[82d934]: NEWS-3.9  Maximize  Restore  History

Download this file

129 lines (87 with data), 4.9 kB

IT++ version 3.9.1                                         Released: 2006-01-28

This is a bug-fix release. Therefore, no new features have been added
since release 3.9.0. Only minor modifications and bug-fixes have been done.


* Fixed a few test programs that use LAPACK, so now they are robust to
  non-unique results and floating-point precision differences on different

* Protocol sources are now included in the MSVC++ project file.

* Fixed bug [1412383]. The dec2bin() functions appended en extra zero in
  front of the bit vector. This wrong behaviour has been fixed.

* Fixed bug [1399822]. The problem was caused by explicit specialisation of
  some templated operators, when HAVE_CBLAS or HAVE_MKL are defined. In such
  a case, the following forward declarations of explicit instantiations
  (with extern) of these operators should be switched off.

* Fixed bug [1396020]. Now itsave() works properly in both Matlab and
  Octave. Thanks to Jordy Potman for providing a patch.

* A lot of other minor bug-fixes. For a full list of them, please check the
  ChangeLog file.

IT++ version 3.9.0                                         Released: 2005-12-28

New features:

* Completely new configuration, compilation and installation approach based
  on the GNU Autotools: Autoconf, Automake and Libtool. External libraries
  (FFTW, CBLAS and LAPACK) installed system-wide are detected by the
  configure script.

* Instead of `itpp-config' script, the `pkg-config'
  (http://pkgconfig.freedesktop.org/wiki/) tool can be used for  linking
  with the IT++ library.

* Implementation of `erf(z)' function for complex argument `z' added.

* New statistical functions `within_tolerance(x, xref, tol)' added.

* New GF2 matrix class added, which is designed for a memory efficient
  representation of binary matrices.

* Polynomials support and filter design functions added.

* Added real Toeplitz matrices using the function toeplitz(mat, mat).
  Generalized also to be valid for non-square matrices.


* Removed support for libharu PDF generation library. External tools can be
  used instead, for instance:
  - Gnuplot - http://www.gnuplot.info/
  - Mathplotlib - http://matplotlib.sourceforge.net/
  - Asymptote - http://asymptote.sourceforge.net/

* Documentation is now provided in HTML format only. It is generated and
  installed automatically if the following tools are available: Doxygen,
  LaTeX, Dvips and Ghostscript.

* BLERC class improved by adding setup and clear flags and non default
  constructor. Now it is possible to measure BLER for changing block sizes,
  e.g. in an adaptive transmission.

* diag() function now takes a second argument stating which diagonal to fill
  with values (Matlab compatible).

* Implemented feature request [1167378]. An error message is displayed if
  the Parser's init() functions can not open a file.


* Fixed bug [1388024]. `fft_real()' and `ifft_real()' functions now work
  correctly for both external libraries: FFTW3 and MKL. They are also
  compatible to Matlab's implementation.

* Fixed test programs failures due to minor precision differences on
  different platforms.

* Fixed bug [1333897]. The problem with	infinite loop in
  `it_file::seek(...)' function has been solved by fixing
  `it_file::write_data_header(...)' function.

* Fixed	bug [1358858]. Now it is possible to use decode_tailbite() method
  for a punctured code.

* Fixed bug [1358861]. Now the destructor properly deletes the fading
  generators dynamically created.

* Fixed bug [1302350]. Corrected behavior of operator<< in `itpp/itconfig.h'
  when the imaginary part equals "minus zero".

* Fixed a bug in the oversampling of the OFDM class.

* Fixed bug [1254814]. Added `itpp::' explicit scope for `it_assert_f',
  `it_warning_f' and `it_error_f' functions. Now the macros that use them
  work properly without `using namespace itpp' in the application file.

* Fixed reopened bug [1094934]. Now endianity of x86_64 platforms can be
  successfully detected.

* Fixed bug [1251878]. `min' and `max' functions did not return indexes
  properly due to the lack of references in functions' arguments.

* Fixed bug [1084246]. Operators and other functions in Vec, Mat and Array
  revised to follow standard C++ conventions. Now operators provide an
  lvalue and also check against self-assignment.

* A lot of other minor bug-fixes. For a full list of them, please check the
  ChangeLog file.


Get latest updates about Open Source Projects, Conferences and News.

Sign up for the SourceForge newsletter:

No, thanks