Menu
▾
▴
[62db13]: / ChangeLog-2007 Maximize Restore History
2200 lines (1572 with data), 87.4 kB
2007-12-30 Erik G. Larsson <erik_g_larsson@users.sourceforge.net>
* doc/local/verification.doc: Documentation update
2007-12-24 Adam Piatyszek <ediap@users.sourceforge.net>
* configure.ac.in, itpp/base/Makefile.am, itpp/base/sources.mk,
itpp/base/zdotusub.f, m4/acx_blas.m4: Removed redundant zdotusub_
Fortran wrapper to zdotu_ function
Now, the proper calling conventions of zdotu_ Fortran function
should be detected properly in all cases. Therefore, the zdotusub_
wrapper function, which was a temporary workaround solution,
should not be required any more.
2007-12-23 Adam Piatyszek <ediap@users.sourceforge.net>
* m4/ax_func_zdotu.m4: Improved zdotu_ calling convention tests
The previous tests did not work properly on Debian unstable (sid)
and SUSE 10.0 64-bit architectures. The improved tests work as
expected.
2007-12-21 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/blas.h, itpp/base/vec.h: Fix wrong zdotu_ calling
convention on Debian amd64
This patch promotes the calling convention of BLAS zdotu_ function
that returns its result through the first function argument. This
is the default method used by older GCC Fortran
compiler (version 3).
On Debian amd64, the configure-time detection of the proper
calling convention results in both methods working. However, the
method that returns the complex result by value, like standard
C/C++ functions, causes failures of blas_test and vec_test.
This patch fixes bug report [1855481].
2007-12-17 Adam Piatyszek <ediap@users.sourceforge.net>
* doc/local/linking.doc: Cosmetic documentation fixes
------------------------------------------------------------------------------
2007-12-15 Adam Piatyszek <ediap@users.sourceforge.net>
* IT++ 4.0.1 released (SVN tag: release-4-0-1)
* NEWS: Release notes for IT++ 4.0.1 added
2007-12-13 Adam Piatyszek <ediap@users.sourceforge.net>
* doc/Makefile.am, doc/tutorial/src/Makefile.am: Build tutorial
examples when doing "make check"
Compile and link time incompatibilities between the library
interfaces and tutorial example programs should now be detected
during the standard check procedure.
* doc/local/installation.doc, doc/local/linking.doc: Documentation
updated to reflect changes in MSVC++ project files
* win32/itpp_acml.vcproj, win32/itpp_acml_tests/*.vcproj,
win32/itpp_mkl.vcproj, win32/itpp_mkl_tests/*.vcproj: Replace
explicit paths with LIB and INCLUDE environment variables
With this change the user is responsible for setting the LIB and
INCLUDE environment variables with proper directories to MKL or
ACML librarary and include files.
Besides, build multi-threaded DLL libraries by default.
* itpp/base/binfile.cpp, itpp/base/copy_vector.h: Include missing
header files to fix compilation using GCC 4.3
2007-12-11 Erik G. Larsson <erik_g_larsson@users.sourceforge.net>
* itpp/comm/ldpc.h, itpp/comm/ldpc.cpp: Improvements to the
generation of regular LDPC codes
2007-12-10 Erik G. Larsson <erik_g_larsson@users.sourceforge.net>
* doc/tutorial/src/ldpc_bersim_awgn.cpp: Changed tutorial so it
counts bit errors in entire block, not only systematic bits
* itpp/comm/ldpc.cpp, tests/ldpc_test.ref: Fixed spelling error
2007-12-10 Adam Piatyszek <ediap@users.sourceforge.net>
* m4/acx_blas.m4: Improve autodetection of ACML libraries
* itpp/base/algebra/qr.cpp: Fix incorrect precalculation of the
workspace size for QR decomp.
The size of the workspace memory is now checked for each of
[dz]gerf_, [dz]orgqr_, [dz]geqp3_ and [dz]ungqr_ funcitions
separately, because they operate on different matrices. Besides,
R is now correctly initialised with A before quering for the
workspace size.
* itpp/base/algebra/svd.cpp: Improvements in calculation of the
SVD workspace size
Before resizing the current memory allocated for SVD workspace,
check if the workspace query was sucessful. If not, use the
default, but not optimum workspace size.
2007-12-09 Simon Wood <sgwoodjr@users.sourceforge.net>
* itpp/base/algebra/qr.cpp, itpp/base/algebra/svd.cpp: Added a
workspace query for each of the SVD and QR routines
The query returns the optimal value of the memory workspace size
used in SVD and QR decompositions. Setting the workspace size to
an optimum value greatly improves the speed of calculations.
2007-12-08 Erik G. Larsson <erik_g_larsson@users.sourceforge.net>
* doc/tutorial/src/ldpc_bersim_awgn.cpp: Fixes in tutorial
examples.
2007-12-07 Adam Piatyszek <ediap@users.sourceforge.net>
* Makefile.am, configure.ac.in, doc/local/linking.doc,
itpp_debug.pc.in: Provide itpp_debug.pc pkg-config's file
This additional file can be used to obtain the debugging compiler
flags and libraries. It is only installed, when IT++ is configured
with the --enable-debug option.
2007-12-06 Adam Piatyszek <ediap@users.sourceforge.net>
* INSTALL, doc/local/installation.doc: Installation manual updated
to reflect recent changes
* itpp/base/math/misc.h, itpp/srccode/audiofile.h: Fix minor
Doxygen warnings
* configure.ac.in, itpp/base/converters.h, itpp/config_msvc.h:
Move GCC_VERSION macro to the generated config.h file
* itpp/comm/convcode.cpp: Replace ambiguous expression with a
correct one
"i = ++i;" is ambiguous in C/C++. For more information, see
question 3.3 of the comp.lang.c FAQ list:
http://c-faq.com/expr/ieqiplusplus.html
* configure.ac.in: Use AC_MSG_NOTICE for printing configuration
summary
Thanks to this change, the configuration summary is automatically
included in config.log, so there is no need to create a separate
config_summary.log file.
* configure.ac.in: Simplify detection of explicit library
dependencies
There is no need to check for "deplibs_check_method" by executing
the "./libtool --config" command. This information is already
available in a cached variable "lt_cv_deplibs_check_method"
provided by the AC_PROG_LIBTOOL macro.
2007-12-05 Adam Piatyszek <ediap@users.sourceforge.net>
* configure.ac.in: Use "host_os" instead of "host" to simplify
case statements
* configure.ac.in: Enable explicit library dependencies on Mac
OS X
Dynamic linking on Mac OS X requires passing explicit library
dependencies to the linker, even if libtool thinks it is not
necessary. Therefore "enable_explicit_deps" is set to "yes"
for this operating system. The solution is taken from pkg-config
Portfile from macports.
* itpp/base/random.cpp: Use tgamma() function instead of
deprecated ambiguous gamma()
* itpp/base/math/elem_math.h: Make sure itpp::gamma() use correct
gamma implementation
BTW, denote that this function name is deprecated due to its
different implementation (and meaning) in BSD and glibc.
* configure.ac.in: std::isfinite, std::isinf and std::isnan are
broken on FreeBSD
The use of these functions on FreeBSD with GCC 3.4.x compilers
causes infinite loops or segmentation faults in bessel_test.
Therefore, do not check for these functions and override them
with local inlined implementations defined in
itpp/base/math/main.h.
* configure.ac.in, itpp/base/bessel/gamma.cpp
itpp/base/math/misc.h itpp/config_msvc.h
itpp/stat/mog_generic.cpp: Add full support of std::isfinite(),
std::isinf() and std::isnan()
Instead of using isnan(), isinf() C99 macros and deprecated
finite() function switch to their equivalents in std library. This
should solve recent portability problems under MinGW/MSYS.
* itpp/base/bessel/hyperg.cpp, itpp/base/bessel/iv.cpp,
itpp/base/bessel/jv.cpp, itpp/base/bessel/k1.cpp,
itpp/base/bessel/kn.cpp: Unify warning messages
* itpp/base/bessel/airy.cpp, itpp/base/bessel/bessel_internal.h,
itpp/base/bessel/chbevl.cpp, itpp/base/bessel/hyperg.cpp,
itpp/base/bessel/i0.cpp, itpp/base/bessel/i1.cpp,
itpp/base/bessel/iv.cpp, itpp/base/bessel/jv.cpp,
itpp/base/bessel/k0.cpp, itpp/base/bessel/k1.cpp,
itpp/base/bessel/kn.cpp, itpp/base/bessel/polevl.cpp: Clean up
"include" and "using namespace" statements
* itpp/base/bessel/bessel_internal.h, itpp/base/bessel/gamma.cpp,
itpp/base/bessel/hyperg.cpp, itpp/base/bessel/iv.cpp,
itpp/base/bessel/jv.cpp, itpp/base/bessel/sources.mk,
itpp/base/bessel/struve.cpp, win32/itpp_acml.vcproj,
win32/itpp_mkl.vcproj: Fixed wrong calculation of bessely()
detected on Intel C++ compiler
When IT++ was compiled with Intel C++ (icpc) compiler, the
bessel_test failed on calculation of bessely() function for
non-integer orders. This bug was caused by the fact that the
ported bessel functions from Cephes Math Library should use
a gamma function, which sets "signgam" global variable. But
instead C99 tgamma() was used, which does not set "signgam".
This patch adds explicit gam() and lgam() definitions, ported
from the original Cephes Math Library, so all bessel functions
in IT++ now use the proper gamma implementations.
2007-12-04 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/optim/newton_search.cpp: Initialize variables with zeros to
suppress compiler warnings
* itpp/base/math/misc.h: Add missing right parenthesis in isnan()
definition
2007-12-03 Adam Piatyszek <ediap@users.sourceforge.net>
* configure.ac.in, itpp/base/math/misc.h, itpp/config_msvc.h: Fix
improper use of HAVE_DECL_XXX definitions
HAVE_DECL_XXX are always defined as 0 or 1, so one can not simply
check if these names are defined.
2007-11-30 Adam Piatyszek <ediap@users.sourceforge.net>
* configure.ac.in, m4/acx_blas.m4, m4/acx_lapack.m4: Do not
require Fortran compiler for BLAS/LAPACK linking
BLAS and LAPACK libraries should be checked even if no Fortran
compiler is available. For instance, BLAS/LAPACK routines from MKL
can be used without additional Fortran libraries. This patch
changes IT++ configuration scripts to make this possible.
* m4/acx_blas.m4: Add support for MKL 10
MKL 10 requires explicit linking to -lguide and -lpthread
libraries. With this patch MKL 8.1.1 or newer versions should be
automatically detected on both 32- and 64-bit (em64t) Linux
systems.
2007-11-26 Adam Piatyszek <ediap@users.sourceforge.net>
* configure.ac.in, itpp/config_msvc.h: Get rid of "min" and "max"
macro definitions
MSVC++ compiler defines these two macros and they might clash with
our proper C++ code using std::min() and std::max() functions
only. This patch removes such macros conditionally in the
generated config.h and config_msvc.h header files.
