Seldon 5.2

Seldon provides matrix and vector structures, and some methods and functions for computations (linear algebra). Seldon is designed to be efficient and convenient, which is notably achieved thanks to template classes. Exception handling and several debug levels are helpful while coding.

Seldon provides several types of matrices: sparse matrices, symmetric matrices, full matrices, hermitian matrices and triangular matrices. Each type includes several formats. E.g. symmetric matrices may be stored as full matrices or in packed form (Blas format).

What's New in This Release: [ read full changelog ]

· Fixed a bug: the methods 'Read' and 'Write' did not work properly under MS
· Windows because the streams were not opened with the option 'binary'.
· Fixed a bug: 'MallocObject' could fail to deallocate or reallocate memory
· under MS Windows because of alignment problems.
· The method 'Fill(const T0& x)' for matrices was called instead of the
· assignment operator which is not inherited in C++. This led to a compilation
· error.
· Set 'data_' to 'NULL' in 'Matrix_Base' copy constructor, so that the copy
· constructors of all derived classes will be safe, even if they call the copy
· constructor of 'Matrix_Base'.
· Fixed a few error messages related to input/output operations.
· The copy constructor of 'Matrix_Sparse' now calls that of 'Matrix_Base' in
· order to avoid a compilation warning.
· Added the prefix 'std::' to 'cout' and 'endl' in the preprocessor directives
· ERR, DISP, DISPLAY and END.
· 'Matrix_Pointers::Resize' does not perform a useless copy anymore when the
· dimensions do not change.