On Mar 24, 3:20 am, Peter Jeremy <peterjer...@optushome.com.au> wrote:
> On 2009-Mar-23 12:38:45 -0700, mabshoff <mabsh...@googlemail.com> wrote:
Hi,
> >> FreeBSD and I've reached the point where I can compile sage-3.4 on
> >> FreeBSD-8/amd64 (using gcc/g++/gfortran 4.3) and get it to start.
>
> >Which FreeBSD release are you using?
>
> As I said, 8-current/amd64.
D'oh :)
> >This is way too much - I get Sage to build and pass all but a few
> >(around 15 or so IIRC) tests with 4 small patches.
>
> Ah. Are these patches available? I looked in the FreeBSD entry on
> the wiki but whilst it talks about various failures for older versions
> of sage, I couldn't find any reference to either sage-3.4 or FreeBSD
> patches.
The patches are flying around somewhere on my hard disk and I need to
dig them out :)
> >Why do you apply fixes from ports? Python compiles out of the box for
> >me on FreeBSD.
>
> The python embedded in sage compiles for me without patches but a lot
> of the failures I am getting point towards a python problem. The
> FreeBSD port for python 2.5.4 includes a number of patches that may
> be relevant to the problems I am having.
I doubt that - at least I have no problem with the python we ship on
FreeBSD 7.
> As a general question on sage packaging, the sources are currently
> totally self-contained. Whilst this has the advantage that all the
> dependencies are in one place, it also has the disadvantage that it
> results in multiple copies of some tools (bzip2, gd, libgcrypt,
> libgpg_error, libpng, mpfr, python, sqlite3 and zlib in my case) being
> installed. Was this done because Linux doesn't have a standard
> package management system across all distros or to allow sage to
> include sage-specific patches in the tools it uses (I notice this is
> done for some tools)? I know that trying to turn sage into a FreeBSD
> port in its current form will encounter resistance due to this (there
> is pressure on other large ports like OpenOffice.org and the Mozilla
> suite to depend on FreeBSD ports rather than embedding equivalent
> functionality).
I personally couldn't care less whether Sage becomes a FreeBSD port or
not. I see little evidence that any packager can keep up with Sage
*and* have a Sage library that passes doctests. Not that I want to
discourage you from trying :)
The failures you saw indicate to me that some problem I did not have
was introduced by using the Python from ports. And that is only the
start IMHO. My main two issues with FreeBSD 7 and earlier releases of
Sage were:
* pyprocessing segfaulting on FreeBSD 7 - this is likely something
that has been fixed in the backported pyprocessing from Python 2.6,
but I have not looked into fixing this
* java not working - we currently cannot use IcedTea or anything like
that since jmol does not work with it and each party involved (IcedTea
and jmol) are pointing their finger at the other party and blaming
them. I installed the SunJDK on FreeBSD 7 and the linux compat layer
to make it run and java just segfaulted for me, so no joy :(
> --
> Peter Jeremy
Cheers,
Michael
> application_pgp-signature_part
> < 1KViewDownload
--~--~---------~--~----~------------~-------~--~----~
To post to this group, send email to sage-devel@googlegroups.com
To unsubscribe from this group, send email to
sage-devel-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URLs: http://www.sagemath.org
-~----------~----~----~----~------~----~------~--~---
