Re: Bug#573538: Packaging again sagemath

classic Classic list List threaded Threaded
6 messages Options
Reply | Threaded
Open this post in threaded view
|

Re: Bug#573538: Packaging again sagemath

Giovanni Mascellani-3
Hi.

Thanks for your feedback. Unfortunately in the meantime upstream's
reaction wasn't really the one I'd liked more[1].

 [1]
http://groups.google.com.au/group/debian-sage/browse_thread/thread/20e092f1b11e6f02

Anyway, maybe these issues can be solved, if we talk with upstream and,
above all, we can find a few people to comaintain the package. Working
with a non collaborative upstream would be very frustrating on such big
package (it is frustrating on much more stupid packages...).

Il 19/07/2010 15:52, Rogério Brito ha scritto:
> BTW, I will be at this year's DebConf and I hope to be able to work on this, at
> least to get something started.

Unfortunately, I won't come (even if I hope to be in Banja Luka 2011).
Good luck! :-)

After announcing my interest, I didn't invest any time in sage, partly
because of upstream's reaction I mentioned above, partly because I
wanted to finish to work on some packages I'm already busy on. However,
I really don't think that sagemath will be part of squeeze, I guess
we'll have to wait at least squeeze+1.

Ciao, Giovanni.
--
Giovanni Mascellani <[hidden email]>
Pisa, Italy

Web: http://poisson.phc.unipi.it/~mascellani
Jabber: [hidden email] / [hidden email]


signature.asc (390 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: Bug#573538: Packaging again sagemath

Rogério Brito
Hi, Giovanni.

On Jul 19 2010, Giovanni Mascellani wrote:
> Thanks for your feedback. Unfortunately in the meantime upstream's
> reaction wasn't really the one I'd liked more[1].

Thank you very much for your reply.

> http://groups.google.com.au/group/debian-sage/browse_thread/thread/20e092f1b11e6f02

I did not know about this reaction, but I can understand where they are
coming from: it seems that sage is updated regularly and the Debian
packaging is rotting in unstable. [1]

[1] http://packages.qa.debian.org/s/sagemath.html

The package is not installable for quite some time and I think that, in
the current situation, it will never be part of any stable release,
which is one of the justifications for it to be removed from the
archives.

> Anyway, maybe these issues can be solved, if we talk with upstream and,
> above all, we can find a few people to comaintain the package.

If I understood them correctly, they are not 100% opposed to the idea of
having sagemath: they are only opposed to having the package too outdated.

> Working with a non collaborative upstream would be very frustrating on
> such big package (it is frustrating on much more stupid packages...).

Well, I think that a public repository for packaging it would be a very
good start (Tim, do you have any?). Some of the big tasks that I would
like to see addressed regarding sagemath would be:

* getting a first draft package done, even if "improper for public
  consumption".
* relaxing the huge amount of dependencies (dropping them to recommends,
  instead).
* modularizing things as much as possible.
* getting patches fed up to other upstream packages.

Of course, packaging sagemath is a very big task, basing myself only on
the list of programs that it embraces.

> Unfortunately, I won't come (even if I hope to be in Banja Luka 2011).
> Good luck! :-)

Thanks. I hope to go also to Banja Luka. :-)

> After announcing my interest, I didn't invest any time in sage, partly
> because of upstream's reaction I mentioned above, partly because I
> wanted to finish to work on some packages I'm already busy on.

OK. I did not have time to play with it yet, but now that I completed
some tasks (read: playing with the port of xpdf to use libpoppler and
other small stuff), I think that I can reserve some time for another
package, especially if we can put everything under a git repository,
which will be convenient for the development.

And I am happy to teach the little that I know about maintaining
projects with git.

> However, I really don't think that sagemath will be part of squeeze, I
> guess we'll have to wait at least squeeze+1.

I am not very ambitious regarding getting it to be part of squeeze, but
just having it in Debian, in a working situation is way better than the
current situation. :-)

Seeing upstream's reaction makes me think that, perhaps, the best option
for packaging sagemath would be to place it in volatile [2], so that it can
always be close to what upstream wants and also what users can use.


[2] http://www.debian.org/volatile/


Regards,

--
Rogério Brito : rbrito@{ime.usp.br,gmail.com} : GPG key 1024D/7C2CAEB8
http://rb.doesntexist.org : Packages for LaTeX : algorithms.berlios.de
DebianQA: http://qa.debian.org/developer.php?login=rbrito%40ime.usp.br


--
To UNSUBSCRIBE, email to [hidden email]
with a subject of "unsubscribe". Trouble? Contact [hidden email]
Archive: http://lists.debian.org/20100719162634.GA8458@...

Reply | Threaded
Open this post in threaded view
|

Re: Bug#573538: Packaging again sagemath

Giovanni Mascellani-3
Hi.

Rogério, you can drop me from recipients list, I'm already subscribed to
the bug and to debian-science.

Il 19/07/2010 18:26, Rogério Brito ha scritto:
>> Anyway, maybe these issues can be solved, if we talk with upstream and,
>> above all, we can find a few people to comaintain the package.
>
> If I understood them correctly, they are not 100% opposed to the idea of
> having sagemath: they are only opposed to having the package too outdated.

Yes, that's my interpretation too. But I think that, in order to have
good communication with upstream, we'll have to show that the package
can be maintained and won't fall forgotten as last time. This won't be
easy. I agree that the first pace is to try again to have a working
package, then understand which are the difficulties and decide whether
we're able to go forth with it.

