Delete old optional packages

[Volker Braun]

Since we once again had a thread about the pains of accidentally installing an old-style optional package, I propose to delete them except the following instead of opening a trac ticket for each one once something bad happened. If there is anything else you want to hit reply...

See the list here: http://files.sagemath.org/spkg/optional/

To keep:

- chomp

- java3d

- phc

- qhull

- kash3

[Jeroen Demeyer]

On 2016-04-20 09:24, Volker Braun wrote:
> I propose to delete them

I disagree. It's not because some packages are broken, that they should
all be removed.

[Dima Pasechnik]

IMHO everything that can be pip-installed can go (e.g. all  b*.spkg are pip-installable). 

But the rest ought to be inspected. E.g.

cunningham_tables-1.0.spkg 
database_jones_numfield-v4.spkg
database_kohel-20060803.spkg

 

ought to be kept.

extra_docs are old docs for external programs and probably can be removed (eg noone needs docs for python 2.5, I hope)

ginv and gnuplotpy are not pip-installable.

OK, so this is how far I went, alphabetically.

Should I go on and produce a complete list of things that can be removed?

(and things that must stay, and things that need to be inspected)

Dima

 

[Dima Pasechnik]

there is a discussion on http://trac.sagemath.org/ticket/20472

along these lines - PyX is an old-style optional package, apparently not

required by anything in Sage.

IMHO it should just be removed, as one can do 'sage --pip install PyX' just fine.

Jeroen instead wants to add yet another new-style optional package...

[mmarco]

In theory, each od these packages should have a mantainer. I think we should, at least, consult them before removing the packages. 

If a package has no mantainer at all... that is already a good reason to remove it from the list of optional packages.

[Dima Pasechnik]

Specifically, Jeroen says that `sage -i blah` is more user-friendly, as opposed to `sage --pip install blah`.

I argue that this is not true, e.g. as `sage -i` is not something that can be reverted, whereas `sage --pip` can.

I could also say that `sage -i blah` on packages which are behind the scene are installed by doing `sage --pip blah`

is an example of misleading the user into thinking that Sage treats `blah` in some special way, something that a commercial 

system would gladly do, but an open-source system should avoid as a plague. 

[William Stein]

Hi,

My one remark for this thread is that it is relatively easy to make
something pip installable. For example, somebody recently complained
that pygsl was only available via downloading files from sourceforce,
so I (very) easily made

https://pypi.python.org/pypi?:action=display&name=pygsl&version=2.1.1

by reading the bottom of:

https://python-packaging.readthedocs.org/en/latest/minimal.html

Now people can do "pip install pygsl".

William

> --
> You received this message because you are subscribed to the Google Groups
> "sage-devel" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to sage-devel+...@googlegroups.com.
> To post to this group, send email to sage-...@googlegroups.com.
> Visit this group at https://groups.google.com/group/sage-devel.
> For more options, visit https://groups.google.com/d/optout.



--
William (http://wstein.org)

[Jeroen Demeyer]

On 2016-04-20 16:26, William Stein wrote:
> Hi,
>
> My one remark for this thread is that it is relatively easy to make
> something pip installable.

For anybody reading William's post, better use this link:

https://python-packaging-user-guide.readthedocs.org/en/latest/

It contains similar information but it is more up-to-date, it describes
current packaging best practices, not those of 4 years ago.

[Volker Braun]

Congratulations, you are now in charge of contacting the maintainers ;-)

[mmarco]

Wow, It was easy to get that promotion!

[Volker Braun]

IMHO pip-type Sage packages should be limited to a) when we need a specific version or b) when it is required as a dependency.

Whatever the official user-facing installation command is (i.e. "sage -i" right now) should just fall back to pip/PyPI when there is no specific package in Sage.

Ideally it will behave like dnf/yum/apt/... and

1. Display what will be installed (including dependencies, free/nonfree license information, package type, ...)

2. Ask whether that was intended (y/N) unless overridden with a commandline switch "-y"

3. Only then do it

[Volker Braun]

On Wednesday, April 20, 2016 at 10:41:53 AM UTC+2, Dima Pasechnik wrote:

database_jones_numfield-v4.spkg

There is a new-style package for that one.

[Travis Scrimshaw]

 

See the list here: http://files.sagemath.org/spkg/optional/

IIRC, there is a more up to date list of packages, including which have been made into new-style, moved to experimental, ready-to-be-removed, and undecided. This would be more useful information IMO.

Best,
Travis

 

[John H Palmieri]

Well, there is the list at http://trac.sagemath.org/ticket/19220, but I don't think that's what you mean. It would be a starting point for packages which can be safely deleted from the server.

--
John
 

[Dima Pasechnik]

indeed: 8-)

build/pkgs/database_jones_numfield$ git blame SPKG.txt

e1900ef6 (Dima Pasechnik 2015-09-09 12:26:18 -0700  1) = database_jones_numfield =

see http://trac.sagemath.org/ticket/19174

whoever can delete the old spkg file, please do so!

