3
0
Fork 0
forked from mirrors/nixpkgs
nixpkgs/pkgs/development/libraries/hyperscan/default.nix
2021-11-17 13:35:26 +01:00

76 lines
2.5 KiB
Nix

{ lib, stdenv, fetchFromGitHub, cmake, ragel, python3
, coreutils, gnused, util-linux, fetchpatch
, boost
, withStatic ? false # build only shared libs by default, build static+shared if true
}:
# NOTICE: pkg-config, pcap and pcre intentionally omitted from build inputs
# pcap used only in examples, pkg-config used only to check for pcre
# which is fixed 8.41 version requirement (nixpkgs have 8.42+, and
# I not see any reason (for now) to backport 8.41.
stdenv.mkDerivation rec {
pname = "hyperscan";
version = "5.4.0";
src = fetchFromGitHub {
owner = "intel";
repo = pname;
sha256 = "sha256-AJAjaXVnGqIlMk+gb6lpTLUdZr8nxn2XSW4fj6j/cmk=";
rev = "v${version}";
};
outputs = [ "out" "dev" ];
buildInputs = [ boost ];
nativeBuildInputs = [
cmake ragel python3
# Consider simply using busybox for these
# Need at least: rev, sed, cut, nm
coreutils gnused util-linux
];
cmakeFlags = [
"-DFAT_RUNTIME=ON"
"-DBUILD_AVX512=ON"
]
++ lib.optional (withStatic) "-DBUILD_STATIC_AND_SHARED=ON"
++ lib.optional (!withStatic) "-DBUILD_SHARED_LIBS=ON";
patches = [
(fetchpatch {
# part of https://github.com/intel/hyperscan/pull/336
url = "https://github.com/intel/hyperscan/commit/e2c4010b1fc1272cab816ba543940b3586e68a0c.patch";
sha256 = "sha256-doVNwROL6MTcgOW8jBwGTnxe0zvxjawiob/g6AvXLak=";
})
];
postPatch = ''
sed -i '/examples/d' CMakeLists.txt
substituteInPlace libhs.pc.in \
--replace "libdir=@CMAKE_INSTALL_PREFIX@/@CMAKE_INSTALL_LIBDIR@" "libdir=@CMAKE_INSTALL_LIBDIR@" \
--replace "includedir=@CMAKE_INSTALL_PREFIX@/@CMAKE_INSTALL_INCLUDEDIR@" "includedir=@CMAKE_INSTALL_INCLUDEDIR@"
'';
meta = with lib; {
description = "High-performance multiple regex matching library";
longDescription = ''
Hyperscan is a high-performance multiple regex matching library.
It follows the regular expression syntax of the commonly-used
libpcre library, but is a standalone library with its own C API.
Hyperscan uses hybrid automata techniques to allow simultaneous
matching of large numbers (up to tens of thousands) of regular
expressions and for the matching of regular expressions across
streams of data.
Hyperscan is typically used in a DPI library stack.
'';
homepage = "https://www.hyperscan.io/";
maintainers = with maintainers; [ avnik ];
platforms = [ "x86_64-linux" ]; # can't find nm on darwin ; might build on aarch64 but untested
license = licenses.bsd3;
};
}