Cephes licence : repackage?

[Julien Puydt]

Hi,

I ask because I stumbled on the following thread :
http://lists.debian.org/debian-legal/2004/12/msg00295.html
after this thread, a debian bug report was closed ; look closely the
associated changelog :
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=287507
the mail mentioned is the following :
https://portal.sbgrid.org/docs/usr_doc/grace-5.1.22/cephes-license.email

It doesn't seem sage got the cephes code from one of those places where
it was released under a GPL licence ; perhaps it would be worth
repackaging?

Snark on #sagemath

[Volker Braun]

Currently, the cephes package only prints:

"We do not install the cephes library on any operating system except Cygwin."

I think boost.math would be a better choice of special function library e.g. for systems where the libc implementation sucks. Also, it seems like cephes is included in scipy distribution (scipy-0.9.p1/src/scipy/special/cephes/) so we ship it already.

Remove the cephes spkg?

[Mike Hansen]

On Sat, May 19, 2012 at 4:58 PM, Volker Braun <vbrau...@gmail.com> wrote:
> I think boost.math would be a better choice of special function library e.g.
> for systems where the libc implementation sucks.

I don't think that boost.math supports the things that we need it to,
for example complex.h.

>Also, it seems like cephes
> is included in scipy distribution (scipy-0.9.p1/src/scipy/special/cephes/)
> so we ship it already.

I believe this is something totally different.

--Mike

[Volker Braun]

On Saturday, May 19, 2012 8:15:06 PM UTC-4, Mike Hansen wrote:

I don't think that boost.math supports the things that we need it to,
for example complex.h.

Oh I see you are saying they don't support complex arguments for all functions.

 

>Also, it seems like cephes
> is included in scipy distribution (scipy-0.9.p1/src/scipy/special/cephes/)

I believe this is something totally different. 

I'm pretty sure this is the same cephes, the sources contain "Cephes Math Library Release 2.3:  March, 1995 Copyright 1984, 1995 by Stephen L. Moshier". 

[Mike Hansen]

On Sat, May 19, 2012 at 5:33 PM, Volker Braun <vbrau...@gmail.com> wrote:
> On Saturday, May 19, 2012 8:15:06 PM UTC-4, Mike Hansen wrote:
>>
>> I don't think that boost.math supports the things that we need it to,
>> for example complex.h.
> Oh I see you are saying they don't support complex arguments for all
> functions.

No -- Cygwin doesn't have C9X complex support nor long double support
(things like cabs or sqrtl). The cephes package provides that.

>> >Also, it seems like cephes
>> > is included in scipy distribution
>> > (scipy-0.9.p1/src/scipy/special/cephes/)
>> I believe this is something totally different.
> I'm pretty sure this is the same cephes, the sources contain "Cephes Math
> Library Release 2.3:  March, 1995 Copyright 1984, 1995 by Stephen L.
> Moshier".

It looks like an older version of the double library for special
functions (src/double/ in the cephes spkg), but that's only a small
part of what is needed.

--Mike

[David Kirkby]

On 20 May 2012 00:58, Volker Braun <vbrau...@gmail.com> wrote:
> Currently, the cephes package only prints:
>
> "We do not install the cephes library on any operating system except
> Cygwin."

I believe when someone else (I forget who), was working on the FreeBSD
port, he needed Cephas too. I believe Sage is now building on FreeBSD,
but it might be using this too.

dave

[Julien Puydt]

Le dimanche 20 mai, David Kirkby a écrit:

> I believe when someone else (I forget who), was working on the FreeBSD
> port, he needed Cephas too. I believe Sage is now building on FreeBSD,
> but it might be using this too.

The following two tickets seem relevant :
http://trac.sagemath.org/sage_trac/ticket/9543
http://trac.sagemath.org/sage_trac/ticket/9601

In fact, at some point Dima thought we could use cephes to fix the
gamma precision issues of the libc on ARM ; that is why I was
investigating cephes more closely.

Snark on #sagemath

[Stephen Montgomery-Smith]

That was me. Yes, please keep cephes in sage.

[Dr. David Kirkby]

It was Peter Jeremy before you working on the FreeBSD port. I knew he used
Cephes. So it seems the comment about

"We do not install the cephes library on any operating system except Cygwin."

is just plain wrong.

Is anyone working on a Cygin port any more? I used to wonder if the Solaris or
Cygwin ports would be completed first, but the race was not as close as I
expected it would be.


Dave

[Julien Puydt]

Le dimanche 20 mai, Stephen Montgomery-Smith a écrit:

> That was me. Yes, please keep cephes in sage.

Notice that I didn't ask for removal -- I only notified that there
might be a licence issue. :-)

Snark on #sagemath

[Stephen Montgomery-Smith]

On 05/20/2012 10:39 AM, Dr. David Kirkby wrote:
> On 05/20/12 03:38 PM, Stephen Montgomery-Smith wrote:
>> On 05/20/2012 06:50 AM, David Kirkby wrote:
>>> On 20 May 2012 00:58, Volker Braun<vbrau...@gmail.com> wrote:
>>>> Currently, the cephes package only prints:
>>>>
>>>> "We do not install the cephes library on any operating system except
>>>> Cygwin."
>>>
>>>
>>> I believe when someone else (I forget who), was working on the FreeBSD
>>> port, he needed Cephas too. I believe Sage is now building on FreeBSD,
>>> but it might be using this too.
>>>
>>> dave
>>
>> That was me. Yes, please keep cephes in sage.
>
> It was Peter Jeremy before you working on the FreeBSD port.

Yes, my apologies if there was any misattribution. I use Peter Jeremy's
patches in the port of sage to FreeBSD.

http://www.freebsd.org/cgi/ports.cgi?query=sage&stype=name&sektion=math

[Stephen Montgomery-Smith]

and also

http://www.freebsd.org/cgi/cvsweb.cgi/ports/math/sage/files/spkg-patch-cephes-2.8_-_big-patch

[Keshav Kini]

"Dr. David Kirkby" <david....@onetel.net> writes:
> Is anyone working on a Cygin port any more? I used to wonder if the
> Solaris or Cygwin ports would be completed first, but the race was not
> as close as I expected it would be.

As far as I remember, our BDFL publicly stated that working on a Cygwin
port is a waste of time, after which people more or less stopped working
on it :)

After some searching, here is the message in question:
http://thread.gmane.org/gmane.comp.mathematics.sage.devel/54804

-Keshav

----
Join us in #sagemath on irc.freenode.net !