There is a bug in the current version of the bindings script.  See line 63 of the lapack_generator.py, which doesn&#39;t close the &quot;&gt;&quot; for the static_assert include.  I don&#39;t have the LAPACK source, so can&#39;t rerun the python script on my own.  I am using Intel C++ 11.0 on windows.<div>
<br></div><div>When I:</div><div>1) Copied over the current sandbox lapack and blas bindings</div><div>2) Merged the numeric bindings traits with an older copy of the traits I had (because the sandbox copy is missing a lot of them, this is my likely problem)</div>
<div>3) Make the text replace to the static_assert  to close it properly</div><div><br></div><div>Then, I get a failure on the matrix traits (perhaps because I have an older version):</div><div><div>1&gt;C:\working\libraries\boost\boost/numeric/bindings/lapack/driver/gesv.hpp(68): error: too few arguments for class template &quot;boost::is_same&quot;</div>
<div>1&gt;          BOOST_STATIC_ASSERT( boost::is_same&lt; typename traits::matrix_traits&lt;</div><div>1&gt;          ^</div><div>1&gt;          detected during:</div><div>1&gt;            instantiation of &quot;void boost::numeric::bindings::lapack::gesv_impl&lt;ValueType&gt;::compute(MatrixA &amp;, VectorIPIV &amp;, MatrixB &amp;, integer_t={int} &amp;) [with ValueType=double, MatrixA=boost::numeric::ublas::matrix&lt;double, boost::numeric::ublas::column_major, boost::numeric::ublas::unbounded_array&lt;double, std::allocator&lt;QuantLib::Real={double}&gt;&gt;&gt;, VectorIPIV=std::vector&lt;ptrdiff_t={int}, std::allocator&lt;int&gt;&gt;, MatrixB=boost::numeric::ublas::matrix&lt;double, boost::numeric::ublas::column_major,</div>
<div>1&gt;                      boost::numeric::ublas::unbounded_array&lt;double, std::allocator&lt;QuantLib::Real={double}&gt;&gt;&gt;]&quot; at line 93</div><div>1&gt;            instantiation of &quot;integer_t={int} boost::numeric::bindings::lapack::gesv(MatrixA &amp;, VectorIPIV &amp;, MatrixB &amp;) [with MatrixA=boost::numeric::ublas::matrix&lt;double, boost::numeric::ublas::column_major, boost::numeric::ublas::unbounded_array&lt;double, std::allocator&lt;QuantLib::Real={double}&gt;&gt;&gt;, VectorIPIV=std::vector&lt;ptrdiff_t={int}, std::allocator&lt;int&gt;&gt;, MatrixB=boost::numeric::ublas::matrix&lt;double, boost::numeric::ublas::column_major, boost::numeric::ublas::unbounded_array&lt;double,</div>
<div>1&gt;                      std::allocator&lt;QuantLib::Real={double}&gt;&gt;&gt;]&quot; at line 61 of &quot;C:\working\etk_lib\test_suite\etk_testsuite\test_ublas.cpp&quot;</div></div><div><br></div><div><br></div><div>
On a quick review of the traits classes, I didn&#39;t see any value_type , but I don&#39;t really know what I am doing...</div><div>When I comment out the STATIC_ASSERT, the bindings work for gesv.  My matrix types are definitely the same here.</div>
<div><br></div><div>Note, here is my test code:</div><div><div>ublas::matrix&lt;double, column_major&gt; A(3,3);</div><div>A(0,0) = 2; A(0,1) = 0; A(0,2) = 0;</div><div>A(1,0) = 0; A(1,1) = 1; A(1,2) = 0;</div><div>A(2,0) = 0; A(2,1) = 0; A(2,2) = 2;</div>
<div><br></div><div>ublas::matrix&lt;double, column_major&gt; b(3,1);</div><div>b(0,0) = 2;</div><div>b(1,0) = 1;</div><div>b(2,0) = 2;</div><div><br></div><div>std::vector&lt;int&gt; ipiv (3);<br></div><div>lapack::gesv(A, ipiv, b);<br>
</div><div>cout &lt;&lt; b;</div><div>//Gives the proper answer after the above changes</div></div>