On Sun, Jul 5, 2009 at 10:58 PM, kilian<koeps...@gmail.com> wrote:
>
> Ondrej,
>
> On Jul 5, 6:28 pm, Ondrej Certik <ond...@certik.cz> wrote:
>>
>> Excellent, I have added you to the project, so just upload your spkg
>> package into the Downloads (hit new download and it should work).
>>
>
> OK, I uploaded it.
>
>> > One thing that I think would be great on the longer run would be to
>> > make package description files,
>> > just like the spkg files but without the src directory, and that
>> > contain the url for the upstream package.
>> > This way, people could contribute package descriptions and you could
>> > include these (small) files in
>> > the distribution. This is how the fink project (http://fink.sf.net)
>> > distributes the patches necessary to
>> > compile a variety of unix programs under OSX.
>>
>> Interesting --- so you think that SPD itself should include most of
>> those small files for most of packages and it would now how to
>> download the src directory, thus installing itself?
>>
>> I think we would have to keep it updated all the time, isn't it easier
>> to just point to the real spkg packages, that are known to work? E.g.
>> create a pool of supported spkg packages and then just do something
>> like:
>>
>> ./spd -i nose
>>
>> and it would try to  download the latest spkg package. In fact, this
>> is how Sage works already (it tries to download the package form
>> sagemath.org).
>
> I thought, the advantage in having a small package description file
> would be that it would be easier to see, what the package actually
> does
> and therefore, it would be easier to monitor many packages from many
> contributors. And it would be easier to put a tree of these
> descriptions under
> source control, maybe even two trees: testing and stable. You could
> still
> keep a copy of the original source tar file as a fall-back solution.
>
> Each package description file would contain the url of the original
> source,
> its MD5 sum, a patch against this source code and maybe the email
> address
> of the package maintainer,
>
> On the other hand, the fact that the new format would deviate from the
> sage
> format would be a clear drawback. Maybe, we should have this
> discussion
> rather on the sage email list...

Let's discuss this on the mailinglist from now on.

Ondrej

--~--~---------~--~----~------------~-------~--~----~
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
-~----------~----~----~----~------~----~------~--~---

Reply via email to