmlm/blob/dev/cran-comments.md. I would like
to get it to CRAN but am afraid it wont go through with the warnings.
Test environments

local OS X install, R 3.3.1
win-builder (devel and release)

R CMD check results

0 errors | 1 warning | 0 notes

Status: 1 WARNING checking compiled code ... WARNING

File ‘bmlm/libs/bmlm.so’:
Found ‘___assert_rtn’, possibly from ‘assert’ (C)
Object: ‘Modules.o’

Compiled code should not call entry points which might terminate R nor
write to stdout/stderr instead of to the console, nor the system RNG.
Win-builder results

Installation time in seconds: 197 Check time in seconds: 108 Status: 1
WARNING

checking whether package 'bmlm' can be installed ... WARNING

Found the following significant warnings:

d:/RCompile/CRANpkg/lib/3.4/BH/include/boost/math/special_
functions/detail/bernoulli_details.hpp:127:36: warning: ISO C++ 1998 does
not support 'long long' [-Wlong-long]

d:/RCompile/CRANpkg/lib/3.4/BH/include/boost/math/special_
functions/detail/bernoulli_details.hpp:151:9: warning: ISO C++ 1998 does
not support 'long long' [-Wlong-long]

See 'd:/RCompile/CRANguest/R-devel/bmlm.Rcheck/00install.out' for details.

Kindly asking if @jgabry https://github.com/jgabry has insight on
these, what I think are C++ and rstanarm related issues.

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
#5, or mute the thread
https://github.com/notifications/unsubscribe-auth/AHb4Qz_VPjG1WqSC7oLyvh4X446xkPOUks5qvAcZgaJpZM4KKVrQ
.

bgoodri · 2016-09-30T03:08:33Z

I have seen the -Wlong-long warning on Windows and we had to put CXX_STD = CXX11 into src/Makevars.win to build it with the C++11 standard. I have not seen the OSX one, but I don't use OSX. What version of clang++ is it and what optimization level?

mvuorre · 2016-09-30T18:01:35Z

Thanks, both of you!

Wlong-long warning with win-builder: I added the src/Makevars.win as in rstanarm (with CXX_STD = CXX11), and I no longer get the warning with win-builder. 👍
clang version on my OS X:

> clang++ --version
Apple LLVM version 7.3.0 (clang-703.0.31)
Target: x86_64-apple-darwin15.6.0
Thread model: posix
InstalledDir: /Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin

I don't know how to answer the question about clang optimization level. The OS X warning might be related to this discussion.

(I am currently writing tests and hoping to add them soon. Thanks for this suggestion!)

mvuorre · 2016-09-30T18:03:31Z

I have reproduced the OS X warning on a newer system (mac OS Sierra)

> clang++ --version
Apple LLVM version 8.0.0 (clang-800.0.38)
Target: x86_64-apple-darwin16.0.0
Thread model: posix
InstalledDir: /Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin

bgoodri · 2016-09-30T22:15:00Z

You have -DBOOST_DISABLE_ASSERTS in the PKG_CPPFLAGS line in the src/Makevars file. @jgabry can you run R CMD check on the dev branch of bmlm?

jgabry · 2016-09-30T23:01:25Z

Yeah I can do it this evening. I'll get back to you.

On Friday, September 30, 2016, bgoodri notifications@github.com wrote:

You have -DBOOST_DISABLE_ASSERTS in the PKG_CPPFLAGS line in the
src/Makevars file. @jgabry https://github.com/jgabry can you run R CMD
check on the dev branch of bmlm?

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
#5 (comment), or mute
the thread
https://github.com/notifications/unsubscribe-auth/AHb4QzKurGGSi-wbRzi1GDNdkIx75d5Fks5qvYnlgaJpZM4KKVrQ
.

jgabry · 2016-10-01T03:58:29Z

@bgoodri for me the dev branch passes R CMD check

bgoodri · 2016-10-01T04:19:15Z

OK. Then I would suspect another copy of Boost is lurking on the original
computer that somehow gets pulled in instead of BH.

On Fri, Sep 30, 2016 at 11:58 PM, Jonah Gabry notifications@github.com
wrote:

@bgoodri https://github.com/bgoodri for me the dev branch passes R CMD
check

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
#5 (comment), or mute
the thread
https://github.com/notifications/unsubscribe-auth/ADOrqr5IL-JHPn9A8whhlWIQTyJG8cykks5qvdplgaJpZM4KKVrQ
.

mvuorre · 2016-10-02T15:56:44Z

OK, it seems like its a local problem on my 2 machines then. I'll investigate and close this issue once I figure out what is causing the warnings.

Thank you both so much!

mvuorre · 2016-10-27T14:28:17Z

I can't replicate the issue on other than my personal computers, so it seems that it's a local problem with C++ / Boost. The issue with the package is therefore resolved.

mvuorre added the help wanted label Sep 29, 2016

mvuorre pushed a commit that referenced this issue Sep 30, 2016

Makevars.win as per #5

de57b0b

mvuorre closed this as completed Oct 27, 2016

mvuorre reopened this Oct 27, 2016

mvuorre closed this as completed Oct 27, 2016

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Build warnings #5

Build warnings #5

mvuorre commented Sep 29, 2016

jgabry commented Sep 29, 2016

jgabry commented Sep 30, 2016

bgoodri commented Sep 30, 2016 via email

mvuorre commented Sep 30, 2016

mvuorre commented Sep 30, 2016

bgoodri commented Sep 30, 2016

jgabry commented Sep 30, 2016

jgabry commented Oct 1, 2016

bgoodri commented Oct 1, 2016

mvuorre commented Oct 2, 2016

mvuorre commented Oct 27, 2016

Build warnings #5

Build warnings #5

Comments

mvuorre commented Sep 29, 2016

Test environments

R CMD check results

Win-builder results

jgabry commented Sep 29, 2016

jgabry commented Sep 30, 2016

bgoodri commented Sep 30, 2016 via email

mvuorre commented Sep 30, 2016

mvuorre commented Sep 30, 2016

bgoodri commented Sep 30, 2016

jgabry commented Sep 30, 2016

jgabry commented Oct 1, 2016

bgoodri commented Oct 1, 2016

mvuorre commented Oct 2, 2016

mvuorre commented Oct 27, 2016