* m4/ax_func_zdotu.m4: Cosmetic changes in printed messages
2007-11-23 Adam Piatyszek <ediap@users.sourceforge.net>
* configure.ac.in, itpp/base/blas.h, itpp/base/vec.h,
itpp/config_msvc.h, m4/acx_blas.m4, m4/ax_func_zdotu.m4: Detect
incompatible g77/gfortran calling conventions of zdotu_
The calling conventions of zdotu_ BLAS function are different
for g77 and gfortran compilers. In the case of g77, the result
is passed using a first argument of zdotu_ function, whereas for
gfortran the result is simply returned.
To avoid building an extra Fortran wrapper, which workarounds
this issue, IT++ needs to know which calling convention should be
used. This patch adds two macros to the configure scripts that
compile and try to run simple C++ programs using zdotu_ BLAS
function, As a result of these checks the correct calling convention
is used in itpp::dot() function. The Fortran zdotusub_ wrapper
function can still be linked and used as a fallback solution.
For more information on this issue, please have a look at this
bug report: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20178
* itpp/base/factory.h: Add explicit constructor definition in a
default Factory class
Without this explicit empty constructor in the Factory class, some
compilers (PGI, HP C/aC++) give warnings or errors on a constant
DEFAULT_FACTORY.
* configure.ac.in, itpp/base/math/misc.h, itpp/config_msvc.h: Add
additional checks for isinf() and isfinite() macros
Without these new checks the compilation might fail on
redeclaration of these macros with inline functions in
itpp/base/math/misc.h. This problem occurs for instance on a
HP-UX system.
2007-11-21 Adam Piatyszek <ediap@users.sourceforge.net>
* configure.ac.in, itpp/base/math/misc.h, itpp/config_msvc.h:
Clean up checks for finite(), isfinite() and isnan()
isnan() is defined as a macro on some systems in <math.h>, so we
should check for it. BTW, removed macro definition of finite(),
which is also defined as inline function in
"itpp/base/math/misc.h".
* configure.ac.in: Removed duplicated checks
AC_HEADER_STDC is invoked when AC_LANG([C++]) is used, whereas
AC_PROG_MAKE_SET is executed by AM_INIT_AUTOMAKE.
* m4/ac_cxx_extern_template.m4: Obsolete macros replaced with the
ones recommended by autoconf
AC_LANG_SAVE + AC_LANG_CPLUSPLUS -> AC_LANG_PUSH([C++])
AC_LANG_RESTORE -> AC_LANG_POP([C++])
AC_TRY_COMPILE -> AC_COMPILE_IFELSE + AC_LANG_PROGRAM
* configure.ac.in, m4/acx_blas.m4, m4/acx_lapack.m4,
m4/acx_fft.m4: Use AS_HELP_STRING macros instead of obsolete
AC_HELP_STRING ones
* Makefile.am, autogen.sh, configure.ac.in,
m4/ac_cxx_extern_template.m4, m4/acx_blas.m4, m4/acx_lapack.m4,
m4/acx_fft.m4: Renamed "config" directory to "m4" to reflect its
contents
* autogen.sh, configure.ac.in: Use "build-aux" directory for
auxiliary building scripts
This directory name is suggested in the autoconf manual as a
standard placeholder for auxiliary build scripts. If this
directory does not exist, create it before invoking autotools.
* autogen.sh: On error return the exit code of the failing command
* autogen.sh: Allow runing autogen.sh using an absolute path
After checking the necessary tools the script now enters the
IT++ source directory for the bootstrap process.
* configure.ac.in: Report all automake warnings and treat them as
errors
This should give the maintainer and people building IT++ from SVN
a better detection of possible bugs in Makefiles.
2007-11-16 Adam Piatyszek <ediap@users.sourceforge.net>
* configure.ac.in, tests/Makefile.am: Bump autoconf requirement to
2.60 and use AC_PROG_SED
The $datarootdir and $docdir variables, which are used for HTML
documentation installation, were introduced in autoconf 2.60.
Therefore we need to check for this version. Besides, manual check
for sed program is now replaced with AC_PROG_SED macro, which
makes sed as a required dependency. Additionally, gdiff and
gnudiff are searched as diff replacement.
* tests/parser_test.cpp: Prevent warnings about string to char*
conversions
GCC >=4.2 uses "-Wwrite-string" in when "-Wall" flag is used and
this results in warnings about deprecated conversions from
constant strings to char*. This patch workarounds this problem by
a bit different initialisation of arrays of char* strings.
2007-11-12 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/binfile.h: Fixed invalid cast from bool to enum
The Apple's g++3.3 compiler from Mac Developer Tools v1.2 reports
an error on a static cast from bool to enum. This patch changes
the problematic code, so it should be accepted by older compilers.
2007-11-02 Adam Piatyszek <ediap@users.sourceforge.net>
* Makefile.am, configure.ac.in: Save configuration summary to
config_summary.log file
* configure.ac.in, itpp-config.in, itpp.pc.in: Improved
itpp-config and itpp.pc scripts
This patch adds an additional configure check to see whether
external libraries' dependencies need to be explicitly added to
the output of "itpp-config --libs" command. As a result it should
be always possible to use "itpp-config --libs" or "pkg-config itpp
--libs" without adding an extra --static switch. Besides,
"itpp-config --libs" should now return exactly the same libraries
as "pkg-config itpp --libs" on Mac OS X.
* doc/tutorial/src/mimoconv.cpp: Updated MIMO with convolutional
coding example program
2007-11-01 Erik G. Larsson <erik_g_larsson@users.sourceforge.net>
* doc/local/installation.doc: Updated the "no root" installation
description
2007-10-30 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/mat.h: operator*(Vec &v, Mat &m) documented as
deprecated
The existence of this operator in IT++ is questionable, since its
behaviour is equivalent to "outer_product(v, m.get_row(0))".
Moreover, the documentation was wrong so many users had problems
using it correctly. This patch marks this operator as deprecated,
both in the documentation and through a run-time warning message.
2007-10-25 Adam Piatyszek <ediap@users.sourceforge.net>
* AUTHORS, doc/local/authors.doc: List of authors updated
* configure.ac.in: Do not use "-Werror" flag for debugging library
by default
This flag was propagated into itpp-config and itpp.pc scripts and
this resulted in reporting each warning in users' programs as an
error.
2007-10-23 Adam Piatyszek <ediap@users.sourceforge.net>
* tests/gf2mat_test.cpp: Suppress compilation warning about string
to "char*" conversion
* itpp/base/converters.cpp, itpp/base/converters.h: Workaround for
compilation errors when using GCC 3.3.x
GCC 3.3.x has problems with compilation of a few explicit
instantiations in converters.cpp from IT++ 4.0.0. This only
happens when -finline-functions flag is used. This flag is
included in -03 optimisations. This patch conditionally disables
these problematic instantiation when GCC version < 3.4.0.
2007-10-22 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/algebra/svd.h, itpp/base/matfunc.h: Documentation of
diag() and svd() updated
* itpp/base/matfunc.h: Fixed a bug in rank() function
The order of checks in a while loop was wrong and it caused
assertion when accessing a non-existing element in a vector of
singular values. This patch fixes this problem.
2007-10-18 Adam Piatyszek <ediap@users.sourceforge.net>
* doc/doxygen_html.cfg.in: Reset TAB_SIZE to 8 spaces
Some of the IT++ codes still use tabs of 8 spaces for indenting.
Thus this patch resets the TAB_SIZE to 8 (default value) in
Doxygen's configuration file.
2007-10-16 Adam Piatyszek <ediap@users.sourceforge.net>
* Makefile.am, VERSION, autogen.sh, configure.ac.in
extras/itpp-config.t2t, itpp-config.1.in, itpp.spec.in:
Changed the way version numbers are handled
This patch fixes the problem of removing the "itpp.spec" and
"itpp-config.1" files during "make distclean". These files should
not be removed, since they are distributed in packages. Since now,
version numbers in these two files and in "configure.ac" are
updated according to a VERSION file content, when "autogen.sh"
bootstrapping script is executed.
------------------------------------------------------------------------------
2007-10-14 Adam Piatyszek <ediap@users.sourceforge.net>
* IT++ 4.0.0 released (SVN tag: release-4-0-0)
* NEWS: Updated for 4.0.0 stable release
2007-10-12 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/converters.h, itpp/base/help_functions.h,
itpp/base/mat.h, itpp/base/matfunc.cpp, itpp/base/matfunc.h,
itpp/base/smat.cpp, itpp/base/smat.h, itpp/base/svec.cpp,
itpp/base/svec.h, itpp/base/vec.cpp, itpp/base/vec.h,
itpp/comm/error_counters.cpp, itpp/comm/interleave.cpp,
itpp/comm/interleave.h, itpp/comm/pulse_shape.cpp,
itpp/comm/pulse_shape.h, itpp/fixed/cfixed.h,
itpp/fixed/fix_operators.h, itpp/fixed/fixed.h,
itpp/signal/filter.cpp, itpp/signal/filter.h,
itpp/signal/transforms.cpp, itpp/signal/transforms.h: Use
HAVE_EXTERN_TEMPLATE where necessary
2007-10-11 Adam Piatyszek <ediap@users.sourceforge.net>
* config/ac_cxx_extern_template.m4, configure.ac,
itpp/config_msvc.h: Check for C++ "extern template" when
configuring IT++
Previously, it was assumed that all C++ compilers except MSVC++
support extern template instantiation. Now this feature is checked
with the configure script and HAVE_EXTERN_TEMPLATE is defined
accordingly.
* doc/doxygen_html.cfg.in: Doxygen configuration updated with
Doxygen 1.4.7
* doc/local/authors.doc, itpp/base/bessel/bessel_internal.h,
itpp/base/blas.h, itpp/base/converters.cpp,
itpp/base/converters.h, itpp/base/copy_vector.h,
itpp/base/factory.cpp, itpp/base/help_functions.cpp,
itpp/base/help_functions.h, itpp/base/itassert.cpp,
itpp/base/mat.cpp, itpp/base/mat.h, itpp/base/matfunc.h,
itpp/base/math/elem_math.h, itpp/base/specmat.cpp,
itpp/base/vec.cpp, itpp/base/vec.h, itpp/comm/convcode.cpp,
itpp/comm/crc.cpp, itpp/comm/ldpc.h,
itpp/comm/rec_syst_conv_code.cpp, itpp/fixed/cfixed.h,
itpp/fixed/fix_factory.h, itpp/fixed/fix_functions.h,
itpp/fixed/fixed.h, itpp/protocol/events.h,
itpp/protocol/front_drop_queue.h, itpp/protocol/packet.h,
itpp/protocol/packet_channel.h, itpp/protocol/packet_generator.h,
itpp/protocol/selective_repeat.cpp,
itpp/protocol/selective_repeat.h, itpp/protocol/signals_slots.h,
itpp/protocol/tcp.cpp, itpp/protocol/tcp.h,
itpp/protocol/tcp_client_server.h, itpp/signal/freq_filt.cpp,
itpp/srccode/audiofile.cpp, itpp/srccode/audiofile.h,
itpp/srccode/gmm.cpp, itpp/srccode/gmm.h,
itpp/srccode/lpcfunc.cpp, itpp/srccode/pnm.cpp,
itpp/srccode/pnm.h, itpp/srccode/vq.cpp, itpp/srccode/vq.h,
itpp/srccode/vqtrain.cpp, itpp/srccode/vqtrain.h,
itpp/stat/mog_diag.h, itpp/stat/mog_diag_em.h,
itpp/stat/mog_diag_kmeans.h, itpp/stat/mog_generic.h: Cleanup of
Doxygen documentation
Missing documentation reported by Doxygen have been marked by "ADD
DOCUMENTATION HERE" string. Besides, code fragments that do not
require documentation have been disabled from parsing by using
"\cond" and "\endcond" Doxygen's commands.
