* \brief Holds information about the various numeric (i.e. scalar) types allowed by Eigen.
*
* \param T the numeric type at hand
* \tparam T the numeric type at hand
*
* This class stores enums, typedefs and static methods giving information about a numeric type.
*
* The provided data consists of:
* \li A typedef \a Real, giving the "real part" type of \a T. If \a T is already real,
* then \a Real is just a typedef to \a T. If \a T is \c std::complex<U> then \a Real
* \li A typedef \c Real, giving the "real part" type of \a T. If \a T is already real,
* then \c Real is just a typedef to \a T. If \a T is \c std::complex<U> then \c Real
* is a typedef to \a U.
* \li A typedef \a NonInteger, giving the type that should be used for operations producing non-integral values,
* \li A typedef \c NonInteger, giving the type that should be used for operations producing non-integral values,
* such as quotients, square roots, etc. If \a T is a floating-point type, then this typedef just gives
* \a T again. Note however that many Eigen functions such as internal::sqrt simply refuse to
* take integers. Outside of a few cases, Eigen doesn't do automatic type promotion. Thus, this typedef is
* only intended as a helper for code that needs to explicitly promote types.
* \li A typedef \c Literal giving the type to use for numeric literals such as "2" or "0.5". For instance, for \c std::complex<U>, Literal is defined as \c U.
* Of course, this type must be fully compatible with \a T. In doubt, just use \a T here.
* \li A typedef \a Nested giving the type to use to nest a value inside of the expression tree. If you don't know what
* this means, just use \a T here.
* \li An enum value \a IsComplex. It is equal to 1 if \a T is a \c std::complex
...
...
@@ -42,10 +75,14 @@ namespace Eigen {
* \li An enum value \a IsSigned. It is equal to \c 1 if \a T is a signed type and to 0 if \a T is unsigned.
* \li An enum value \a RequireInitialization. It is equal to \c 1 if the constructor of the numeric type \a T must
* be called, and to 0 if it is safe not to call it. Default is 0 if \a T is an arithmetic type, and 1 otherwise.
* \li An epsilon() function which, unlike std::numeric_limits::epsilon(), returns a \a Real instead of a \a T.
* \li An epsilon() function which, unlike <a href="http://en.cppreference.com/w/cpp/types/numeric_limits/epsilon">std::numeric_limits::epsilon()</a>,
* it returns a \a Real instead of a \a T.
* \li A dummy_precision() function returning a weak epsilon value. It is mainly used as a default
* value by the fuzzy comparison operators.
* \li highest() and lowest() functions returning the highest and lowest possible values respectively.
* \li digits10() function returning the number of decimal digits that can be represented without change. This is
* the analogue of <a href="http://en.cppreference.com/w/cpp/types/numeric_limits/digits10">std::numeric_limits<T>::digits10</a>
* which is used as the default implementation if specialized.
* \param SizeAtCompileTime the number of rows/cols, or Dynamic
* \param MaxSizeAtCompileTime the maximum number of rows/cols, or Dynamic. This optional parameter defaults to SizeAtCompileTime. Most of the time, you should not have to specify it.
* \param IndexType the interger type of the indices
* \tparam SizeAtCompileTime the number of rows/cols, or Dynamic
* \tparam MaxSizeAtCompileTime the maximum number of rows/cols, or Dynamic. This optional parameter defaults to SizeAtCompileTime. Most of the time, you should not have to specify it.
* \tparam _StorageIndex the integer type of the indices
*
* This class represents a permutation matrix, internally stored as a vector of integers.
*
* \sa class PermutationBase, class PermutationWrapper, class DiagonalMatrix
// We have a 1x1 matrix/array => the argument is interpreted as the value of the unique coefficient (case where scalar type can be implicitely converted)
// if so, then we must take care at removing the call to nested_eval in the specializations (e.g., in permutation_matrix_product, transposition_matrix_product, etc.)
* \brief A matrix or vector expression mapping an existing expressions
*
* \tparam PlainObjectType the equivalent matrix type of the mapped data
* \tparam Options specifies whether the pointer is \c #Aligned, or \c #Unaligned.
* The default is \c #Unaligned.
* \tparam StrideType optionally specifies strides. By default, Ref implies a contiguous storage along the inner dimension (inner stride==1),
* but accept a variable outer stride (leading dimension).
* This can be overridden by specifying strides.
* The type passed here must be a specialization of the Stride template, see examples below.
*
* This class permits to write non template functions taking Eigen's object as parameters while limiting the number of copies.
* A Ref<> object can represent either a const expression or a l-value:
* \code
* // in-out argument:
* void foo1(Ref<VectorXf> x);
*
* // read-only const argument:
* void foo2(const Ref<const VectorXf>& x);
* \endcode
*
* In the in-out case, the input argument must satisfies the constraints of the actual Ref<> type, otherwise a compilation issue will be triggered.
* By default, a Ref<VectorXf> can reference any dense vector expression of float having a contiguous memory layout.
