"call of overloaded ‘sqrt(int)’ is ambiguous" using g++ 7.2.0 #2478
Comments
|
It is a Stan (Math) issue and has already had several issues filed about it over the years, such as stan-dev/math#712 |
|
I spun up a docker image and tested it with Ubuntu with g++ 7.2.0. It works fine with CmdStan. I'm trying to reproduce the error in RStan, but having trouble installing RStan under that configuration. |
|
if you think I could be of any help (I am a really entry-level STAN user) let me know! |
|
@zottelef, thank you! There is something you could try to help us figure this out. Can you:
What I'm trying to tease apart is whether this is a configuration problem in RStan or CmdStan / Stan. |
|
@bgoodri, this is the result I have from building from CmdStan (tagged version v2.17.1): No compiler error. |
|
@zottelef, please let me know if you run into any trouble getting that up and running. It took me less than 30 minutes, but I know CmdStan inside and out. |
|
Thanks. Yikes. How in the world would R be messing things up? Is it something in the Makevars?
|
|
|
The solution's the same here. Write our own sqrt() and make
sure we call that.
- Bob
… On Feb 28, 2018, at 2:16 PM, bgoodri ***@***.***> wrote:
***@***.***:/opt/cmdstan$ make /tmp/sqrt
--- Translating Stan model to C++ code ---
bin/stanc /tmp/sqrt.stan --o=/tmp/sqrt.hpp
Model name=sqrt_model
Input file=/tmp/sqrt.stan
Output file=/tmp/sqrt.hpp
--- Linking C++ model ---
g++-7 -Wall -I . -isystem stan/lib/stan_math/lib/eigen_3.3.3 -isystem stan/lib/stan_math/lib/boost_1.64.0 -isystem stan/lib/stan_math/lib/cvodes_2.9.0/include -std=c++1y -DBOOST_RESULT_OF_USE_TR1 -DBOOST_NO_DECLTYPE -DBOOST_DISABLE_ASSERTS -DBOOST_PHOENIX_NO_VARIADIC_EXPRESSION -I src -I stan/src -isystem stan/lib/stan_math/ -DEIGEN_NO_DEBUG -DFUSION_MAX_VECTOR_SIZE=12 -DNO_FPRINTF_OUTPUT -pipe -lpthread -O3 -o /tmp/sqrt src/cmdstan/main.cpp -include /tmp/sqrt.hpp stan/lib/stan_math/lib/cvodes_2.9.0/lib/libsundials_nvecserial.a stan/lib/stan_math/lib/cvodes_2.9.0/lib/libsundials_cvodes.a
In file included from <command-line>:0:0:
/tmp/sqrt.hpp: In member function ‘void sqrt_model_namespace::sqrt_model::ctor_body(stan::io::var_context&, unsigned int, std::ostream*)’:
/tmp/sqrt.hpp:125:240: error: call of overloaded ‘sqrt(int)’ is ambiguous
:model::rvalue(qr_Q(x), stan::model::cons_list(stan::model::index_omni(), stan::model::cons_list(stan::model::index_min_max(1, K), stan::model::nil_index_list())), "qr_Q(x)"),sqrt((N - 1))));
^
In file included from /usr/include/features.h:419:0,
from /usr/include/x86_64-linux-gnu/bits/libc-header-start.h:33,
from /usr/include/stdint.h:26,
from /usr/lib/gcc/x86_64-linux-gnu/7/include/stdint.h:9,
from stan/lib/stan_math/stan/math/memory/stack_alloc.hpp:6,
from stan/lib/stan_math/stan/math/rev/core/autodiffstackstorage.hpp:4,
from stan/lib/stan_math/stan/math/rev/core.hpp:4,
from stan/lib/stan_math/stan/math/rev/mat.hpp:4,
from stan/lib/stan_math/stan/math.hpp:4,
from stan/src/stan/model/model_header.hpp:4,
from /tmp/sqrt.hpp:3,
from <command-line>:0:
/usr/include/x86_64-linux-gnu/bits/mathcalls.h:149:1: note: candidate: double sqrt(double)
__MATHCALL (sqrt,, (_Mdouble_ __x));
^
In file included from stan/lib/stan_math/lib/boost_1.64.0/boost/config/no_tr1/cmath.hpp:21:0,
from stan/lib/stan_math/lib/boost_1.64.0/boost/math/tools/config.hpp:19,
from stan/lib/stan_math/stan/math/rev/core/var.hpp:7,
from stan/lib/stan_math/stan/math/rev/core/gevv_vvv_vari.hpp:5,
from stan/lib/stan_math/stan/math/rev/core.hpp:12,
from stan/lib/stan_math/stan/math/rev/mat.hpp:4,
from stan/lib/stan_math/stan/math.hpp:4,
from stan/src/stan/model/model_header.hpp:4,
from /tmp/sqrt.hpp:3,
from <command-line>:0:
/usr/include/c++/7/cmath:463:3: note: candidate: constexpr float std::sqrt(float)
sqrt(float __x)
^~~~
/usr/include/c++/7/cmath:467:3: note: candidate: constexpr long double std::sqrt(long double)
sqrt(long double __x)
^~~~
