New package: FPLLL

New package: FPLLL

[Andreas Enge]

[-- Attachment #1: Type: text/plain, Size: 823 bytes --]

Here is a recipe:

(define-public fplll
  (package
   (name "fplll")
   (version "4.0.1")
   (source (origin
            (method url-fetch)
            (uri (string-append
                  "http://perso.ens-lyon.fr/damien.stehle/fplll/libfplll-" 
version ".tar.gz"))
            (sha256 (base32
                     "122bpqdlikshhd7nmq0l5qfc0agyk7x21gvplv1l9hb77l8cy9rw"))))
   (build-system gnu-build-system)
   (inputs `(("gmp" ,gmp)
             ("mpfr" ,mpfr)))
   (synopsis "fplll, a library for LLL-reduction of euclidean lattices")
   (description
    "fplll LLL-reduces euclidean lattices. Since version 3, it can also 
solve the shortest vector problem.")
   (license "LGPLv2.1")
   (home-page "http://perso.ens-lyon.fr/damien.stehle/fplll/")))

It could go into multiprecision.scm, or get its own file.

Andreas

[-- Attachment #2: Type: text/html, Size: 4841 bytes --]

Re: New package: FPLLL

[Ludovic Courtès]

Hi Andreas!

Andreas Enge <andreas@enge.fr> skribis:

> Here is a recipe:

Thanks!  I could add it, but in the long run I’d prefer patches as
produced by ‘git format-patch’, which has the advantage of being readily
applicable, and preserving authorship.  Would it be possible for you?

>    (license "LGPLv2.1")

It seems to be LGPLv2.1+ rather.

> It could go into multiprecision.scm, or get its own file.

Yes, multiprecision.scm would seem like a good option to me.  WDYT?

Thanks,
Ludo’.

Re: New package: FPLLL

[Nikita Karetnikov]

>>    (license "LGPLv2.1")

> It seems to be LGPLv2.1+ rather.

Please use (guix licenses) instead.

Nikita

Re: New package: FPLLL

[Andreas Enge]

[-- Attachment #1: Type: text/plain, Size: 1334 bytes --]

Am Mittwoch, 12. Dezember 2012 schrieb Ludovic Courtès:
> >    (license "LGPLv2.1")
> It seems to be LGPLv2.1+ rather.

The file README.html states:
"fplll is distributed under GNU Lesser General
Public License (v2.1) as published by the Free Software Foundation."

I checked two source files, which say
"or (at your option) any later version".
I suppose this is authoritative.

But I will ask the authors to clarify.

Am Mittwoch, 12. Dezember 2012 schrieb Nikita Karetnikov:
> Please use (guix licenses) instead.

Could you provide an example on how to do this? As far as I could see, all
files in distro/packages still use strings for the licenses.

Am Mittwoch, 12. Dezember 2012 schrieb Ludovic Courtès:
> > It could go into multiprecision.scm, or get its own file.
> Yes, multiprecision.scm would seem like a good option to me.  WDYT?

It depends on whether you would rather go for one file per package (maybe
to be organised in subdirectories), or one file per group of packages of a
similar category.

Even in the latter case, it looks like multiprecision.scm is not a very
good place, since it regroups the dependencies of gcc. Maybe algebra.scm?
It could also contain mpfrcx, gmp-ecc and maybe pari (which does not
exactly follow gnu build standards, so I would need some help).

Andreas

[-- Attachment #2: Type: text/html, Size: 6793 bytes --]

Re: New package: FPLLL

[Nikita Karetnikov]

>> Please use (guix licenses) instead.

> Could you provide an example on how to do this?

Just add the following to the corresponding sections:

  #:use-module (guix licenses)

  (license lgpl2.1+)

> As far as I could see, all files in distro/packages still use
> strings for the licenses.

I'll send a patch soon. (I'm trying to be careful because it's easy to
break something when you change so many things.)

Nikita

Re: New package: FPLLL

[Andreas Enge]

[-- Attachment #1.1: Type: text/plain, Size: 422 bytes --]

Patch attached for a new file algebra.scm containing fplll and mpfrcx.

I also tried to add gmp-ecm, but it is only downloadable over https://,
which is not yet implemented.

Then I would like to add Pari/GP, but it uses non-standard commands
("./Configure" instead of "./configure", "make all" instead of "make",
"make dobench" instead of "make check"; only "make install" is the same).
Can this be done easily?