* itpp/base/converters.cpp, itpp/base/converters.h: Moved
templated functions' definitions into a header file
2007-10-10 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/vec.cpp, itpp/base/vec.cpp: Removed unnecessary
conditional checks for MKL & MSVC++
2007-10-09 Adam Piatyszek <ediap@users.sourceforge.net>
* INSTALL, configure.ac, doc/Makefile.am,
doc/local/installation.doc: Removed redundant "--with-docdir=DIR"
switch
This switch is not needed any more, because recent autoconf
provides "--docdir=DIR" switch with the same functionality by
default. Installation manual updated accordingly. Thanks to
Guenter Dannoritzer for reporting this problem.
* itpp/base/matfunc.h: Disabled incorrect specialisation of rank()
function
The rank() function specialisation for binary GF(2) matrices
returned incorrect results. Now it is disabled - the
specialisation results in an error message.
------------------------------------------------------------------------------
2007-10-06 Adam Piatyszek <ediap@users.sourceforge.net>
* IT++ 4.0.0-rc2 released (SVN tag: release-4-0-0-rc2)
* NEWS: Updated to reflect recent changes
2007-10-05 Adam Piatyszek <ediap@users.sourceforge.net>
* configure.ac: Changed version number of the trunk sources
The version number of SVN trunk sources is now of the form:
"major.minor-dev", e.g. "4.1-dev".
* itpp/base/matfunc.h, tests/matfunc_test.cpp,
tests/matfunc_test.ref: Added templated rank() function
The rank() function calculates the rank of a matrix by counting
the number of non-zero singular values. LAPACK's SVD function is
thus required by this function. Thanks to Martin Senst for
submitting the patch.
* INSTALL, doc/local/installation.doc: Installation manual updated
to reflect recent changes
2007-10-04 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/binary.cpp, itpp/base/binary.h: Check for allowed 0 or
1 values in ">>" operator of bin class
Without this patch it was possible to initialise binary vectors or
matrices with values other than 0 or 1, when linking to the IT++
library build without "-DNDEBUG" flag.
BTW, performed a set of cosmetic changes in bin class operators
and constructors.
2007-10-02 Adam Piatyszek <ediap@users.sourceforge.net>
* INSTALL, doc/local/installation.doc, config/acx_fft.m4: Added
--with-fft-include=DIR option to the configure command
This option can be used instead of CPPFLAGS to pass non standard
paths to include files required by IT++ FFT routines (using ACML,
MKL or FFTW).
* config/acx_blas.m4, config/acx_fft.m4, config/acx_lapack.m4:
Cosmetic changes in --with-PACKAGE strings of the configure
command
2007-10-01 Adam Piatyszek <ediap@users.sourceforge.net>
* tests/Makefile.am: All test programs needs to be declared in
ALL_TESTS
Without this change, some reference files (*.ref) might not have
been included in the distribution package when --without-lapack or
--without-fft switch was used.
* config/acx_fft.m4: Do not set blas_xxx_ok=yes when detecting FFT
libraries
This patch fixes FFT library detection macros, so they do not set
blas_mkl_ok or blas_acml_ok variables. Without this fix, it might
happened that even if BLAS detection was disabled (e.g. by using
`--without-blas'), an ACML or MKL BLAS library was indicated as
found.
* config/acx_blas.m4, itpp/base/blas.h, itpp/base/vec.h,
itpp/config_msvc.h: Fixed failing blas_test and vec_test when
linking with MKL
There was a problem with using zdotusub_ Fortran wrapper with
Intel's MKL library, which resulted in segmentation faults in
blas_test and vec_test programs. Now zdotusub_ is defined as
zdotu_ when MKL BLAS is used, and zdotusub.f is not built in such
a case. BTW, the detection of MKL, ACML and ATLAS BLAS libraries
is improved (additional HAVE_BLAS_MKL, HAVE_BLAS_ACML and
HAVE_BLAS_ATLAS definitions are available).
* configure.ac, itpp/base/Makefile.am: Fixed linking problem with
zdotu_ when not using BLAS
If an external BLAS library is not used, do not compile and link
the zdotusub_ Fortran wrapper, since it requires an external
zdotu_ function to exist.
------------------------------------------------------------------------------
2007-09-28 Adam Piatyszek <ediap@users.sourceforge.net>
* IT++ 4.0.0-rc1 released (SVN tag: release-4-0-0-rc1)
* NEWS: Updated to reflect recent changes
* tests/itfile_test.cpp: Fixed a minor bug in itfile_test program
This test program opened the reference data file from source
directory in read/write mode, which was wrong, because source
directory can be located on read-only file system. This patch
fixes this bug.
2007-09-27 Adam Piatyszek <ediap@users.sourceforge.net>
* tests/modulator_nd_test.cpp, tests/modulator_nd_test.ref: Test
program updated to use ZF Log-MAP demodulation method
2007-09-26 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/comm/modulator_nd.cpp, itpp/comm/modulator_nd.h:
Implemented Zero-Forcing (ZF) Log-MAP soft demodulation method
Zero-Forcing approximate demodulation method can be used instead
of a brute-force optimum enumeration of all constellation points
in both real- and complex-valued MIMO demodulators. Thanks to Erik
G. Larsson, who provided the draft implementation and reviewed the
code.
2007-09-07 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/comm/modulator_nd.cpp, itpp/comm/modulator_nd.h,
tests/modulator_nd_test.cpp: Interface improvements in
Modulator_ND classes
The interfaces of Modulator_ND class and its descendants were
redesigned to be more similar to 1D and 2D modulators' interfaces.
BTW, revised the documentation and performed a set of cosmetic
changes.
2007-09-06 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/comm/channel.h: Documentation improvements in AWGN_Channel
Described clearly what is the meaning of the AWGN_Channel's
"noisevar" input parameter for real- and complex-valued signals.
2007-09-04 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/math/elem_math.h: Added sign_i() and sgn_i() functions
These new functions calculates the sign of an integer or double
argument and return -1, 0 or 1 integer value.
2007-08-30 Adam Piatyszek <ediap@users.sourceforge.net>
* configure.ac, itpp/base/math/log_exp.h, itpp/base/specmat.cpp,
itpp/comm/galois.cpp, itpp/comm/modulator_nd.cpp,
itpp/signal/sigfun.cpp: Improved log2() and log1p() portability
Additional checks for log2() and log1p() functions have been added
to the configure script. If they are not found, local
implementations are used. Besides, log2() is now in a global
scope, not in the "itpp" namespace, which is consistent with GCC.
BTW, removed unnecessary checks for C++ standard header files.
* itpp/itmex.h: Added missing mxArray2string() and
string2mxArray() functions
Thanks to Bogdan Cristea for reporting this issue and providing
the patch.
* doc/local/linking.doc, extras/itpp-config.t2t, itpp-config.1.in,
itpp-config.in, itpp.pc.in: Improvements in itpp-config and
pkg-config's itpp.pc scripts
These changes are intended to follow the distinction between
linking to a static and shared version of the library, which is
used in pkg-config's itpp.pc file. By default "itpp-config --libs"
returns "-I${prefix}/lib -litpp", whereas when "--static" is added
in front of "--libs", the returned string is extended with a set
of external library names required for proper static linking.
Moreover, an additional "--debug" switch has been added, which
changes the output of "itpp-config --cflags" and "itpp-config
--libs" by providing debugging flags and debugging library name.
For backward compatibility "--cflags-opt", "--cflags-debug",
"--libs-opt" and "--libs-debug" still exist, but they are not
documented and their use is deprecated.
2007-08-28 Adam Piatyszek <ediap@users.sourceforge.net>
* config/acx_blas.m4, config/acx_fft.m4, config/acx_lapack.m4,
configure.ac, itpp-config.in, itpp.pc.in: Break configuration when
BLAS, LAPACK or FFT not found
If the required external libraries are not found, stop
configuration process and print an error message informing that
one can use "--without-{blas,lapack,fft}" option, but IT++
functionality will be limited. BTW, removal of redundant spaces
from CXXFLAGS_OPT, CXXFLAGS_DEBUG CPPFLAGS, LDFLAGS and LIBS has
been implemented in the configure script.
2007-08-27 Adam Piatyszek <ediap@users.sourceforge.net>
* configure.ac, tests/Makefile.am, tests/cholesky_test.cpp,
tests/det_test.cpp, tests/eigen_test.cpp, tests/fastica_test.cpp,
tests/filter_design_test.cpp, tests/freq_filt_test.cpp,
tests/inv_test.cpp, tests/ls_solve_test.cpp, tests/lu_test.cpp,
tests/matfunc_test.cpp, tests/modulator_nd_test.cpp,
tests/poly_test.cpp, tests/qr_test.cpp, tests/schur_test.cpp,
tests/sigfun_test.cpp, tests/stat_test.cpp, tests/svd_test.cpp,
tests/transforms_test.cpp: Made LAPACK and FFT dependent tests
conditional
If linking to an optional LAPACK or FFT external library is not
configured, do not compile and run the test programs that depends
on these libraries. This results in all tests passed when
"--without-{blas,lapack,fft}" configure switch is intentionally
used.
* itpp/base/factory.cpp, itpp/base/factory.h,
itpp/fixed/fix_factory.cpp: Cosmetic changes in memory allocation
Instead of using "new char[size]" for raw memory allocation, use
more generic "operator new(size)". Similarly, use "operator
delete(p)" instead of "delete [] (p)". This makes the code
consistent with 16-byte aligned create_elements() and
destroy_elements() specialisations.
* itpp/base/array.h, itpp/base/factory.cpp, itpp/base/factory.h,
itpp/base/mat.h, itpp/base/vec.h, itpp/fixed/fix_factory.cpp:
Improved memory allocation routines
Following the idea proposed in feature request [1277721]
implementation of create_elements() functions was improved. First
the memory for an Array, Vec or Mat is allocated and then the
constructor of each element is executed, but only if necessary.
For simple types like bin, int, double, etc., the constructor is
not invoked. Moreover, in the previous implementation the
constructor of each Array<Array>, Array<Vec> or Array<Mat> element
was invoked twice. Now this is fixed. Thanks to George Jongren for
submitting this feature request.
* itpp/base/factory.cpp, itpp/base/factory.h,
itpp/base/sources.mk, itpp/signal/transforms.cpp,
win32/itpp_acml.vcproj, win32/itpp_mkl.vcproj: Solved FFTW
alignment issue
By implementing new specialisations of create_elements() and
destroy_elements() functions, 16-byte memory alignment of Array,
Vec and Mat data is used for double and complex<double> types.
With this new implementation, fft(), ifft(), dct() and idct()
functions using FFTW external library run up to 30% faster than
with non-aligned data structures. This fix closes feature
request [1418707].
* itpp/base/array.h, itpp/base/factory.h, itpp/base/mat.h,
itpp/base/vec.h: Added destroy_elements() function
The destroy_elements() function accompanies create_elements() used
for memory allocation. Future changes to memory allocation
scheme (e.g. memory alignment) for Array, Vec and Mat classes will
only require redefinition of these two methods.