/usr/include/c++/7/cmath:475:5: note: candidate: constexpr typename __gnu_cxx::__enable_if<std::__is_integer<_Tp>::__value, double>::__type std::sqrt(_Tp) [with _Tp = int; typename __gnu_cxx::__enable_if<std::__is_integer<_Tp>::__value, double>::__type = double]
sqrt(_Tp __x)
^~~~
In file included from stan/lib/stan_math/stan/math/rev/scal/fun/pow.hpp:5:0,
from stan/lib/stan_math/stan/math/rev/scal/fun/inc_beta.hpp:7,
from stan/lib/stan_math/stan/math/rev/scal/fun/grad_inc_beta.hpp:12,
from stan/lib/stan_math/stan/math/rev/scal.hpp:50,
from stan/lib/stan_math/stan/math/rev/arr.hpp:9,
from stan/lib/stan_math/stan/math/rev/mat.hpp:13,
from stan/lib/stan_math/stan/math.hpp:4,
from stan/src/stan/model/model_header.hpp:4,
from /tmp/sqrt.hpp:3,
from <command-line>:0:
stan/lib/stan_math/stan/math/rev/scal/fun/sqrt.hpp:50:16: note: candidate: stan::math::var stan::math::sqrt(const stan::math::var&)
inline var sqrt(const var& a) {
^~~~
In file included from stan/lib/stan_math/stan/math/prim/mat.hpp:209:0,
from stan/lib/stan_math/stan/math/rev/mat.hpp:12,
from stan/lib/stan_math/stan/math.hpp:4,
from stan/src/stan/model/model_header.hpp:4,
from /tmp/sqrt.hpp:3,
from <command-line>:0:
stan/lib/stan_math/stan/math/prim/mat/fun/sqrt.hpp:32:5: note: candidate: typename stan::math::apply_scalar_unary<stan::math::sqrt_fun, T>::return_t stan::math::sqrt(const T&) [with T = int; typename stan::math::apply_scalar_unary<stan::math::sqrt_fun, T>::return_t = double]
sqrt(const T& x) {
^~~~
make: *** [make/models:12: /tmp/sqrt] Error 1
—
You are receiving this because you modified the open/close state.
Reply to this email directly, view it on GitHub, or mute the thread.
|
|
@bgoodri, is that compiler installed on the linux box? If so, I think we can run a test in Stan (compile all models) with Jenkins that will test that things are C++ compilable with that compiler. |
|
(I can't reproduce the failure with g++ 7.2.0.) |
|
Yes
…On Wed, Feb 28, 2018 at 4:02 PM, Daniel Lee ***@***.***> wrote:
@bgoodri <https://github.com/bgoodri>, is that compiler installed on the
linux box?
If so, I think we can run a test in Stan (compile all models) with Jenkins
that will test that things are C++ compilable with that compiler.
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#2478 (comment)>, or mute
the thread
<https://github.com/notifications/unsubscribe-auth/ADOrqiVqlorn9SiB9foEJPc_h1wxV-ydks5tZb7QgaJpZM4SWwLs>
.
|
|
@bgoodri, mind doing me a favor and seeing if this test fails in Stan? (after setting We've written out a lot of Stan programs that we assume to be compilable by the C++ compiler and the two tests in there make sure they're parsed properly and that they compile. If that catches when it doesn't work, we can put that into a Jenkins job and make sure we are able to keep compiling with changes to the code base. |
|
@syclik : sorry for the late answer. I did not notice the advances in the topic. Really sorry about that. |
|
Thank you! |
|
OK, this is the one I meant to reopen, as the problem's in this repo. |
|
Summary: Description: Reproducible Steps: Current Output: and Expected Output: Current Version: |
|
This very likely isnt an issue anymore and any chance of ambiguity in the future will be the handled by stan-dev/math#1692 anyways. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Summary:
A valid Stan program is failing to C++ compile on Linux using g++ 7.2.0.
This is failing from RStan, so we need to see if it's a Stan problem or an RStan problem.
(Originally reported by @zottelef on Discourse: http://discourse.mc-stan.org/t/qr-reparametrization-error/3426/9)
Description:
This program generates C++ code and then fails to compile.
Reproducible Steps:
We need to determine if this is an RStan or a Stan problem. This error was reported using:
Current Output:
The relevant part of the compiler error is:
Here's the rest of the reported compiler error:
Expected Output:
No error.
Current Version:
v2.17.1
The text was updated successfully, but these errors were encountered: