nixpkgs/pkgs/development/libraries/mpfr/default.nix
2023-08-22 12:45:13 +01:00

78 lines
2.3 KiB
Nix

{ lib
, stdenv
, fetchurl
, gmp
, writeScript
}:
# Note: this package is used for bootstrapping fetchurl, and thus
# cannot use fetchpatch! All mutable patches (generated by GitHub or
# cgit) that are needed here should be included directly in Nixpkgs as
# files.
stdenv.mkDerivation rec {
version = "4.2.1";
pname = "mpfr";
src = fetchurl {
urls = [
"https://www.mpfr.org/${pname}-${version}/${pname}-${version}.tar.xz"
"mirror://gnu/mpfr/${pname}-${version}.tar.xz"
];
hash = "sha256-J3gHNTpnJpeJlpRa8T5Sgp46vXqaW3+yeTiU4Y8fy7I=";
};
outputs = [ "out" "dev" "doc" "info" ];
strictDeps = true;
# mpfr.h requires gmp.h
propagatedBuildInputs = [ gmp ];
configureFlags = lib.optional stdenv.hostPlatform.isSunOS "--disable-thread-safe"
++ lib.optional stdenv.hostPlatform.is64bit "--with-pic"
++ lib.optionals stdenv.hostPlatform.isPower64 [
# Without this, the `tget_set_d128` test experiences a link
# error due to missing `__dpd_trunctdkf`.
"--disable-decimal-float"
];
doCheck = true; # not cross;
enableParallelBuilding = true;
passthru = {
updateScript = writeScript "update-mpfr" ''
#!/usr/bin/env nix-shell
#!nix-shell -i bash -p curl pcre common-updater-scripts
set -eu -o pipefail
# Expect the text in format of '<title>GNU MPFR version 4.1.1</title>'
new_version="$(curl -s https://www.mpfr.org/mpfr-current/ |
pcregrep -o1 '<title>GNU MPFR version ([0-9.]+)</title>')"
update-source-version ${pname} "$new_version"
'';
};
meta = {
homepage = "https://www.mpfr.org/";
description = "Library for multiple-precision floating-point arithmetic";
longDescription = ''
The GNU MPFR library is a C library for multiple-precision
floating-point computations with correct rounding. MPFR is
based on the GMP multiple-precision library.
The main goal of MPFR is to provide a library for
multiple-precision floating-point computation which is both
efficient and 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).
'';
license = lib.licenses.lgpl2Plus;
maintainers = [ ];
platforms = lib.platforms.all;
};
}