BTW, removed inlining of alloc(), free(), create_elements() and
destroy_elements() functions, which does not give any real
performance gain but the code is bigger than necessary.
2007-08-26 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/mat.h, itpp/base/vec.cpp, itpp/base/vec.h,
itpp/fixed/cfix.cpp, itpp/fixed/cfix.h, itpp/fixed/fix.cpp,
itpp/fixed/fix.h: Improvements in string parsers
Added support for NaN (not a number) and Inf (infinity) for parsed
Vec<double> arguments. This fixes feature request [1066711].
Changed "bool Vec<>::set()" into "void Vec<>::set(), because
returned value was always true. Reused Vec<int>::set()
specialisation in Vec<short int>::set() implementation. Simplified
implementation of Vec<int>::set() specialisation.
2007-08-25 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/mat.h, tests/mat_test.cpp, tests/mat_test.ref:
Added set_rows() and set_cols() methods to Mat
These new set functions match existing get_rows() and get_cols()
methods. This patch fixes feature request [711263].
2007-08-24 Andreas Wolfgang <andreas-w@users.sourceforge.net>
* extras/itsave.m Fixed script to be compatible with Matlab V7.3
Since Matlab does not support sizeof(...), it was replaced by a new
function itsizeof(...). In addition logical expressions have been
made Matlab compatible (only variables with identical dimensions
are used in logical expressions). The patch was tested using
Matlab V7.3 and Octave V2.9.9.
2007-08-22 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/random.h: Fine tuning of random generators
When generating a vector or matrix of random samples, it is more
efficient to use "*=" and "+=" operators on the same allocated
memory pool than create temporary copies of vectors and matrices.
Moreover, for double and complex<double> types, "*=" operator can
employ optimised BLAS implementation.
* itpp/comm/channel.cpp, itpp/comm/channel.h: Fine tuning of
AWGN_Channel operator() performance
Instead of using randn() and randn_c() global functions,
Normal_RNG and Complex_Normal_RNG generators are declared as class
objects. This reduces the number of generator's constructor
executions to minimum.
2007-08-23 Adam Piatyszek <ediap@users.sourceforge.net>
* tests/Makefile.am: Return non-zero exit status on make check
failure
If not all test programs pass, output a short status message with
the number of tests failed and return non-zero exit status to
make.
2007-08-21 Adam Piatyszek <ediap@users.sourceforge.net>
* doc/local/installation.doc, doc/local/linking.doc: Documentation
updated to reflect recent changes
* win32/itpp_acml_tests/fastica_test.vcproj,
win32/itpp_acml_tests/itfile_test.vcproj,
win32/itpp_acml_tests/parser_test.vcproj,
win32/itpp_mkl_tests/*.vcproj: Final problems fixed in the MSVC++
project files for test programs
The fastica_test, itfile_test and parser_test projects require
additional preprocessor definitions of some test data files.
Besides, all project files in win32/itpp_mkl_tests directory were
missing ProjectGUID, which is normally automatically added by the
MSVC++ IDE environment. All test programs should now compile and
run without any problems.
2007-08-20 Adam Piatyszek <ediap@users.sourceforge.net>
* extras/acml_tests_vcproj.template,
extras/mkl_tests_vcproj.template, win32/**/*.vcproj: Removed
versions from MKL and ACML directories
To prevent unnecessary updates of MSVC++ project files, the
installation directories of MKL and ACML libraries under windows
do not contain version numbers, i.e. "C:\Program Files\Intel\MKL"
and "C:\Program Files\AMD\acml". MKL and ACML libraries should be
installed into these directories in order to work with the
prepared project files without any modifications.
* itpp/base/mat.cpp, itpp/base/mat.h, itpp/base/vec.cpp,
itpp/base/vec.h: Fixed problems with comma in Vec::set() parser
The previous string parser implementation in Vec class was not
fully functional when using MSVC++ compiler. The problem was that
commas just after a parsed value caused application crash. For
instance "1 , 0 1 " was parsed fine, but parsing "1, 0,1" resulted
in a crash. Fixed this issue by replacing all commas with spaces
before actual parsing is done, except the complex vector case,
where "(1,0.2)" are valid complex values. BTW, the Mat::set()
parsing function was rewritten to reuse the Vec::set() method.
* extras/acml_tests_vcproj.template, extras/gen_vcproj_tests.sh,
extras/mkl_tests_vcproj.template, win32/itpp_acml.vcproj,
win32/itpp_acml_tests/*.vcproj, win32/itpp_acml_tests/Makefile.am,
win32/itpp_acml_tests/itpp_acml_tests.sln, win32/itpp_mkl.vcproj,
win32/itpp_mkl_tests/*.vcproj, win32/itpp_mkl_tests/Makefile.am,
win32/itpp_mkl_tests/itpp_mkl_tests.sln: Prepared MSVC++ project
files for test programs
These project files were generated from two template files with
"gen_vcproj_tests.sh" script. Then, they were included in
appropriate solutions: "itpp_acml_tests.sln" and
"itpp_mkl_tests.sln", for external ACML and MKL libraries
respectively.
* itpp/base/blas.h, itpp/base/vec.cpp, itpp/base/vec.h,
itpp/base/zdotusub.f: Modified zdotusub_() function to match ACML
interface
The Fortran wrapper zdodusub_() of BLAS zdotu_() function used
CBLAS inteface, which was incompatible with ACML implementation.
Fixed this, so the native ACML function can be used under MSVC++
compiler. Besides, disabled this function when using IT++ with
Intel MKL under MSVC++.
* tests/fastica_test.cpp, tests/itfile_test.cpp,
tests/parser_test.cpp, tests/turbo_test.cpp: Cosmetic changes in a
few test programs
Compilation of turbo_test and itfile_test programs under MSVC++
resulted in some warnings related to the use of variable `i' and
initialisation of a float variable `f_ref'. These issues are now
fixed. BTW, error messages in fastica_test, itfile_test and
parser_test were changed to use `cerr' output stream, instead of
`cout'.
2007-08-18 Adam Piatyszek <ediap@users.sourceforge.net>
* tests/rand_test.cpp, tests/rand_test.ref: Fixed a bug in
rand_test program
The rand_test implementation was not deterministic, which resulted
in a failed test result on SunOS SPARC platform. Generation of
scalars, vectors and matrices of random numbers in separate "cout"
commands fixes this problem.
* doc/local/linking.doc: Documentation updated
Reflected recent changes in the pkg-config's itpp.pc file in the
"Linking with IT++" documentation. BTW, revised the section about
using itload.m and itsave.m m-files in Matlab/Octave.
* itpp/base/vec.cpp, itpp/base/vec.h: Fixed regressions related to
the specialised operator*()
The templated friend operator*() for two vectors, which calculates
the dot product, needs to be instantiated in the same way as the
dot() function. Besides, the declaration of the friend operator*()
included the definition. This definition is now moved below dot()
function definition. Without these two changes, a segmentation
fault occurred in vec_test under Cygwin.
* itpp.pc.in: Improvements in pkg-config's itpp.pc script
"Libs" variable settings in the itpp.pc pkg-config's file are now
split into "Libs" and "Libs.private". The latter one is only used
when doing a static linking (using "--static" pkg-config switch).
* config/acx_blas.m4, config/acx_fft.m4, config/acx_lapack.m4:
Cosmetic changes - deleted trailing whitespace
* configure.ac: Let the user override the default CXXFLAGS_DEBUG
The CXXFLAGS_DEBUG environment variable can be used to override
the default settings of this flag. BTW, added "-Werror" to the
default debug flags.
2007-08-13 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/random.cpp, itpp/base/random.h, tests/rand_test.cpp,
tests/rand_test.ref, tests/source_test.cpp, tests/source_test.ref:
Simplified and optimised RNG implementations
Rayleigh_RNG and Rice_RNG now use Normal_RNG for efficient
generation of samples (new implementation is about 4 times
faster). Minor code optimisations in Laplace_RNG, Weibull_RNG and
AR1_Normal_RNG. Test programs for RNG and deterministic sources
updated accordingly.
2007-08-11 Adam Piatyszek <ediap@users.sourceforge.net>
* Makefile.am: Added NEWS-3.99 to EXTRA_DIST
* tests/stat_test.cpp: Check for LAPACK
This test requires svd() function, which is provided by an
external LAPACK library. Therefore we have to check for
HAVE_LAPACK definition like in other LAPACK-related test programs.
* configure.ac: Version bump
* NEWS, NEWS-3.99: Renamed NEWS file to NEWS-3.99
* NEWS-3.10: Updated to include 3.10.12 release notes
------------------------------------------------------------------------------
2007-08-10 Adam Piatyszek <ediap@users.sourceforge.net>
* IT++ 3.99.3.1 released (SVN tag: release-3-99-3-1)
* NEWS: Updated to reflect recent changes
* configure.ac: Version bump
* configure.ac, itpp/base/Makefile.am, itpp/base/blas.h,
itpp/base/sources.mk, itpp/base/vec.h, itpp/base/zdotusub.f: Fixed
a bug related to zdotu_() BLAS function
Wrong declaration of zdotu_() BLAS function interface caused
segmentation faults when IT++ was compiled with
"-fomit-frame-pointer" flag. To fix this problem a Fortran wrapper
function zdotusub_() has been added to IT++. This wrapper function
comes from NetLib's CBLAS package.
------------------------------------------------------------------------------
2007-08-10 Adam Piatyszek <ediap@users.sourceforge.net>
* IT++ 3.99.3 released (SVN tag: release-3-99-3)
* NEWS: Updated to reflect recent changes
2007-08-09 Adam Piatyszek <ediap@users.sourceforge.net>
* configure.ac, win32/Makefile.am,
win32/itpp_acml_tests/Makefile.am,
win32/itpp_acml_tests/array_test.vcproj,
win32/itpp_acml_tests/bch_test.vcproj,
win32/itpp_acml_tests/bessel_test.vcproj,
win32/itpp_acml_tests/blas_test.vcproj,
win32/itpp_acml_tests/itpp_acml_tests.sln, win32/itpp_mkl.vcproj,
win32/itpp_mkl_tests/Makefile.am,
win32/itpp_mkl_tests/array_test.vcproj,
win32/itpp_mkl_tests/bch_test.vcproj,
win32/itpp_mkl_tests/bessel_test.vcproj,
win32/itpp_mkl_tests/blas_test.vcproj,
win32/itpp_mkl_tests/itpp_mkl_tests.sln: Added MSVC++ project
files for test programs
This patch includes MSVC++ project files for easy building and
linking a few test programs with IT++ and MKL or ACML libraries.
The project files are configured to use 32-bit versions of MKL
9.1.025 and ACML 3.6.0 PGI libraries installed into their default
locations on Windows platforms.
2007-08-08 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/blas.h, itpp/base/vec.h, tests/blas_test.cpp,
tests/blas_test.ref: Fixed a bug in dot() for complex arguments
Declaration of BLAS zdotu_() method changed so, the first argument
now passes the complex result. This fixes segmentation fault of
blas_test and vec_test when using GCC 3.4.5 under SunOS 5.9 on
SPARC. It also makes workaround for MSVC++ warning C4190
redundant.
* tests/mat_test.cpp, tests/mat_test.ref: Improvements in mat_test
Perform more consistent tests of the templated Mat class for bin,
int, double and std::complex<double> arguments.