Andreas

[-- Attachment #1.2: Type: text/html, Size: 2569 bytes --]

[-- Attachment #2: 0001-distro-Add-mpfrcx-and-fplll.patch --]
[-- Type: text/x-patch, Size: 4066 bytes --]

From cee62de2b42d26c0ada1b2d8ecf536199dd0a778 Mon Sep 17 00:00:00 2001
From: Andreas Enge <andreas@enge.fr>
Date: Wed, 19 Dec 2012 23:50:52 +0100
Subject: [PATCH]     distro: Add mpfrcx and fplll.

    * distro/packages/algebra.scm: New file.
---
 Makefile.am                 |    1 +
 distro/packages/algebra.scm |   75 +++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 76 insertions(+)
 create mode 100644 distro/packages/algebra.scm

diff --git a/Makefile.am b/Makefile.am
index 3d16a16..7a9bc1c 100644
--- a/Makefile.am
+++ b/Makefile.am
@@ -44,6 +44,7 @@ MODULES =					\
   guix.scm					\
   distro.scm					\
   distro/packages/acl.scm			\
+  distro/packages/algebra.scm			\
   distro/packages/attr.scm			\
   distro/packages/autotools.scm			\
   distro/packages/base.scm			\
diff --git a/distro/packages/algebra.scm b/distro/packages/algebra.scm
new file mode 100644
index 0000000..1e1f093
--- /dev/null
+++ b/distro/packages/algebra.scm
@@ -0,0 +1,75 @@
+;;; Guix --- Nix package management from Guile.         -*- coding: utf-8 -*-
+;;; Copyright (C) 2012 Andreas Enge <andreas@enge.fr>
+;;;
+;;; This file is part of Guix.
+;;;
+;;; Guix is free software; you can redistribute it and/or modify it
+;;; under the terms of the GNU General Public License as published by
+;;; the Free Software Foundation; either version 3 of the License, or (at
+;;; your option) any later version.
+;;;
+;;; Guix is distributed in the hope that it will be useful, but
+;;; WITHOUT ANY WARRANTY; without even the implied warranty of
+;;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+;;; GNU General Public License for more details.
+;;;
+;;; You should have received a copy of the GNU General Public License
+;;; along with Guix.  If not, see <http://www.gnu.org/licenses/>.
+
+(define-module (distro packages algebra)
+  #:use-module (distro)
+  #:use-module (distro packages multiprecision)
+  #:use-module (guix licenses)
+  #:use-module (guix packages)
+  #:use-module (guix download)
+  #:use-module (guix build-system gnu))
+
+
+(define-public mpfrcx
+  (package
+   (name "mpfrcx")
+   (version "0.4.1")
+   (source (origin
+            (method url-fetch)
+            (uri (string-append
+                  "http://www.multiprecision.org/mpfrcx/download/mpfrcx-"
+version ".tar.gz"))
+            (sha256 (base32
+                     "1rrc75chxyicqjgg5mfhgbz7p9mx1fgh0qlx14a82m25vfhifnd1"))))
+   (build-system gnu-build-system)
+   (inputs `(("gmp" ,gmp)
+             ("mpfr" ,mpfr)
+             ("mpc"  ,mpc)))
+   (synopsis "mpfrcx, a library for the arithmetic of univariate polynomials
+over arbitrary precision real or complex numbers")
+   (description
+    "mpfrcx is a library for the arithmetic of univariate polynomials over
+arbitrary precision real (mpfr) or complex (mpc) numbers, without control
+on the rounding. For the time being, only the few functions needed to
+implement the floating point approach to complex multiplication are
+implemented. On the other hand, these comprise asymptotically fast
+multiplication routines such as Toom–Cook and the FFT. ")
+   (license lgpl2.1+)
+   (home-page "http://mpfrcx.multiprecision.org/")))
+
+
+(define-public fplll
+  (package
+   (name "fplll")
+   (version "4.0.1")
+   (source (origin
+            (method url-fetch)
+            (uri (string-append
+                  "http://perso.ens-lyon.fr/damien.stehle/fplll/libfplll-"
+version ".tar.gz"))
+            (sha256 (base32
+                     "122bpqdlikshhd7nmq0l5qfc0agyk7x21gvplv1l9hb77l8cy9rw"))))
+   (build-system gnu-build-system)
+   (inputs `(("gmp" ,gmp)
+             ("mpfr" ,mpfr)))
+   (synopsis "fplll, a library for LLL-reduction of euclidean lattices")
+   (description
+    "fplll LLL-reduces euclidean lattices. Since version 3, it can also
+solve the shortest vector problem.")
+   (license lgpl2.1+)
+   (home-page "http://perso.ens-lyon.fr/damien.stehle/fplll/")))
-- 
1.7.10.4

Re: New package: FPLLL

[Ludovic Courtès]

Hi!

Andreas Enge <andreas@enge.fr> skribis:

> Patch attached for a new file algebra.scm containing fplll and mpfrcx.

