Errors using CppAD with Eigen 3.3.3 about cppad

[bradbell]

I have verified this problem. In addition, I found the following code in eigen-3.3.3
Eigen/src/Core/IO.h

// NOTE: This helper is kept for backward compatibility with previous code specializing
// this internal::significant_decimals_impl structure. In the future we should directly
// call digits10() which has been introduced in July 2016 in 3.3.
template
struct significant_decimals_impl
{
static inline int run()
{
return NumTraits::digits10();
}
};

[bradbell]

The cppad/example/cppad_eigen.hpp file has been converted to
the newer version of the eigen requirements for custom scalars;
see the following commits:
a6a6965
e7a2df5
5f5465f
For more details, see the heading 05-12 in the file
https://github.com/coin-or/CppAD/blob/master/omh/whats_new/whats_new_17.omh

Please run your tests using the current master branch to see if you get any
unexpected results. If not, please close this issue.

[braunm]

I am now getting the following error after >make check example:

[ 87%] Building CXX object test_more/CMakeFiles/test_more.dir/mul_cskip.cpp.o
[ 87%] Building CXX object test_more/CMakeFiles/test_more.dir/mul_eq.cpp.o
[ 87%] Building CXX object test_more/CMakeFiles/test_more.dir/mul_level.cpp.o
[ 88%] Building CXX object test_more/CMakeFiles/test_more.dir/mul_cond_rev.cpp.o
[ 88%] Building CXX object test_more/CMakeFiles/test_more.dir/mul_zdouble.cpp.o
In file included from /Users/braunm/cppad/test_more/mul_zdouble.cpp:15:
In file included from /Users/braunm/cppad/cppad/cppad.hpp:22:
In file included from /Users/braunm/cppad/cppad/base_require.hpp:174:
/Users/braunm/cppad/cppad/core/numeric_limits.hpp:210:52: error: no member named 'digits10' in 'CppAD::numeric_limits<CppAD::zdouble>'
        static const int digits10 = numeric_limits<Base>::digits10;
                                    ~~~~~~~~~~~~~~~~~~~~~~^
/Users/braunm/cppad/cppad/core/forward.hpp:136:32: note: in instantiation of template class 'CppAD::numeric_limits<CppAD::AD<CppAD::zdouble> >'
      requested here
                        taylor_[C * j + k] = CppAD::numeric_limits<Base>::quiet_NaN();
                                                    ^
/Users/braunm/cppad/test_more/mul_zdouble.cpp:47:7: note: in instantiation of function template specialization 'CppAD::ADFun<CppAD::AD<CppAD::zdouble>
      >::Forward<CppAD::vector<CppAD::AD<CppAD::zdouble> > >' requested here
                a1f.Forward(0, a1x);
                    ^
In file included from /Users/braunm/cppad/test_more/mul_zdouble.cpp:15:
In file included from /Users/braunm/cppad/cppad/cppad.hpp:63:
In file included from /Users/braunm/cppad/cppad/core/ad_fun.hpp:701:
In file included from /Users/braunm/cppad/cppad/core/fun_eval.hpp:15:
/Users/braunm/cppad/cppad/core/forward.hpp:136:52: error: incomplete definition of type 'CppAD::numeric_limits<CppAD::AD<CppAD::zdouble> >'
                        taylor_[C * j + k] = CppAD::numeric_limits<Base>::quiet_NaN();
                                             ~~~~~~~~~~~~~~~~~~~~~~~~~~~^~
/Users/braunm/cppad/test_more/mul_zdouble.cpp:47:7: note: in instantiation of function template specialization 'CppAD::ADFun<CppAD::AD<CppAD::zdouble>
      >::Forward<CppAD::vector<CppAD::AD<CppAD::zdouble> > >' requested here
                a1f.Forward(0, a1x);
                    ^
2 errors generated.
make[3]: *** [test_more/CMakeFiles/test_more.dir/mul_zdouble.cpp.o] Error 1
make[2]: *** [test_more/CMakeFiles/test_more.dir/all] Error 2
make[1]: *** [CMakeFiles/check.dir/rule] Error 2
make: *** [check] Error 2

[bradbell]

The member variable digits10 was missing from the zdouble numeric limits. I am not sure why mul_zdouble.cpp passed its test on my system. I added a test of all the zdouble numeric limits and I think that I have fixed the problem with commit
cf241d4

[bradbell]

I am closing this issue because I do not know of any outstanding problems using eigen-3.3.3.