2007-08-06 Adam Piatyszek <ediap@users.sourceforge.net>
* AUTHORS, doc/Makefile.am, doc/images/Makefile.am,
doc/images/newcom_logo.png, doc/local/authors.doc,
doc/local/itpp_header.html: NEWCOM logo removed
Since the NEWCOM project ended in March 2007, the existence of its
logo in IT++ HTML documentation is not necessary any more.
* tests/vec_test.cpp, tests/vec_test.ref: Improvements in vec_test
Perform more consistent tests of the templated Vec class for bin,
int, double and std::complex<double> arguments.
* itpp/base/mat.h, itpp/base/vec.h: Code cleanup in Vec/Mat's
multiply and division operators
Used BLAS-aided scal_vector() method in multiplication by constant
operator in Mat class. Replaced duplicated implementation with
executions of inlined multiplication operators in both Vec and Mat
classes. Simplified Vec's division operators by removal of
unnecessary checks for special cases.
* itpp/base/vec.h: Added missing initialization command
Output matrix has to be initialised with zeros in BLAS-based
Vec::outer_product() function specializations.
2007-08-03 Erik G. Larsson <erik_g_larsson@users.sourceforge.net>
* itpp/comm/ldpc.h: Documentation updates
2007-07-21 Adam Piatyszek <ediap@users.sourceforge.net>
* AUTHORS, doc/local/authors.doc, itpp/base/sort.h,
tests/Makefile.am, tests/sort_test.cpp, tests/sort_test.ref:
Sorting functions revised
The sort() and sort_index() functions can now use four different
algorithms: Introsort, Quicksort, Heapsort and Insertion Sort,
which are implemented in a common Sort class. The Introsort is
the default method. The revised implementation is much faster
because it uses pointers for accessing the data to be sorted.
Sorting test program is included in this commit as well.
Thanks to Mark Dobossy for this improved implementation.
This commit closes feature request [1746876].
2007-07-20 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/blas.h, itpp/base/copy_vector.h, itpp/base/vec.h:
Added scal_vector() and axpy_vector() low level functions
These two functions uses BLAS methods for double and
std::complex<double> types of a vector. The function scal_vector()
realizes "x *= alpha", whereas axpy_vector() realizes
"y = alpha*x + y". Vector x and y are of the same size.
BTW, improved performance of Vec's "*=" and "/=" operators with
scalar arguments by using scal_vector() method.
* INSTALL, README, configure.ac, doc/doxygen_html.cfg.in,
doc/local/index.doc.in, doc/local/installation.doc,
itpp/base/blas.h, itpp/base/copy_vector.h, itpp/base/mat.cpp,
itpp/base/mat.h, itpp/base/sources.mk, itpp/base/vec.cpp,
itpp/base/vec.h, itpp/config_msvc.h, itpp/signal/transforms.cpp,
tests/Makefile.am, tests/blas_test.cpp, tests/blas_test.ref,
win32/itpp_acml.vcproj, win32/itpp_mkl.vcproj: Removed CBLAS
dependency
To limit the number of external libraries required by IT++ CBLAS
interfaces have been replaced with a Fortran BLAS. Besides, vec
and cvec specialisation of the outer_product() method using BLAS
functions has been added. To verify BLAS-aided routines, an extra
blas_test has been added.
This commit closes feature request [1746876].
2007-07-19 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/mat.h: Fixed a bug in set_size() method
When resizing a smaller matrix into a bigger one with copying the
content, the unused entries were not initialised to zeros
properly.
Besides, the use of a copy_vector() function instead of nested
loops with element by element copying increases the overall
performance of various operations on matrices. Therefore, this
patch improves performance of set_size(), transpose(),
hermitian_transpose(), concat_horizontal() and concat_vertical()
functions.
* itpp/base/vec.h : Fixed assertion check in operator "/="
2007-07-18 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/vec.h: Use copy_vector() where possible
This patch improves the consistency and possibly the performance
of various Vec's methods by using copy_vector() instead of
explicit copying of elements in a loop. It also fixes a minor bug
in "/=" operator for the case of division of a vector with itself,
i.e. "v /= v;". Finally, this patch adds missing friend
declarations for concat() functions and improves if statements in
the alloc() function.
2007-07-17 Adam Piatyszek <ediap@users.sourceforge.net>
* **/*: Removed trailing whitespace
Trailing whitespace (spaces and tabs at the end of each line) does
not contain any information. Moreover, it sometimes causes
conflicts when applying patches or merging branches. Therefore,
let's get rid of it from the code and other text files.
Trailing whitespace can be easily removed with sed or perl and
their regular expressions:
% sed -i -e 's/[ \t]*$//' source.cpp
% perl -i -p -e 's/[ \t]*$//' source.cpp
2007-07-16 Adam Piatyszek <ediap@users.sourceforge.net>
* doc/local/documentation.doc, extras/itload.m, extras/itsave.m,
itpp/**/*.cpp, itpp/**/*.h, tests/*.cpp: Removed $Revision: 1418 $ and
$Date: 2008-02-15 15:03:10 +0100 (Fri, 15 Feb 2008) $ keywords
The $Revision: 1418 $ and $Date: 2008-02-15 15:03:10 +0100 (Fri, 15 Feb 2008) $ keywords in each file comment sometimes
cause errors when trying to apply patches or merged branches.
Therefore, IT++ sources will not use them any more.
* itpp/**/*.cpp, itpp/**/*.h: Removed trailing whitespace
Trailing whitespace (spaces and tabs at the end of each line) does
not contain any information. Moreover, it sometimes causes
conflicts when applying patches or merging branches. Therefore,
let's get rid of it from the code.
* itpp/base/mat.h: Improved del_rows() performance
The more optimum but not working implementation of del_rows() has
been fixed and uncommented, so it now replaces the slow execution
of del_row() method for each deleted row in a loop.
2007-07-11 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/comm/commfunc.cpp: Compilation warning suppressed by
initializing the lambda variable
2007-07-04 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/matfunc.cpp, itpp/base/matfunc.h: Added repmat()
function for vectors
Matrin Senst provided a new repmat() implementation which
concatenates a multiple number of vectors. It extends the
functionality of recently added repmat() functions for matrices.
2007-06-29 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/signal/resampling.h: Minor spelling fixes in documentation
* tests/*.cpp: Copyright dates updated
* configure.ac: Version number updated
* itpp/base/matfunc.cpp, itpp/base/matfunc.h,
tests/matfunc_test.cpp, tests/matfunc_test.ref: Added repmat()
function (feature request [1744527])
This functions is a clone of a Matlab's repmat function. It
expands an input matrix or vector to a larger matrix by
concatenating (m x n) copies of the input matrix (vector). Thanks
to Mark Dobossy for providing a patch.
2007-06-22 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/comm/ldpc.cpp: Improved LDPC binary files versioning
A global static const variable LDPC_binary_file_version has been
defined, which is now used in load/save methods of LDPC_Generator
and LDPC_Code classes. Besides, H_defined and G_defined boolean
variables are loaded/saved using the new stream operators of
it_file and it_ifile classes.
* TODO, itpp/base/binfile.cpp, itpp/base/binfile.h,
itpp/base/itfile.cpp, itpp/base/itfile.h, tests/itfile_test.cpp,
tests/itfile_test.ref, tests/itfile_test_data.it: Added support
for bool variables in it_file class
It is now possible to read and write bool variables using standard
stream operators ">>" and "<<" of the it_file and it_ifile
classes. Boolean variables are stored in a binary file using a
char datatype.
------------------------------------------------------------------------------
2007-06-20 Adam Piatyszek <ediap@users.sourceforge.net>
* IT++ 3.99.2 released (SVN tag: release-3-99-2)
* NEWS, TODO: Updated to reflect recent changes
* itpp/comm/ldpc.cpp: Fixed a minor bug reported as error in
MSVC++
The implicit cast of int variables to bool results in compilation
errors under MS Visual C++ .NET 2005 compiler. The applied fix is
a workaround solution. In future fstream operators ">>" and "<<"
for bool variables should be implemented in it_file class.
* configure.ac: Removed "-fno-exceptions" flag from CXXFLAGS
This conditional flag was added to CXXFLAGS and CXXFLAGS_DEBUG
when "--enable-exceptions" switch was passed to configure.
However, it is better not to use it by default, since it might
cause problems when linking the IT++ library with other programs
that use exceptions.
2007-06-19 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/specmat.cpp, itpp/base/specmat.h: Fixed bug [1739500]
in hadamard() function
The initial recursion for creating Hadamard matrices has to start
with a 1x1 matrix H = [1], not with a 2x2 matrix. This fix
provides correct solution for imat H = hadamard(1). Thanks to
Matthias for reporting this problem.
2007-06-18 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/math/elem_math.cpp: Fixed bug [1739267] in binom() and
binom_i()
The binomial coefficient was calculated wrongly for n = k. Instead
of 1 the functions result was n+1.
2007-06-16 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/comm/ldpc.cpp, itpp/comm/ldpc.h: Added B-LDPC Codes
implementation
B-LDPC Codes implementation in IT++ consists of a BLDPC_Parity and
BLDPC_Generator classes. The former one can be used for
constructing the parity-check matrix from so-called base-matrix.
Base matrix can be saved and loaded from a text file. The
BLDPC_Generator class implements an efficient encoding algorithm
using a preprocessed version of the parity-check matrix.
2007-06-10 Adam Piatyszek <ediap@users.sourceforge.net>
* INSTALL, config/acx_blas.m4, config/acx_fft.m4,
config/acx_lapack.m4, configure.ac, doc/local/installation.doc,
itpp/config_msvc.h, itpp/signal/transforms.cpp:
Improvements in MKL detection macros
The latest MKL release 9.1.018 has different naming of a lapack
shared library (libmkl_lapack32.so and libmkl_lapack64.so ->
libmkl_lapack.so). Cosmetic changes in MKL's FFT macro and
variable names (MKL8 -> MKL)
2007-06-04 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/srccode/pnm.cpp: The functions for reading ppm and pgm
images have to use binary mode when opening files. This fixes bug
[1730610]. Thanks to Thomas Andersson for reporting this problem.
------------------------------------------------------------------------------
2007-06-02 Adam Piatyszek <ediap@users.sourceforge.net>
* IT++ 3.10.11 released (SVN tag: release-3-10-11)
(merged rev. 1050:1051 from itpp-3-10 branch)
* NEWS-3.10: Updated to reflect recent changes
(merged rev. 1050:1051 from itpp-3-10 branch)
* itpp/base/vec.h: The outer_product() function now uses an extra
argument "hermitian", which controls the calculation for complex
vectors. By default "hermitian" is set to false, which results in
(x * y^T) calculation. If "hermitian" is true, (x * y^H) is
calculated. This solution is backward compatible with 3.10.x
interface of outer_product(). Final fix for bug [1727923].
2007-05-29 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/vec.h: Added specialisation of the outer_product()
method for complex vector arguments. This fixes bug [1727923].
Thanks to Yngve Selen for reporting this issue.
* itpp/signal/transforms.cpp: Fixed bug [1724746], i.e. a problem
with modified input data in ifft_real() function. Additional
FFTW_PRESERVE_INPUT flag used during FFTW plan creation solves
this problem. Thanks to Klaas Hofstra for reporting this bug.
