Re: [sage-combinat-devel] Re: Categories

[William Stein]

On Tue, Jul 7, 2009 at 9:35 AM, Anne Schilling<an...@math.ucdavis.edu> wrote:
>
> Hi!
>
> Let me reiterate Nicolas' message that it would be really
> great to have his category code (and hopefully the root system
> and affine crystal patches) integrated into Sage before FPSAC
> which starts in less than two weeks from now! I spent a serious
> amount of time getting the affine crystal code ready for FPSAC,
> and I know Nicolas did as well for the category/root system
> patches.
>
> I was at a conference in Ottawa last week and gave a short demo
> on the affine crystal code there. Some people wanted to start
> using the code, but had trouble installing Sage-combinat (on top
> of Sage) since it requires additional development tools.
>
> If all of this code was already in Sage, I am sure things will
> go much smoother at FPSAC and the *-Combinat days following it.
>

I propose that before FPSAC we just make two releases of Sage in parallel:

(1) sage-4.1.tar

(2) sagecombinat-4.1.tar

Version (1) will have passed the full Sage referee process, testing
process, etc.
Version (2) will have not, but will include *all* of *-Combinat in it.
It would not be too much extra work for me to build binaries for
both. The source and binaries can be posted in a directory on
sage.math, so it's still easy for people to get.

Thoughts?

This would mean making something easily available for users that has
not just the
category code but everything else in *-combinat. To do this, could
somebody ASAP build sage-4.1.rc0.tar, pull and rebase all the
*-combinat stuff against it?

-- William

[William Stein]

On Tue, Jul 7, 2009 at 11:28 AM, Nicolas M.
Thiery<Nicolas...@u-psud.fr> wrote:
>
>        Dear William,

> Ah, some sort of preview release. This could a  way to handle the
> situation. Thanks for offering this idea!

I'm glad you find it appealing.

> It's probably best to not put all of *-Combinat, but only those stable
> parts that are waiting for reviews Sage. As a side effect this will
> make for a dry-run release, checking that their integration indeed
> goes smoothly (or not!).

Yep.

>
> Just out of curiosity: do you think sage -upgrade would be able to
> handle the later transition sagecombinat-4.1 -> sage-4.2?

I don't know, since sage-4.2 doesn't exist. However,

sage -upgrade a/specific/url

will definitely be able to handle sagecombinat-4.1 --> sagecombinat-4.2.

>> To do this, could somebody ASAP build sage-4.1.rc0.tar, pull and
>> rebase all the *-combinat stuff against it?
>

> Will do this afternoon, and put some guards to mark those part we want
> to integrate into sagecombinat-4.1.
>

Thanks. I'll think of you as the "release manager" for putting
together the Sage library code that goes from sage-4.1.rc0 to
sagecombinat-4.1.rc0. This could mean that you provide a command I
can type that will take an installed sage-4.1.rc0 and make it into
what you want sagecombinat-4.1.rc0 to be.

William

[Nick Alexander]

> (1) sage-4.1.tar
>
> (2) sagecombinat-4.1.tar

+1 to this idea. I'm getting worried about this giant patch bomb
ticking down.

Nick