[PATCH] Update numactl

[PATCH] Update numactl

[Andreas Enge]

[-- Attachment #1: Type: text/plain, Size: 317 bytes --]

The attached patch updates numactl. My secret hope is that it will now build
on arm (together with potentially 15 dependent packages). I successfully
compiled hwloc with it.

Is the comment about the test suite still valid? I tried to run the tests,
they require taskset (which is available in util-linux).

Andreas


[-- Attachment #2: 0001-gnu-numactl-Update-to-2.0.10.patch --]
[-- Type: text/plain, Size: 2967 bytes --]

From 9c620c228a99335798212f37d3788a9bfa353a78 Mon Sep 17 00:00:00 2001
From: Andreas Enge <andreas@enge.fr>
Date: Sat, 17 Oct 2015 21:42:42 +0200
Subject: [PATCH] gnu: numactl: Update to 2.0.10.

* gnu/packages/linux.scm (numactl)[source]: Update to 2.0.10.
  [arguments]: Adapt phases to use the autotools machinery.
  [native-inputs]: Add autotools packages.
---
 gnu/packages/linux.scm | 38 +++++++++++++-------------------------
 1 file changed, 13 insertions(+), 25 deletions(-)

diff --git a/gnu/packages/linux.scm b/gnu/packages/linux.scm
index 4ee4234..d4b9229 100644
--- a/gnu/packages/linux.scm
+++ b/gnu/packages/linux.scm
@@ -1389,7 +1389,7 @@ file system is as easy as logging into the server with an SSH client.")
 (define-public numactl
   (package
     (name "numactl")
-    (version "2.0.9")
+    (version "2.0.10")
     (source (origin
               (method url-fetch)
               (uri (string-append
@@ -1398,35 +1398,23 @@ file system is as easy as logging into the server with an SSH client.")
                     ".tar.gz"))
               (sha256
                (base32
-                "073myxlyyhgxh1w3r757ajixb7s2k69czc3r0g12c3scq7k3784w"))))
+                "0qfv2ks6d3gm0mw5sj4cbhsd7cbsb7qm58xvchl2wfzifkzcinnv"))))
     (build-system gnu-build-system)
+    (native-inputs
+     `(("autoconf" ,autoconf)
+       ("automake" ,automake)
+       ("libtool" ,libtool)))
     (arguments
-     '(#:phases (alist-replace
-                 'configure
-                 (lambda* (#:key outputs #:allow-other-keys)
-                   ;; There's no 'configure' script, just a raw makefile.
-                   (substitute* "Makefile"
-                     (("^prefix := .*$")
-                      (string-append "prefix := " (assoc-ref outputs "out")
-                                     "\n"))
-                     (("^libdir := .*$")
-                      ;; By default the thing tries to install under
-                      ;; $prefix/lib64 when on a 64-bit platform.
-                      (string-append "libdir := $(prefix)/lib\n"))))
-                 %standard-phases)
-
-       #:make-flags (list
-                     ;; By default the thing tries to use 'cc'.
-                     "CC=gcc"
-
-                     ;; Make sure programs have an RPATH so they can find
-                     ;; libnuma.so.
-                     (string-append "LDLIBS=-Wl,-rpath="
-                                    (assoc-ref %outputs "out") "/lib"))
-
+     '(#:phases
+       (modify-phases %standard-phases
+         (add-after
+          'unpack 'autogen
+          (lambda _
+            (zero? (system* "sh" "autogen.sh")))))
        ;; There's a 'test' target, but it requires NUMA support in the kernel
        ;; to run, which we can't assume to have.
        #:tests? #f))
+
     (home-page "http://oss.sgi.com/projects/libnuma/")
     (synopsis "Tools for non-uniform memory access (NUMA) machines")
     (description
-- 
2.5.0

Re: [PATCH] Update numactl

[Andreas Enge]

On Sat, Oct 17, 2015 at 09:49:17PM +0200, Andreas Enge wrote:
> The attached patch updates numactl. My secret hope is that it will now build
> on arm (together with potentially 15 dependent packages). I successfully
> compiled hwloc with it.

It does not build on arm. On debian, hwloc does not depend on libnuma
on the armel and armhf architectures:
   https://packages.debian.org/jessie/libhwloc5
So maybe we could make it an input only for non-arm architectures?

Andreas

Re: [PATCH] Update numactl

[Ludovic Courtès]

Andreas Enge <andreas@enge.fr> skribis:

> Is the comment about the test suite still valid? I tried to run the tests,
> they require taskset (which is available in util-linux).

I don’t know; I wouldn’t be surprised if it’s still valid.

> From 9c620c228a99335798212f37d3788a9bfa353a78 Mon Sep 17 00:00:00 2001
> From: Andreas Enge <andreas@enge.fr>
> Date: Sat, 17 Oct 2015 21:42:42 +0200
> Subject: [PATCH] gnu: numactl: Update to 2.0.10.
>
> * gnu/packages/linux.scm (numactl)[source]: Update to 2.0.10.
>   [arguments]: Adapt phases to use the autotools machinery.
>   [native-inputs]: Add autotools packages.

LGTM.

> +    (native-inputs
> +     `(("autoconf" ,autoconf)
> +       ("automake" ,automake)
> +       ("libtool" ,libtool)))

Pfff…  Could you email upstream and ask them to run ‘make dist’ next
time?  It’s really weird: They used to do it, and it looks as if they
had forgotten.  :-/

Thanks,
Ludo’.

Re: [PATCH] Update numactl

[Andreas Enge]

On Sun, Oct 18, 2015 at 07:35:44PM +0200, Ludovic Courtès wrote:
> Pfff…  Could you email upstream and ask them to run ‘make dist’ next
> time?  It’s really weird: They used to do it, and it looks as if they
> had forgotten.  :-/

No, they did not use the autotools and had just a Makefile.

Andreas

Re: [PATCH] Update numactl

[Andreas Enge]

On Sat, Oct 17, 2015 at 11:18:25PM +0200, Andreas Enge wrote:
> It does not build on arm. On debian, hwloc does not depend on libnuma
> on the armel and armhf architectures:
>    https://packages.debian.org/jessie/libhwloc5
> So maybe we could make it an input only for non-arm architectures?

What do you think of this suggestion?

Andreas

Re: [PATCH] Update numactl

[Ludovic Courtès]

Andreas Enge <andreas@enge.fr> skribis:

> On Sat, Oct 17, 2015 at 11:18:25PM +0200, Andreas Enge wrote:
>> It does not build on arm. On debian, hwloc does not depend on libnuma
>> on the armel and armhf architectures:
>>    https://packages.debian.org/jessie/libhwloc5
>> So maybe we could make it an input only for non-arm architectures?
>
> What do you think of this suggestion?

What’s the rationale for this special case?  That libnuma doesn’t
support ARM?

If that’s the case, the suggestion sounds good to me.

Ludo’.