Group :: System/Libraries
RPM: liblevmar
Main Changelog Spec Patches Sources Download Gear Bugs and FR Repocop
# BEGIN SourceDeps(oneline):
BuildRequires: gcc-c++
# END SourceDeps(oneline)
BuildRequires: libblas-devel
%add_optflags %optflags_shared
%define oldname levmar
# see https://bugzilla.altlinux.org/show_bug.cgi?id=10382
%define _localstatedir %{_var}
# SOlib major and minor version
%global major 2
%global minor 2
Name: liblevmar
Version: 2.5
Release: alt1_14
Summary: Levenberg-Marquardt nonlinear least squares algorithm
URL: http://www.ics.forth.gr/~lourakis/levmar/
Source0: http://www.ics.forth.gr/~lourakis/levmar/levmar-%{version}.tgz
# Patch to fix compilation of the shared library and compile the demo program
# Sent to levmar author on 28-Jan-2010
Patch0: levmar-shared.patch
License: GPLv2+
Group: System/Libraries
BuildRequires: dos2unix
BuildRequires: liblapack-devel
Provides: levmar = %{version}-%{release}
%description
levmar is a native ANSI C implementation of the Levenberg-Marquardt
optimization algorithm. Both unconstrained and constrained (under linear
equations, inequality and box constraints) Levenberg-Marquardt variants are
included. The LM algorithm is an iterative technique that finds a local
minimum of a function that is expressed as the sum of squares of nonlinear
functions. It has become a standard technique for nonlinear least-squares
problems and can be thought of as a combination of steepest descent and the
Gauss-Newton method. When the current solution is far from the correct on,
the algorithm behaves like a steepest descent method: slow, but guaranteed
to converge. When the current solution is close to the correct solution, it
becomes a Gauss-Newton method.
%package devel
Summary: Development files for levmar library, and demo program
Group: Development/Other
Requires: liblevmar = %{version}-%{release}
Provides: levmar-devel = %{version}-%{release}
%description devel
Development files for the levmar library, and demo program.
%prep
%setup -n %{oldname}-%{version} -q
%patch -P 0 -p1 -b .shared
dos2unix -k README.txt
%build
mkdir sobj
make CFLAGS="%{optflags} -funroll-loops -fPIC" %{?_smp_mflags} -f Makefile.so
%install
install -D -p -m 755 sobj/liblevmar.so.%{major}.%{minor} %{buildroot}%{_libdir}/liblevmar.so.%{major}.%{minor}
install -D -p -m 644 levmar.h %{buildroot}%{_includedir}/levmar.h
install -D -p -m 755 lmdemo %{buildroot}%{_bindir}/lmdemo
ln -s liblevmar.so.%{major}.%{minor} %{buildroot}%{_libdir}/liblevmar.so.%{major}
ln -s liblevmar.so.%{major}.%{minor} %{buildroot}%{_libdir}/liblevmar.so
%files
%doc README.txt LICENSE
%{_libdir}/liblevmar.so.%{major}.%{minor}
%{_libdir}/liblevmar.so.%{major}
%files devel
%{_includedir}/levmar.h
%{_libdir}/liblevmar.so
%{_bindir}/lmdemo
%changelog
…
Full changelog you can see here
BuildRequires: gcc-c++
# END SourceDeps(oneline)
BuildRequires: libblas-devel
%add_optflags %optflags_shared
%define oldname levmar
# see https://bugzilla.altlinux.org/show_bug.cgi?id=10382
%define _localstatedir %{_var}
# SOlib major and minor version
%global major 2
%global minor 2
Name: liblevmar
Version: 2.5
Release: alt1_14
Summary: Levenberg-Marquardt nonlinear least squares algorithm
URL: http://www.ics.forth.gr/~lourakis/levmar/
Source0: http://www.ics.forth.gr/~lourakis/levmar/levmar-%{version}.tgz
# Patch to fix compilation of the shared library and compile the demo program
# Sent to levmar author on 28-Jan-2010
Patch0: levmar-shared.patch
License: GPLv2+
Group: System/Libraries
BuildRequires: dos2unix
BuildRequires: liblapack-devel
Provides: levmar = %{version}-%{release}
%description
levmar is a native ANSI C implementation of the Levenberg-Marquardt
optimization algorithm. Both unconstrained and constrained (under linear
equations, inequality and box constraints) Levenberg-Marquardt variants are
included. The LM algorithm is an iterative technique that finds a local
minimum of a function that is expressed as the sum of squares of nonlinear
functions. It has become a standard technique for nonlinear least-squares
problems and can be thought of as a combination of steepest descent and the
Gauss-Newton method. When the current solution is far from the correct on,
the algorithm behaves like a steepest descent method: slow, but guaranteed
to converge. When the current solution is close to the correct solution, it
becomes a Gauss-Newton method.
%package devel
Summary: Development files for levmar library, and demo program
Group: Development/Other
Requires: liblevmar = %{version}-%{release}
Provides: levmar-devel = %{version}-%{release}
%description devel
Development files for the levmar library, and demo program.
%prep
%setup -n %{oldname}-%{version} -q
%patch -P 0 -p1 -b .shared
dos2unix -k README.txt
%build
mkdir sobj
make CFLAGS="%{optflags} -funroll-loops -fPIC" %{?_smp_mflags} -f Makefile.so
%install
install -D -p -m 755 sobj/liblevmar.so.%{major}.%{minor} %{buildroot}%{_libdir}/liblevmar.so.%{major}.%{minor}
install -D -p -m 644 levmar.h %{buildroot}%{_includedir}/levmar.h
install -D -p -m 755 lmdemo %{buildroot}%{_bindir}/lmdemo
ln -s liblevmar.so.%{major}.%{minor} %{buildroot}%{_libdir}/liblevmar.so.%{major}
ln -s liblevmar.so.%{major}.%{minor} %{buildroot}%{_libdir}/liblevmar.so
%files
%doc README.txt LICENSE
%{_libdir}/liblevmar.so.%{major}.%{minor}
%{_libdir}/liblevmar.so.%{major}
%files devel
%{_includedir}/levmar.h
%{_libdir}/liblevmar.so
%{_bindir}/lmdemo
%changelog
…
Full changelog you can see here