2007-05-22 Adam Piatyszek <ediap@users.sourceforge.net>
* tests/schur_test.cpp, tests/schur_test.ref: Schur decomposition
is not unique, therefore the comparision of the non-zero norm of
the lower triangular part of matrix T was not a good idea. Fixed
this problem in the schur_test program.
2007-05-15 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/signal/transforms.cpp, itpp/signal/transforms.h: Fixed
linking errors when trying to use Hadamard and Walsh-Hadamard
transform functions. This fixes bug report [1719146] submitted by
Stephan Ludwig.
2007-05-03 Adam Piatyszek <ediap@users.sourceforge.net>
* AUTHORS, doc/local/authors.doc: Sean Finney added to the list of
contributors
* itpp/base/random.h, tests/channel_test.ref,
tests/cholesky_test.ref, tests/det_test.ref, tests/eigen_test.ref,
tests/filter_design_test.ref, tests/filter_test.ref,
tests/inv_test.ref, tests/ls_solve_test.ref, tests/lu_test.ref,
tests/matfunc_test.ref, tests/modulator_nd_test.ref,
tests/modulator_test.ref, tests/poly_test.ref, tests/qr_test.ref,
tests/schur_test.ref, tests/source_test.ref, tests/svd_test.ref,
tests/transforms_test.ref: New implementation of the
Complex_Normal_RNG class, which now uses the improved Normal_RNG
for generating the real and imaginary parts of the complex random
sample. The new generator is about four times faster than the
previous implementation. This closes feature request [1711727].
Thanks to Sean Finney for providing a patch.
* itpp/comm/ldpc.cpp, itpp/comm/ldpc.h: Now the decode() methods
return only systematic bits, to be symetric with the encode()
methods. New methods decode_soft_out() added, which return soft
LLR values for all bits in a codeword.
2007-05-02 Erik G. Larsson <erik_g_larsson@users.sourceforge.net>
* itpp/tests/commfunc_test.cpp, itpp/tests/commfunc_test.ref:
Added missing test program
* itpp/comm/commfunc.cpp: Fixed missing memory allocation
2007-04-30 Erik G. Larsson <erik_g_larsson@users.sourceforge.net>
* itpp/comm/commfunc.h, itpp/comm/commfunc.cpp: Added a new
function "waterfilling()" to solve water-filling problems
2007-04-29 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/random.cpp, itpp/base/random.h,
tests/cholesky_test.ref, tests/circular_buffer_test.ref,
tests/det_test.ref, tests/eigen_test.ref,
tests/filter_design_test.ref, tests/filter_test.ref,
tests/histogram_test.ref, tests/inv_test.ref, tests/ldpc_test.cpp,
tests/ldpc_test.ref, tests/ls_solve_test.ref, tests/lu_test.ref,
tests/matfunc_test.ref, tests/mat_test.ref,
tests/modulator_nd_test.ref, tests/modulator_test.ref,
tests/poly_test.ref, tests/qr_test.ref, tests/rand_test.ref,
tests/schur_test.ref, tests/sigfun_test.ref,
tests/source_test.ref, tests/sparse_test.ref, tests/stat_test.ref,
tests/svd_test.ref, tests/transforms_test.ref,
tests/turbo_test.ref, tests/vec_test.ref: New implementation of
the Normal_RNG sample() function, which uses the so-called
Ziggurat algorithm. The new generator is about five times faster
than the previous one using the Box-Mueller method. This fixes
feature request [1707969]. Thanks to Sean Finney for providing a
patch.
2007-04-26 Adam Piatyszek <ediap@users.sourceforge.net>
* doc/tutorial/src/ldpc_bersim_awgn.cpp, itpp/comm/ldpc.cpp,
itpp/comm/ldpc.h, tests/ldpc_test.ref: Fixed bugs in
LDPC_Code::load_code() and LDPC_Code::save_code() functions. It
could happen that LDPC generator object might be left
uninitialised when loading the coded from a binary file. Removed
unnecessary integrity_check() execution from the load_code()
function. setup_decoder() method replaced with three separate
methods, which are more flexible to use. Tutorial example updated
to reflect recent changes.
2007-04-23 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/itmex.h: Fixed bug [1703731] by implementing the missing
double2mxArray() and double_complex2mxArray() functions. Thanks to
Danilo Zanatta Filho for providing a patch.
2007-04-18 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/**/*.{cpp,h}: Copyright dates updated to `1995-2007'
* itpp/base/svec.h, itpp/protocol/selective_repeat.cpp,
itpp/stat/mog_diag.cpp: Once again added missing #include
<cstdlib> statements
2007-04-17 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/itassert.cpp, itpp/base/math/elem_math.h,
itpp/base/parser.cpp, itpp/comm/error_counters.cpp,
itpp/comm/llr.h, itpp/protocol/selective_repeat.cpp,
itpp/protocol/tcp.cpp, itpp/srccode/vq.cpp: Added missing #include
<cstdlib> and <limits> header files. This fixes problems with
building using GCC 4.3 snapshot. Thanks to Martin Michlmayr from
the Debian project for reporting this bug.
(merged rev. 1005:1006 from the itpp-3-10 branch)
2007-04-12 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/itassert.cpp: Fixed a problem with not displayed
assert and error messages, which was introduced in rev. 982.
------------------------------------------------------------------------------
2007-04-10 Adam Piatyszek <ediap@users.sourceforge.net>
* IT++ 3.10.10 released (SVN tag: release-3-10-10)
(merged rev. 999:1000 from the itpp-3-10 branch)
* NEWS-3.10: Updated to reflect recent changes
(merged rev. 999:1000 from the itpp-3-10 branch)
* TODO: Updated to reflect recent changes in IT++ file format
(merged rev. 986:997 from the it_file-v3 temporary branch)
* extras/itload.m, extras/itsave.m: Updated to be compatible with
IT++ file format version 3. Implemented missing load functions for
Array<Vec<T> >, Array<Mat<T> > and Array<string> types. Function
itsave() detects the type of a saved vector or matrix and uses
bvec, bmat, ivec, imat, vec, mat, cvec or cmat types respecively.
(merged rev. 986:997 from the it_file-v3 temporary branch)
* itpp/base/binfile.cpp, itpp/base/binfile.h,
itpp/base/gf2mat.cpp, itpp/base/itfile.cpp, itpp/base/itfile.h,
tests/itfile_test.cpp, tests/itfile_test.ref,
tests/itfile_test_data.it:
Major changes in IT++ file format related classes:
- the new file format (version 3) is incompatible with version 2,
which was used in IT++ 3.10.x series; for backward compatibility
the previous format can be handled using the it_file_old and
it_ifile_old classes, which are considered as deprecated and will
be removed from the library in future
- data is always saved using a little endian byte ordering; files
files written on machines using big or little endian ordering
should be identical; endianity variable removed from the
data_header structure
- all size variables are stored using fixed-width 64-bit unsigned
integer type, and therefore are independent of the architecture
word size (32- or 64-bit)
- short and int data types are explicitly casted to fixed-with
types int16_t, uint16_t, int32_t and uint32_t, since C/C++
standards do not assume fixed widths of these types
- cleaner and more efficient implementation of the
write_data_header() function
- added missing implementation of a pack() function, which compacts
an opened it_file by removing free space between data blocks
- added optional description filed which can be saved with each
variable
- added missing operators and methods for reading and writing svec
and smat data types
(merged rev. 986:997 from the it_file-v3 temporary branch)
* tests/timer_test.cpp, tests/timer_test.ref: CPU_Timer test is
now independent of a Real_Timer test. This should solve test
failures in some specific cases (high load of CPU, etc.).
(merged rev. 993:994 from the itpp-3-10 branch)
2007-04-07 Erik G. Larsson <erik_g_larsson@users.sourceforge.net>
* itpp/ldpc/ldpc.cpp, tests/ldpc_test.ref: "<=" changed to "<" in
iteration count in decoder. Thanks to M. Senst for reporting this.
2007-04-06 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/vec.cpp, tests/vec_test.cpp, tests/vec_test.ref: Fixed
a minor bug in the set() method, when parsing an "a:b" or "a:b:c"
fromat. Due to precision problems the last value from some
specific ranges was not included in the result, e.g. "0:0.2:3"
resulted in a vector with 2.8 at the end (instead of 3). Thanks to
Martin Senst for reporting this issue.
2007-04-04 Adam Piatyszek <ediap@users.sourceforge.net>
* tests/itfile_test.cpp, tests/itfile_test_data.it,
tests/itfile_test.ref: IT++ file format test program improved
2007-04-03 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/itassert.cpp: Fixed a compilation problem when IT++ is
configured with the "--enable-exceptions" switch
* itpp/base/itfile.cpp: Fixed bug [1693384] caused by a missing
explicit cast of string size to int
* configure.ac, itpp.spec.in: Minor cleanups and improvements
* AUTHORS, doc/local/authors.doc: Added Gunter Dannoritzer as
openSUSE distribution packages maintainer
2007-04-01 Erik G. Larsson <erik_g_larsson@users.sourceforge.net>
* itpp/comm/ldpc.h: Documentation updates
2007-03-30 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/comm/ldpc.cpp, itpp/comm/ldpc.h: Fixed a small bug in the
integrity_check() function. Some of the private data and methods
made protected to allow access by the classes who might inherit
from the LDPC_Code.
2007-03-27 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/gf2mat.cpp, itpp/base/gf2mat.h: Added set_size()
method to the GF2mat class, which is compatible with Mat's
set_size() function.
* configure.ac: Version bump
------------------------------------------------------------------------------
2007-03-22 Adam Piatyszek <ediap@users.sourceforge.net>
* IT++ 3.99.1 released (SVN tag: release-3-99-1)
* NEWS: Updated to reflect recent changes
* doc/local/verification.doc: Added a new verification report for
MSVC++ .NET and IT++ with ACML
2007-03-21 Adam Piatyszek <ediap@users.sourceforge.net>
* win32/itpp_acml.sln, win32/itpp_acml.vcproj, win32/itpp_mkl.sln,
win32/itpp_mkl.vcproj, doc/local/installation.doc,
doc/local/linking.doc, itpp/config_msvc.h, win32/itpp.sln,
win32/itpp.vcproj, win32/Makefile.am: Added a new MSVC++ project
file for linking IT++ with ACML (PGI build for Windows).
Installation manual and "Linking with IT++" document updated to
reflect these changes.
2007-03-20 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/random.cpp, itpp/base/random.h: Added
rand53_01_lclosed() function for 53-bit resolution [0,1) random
numbers generator. Global static initialization flag moved into
Random_Generator class.
2007-03-09 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/comm/ldpc.cpp, itpp/comm/ldpc.h: "std::string" replaced
with "const std::string&" in function arguments to prevent passing
by value. Cosmetic changes related to line wrapping.
2007-03-08 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/comm/ldpc.cpp, itpp/comm/ldpc.h: "init_state" variable
renamed to "init_flag". Cosmetic changes in the integrity_check()
function.
2007-03-07 Adam Piatyszek <ediap@users.sourceforge.net>
* doc/tutorial/src/ldpc_bersim_awgn.cpp,
doc/tutorial/src/ldpc_gen_codes.cpp, itpp/comm/ldpc.cpp,
itpp/comm/ldpc.h, tests/ldpc_test.cpp: LDPC_Parity_Matrix class
split into a set of classes, which inherits from the new base
class LDPC_Parity. Test and tutorials programs updated
accordingly.
