Provide nixpkgs.lib in bscpkgs outputs

Currently, we can use bscpkgs similarly to nixpkgs either through
the flake outputs or with import bscpkgs:

```nix
# currently supported:
bscpkgs.legacyPackages.x86_64-linux.hello
let pkgs = import bscpkgs { system = "x86_64-linux"; }; in pkgs.hello
```
The missing piece is nixpkgs.lib (not pkgs.lib, the system agnostic
one). The workaround is to do bscpkgs.inputs.nixpkgs.lib instead. We can
simplify this by forwarding the lib to our outputs.

This enables us to use bscpkgs as a drop-in
replacing the inputs to our flake from nixpkgs to bscpkgs.
(inputs.nixpkgs.url = "<*BSC*pkgs url>").


Reviewed-by: Rodrigo Arias Mallo <rodrigo.arias@bsc.es>
Tested-by: Aleix Boné <abonerib@bsc.es>

overlay.nix

@@ -18,7 +18,7 @@ let
     clangOmpss2NodesOmpv = callPackage ./pkgs/llvm-ompss2/default.nix { ompss2rt = final.nodes; openmp = final.openmpv; };
     clangOmpss2Unwrapped = callPackage ./pkgs/llvm-ompss2/clang.nix { };
     #extrae = callPackage ./pkgs/extrae/default.nix { }; # Broken and outdated
-    gpi-2 = callPackage ./pkgs/gpi-2/default.nix { };
+    #gpi-2 = callPackage ./pkgs/gpi-2/default.nix { }; # Broken: https://jungle.bsc.es/git/rarias/bscpkgs/issues/7
     intelPackages_2023 = callPackage ./pkgs/intel-oneapi/2023.nix { };
     jemallocNanos6 = callPackage ./pkgs/nanos6/jemalloc.nix { };
     lmbench = callPackage ./pkgs/lmbench/default.nix { };
@@ -44,7 +44,7 @@ let
     stdenvClangOmpss2Nanos6 = final.stdenv.override { cc = final.clangOmpss2Nanos6; allowedRequisites = null; };
     stdenvClangOmpss2Nodes = final.stdenv.override { cc = final.clangOmpss2Nodes; allowedRequisites = null; };
     stdenvClangOmpss2NodesOmpv = final.stdenv.override { cc = final.clangOmpss2NodesOmpv; allowedRequisites = null; };
-    tagaspi = callPackage ./pkgs/tagaspi/default.nix { };
+    #tagaspi = callPackage ./pkgs/tagaspi/default.nix { }; # Broken due gpi-2
     tampi = callPackage ./pkgs/tampi/default.nix { };
     wxparaver = callPackage ./pkgs/paraver/default.nix { };
   };

pkgs/gpi-2/default.nix

@@ -34,8 +34,6 @@ stdenv.mkDerivation rec {
 
   enableParallelBuilding = true;
 
-  patches = [ ./rdma-core.patch ./max-mem.patch ];
-
   preConfigure = ''
     patchShebangs autogen.sh
     ./autogen.sh

pkgs/gpi-2/max-mem.patch

@@ -1,10 +0,0 @@
---- a/tests/tests/segments/max_mem.c	2025-09-12 13:30:53.449353591 +0200
-+++ b/tests/tests/segments/max_mem.c	2025-09-12 13:33:49.750352401 +0200
-@@ -1,5 +1,7 @@
- #include <test_utils.h>
- 
-+gaspi_size_t gaspi_get_system_mem (void);
-+
- /* Test allocates 45% of system memory and creates a segment that
-    large or if several ranks per node exist, divided among that
-    number */

pkgs/gpi-2/rdma-core.patch

@@ -1,12 +0,0 @@
---- a/src/devices/ib/GPI2_IB.h	2025-09-12 13:25:31.564181121 +0200
-+++ b/src/devices/ib/GPI2_IB.h	2025-09-12 13:24:49.105422150 +0200
-@@ -26,6 +26,9 @@ along with GPI-2. If not, see <http://ww
- 
- #include "GPI2_Dev.h"
- 
-+/* Missing prototype as driver.h is now private */
-+int ibv_read_sysfs_file(const char *dir, const char *file, char *buf, size_t size);
-+
- #define GASPI_GID_INDEX   (0)
- #define PORT_LINK_UP      (5)
- #define MAX_INLINE_BYTES  (128)