Added mpfr.
authorRichard W.M. Jones <rjones@redhat.com>
Wed, 28 Jan 2009 17:41:57 +0000 (17:41 +0000)
committerRichard W.M. Jones <rjones@redhat.com>
Wed, 28 Jan 2009 17:41:57 +0000 (17:41 +0000)
.hgignore
mpfr/mingw32-mpfr.spec [new file with mode: 0644]

index 9102bd0..85413e6 100644 (file)
--- a/.hgignore
+++ b/.hgignore
@@ -81,6 +81,7 @@ libxml2/libxml2-2.6.32.tar.gz
 libxml2/libxml2-2.7.1.tar.gz
 libxml2/libxml2-2.7.2.tar.gz
 libxslt/libxslt-1.1.24.tar.gz
+mpfr/mpfr-2.3.2.tar.bz2
 nsis/nsis-2.19-src.tar.bz2
 nsis/nsis-2.39-src.tar.bz2
 nspr/nspr-4.7.2.tar.gz
diff --git a/mpfr/mingw32-mpfr.spec b/mpfr/mingw32-mpfr.spec
new file mode 100644 (file)
index 0000000..9f482b0
--- /dev/null
@@ -0,0 +1,100 @@
+%define __strip %{_mingw32_strip}
+%define __objdump %{_mingw32_objdump}
+%define _use_internal_dependency_generator 0
+%define __find_requires %{_mingw32_findrequires}
+%define __find_provides %{_mingw32_findprovides}
+
+# Define this to run tests (requires Wine, and won't work
+# inside mock or Koji).
+%define run_tests 0
+
+Name:           mingw32-mpfr
+Version:        2.3.2
+Release:        1%{?dist}
+Summary:        Windows port of multiple-precision floating-point computations
+
+License:        LGPLv2+ and GPLv2+ and GFDL
+Group:          Development/Libraries
+
+URL:            http://www.mpfr.org/
+Source0:        http://www.mpfr.org/mpfr-current/mpfr-%{version}.tar.bz2
+
+BuildRoot:      %{_tmppath}/%{name}-%{version}-%{release}-root-%(%{__id_u} -n)
+BuildArch:      noarch
+
+BuildRequires:  mingw32-filesystem >= 41
+BuildRequires:  mingw32-gcc
+BuildRequires:  mingw32-binutils
+
+BuildRequires:  mingw32-dlfcn
+BuildRequires:  mingw32-gmp >= 4.2.1
+
+# Native package requires these, but I can't see how they are used.
+#BuildRequires:  autoconf
+#BuildRequires:  libtool
+
+Requires:       mingw32-gmp >= 4.2.1
+
+
+%description
+The MPFR library is a C library for multiple-precision floating-point
+computations with "correct rounding". The MPFR is efficient and 
+also has a well-defined semantics. It copies the good ideas from the 
+ANSI/IEEE-754 standard for double-precision floating-point arithmetic 
+(53-bit mantissa). MPFR is based on the GMP multiple-precision library.
+
+This is the development package for Windows cross-compilation.
+
+
+%prep
+%setup -q -n mpfr-%{version}
+
+
+%build
+%{_mingw32_configure} \
+  --disable-assert \
+  --disable-static \
+  --enable-shared
+make %{?_smp_mflags}
+
+
+%check
+%if %run_tests
+# Bloody Wine ...  'make check' doesn't work because Wine no longer
+# looks on $PATH for libraries.  Instead we have to run the programs
+# by hand from the .libs directory.
+make %{?_smp_mflags} check
+pushd .libs
+for f in ../tests/.libs/*.exe; do
+  srcdir=$(pwd)/../tests/ wine $f
+done
+popd
+%endif
+
+
+%install
+rm -rf $RPM_BUILD_ROOT
+make DESTDIR=$RPM_BUILD_ROOT install
+
+# We don't want the documentation, since that is available
+# in the native Fedora package.
+rm -r $RPM_BUILD_ROOT%{_mingw32_infodir}
+
+
+%clean
+rm -rf $RPM_BUILD_ROOT
+
+
+%files
+%defattr(-,root,root)
+%doc COPYING COPYING.LIB README
+%{_mingw32_bindir}/libmpfr-1.dll
+%{_mingw32_libdir}/libmpfr.dll.a
+%{_mingw32_libdir}/libmpfr.la
+%{_mingw32_includedir}/mpf2mpfr.h
+%{_mingw32_includedir}/mpfr.h
+
+
+%changelog
+* Wed Jan 28 2009 Richard W.M. Jones <rjones@redhat.com> - 2.3.2-1
+- Initial RPM release (on behalf of Ralf Corsepius).