2007-03-06 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/svec.h: Added a new method get_nz_indices(), which
returns the indices of non-zero values in a sparse vector.
2007-03-02 Adam Piatyszek <ediap@users.sourceforge.net>
* config/acx_blas.m4: Fixed a bug in BLAS detection, when an
explicit library is passed to configure using "--with-blas=..."
option.
2007-03-01 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/gf2mat.cpp, itpp/base/gf2mat.h: Portability improved
by replacing "unsigned short int" with 8-bit "unsigned char" type
for storing data bits. Implemented missing GF2mat constructor that
uses bmat as an argument. A few other cosmetic changes and
documentation improvements done.
* itpp/base/itfile.cpp, itpp/base/itfile.h: Added missing
read/write operators for the char type
2007-03-01 Erik G. Larsson <erik_g_larsson@users.sourceforge.net>
* itpp/comm/modulator_nd.h: Added a function set_llrcalc()
2007-02-28 Adam Piatyszek <ediap@users.sourceforge.net>
* tests/Makefile.am: Test programs are now linked to the optimised
library by default. Fixed the problem of missing "*.ref" files in
the distribution package, when some modules were disabled during
the configuration process.
* tests/gf2mat_test.cpp: Added EXTENSIVE_TESTS macro definition,
which turns on/off extensive tests.
2007-02-22 Adam Piatyszek <ediap@users.sourceforge.net>
* doc/tutorial/src/ldpc_bersim_awgn.cpp: Changed the Nbits type
to 64-bit int64_t
* itpp/comm/ldpc.cpp: Fixed a problem in the integrity_check()
function when no generator is defined.
2007-02-21 Adam Piatyszek <ediap@users.sourceforge.net>
* doc/tutorial/src/ldpc_bersim_awgn.cpp,
doc/tutorial/src/ldpc_gen_codes.cpp, itpp/comm/ldpc.cpp,
itpp/comm/ldpc.h, tests/ldpc_test.cpp: LDPC_Generator abstract
base class added. It provides a common interface for user defined
generators. LDPC_Generator_Matrix replaced with the
LDPC_Generator_Systematic class derived from the LDPC_Generator.
Improved the interface of LDPC_Code class. Tutorial examples and
test program updated accordingly.
2007-02-18 Erik G. Larsson <erik_g_larsson@users.sourceforge.net>
* itpp/comm/ldpc.h, itpp/comm/ldpc.cpp, tests/ldpc_test.cpp: Added
a feature to generator matrix computation (minor interface change)
2007-02-17 Adam Piatyszek <ediap@users.sourceforge.net>
* config/acx_blas.m4, config/acx_cblas.m4, config/acx_fft.m4,
config/acx_lapack.m4, configure.ac: Improved the speed and
portability of external libraries detection macros
2007-02-16 Erik G. Larsson <erik_g_larsson@users.sourceforge.net>
* itpp/comm/ldpc.cpp: Fixed some assertions
2007-02-16 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/comm/modulator.h: Fixed bug [1661419] by adding default
constructors to the QAM, PSK, PAM_c and PAM modulator classes.
Thanks to Jordy Potman for reporting this bug.
2007-02-12 Adam Piatyszek <ediap@users.sourceforge.net>
* doc/Makefile.am, Makefile.am: Changed the naming scheme of
snapshots to "itpp-VERSION_preYYYYMMDD"
------------------------------------------------------------------------------
2007-02-10 Adam Piatyszek <ediap@users.sourceforge.net>
* IT++ 3.10.9 released (SVN tag: release-3-10-9)
(merged rev. 929:930 from the itpp-3-10 branch)
* NEWS: Updated to include recent changes
(merged rev. 929:930 from the itpp-3-10 branch)
2007-02-09 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/gf2mat.cpp: Replaced it_info() with it_info_debug() to
disable information messages when using an optimized library
2007-02-08 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/mat.cpp, itpp/base/mat.h, itpp/base/vec.cpp,
itpp/base/vec.h, itpp/signal/freq_filt.cpp: Fixed improper changes
in template specializations
* doc/doxygen_html.cfg.in, itpp/base/binary.h,
itpp/base/circular_buffer.h, itpp/base/gf2mat.h,
itpp/base/itfile.h, itpp/base/mat.cpp, itpp/base/mat.h,
itpp/base/parser.h, itpp/base/vec.cpp, itpp/base/vec.h,
itpp/comm/channel.h, itpp/comm/channel_code.h,
itpp/comm/convcode.cpp, itpp/comm/convcode.h, itpp/comm/ldpc.cpp,
itpp/comm/ldpc.h, itpp/comm/llr.h, itpp/comm/modulator.h,
itpp/comm/modulator_nd.h, itpp/comm/reedsolomon.h,
itpp/comm/spread.h, itpp/comm/turbo.h, itpp/fixed/cfix.h,
itpp/fixed/cfixed.h, itpp/fixed/fix.h, itpp/fixed/fix_base.h,
itpp/fixed/fix_factory.h, itpp/fixed/fixed.h,
itpp/optim/newton_search.h, itpp/signal/fastica.cpp,
itpp/signal/filter.h, itpp/signal/filter_design.h,
itpp/signal/freq_filt.cpp, itpp/signal/freq_filt.h,
itpp/signal/poly.h, itpp/signal/sigfun.h,
itpp/signal/transforms.cpp, itpp/stat/histogram.h: Fixed Doxygen
warnings by adding or correcting the documentation syntax
* doc/local/installation.doc: Updated CXXFLAGS description
* configure.ac, Makefile.am: "upload" target removed
* itpp/comm/rec_syst_conv_code.cpp,
itpp/comm/rec_syst_conv_code.h: "INF" preprocessor definition
replaced with a private constant "infinity"
* itpp/comm/bch.h, itpp/comm/channel_code.h, itpp/comm/convcode.h,
itpp/comm/egolay.h, itpp/comm/hammcode.h, itpp/comm/ldpc.h,
itpp/comm/punct_convcode.h, itpp/comm/reedsolomon.h,
itpp/comm/turbo.h: Changed get_rate() functions into const ones
2007-02-07 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/fastmath.h, itpp/base/gf2mat.h, itpp/base/ittypes.h,
itpp/base/mat.h, itpp/base/matfunc.h, itpp/base/math/elem_math.h,
itpp/base/math/error.cpp, itpp/base/math/integration.cpp,
itpp/base/operators.h, itpp/base/parser.h, itpp/base/smat.h,
itpp/base/svec.h, itpp/base/vec.h: Fixed Doxygen warnings by
adding or correcting the documentation
2007-02-06 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/itassert.h: Added missing documentation for macros
* itpp/base/converters.cpp, itpp/base/converters.h: Fixed Doxygen
errors and improved the implementation of to_cvec() and to_cmat()
functions
2007-02-05 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/array.h, itpp/base/binfile.cpp, itpp/base/binfile.h,
itpp/base/converters.cpp: Documentation fixes related to Doxygen
warnings and errors
* itpp/base/algebra/ls_solve.h: Fixed improper declarations of
ls_solve_od() and ls_solve_ud() methods
* itpp/base/itfile.h: Replaced "uint32_t" with "unsigned int",
which is a workaroud for the compilation problem under Cygwin.
2007-02-04 Erik G. Larsson <erik_g_larsson@users.sourceforge.net>
* itpp/comm/ldpc.cpp, tests/ldpc_test.ref: Updated information
messages
2007-02-04 Adam Piatyszek <ediap@users.sourceforge.net>
* tests/Makefile.am: Redirecting stderr to /dev/null during "make
check"
* itpp/base/itassert.h, itpp/base/itassert.cpp: Added
it_info_no_endl() and it_info_no_endl_debug() macros, which does
not append "std::endl" at the end of message.
2007-02-03 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/algebra/ls_solve.cpp: Minor fixes to prevent "unused
variable" warnings when NDEBUG flag is defined.
2007-02-03 Erik G. Larsson <erik_g_larsson@users.sourceforge.net>
* itpp/base/gf2mat.cpp, itpp/comm/llr.cpp, itpp/comm/ldpc.cpp:
Updated some assertions and information messages
2007-02-02 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/algebra/det.cpp, itpp/base/algebra/eigen.cpp,
itpp/base/algebra/inv.cpp, itpp/base/algebra/ls_solve.cpp,
itpp/base/algebra/lu.cpp, itpp/base/algebra/schur.cpp,
itpp/base/array.h, itpp/base/binary.h,
itpp/base/circular_buffer.h, itpp/base/converters.cpp,
itpp/base/gf2mat.cpp, itpp/base/gf2mat.h, itpp/base/mat.cpp,
itpp/base/matfunc.h, itpp/base/mat.h, itpp/base/operators.cpp,
itpp/base/smat.h, itpp/base/specmat.cpp, itpp/base/stack.h,
itpp/base/svec.h, itpp/base/vec.h, itpp/comm/commfunc.cpp,
itpp/comm/convcode.cpp, itpp/comm/galois.h, itpp/comm/ldpc.cpp,
itpp/comm/ldpc.h, itpp/comm/llr.cpp, itpp/comm/llr.h,
itpp/comm/modulator.cpp, itpp/comm/modulator.h,
itpp/fixed/cfix.cpp, itpp/fixed/cfix.h, itpp/fixed/fix_base.cpp,
itpp/fixed/fix.cpp, itpp/fixed/fix_functions.h, itpp/fixed/fix.h,
itpp/fixed/fix_operators.cpp, itpp/signal/resampling.h,
itpp/signal/sigfun.cpp, itpp/signal/transforms.cpp,
itpp/srccode/audiofile.cpp, itpp/srccode/pnm.cpp:
Mass replace of the depreceted it_assert0() and it_assert1()
macros with it_assert_debug()
* itpp/base/itassert.h: Fixed a small bug in the it_info_debug()
macro definition. Documentation improved.
2007-02-01 Adam Piatyszek <ediap@users.sourceforge.net>
* configure.ac, itpp/base/itassert.cpp, itpp/base/itassert.h:
Improved and reorganised existing error handling macros:
- using standard NDEBUG flag for controlling it_assert_debug()
macro, which replaced a three-level ASSERT_LEVEL definition
- for backward compatibility it_assert0() and it_assert1() now
behaves the same as it_assert_debug()
- added new it_info() and it_info_debug() macros for printing
information messages; it_info_debug() is also dependent on the
NDEBUG compile-time definition
- added it_error_msg_style() function for changing the format of
it_assert(), it_error() and it_warning() messages; if style is set
to Minimum, file name and line number of the error/assert/warning
macro is not printed
2007-02-01 Erik G. Larsson <erik_g_larsson@users.sourceforge.net>
* itpp/comm/ldpc.cpp: Fix in the standard constructor
2007-01-31 Erik G. Larsson <erik_g_larsson@users.sourceforge.net>
* itpp/comm/ldpc.h, itpp/comm/ldpc.cpp: Added some initialization
functions for completeness of the class hierarchy
2007-01-31 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/comm/punct_convcode.h: Protected inheritance of the
Convolutional_Code class replaced with a public one. This fixes
the problem of assigning the Punctured_Convolutional_Code class
object to a pointer of Channel_Code class.