>> Working with a non collaborative upstream would be very frustrating on
>> such big package (it is frustrating on much more stupid packages...).
>
> Well, I think that a public repository for packaging it would be a very
> good start (Tim, do you have any?). Some of the big tasks that I would
> like to see addressed regarding sagemath would be:
>
> * getting a first draft package done, even if "improper for public
>   consumption".
> * relaxing the huge amount of dependencies (dropping them to recommends,
>   instead).
> * modularizing things as much as possible.
> * getting patches fed up to other upstream packages.
>
> Of course, packaging sagemath is a very big task, basing myself only on
> the list of programs that it embraces.
I fully agree with this program.

> OK. I did not have time to play with it yet, but now that I completed
> some tasks (read: playing with the port of xpdf to use libpoppler and
> other small stuff), I think that I can reserve some time for another
> package, especially if we can put everything under a git repository,
> which will be convenient for the development.
>
> And I am happy to teach the little that I know about maintaining
> projects with git.

+1 for git, is my preferred VCS too.

>> However, I really don't think that sagemath will be part of squeeze, I
>> guess we'll have to wait at least squeeze+1.
>
> I am not very ambitious regarding getting it to be part of squeeze, but
> just having it in Debian, in a working situation is way better than the
> current situation. :-)

Of course! :-)

> Seeing upstream's reaction makes me think that, perhaps, the best option
> for packaging sagemath would be to place it in volatile [2], so that it can
> always be close to what upstream wants and also what users can use.

I'm not sure it's a good idea to put such a complex package in volatile:
it's very difficult to make an update without disrupting things, I guess.

Anyway, I think the main problem is not having an old package, but one
that is not working. Having an old package with perhaps old, but
working, features seems to be tolerable (and, as a matter of fact, it's
what happen to most stable packages).

Ciao, Giovanni.
--
Giovanni Mascellani <[hidden email]>
Pisa, Italy

Web: http://poisson.phc.unipi.it/~mascellani
Jabber: [hidden email] / [hidden email]


signature.asc (390 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: Bug#573538: Packaging again sagemath

Andreas Tille-5
In reply to this post by Rogério Brito
On Mon, Jul 19, 2010 at 01:26:34PM -0300, Rogério Brito wrote:
>
> Well, I think that a public repository for packaging it would be a very
> good start (Tim, do you have any?).

Why not using

   svn://svn.debian.org/svn/debian-science/packages/sagemath/  or
   git://git.debian.org/git/debian-science/packages/sagemath.git  ?

Once you are using this list for communication it sounds the natural
location for a packaging repository.

> Some of the big tasks that I would
> like to see addressed regarding sagemath would be:
>
> * getting a first draft package done, even if "improper for public
>   consumption".
> * relaxing the huge amount of dependencies (dropping them to recommends,
>   instead).
> * modularizing things as much as possible.
> * getting patches fed up to other upstream packages.

Sounds reasonable

       Andreas.

--
http://fam-tille.de


--
To UNSUBSCRIBE, email to [hidden email]
with a subject of "unsubscribe". Trouble? Contact [hidden email]
Archive: http://lists.debian.org/20100720075644.GD1097@...

Reply | Threaded
Open this post in threaded view
|

Re: Bug#573538: Packaging again sagemath

Johan Gronqvist
In reply to this post by Giovanni Mascellani-3
Hi List,

As a normal user who has read the emails, but not taken part in the
discussion due to lack of time for commitments, I would like to comment.

2010-07-19 18:54, Giovanni Mascellani skrev:
> Yes, that's my interpretation too. But I think that, in order to have
> good communication with upstream, we'll have to show that the package
> can be maintained and won't fall forgotten as last time.


As I read the reaction from upstream, they were frustrated that people
installed the package with the expectation that it should be current and
just work.

As Tim stated, I think the problem was that it entered into a repository
where people expect such things, and that this happened when it was
released as part of a stable ubuntu release. That is, the problem was
not primarily debian-related.

The response and proposals from upstream also seem to argue about ubuntu
rather than debian.

IMHO and AFAIUI:

For stable debian, people probably do not expect packages to be up-to-date.
For unstable debian, people probably do not expect packages to be bug-free.
For experimental debian, people probably do not install packages without
knowing what they are doing, and without expecting them to work without
problems.

It was also stated (by Tim?) that due to a process involving the
NEW-queue (I think), removing the package from debian and then
reintroducing it again could be much more work than moving to
experimental and working from there.


I would appreciate getting an up-to-date sage-package in debian again,
so I hope you find time and interest enough to continue the work.

Regards

Johan


--
To UNSUBSCRIBE, email to [hidden email]
with a subject of "unsubscribe". Trouble? Contact [hidden email]
Archive: http://lists.debian.org/i23mnv$jat$1@...

Reply | Threaded
Open this post in threaded view
|

Re: Bug#573538: Packaging again sagemath

Giovanni Mascellani-3
In reply to this post by Andreas Tille-5
Il 20/07/2010 09:56, Andreas Tille ha scritto:
> Why not using
>
>    svn://svn.debian.org/svn/debian-science/packages/sagemath/  or
>    git://git.debian.org/git/debian-science/packages/sagemath.git  ?
>
> Once you are using this list for communication it sounds the natural
> location for a packaging repository.

I just initialized a GIT repository with the old versions of sagemath
already loaded (using pristine-tar). I got the old version from
snapshot.debian.org.

Giovanni.
--
Giovanni Mascellani <[hidden email]>
Pisa, Italy

Web: http://poisson.phc.unipi.it/~mascellani
Jabber: [hidden email] / [hidden email]


signature.asc (390 bytes) Download Attachment