 

[Travis Scrimshaw]

I was thinking of this:

https://wiki.sagemath.org/Classify%20old-style%20packages

Best,
Travis

[Dima Pasechnik]

unclassified...

classified...

secret... (password-protected ?)

top secret... (invisible ?)

[Simon King]

Hi Volker,

On 2016-04-20, Volker Braun <vbrau...@gmail.com> wrote:
> Since we once again had a thread about the pains of accidentally installing
> an old-style optional package, I propose to delete them except the
> following instead of opening a trac ticket for each one once something bad
> happened. If there is anything else you want to hit reply...

If I understand correctly, I have to point potential users to
the server in Jena if they want to download and install my group
cohomology spkg, and installing the downloaded old style spkg would
still work, right? I am in the process of turning it into a new style
package, at the same time rebasing it on top of a more recent version
of third party code that has thoroughly changed the API --- which sucks.

So, currently, the old style package is the only option.

Best regards,
Simon

[mmarco]

I have emailed the maintainers whose email address I could locate. someone knows how can I contact Michael Abshoff, Soroosh Yazdani
 or Mitesh Patel? 

[John Cremona]

On 21 April 2016 at 09:58, mmarco <mma...@unizar.es> wrote:
> I have emailed the maintainers whose email address I could locate. someone
> knows how can I contact Michael Abshoff, Soroosh Yazdani
> or Mitesh Patel?

Michael Abshoff was heavily involved in Sage in the early years and so
got put as a maintainer of a lot of spkgs when they were first
created. He no longer contributes to Sage so I think his name should
be removed from package maintainer lists, possibly except valgrind.

John

>
>
> El miércoles, 20 de abril de 2016, 19:04:09 (UTC+2), mmarco escribió:
>>
>> Wow, It was easy to get that promotion!
>>
>>
>> El miércoles, 20 de abril de 2016, 18:18:00 (UTC+2), Volker Braun
>> escribió:
>>>
>>> Congratulations, you are now in charge of contacting the maintainers ;-)
>>>
>>> On Wednesday, April 20, 2016 at 11:19:29 AM UTC+2, mmarco wrote:
>>>>
>>>> In theory, each od these packages should have a mantainer. I think we
>>>> should, at least, consult them before removing the packages.
>>>> If a package has no mantainer at all... that is already a good reason to
>>>> remove it from the list of optional packages.
>>>>

[mmarco]

Ok, so that is a good reason to delete those packages, unless somebody steps in as a maintainer.

[William Stein]

Similarly remove M Patel for similar reasons.  

--
Sent from my massive iPhone 6 plus.

[David Roe]

I've CCed Soroosh.  I'm not sure which package you were referring to.

David

[David Roe]

Actually, it looks like this address might be better.

David

[John H Palmieri]

It looks like this referred to mpi4py, which has already been converted to a new-style package.

  John

[mmarco]

Several of the maintainers answered me telling that it is OK to remove those packages. The rest of them either couldn't be contacted or did not answer at all. So it is safe to delete the packages.

[Thierry Dumont]

Mhhh, I have colleagues who use "lie".

An other package: coxeter. It was written by a colleague and friend of
mine, Fokko Ducloux, who passed away now 10 years ago. If I remember
well, people from the Sage-Combinat group wanted to work on it. What is
the situation ?

yours
t.

> --
> You received this message because you are subscribed to the Google
> Groups "sage-devel" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to sage-devel+...@googlegroups.com

> <mailto:sage-devel+...@googlegroups.com>.

> To post to this group, send email to sage-...@googlegroups.com

> <mailto:sage-...@googlegroups.com>.

[Travis Scrimshaw]

Mhhh, I have colleagues who use "lie".

It has been converted to a new-style skpg, but is experimental because it needs an overhaul on its build system. It is on my todo list, but unfortunately not very high. Also I need to learn how to work with autotools (not just use the final product). Any help here would be appreciated.

An other package: coxeter. It was written by a colleague and friend of
mine, Fokko Ducloux, who passed away now 10  years ago. If I remember
well, people from the Sage-Combinat group wanted to work on it. What is
the situation ?

That would be me in particular. I've created a github page for it: https://github.com/tscrim/coxeter. Again, I haven't had time to really work on it. I've started to write a test suite for it (most importantly, doing the timings) and cleaning up the code, but more work is needed. Help is welcomed.

Best,
Travis

 

[Dima Pasechnik]

Lie is a new style package. See http://files.sagemath.org/spkg/upstream/lie/index.html

What will be removed is just an old spkg file...

An other package: coxeter. It was written by a colleague and friend of
mine, Fokko Ducloux, who passed away now 10  years ago. If I remember
well, people from the Sage-Combinat group wanted to work on it. What is
the situation ?

yours
t.
> --
> You received this message because you are subscribed to the Google
> Groups "sage-devel" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to sage-devel+...@googlegroups.com

> <mailto:sage-devel+unsub...@googlegroups.com>.

[Alejandro Serrano Mena]

Dear all,

I was the uploader of the `sip` and `PyQt4` packages. This was long ago and right now I cannot take the effort of upgrading them. As far as I know, they haven't been used for any other thing that experimentation...