* itpp/base/gf2mat.cpp: Fixed improper reading and writing
functions of the alist file format
2007-01-30 Adam Piatyszek <ediap@users.sourceforge.net>
* configure.ac, extras/itpp-config.t2t, extras/txt2man.sh,
itpp-config.1.in, Makefile.am: Added man page for itpp-config
script
2007-01-29 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/comm/ldpc.cpp: Replaced long with int, since long and int
have the same sizes on most IT++ supported architectures. If there
is a need for 64-bit integers, (u)int64_t type should be used
instead of long.
2007-01-28 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/random.cpp, itpp/base/random.h: Replaced "unsigned
long" type of "seed" argument of function reset() with "unsigned
int", because the rest of Random_Generator implementation uses
"unsigned int" type.
* itpp/base/itfile.cpp, itpp/comm/convcode.cpp,
itpp/comm/convcode.h, itpp/comm/interleave.h,
itpp/comm/punct_convcode.cpp, itpp/comm/punct_convcode.h,
itpp/comm/turbo.cpp, itpp/comm/turbo.h, itpp/signal/resampling.h,
itpp/srccode/gmm.cpp: Replaced long with int, since long and int
have the same sizes on most IT++ supported architectures. If there
is a need for 64-bit integers, due to portability
reasons (u)int64_t type should be used instead of long.
2007-01-27 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/bessel/hyperg.cpp, itpp/comm/modulator_nd.cpp: Fixes
to avoid warnings on possibly uninitialised variables
2007-01-26 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/math/elem_math.cpp, itpp/base/math/elem_math.h:
Replaced long with int in binom_i() and gcd() functions, since
long and int have the same sizes on most IT++ supported
architectures. If there is a need for 64-bit integers, (u)int64_t
type should be used instead of long.
2007-01-26 Erik G. Larsson <erik_g_larsson@users.sourceforge.net>
* itpp/comm/ldpc.cpp: Fix to avoid compiler warning
2007-01-26 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/binfile.cpp: Fixed a minor bug, which caused infinite
loop of itfile_test on Solaris SPARC with GCC 3.4.5
* itpp/base/mat.h, itpp/base/vec.h, itpp/comm/crc.cpp,
itpp/comm/crc.h: Added some missing const keywords, as suggested
in feature request [1644569]. Thanks to Christian for providing a
patch.
* itpp/base/binfile.cpp, itpp/base/binfile.h, itpp/base/itfile.h:
long type replaced with int or int64_t where necessary due to
portability issues. Fixed assumptions of type sizes in << and >>
operators replaced with size dependent read_endian() and
write_endian() templated functions. Minor cosmetic changes.
2007-01-25 Adam Piatyszek <ediap@users.sourceforge.net>
* configure.ac, itpp/base/ittypes.h, itpp/config_msvc.h,
win32/itpp.vcproj: Improved portability by adding checks for sizes
of short, int, long and "long long" integer types. Assuming that
short has always 2 bytes and int has always 4 bytes. Instead of
long or "long long" types, an int64_t type should be used, which
has always 8 bytes on both 32- and 64-bit platforms. MSVC++
project file and configuration header updated accordingly.
2007-01-24 Adam Piatyszek <ediap@users.sourceforge.net>
* AUTHORS, doc/local/authors.doc: Added Kumar Appaiah (Debian
packages maintainer) to the list of authors
* itpp/base/sort.h: Minor documentation improvements
* tests/Makefile.am: Added missing "tests/itfile_test_data.it"
file
* itpp/base/sources.mk: Added missing "ittp/base/ittypes.h" file
2007-01-23 Adam Piatyszek <ediap@users.sourceforge.net>
* configure.ac: IT++ version incremented
2007-01-23 Erik G. Larsson <erik_g_larsson@users.sourceforge.net>
* itpp/comm/modulator_nd.cpp: Fixed an incorrect assertion check
2007-01-23 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/itfile.cpp, itpp/base/itfile.h, itpp/base/ittypes.h,
itpp/fixed/fix_base.cpp, itpp/fixed/fix_base.h, itpp/itbase.h:
Improved inter-platform portability by using fixed-length integer
types where needed.
* itpp/base/binfile.cpp, itpp/base/binfile.h: Documentation
improvements and minor cosmetic changes
* tests/itfile_test.cpp, tests/itfile_test_data.it,
tests/itfile_test.ref: Improved it_file tests
2007-01-22 Adam Piatyszek <ediap@users.sourceforge.net>
* tests/itfile_test.cpp, tests/itfile_test_data.it,
tests/itfile_test.ref, tests/Makefile.am: Added test file for a
binary it_file format. It is supposed to detect wrong endianness
detection on various platforms.
(merged rev. 860:863 from the itpp-3-10 branch)
* itpp/comm/modulator.cpp: Fixed wrong assertion check in QAM
modulator set_M() function
------------------------------------------------------------------------------
2007-01-21 Adam Piatyszek <ediap@users.sourceforge.net>
* IT++ 3.99.0 released (SVN tag: release-3-99-0)
* Makefile.am: Added missing files to EXTRA_DIST
2007-01-20 Adam Piatyszek <ediap@users.sourceforge.net>
* ChangeLog-2006: ChangeLog entries from year 2006 moved to a
separate file
* doc/local/verification.doc: Updated to include tests under
Solaris SunOS with ATLAS
2007-01-19 Adam Piatyszek <ediap@users.sourceforge.net>
* doc/local/verification.doc: Updated to include recent tests
under Cygwin/Windows XP and Solaris SunOS
2007-01-18 Adam Piatyszek <ediap@users.sourceforge.net>
* doc/local/verification.doc: Updated to include recent test under
Gentoo Linux with Intel MKL
2007-01-18 Erik G. Larsson <erik_g_larsson@users.sourceforge.net>
* itpp/comm/ldpc.h, itpp/comm/ldpc.cpp: Changed one function name
and updated documentation
* doc/local/verification.doc: Updated
2007-01-17 Adam Piatyszek <ediap@users.sourceforge.net>
* doc/local/verification.doc: Updated to reflect recent tests
* NEWS: Consistancy and language improved. Thanks to Erik for his
valuable comments.
2007-01-17 Erik G. Larsson <erik_g_larsson@users.sourceforge.net>
* itpp/base/gf2mat.h, itpp/base/gf2mat.cpp, itpp/comm/ldpc.cpp:
Changed int to bool and vice versa for code portability
2007-01-17 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/random.h: Fixed bug [1635988] by implementing the
twist() function in a similar way as the original TWIST() macro
from Takuji Nishimura's and Makoto Matsumoto's C sources.
* win32/itpp.vcproj: Updated to reflect recent changes in source
files' locations
2007-01-16 Erik G. Larsson <erik_g_larsson@users.sourceforge.net>
* itpp/comm/ldpc.cpp, tests/ldpc_test.ref: Changed rand() to
randi() for code portability and to use it++ routines for random
numbers
* itpp/comm/llr.h: C convention for INT_MAX changed to C++
2007-01-16 Adam Piatyszek <ediap@users.sourceforge.net>
* tests/Makefile.am: Fixed non-portable parts of the Makefile
reported by the latest stable automake 1.10
2007-01-14 Adam Piatyszek <ediap@users.sourceforge.net>
* NEWS, NEWS-3.10: NEWS file with 3.10 series information copied
to NEWS-3.10. Preliminary information for 3.99.x
development releases prepared in NEWS.
------------------------------------------------------------------------------
2007-01-14 Adam Piatyszek <ediap@users.sourceforge.net>
* IT++ 3.10.8 released (SVN tag: release-3-10-8)
(merged rev. 828:830 from the itpp-3-10 branch)
2007-01-13 Adam Piatyszek <ediap@users.sourceforge.net>
* NEWS: Updated to reflect recent changes
(merged rev. 828:830 from the itpp-3-10 branch)
* itpp/comm/ldpc.cpp, itpp/comm/ldpc.h, tests/channel_test.cpp,
tests/ldpc_test.cpp: Added missing SVN properties
2007-01-13 Erik G. Larsson <erik_g_larsson@users.sourceforge.net>
* itpp/comm/ldpc.cpp, itpp/comm/ldpc.h, itpp/comm/modulator_nd.h,
itpp/comm/modulator_nd.cpp: Code cleanup, removing local
declarations, etc.
2007-01-11 Erik G. Larsson <erik_g_larsson@users.sourceforge.net>
* itpp/comm/ldpc.cpp, tests/ldpc_test.ref: Minor enhancements
* itpp/base/specmat.h, itpp/base/specmat.cpp: Added function
"zigzag_space()"
2007-01-09 Conrad Sanderson <conrad_s@users.sourceforge.net>
* itpp/base/itassert.cpp: Changed "Assertation" to "Assertion"
2007-01-07 Erik G. Larsson <erik_g_larsson@users.sourceforge.net>
* AUTHORS, doc/local/authors.doc, itpp/comm/ldpc.h,
itpp/comm/ldpc.cpp, itpp/itcomm.h, itpp/tests/ldpc_test.cpp,
itpp/tests/ldpc_test.ref, itpp/comm/sources.mk, tests/Makefile.am,
doc/tutorial/tutorial.doc, doc/tutorial/sources.mk,
doc/tutorial/src/sources.mk, doc/tutorial/ldpc_gen_codes.doc,
doc/tutorial/ldpc_bersim_awgn.doc,
doc/tutorial/src/ldpc_gen_codes.cpp,
doc/tutorial/src/ldpc_bersim_awgn.cpp: Added first version of the
LDPC codec class
* itpp/comm/llr.h, itpp/comm/llr.cpp, itpp/tests/llr_test.ref,
itpp/tests/modulator_nd_test.ref: Minor updates
* itpp/base/smat.h: Minor efficiency improvements
* itpp/comm/error_counters.h: Added function get_total_bits() and
get_total_blocks()
* itpp/base/gf2mat.h: Documentation updates
* TODO: Updated
2007-01-05 Erik G. Larsson <erik_g_larsson@users.sourceforge.net>
* itpp/base/gf2mat.h: Clarification in documentation
2007-01-04 Erik G. Larsson <erik_g_larsson@users.sourceforge.net>
* itpp/base/gf2mat.h, itpp/base/gf2mat.cpp, tests/Makefile.am:
Code readability and documentation improvements. Minor fixes.
2007-01-03 Erik G. Larsson <erik_g_larsson@users.sourceforge.net>
* itpp/base/gf2mat.h, itpp/base/gf2mat.cpp, tests/gf2mat_test.cpp,
tests/gf2mat_test.ref: Added support for "alist" representation
(based on code provided by Adam P) of sparse GF(2)
matrices. Implemented minor improvements.
* doc/tutorial/src/mimoconv.cpp: Minor improvements
* itpp/base/random.h: Minor documentation improvement
2007-01-02 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/mat.h: Implemented feature request [1216416]. Multiply
operator `*' of a vector `v' and single-row matrix `m' now works
for any size of vector `v'. This change is backward compatible
with the previous implementation.
* itpp/base/array.h: Fixed a small bug caused by a missing "const"
keyword in in_range() inline function.
2007-01-01 Adam Piatyszek <ediap@users.sourceforge.net>
* itpp/base/array.h, tests/array_test.cpp, tests/array_test.ref,
TODO: Added missing left(), right() and mid() methods to the Array
class.