* Likewise, a Ref<MatrixXf> can reference any column major dense matrix expression of float whose column's elements are contiguously stored with
* the possibility to have a constant space inbetween each column, i.e.: the inner stride mmust be equal to 1, but the outer-stride (or leading dimension),
* can be greater than the number of rows.
*
* In the const case, if the input expression does not match the above requirement, then it is evaluated into a temporary before being passed to the function.
* Here are some examples:
* \code
* MatrixXf A;
* VectorXf a;
* foo1(a.head()); // OK
* foo1(A.col()); // OK
* foo1(A.row()); // compilation error because here innerstride!=1
* foo2(A.row()); // The row is copied into a contiguous temporary
* foo2(2*a); // The expression is evaluated into a temporary
* foo2(A.col().segment(2,4)); // No temporary
* \endcode
*
* The range of inputs that can be referenced without temporary can be enlarged using the last two template parameter.
* Here is an example accepting an innerstride!=1:
* \code
* // in-out argument:
* void foo3(Ref<VectorXf,0,InnerStride<> > x);
* foo3(A.row()); // OK
* \endcode
* The downside here is that the function foo3 might be significantly slower than foo1 because it won't be able to exploit vectorization, and will involved more
* expensive address computations even if the input is contiguously stored in memory. To overcome this issue, one might propose to overloads internally calling a
* template function, e.g.:
* \code
* // in the .h:
* void foo(const Ref<MatrixXf>& A);
* void foo(const Ref<MatrixXf,0,Stride<> >& A);
*
* // in the .cpp:
* template<typename TypeOfA> void foo_impl(const TypeOfA& A) {
* ... // crazy code goes here
* }
* void foo(const Ref<MatrixXf>& A) { foo_impl(A); }
* void foo(const Ref<MatrixXf,0,Stride<> >& A) { foo_impl(A); }
AlignmentMatch=(int(traits<PlainObjectType>::Alignment)==int(Unaligned))||(DerivedAlignment>=int(Alignment)),// FIXME the first condition is not very clear, it should be replaced by the required alignment
* \brief A matrix or vector expression mapping an existing expression
*
* \tparam PlainObjectType the equivalent matrix type of the mapped data
* \tparam Options specifies the pointer alignment in bytes. It can be: \c #Aligned128, , \c #Aligned64, \c #Aligned32, \c #Aligned16, \c #Aligned8 or \c #Unaligned.
* The default is \c #Unaligned.
* \tparam StrideType optionally specifies strides. By default, Ref implies a contiguous storage along the inner dimension (inner stride==1),
* but accepts a variable outer stride (leading dimension).
* This can be overridden by specifying strides.
* The type passed here must be a specialization of the Stride template, see examples below.
*
* This class provides a way to write non-template functions taking Eigen objects as parameters while limiting the number of copies.
* A Ref<> object can represent either a const expression or a l-value:
* \code
* // in-out argument:
* void foo1(Ref<VectorXf> x);
*
* // read-only const argument:
* void foo2(const Ref<const VectorXf>& x);
* \endcode
*
* In the in-out case, the input argument must satisfy the constraints of the actual Ref<> type, otherwise a compilation issue will be triggered.
* By default, a Ref<VectorXf> can reference any dense vector expression of float having a contiguous memory layout.
* Likewise, a Ref<MatrixXf> can reference any column-major dense matrix expression of float whose column's elements are contiguously stored with
* the possibility to have a constant space in-between each column, i.e. the inner stride must be equal to 1, but the outer stride (or leading dimension)
* can be greater than the number of rows.
*
* In the const case, if the input expression does not match the above requirement, then it is evaluated into a temporary before being passed to the function.
* Here are some examples:
* \code
* MatrixXf A;
* VectorXf a;
* foo1(a.head()); // OK
* foo1(A.col()); // OK
* foo1(A.row()); // Compilation error because here innerstride!=1
* foo2(A.row()); // Compilation error because A.row() is a 1xN object while foo2 is expecting a Nx1 object
* foo2(A.row().transpose()); // The row is copied into a contiguous temporary
* foo2(2*a); // The expression is evaluated into a temporary
* foo2(A.col().segment(2,4)); // No temporary
* \endcode
*
* The range of inputs that can be referenced without temporary can be enlarged using the last two template parameters.
* Here is an example accepting an innerstride!=1:
* \code
* // in-out argument:
* void foo3(Ref<VectorXf,0,InnerStride<> > x);
* foo3(A.row()); // OK
* \endcode
* The downside here is that the function foo3 might be significantly slower than foo1 because it won't be able to exploit vectorization, and will involve more
* expensive address computations even if the input is contiguously stored in memory. To overcome this issue, one might propose to overload internally calling a
* template function, e.g.:
* \code
* // in the .h:
* void foo(const Ref<MatrixXf>& A);
* void foo(const Ref<MatrixXf,0,Stride<> >& A);
*
* // in the .cpp:
* template<typename TypeOfA> void foo_impl(const TypeOfA& A) {
* ... // crazy code goes here
* }
* void foo(const Ref<MatrixXf>& A) { foo_impl(A); }
* void foo(const Ref<MatrixXf,0,Stride<> >& A) { foo_impl(A); }