Issues with plotting on Intel Mac 64 bit?

[Simon King]

Hi sage-devel!

Trying to follow Mike's advice from
http://groups.google.com/group/sage-devel/browse_thread/thread/917f03cb17de06b6
I did some testing of show.

When I logged into bsd.math via ssh, I get the following in sage-4.1.1
in 64 bit mode:
sage: from sage.plot.plot import circle
sage: C = circle((0,1),1)
sage: show(C)
/scratch/king/SAGE64/sage-4.1.1/local/bin/sage-sage: line 199: 62601
Abort trap sage-ipython "$@" -i

This is actually not only an error that is raised: Sage dies!

The same code works without problem on bsd.math in 32 bit. Of course I
don't see a picture, but I do get a picture on my local machine
(openSUSE 11.0).

What is happening there?

Best regards,
Simon

[Mike Hansen]

Hello,

On Thu, Oct 8, 2009 at 4:43 PM, Simon King <simon...@nuigalway.ie> wrote:
> When I logged into bsd.math via ssh, I get the following in sage-4.1.1
> in 64 bit mode:
>  sage: from sage.plot.plot import circle
>  sage: C = circle((0,1),1)
>  sage: show(C)
>  /scratch/king/SAGE64/sage-4.1.1/local/bin/sage-sage: line 199: 62601
> Abort trap              sage-ipython "$@" -i

This is one of the outstanding issues for Sage on 10.6. See #7022
which should / may fix the specific issue that you reported. However,
there are plenty of ones that we don't have a fix for -- see #7095. I
think there are a couple threads on sage-devel about this.

--Mike

[Jason Grout]

Mike Hansen wrote:

However,
> there are plenty of ones that we don't have a fix for -- see #7095. I
> think there are a couple threads on sage-devel about this.

Should we be holding up 4.1.2 for issues that (according to what you
say) don't have any solution, when the issue in question is supporting a
new platform (at least, a new release) that admittedly breaks lots of
software out there? Or should we release 4.1.2, declare OSX 10.6 not
supported for this release, and then follow up with 4.2 when we have
fixes for 10.6?

There's an awful lot already in 4.1.2, and a lot more patches are being
held up which can't be merged because 4.1.2 is frozen for who knows how
long. There are 55 "positive review" patches right now.

Jason

--
Jason Grout

[kcrisman]

> Should we be holding up 4.1.2 for issues that (according to what you
> say) don't have any solution, when the issue in question is supporting a
> new platform (at least, a new release) that admittedly breaks lots of
> software out there?  Or should we release 4.1.2, declare OSX 10.6 not
> supported for this release, and then follow up with 4.2 when we have
> fixes for 10.6?

This is a good question.

> There's an awful lot already in 4.1.2, and a lot more patches are being
> held up which can't be merged because 4.1.2 is frozen for who knows how
> long.  There are 55 "positive review" patches right now.

And to be honest, there is a lot of new functionality, upgrades - 4.2
might not be an inappropriate designation for Snow Leopard support + a
few other goals of various sorts.

- kcrisman

[William Stein]

On Thu, Oct 8, 2009 at 5:41 PM, Jason Grout <jason...@creativetrax.com> wrote:
>
> Mike Hansen wrote:
>
> However,
>> there are plenty of ones that we don't have a fix for -- see #7095.  I
>> think there are a couple threads on sage-devel about this.
>
>
> Should we be holding up 4.1.2 for issues that (according to what you
> say) don't have any solution, when the issue in question is supporting a
> new platform (at least, a new release) that admittedly breaks lots of
> software out there?  Or should we release 4.1.2, declare OSX 10.6 not
> supported for this release, and then follow up with 4.2 when we have
> fixes for 10.6?
>

Jason, to move your suggestion forward, please referee this patch:

http://trac.sagemath.org/sage_trac/ticket/7193

William

[Jason Grout]

I don't have access to any 10.6 machines; otherwise I would.

Can someone referee this patch?

Jason

[Jason Grout]

Jason Grout wrote:
> William Stein wrote:
>> On Thu, Oct 8, 2009 at 5:41 PM, Jason Grout <jason...@creativetrax.com> wrote:
>>> Mike Hansen wrote:
>>>
>>> However,
>>>> there are plenty of ones that we don't have a fix for -- see #7095. I
>>>> think there are a couple threads on sage-devel about this.
>>> Should we be holding up 4.1.2 for issues that (according to what you
>>> say) don't have any solution, when the issue in question is supporting a
>>> new platform (at least, a new release) that admittedly breaks lots of
>>> software out there? Or should we release 4.1.2, declare OSX 10.6 not
>>> supported for this release, and then follow up with 4.2 when we have
>>> fixes for 10.6?
>>>
>> Jason, to move your suggestion forward, please referee this patch:
>>
>> http://trac.sagemath.org/sage_trac/ticket/7193

Looking at the patch, shouldn't we use the python platform module to
test for platform identity?

http://docs.python.org/library/platform.html

In particular, this lets us test for 10.6 specifically:

>>> platform.mac_ver()
('10.6.1', ('', '', ''), 'i386')

We can just print a warning out when the person is running 10.6. On the
other hand, it might be nice to warn someone that plans on upgrading
that Sage will not work (for now) if they upgrade to 10.6. Should the
warning be printed out for all OSX users, or just for people running on
10.6?

Thanks,

Jason

--
Jason Grout