depot/third_party/nixpkgs/pkgs/tools/compression/lz4/default.nix

75 lines
2.3 KiB
Nix
Raw Normal View History

{ lib, stdenv, fetchFromGitHub, valgrind, fetchpatch
, enableStatic ? stdenv.hostPlatform.isStatic
, enableShared ? !stdenv.hostPlatform.isStatic
}:
stdenv.mkDerivation rec {
pname = "lz4";
version = "1.9.3";
src = fetchFromGitHub {
sha256 = "1w02kazh1fps3sji2sn89fz862j1199c5ajrqcgl1bnlxj09kcbz";
rev = "v${version}";
repo = pname;
owner = pname;
};
patches = [
(fetchpatch { # https://github.com/lz4/lz4/pull/972
name = "CVE-2021-3520.patch";
url = "https://github.com/lz4/lz4/commit/8301a21773ef61656225e264f4f06ae14462bca7.patch";
sha256 = "0r1cwpqdkdc8im0pf2r5jp7mwwn69xcw405rrk7rc0mpjcp5ydfk";
})
(fetchpatch { # https://github.com/lz4/lz4/pull/973
name = "avoid-null-pointer-dereference.patch";
url = "https://github.com/lz4/lz4/commit/29a6a1f4941e7243241fe00d6c13b749fd6b60c2.patch";
sha256 = "0v5yl5hd3qrfm3xm7m06j4b21qwllb4cqkjn2az7x1vnzqgpf8y7";
})
];
# TODO(@Ericson2314): Separate binaries and libraries
outputs = [ "bin" "out" "dev" ];
buildInputs = lib.optional doCheck valgrind;
enableParallelBuilding = true;
makeFlags = [
"PREFIX=$(out)"
"INCLUDEDIR=$(dev)/include"
"BUILD_STATIC=${if enableStatic then "yes" else "no"}"
"BUILD_SHARED=${if enableShared then "yes" else "no"}"
"WINDRES:=${stdenv.cc.bintools.targetPrefix}windres"
]
# TODO make full dictionary
++ lib.optional stdenv.hostPlatform.isMinGW "TARGET_OS=MINGW"
;
doCheck = false; # tests take a very long time
checkTarget = "test";
# TODO(@Ericson2314): Make resusable setup hook for this issue on Windows.
postInstall =
lib.optionalString stdenv.hostPlatform.isWindows ''
mv $out/bin/*.dll $out/lib
ln -s $out/lib/*.dll
''
+ ''
moveToOutput bin "$bin"
'';
meta = with lib; {
description = "Extremely fast compression algorithm";
longDescription = ''
Very fast lossless compression algorithm, providing compression speed
at 400 MB/s per core, with near-linear scalability for multi-threaded
applications. It also features an extremely fast decoder, with speed in
multiple GB/s per core, typically reaching RAM speed limits on
multi-core systems.
'';
homepage = "https://lz4.github.io/lz4/";
license = with licenses; [ bsd2 gpl2Plus ];
platforms = platforms.all;
};
}