[gentoo-science] sage status

[gentoo-science] sage status

[François Bissey]

Hi all,

so sage-4.4 is coming soon. If we stick to our plans sage-4.4 will
leave /opt to join other non-binary packages under /usr.
* we have completely split sage... this means we don't rely anymore on 
sage build system in any way.
* thanks to contribution of Steve Trogdon we have identified a number of
issues on amd64 and some with a larger scope.

* for those of you living on ~arch and are getting headaches because
you cannot update python to version 2.6.5, I have committed a patched
python-2.6.5 to the overlay. This is not supported by the sage devs so 
there may be problems.

*Steve found an issue with matplotlib. If you are using matplotlib with gtk
enabled _and_ the proprietary nvidia drivers sage can break.

*two people have come forward with a problem related to the way sage 
is interacting with gap. If you have that kind of problem please tell us
as we would really like to figure out the cause of the problem and the two
cases we have are not enough.

Francois

Re: [gentoo-science] sage status

[Thomas Kahle]

[-- Attachment #1: Type: text/plain, Size: 890 bytes --]

Hi,

> * for those of you living on ~arch and are getting headaches because
> you cannot update python to version 2.6.5, I have committed a patched
> python-2.6.5 to the overlay. This is not supported by the sage devs so 
> there may be problems.

For those of us living in x86, is it save to use python-2.6.4-r99 as the
system python??
I'm slightly worried to use a sage-patched python on global scale...

Regards,
Thomas

> 
> *Steve found an issue with matplotlib. If you are using matplotlib with gtk
> enabled _and_ the proprietary nvidia drivers sage can break.
> 
> *two people have come forward with a problem related to the way sage 
> is interacting with gap. If you have that kind of problem please tell us
> as we would really like to figure out the cause of the problem and the two
> cases we have are not enough.
> 
> Francois


-- 
Thomas Kahle


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 380 bytes --]

Re: [gentoo-science] sage status

[Christopher Schwan]

Hi,

On Monday 26 April 2010 01:22:48 Thomas Kahle wrote:
> Hi,
> 
> > * for those of you living on ~arch and are getting headaches because
> > you cannot update python to version 2.6.5, I have committed a patched
> > python-2.6.5 to the overlay. This is not supported by the sage devs so
> > there may be problems.
> 
> For those of us living in x86, is it save to use python-2.6.4-r99 as the
> system python??
> I'm slightly worried to use a sage-patched python on global scale...
> 
> Regards,
> Thomas

that is a good point - we are also not happy to ship our own python, but 
unfortunately a lot of Sage functions are not working without it, as you 
already may know.

The difference to the non -r99 version is only a small patch, which is waiting 
to be integrated into python (see http://bugs.python.org/issue7689 for an 
exhaustive description and patch). I see no reason why this patch should not 
be safe - its adding new functionality and does not break existing code. I am 
using python-2.6.4-r99 since we were introducing it and did not notice any 
problems.

We are hoping that the python people integrate this as soon as possible, but 
if you are looking at the bug report date you can see that it already is 3 1/2 
months old - does anybody know how to speed up the process ?

Christopher

Re: [gentoo-science] sage status

[François Bissey]

> For those of us living in x86, is it save to use python-2.6.4-r99 as the
> system python??
> I'm slightly worried to use a sage-patched python on global scale...

Hi Thomas.

I am on x86 and I have been using it on my system for as long as I have
put it in the overlay. 
The patch really concern 3 files (including a test file) in the whole of 
python. It shouldn't really have any side effects as it brings more 
functionality to python without removing anything. 
I am not even sure who/what use pickling apart from sage.

Francois