Applied, thanks.

> I also tried to add gmp-ecm, but it is only downloadable over https://,
> which is not yet implemented.

Oh.  Would be great if they had an FTP or HTTP mirror.

> Then I would like to add Pari/GP, but it uses non-standard commands
> ("./Configure" instead of "./configure", "make all" instead of "make",
> "make dobench" instead of "make check"; only "make install" is the same).
> Can this be done easily?

Roughly, yes.  See how this is done for Perl, for instance: the idea is
to replace the standard ‘configure’ phase (which can be seen in
gnu-build-system.scm) by a custom one, via the #:phases argument.

HTH,
Ludo’.

Re: New package: FPLLL

[Andreas Enge]

[-- Attachment #1: Type: text/plain, Size: 482 bytes --]

Am Donnerstag, 20. Dezember 2012 schrieb Ludovic Courtès:
> > I also tried to add gmp-ecm, but it is only downloadable over
> > https://, which is not yet implemented.
> Oh.  Would be great if they had an FTP or HTTP mirror.

Would it be difficult to implement in guix? There are guile bindings
in gnutls. One cannot reasonably tell people who use crypto to not do so.

We could mirror the source elsewhere, but my own web server also
forces the use of TLS...

Andreas

[-- Attachment #2: Type: text/html, Size: 2719 bytes --]

Re: New package: FPLLL

[Ludovic Courtès]

Andreas Enge <andreas@enge.fr> skribis:

> Am Donnerstag, 20. Dezember 2012 schrieb Ludovic Courtès:
>> > I also tried to add gmp-ecm, but it is only downloadable over
>> > https://, which is not yet implemented.
>> Oh.  Would be great if they had an FTP or HTTP mirror.
>
> Would it be difficult to implement in guix? There are guile bindings
> in gnutls.

Right, so that’s definitely doable.  It just has to be done.  :-)

> One cannot reasonably tell people who use crypto to not do so.

OTOH, what matters most here is the authenticity and integrity of the
tarball, and TLS doesn’t help here.

(FWIW I’m also fond of EFF’s HTTPS-Everywhere; I just don’t think it’s
this relevant here, esp. since we’ll eventually have content-based
mirrors as on hydra.nixos.org.)

Thanks,
Ludo’.

Re: New package: FPLLL

[Andreas Enge]

[-- Attachment #1: Type: text/plain, Size: 326 bytes --]

Am Freitag, 21. Dezember 2012 schrieb Ludovic Courtès:
> OTOH, what matters most here is the authenticity and integrity of the
> tarball, and TLS doesn’t help here.

Agreed, we do not gain anything with TLS here, since we hardcode the hash
values. I just do not see myself asking people to turn off TLS...

Andreas

[-- Attachment #2: Type: text/html, Size: 1976 bytes --]

Re: New package: FPLLL

[Ludovic Courtès]

Andreas Enge <andreas@enge.fr> skribis:

> Agreed, we do not gain anything with TLS here, since we hardcode the hash
> values. I just do not see myself asking people to turn off TLS...

Yeah, this I can understand.  :-)

Ludo’.

Re: New package: FPLLL

[Ludovic Courtès]

Hi!

Andreas Enge <andreas@enge.fr> skribis:

> Am Mittwoch, 12. Dezember 2012 schrieb Ludovic Courtès:
>> >    (license "LGPLv2.1")
>> It seems to be LGPLv2.1+ rather.
>
> The file README.html states:
> "fplll is distributed under GNU Lesser General
> Public License (v2.1) as published by the Free Software Foundation."
>
> I checked two source files, which say
> "or (at your option) any later version".
> I suppose this is authoritative.

Yes, it is.  (If there were no statements at all in source files, and
just the standard COPYING.LESSER, then it would be “or later” too,
because that’s what the license says.)

> Am Mittwoch, 12. Dezember 2012 schrieb Ludovic Courtès:
>> > It could go into multiprecision.scm, or get its own file.
>> Yes, multiprecision.scm would seem like a good option to me.  WDYT?
>
> It depends on whether you would rather go for one file per package (maybe
> to be organised in subdirectories), or one file per group of packages of a
> similar category.

I have mixed feelings.  Categories would often be hard to determine
correctly and intelligently while avoiding circular references, except
in some cases like this one.

> Even in the latter case, it looks like multiprecision.scm is not a very
> good place, since it regroups the dependencies of gcc. Maybe algebra.scm?
> It could also contain mpfrcx, gmp-ecc and maybe pari (which does not
> exactly follow gnu build standards, so I would need some help).

Yes, makes sense to me.  We could rename multiprecision to algebra, then.

Ludo’.